Conversation Analysis (Pragmatics & Beyond New)

Conversation Analysis

Pragmatics & Beyond New Series Editor Andreas H. Jucker University of Zurich, English Department Plattenstrasse 47, CH-8032 Zurich, Switzerland e-mail: [email protected]

Associate Editors Jacob L. Mey University of Southern Denmark

Herman Parret Belgian National Science Foundation, Universities of Louvain and Antwerp

Jef Verschueren Belgian National Science Foundation, University of Antwerp

Editorial Board Shoshana Blum-Kulka Hebrew University of Jerusalem

Catherine Kerbrat-Orecchioni University of Lyon 2

Jean Caron Université de Poitiers

Claudia de Lemos University of Campinas, Brazil

Robyn Carston University College London

Marina Sbisà University of Trieste

Bruce Fraser Boston University

Emanuel Schegloff University of California at Los Angeles

Thorstein Fretheim University of Trondheim

Deborah Schiffrin Georgetown University

John Heritage University of California at Los Angeles

Paul O. Takahara Kansai Gaidai University

Susan Herring University of Texas at Arlington

Sandra Thompson University of California at Santa Barbara

Masako K. Hiraga St.Paul’s (Rikkyo) University

Teun A. Van Dijk Pompeu Fabra, Barcelona

David Holdcroft University of Leeds

Richard J. Watts University of Berne

Sachiko Ide Japan Women’s University

Volume 125 Conversation Analysis: Studies from the first generation by Gene H. Lerner

Conversation Analysis Studies from the first generation

Gene H. Lerner University of California, Santa Barbara

John Benjamins Publishing Company Amsterdam
/
Philadelphia

8

TM

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.

Library of Congress Cataloging-in-Publication Data Conversation analysis : studies from the first generation / [edited by] Gene H. Lerner. p. cm. (Pragmatics & Beyond, New Series, issn 0922-842X ; v. 125) Includes bibliographical references and indexes. 1. Convesation analysis. I. Lerner, Gene H. II. Pragmatics & beyond; new ser. 125. P95.45. C6644 2004 302.3´46-dc22 isbn 90 272 5367 6 (Eur.) / 1 58811 538 0 (US) (Hb; alk. paper) isbn 90 272 5368 4 (Eur.) / 1 58811 539 9 (US) (Pb; alk. paper)

2004050204

© 2004 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa

Basically what I have to sell is the sorts of work I can do. I don’t have to sell its theoretical underpinnings, its hopes for the future, its methodological elegance, its theoretical scope, or anything else. I have to sell what I can do, and the interestingness of my findings. Harvey Sacks

Harvey Sacks 1935–1975

1967

c. 1974

Contents

Introductory remarks 1 Gene H. Lerner Glossary of transcript symbols with an introduction 13 Gail Jefferson I. Taking turns speaking An initial characterization of the organization of speaker turn-taking in conversation 35 Harvey Sacks A sketch of some orderly aspects of overlap in natural conversation 43 Gail Jefferson II. Implementing actions Answering the phone 63 Emanuel A. Schegloff Investigating reported absences 109 Anita Pomerantz “At first I thought” 131 Gail Jefferson III. Sequencing actions Pre-announcement sequences in conversation 171 Alene Kiku Terasaki Collaborative turn sequences 225 Gene H. Lerner The amplitude shift mechanism in conversational closing sequences 257 Jo Ann Goldberg Index 299

Introductory remarks Gene H. Lerner

Harvey Sacks once noted, “One commonly tends to avoid making ‘obvious’ observations because it is not obvious what thereafter is to be done with them” (Sacks 1987). In part, Sacks’ brilliance resided in his ability to detect the organization of action that underpins social life from the obvious, mundane details of conversation and other human conduct. Yet, beyond Sacks’ seminal contributions to the way we conceive of language, interaction and culture, one of his most important legacies can be found in the ways he worked up empirical materials – ways of working that other researchers could adopt and adapt. It is this clarity of purpose and method – embodied in Sacks’ investigations of conversation more than offered up as explicit research policies – that was necessary for a new discipline to emerge; without it even the most original work would fade. This volume collects early applications of the ground-breaking approach first pioneered by Sacks. His innovative ways of working will be found at the core of each chapter – not to mention the hundreds and hundreds of papers across perhaps a dozen disciplines that have been produced around the world since Sacks first began circulating the results of his investigations. From the mid-1960s up until his untimely death in November 1975, Harvey Sacks led a small group of researchers and personally trained a single generation of graduate students. What distinguishes this collection is that each of the contributors worked directly with Sacks as a collaborator or was trained by him at the University of California, or both. It was in dialogue with Schegloff, beginning while both were still graduate students, that many of Sacks’ most important insights about conversation as a possible site for organized social conduct first took shape. And it was Schegloff ’s move to work with, as Sacks once put it, “masses of data” that added a crucial dimension to the development of Conversation Analysis as a distinct discipline. In my view, without this move it would have been nearly impossible



Gene H. Lerner

for others to develop the kind of insight into human conduct that Sacks was so able to extract from single cases.1 Jefferson’s entry onto the scene was unique. Having enrolled in Sacks’ course at UCLA in 1965 to fulfill one remaining requirement for her B.A. degree in Dance, she stayed on after graduation continuing to audit his lectures, and began making transcripts for him of his lectures and of the tape-recorded conversations he worked with.2 In 1966, Sacks then found a way to pay her for her work at UCLA, and so Jefferson began her distinguished research career with the improbable title of ‘clerk/typist’. She continued making transcripts for Sacks and pursuing her own research interests at UC Irvine, following a move to the Irvine campus by Sacks and several of his students. (It was only in 1970 when UC Irvine could no longer keep her on as a clerk/typist that Sacks persuaded her to formally enter the graduate program there.) Through her sustained and concentrated association with the details of the recorded data – by coming to terms with those details and revealing some of the fine grained orderliness found in those details – Jefferson has provided Conversation Analysis with a wealth of resources for its development into the meticulously empirical discipline that it aims to be. The remaining contributors, Pomerantz, Goldberg, Terasaki and Lerner, all received training from Sacks as graduate students: Pomerantz and Goldberg first at UCLA and then at UC Irvine, Terasaki and Lerner at UC Irvine.3 A constant theme in that training was Sacks’ complete intellectual openness as to where the work and graduate students’ interest in it would go. Pomerantz was simply asked to “find an instance of a something” and she returned with a something (a compliment) that developed into her influential work on the preference/dispreference organization in the composition of second assessment turns. When Terasaki developed a strong interest in formal linguistics, Sacks encouraged that interest by suggesting to Terasaki that she act as the “formal linguistics person” in the group. (A trace of that interest can be seen in her contribution to this volume.) When Goldberg began exploring the use of an acoustic measuring instrument, that was incorporated into an investigation of action sequencing. Finally, an incident that occurred at the close of my very first encounter with Sacks could be considered emblematic of this openness.4 During that meeting I had described my data to him: video recordings of parent-child interaction. After arranging another meeting, and as we were walking out of the building, he made a simple suggestion: try transcribing a bit of the data. Yet, rather than tell me about Jefferson’s, by then, proven approach to transcription, he advised me to try to figure out a way to transcribe the data – and then

Introductory remarks

somewhat playfully suggested that I might be able to invent a whole new way to transcribe. The radical nature of the proposal was not lost on me – everything is possible, there are no pre-established limitations on how to proceed. There is nothing between you and the data. This was, of course, exactly what I had wanted to hear.

Overview In addition to these introductory remarks, this volume also includes a second introductory chapter by Jefferson that considers the value of carefully produced transcripts and presents the set of transcription conventions she originated. The eight contributions that make up the main body of the volume have been collected into three sections on turn taking, action formation and action sequencing. I first describe Jefferson’s contribution, and then introduce the chapters in each of the three sections of the volume. Taken together this collection offers a sampling of Conversation Analytic inquiry from its early years, while nevertheless presenting research of contemporary significance.5 The transcription system developed by Jefferson for Conversation Analytic research is the internationally recognized “gold standard” for transcribing the interactionally relevant features of talk-in-interaction. In her first of three contributions to this volume (specially produced for this collection) Jefferson expresses her views on transcribing and transcripts and then produces an updated compendium of transcript symbols tailored for computer-based use. However, it is not quite accurate to simply say “Jefferson expresses her views on transcribing. . .” Rather she delivers a powerful empirically-based lesson on the importance of putting “all that stuff ” in our transcripts by comparing the analytic results of working up and then working with a painstakingly detailed transcript with transcripts that do not capture many of the productional features of the talk. She shows the reader the value of transcripts that pay attention to the positioning and design features of talk-in-interaction not by means of theoretical pronouncements but as a practical matter in that such transcripts can have specific analytic payoffs. Particularly instructive is her treatment of incipient laugher. Though capturing the productional particulars of laugh tokens (is it “heh” or “hih” or “hn”?) or the exact placement of a plosive breath within a word takes careful and repeated listening, Jefferson shows just how crucial this can be to an understanding of what is going on, and how ‘what is going on’ is getting done.





Gene H. Lerner

The chapters in the main body of this volume advance our understanding of three forms of organization generic to conversation: turn taking, action formation and action sequencing. I now take up each of these forms of organization in turn.

Taking turns speaking Turn taking practices organize the allocation of opportunities to participate in conversation and the turn-constructional forms such participation take. Understanding turn taking for conversation and other forms of talk-in-interaction is key to understanding human conduct, because most actions carried out through talking are shaped by the organization of that talk into speaking turns: it shapes how speakers compose their contributions, it shapes where they position those contribution in the ongoing interaction, and it shapes when they get to participate. Part One features two early attempts to come to terms with turn taking phenomena for conversation. It is a distinct honor to be able to publish Harvey Sacks’ original manuscript of what later became “A Simplest Systematics for the Organization of Turn Taking for Conversation” (Sacks, Schegloff, & Jefferson 1978).6 The 1974 version of “A Simplest Systematics ...” is without a doubt the most cited article in the field of Conversation Analysis (and is by far the most-cited article Language has ever published (Joseph 2003)).7 I believe it would be fair to say that this paper has contributed significantly to establishing Conversation Analysis as a distinct discipline. In this early attempt, Sacks first compares turn taking for conversation to turn taking for other forms of talk-in-interaction, and then lays out a basic systematics for turn taking for conversation. The present chapter not only gives us a glimpse of Sacks’ original thinking on this topic but also is an extraordinarily clear explication of many of the essential elements of turn-taking organization. (Though, interestingly, here he focuses almost exclusively on turn allocation, with only the briefest mention of turn construction.) With its prominent focus on comparative speech exchange systems, it has the effect of affording readers a new perspective on a classic contribution to Conversation Analysis. In the second chapter of this section, Jefferson addresses moments of problematic participation, when more than one party is speaking at once. Her contribution on overlapping speech in conversation derives from and complements Sacks’ work on turn taking.8 It is only once we understand the organization of turn taking, that simultaneous speech or overlap can be properly understood. Jefferson lays out the systematic forms of overlap competition

Introductory remarks

and introduces the important conceptual distinction between marked and unmarked overlap competition. Almost half of the paper is devoted to a systematic explication of methods for “post-overlap retrieval” of overlapped talk. Here she lays out in detail the consequentiality of overlap for the trajectory of subsequent talk.

Implementing actions Practices implement actions – that is, the practical actions speakers accomplish through talking in interaction are formed up by speaking in particular ways in particular places in an emerging conversation. In this section, the authors spotlight the ways speakers do such things as answer the phone, determine if a student absent from school was truant, and report what they were thinking when something unusual happened. Schegloff ’s contribution, “Answering the Phone,” is a companion piece to his classic paper on Summons-Answer sequences (Schegloff 1968). This chapter deepens our understanding of sequence organization, since it describes, in detail, members’ interpretive work – sequential interpretive work – that underpins the resulting routine conduct. Answering the phone may seem like an inconsequential matter if one only considers the small range of ways phone calls are answered, but its very simplicity allows Schegloff to consider crucial elements in members’ interpretations of their situated social-interactional circumstances that have broad application for studying talk-in-interaction. Rather than stipulating the circumstances and identities of answerers, he warrants their inclusion by showing how identities become relevant and circumstances become consequential for action. Here Schegloff leverages the simple act of answering the phone into a careful description of the range of social and interactional matters that members’ must take into account in determining how to respond. “Investigating Reported Absences” is an important contribution to our understanding of “talk at work” – or perhaps more precisely “talk as work.” Here, Pomerantz describes the methods used by an institutional agent (a school attendance office clerk) for carrying out the institutionally mandated investigation of students’ reported absence from school, to determine if an absence is “excused” or “unexcused.” She reveals the delicate ways such investigations proceed, especially in cases where unexcused absence might be suspected, so as to maintain a stance of neutrality – that is, she shows how clerks engage in investigatory activities in ways that do not reveal a presumption of guilt. Moreover, Pomerantz shows that what participants understand to be at stake bears





Gene H. Lerner

on the practices they employ. She is able to ground participants’ taken-forgranted knowledge through a careful analysis of the methods the participants use to carry out tasks related to their institutionally relevant identities. Here, we also see how official records are employed as part of routine procedures designed to enforce institutional control. Jefferson’s chapter in this section, “At First I Thought” explores a conversational device that demonstrates how reported thoughts can be subject to social organization – in this instance sustaining the social world as an altogether ordinary place by showing one’s “commitment to the normal.” She first traces Sacks’ developing interest in this matter – based for the most part on reports found in newspaper stories. Then, using both newspaper reports and conversational data, Jefferson shows how people routinely select “first thoughts” to report, that are appropriate to local circumstances and/or category membership. Or at least they seem so, when compared to the extraordinary events that turn out to have been the case, on second thought – “At first I thought X, then I realized Y.” She demonstrates how this action is formulated through the analysis of both cases in which the canonical format is present, and ones in which it is absent – but has nonetheless left its fingerprints. Most striking are those cases that reveal a commitment to an extreme new reality as now normal by reporting such first thoughts as a terrorist attack for something that turns out to have been a fire at a tire dump.

Sequencing actions Activities carried out in conversation are organized into sequences of actions. The organization of activities into sequences of actions shapes participation. Each course of action structures opportunities to participate within it. And as participants make relevant various subsequent opportunities to participate through their actions, they thereby organize their activities moment by moment. These sequences of actions turn out to embody recurrent patterns of actions with their own organizational features. In this section, each chapter examines an aspect of the ordering of actions in conversation, and each was begun as part of dissertation research directed by Sacks. Terasaki examines the operation of one type of preliminary (or “pre”) sequence – a small sequence of actions designed to come before another, main action sequence. Lerner describes one form of retrospectively initiated (or “retro”) sequence – the collaborative turn sequence. And Goldberg shows how the loudness or amplitude of an utterance is employed to mark the type of place in a sequence that utterance is designed to occupy.

Introductory remarks

For almost 30 years, Terasaki’s paper has remained the standard reference on pre-announcement sequence organization.9 In this chapter, Terasaki shifts the linguistically motivated interest in the “given/new” distinction for information at the level of single utterances, to participants’ orientation to news at the level of sequences of action. First, she demonstrates that position can be crucial for “new information” to be responded to as news. The import of this is that one cannot rely solely on content to determine if something is treatable as news. Then, drawing on the general organizing principle of recipient design (that speakers tailor their talk to fit their recipients in many ways), Terasaki shows how news delivery sequences are shaped, in part, by an orientation to determining whether possible news is news for a deliverer’s current recipient. She provides a detailed account of the linguistic composition and sequential organization of pre-announcement sequences and their connection to the composition of subsequent news delivery turns. Lerner’s contribution demonstrates another connection between syntax and social action – in this case between turn construction and action sequencing. He examines the sequences of action that can be launched when one speaker completes the in-progress turn-constructional unit (TCU) of another speaker. Special attention is given to the range of sequence-responding actions that are employed once a speaker completes the TCU-in-progress of another speaker, with a special focus on the practices that can stand as alternatives to responding directly to such completions. Lerner shows that one outcome of anticipatory completion of another participant’s turn can be the ratification of that completion, thus resulting in a collaboratively produced TCU, but that there are also routinely used methods for disregarding the proposed completion. Finally, empirical materials seem to indicate that, although acceptance and rejection of an anticipatory completion are response alternatives, rejection rarely happens. This is so because it is always possible to disregard a proffered completion. Thus, outright rejections seem to be employed, for the most part, when the proffered completion is produced as, or treated as, a non-serious rendition of the projected completion. In the final contribution to this section, Goldberg shows us one way that prosody can be used to position turns at talk within their sequential environment. This work extends her previously published research (Goldberg 1978) on amplitude shift (the change in loudness from one utterance by a speaker to the next utterance by the same speaker) in sequences of turns at talk, by showing how amplitude shifts furnish a vocal resource for the organization of closing sequences in telephone calls. Goldberg examines how amplitude shift in conversational closings can be used as a device to sustain the engagement





Gene H. Lerner

with prior talk or demonstrate disengagement from it. She contributes to our understanding of sequence organization in general by describing how vocal resources are employed in sequence initiation and sequence suspension, and she contributes to our understanding of closing sequences in particular by describing how amplitude shift figures into their coherence as a sequence unit, and how it figures into procedures for moving out of closings. I conclude these introductory remarks on a somewhat personal note, with a short sketch of my first meeting with Harvey Sacks. I have already referred to this meeting briefly in the opening section.

On meeting Harvey Sacks – A concluding remark I first met Harvey Sacks in the course of investigating doctoral programs. I was on the verge of completing a Master’s degree in psychology, when I began to look for a doctoral program that would allow me to extend my thesis research on parent-child interaction. A fellow graduate student, who had gone on to UC Irvine’s School of Social Sciences the previous year, suggested that I might look into the graduate program there. I found his description of the openness and flexibility of a School without departments or pre-established requirements enormously appealing. On my first trip to the Irvine campus to see if I could find members of the faculty whose work matched my interest in parent-child interaction (and a more general interest in methods for measuring and modeling small group interaction), I talked first to a faculty member involved in mathematical modeling. However, as soon as I mentioned that my data consisted of videotapes of interaction, I was immediately directed to Harvey Sacks as the one person I should speak to about my work, and about graduate school. I tracked Sacks down in his “Lab” which consisted of a windowless room with a single bed (without bedclothes), a few reel-to-reel tape recorders on a table, a typewriter on a small typing table in the center and the walls decorated with large pieces of what looked like butcher paper, each covered with a few lines of hand printed text. I discussed my interests with him – and in particular my interest in and dissatisfaction with currently available methods for measuring small group interaction. Sacks really did not talk much about his work.10 However, the most important result of that meeting (other than the business about transcribing) was that I left with a copy of the recently published article, “Opening Up Closings” (Schegloff & Sacks 1973), and an agreement that we would talk again. It would

Introductory remarks

be an understatement to say that I was not prepared for what I encountered in that paper. What I found there was an entirely new way of thinking and an entirely new way of working with data. Even this many years later, it is completely clear to me what I found so powerful. There were two things I took away from my first reading of that first paper: the presentation of a formal structural argument, but without any sign of quantification; and even more important to me at the time, here was an empirically grounded proof of a problem – not the solution to a problem, but of the existence of an interactional problem, “the closing problem.” On my next visit, I believe I brought a couple of reels of my taped data that we looked at, and I brought something that actually turned out to be more important to Sacks’ decision to admit me into the program. What happened was that I brought out a transcript – admittedly short and rough by today’s standards, but carefully done – of the bit of tape I wanted to show him. (I should say that I had not followed Sacks’ advice by inventing my own method of transcribing, but followed Jefferson’s guidelines once I encountered then in “Opening Up Closings”.) Actually, I think he would have admitted me because I was interested in his work and his way of working, but his whole demeanor changed at this point. He seemed to relax, and it was my impression that he took this as an important sign of my seriousness – of my willingness to work. I abandoned my nearly completed thesis and entered the doctoral program at UC Irvine in the Fall of 1974 to study with Sacks. The time I spend studying with him was regrettably way too short. His death was a sharp loss, still felt by all who worked with him. But when I reflect upon it, I can see that I have been, in a sense, studying with him ever since.

Notes . In fact, in an addendum to the same correspondence in which the “masses of data” comment was made, Sacks makes the unqualified statement, “The structures for particulars direction doesn’t work.” (This is taken from a handwritten letter from Sacks to Schegloff, dated March 1974, and circulated in 1988.) . At this point Jefferson also began to do some unofficial teaching, working with those of Sacks’ undergraduate students who were not doing well in his course. . Pomerantz, like Jefferson, completed her dissertation with Sacks at UC Irvine. After Sacks’ death Goldberg and Lerner completed theirs with Schegloff, while Terasaki did not. . I will expand the description of that meeting a bit at the end of these introductory remarks.





Gene H. Lerner

. Although some editing was done to prepare the manuscripts for publication, there was no attempt to change the style of the text. These are truly “first generation” papers in terms of style, content and aim. . I would like to thank Manny Schegloff, as Sacks’ literary executor, for bringing the original draft of the turn taking paper to my attention and offering it for publication. . Although the 1974 article published in Language is probably the better known version of the turn taking paper, it is the so-called “variant version” later published in the Schenkein (1978) collection that stands as the definitive text. In comparison, some editorially induced changes were made for publication in Language. . I believe it was the import of Jefferson’s work on overlap, as direct evidence for party’s orientation to turn taking, that is probably the main reason Sacks listed her as a co-author of the initial draft of the turn taking paper published in this volume – although she was not directly involved in its drafting. Some of the research on overlap Jefferson had been conducting at the time became part of her 1972 dissertation and was published the next year as “A Case of Precision Timing in Ordinary Conversation: Overlapped Tag-Positioned Address Terms in Closing Sequences” (Jefferson 1973). . This chapter is a slightly revised version of a paper originally circulated as Social Science Working Paper #99 at the University of California, Irvine in 1976 where it was awarded the Alice Macy Prize. . Sacks did ask if I was familiar with Goffman’s work, but Garfinkel did not come up at all in that meeting – and he actually only mentioned Garfinkel to me once while we were discussing my interest in children. He mentioned that he had been loosely connected to a research proposal of Garfinkel’s concerning children, but then stated that his relationship to that stuff was “purely historical.”

References Goldberg, Jo Ann (1978). “Amplitude shift: a mechanism for the affiliation of utterances in conversational interaction”. In J. N. Schenkein (Ed.), Studies in the organization of conversational interaction (pp. 199–218). New York: Academic Press. Jefferson, Gail (1973). “A case of precision timing in ordinary conversation: overlapped tagpositioned address terms in closing sequences”. Semiotica, 9, 47–96. Joseph, Brian D. (2003) “The Editor’s Department: Reviewing Our Contents”. Language, 79 (3), 461–463. Sacks, Harvey (1987). “On the preferences for agreement and contiguity in sequences in Conversation”. In Button, G., J. R. E. Lee (Eds.), Talk and social organization (pp. 54–69). Clevedon: Multilingual Matters. Sacks, Harvey, Schegloff, Emanuel A., & Jefferson, Gail (1974). “A simplest systematics for the organization of turn-taking for conversation”. Language, 50 (4), 696–735. Sacks, Harvey, Schegloff, Emanuel A., & Jefferson, Gail (1978). “A simplest systematics for the organization of turn taking for conversation”. In J. Schenkein (Ed.), Studies in the Organization of Conversational Interaction (pp. 7–57). New York: Academic Press.

Introductory remarks

Schegloff, Emanuel A. (1968). “Sequencing in conversational openings”. American Anthropologist, 70, 1075–1079. Schegloff, Emanuel A., & Sacks, Harvey (1973). “Opening up closings”. Semiotica, 7, 289– 327. Schenkein, Jim (1978). Studies in the Organization of Conversational Interaction. New York: Academic Press.



Glossary of transcript symbols with an introduction Gail Jefferson

Although I’d probably rather transcribe than any do any other part of the work (analyzing, theorizing, lecturing, teaching, etc.), the one thing I’d rather not do is talk about transcribing. It’s not a topic. You might as well talk about typewriting. Transcribing is just something one does to prepare materials for analysis, theorizing, etc. Do the best you can, but what is there to talk about? On the other hand, there might be something to talk about if we compare a 1964 transcript by Harvey Sacks, (1) [Sacks GTS trans:1964] A. I started work at a buck thirty an hour and he said if I work a month you geta buck thirty five an hour and every month there be a raiseT. Howd you get the job? A. I just went down there and asked him for it. T. Last week you were mentioning something about the fact that youA. I got lost in one job, yeah. T. Got lost in it, and your fatherC. He know your father? Yeah. [ A. Sure he knows my father, but my father’s got nothing to do with it. [ C. Ok he gave you the job. A. No, he’s got nothing to do with it. Huh uh, My father’s not buying beer anymore

with a transcript that I did in 1985.



Gail Jefferson

(2) [GTS:I:2:3:R:1-5:3-4] Ken: I started workin etta buck thirty en hour (0.4) Ken: en’e sid that if I work fer a month: yih getta buck,h ·h thi[rty ↓fi:ve= (Dan): [((sniff)) Ken: =’n hour en (·) ev’ry month he uh ( ) he rai[ses you ]◦ ( )◦ ] Dan: [How’dju]g e t th]e jo:b, (1.0) Ken: ↑I js wen’ down there’n ↓a:st eem for it (1.8) ◦ ◦ Dan: Cz la:st week you were mentioning something about th’ fa:ct ◦ thet you ↓u[h◦ Ken: [I got ul u. -got (·) lost in one jub=↓Yea:h. (0.5) (i t bo:th[ered) ] Dan: Got lo:st innit ↓e[r (y’r fa:th[e r ) ]◦ ( )◦ Ken: [ w h h h h [h h h h]h h Al: [Dz ’ee ] know yer father? (0.2) Al: ↓Yah. (0.6) Ken: Sure ’ee knows my father [ b’t my f a t h e r’s g’t] nothina ] do with it. Al: [( ) they gave you] th’ jo:b,] (0.7) Ken: No: he’s got nothin d’do w’th it. Huh-uh, my fa(h)ather’s not buyin beer innymo:re

Fragment (1) is concise, and readable; fragment (2) is a nightmare. Or, for example, we could look at a recently produced, succinct and readable transcript from Stanley I. Kutler’s Abuse of Power (1997:253),1 (3) [AbPow:253] ((re the “seamy things” Hunt says he did for the White House Plumbers)) DEAN: I don’t know the full extent of it, PRESIDENT NIXON: I don’t know about anything else. → DEAN: I don’t know either, and I [laughs] almost hate to learn some of these things.

and ask why I feel compelled to produce something like this:2

Glossary of transcript symbols

(4) [Jeff:Canc:40:10-20] Dean: I ↑don’t kno:w th˙e (·) full extent ↓’v it.↓ (0.7) ◦ Dean: ↓Uh:::eh◦ (0.9) Nixon: ◦ I don’noo◦ ’bout anything else exchh e[pt Dean: → [I don’t either in I: ◦ w’d (h)als(h)o → hhate tuh learn [some a’] these thi]ngs. ·hh·hh·hh·hh Nixon: [W e l l ] y a : h ] (0.2) Dean: So ↑That’s,hhhh that’s that situation.

Why put all that stuff in? Well, as they say, because it’s there. Of course there’s a whole lot of stuff “there,” i.e., in the tapes, and it doesn’t all show up in my transcripts; so it’s because it’s there, plus I think it’s interesting. Things like overlap, laughter, and ‘pronunciational particulars’, (what others call ‘comic book’ and/or stereotyped renderings), for example. My transcripts pay a lot of attention to those sorts of features. What good are they? I suppose that could be argued in principle, but it seems to me that one cannot know what one will find until one finds it, so what I’ll do is show some places where attention to such features turned out to be fruitful. First of all, we can compare bits of fragments (1) and (2), and then do the same with fragments (3) and (4). Starting off with fragments (1) and (2), I’ll focus on the most irritating segment of the latter, and see what might be gotten from it. Recalling three lines from fragment (1): A: I got lost in one job, yeah. T: Got lost in it, and your fatherC. He know your father? Yeah.

and the jumble from fragment (2): Ken: I got u.l-got (·) lost in one jub=↓Yea:h. (0.5) (i t bo:th[ered) ] Dan: Got lo:st innit ↓e[r (y’r fa:th[e r ) ]◦ ( )◦ Ken: [ w h h h h [h h h h]h h Al: [Dz ’ee ] know yer father? (0.2) Al: ↓Yah.





Gail Jefferson

What is so clearly to be made of this bit in fragment (1) is that the therapist mentions patient A’s father, and the mention of A’s father triggers a question by patient C: T: ...and your fatherC. He know your father?

In fragment (2) the issue is drastically obscured, first of all by the sheer difficulty of reading through all that stuff, now including some sort of whoosh of breath by Ken (A), but mainly by the addition of an alternative hearing for “your father”, i.e., “it bothered”. Surely that could be resolved by looking at the context: We have the therapist, (Dan/T), mentioning the word “father”, which prompts one patient (Al/C) to ask the other (Ken/A) a question about that father, starting up before the therapist finishes whatever he was going to say. The alternative is rather improbable, i.e., that Dan has said something about Ken’s having been “bothered”, and it just so happens that at that moment Al decides to cut in with a tangential question about Ken’s father. Logic plumps for “father”; so much so that I only kept the alternative in as a matter of principle. I wished that I could unhear it. But I do hear it; it would be irresponsible to leave it out even though it not only messes up the transcript but it’s so improbable that I’m embarrassed by it. But on thinking about it, yes, it’s more than likely that the “father” alternative is what’s happening with Al, but that doesn’t tell us what Dan said. In our materials we sometimes can clearly hear one speaker saying something, and find that a coparticipant somehow has heard something else. Once in a while, one or the other mentions the mishearing. For example: (5) [Frankel:GS:X] ((re a plant Alan is trying to sell to Nell)) Alan: → Still growing. It’s got buds ’n everything else on it. Nell: Oh has i:t? Alan: → Buds. No[t bugs. ] Nell: → [Oh bu:d]s. I thought you said bugs. Alan: No. I don’t see any bugs. It might have, but I can’t see any

It’s a phenomenon. And it very likely happens more often than we have access to in our materials, because people don’t always, or even routinely, mention the fact that they did a mishearing (sometimes they don’t realize it, sometimes they do realize it but it doesn’t seem to matter, etc.). What the “bothered” alternative does is to raise the possibility that such a thing has happened here: Dan is saying something about Ken having been

Glossary of transcript symbols

“bothered”, which Al hears as his saying something about Ken’s “father”, which inspires the question he then and there asks (‘then and there’ in fine detail, i.e., starting up after the first syllable of what he hears as the trigger-word, “fa:th[er”).3 I’ll just mention one other difference between the two fragments. One of them shows an ‘and’ while the other shows an ‘or’ (pronounced “er”). A. T. →

I got lost in one job, yeah. Got lost in it, and your father– versus

Ken:

Dan: → Ken:

I got ul u.-got (·) lost in one jub=↓Yea:h. (0.5) (it bo:thered) Got lo:st innit ↓e[r ( y’r fa:ther ) ◦ ( )◦ [w h h h h h h h h h

So what, “and” or “or”. Well, “and” is accepting a proffered version of what happened, and augmenting it: ‘You got lost in it and it bothered...’ / ‘You got lost in it and your father...”. If it’s “or”, then while repeating the proffered version of what happened may acknowledge it as reasonable, the “or” projects an alternative version; not that you “got lost in it”, but that “...it both[ered ( )...” / but that “...your fath[er ( )...”. So, the difference between “and” and “or” is a difference between two polar activities: One transcript shows the therapist accepting, the other shows him rejecting, the patient’s version of what had happened. Turning to fragments (3) and (4), focusing on the occurrence of laughter. In fragment (3), we are shown some talk in progress, in the course of which Dean “[laughs]”: PRESIDENT NIXON: I don’t know about anything else. DEAN: I don’t know either, and I [laughs] almost hate to learn some of these things.

In fragment (4) we are shown some details of those two utterances: Nixon: ◦ I don’noo◦ ’bout anything else exchh e[pt Dean: [I don’t either in I: ◦ w’d (h)als(h)o hhate tuh learn some a’ these things.

For one, fragment (4) suggests that Nixon is going on to mention something else he knows about, i.e., we now have the word ‘except’. And we have Dean starting up within that word, at “exce...”, where after, Nixon stops. One thing that might be happening here is that Dean hears, in “exce...”, the word ‘except’





Gail Jefferson

forming up, and starts to talk at that point. This ‘recognitional- response’ is a not-uncommon phenomenon. Here are a couple of cases.4 (6) [Rah:II:11:R] ◦ Jessie: two pihleece cah:rs’d◦ stopped outsi:de.= Ada: =eeYe::[s? Jessie: → [·h An’ that whether he thow:t thet I ed’n ac[cidn’t] [ohr someth]ing... Ada: → [I : : : :]:[ k n e o : w ]

In fragment (6) above, just as Jessie starts to say the dreaded word “ac[cident”, Ada starts her comforting “I:::: kno:w”. And in fragment (7) below, just as Mr. Bryant is pointing out that with the credit note his firm is offering Miss Sokol, their services won’t “co[st” her anything, she starts up with her acceptance of his offer: (7) [SCC:DCD:37:R] Bryant: → ...but ·hh if you’ve gohtta credit neote ih weon’t co[st you anything anyw]ay. Sokol: → [ Wul owright the:n, ] Sokol: That’s faiuh.

Similarly, with “exce...”, Dean may hear the word ‘except’ forming up. Hearing that, he may hear that Nixon is starting to mention something else, knowledge of which Dean doesn’t want to be burdened with. It may be that Dean moves then and there to stop any possible revelations by cutting in on the alerting word ‘except’, prior to its completion, with an agreement that specifically ignores its projection of further things (i.e., his “I don’t either” targets Nixon’s initial proposal, “I don’t know about anything else”), and then goes on to announce his unwillingness to know any more. Then there is the laughter. It appears that there is not merely laughter in the utterance with which Dean may be declining to hear what Nixon started to say, but that Nixon, in projecting such talk, himself produces something that might be taken for laughter. That is, fragment (4) shows that he produces the word “except” with some breathiness and an unvocalized mid-vowel: “exchh ept”. Again, here is a possible case of a recurrent phenomenon: A recipient of someone’s utterance may treat something in that utterance as laughter, and respond with laughter of his own. Here’s a simple case:5

Glossary of transcript symbols

(8) [Schenkein:II] Bill: Dju watch by any chance Miss International Showcase las’ night? Ellen: n:No I didn’[I w’z reading myBill: → [You missed a really great pro(h)[gram. Ellen: → [O(hh)h i(h)t wa(hh)s?= Ellen: =ehh heh heh heh!

Here, just after a plosive breath occurs in the course of Bill’s utterance, Ellen produces a laughter-loaded response. In fragment (4), Dean may be doing something similar, but less directly – and less transparently – reciprocal. That is, he does not immediately start to laugh, but first produces a bit of talk. He’s not, then, to be heard as slavishly laughing just because the President is laughing. As well, the work he’s doing with that bit of talk may be enhanced by his not only ignoring the word “except”, but declining to reciprocate the laughter with which the problematic word is being produced. Furthermore, his delayed laughter, while perhaps taking Nixon’s prior laughter as a warrant (i.e., it is all right for him to laugh since the President has just laughed), can, by its delay, be heard for the work it may independently be doing, e.g., as tempering his declination to hear what the President was perhaps about to tell him.6 Yet another point can be made. It is possible that the transcriber of fragment (3), by not attending the details of the laughter, but just noting that it occurred, misheard the talk in which that laughter occurs, i.e., fragment (3) shows Dean saying “I [laughs] almost...”, while fragment (4) shows him saying “I: w’d (h)als(h)o...” (in standard orthography, “I would also”). In an early paper on laughter (Jefferson 1985), I proposed that laughter can make an utterance less (or un-) intelligible, and may possibly be used for that feature. In the case at hand, Dean’s laughter, while not being deployed for that purpose, may have made the utterance that I hear as “I would also...” less intelligible, resulting in its being heard by another transcriber as “I almost...”. Here’s another sort of thing that can be noticed in a comparison of fragments (3) and (4). In the former, Dean, shown of course in standard orthography, is saying “...I [laughs] almost hate to learn some of these things.” In the latter, among other pronunciational particulars, he’s shown as saying “tuh learn”. In an early paper on this issue (Jefferson 1983), I point out that “the sort of ‘comic book’ orthography I use (e.g., for ‘What are you doing?’, ‘Wutche doin’?’) is considered objectionable in that it makes the speakers look ‘stupid’;





Gail Jefferson

it seems to caricature them rather than illuminate features of their talk”, and that “experts on phonetics such as William Labov, propose that someone who, for example, says ‘dat’ instead of ‘that’, is not producing defective English but is speaking correctly in his dialect, and thus should not be transcript-displayed as producing an object which is commonly treated as defective.” In that paper, as part of my defense of ‘pronunciational particulars’ I show several fragments in which speakers of one or another ‘dialect’ can be found to be varying their pronunciations; for example, a member of the California motorcycle gang, the Hell’s Angels, produces “them” followed by “dere” (‘their’): (9) [KPFK:GJ] Joe: I tell them right t’dere face . . .

For example, a Bronx janitor produces several versions of ‘there’ (and ‘they’re’): “theyuh”, “they’re”,“deyuh”, and “dere”, while consistently using “th” for (two different versions of) ‘the’: (10) [Frankel:USI:117:R] Vic: We get in they:uh (0.5) en they’re uh (0.2) the tu- (·) u. (·) t-two ↓guys uh deyuh, ’n me ’n James ↑Wal↓kuh’s dere ’n th’ broa:d is in th’ bed.

For example, A Philadelphia-Italian meatcutter uses both “dis” and “this”: (11) [Goodwin:M:3-4] Frank: → Fu(h)ck you. Dis g[uy is– Joe: [Hey wait. . . Frank: → I definitely ain’t goi(h)in’ ou(h)t with this chick again.

And, for example, in a fragment of a transcript I made of a tape collected by Labov, we find both Mez, a member of a black teenage street gang, and B.J., a black social worker/ ex-street gang member, producing “d” and “th”; Mez saying “Hey lookih dat”, “Who’s that. Who’s that punk right there.”, (and also producing a vowel-begun version of ‘that’s’, “Ahz Davey More.”), B.J., saying things like “y’all see these pictichiz...”, “y’know who dat is”, and “That’s Davey More”.

Glossary of transcript symbols

(12) [Labov:Jets:3-5] Mez:→ Hey lookih dat. . B.J.: → Jo man, we gonna talk about fights man y’all see these pictchiz → up here man y’know who dat is, Mez:→ Who’s that. Who that punk right there. Mayall: Mothuhfuck[im!! B.J.: → [A punk, [That’s– Mayall: [He stinks! Alex: [Shuddup Mayall!! Leonard: [Man- eh heh! [khhkhkhkhkhkhkh[khkhkh B.J.: [Davey More. [ Mez:→ [Ahz Davey More B.J.: → Th-that’s Davey More en that’s- Benny Kid Paret.

I then go on to locate some possible systematicities and interactional phenomena that can be found by looking at pronunciational details, concluding that by omitting such particulars from our transcripts, we are obliterating a potentially fruitful data base. Now, that paper specifically focused on talk by people who stereotypically mispronounce (or, as Labov has it, correctly-for-their-dialect pronounce) various words, who can be found to do correct (or incorrect-for-their-dialect) pronunciations, as well. But if we look at the talk of John Dean, surely an epitome of WASP, middle class, etc. etc., we also find variation. So, for example, in a fragment from Abuse of Power, Dean uses the word “to”, three times within a short spate of talk: (13) [AbPow:247-248]

⇓ ⇓ DEAN: It started with an instruction to me from Bob Haldeman to see if we couldn’t set up a perfectly legitimate campaign intelligence operation over at the Re-Election Committee. PRESIDENT NIXON: Hm-hm. ⇓ DEAN: Not being in this business, I turned to somebody who had been in this business, Jack Caulfield, who is, I don’t know if you remember Jack or not. He was... My transcript shows that on each of the three occasions of its occurrence, Dean pronounces the word ‘to’ differently: “to”, “tuh”, and “dih”:





Gail Jefferson

(14) [Jeff:Canc:7:27-8:9] Dean: → ↑Started with (1.0) en instruction to me:: (0.9) from Bob ↓Haldeman. → (0.4) tuh see if we couldn’t set up ay perfectly le↓git’mate (0.3) campaign intelligence operation over et the ReElection Committee. (·) Nixon: Mmhm, (0.8) Dean: → Not being in this business?=I turned dih somebuddy who: ·h had been in this business: (·) Jack Cau’field who: w’z I don’ know if you r’member Ja:ck er not he w’z...

We can at least note that the ‘correct’ pronunciation, “to”, coincides with a stress on that word, while the two ‘incorrect’ pronunciations, “tuh” and “dih” occur at points where the word is not being stressed. Also, we can at least account for the occurrence of the ‘d’-begun “dih”, as conditioned by the preceding, ‘d’-ended word, “turned”. Who knows what other orderlinesses will emerge as attention is given to such details? Harvey Sacks often spoke of the possibility of “order at all points”, and in one of his lectures (Sacks 1966 [1992: I: 238]) tells us that “we’re dealing with something real and powerful. And not just grossly powerful, like, it provides for the rate of industrial development, but it provides for little tiny things that God might have overlooked, perhaps.” Most of the things Sacks dealt with, while not as gross as the rate of industrial development, were on a larger scale than tracking the variations in a Bronx janitor’s pronunciations of ‘there’, or John Dean’s pronunciations of ‘to’. But now and then, something in one of our more detailed transcripts would catch his attention, and we’d be treated to “a little something I find enormously amusing”, having to do with someone’s spelling out their name in a traffic court, producing it as “M-a-u-e-r, (pause) h-a-n.”, the point being that “pause distributions are kind of important...in this sense at least: You can do them wrong.”, which led to a consideration of the “normative” character of pausings (Sacks 1968 [1992: I: 784). Or, for example, we’d be told of “a fascination” he had with “[t]hings like ‘didje’ and ‘wanche”’, and his having “noticed several differences between them”, which led to an exploration of the phenomenon of “transformation” in actual talk (Sacks 1969 [1992: II: 137– 139]). (I have a feeling that Sacks’ specifically formulating this sort of material as something he found “enormously amusing”, or had “a fascination” with, was dealing with their on-the-face-of-it utter trivialness, even for the sort of mundane stuff he was known to work with. These “little tiny things” were perhaps beginning to be just a bit too tiny for comfort.)

Glossary of transcript symbols

But they’re “there” in the talk recorded on the tapes, and many of them are captured in the transcripts that use the system explicated in the ensuing glossary of transcript symbols. Some of them have led to the discovery of ranges of orderlinesses; most of them are yet to be explored.

Notes . In his introduction, Kutler says “I have edited the conversations with an eye toward eliminating what I believe insignificant, trivial, or repetitious . . . and often have omitted dutiful choruses of agreement by those present unless I believed them particularly important. The dialogue of innumerable uses of ‘right’, ‘yeah’, ‘okay’ often has been dropped . . . The ‘uhs’ and ‘ahs’ usually have been eliminated”. (p. vii) . This and subsequent fragments titled and referred to as “Jeff:Canc” are from a transcript I did in October, 2000 of the March 21, 1973 ‘Cancer on the Presidency’ meeting between Nixon and Dean. . Another possibility it raises is that Dan did say “bothered”, that Al did hear it as such, and that the sound- similarity of the first syllables of “both[ered” and “fath[er” triggered Al’s question to Ken about “father”. . For a discussion, and further cases, of this phenomenon, see Jefferson (1984: esp. 28–29). In its way, fragment (2) is also a case of ‘recognitional’ onset of talk, if not specifically of recognitional ‘response’. . A version of fragment (8), and a discussion of this phenomenon can be found in Jefferson (1979: 82–83). . Transcribing the Watergate materials, I’ve been struck by the delicacy of some of Dean’s interactional work. I didn’t notice anything of that nature in this particular fragment until I focused on it for this exercise, having chosen the fragment simply because it gave more detail to the Abuse of Power version, “...and I [laughs] almost hate to learn some of these things”.

References Jefferson, G. (1979). “A technique for inviting laughter and its subsequent acceptancedeclination”. In G. Psathas (Ed.), Everyday Language: Studies in Ethnomethodology (pp. 79–96). New York: Irvington. Jefferson, G. (1983). “Issues in the transcription of naturally-occurring talk: Caricature versus capturing pronunciational particulars”. In Tilburg Papers in Language and Literature, 34, 1–12. Tilburg: Tilburg University. Jefferson, G. (1984). “Notes on some orderlinesses of overlap onset”. In V. D’Urso & P. Leonardi (Eds.), Discourse Analysis and Natural Rhetorics (pp. 11–38). Padua: Cleup Editore.





Gail Jefferson

Jefferson, G. (1985). “An exercise in the transcription and analysis of laughter”. In T. van Dijk (Ed.), Handbook of Discourse Analysis, Vol. 3: Discourse and Dialogue (pp. 24–34). London: Academic. Kutler, Stanley I. (1997). Abuse of Power. New York: Touchstone. Sacks, Harvey (1992). Lectures on Conversation, Volumes I and II, G. Jefferson (Ed.). Cambridge, MA: Blackwell.

Glossary of transcript symbols //

[

]

=

Double obliques indicate the point at which a current speaker’s talk is overlapped by the talk of another. ((No longer in use.1 )) Louise: ’N how t//all are you, Al, Roger: How tall’r you Al. A left bracket indicates the point of overlap onset. ((The currently-used alternative to the double obliques. Note also, a change in descriptive language.)) Louise: ’N how tall [are you, Al, Roger: [How tall ’r you Al. A right bracket indicates the point at which two overlapping utterances end, if they end simultaneously, or the point at which one of them ends in the course of the other. It also is used to parse out segments of overlapping utterances. Louise: ’N how t[a l l u h r ] you ]↓A:]l, Roger: [How tall ’r] ↓you] Al, ] Equal signs indicate no break or gap. A pair of equal signs, one at the end of one line and one at the beginning of a next, indicate no break between the two lines. Maggie: . . .en ’e weighs about a hunnerd ’n thirdy five pounds.= Ronald: =AAUUGH! WHADDA L-LIE! The pair is also used as a transcript convenience when a single speaker’s talk is broken up in the transcript, but is actually through-produced by its speaker. Pammy: Yeah well okeedoe= Myra: =[Yeah. Pammy: [I j’s thought I’d ask

1. The asterisk (*) was used by some transcribers to indicate termination of simultaneous speech in conjunction with double obliques (//). (Ed.)

Glossary of transcript symbols

In this case, Pammy’s utterance is produced as “Yeah well okeedoe I j’s thought I’d ask”, with Myra’s “Yeah” starting up immediately upon completion of “okeedoe”, and simultaneously with “I j’s. . .”. A single equal sign indicates no break in an ongoing piece of talk, where one might otherwise expect it, e.g., after a completed sentence. Ehrlichman: . . . so I said I jis’ find that hard to ima↓gine.=Now (0.4) ·p ↑since ↓then I’ve retained coun↓sel. (0.0) Numbers in parentheses indicate elapsed time by tenths of seconds. Al: . . .j’s be a lot’v (shh) lotta work- lotta hassle. (0.2) Al: =[Well, Roger: [Well if yer goin’ t’ all that trouble, –– Double dashes indicate a short, untimed interval without talk, e.g., a ‘beat’. ((no longer in use)) Vic: I’m intuh my thing, intuh my – – attitude against othuh pih- ·hh (·) A dot in parentheses indicates a brief interval (± a tenth of a second) within or between utterances. Mrs A: ’Ello:? Guy: ’Ello is Curly there? → (·) Mrs A: → Oo jis (·) e˙ -Who:? Guy: Johnny?h An[sin?] Mrs A: [Oo j]ist ↑a minnih, _____ Numbers in parentheses bracketing several lines of transcript indicate time | (0.0) elapsed between the end of the utterance or sound in the first bracketed | _____ line and the start of the utterance or sound in the last bracketed line. Mrs A: _____ Oo jist ↑a minnih, (0.6) Kid: (1.2) ( [) | Mrs A: _____ [It’s fer you dea:r, ____ Underscoring indicates some form of stress, via pitch and/or amplitude. A short underscore indicates lighter stress than does a long underscore. Ehrlichman: Well Dean has: uh:,h totally coop’rated with the U.S. Attorney. :: Colons indicate prolongation of the immediately prior sound. The longer the colon row, the longer the prolongation.





Gail Jefferson

:__

↑↓

Mike: ↑iYeh it’s all in the ↑chair= Mike: =↓all th[at ◦ junk’s in the chair.◦ ↓] Vic: [W o : : : : : : : : : : ] ↓ : : : w : .= Vic: =I din’ know that? Combinations of underscore and colons indicate intonation contours. Basically, the underscore ‘punches up’ the sound it occurs beneath. wo:rd If a letter preceding a colon is underscored, the sound represented by that letter is ‘punched up’, i.e., an underscored letter followed by a colon indicates an ‘up-to-down’ contour. Kalmbach: Hi:.= Ehrlichman: =How’r you:. wo:rd If the colon is underscored, then the sound at the point of the colon is ‘punched up’, i.e., a letter followed by an underscored colon indicates a ‘down-to-up’ contour. Emma: Is SA:M there with [yuh?] Lottie: → [Y e : ]a h,= Emma: → =Uh ha[:h, Lottie: [Uh huh wo:rd If underscoring occurs prior to the vowel preceding the colon, then the entire word is ‘punched up’, i.e., the colon indicates prolongation only; there is no mid-word shift in pitch. Vic: ’M not saying he works ha:rd. In multi-syllabic words, if the consonent is underscored, then all syllables thereafter are ‘punched up’. Ehrlichman: He said e˙ -I came dih you:,hh fr’m Mitchell,hh en I sai:d,h uh↓ : Mitchell needs money? Here, the first mention of “Mitchell”, with only the initial consonant underscored, is produced with the entire word ‘punched up’, while in the second mention, “Mitchell” with the underscored vowel, pitch drops at the second syllable. Likewise, the entire word “money” with only the initial ‘m’ underscored, is ‘punched up’. Arrows indicate shifts into especially high or low pitch. ◦ Dan: Thet’s a good ↑ques↓tion.◦ (0.6) Louise: ↑↑Thank ↓you.

Glossary of transcript symbols

.,??

Punctuation markers are used to indicate ‘the usual’ intonation. (The italicized question-mark [?] substitutes for the question-mark/comma of my non-computer transcripts, and indicates a weaker rise than that indicated by a standard question-mark.2 ) These symbols usually occur at appropriate syntactical points, but occasionally there are such displays as the following. Maggie: Oh I’d say he’s about what.=five three enna ha:lf?=aren’t chu Ronald, Sometimes, at a point where a punctuation marker would be appropriate, there isn’t one. The absence of an ‘utterance-final’ punctuation marker indicates some sort of ‘indeterminate’ contour. WORD Upper case indicates especially loud sounds relative to the surrounding talk. Kalmbach: I returned it ’n went over the:re (·) tih↑da:y, (0.5) A::ND uh (0.8) he said the ↑rea:son thet. . . ◦ word◦ Degree signs bracketing an utterance or utterance-part indicates that the sounds are softer than the surrounding talk. Leslie: But we were ↑very ↓sorry to hea:r (·) that uh (·) ◦ your mother◦ had (·) died is that ri:ght Phi[↑lip? Philip: [Yeah. ... Ava: B’t the point is Jessie don’t fehget no:w. ·h (0.3) eh:m (·) ’e w’ so: close t’◦ Mart’n◦ wa:sn’t ↓ ’e. * Asterisk In some transcripts, the asterisk indicates percussive non-speech sounds, e.g., as in the following fragment, a fist thumping a table. Vic: BU(h)D I’M NO(h)T I(h)NTUH THA(h)*T! * * In non-computer transcripts, the asterisk indicates ‘creaky voice’. (In computer transcripts, I’ve stopped tracking ‘creaky voice’ and am using the asterisk for another phenomenon.) Emma: En ar air co.nditioner went out. comin’ ba:ck so Go*:d.= Lottie: =◦ O[h:: G*o:d.◦ Emma: [itEmma: Oh:: God ih w’z hot. . .

2. The inverted question-mark (¿) is also used as a substitute for the question-mark/comma by some transcribers. (Ed.)





Gail Jefferson

t*, d*

t, d

a.,˙e,ï

In early computer transcripts, an asterisk following a consonant replaces the single sub- or superimposed dot which serves as a ‘hardener’ in my non-computer transcripts. Kalmbach: I w’jist (·) understa:nd thet* uh: you en I are dehabs’ooly dihgether on tha:t, Ehrlichman: No question about it*?=uh hHerb In this case, while Kalmbach produces “jist” and “tha:t,” with the American-standard, soft ‘t’, the ‘t’ in his “thet*” is crisp, dentalized, i.e., ‘hard’. Similarly, while Ehrlichman produces “about” with the soft ‘t’, the ‘t’ in his “it*” is ‘hard’. In more recent computer transcripts, a boldface consonant replaces the single sub- or superimposed dot which serves as a ‘hardener’ in my noncomputer transcripts. The above fragment would now be shown as: Kalmbach: I w’jist (·) understa:nd thet uh: you en I are dehabs’ooly dihgether on tha:t, Ehrlichman: No question about it?=uh hHerb When a single dot is not available, two dots over a vowel replace the single sub- or superimposed dot which, as well as a ‘hardener’, serves as a ‘shortener’ in my non-computer transcripts. Ehrlichman: e˙ -he:: told me::? . . .an:d uh,h ï-he sid . . . Here, while conceivably the ‘e-’ in “e-he” and the ‘i-’ in “i-he” could be read as long sounds, “ee” and “eye”, the single dot over the ‘e-’ and the double dots over the ‘i-’, confirm that those sounds are short. I don’t show the sounds as “eh” and “ih” because they are more fleeting than those spellings indicate. The dots do an additional job in transcripts where I use non-standard orthography. Many words get a range of oddball spellings, in keeping with the range of pronunciations they are subject to. On occasion such a word appears in its standard spelling. If that word carries the dot(s), it means that while such a spelling could be the result of a lapse of transcriber concentration, in this case the standard spelling does indicate the way the word was pronounced. Emma: En ar air co.nditioner went out. comin’ ba:ck Here, while ‘air conditioner’ is routinely pronounced as ‘air c’nditioner’, it is being given a fully formed vowel, shown as “co.nditioner”. ... (a) Kalmbach: Ehm: I:’m uh scheduled fo.r ↑two duhmorrow afternoo:n. . .

Glossary of transcript symbols

(b) Kalmbach: . . .he said the ↑rea:son thet wz: u.-fer the ca:ll w’z. . . In this case, while at point (b) Kalmbach is shown pronouncing the word ‘for’ as “fer”, the dot below the ‘o’ in “for” at point (a) indicates that it’s not that the transcriber had simply written the word in its standard orthography, but that it is there fully pronounced as “for”. (b) A parenthesized italicized letter replaces the parenthesized letter with a sub- or superscribed degree sign which, in my non-computer transcripts, indicates an ‘incipient sound’. Emma: you couldn’ ev’n putcher hand ou:ts:I:de the CAR ih jiz’(b)bu:rn. ... Ehrlichman: But they- (·) th˙e(p) the point is. . . whord An italicized ‘h’ appearing in such a word as ‘which’ ‘where’, ‘what’ ‘when’, ‘whether’, etc., indicates that while such words are often produced with the ‘h’ silent (as if they were the words ‘witch’, ‘wear’, ‘wen’, ‘weather’, etc.), in this case the ‘h’ was sounded. Ehrlichman: En I said well Joh:n what ’n the world er yih talking ↓about*. . . Ehrlichman: See ↑what they’ve said duh Dean is. . . While at one point in a conversation Ehrlichman pronounces the word ‘what’ with the ‘h’ sounded, at a later point, the ‘what’ is produced with no ‘h’. word An italicized letter replaces the sub- or superscribed degree sign which, in my non-computer transcripts, indicates unvoiced production. Ehrlichman:→ He said we:ll?=hmhh e˙ -I came dih you:,hh fr’m Mitchell,hh en I sai:d*,h uh↓ : Mitchell needs money? (Kalmbach):→ (◦◦ Right◦◦ ) Ehrlichman: Uh::: could*=uh we::: ca::ll Herb Kalmbach en ask ’im duh raise ↓some. Kalmbach: → ◦◦ Yeah.◦◦ <word A pre-positioned left carat is a ‘left push’, indicating a hurried start; in effect, an utterance trying to have started a bit sooner than it actually did. This can be heard, for example, as a compressed onset of the utterance or utterance-part in question. A common locus of this phenomenon is ‘self repair’.





Gail Jefferson

Ruth:

Monday nights we play, (0.3) < Right/left carats bracketing an utterance or utterance-part indicate that the bracketed material is speeded up, compared to the surrounding talk. <> Left/right carats bracketing an utterance or utterance-part indicate that the bracketed material is slowed down, compared to the surrounding talk. ·hhh A dot-prefixed row of ‘h’s indicates an inbreath. Without the dot, the ‘h’s indicate an outbreath. wohhrd A row of ‘h’s within a word indicates breathiness. In some transcripts the ‘h’s are italicized, in some not. Colson: . . .a ghhuy wh(h)o’s olso totally loyal< And in computer transcripts, the ‘h’s are sometimes also superscribed. Th:e ↑thing thet worries me is the(p)(·) is the phh ossibility’v. . . (h) Parenthesized ‘h’ indicates plosiveness. This can be associated with laughter, crying, breathlessness, etc. Jim: Don’t sound so (h)amp(h)itious fer Ch(h)rise’sake (h)ih suh ·hh sou’ l(h)i’ yuh k(h)uh g(h)o tuh sleep ’n the pho(h)one. ... Maggie: I j’st ran up th’ stai(hh)rs that’s wh(h)y I’m huffing en puffing. £ The pound-sterling sign indicates a certain quality of voice which conveys ‘suppressed laughter’ (various transcripts have other symbols, e.g., the Dutch guilder-sign which no longer exists). Ken: ahh ha £I don’ know who’s payin’ fer thi(h)s I ↑think ih my fa:-ther.£ wghord A ‘gh’ stuck into a word indicates gutteralness. In some transcripts the ‘gh’ is italicized, in others, not. Mike: Ah don’ think ’ee lives onna ↓groun’ flo↓ :h. (0.3) James: The: ghghroun’ flo’ Colson:

Glossary of transcript symbols

In this case, a speaker with phleghm in his throat is saying “the ground floor”, with the word ‘ground’ heavily gutteralized. () Empty parentheses indicate that the transcriber was unable to get what was said. The length of the parenthesized space reflects the length of the ungotten talk. Mike: No. (0.4) Mike: ( ), In the speaker-designation column, the empty parentheses indicate transcriber’s inability to identify a speaker. Roger: Paz’m z’m Miller Highlit*e.h (0.5) ( ): hnh Yhehh (word) Parenthesized words and speaker designations are especially dubious. (Mike) [(Lee me alone.)] Carol: [Mike I know yu]h’ love’ m (blerf) Nonsense syllables are sometimes provided, to give at least an indication of various features of the un-gotten material. Nixon: Jerry sh’d talk to ↑Witnaw. (0.5) And uh: (·) jis brace eem ’n tell ’im tih (·)(offih sebbatikiss). . . (Ø) A nul sign indicates that there may not be talk occurring; that what is being heard as possibly talk might also be ambient noise. Nixon: _____ ◦ (Well ah’ll protect chu but* uh)◦ | ( Ø ): (0.7) (Okay.) | Nixon: __|__ ◦ (Thet uh)◦ thet’s that’s why:. (0.9) ◦◦ I◦◦ (0.5) can’t let chu ↓go (0.2) go dow:n. (( )) Doubled parentheses contain transcriber’s descriptions. Ray : ehh-heh-heh-heh-heh-he:h-eh= Maggie: =((dainty snort)) ... Vic: ((dumb slob voice)) Well we usetuh do dis



P I

Taking turns speaking

An initial characterization of the organization of speaker turn-taking in conversation Harvey Sacks

.

A positioning of conversation among the speech-exchange systems1

1.0 While the talk that participants to any conversation do is quite variably distributed among participants, one massively evident social organizationally relevant orderliness their talk’s distribution exhibits is the taking of turns at talking. Though speakers change, it is overwhelmingly true that one party talks at a time in conversations, and that feature of conversation is preserved across variation in the number of parties to a conversation, its length, the relative amount each party talks, the relative size of their turns, etcetera. The feature’s preservation must take work, the taking of turns must be organizationally achieved. Here, on the basis of audio recordings collected from naturally occurring conversations, we attempt to characterize, in its simplest systematic form, the organization of the taking of turns at talking in conversation. 1.1 That talk proceeds in a one party at a time fashion while speaker change recurs is not unique to conversation: It is massively present as well for debates, meetings, press conferences, plays, therapy sessions, interviews, trials, etcetera, although these latter differ from conversation in how the feature is preserved. Nor is the feature unique to a particular linguistic or social community. It is evidently exhibited in conversation, meetings, etcetera for e.g., societies whose languages and systems of social organization quite drastically differ. 1.2 Examination of a range of the various techniques used for allocating turns in conversation permits an initial generalization which is of very considerable import for the character of turn-taking’s organization in conversation and also



Harvey Sacks

because it is plainly not correct for other2 than conversation. The generalization is: The techniques for allocating turns in conversation operate in a one turn at a time fashion: On each operation of a turn allocating technique one turn only is allocated. Alternatives to such a mode of operation are readily found. In debates, for example, the ordering of all turns is pre-allocated, by formula, be reference to ‘pro’ and ‘con’ positions. In contrast to both debates or conversation, meetings that have a chairperson partially pre-allocate turns, and provide for the allocation of unallocated turns via use of the pre-allocated turns. Thus, the chairperson has rights to talk first and after each other speaker and can use each such turn to allocate next-speakership. 1.3 One structural possibility that the foregoing comparison suffices to suggest is that turn-taking systems, or at least the class of them whose members each preserve one party talks at a time, are, with respect to their allocational arrangements, linearly arrayed. Thus, the foregoing may be restated as proposing a linear array in which one polar type involves ‘one turn allocation at a time’, the other pole involves ‘pre-allocation of all turns’, and medial types involve various mixes of pre-allocational and local allocational means. That the types so array themselves turns out to permit them to be compared directly in relevantly functional terms: As, one pole organizationally permits a maximization of the size of the set of potential next speakers to each current turn, whereas the other permits maximization of equalization of the number of turns each party gets, maximization, organizationally, of either of these being, generally, inconsistent with maximization of the other. Since a maximization of either of such functions is (otherwise than where two parties are involved) inconsistent with maximization of the other, it is specifically notable that the array of types is co-present, that where one allocation at a time is used, partially pre-allocational and fully pre-allocational means are also organizationally available and used. 1.4 Given the linear array, the polar position of conversation, and the function that position permits maximization of, characterization of the organization of turn-taking in conversation takes on more than merely ethnographic interest. For, occupying such a functionally interesting structural position the turn-taking system we describe is at least one representative of how this polar possibility may be organizationally achieved.

The organization of turn-taking in conversation

1.5 All turn-taking systems on the linear array use turns and preserve one party talking at a time. While they do each specify these differently, and while our systematic characterization will be for conversation only, one further generalization and the orderliness of the difference it leads to noticing may be mentioned. For all positions ‘turns’ are at least partially organized via language specific constructional formats, e.g., syntactic construction.3 For all positions, turns, in sentence terms, can begin with sentence starts and close with sentence ends, however many sentences they are composed of. And, while turns are not specifiable in terms of some number of sentences (debates, for example, specify turns in terms of elapsed time), it does seem correct that turn size is correlated with position on the linear array, increasing from conversation, through meetings, to debates.

. A simplest systematics for the organization of turn-taking in conversation For conversation, preservation of ‘one party talking at a time’ is organizationally primary. Turn-allocational and turn-constructional techniques are used which provide for just one next speaker and for a minimization of both ‘gaps’ and ‘overlaps’ between turns. The turn allocational techniques are distributed into two groups; those that allocate next speakership via current speaker selecting a next, and those that use self-selection for next speakership. The use of these two sorts of allocational techniques poses a range of integrational problems, for the two are not directly compatible. If, for example, both were usable on any occasion in which one of them was usable, then the possibility of more than one party being selected will have been provided for by the very techniques whose operation should yield only one next speaker. That is, if both were used, each technique involving a different party using it, then unless it happened that current speaker selected that next speaker who was self-selecting, more than one will have been selected. Furthermore, the minimization of both gaps and overlaps also poses integrational problems. Except when a simultaneous no gap and no overlap transition between speakers is achieved, minimization of either one is not directly compatible with minimization of the other. Means for reducing gaps encourage overlaps, and vice versa. While there are constructional and allocational techniques whose combination does methodically permit achieving just one next speaker and a simultaneously no gap and no overlap transition between





Harvey Sacks

current speaker and that next, such techniques are neither generic nor even prototypic. Therefore, while the potentially strongest version of ‘one party talking at a time’ – that exactly one party (i.e., at least and no more than one party) talks at a time, across change of speakers – is methodically achievable, the slightly weaker version, which minimizes both gaps and overlaps between turns while providing for one next speaker is, in general, the way that ‘one party talks at a time’ is specified for conversation. It is that conjunction of features which defines the general problem that the organization of turn-taking in conversation systematically solves. For turn-taking in conversation to be organized, some way of integrating the two techniques is required so as to preserve one-talking-at-a-time across change of speakers while minimizing both gaps and overlaps between turns, for turns that use sentential constructional formats. The following set of rules seems basic. 1. If a current-speaker-selects-next-speaker technique is used, then the party its use selects has rights to, and is obliged to, take next turn to speak, and all others are excluded. 2. If a current-speaker-selects-next-speaker technique is used, then on the next possible completion of the sentence current speaker is constructing, transition should occur; i.e., current speaker should stop and next speaker should start. 3. If, by any next possible completion of the current sentence of a turn, current-speaker-selection of a next has not been done, self-selection may – but need not – be instituted, with first starter acquiring rights to a turn at talk. 4. On any next possible completion of some current sentence, current speaker may stop, but unless he has done selection he need not stop unless another has self-selects.4 The foregoing set of rules powerfully integrate the two turn-allocative technique groups. An ordering of their usabilities can make them compatible, but the rules do rather more than that. Rule 1 may initially be appreciated as giving first-option status to the use of current-speaker-selection of a next speaker. Then, by specifying conditions under which that first option is lifted and self-selection is made potentially usable, Rule 3 serves to constrain how far a turn at talk can proceed before current-speaker-selection is done – if the chance to do it is to be assured. Unless such selection is done before the first possible completion of the first sentence

The organization of turn-taking in conversation

in a turn, the chance to allocate next turn is not assured. For at that point, Rule 3 becomes operational, and current speaker’s turn may be terminated by a self-selecting speaker. Rule 3 makes self-selection optional. Rule 4 permits current-speaker continuation if current-speaker-selection-of-a-next has not been done. Then, Rule 1 may further be appreciated as providing for the within-turn recurrence of current-selection as a first option; that is, after any last possible completion and before any next. And the recurrent usability of Rule 1 serves to constrain when, in a turn, self-selection must be done if the chance to use it to acquire next turn is to be assured. Unless a potential next speaker uses selfselection on his first chance he may lose the chance to acquire next turn, for thereafter, current speaker can continue his turn, and can, furthermore, use current-selection before self-selection recurs as a possibility. But the possibility of self-selection does recur unless current-selection-of-a-next has been done, within each recurred continuation of a turn, before the next possible completion of its current sentence. So, the possibility of self-selection does continue to constrain the usability of selection by current speaker, by providing discrete positions at which self-selection may be done, thereby potentially cutting off current speaker’s chance to allocate next turn. In fine, the rules integrate the techniques powerfully by providing that the possibilities of self-selection constrain the placement of current-speakerselection if it is to be assured, and by providing that the possibility of currentspeaker-selection be a recurring first option that thereby constrains the placement of self-selection, and by providing that self-selection recurs discretely as a possibility, which constrains the placement of current-speaker-selection, ad infinitum, for a turn or some series of turns.

Achieving minimization of gap and overlap Since Rule 2 provides that if current-speaker-selection of a next speaker is done, transition should occur on next possible completion – which can be on the first possible completion of the first sentence in a turn – and since Rule 3 can be employed to permit self-selection on a first possible completion, a general minimal turn unit exists, one that is consistent with the operation of either allocative technique. That unit is: a sentence to its first possible completion. Since Rule 3 is optional, and since Rule 4 allows continuation beyond any next possible completion if current-speaker-selection has not been done and selfselection has not been ventured, turns can be subject to recurrently terminable





Harvey Sacks

expansion beyond that minimal size. There is, then, no general specification of turn size for conversation. Actual turn sizes depend on the use of the allocative techniques to occasion transition. Adapting actual turn sizes to the use of the allocational techniques is not only crucial to the integration of those techniques, but it also enables that integration to be made consistent with a minimization of both gaps and overlaps between turns. In order to develop the latter, some specification of ‘gap’ and ‘overlap’ must be introduced. We do that by offering a differentiation between ‘gap’ and ‘pause’ – these constituting the two ways that less-than-one-speaker-ata-time occurs – and ‘overlap’ and ‘interruption’ – these constituting the two ways that more-than-one-speaker-at-a-time occurs. The distinctions are further motivated by the fact that the turn-taking organizational system we are characterizing concerns itself with ‘gaps’ and ‘overlaps’ as systematically-based possibilities, but not with ‘pauses’ and ‘interruptions’. A ‘gap’ is a between-turn silence and differs from a ‘pause’ which is a within-turn silence. Turn-taking-organizational preservation of gap minimization is not concerned with pauses. Then, by virtue of Rule 4, some possible gaps may be rendered pauses. Current speaker having reached a possible completion and having stopped to allow for self-selection may, none having ventured it, produce a ‘continuation’ of his turn, continuing its possibly complete sentence or starting another. Having done that, the silence that would have been a gap had transition occurred after it, will be turned into a pause. Where transition is involved – whether self-selectionally or via current speaker’s selection of a next – gaps may be avoided or reduced by newly starting speakers via the use of available techniques for turning the potential gap, or part of it, into a pause. Such a technique as involves [“Uh” + pause + sentence] is a means for transferring a silence from before, to within, the talk of a turn. A speaker can then start a turn with a sequence of filled and unfilled pauses before he is prepared to produce the sententially organized talk with which his turn will be occupied, and by so starting he can avoid or reduce gap size. If by ‘interruption’ we understand such a start as is projected to occur within another’s turn, then interruption differs from ‘overlap’, if we understand by the latter such a start as is projected to occur on another’s possible completion while intendedly avoiding a gap. The sources of overlap are then independent of the sources of interruption, involving, for example, a correct appreciation of the forthcomingness of possible completion, where the intention to avoid a gap misfires for any of a large range of reasons, of which the following are offered merely instantially.

The organization of turn-taking in conversation

Current speaker might extend the length of a correctly appreciated last word, and next speaker, intendedly projecting his start to occur on that word’s completion may then produce an overlap. Or, current speaker may append to some correctly appreciated possible completion such a word as can optionally close a turn – for example, an address term – and intending next speaker, being unprepared for that and attentive to the forthcomingness of a completion, may overlap that address term. As either of these possibilities can be engendered by a coincidence of current speaker attempting to avoid a gap while intending next speaker is attempting to avoid overlap, their occurrence is irremediable. But interruption, as it involves a start that is projected to occur within another’s turn, does not have the minimization of gaps as a basis or justification for its occurrence. If that start had not been done, a gap would not expectably occur. The illegality of interruption derives directly from the ‘no more than one’ facet of that feature which is organizationally basic to conversation and also to debates, interviews, meetings, etcetera, and its problematic status is not particular to conversation or its turn-taking organization, as overlap is or can be. Given the foregoing specifications, we may return to the problem of this section. By virtue of Rule 2, where current-speaker-selection of a next has been used, overlap may be avoided. The selected speaker, having exclusive rights to a turn, is not motivated to start before completion has occurred, as a means of assuring his acquisition of that turn. Therefore, current selection of a next can allow for gaps. Alternatively, by giving rights to a turn to first starter on a possible completion that has not involved current speaker’s selection of a next, a minimization of gaps is provided for, and thereby some tendency to develop overlaps is engendered. These two tendencies – towards gaps if selection by current has been used, and towards overlaps if self-selection is used – seem evident from an examination of conversation. There are, however, means for minimizing each of them. Where current-speaker-selection of a next is employed, transition is treated as having occurred on current speaker’s completion according to Rule 2, and an ensuing gap is treated as attributable to the selected next speaker. Such a gap is usable for appreciating his talk, as hesitant and the like; an appreciation which he can avoid by producing his talk as quickly as possible, thus avoiding having a gap attributable to it. Alternatively, a gap attributable to a selected next speaker may be appreciated as engendered by some feature of current speaker’s talk. And there is a set of techniques for reducing the size of such a developable gap which does not rely on the occurrence of talk by the selected speaker – on his motivation to limit the size of a gap in order to avoid such a





Harvey Sacks

reading of his talk as that gap’s size permits. For example, on the occurrence of some gap, the selecting speaker may use repeats and various other specifically second selectors to re-occasion transition, and perhaps to re-occasion it in a gap-minimizing fashion. Where it is self-selection that is being ventured, one sort of bound on such a positioning of its start as results in overlap is that only such an overlap may be done as cannot be construed as constituting an interruption. The range of rules which penalize interruptions are available controls on starts that can be construed as an interruption. For example, the interruption producing a situation in which at least two are talking, a return to only one talking is achievable by one of them stopping. It appears that the rule provides that the interrupter should stop in a situation of two or more talking with interruption as its source. A self-selector finding himself in such a position will have had his move to acquire next turn renderable ineffective by the continuation current speaker does after the self-selector has stopped. If he does not stop, his talk is subject to a variety of other remedial actions which interruptions permit – for example, complaints, interruptive complaints, requests for apologies, and the like.

Notes . The original heading for this section seems to read, “The Position of Conversation among the Speech-Exchange Systems.” Sacks has then erased “The” and replaced it with “A” and appended “ing” to “Position.” (Ed.) . A marginal note seems to indicate that Sacks is questioning this claim. Gail Jefferson (personal communication) has suggested that the marginal note may offer “all other” or “some other” as a replacement here. (Ed.) . There are two extant versions of the so-called “little turn-taking paper.” In the second version, beginning with this paragraph, the remainder of the manuscript has been retyped incorporating handwritten corrections from the first version and making several other mostly, but not entirely, stylistic changes. (There are a couple of handwritten notes on this version as well.) The remainder of the text published here follows the retyped version, since it incorporates earlier changes Sacks made to the original typescript. One stylistic change readers will notice is that the remainder of the paper does not retain the paragraph numbering system found in the first part of the paper and found throughout the first version of the paper. Also, there were only two section headings in the first version (here given as Sections 1 & 2), while the second version contains three section headings. One important substantive change made in the first sentence in this paragraph is the replacement of the phrase “i.e., sentential construction” with the more general formulation “e.g., syntactic construction”. (Ed.) . Note that “unless another self selects” is a handwritten addition to the first version that was incorporated into the typescript of the second version. (Ed.)

A sketch of some orderly aspects of overlap in natural conversation Gail Jefferson

Overlapping talk initially appears to be messy – a product of people’s not much attending each other. Our transcription procedures make it possible to focus on such places in conversation, and to find preliminary indications of intense co-attention and orderliness. An array of conditions can be sketched out, each with orderly procedures attendant to it. Overlap Onset can be the product of devices assigned to perform specifiable activities. Within-overlap talk can involve procedures for resolving overlap and/or for competition for the turn space. Post-overlap talk can be examined for its relationship to the prior overlapping talk, and can involve procedures for retrieving talk potentially impaired by its occurrence in overlap. 0. For a number of years transcripts have been produced which capture the precise point at which an overlap of two utterances begins. In this paper the point of overlap-onset is indicated by a left-hand bracket in the ongoing utterance with the starting utterance positioned on the next line down, e.g.: Fran: Jim:

He’s not gunnuh li:sten [tuh tha::t, [I’m not sayin-

More recently, attention is being paid to the point at which overlap ends, and in this paper overlap resolution is indicated with a right-hand bracket. Fran: Jim:

He’s not gunnuh li:sten [tuh tha: :t, ] [I’m not say]in-



Gail Jefferson

Some results have been reported (e.g., Jefferson 1973; Schegloff 1987 [1973]). Basically, it appears that overlap can be an orderly phenomenon, and, as inquiry proceeds, its orderliness is turning up in finer and finer detail. I. It has been found that overlap-onset can be the product of systematic procedures. Specifically, a party can precision-place his talk in the course of another’s, can select and hit a target point. For example: 1. A display of independent knowledge of what is about to be said can be achieved by starting to talk just as some object comes due in an ongoing utterance. Joe: Carol:

So he come[s home one night’n the sonofa] bitch [bit him. [ heh heh heh heh heh heh ] [bit hi:m

2. A display of recognition of what is in the course of being said can be achieved by starting to talk midway through the recognized object. Caller: Desk: Desk:

Fire Department, out at the Fairview Food[mart there’s a[Yes. We’ve already got the uh call on that ma’am,

3. And it appears that a not infrequently targeted starting point is the moment of completion of an ongoing utterance, no sooner and no later. That activity is a “latch”, and is indicated by equal signs at the end of the prior and beginning of the next utterance. Earl: Bud:

How’s everything look.= =Oh looks pretty goo:d,

That procedure is not only used by interacting parties but can be a noticed and reported event. Vic:

uh- at the beginning of it all befaw any chair moved outta ’ere I says Carol you want some a’ th’m barbuh chairsNo.

And such a procedure provides a systematic locus of overlap. For example, when an ongoing speaker turns out to have stretched his last syllable and a next speaker is starting up in latch position, overlap occurs. The prolongation of a sound is indicated by colons. Joe: Mike:

Just like tha:[:t. [Right.

Overlap in natural conversation

Or, for example, when a possible completion point turns out not to be the point of actual completion and ongoing speaker appends a syntactically coherent next utterance component while a next speaker is starting up in latch position, overlap occurs. Bert: Fred:

Uh you been down here before [havenche. [Yeh.

4. Another routine locus of overlap is post a possible completion and a pause. Regularly enough, more than one party simultaneously start to talk. This feature holds for relatively long pauses, for example, seven tenths of a second. Timing is shown in parentheses between the two utterances. Ava: Ava: Bee:

He’n Jo were like on the outs, yihknow? (0.7) [[So uh, [[They always a(h)re hhh!

It holds as well for relatively short pauses, for example, one and a half tenths of a second. James: Vic: James:

(I’ll) set it dehr own the sidewalk. (0.15) [[No. [[Izzat ehkay No.

II. There are, then, various indications that overlap-onset can be a systematically generated occurrence. This leads to inquiry into possible systematicity within overlap: i.e., examination of how parties deal with the fact that they are producing more-than-one-party-at-a-time talk when a fundamental feature of conversation is that one party talks at a time (Sacks, Schegloff, Jefferson:1974). A basic course of action is to resolve the overlap; i.e., one of the overlapping parties drops out, i.e., stops talking, and a state of one-at-a-time is (re)established. This feature appears to hold across a set of starting positions for overlap, e.g., one of two simultaneously starting parties can be found to drop out. Tracy: Lady:

[[T[[Ye:s.

First starter can be found to drop out as second starter begins. Essie: Janet:

I think Cookie [ta[I didn’ even know’e was i::ll.





Gail Jefferson

And a second starter can be found to drop out immediately after an attempted start. Dan: Louise:

May [be yer brother is . . . [( )-

Multiple serial starts and drop outs can be found in association with an ongoing utterance which has several points of possible completion plus continuation by ongoing speaker, with a next speaker attempting to start at latch position for each possible completion point. In this fragment the equal signs indicate no elapsed time between one utterance component and a next. The utterance is produced as a single continuous object by its speaker and has been decomposed for clarity in this fragment. Polly: Janet: Polly: Janet: Polly: Janet:

I jus’ thought it was so kind of stupid= =[[Y[[I didn’ even say anything= =[[Eh[[when I came ho:me. (0.3) Well Essie jus’ called ’n I- an’ I aftuh call ’er back . . .

(This particular instance contains a possible consequence of the serial starts and dropouts at possible completion points. Notice that the party who had been starting up at each possible completion point (and who thereafter permits a pause before a next attempt post a next possible completion), herself produces an utterance with a first possible completion and a continuation (“Well Essie jus” called” + “‘n I aftuh call’er back. . .”). Note that at the juncture point of those two segments, i.e., just the sort of point in her own utterance at which she had been attempting entry into the turn space occupied by the prior utterance, she produces a cutoff and restart: “Well Essie jus’ called [’n I- an’ I] aftuh call ’er back. . .” It appears that the prior juncture-point problem has been carried over into a subsequent, otherwise undisrupted, utterance.) III. While a fundamental feature of conversation is that one party talks at a time, and a basic procedure for achieving such a state from a state of overlap is that one party drops out, we also find that it is not always unequivocal for participants who shall drop out. For example, we find each party to an overlap dropping. Thereafter, following a ‘micropause’ (for now, roughly, an untimed pause of less than 2/10 second, indicated by a dot in parenthesis), one of them takes and is given the turn space.

Overlap in natural conversation

Johnson: [[IRoberts: [[Uh(·) Johnson: I heard uh first that there was really some water in. . .

And in some cases both parties drop, both parties start up again after a micropause, both parties drop again, and subsequently one takes and is given the turn space. Edna: Bud: Edna: Bud: Bud:

[[Well[[Un(·) [[uh[[less(·) If ev’rything goes well.

And in the following, after an initial pair of drop/restarts, both parties pursue their utterances a bit further before each again drops. Subsequently one party takes and is given the turn space. Tracy: Lady: Lady: Tracy: Lady: Tracy: Lady:

But [had[But (·) [[this[[hh Ha(·) [[Ni:neteen s e]v e n]ty I, [[Hadju finished] ( )] (·) was:: had ’n invitation to, (0.3) Bucking’m Palace,

IV. Given that parties can and do drop out of overlap almost instantly, it becomes an observable event when one or both persevere beyond an initial drop point. And given a possible initial equivocalness as to which of them shall drop out, perseverance can be seen as negotiation for which of them shall drop, one (or both) parties indicating to the other that he is not dropping and the other should. Some systematic procedures for negotiation within overlap can be sketched, and two types of procedures can be isolated: Within-word pronunciational adjustments; i.e., manipulation of the sounds of the word a speaker is currently producing (its speed, pitch, amplitude, etc.), and within-utterance segmental





Gail Jefferson

adjustments; i.e., manipulation of larger parts of the utterance a speaker is currently producing (its words, clauses, phrases, etc.).

1. Pronunciational adjustments (a) “Stutters”. Depending upon our transcripts’ detailedness, stutters are roughly or precisely available for their relationship to overlapping talk. Some collection of them appear to be roughly or precisely coterminous with that overlapping talk, and the subsequent, unstuttered portion of the utterance coincides with overlap resolution. Johnson: [[I’m glad to hear it. Roberts: [[But- uh- uh, uh understand that um Franklin . . . ................................................... Gladys: [[EnEdna: [[En you need [som:e ]uh,] Gladys: [S- s- ] sh:]redded lettuce?

(b) “Stretches”. In the following fragments a word is prolonged, and the prolongation roughly or precisely coterminates with the utterance it overlaps, the production of the subsequent portion of the utterance coinciding with overlap resolution. Ken: Heck [a lottaRoger: [Les:::::::try it! ................................................... Carol: Tha:[t’s what they sa:y, ] Denise: [will soo:::::::::::::: n] learn.

The observable cotermination of stutter or stretch with the overlapping talk leads to a posing of these objects as resources for overlap management. They can be extended across the span of the overlap, permitting subsequent utterance components to be produced clear of overlap.

2. Segmental adjustments A similar provision for some part of an overlapped utterance’s occurrence in the clear can be observed for these devices. While the pronunciational adjustments are perhaps addressed to preserving talk across an overlap, the segmental adjustments appear to be explicit attempts by one party to claim a turn space occupied by another. So, for example, in the following fragments we can have

Overlap in natural conversation

an intuitive sense that one party is Turn Occupant, the other Turn Claimant, with Turn Occupant simply producing an utterance while Turn Claimant produces serial segmental adjustments. Mike: Vic: Mike: James: Mike:

[[Th[[Y’know I cut [m y s e l f [o n y o[u : r f r e a k i n [gla:ss,] [Th’ least ’e [ c’do- [Th’ least ’e coulda [do::ne,] (·) [[ Y e : h , ] [[Least ’e c]’d’v done w’z come dow:n en letchu know what happen’

Of the two overlapping utterances, one has a phrase recycled over the continuous course of another, permitting some projectable part of the recycled utterance to occur clear of overlap. The same feature holds for the following fragment. Ken: Roger: Roger: Ken: Ken:

No, they’re women who’v devo[ded their l[They’re women that hadda= =[[bad love [life’n became nuns.hh [heh hh ! [[their [their life[their life, to uh (0.6) the devotion of the church.

(In this particular case there is some indication that projection of a component and reservation for overlap-cleared production can be an abstract matter; i.e., whether or not the component’s actual content is known in advance, is formed up for speech-delivery, etc., it can, in principle, as not yet more than a projected component type, be reserved for overlap clearance by a within-overlap recycling of its prefatory segment(s).) In these fragments, while Turn Claimant’s work is obvious, Turn Occupant’s talk, as work, is not transparent, i.e., it is not obvious that Turn Occupant is doing something vis-à-vis the state of overlap in which his talk is occurring, by constructing a single continuous utterance. To get a sense of it as work, we might consider the sheer acoustic battering Occupant’s talk is subjected to. More to the point, we might consider each recycle by Turn Claimant as a request that Occupant relinquish the turn space, and each ‘continuation’ by Turn Occupant as a declination to relinquish. There are, then, two distinct forms of overlap competition occurring in these fragments; one announces a trouble and attempts to remedy it (via the recycled utterance segments) and one neither recognizes nor attempts to remedy





Gail Jefferson

an observable trouble. We might call the former procedure Marked Competition and the latter Unmarked Competition. Both do competition, but one announces it is doing that and the other proceeds as if nothing of the sort were occurring. Such a distinction can be informative about procedures used by various parties to an overlap competition. It turns out to inform inquiry into activities subsequent to overlap, as well. V. Once an overlap has been resolved there can be a problem: What, if any, of the talk which occurred in overlap shall have been ‘heard’; i.e., shall have consequence for subsequent talk? There seems to be a collection of procedures by which talk that is possibly hearing-understanding impaired via the state of overlap in which it occurred can be retrieved. These procedures fall into two types: Self-Retrieval, via which a party to an overlap provides for his own talk’s consequence, and Other-Retrieval, via which a party to an overlap provides for someone else’s, not his own, talk’s consequence. Further, these two groups each have Marked and Unmarked forms; forms which announce trouble and explicitly retrieve talk out of the prior overlap, and forms which do not recognize trouble nor explicitly retrieve talk out of overlap.

1. Marked self-retrieval: Restart A party to an overlap, having dropped out to resolve overlap and re-establish a state of one party talking at a time, can, upon the other party’s dropping out or reaching completion, retrieve the utterance he himself had potentially relinquished by restarting it. This procedure holds across the three positions for overlap-onset; i.e., at simultaneous starts: Edna: Olive: Edna:

[[Hy [[Yeah. (·) Hide it.

At first starter’s dropping out as second speaker starts: Rich: Carol: Rich:

I think if [ you[Am I right? (·) If you bring it intuh them.

Overlap in natural conversation

And at second starter’s dropping out after an attempted start: Louise: Roger:

I w’nd’ if [rilly is someone back the[re, [That[That attracted their attention,

This retrieval device appears in affiliation with overlap competitive procedures; i.e., a party may continue to talk until the other drops out and immediately thereafter perform a restart (see Schegloff 1987 [1973] for a detailed consideration of this procedure). For example: Ann: [[ H e : ha- ] Marty: [[Course wi-] widespread is a double edged swo:rd. ................................................... Bee: T! [ Except thet cl]u. Ava: [that’s not ba:d.] (0.2) That class is suh:: yihknow this is the Indian class. . . Bee: ................................................... Fran: He’s not gunnuh li:sten [tuh tha: :t, ] Jim: [I’m not say]in- I’m not sayin that . . .

2. Unmarked self-retrieval: Continuation A party to an overlap, having dropped out to resolve overlap and re-establish a state of one party talking at a time, can, upon the other party’s dropping out or reaching completion, retrieve the utterance he had potentially relinquished by continuing from the point of dropout. Whereas the prior retrieval device announces, and specifically pulls a word out of, overlap, this device proposes that the overlap, as an event which might have consequence for the coherent production of an utterance, was of no consequence. For example: Fred: Bert:

[[ The[[Yah(·) Fred: waves ’r about tuh wash us away. ................................................... Roger: I happen tuh wear buloo jeans constantly. (0.3) Ken: Well, (·)





Gail Jefferson

Roger: Ken: Roger:

Even [in[so do I now, (·) formal occasions, y’know? hheh hh!

The materials considered earlier as instances of problems as to who should drop out (p. 47), can be re-examined for the presence of marked and unmarked self-retrieval. So, for example, in one case both parties appear to be doing unmarked self-retrieval; i.e., continuation, one constructing a continuous “Well uh” while the other constructs a continuous “Unless”. Edna: Bud: Edna: Bud:

[[ Well[[Un(·) [[ uh[[less-

And in another case, one party does unmarked self-retrieval; i.e., constructs a continuous “But this. . .” and then a continuous “Ni:neteen seventy I, was:::...” across overlap and micropause, while the other does marked self-retrieval, restarting each time, “But had-”, “‘hh Ha-”, “Hadju finished ( )”. Tracy: Lady: Lady: Tracy: Lady: Tracy: Lady:

But [had [But(·) [[ this[[ ·hh Ha (·) [[ Ni:neteen seventy I, [[ Hadju finished ( ) (·) was:: had ’n invitation . . .

3. Marked other-retrieval: Repeat request This procedure appears to have some regularity of occurrence after competitive overlap. One of the competing parties announces trouble and explicitly initiates repair procedures by requesting a repeat of his co-competitor’s overlapped utterance. In the following fragment, each party displays Turn Occupancy and the continuously produced utterances reach completion simultaneously. There-

Overlap in natural conversation

after, one way to characterize the conditions under which it is decided who should yield is to find that a continuation of prior talk has yielded to initiation of a new topic; the explanation “Before he gets home” yielding to “You goin up’n getcher hair fixed tuhday,”. (Whether this is a generalizable feature of overlap management remains to be seen.) Gladys: Edna: Gladys: Gladys: Edna: Gladys:

En then you could return it uhb, oh along about noon. (·) You goin up ’n get [cher hair fixed tuhday, ] [Before he gets home. ] (0.5) What deah? (·) Yer goin up t’day en getcher hair fi [xed. [Oh no.

In the following, each party displays Turn Occupancy, with one party’s utterance extending beyond overlap resolution. Perhaps a way to characterize the conditions under which it is decided who should yield is to find that First Stopper has yielded to Last Stopper. (Again, whether this is a generalizable feature of overlap management remains to be seen.) Lil: Lil: Tony: Lil: Tony: Lil:

A:n’ I’m hoping a lo:t,= =·hh thet [you’ll] [do a few a’ them,] [(Yeh)] I [ bet yooer t e r r] ibly hoping, (0.2) t! Hu:h? (0.2) t! I bet yer terribly [hopi [ng. [·hhh [I’m terribly terribly hoping.

In the following two fragments one party displays himself to be Turn Occupant, the other Turn Claimant. In each case Turn Occupant is also Unmarked Competitor and is also First Stopper, yielding to Turn Claimant who is also Marked Competitor and is also Last Stopper. In the first fragment Turn Occupant/Unmarked Competitor/First Stopper produces an utterance plus continuous, extended laughter while Turn Claimant/Marked Competitor/Last Stopper produces a complete question repeated three times over the course of the continuous display.





Gail Jefferson

Mary: Sue: Mary: Sue: Mary: Sue: Mary: Sue: Mary:

[[ I s i d I get mo]re the n h(h)alf= [[Wuh year was it?] =hh(h)unhh heh [hehh hnnnn hnnnn ] [ Wh’t year was i:t? ] (·) u.-hnnn ·huh ·hu[h ·hnnnn!] [Wuh year ] was it? (·) Ha:h? (·) What w-year wa[ s it? ] [Ni:ne] teen fifty.

And in the second fragment Turn Occupant. . . etc produces a single coherent utterance: “Okay duh soopuh. Freak it. He’s a bitch he didn’pud in duh light own dih sekkin flaw, ·hh man tell im. Y’kno:w?” while Turn Claimant... etc. produces revisions of a same question: “Y’ didn’t getta holda-”, “Listen man.”, “Y’ couldn’t gitta hol-”, “Jim wasn’ home uh what.” Vic: Vic: Mike: Vic: Mike: Vic: Mike: Vic: Mike: Vic: Vic: Mike: Vic:

It’s, the attitude of people! (1.0) Okay[ [Y’[didn’t getta holda-] [d u h s o o p u h ] (·) Listen [man. [Freak it. He’s a bitch he didn’ pud in duh light own dih sekkin flaw, ·hh= =Y’ couldn’t gitta ho[l[Man tell im. (·) Jim[wasn’ home] uh what. [ Y’ kno:w? ] (0.5) Hah? (·) Jim wasn’ home [◦ (Whenyih wenovuh deh)◦ ] [I d i d n’ g o b y t h e h. ]

(This particular case is transparent for the constructedness of the competitive continuous coherent utterance, with the bulk of its components starting

Overlap in natural conversation

up after co-competitor has himself gotten started. Further, the issue of who shall have been First Stopper appears to be systematically negotiable, in this case via Appendor Questions; Turn Occupant’s “Y’ kno:w?” followed by Turn Claimant’s “or what.”.)

4. Unmarked other-retrieval: Acknowledgment and embedded repeat These devices appear to have their home in situations of minimal overlap, where competition is not marked or protracted. (a) Acknowledgment. With the post overlap-resolution proffering of an acknowledgment token (“Yeah”, “Uh huh”, etc.), one party treats the other’s overlapped utterance as if it had occurred in the clear; does not recognize the consequence of, or explicitly retrieve the object from, overlap, but simply responds to a prior utterance Bea:

Come to think of it, I think I can manage uh:: otherwise. (·) Francis: Well [ten peopleBea: [So don’t (·) Bea: Un huh, because part of them were going to drink coffee, ................................................... Martha: . . .because she w- you know, was [in the house[ Bea: [so near[Yes. ................................................... Ray: Okay Maggie,= Maggie: =Okay [Love Ray: [See yuh then Yeah.

The acknowledgment token does appear in conjunction with competitive talk, serving as a minimal acknowledgment of a co-competitor’s utterance, but not directed to its subsequent consequence. So, for example, in the following fragment an acknowledgment token turns out to preface a restart i.e., a marked self-retrieval. Helen: Bill: Helen: Bill:

Specially when they [don’ know what they’re doin]= [ N o t b e i n g b- ] =in the [f i r s t ] place.] [Yeh no]t being] bulat’nt.

And in the following, an acknowledgment token turns out to preface a revised a restart.





Gail Jefferson

Olive: Edna: Edna:

I didn’ know what time you were gonna get down, so I wen’ out [shopping yester[day, [Oh::: we got- [Yeah= =We didn’t get down till about- (0.6) Oh I don’t know, six uh’clock . . .

The acknowledgement token as accessory to competition may import some function from its operation as a recognitional in ‘recognition-placed’ overlap (see page 44, point 2, and Jefferson 1973); i.e., proposes that hearingunderstanding of the overlapped talk was achieved, and the talk so far was sufficient for recognition of the projected line, so co-competitor need say no more. (b) Embedded repeat. With this device a party can retrieve another’s overlapped talk by incorporating the other’s possibly-unheard materials into his own next utterance, resulting in an undisrupted flow of talk. For example: Clara: May:

Isn’t that place [something. [I tell yuh. (·) May: It is really something. ................................................... Lottie: How come yih didn’ stay. Oh ih w’zis [too hot huh, Emma: [OH-:: there(·) Emma: Jus’ too hot Lottie, an’ it was. . .

And the combination of acknowledgment and repeat provides for an unequivocal retrieval of another party’s talk. Ken: I-I made pretty good time, but it’s [tiresome. Louise: [But it was one pers-= Louise: =Yeah it’s tiresome. ................................................... Carol: [[HeDenise: [[He’s ser[vin Jody.] Carol: . [ serves? ] Carol: [[ Jo:dy? Denise: [[Yeh! (.) Denise: He serves Jody.

Overlap in natural conversation

(In this case both parties display Turn Occupancy, one with a serially constructed single coherent utterance: “he + ser:ves? + Jo:dy?”, the other with a through-constructed single coherent utterance: “He’s servin Jody.” Note that the party who provides the acknowledgement token thereafter produces the other party’s version of that same statement; i.e., shifts from “He’s servin Jody” to “He serves Jody.”) In the following it appears that each party is attempting to acknowledgmenttoken retrieve the other’s version of a same statement (“drops” vs. “dump”). The matter is resolved with a combined acknowledgment token/repeat. Loren:

Loren: Kate: Loren: Kate: Loren:

. . .people thet jus’ come en 1eave their little kids et th’ skating rink’s appalling. (pause) Y’know[drops them ] [dump them ]= =Yeah. (·) Yeah. (·) Yeah dump. Y’ know. These car pools just pour out five and ten kids.

It should be noted that the other-retrievals, in both marked and unmarked forms, by putting a speaker at the service of another’s talk, simultaneously stand as an offer to delete one’s own utterance; to withdraw it from contention for consequence in the ensuing talk. This opens up such questions as: can the currently withdrawn materials reappear in the conversation, do they have some continuing relevance, can they achieve consequence for subsequent talk, and if so, how? One rather curious fragment will be shown as an indication of the possible continuing relevance of currently withdrawn talk. In this fragment a Last Stopper (Frank) does a repeat request (“Huh?”) and the request is declined by First Stopper (Kitty), who, instead of reproducing her own talk (“...they just rescued-”), retrieves, via a repeat, the overlapped part of Last Stopper’s utterance (“Oh God”). Subsequently the initially withdrawn, now repeat-requestdeclined materials (that “they had rescued a lot of people”) reappear.





Gail Jefferson

Kitty: Kitty: Frank: Frank: Kitty: Frank: Kitty:

Frank:

They said they’ve been, huge wa:ves.= =en they [jus’ rescued[Oh::::: God. Over the pie:r, (·) Huh? (·) Oh Go::d. (0.2) Excep’ tuhday it’s t-calm’ down, I guess ih was a storm out et [sea, [Yeah, they said thet it was gah- it hed gotten nicer but they ed rescued a lotta people, en Newport Beach wz spozetuh have, (0.2) huge waves. (·) Oh they w’ clear over the pier.

The entire sequence is reproduced and reorganized with the “rescue” segment specifically removed from contention for consequence by now being placed prefatory to the segment about “. . .huge waves...”, “. . .over the pier...”. The fact that such a procedure is undertaken in this fragment suggests that in general, overlapped non-retrieved talk is not thereby extinguished, but may remain relevant, to-be-resolved. That such is the case remains to be seen.

VI. Conclusion With a transcript device that captures the precise positioning of overlapping talk it is possible to focus on places in conversation where it appears that a lot of messy talk is going on and people are not much attending each other (i.e., situations of overlapping talk), to find some preliminary indications of intense orderliness and co-attention which can be sketched out. The sketch presented here consisted of an array of conditions, each with orderly procedures attendant to it. (1) Overlap Onset can be the product of systematic procedures, those procedures constituting means of performing specifiable activities; specifically, a party can precision-place his talk in the course of another’s, can select and hit a target point. With such a capability a display of independent knowledge of what is about to be said can be achieved by starting to talk just as some object comes due in an ongoing utterance; a display of recognition of what is in the course of being said can be achieved by starting to talk midway through the recognized object; and it appears that a not infrequently targeted starting

Overlap in natural conversation

point is the moment of completion of an ongoing utterance, this last resulting in overlap should the currently speaking party continue talking beyond that point. (2) Within-Overlap Talk can involve systematic procedures for resolving overlap and/or attempted solutions to a problem which arises as to who should drop out. It can also involve systematic procedures for competing within overlap, negotiating for who shall drop out via, e.g., pronunciational and segmental adjustments. Further, it appears that participants routinely distribute their talk into displays of Turn Occupancy (with a single coherent continuous utterance) and Turn Claimancy (with a repeated recycle of an utterance component). These activities can be examined for their status as Marked and Unmarked competitive forms; i.e., for their explicit attention to, or displayed dismissal of, the fact of overlap and the trouble it might cause for hearing-understanding. (3) Post-Overlap Talk can be investigated for its relationship to the prior overlapping talk, and can involve systematic procedures for retrieving talk potentially not heard due to its occurrence in overlap; i.e., for providing for its consequence in subsequent talk. And these procedures seem to be distributed into types: Self-Retrieval and Other-Retrieval, each type with its marked and unmarked forms; restarts constituting marked self-retrieval, continuations constituting unmarked self-retrieval, repeat-requests constituting marked otherretrieval, and acknowledgment tokens and/or embedded repeats constituting unmarked other-retrieval. These procedures show up in their simplest forms and in more complex versions, further enriching the array of resources available and used for the resolution of overlap and subsequent retrieval of overlapped talk.

References Jefferson, Gail (1973). A Case of Precision Timing in Ordinary Conversation: Overlapped Tag-Positioned Address Terms in Closing Sequences. Semiotica, 9, 47–96. Schegloff, Emanuel A. (1987) [1973]. Recycled Turn Beginnings: A Precise Repair Mechanism in Conversation’s Turn-taking Organization. In G. Button & J. R. E. Lee (Eds.), Talk and Social Organization (pp. 70–85). Clevedon, England: Multilingual Matters. Sacks, Harvey, Schegloff, Emanuel A., & Jefferson, Gail (1974). A Simplest Systematics for the Organization of Turn Taking for Conversation. Language, 50 (4).



P II

Implementing actions

Answering the phone* Emanuel A. Schegloff

It has been suggested1 that the production of an answer to a summons is a selectional issue. It is selectional not only in the interactionally interesting sense, whereby that an answerer can select a problematic answer can serve as a constraint on a prospective summoner to not summon if such an answer is possibly relevant, and an answerer can rely on a summoner having been attentive to such considerations in nonetheless producing his summons; the selectionality remains relevant on the proximate level. Having analyzed an utterance as a summons, a summoned in answering is involved in selecting an answer. The initial selection is between the class of clearance cue answers and the class of problematic answers. It will be useful to review and elaborate some of the resources relevant to this selection in face-to-face settings.2

Answering a summons in face-to-face settings In selecting an answer, one set of resources is the availability to the summoned of the summons as a selected summons. The prospective answerer can see in the actually employed summons utterance a set of practices and selections which produced it as an outcome. On the one hand, in the fact of the production of a summons, he may, as was suggested above, see the outcome of relevant pre-assessments by summoner of the comparative priorities of any ongoing courses of action and the course of action his summons pre-sequences, as well as an assessment of the possible propriety of temporal fit between those courses of action if relevant (e.g., if it is as an insertion interruption that the pre-sequenced activity is to be accomplished). The summoned may thereby be furnished in turn with a sense of the sort of activity that is being pre-sequenced, being entitled to expect via the summoner’s analysis that it has such claims of priority and temporal fit that would warrant its introduction into an ongo-



Emanuel A. Schegloff

ing setting such as both summoner and summoned may inspect. In seeing the fact of the production of a summons as the outcome of such pre-assessment practices by summoner, a prospective answerer is furnished information about some features of the pre-sequenced course of action. Other sorts of information, both about the character of the presequenced course of action and about the summoner (and about the former by virtue of the latter), may be furnished by an analysis of the sort of summons employed, by attending not only that a summons was produced, but the selection that was made to employ the actual item that was produced, and the sub-class of terms from which it was selected. To have selected a summons from the class of “courtesy terms” (i.e., “excuse me,” “pardon me,” etc.), may be heard to display an analysis by summoner that summoned is to him a “stranger.” The use of such a term as a summons is a vehicle for displaying a projected “type” of conversation, one between “unacquainteds.” Since “stranger” is a symmetrical “relational category” (Sacks 1972a), then if summoned is a stranger to summoner, the converse holds as well. To be summoned by such a selected term can then be informative to summoned that summoner is not an “acquainted,” and that the pre-sequenced activity may initially be presumed to be one which unacquainteds are entitled to do,3 that order of identification of the parties being made relevant by the selection of a class of terms whose selection turns on just such an identification. Alternatively, to be summoned by a term from the sub-class “terms of address” may itself be variously informative, as may the selection from among the various sub-classes of that sub-class. For example, to be summoned by name can display at least that order of acquaintedness as “knowing my name,” while the particular name selected may display a claim of relative status, intimacy, solidarity, or membership in some class which entitles use of such a form of address,4 the selection of “Bill” rather than “Mr. Smith” allowing summoned important resources for analysis. Similar resources may be provided by the use of non-name terms of address, for example, occupational titles. That summoned parties will employ the term used to summon them to identify the summoner may then be used by others to “kid around” by employing summons terms which intendedly lead to incorrect inferences, as in the following observation: On Broadway; one policeman walking about 15 yards behind another who is “unaware” of him. Policeman: Officer . . . officer . . help, police . . . help. Bystander smiles; bystander and policeman exchange smiles.

Answering the phone

Summonses, then, can display and make available for summoned parties’ analysis how they were selected, given that that they are selected is something summoned can orient to. They display features relevant to such identifications as “acquainted-stranger,” “relationship to summoned,” etc., by displaying identifications summoner has made of summoned on such dimensions. Such identifications are relevant also to inferences concerning the sort of activity the summons may be pre-sequencing, or they may serve to exclude some domains of action from prospect. Together with the evidence the fact of doing a summons may be taken to give with respect to the claimed priority and temporal fit of the pre-sequenced activity, these can provide one set of resources, interactional resources concerning the relative states of the parties, for a prospective answerer for selecting between the classes “clearance cues” and “problematic answers.” If the initial selection should be the class “problematic answers,” there may be further selection within that class, for example to fit the problematic answer to the setting, as for example in selecting between “Just a minute” and some formulation of ongoing course of action. If the initial selection is the class “clearance cues,” there does not appear to be a selectional issue within the class; there does not appear to be a relevant selection made, for example, between “yeah” and “what.” (There may be settings in which such selection may be specifically required, as when military ranks are established as the relevant identifications, a subordinate may be required to display his recognized subordinate status by answering “Yes, sir.”) While the selection of a clearance cue answer provides the basis for the occurrence of a next action or utterance by summoner, the particular answer employed is not a basis for fitting next utterance, for special selection of sort of utterance, for example, and does not, in that sense, set constraints on what can be done next. A summoned party may, from his analysis of the summons and the setting, decide to clear the summoner to continue with the sort of activity he may take it summoner was pre-sequencing (if he had, indeed, any inference in that regard), but a clearance cue answer in effect clears summoner to proceed without displayed constraint; it does not display what answerer takes it the order of pre-sequenced activity may be, and does not constrain it beyond whatever constraint may have been involved in summoner’s preassessments. It provides, so to speak, a carte blanche.





Emanuel A. Schegloff

Answering the summons on the telephone The features of the opening of telephone conversations are somewhat different.5 The telephone ring, as a form of summons, is not treated by members as displaying selectionality, i.e., as making available to summoned parties the practices whereby it was selected. While members may come to see, in the developing course of the conversation, that a telephone conversation was selected over, e.g., a visit, they do not treat the telephone ring as selected from among a set of ways of doing a summons. The telephone ring stands as a standardized summons. While there are resources upon which one confronted with a ringing phone can base inferences regarding the summoner and the possible character of the pre-sequenced activity, such as the time of day (calls at 3 A.M. to one known not to be normally awake then being inferably high priority), pending business, regular calling practices, “who’s due to call,” etc. (some of which will be elaborated below), displayed selection procedure of the summons is not one of them, and typically they provide at best “good guesses” about what may follow an answer, and are not treated as providing adequate bases for action.6 Similarly, the telephone ring as summons is not treated as evidencing a preassessment by summoner of comparative priorities or temporal fit between pre-sequenced and possibly ongoing courses of action, no such comparative alignment being envisioned (except, again, in cases where it can be taken that summoner must “surely” have known that he was “interrupting,” as with a call at 3 A.M.). There seem to be no interactional resources, then, in analyzing the selection of the summons, and typically in the fact of its production, for deciding between a clearance cue answer and problematic answer (though there may be other bases for not selecting a clearance cue, e.g., ongoing course of action has such priority that it is treated as non-interruptable; in that case, however, the outcome is “no answer”), and although there may be other bases of inference that are relevant, they are typically not treated as adequate grounds for answering the telephone with a problematic answer. Indeed, in the absence of interactionally based grounds, problematic answers are rarely used in telephone conversation.7 In telephone conversation, there does not seem to be a selection issue between clearance cue and problematic answers. The alternative is not, however, a carte blanche clearance cue as the sole possibility. While such a nonconstraining cue as “yeah” is used in telephone conversations as well as in face-to-face settings, it is used in a restricted set of circumstances, and marks the conversation as characterized by those circumstances. In telephone conversation, however, there is a selection possibility within the class of clearance cue

Answering the phone

answers. Aside from “yeah,” telephone summonses may be answered by “hello” or with some self-identification by answerer (there are other possibilities, but these are the central ones). This chapter will be concerned with the selection within the class of clearance cue answers, and with what each selection may be said to accomplish. With regard to the latter issue, it will be suggested that there are answer resources which can introduce some constraints on what is to be done in the following utterances in the conversation. The theme of that discussion may be anticipated here: it is that the answer is fitted not so much to the summons (though it is that too) as to features of the setting in which the answerer is located, and to which summoner is presumed to be oriented in calling.8 Before proceeding to a discussion of the clearance cue answers available on the phone, their selection, and their use, we may note several consequences of the preceding discussion. “Hello” and self-identifications (or selfformulations) are not typically used answers to summonses in unmediated interaction. The class of clearance cue answers is, then, partitioned into bounded sub-classes, relevant to the “telephone/face-to-face” distinction. The ring of the phone is not merely the way one initiates a conversation on the telephone; it types the conversation as a telephone conversation with consequences such as the following: a) no selection issue between clearance cue answers and problematic answers; b) a selection issue within the class clearance cue answers; c) the selection includes possibilities whose use is restricted to the phone, i.e., answers which display, and themselves type, the prospective course of action as a telephone conversation; d) the selection is not limited to the restricted subclass of telephone-specific answers; and within that sub-class there remains a selection issue. Accordingly, the claim may be warranted that the distinction that has hitherto been employed between face-to-face or unmediated interactions and telephonic ones is not an analyst’s constructive distinction or an empirically describable but unused one, but is based on a members’ distinction which has interactional consequences.9

“Yeah” It was suggested above that there are settings in which the non-constraining clearance cue answers “yeah” and “hi,” regularly used in face-to-face interaction, are used also in telephonic interaction. These are settings in which the information available to a summoned party through the displayed selection of a summons in face-to-face interaction is available through other resources.





Emanuel A. Schegloff

They may be employed by answerers when that information may be treated as providing them not with “good guesses” but with reliable bases of inference about the summoner (caller), either who he is (when personal identity is relevant) or what sort he is (when category is relevant). The latter interest may be satisfied in the case of “inter-com” telephones, where the ring of telephone can be taken as evidence, mechanically assured, of the sort of caller, the category of caller, who is calling – i.e., a co-member of the organization. When such analysis is available to the summoned, he may answer “yeah,” as in the following materials selected from the corpus →

→

→

#500 D: C: #391 D: C: #350 D: C:

Yes. Do you want your private ambulance up there? etc. Yeah Listen, I got a hold of Colonel _______ . . . etc. Yes. You’d better put emergency 5 in on that, we’ve got . . . etc.

Or the following transfer of call by intercom, drawn from another organization: L7: S7: L8: S8: →

H9:

No, listen, do you want to talk to Hal, he can probably be more technical. O.K. O.K. Just a minute. Right. (pause) Yeah. (Radio SB)

The former interest, where information is seen to be available not about the sort of summoner who is calling, but about “who it is,” may be satisfied when one party to a conversation in progress proposes to hang up and “call right back.” When, the connection having been broken, the phone rings after such an interval as the summoned may find would have accommodated the proposed intervening activity (e.g., redialing, getting information, making arrangements, etc.), the answerer may answer “yeah” or “hi” (re: “hi” as compared to “hello,” see below, p. 102, n. 14). Thus, in the following data, E has called J to inquire about an event; J has no information on it, but P does:

Answering the phone

E: J: E: J:

Well, call ‘er can’ call me back. O.K. Guhbye Yeah. (Trio, I)

After a three minute 37 second conversation with P, J calls E back: →

E: J:

Yah? Well, she doesn’t know. (Trio, III)

These circumstances, in which answers regularly used in face-to-face interaction are used on the phone as well, are ones in which at least some of the information about summoners available in face-to-face interaction, in part through the displayed selection of the summons, are available in other ways in the telephone ring. Whether by some mechanical attribute of the telephone instrument which may show the call to be intercom, or by the temporal organization of the telephone’s ring with respect to some earlier completed or suspended conversation, the summons is analyzable to find what sort of conversation is thereby being initiated. The crucial feature is the availability of information in the summons, and not the sheer matter of whether the summoning is done by mechanical ring or lexical term. And the relevant information may be analyzed not only in the displayed selectedness of the summons, but in its temporal placement. This discussion has been meant to indicate the dimensions of analysis relevant to a prospective answerer in selecting answers such as “yeah” or “hi.” On the other hand, as the selection of a summons in face-to-face interaction displays for analysis in the summons selected the procedure whereby it was selected, so, given that there is a selection issue for answers,10 the answer selected may display to the caller the method whereby it was selected. The answer selected may then be said to “type” the conversation as intercom, expected, resumed, returned, scheduled, etc., or more generally “foreknown.” (A basis for deciding between some of these possibilities, these varieties of “foreknown,” will be sketched below.) A caller who is answered with a “yeah” or “hi” may hear in that answer that answerer was “expecting” a particular call, and takes it that the present ring is the ring of that call. The answer is selected to “go with,” it is “fitted” to, the type of conversation that is presumptively being initiated, and in being selected for such a fit it begins the process of constituting a conversation of that type. There are other features which may mark this type, i.e., other ways of producing the talk of the conversation that constitute it, or





Emanuel A. Schegloff

partially constitute it, or are consistent with its constitution as an intercom call, a resumed conversation, etc. For example, in the intercom and personal “call-back” calls cited earlier (numbers 500, 391 and 350; Trio III), note that the callers give no self-identification, and that callers’ first utterances contain classic indexical expressions (e.g., “there,” “that”), whose referent is not in “this conversation.” Neither of these features is distinctive to intercom, or resumed, or scheduled, or expected calls; no caller identification is routine in calls to the police; and the use of indexicals without referrent is characteristic of talk even among strangers following, for example, dramatic public events (e.g., disasters are often initially referred to as “up there,” in the aftermath of both Kennedy assassinations, strangers might ask “How is he doing?” without creating a puzzle). But when produced as co-occurrent features together with others, they may constitute a type of conversation. One important feature of this “foreknown-ness” is that the relevance of a presumed shared “agenda” is invoked by such an answer. Whether an agenda of “topics” (as may be the case in a resumed or returned “personal call”) or an agenda of “sorts of topics” (as may be the case in intercom calls), “yeah” or “hi” may invoke a mutual orientation to “what we know we (were) talk(ing) about” without displaying the parameters of that agenda in the answer.11 An orientation to an agenda of this sort is built on, and implies, an identification and formulation of the “we” for whom it is an agenda, either specifically identified co-participants, or sorts of participants (categories of them). Further, invoking the relevance of a preset agenda without explicating its basis is built on, displays, and requires reliance on the “mutually oriented to” character of the agenda, its interactional status. Again then we see that the “carte blanche” form of clearance cue has its use tied to the availability of interactional resources (for example, with regard to the data from Trio I above, the expectations that the “call back” will be the “earliest possible call back,” i.e., a call back “as soon as the other call is completed,” rather than a call back at some indefinite future). This mutually oriented-to agenda is ordered at least with respect to “first topic;” the orientation to a shared agenda is in the first instance an orientation to a shared priority topic. Such a shared priority topic may then supply, and be relied on to supply, the unexplicated referents of indexical expressions or “pro-terms.” The use, and analyzability, of such pro-terms without explicated referents is a way in which parties can show one another their sensitivity to what is “on one another’s minds.” For the achievement of this demonstration, “no explication of referents” is required; it is not so much that members can “manage” without explication, but to show their “common orientation” in this way, it may be necessary to be able to deal with the pro-terms with no expli-

Answering the phone

cated referents. The use of unexplicated pro-terms in a dramatic public event such as a disaster or assassination is thus a way whereby persons, otherwise unacquainted, can show one another a mutually oriented, commonly focused priority topic – “the event” – thereby displaying “what’s on everyone’s mind”; not only everyone displaying what is on his mind, but everyone displaying what they take it is on everybody else’s mind as well. And in being routinely correct in that regard, in the routine interpretability of initial remarks with unexplicated pro-terms, persons may achieve that “sense of community,” “solidarity,” “together across status boundaries,” etc., so commonly remarked upon in the literature and reportage concerning dramatic public events. The same mechanism may be displayed and relied on in less public occasions, as when a husband returning home on the day promotions were to be announced is greeted by his wife with “Well?” In many interactional settings, one party can show attention to another party with regard to “what’s on your mind?” once he knows who “you” are. Thus, in a call from a reporter to a fire department: → →

A: B: A: B: A: B:

City Fire Department, Livingston speaking. This is Carrie Fortune. It was a garbage rack. Oh(hh)kay. Thank//you heh heh heh Ri(hh)ght G’bye. (AFD, I, 16)

With “yeah” and “hi” it is foreknowledge of “who” or “what category” that allows this mechanism to be brought to bear on the selection of an answer to the summons. (Here, knowing “who” or “what category” does not indicate that for any value of “who” or “what category,” as long as it is known, such an answer may be selected. The “who” and the category known in this way are ones for whom an agenda is also known, and for whom a “yeah” or “hi” may be in order.) But in selecting such an answer, the prospective conversation is typed as one in which some pre-arrangement is relevant, either an agreement which is now being realized, or a “naturally” arrived at concordance of interest, mutually oriented to. To say that answering the phone “yeah” or “hi” types the conversation is not, however, to propose that the typing is thereby finalized, or that no further attention need be given to producing the conversation “to type.” In projecting the conversation as a particular type of conversation, some considerations and constraints are introduced to which the co-participants are to be oriented





Emanuel A. Schegloff

in constructing the conversation in its course, considerations and constraints which in the case of any particular conversation require attention to the particulars of that conversation, at that time, with those personnel, that history, etc., in particular, with that shared priority topic. For example, as the data cited earlier in which “yeah” is the selected answer suggest, when answerer has selected “yeah” as the answer, then the “business” of the conversation should be done immediately by the caller, that is, in the very next turn in the conversation. (E.g., “Well, she doesn’t know.”) Where a shared orientation to priority topic has been invoked by the called’s answer, other possible “moves” by caller in the next slot may show that the expectation on which the answerer based his selection of answer is disappointed. When it is recalled that such an utterance in the next slot regularly involves the use of unexplicated pro-terms, we can see that if the caller is not one who has a pre-arrangement, is not a proper category member, or is not co-oriented to a priority topic, he may not be in a position to continue “to type.” It is by the answerer’s invoking a priority co-oriented topic or sort of topic without explicating its basis, and by setting up the relevance of an immediately next utterance consistent with that usage, that a caller may get the sense that he is involved in an “expected” call. Should he be party to the “pre-arrangement” (whether “formally agreed” or “naturally occurring”), he will be in a position to recognize that it is indeed for him that the “hi” or “yeah” is meant, and will have the resources to accomplish a fitted next utterance. Other callers may take it that answerer could not know who or what sort was calling, or that answerer’s warrantable expectation will prove wrong (as when an outside call is transferred over the intercom). By not being a party to any pre-arrangement, not sharing an orientation to a priority topic or sort of topic, and being confronted with the current operation of such pre-arrangements and co-oriented topics, such an answer can make of such a caller, right off, an “outsider.” Answerer, however, is at this point in the conversation as yet unaware of that fact. While a “proper” caller, i.e., one for whom the “yeah” or “hi” answer is on that occasion proper, appropriately in the next slot turns to the priority topic, co-orientation to which is presumptively displayed by the answer, “non-proper caller’s” job in the next turn is to identify themselves (even if they might otherwise not have done so; even if, for example, such a caller might – was entitled to – have sought voice recognition by saying only “hello,” had the initial answer been “hello”); for they are not so much identifying themselves, as showing themselves to be other than the answerer apparently expected, or could have been expected to expect. So, for example, in the conversation re-

Answering the phone

ported as Radio SB above, while the call is received by “H” as an intercom, it is not a co-member of the organization on the line. The sequel then is:

→

H: S: H: S: H:

Yeah. Hi, Hal? Yeah, what can I do for ya? Marty Anderson here. Yeah.

And answerers who discover that they are not talking to the one they expected, and hearing that their answer may well have made of the actual caller an “outsider to an arrangement,” may undertake to repair matters by accounting for their initial answer and citing their expectations. We may summarize this discussion as follows. The telephone ring types a forthcoming conversation as a telephone conversation (both caller and called attend the rings – and attend each others’ orientation to the rings – through the use of “number of rings” as a measurement system for the temporal relation of answer to summons, used as a basis for inferences regarding “absence” or eagerness [cf. the subsequently published Schegloff 1986]), establishing thereby the relevance of a selection problem for clearance cue answers. Answers such as “yeah,” “hi,” etc., are, then, produced and heard to be produced as the outcome of a selection procedure. Their selection presumptively types the prospective conversation as “foreknown,” as one in which the answerer takes it he has warrantable information about the caller and the prospective course of action. Such answers type the prospective conversation “presumptively” because there may not be a convergence between caller and called on the type, and caller may undertake to correct what he sees as the answerer’s “mistake,” thereby possibly transforming the type of the call in his first utterance.12 “Yeah” and “Hi” are, however, marked forms, and the settings in which they are selected, in which answerers treat their “information” in regard to caller as warranting such a selection, are not typical. How do answerers proceed when the ring of the phone does not furnish them information adequate to the selection of such an answer, when they may not feel they have “reliable” information on the caller and therefore on the presequenced course of action, and when therefore a carte blanche answer such as is used in face-to-face interaction may not be appropriate? And what are the interactional consequences of alternative selections? In the following pages we examine two: “hello” and some form of self-identification.





Emanuel A. Schegloff

“Hello” A first question to be addressed with respect to “hello” as an initial utterance in telephone conversation is whether it is an answer, whether it is as an answer to a summons that it is to be analyzed. The question arises in the following way. We noted earlier that “hello” has as one of its prominent uses that of “greetings.” It is further clearly the case that not any use of “hello” is a greeting, or a possible greeting. When a conversation is temporarily adjourned, as when one co-participant goes to check something, the resumption may be marked by an exchange of “hellos”. There, whether they are treated as availability signs, or as a summons-answer exchange, the “hellos” are not used to do greetings. Whether “hello” is a greeting, or a possible greeting, seems to turn on a combination of the lexeme and its placement in the conversation. Sacks (1975) has proposed that an utterance from the class that can be greetings (of which “hello” is a member), when placed at the beginning of a conversation (e.g., in the first slot, or in the first exchange; the argument may hold for slots further in the conversation, but should hold at least for “first slot”), constitutes a greeting. The production of an utterance as a greeting, its analysis as a greeting by coparticipant, or a search for greetings (e.g., a finding that one was not greeted) turn on a combination of a “place” in the conversation where greeting would occur were it to occur, and a class of utterances, a member of which is placed in that “place.” Greeting term in first position, then, specifies a “greeting”; in this analysis, the initial “hello” in a telephone conversation is not, or is not in the first place, or is not only, an answer to a summons, but a “greeting.” There are (or would be) describable virtues to using a greeting as the initial utterance in a telephone conversation. As has been noted, greetings come in pairs; they are properly organized as utterance pairs, the occurrence of the first making a second (or return) conditionally relevant. In doing a greeting as his initial utterance, a called party might thereby set constraints on what a caller could properly do in his first utterance, making at least a greeting-return the first-order relevant action for caller. For a called party who, to this point, had available as resources only the ring of the phone, some basis might thereby be made available for some identification of the caller and the order of activity that might be forthcoming, by affording the called at least an opportunity for voice recognition. Several considerations make this analysis of the initial “hello” in some telephone conversations unsatisfactory, unless modified. First, in conversations in which the initial “hello” is not answered with a return greeting, a return greeting may nonetheless not be found to be missing.

Answering the phone

For example: → →

A: B: A:

Hello/ Marty Rosenthal calling collect, will you accept the charge/ Yes, certainly, operator. (CF, p. 22)

→ →

A: B: A: B: A: B:

Hello Are you awake/ Yeah I // dis got up I – Oh didjuh/ Yeah hh Weh gooud. I’m alone. (NB: IV:3, p. 1)

→ →

A: B: A: B: A: B: A:

Hello Is Jessie there/ (No) Jessie’s over et ’er gramma’s fer a couple days. A’right thankyou Yer welcome/ Bye Dianne/ (NB: 9/10/68; c. 1, p. 1)

Secondly, one feature of greetings and greeting exchanges appears to be “one per party per occasion, if reciprocated.”13 Although after a greeting exchange parties may employ what Sacks (1975) has termed “greeting substitutes” (such as “How are you?”), they should not continue with additional greeting per se. Yet there are substantial materials available in which after the initial “hello,” a caller “identifies himself,” is “recognized” by answerer, and/or is found by answerer to be one with whom he is in “informal” relations, or a “friend,” and gets from answerer the greeting form that appears to be selected by reference to showing recognition, or being consistent with “informal” or “friendly” relations, i.e., “hi.”14 For example:

→

A: B: A:

Hello. Marty, Al. Oh, hi. (CF, p. 24)

A: B: A:

Hello/ Martin/ Yeah/





Emanuel A. Schegloff

→

B: A:

Hi, this is Sophia Oh, hi. (CF, p. 25)

→

A: B: A: B: A:

Hello Eddy Yeh Guy Huston. Hi Guy. (NB, I, l, p. l)

If the initial “hello” is analyzed straightforwardly as a greeting, then, given that return greetings are conditionally relevant, the non-occurrence of a return should be its absence, notable, actable upon, etc., which does not appear to be the case. Since the pair organization of greetings and their non-repeatability seem amply supported by a wide variety of data, it does not appear that the initial “hello” of some telephone conversations is treated by co-participants, or is to be treated by analysts, as straightforwardly a greeting. The grounds for treating the initial “hello” as an answer to a summons have been developed in Schegloff (1968); the features of greetings do not provide for preferring a “greeting analysis” to an “answer analysis.” The initial “hello,” then, is treated as, and is to be treated as, in the first place an answer to the telephone ring’s summons. Although it is not in the first place a greeting, there is some empirical basis for finding the relevance of greetings in the analysis of the initial “hello.” On the one hand, caller’s first utterance may begin with a greeting term which then seems to operate as the second part of a pair, as in: → →

A: B: A: B: A:

Hello/ Hello. I’m trying to locate Professor ___________, is he there/ No, he’s not. Thank you Mmm (EAS, FN)

On the other hand, while the occurrence of the answerer’s initial “hello” does not entail any subsequent greeting by the answerer, it is also the case that caller’s first utterance may be a greeting alone, that answerer may not answer the greeting, and no greeting return be found absent, as in the following:

Answering the phone

→ →

A: B: A: B:

Hello. ‘Allo. G’you–your roomate talks forever¿ or is it you? The roomie . . . (etc.) (CF, p. 8)

In such an occurrence, and in ones likes it, it appears that the sequence may emerge non-violative because the “possible greeting” status of the initial “hello” is exploited. The foregoing arguments do not entail that the initial “hello” cannot be a greeting, only that it need not be, and that when used as the initial utterance in a telephone conversation, greeting need not be the firstorder activity that the utterance is accomplishing. Above all, the. initial “hello” is an answer to a summons. Whatever else it accomplishes, it accomplishes (unless specially modified) the completion of an SA sequence [summons-answer sequence] and the establishment of availability; that it accomplishes at least that for co-participants can be seen in the invariable immediate relevance of further talk (and the near invariable occurrence of further talk; when nonoccurring, it is relevantly absent15 ) in closely paced order directly after the utterance. In the use of a possible greeting term to accomplish the answer, and by virtue of the slot for an answer being, in telephone conversation a slot in which a greeting term should be placed to do greeting, the utterance has a possible analysis in addition to that of “answer” (a possible analysis, that is, both for co-participants and for analysts), and that is “possible greeting.” By “possible greeting” is intended that whether it is a greeting or not, whether it accomplishes greeting or not, can turn on what follows it, on whether co-participants convert, or can be seen to have converted, its status as a “possible greeting” into “no greeting” or “actual greeting.” Thus, if it is followed by a caller’s greeting and no further greeting by called, then the latter will not be seen as absent, the possible greeting status of the initial “hello” serving as called’s greeting (as in the data from “CF, p. 8” cited above). If, however, the sequence develops with a subsequent possible greeting by called, then the first may be seen not to have been a greeting, and the non-repeatability constraint is not violated (as in the data cited as “CF, p. 24,” “CF, p. 25,” and “NB, I, l, p. l,” above). We may also then expect that when the phone is answered “hi” in the usage described in a previous section of this chapter, then the non-occurrence of a greeting from caller in the next slot may not involve its absence, the “hi” serving as an answer to the summons of a particular type, and (though it is a greeting term in first position) not as a greeting.





Emanuel A. Schegloff

If the status of “hello” is as a “possible greeting,” then the “constraints” on a next slot suggested earlier as the possible virtue of using a greeting as an initial utterance are very weak constraints indeed, if they are constraints at all. For while they make a return greeting “possibly relevant,” should there be no return greeting it need not be found absent. Are there then no constraints or no specially relevant actions to be done in the caller’s first turn, which “hello” may be seen to occasion? There does seem to be considerable orderliness to what is done in the turn(s) following “hello,” and conversation seems to be organized to provide for that orderliness (rather than, for example, it being an order which no special organization is designed to achieve; on the contrast, see Sacks 1992, passim). However, as compared to the hypothetical orderliness involved in the “greeting” constraints which are tied to the utterance-to-utterance order of organization, the orderliness to be discussed below – involving the relevance of reasons or “identification work” after the initial “hello” – is to be understood by reference to the overall structure of conversation. It is not, then, that the initial “hello” makes “identification work” relevant, but that the initial “hello” may be the occasion for the relevance of “identification work.” It is a feature of the overall structural organization of conversations that identification work or delimited alternative activities are relevant at their beginnings; “hello” may occasion that work, or its relevance, by establishing the availability of the parties, thus meeting a critical condition for beginning. Once the co-participants can proceed, they proceed with the relevant activities, e.g., identification, but those activities are made relevant not by “hello” (as a return greeting might be said to be made relevant by “hello”), but by reference to overall structural considerations. To understand the kind of answer “hello” is on the telephone, the work it does, and the work it occasions, we need at least to sketch some features of overall conversational organization which its use seems to invoke. In describing the overall structural features relevant to the present discussion, we will need to digress a bit in an attempt to specify some features of a class of conversations of which telephone conversations are members, to set the context which makes relevant the doing of identification work.

Designed and by-product conversations The present discussion is intended to give some depth and perspective to several points, which can be supported on quite other bases. They are that in telephone conversations (and the relevance of that as a type for members has been established above) identification work and/or reasons for the contact are

Answering the phone

relevant actions at the very beginning, i.e., directly following the answer. Empirically, one and/or the other are regularly and massively found there. That they are relevant there will be suggested below by showing that the method whereby they are sometimes accomplished turns on their being relevant (see discussion of voice recognition below). The present discussion is in the interests of showing that this relevance is to be understood by reference to certain overall organizational features, and that these are not overall organizational features specific to telephone conversation per se, but organizational features of a type of conversation – what I am here terming “designed” conversation – of which all telephone conversations are treated by members as instances, but of which some face-to-face conversations are treated as instances as well. The discussion will, therefore, begin by limning the sense of the distinct types in face-to-face interaction, although the lack of specific materials and analyses precludes systematic description here. An initial sense of the distinction between designed and by-product conversations or encounters may be provided by the following examples. One person passing another may exchange greetings alone with him, or may do so in passing his office if his door happens to be open (see Goffman (1953: 159– 161) for examples from an isolated rural community). Such minimal exchanges do not characterize occasions on which one has entered the office of another, or “approached” him, nor do we find telephone conversations which consist of no more than an exchange of greetings. When persons “accidentally” encounter one another, it appears, they may be at liberty to confine their remarks to an exchange of greetings (Goffman 1953: 485 points out that length of salutation may depend “on the period that had elapsed since the last salutation and the period that seemed likely before the next”; but a minimal exchange is possible); when there is a “planned” or “intended” encounter, more than a minimal pair is done. The notion of “accidental” and “intentional” encounters implicitly presumes that the encounterers are acquainted; non-acquaintances do not “accidentally” encounter each other, they do not encounter at all. “Accidental” and “intentional” are the acquainted’s version of “by-product” and “designed.” That members may treat as relevant whether an action has been a produced designedly or as by-product, as a general distinction, can be seen as well in Sacks’ observation (1992a: 792–793) that there are important differences in the treatments accorded “interactionally generated” invitations and what night be called “designed” invitations. When an invitation is extended by a “called” party, it can be seen that the conversation was not designed for the invitation’s achievement, the conversation having been initiated by the other; such an invitation may be seen to be interactionally generated, while caller’s invitations may





Emanuel A. Schegloff

(though they need not invariably) be seen as designed. While interactionally generated invitations may be treated by offering “counter-invitation” (“why don’t you come over here?”), cautioning against elaborate preparation, etc., designed invitations do not seem to be properly treated that way. So, members can orient to the designed versus by-product features of an action. The argument here is that conversation as a whole may be similarly analyzed, or more particularly, conversation beginnings may be, and different relevancies be associated with different findings (The point about specifying the discussion to conversational beginnings is that once initiated, transformations of conversational type are possible, so that what may have been initiated as one type may be transformed into another. That is, there are ways of transforming many sorts of conversational “type.” But unless transformed, an initiated “type” holds; cf. the discussion of “expectable monotopicality” below, at pp. 104–105, n. 23.) One method by which the “designed” or “by-product” character of a prospective conversation may be displayed (by possible initiator) and analyzed (by possible recipient) turns on the temporal relation between certain features of the setting, e.g., the achievement and acknowledgment of co-presence, and the attempted initiation of conversation. Conversation initiation may be accomplished as “by-product” when persons find each other to be co-present, do co-presence acknowledgments (e.g., see Goffman 1963a: 83–88 on signs of civil inattention), and do not in closely paced order16 undertake conversation. Co-presence acknowledgements may then serve as a temporal marker; if it does not directly occasion an attempted initiation, the co-presence will not be seen to have been designed and achieved in the interests of conversation. If conversation is initiated in close order, then the co-presence may be seen to have been designed and achieved to allow it. Where the finding and acknowledgment of co-presence is not directly followed by attempted conversation initiation, where co-presence is seen as by-product and not designed, for example when two persons come to be standing at a bus stop, sitting in a waiting room, occupying adjacent seats on an airplane, be juxtaposed in a queue, etc., they may accomplish and be seen to accomplish the establishment of a “base” upon arrival, coming to a stop, sitting down, setting down portable belongings, etc., establishing that their presence there is designed for locally relevant activities, and not for the co-presence with the other it accomplishes as a by-product. Such “establishing a base” may involve as well adjustments in pace of walking on arrival, respect for micro-ecological space boundaries (see Sommers 1959, 1969; Hall 1959; Goffman 1963), body and face positioning to avoid direct confrontation (Scheflen 1963, 1964), eye aversion (Kendon 1967), neutral facial expression (Birdwhistell 1970), etc.

Answering the phone

By contrast, a different configuration of deportment may be produced and seen as “doing preliminaries” to conversation, as “approaching someone” or “being approached.” When one walks a path aimed at another, does not vary or slacken pace as distance diminishes while giving no signs of veering to avoid collision, facing directly at other, incipiently smiling, crossing micro-ecological boundaries and entering conversational range, positioning body vis-a-vis, then that preliminaries to starting up a conversation are being accomplished may be available to the one whose path is thus occupied, who is thus confronted, etc. Persons may sometimes see preliminaries being done in their direction, and “brace themselves” for the initiation of conversation, only to find that the preliminaries were being done “to” one directly behind them in a same path with the initiator. But where preliminaries are seen, the conversation that is then initiated may be analyzed as a designed, rather than a byproduct, conversation. (To be sure, it may be seen that another has, by an orientation to these ways of seeing designed and by-product conversations, designed one to seem a by-product, as when women see “chance” and “casual” encounters “engineered” by males; but the rights to see by-product conversations as designedly engineered may be limited to those for whom there may be reason for such engineering, as in the courtship case, or persons of high prestige or reputation. Others do so at the risk of being seen as paranoid.) While the above discussion is by no means an adequately detailed, systematic, or methodic account of the “doing” and “analyzing” of conversation initiations as “designed” or “by-product,” it may suffice for the present occasion to suggest the types involved and some basis for seeing that they are relevantly discriminated by members. (Where the parties are entitled to mutual recognition as acquaintances, then the discrimination may be formulated, as noted above, as “accidental” versus “intentional,” the issue being whether their “co-presence” is designed or chance.) The discrimination is in point, because, as types, “designed” and “by-product” conversations have different overall structural organizations. The discussion here will largely confine itself to “designed” conversations, for telephone conversation seems to be invariably treated as “designed,” so it is the overall structural features of “designed” conversation that will be relevant here. When a conversation is initiated and seen to be initiated as a designed conversation, a relevant matter attended to as a task for the conversation at its beginning is establishing adequate grounds for its undertaking, establishing its entitlement.17 One basis for its entitlement may be the announcement of a legitimate reason for initiating the conversation. While I cannot here provide an adequate account of what constitutes a “legitimate” reason, that being in any





Emanuel A. Schegloff

case a matter that may be contested by the co-participants, several sorts of reasons which may claim a sort of prima facie legitimacy have been suggested by others; for example, Goffman’s (1963: 128) observations that adequate warrant may be claimed by using the interests of the other as the grounds for starting the conversation (e.g., “You dropped your wallet”; “Your purse is open”), by citing “free needs” (e.g., coin change, the time, a match, directions, etc. ibid., 130), and “ritually impaired” objects which may be fair topic for anyone’s comment (e.g., children and dogs, ibid., 126). Whatever the features that constitute legitimate grounds, initiators of designed conversations regularly provide them in the beginnings of conversational openings (i.e., as their first utterance, or as their second utterance, if their first was a summons). Alternative adequate grounds for starting up a conversation (I will address below the issue of whether the following should stand as alternative to reasons, or vice versa) is reciprocal entitled identification of the parties as “acquainted.” Parties who may relevantly identify each other as acquaintances may need no reason to serve as adequate grounds for starting a conversation (they are not precluded from offering reasons; however, by virtue of not needing them, having them assumes a different status, as will be discussed below). In face-to-face interaction, of course, acquaintanceship may be established by visual inspection. To say that it may be is not, however, to provide a basis for its occurrence, or its relevance. The clarification of the issue of the relevance of identifications by parties of one another as “acquainted,” and the visual-inspection recognition whereby it is achieved, will require a digression, one whose resources will be used not only for the present discussion but for subsequent issues as well.18 The relevance of “acquainted” is problematic because it is one of an indefinitely large collection of identifications, or identification types, that could be made of any parties. Sacks has shown that for the identification of a member, there is at least more than one identification term (or categorization), drawn from a collection of terms (as the term “male” is drawn from the collection “sex”; or the term “plumber” is drawn from the collection “occupations” which also includes “doctor”, “lawyer,” etc.), which is “correct” for a member. That is, there are at least two collections (i.e., age and sex) which have terms one of which will be correct for any member of any unspecified population (in fact there are many more than two that have this property, and an indefinite number which will have correct terms if there is prior specification of the population to be identified). As a consequence, identifications are not adequately warranted by their “correctness” (in some correspondence sense of correctness, whereby “male” is correct if the object so identified is male), for alternate iden-

Answering the phone

tifications would be correct as well. In proposing an identification, then, some procedure whereby that identification term, from that collection of terms, is found relevant is required. By reference to this argument, it is not enough that two members are “acquainted” for us to so identify them, or for them to be asserted to so identify themselves; some procedure whereby the relevance of that identification, and the collection from which it is drawn, the collection of paired-relationship terms (called “R” by Sacks), is established is required. While it has been asserted above that for the receiver of a conversation initiation, or indeed one who sees preliminaries being done, the collection “acquainted-unacquainted” is made relevant, unacquainteds requiring reasons and acquainteds not, on the one hand it is not clearly the case that on any occasion of being addressed the collection “acquainted-unacquainted” is of first-order relevance, and on the other no provision is thereby made for the relevance to a prospective first speaker of “acquainted-unacquainted” as the possibly adequate grounds for his starting a conversation. We may, therefore, outline some considerations which warrant a general first-order relevance of the collection “acquainted-unacquainted,” and the identification “acquainted” in particular. While greetings may be appropriate between members formulated by a wide variety of identifications, greeting exchanges seen to be obligated, and minimal greeting exchanges allowed, between encountering acquainteds. A greeting exchange may serve as an acknowledgment of recognition, and as acknowledgment of co-presence, for acquainteds. If acquainteds see each other to have seen each other, they ought to acknowledge reciprocal recognition and acquaintanceship. While modified by specifications concerning entitlement to reciprocal recognition, status orderings on who greets first, and the like, members regularly have others to whom they will owe a greeting, indeed a first greeting, if co-present. The relevance of this rule can be seen in that failure to do an initial greeting when the rule is relevant may entail an official absence. While second greetings, or return greetings may be found absent when non-occurring by virtue of the pair organization of greetings, and the conditional relevance of a second on the occurrence of the first, such a rule does not provide for the official absence of first greetings; yet first greetings may be found absent.19 They may warrant remarks or inferences such as “He didn’t even say ‘hello,’” “He (you) didn’t (don’t) recognize me”), or, if those are not supportable, “He’s angry,” etc. Members can then be “held responsible” for not doing first greetings, by reference to an obligation they may have to acknowledge recognition and co-presence with those entitled to recognition as acquaintances. Such an obligation, and





Emanuel A. Schegloff

an orientation to meeting it, seem to entail for members monitoring the environment for entitled acquainteds, whose co-presence, or mutually ratified co-presence, might occasion the relevance of a greeting.20 The relevance of the collection “acquainted-unacquainted,” or more correctly of the identification “acquainted,” may be provided by the possible obligations that recognition may entail. (The monitoring may have no interest in the identification “unacquainted.”21 It may therefore be more correct to see the initial adequate grounds for starting a conversation to be the identification “acquainted,” reasons serving as an alternative if that identification cannot be established. This would reverse the order of alternatives suggested earlier. I see no basis now, however, for preferring either version). If the foregoing sketch may be taken to provide some warrant for the relevance of identification as “acquainteds” as a first-order relevant identification, then a basis may have been provided for its availability as possible adequate grounds for starting up a conversation, the point which prompted this excursus in the first place. Acquaintanceship recognition being available to visual inspection in face-to-face interaction, a basis for accomplishing it is thereby provided.22 There is an additional basis for seeing the relevance of the category “acquainted” as a first-order identification, i.e., if a conversation is to be started with a summons, then one selection issue for summonses turns on the acquainted-stranger alternation, “excuse me” being a term selected for unacquainteds. While I have urged that “acquainted” is an identification of first-order relevance to the beginning of conversation (although it should be recalled that discussion is intendedly limited to designed conversation), other identifications may be relevant at the beginning. “Acquainted” as an identification is special in that it may serve as an alternative to reasons, reasons not being needed if acquaintanceship is established. But identifications may be relevant and offered even though they do not lift the requirement of reasons. In particular, identifications may be relevant as orientations to reasons, and as possible modifiers on the sorts of reasons that will serve as legitimate grounds for the conversation. That is, we earlier reported observations on sorts of reasons which non-formulated members, “anybodies,“ might offer as legitimate bases for conversation. Some reasons not legitimate for unidentified initiators of conversation might be legitimate for formulated members, initiators identified in particular ways. Prospective reasons may, therefore, make identification (i.e., self-identification by initiator) relevant.

Answering the phone

Similarly, while reciprocally recognized acquaintanceship in itself may serve as adequate grounds for starting up a designed conversation (and therefore, for achieving co-presence), acquaintances may have reasons too. However, not having to have a reason makes having a reason a different sort of phenomenon. When acquaintances (where acquaintanceship is established as relevant) offer reasons, the reasons are not for starting up a conversation, but for starting a conversation “now.” Where acquaintanceship serves as adequate grounds, however, so that no reason is required, having a reason on each occasion of starting a conversation (calling, dropping in, going over, etc.) may dim the relevance or adequacy of the acquaintanceship per se as a basis for conversation. As Sacks observed (1992b: 163–166), persons may hesitate to call with a reason if they find they have recently called only with reasons, and may make a point of calling for no reason. In announcing that they are “dropping in to say hello,” “calling to find out how you are,” or noting that “I haven’t seen you lately,” initiators of conversations may show that they not only do not have a reason for starting up (which as acquainteds they do not need) but that they don’t have a reason for calling “now.” They thereby show that while any conversation might be said to occur in some “now,” it is not by reference to any particular “now” that they are starting a conversation. Indeed, the character of some members’ acquaintanceship may be defined by the regular “no reason” initiation of conversations, where that there is no reason, that there is no “now” for the conversation, need no longer be asserted, as with persons who “talk every day.” To start a conversation, to design a conversation, with no reason, and with no reason for doing it “now,” may thus display that one is thinking of the other without occasioned reasons or interests in doing so, as a matter of pure acquaintanceship. The question, then, is not: is identification relevant and/or is a reason relevant, but, by virtue of an identification, a reason may or may not be; by virtue of a reason, an identification may or may not be. Accomplishing one or another or both, however, are relevant actions at the beginnings of designed conversations as a feature of their overall structural organization. There are, however, independent grounds for identification if its outcome in the collection “acquainted-unacquainted” is “acquainted,” i.e., acquaintanceship should be recognized and acknowledged, if possible. There are, of course, conversation starts between unacquainteds that do not display reasons in the beginning of the conversation. By-product conversations may not have reasons in the beginning, and although our main interest is in designed conversations, it may be in point here to note some features





Emanuel A. Schegloff

of by-product conversations. They may have reasons offered at the beginning when started by unacquainteds, but if they do not, they will regularly display in their beginnings, in the first or second utterance by initiator, a legitimate basis for the conversation’s start (again, once initiated, the conversations may be transformed, so the basis for its start may not constrain its development over its course). One feature of by-product conversation, it may be recalled, is that co-presence is not seen as having been designed and achieved in the interests of the conversation, but rather, the parties pursuing their own courses of action, co-presence is a by-product. Such occasions of co-presence are frequently ones that have elsewhere been characterized as ones in which a continuing state of incipient talk may exist (Schegloff & Sacks 1973). One feature of such scenes may be that their temporal boundary is fixed independent of the internal development of conversations that may be ongoing. Occupants of adjacent seats on an airplane, those waiting in waiting rooms, at bus-stops, for elevators, at ski lifts, etc., are in such settings. When by-product conversations are started in such settings, initial utterances regularly display attention to the comembership of initiator and target in the setting in which the conversation is started (in the case of airplane seatmates, for example, by utterances invoking the relevance of the plane, the city left, the city of destination, the stewardess, the weather, etc.). By invoking or displaying attention to features of the setting in which the parties are co-members, or searching out features of the setting by reference to which initiator and target are co-members (Goffman 1963a: 133), a basis for the conversation is claimed, what may be termed “acquaintanceship substitutes.” But such bases for conversation do not seem to be invoked for designed conversations, as warranting their beginnings (again, once initiated, such acquaintanceship substitutes may be invoked as grounds for transforming the conversation). I have urged that members orient to a distinction in producing and analyzing conversation initiations between designed and by-product conversation; that designed conversation has as a feature of its overall structural organization the relevance right off at the beginning of the conversation of adequate grounds for its conduct;23 that adequate grounds may be provided by reciprocally recognized acquaintanceship, or by legitimate reasons (where legitimacy may be related to identifications of the parties other than in terms of acquaintanceship); and that therefore, in designed conversations, establishing acquaintanceship or reasons will be a relevant bit of work at the very beginning. The relevance of this discussion to the matter at hand is that telephone conversation appears invariably to be treated as designed conversation. In having to organize a course of action (e.g., most obviously, in dialing) to achieve its initiation, and in be-

Answering the phone

ing seen to have done so, the caller is seen to have designed the conversation’s occurrence. That is available to the called in the ring of the phone. For telephone conversation, then, it appears that the overall structural organization of designed conversation is relevant, and that directly at its beginning the establishment of acquaintanceship or the offering of reasons is the relevant matter. “Hello” as an answer may thus be seen to occasion one or the other, and the former in preference to the latter if possible. That acquaintanceship identification, or as I shall refer to it, “identification work” (to take note of the fact that “identification work” may be done though no identifications are asked for or offered, as in the case of voice identification to be discussed shortly) is relevant by reference to overall structural organization seems required to understand routinely produced sequences directly after “hello,” for which “hello’ does not provide any ready basis of understanding on an utterance-to-utterance level of organization. While the answerer’s initial “hello” does not prospectively invoke identifications or reasons by way of utterance-to-utterance organization, but rather occasions their relevance through the overall structural organization of designed conversation, a caller doing identification work or reasons must accomplish them so as to display attention to the utterance-to-utterance level of organization. The actions made relevant by overall structural considerations are produced with respect for, and with attention to, orderliness at other levels of organization. The utterances produced by caller to do identification work or reasons are thus “fitted to” the initial utterance which they follow; they are produced to follow the answerer’s initial utterance. And they are produced by an orientation to whatever “type” the initial utterance may have projected for the conversation, either to be consistent with it and so continue to constitute the conversation as one of that type, or to transform the type (as caller’s first utterance after “yeah” or “hi” may serve to further constitute a “fore-known,” shared-priority-but-unexplicated-topic type conversation, or to transform it). For each of the possibly relevant actions for caller in his first turn, there is a variety of techniques for accomplishing it. For example, identification work can be accomplished by voice recognition or by “doing identification,” and each of those has ways in which it can be accomplished. But since selection among those techniques in part turns on considerations of “fit” to the initial answer, it will be best to defer a discussion of such sequelae until we have completed an examination of the range of answer resources, what they seem to accomplish and how they are selected, to which caller’s first utterance is fitted (as utterance and as type). In doing so, we will have occasion to examine some interactional resources employed by answerer to select an answer (not to select





Emanuel A. Schegloff

between clearance cue answers and problematic answers, for, as was suggested above, the interactional resources for that selection appear unavailable in telephone conversation beginnings; but to select between possible clearance cue answers), and whose use for the selection is displayed in the answer, and is employed by caller in producing his initial utterance. We will then be in a position to return to what happens after the initial answer, and to work through the accomplishment of the identification work or reasons which the initial “hello” may occasion. Here we can note that in occasioning the relevance of identification work and/or reasons as next actions, the initial “hello” does introduce constraints on the turn immediately following it. However, in setting these constraints, it is left to the caller to select a reason or an identification to introduce, and to choose the manner of its introduction. No constraints are explicated in the initial answer which would set boundaries on a class of legitimate reasons or identifications. “Hello” provides a license to talk to callers identified by a wide range of identifications and with a wide variety of reasons, as long as they properly provide their identifications or reasons at the beginning. This is well fitted to the use of “hello” as the typical answering form at “home phones” or “personal phones,” where a wide variety of callers may properly call with a wide variety of prospective conversations, there being no a priori topical restrictions relevant. To say that “hello” is the “typical” form used at “home” or “personal” phones is to say that this typicality is produced by answerers’ use of their setting as one basis for selecting an answer. The selection procedures involved will be discussed in the last section of this chapter. It is also, however, to suggest that “hello” “types” the conversation it begins as a “personal” conversation, and this is an incorrect suggestion. “Hello” is the unmarked form of answer to the telephone; whereas “yeah” or “hi” may type a prospective conversation as “expected,” and a self-identification form of answer, such as “Police Desk” may type it as “business” (as we shall see below), “Hello” may show that it is not “specifically expected,” and is not “specifically business,” but is not selected specifically to show those negative features. While not selected for those features, it may nonetheless come to be seen that it is one of those features that is most consequential on some occasion of use. Thus: →

A: B: A: B: A:

Hello “Hello::?” Yeah. “Hello.” Wuh–is this 293-4673 No, it’s 293-4637..

Answering the phone

B:

→

A: B:

Oh, I’m awfully sorry. ... Am I supposed to be a business firm? Yes. That’s right. That’s exactly right. I’m calling my office. They never answer with “hello::.” (GZ, p. 27)

“Hello,” then, in allowing “possible business” and “possible personal” conversations, may type the conversation only as “non-typed,” as not in advance setting priority or exclusively relevant topics, conversation types, or legitimate identifications for callers. While there are constraints on callers to provide either reasons or identifications, no constraints are introduced by the “hello”form answer on the accomplishment of those actions (though constraints may otherwise be operative, e.g., those provided by the way in which “hello” is done). In this respect, self-identification forms of answer may differ.

“Police Desk” A first observation about self-identifications as answers to telephone summonses, as the answerer’s initial utterance, is that they are preemptive identifications. Identifications, it was noted earlier, are made relevant in designed conversations as one way of possibly establishing the entitlement for starting up a conversation. Accordingly, the work of identification (or of giving reasons) was the initiator’s work – in the case of telephone conversation, the caller’s work, for it is his entitlement to have started the conversation that may be at issue. The locus of that work, of the relevance of identifications and/or reasons, was thus initiator’s (caller’s) first utterance after the possible start of a conversation has been established; where an initiator’s first utterance was a summons, then it might be in his second utterance that identifications and/or reasons would be relevant. On the telephone, since the summons is not accomplished by an utterance, it is in fact in caller’s first utterance that that work is relevant. Answering the telephone with a self-identification is preemptive because it does the work of identification before the turn-taking organization has provided caller his first opportunity for doing so. At caller’s first turn, he is not then able to do an initial identification; any identification done there will stand in some relationship to the preemptive identification already accomplished in the answer.





Emanuel A. Schegloff

An answer such as “Police Desk” accomplishes an identification of a particular kind, and thereby makes the identification issue the provision of identifications of a particular kind. Of the identifications that may stand as adequate grounds for starting up a conversation, one collection was noted above to be specially relevant, i.e., acquainted-unacquainted. Identification in terms of this collection, in particular identification as “acquainted,” involves a personal identification; together with many other kinds of identification that may be offered as part of the work of warranting a conversation, it identifies a particular initiator or caller, it addresses the issue “who’s calling,” and may, for example, use a personal name to accomplish the identification. Self-identification answers of the form of “Police Desk” make the relevant sort of identification not personal but categorical. Identificational interest is then not in “who in particular is calling” but in the “sort of caller.” After such an answer and the type of identification it makes relevant (and the type of conversation thereby initiated), no personal self-identification by caller need be given or relevant (e.g., no personal name), and it may not be asked for over the whole conversation. For example: D: C: D:

C: D: D: C: D: C: D: C: D: C:

Police Desk. Uh, I have a large uh Pontiac station wagon. Do you think you could use it? Yes, sir, the cattle barn is the emergency first aid station set up there and uh if you go to 73rd and Arena, the main gate, and uh tell the officer that you were sent by headquarters to the cattle barn. I’ll be right out there. Thank you. (IPD, #19) Police Desk Uh uh a car accident at 4700 East Lincoln, Taylor’s Lane and Lincoln. 4700 East Lincoln? Yes. Anybody hurt there? I don’t believe so. O.K. . . . I’ll have somebody check. Thank you. (IPD, #83)

Indeed when a personal identification is given, and furthermore is one inviting recognition as an acquaintance (that the form used in the following data is such a form will be shown below), it is not picked up when offered.

Answering the phone

→ →

D: C: D: C: D:

Police Desk, what is it please? Un it’s Watson. Huh? Watson? Yeah. (IPD, #466)

Personal identification, even identification as acquaintance, is not of first-order relevance in conversations begun with a self-identification answer. Unlike calls answered with “hello” in which caller’s first utterance may be inspected for possible acquaintanceship identification, for example, through voice recognition (Schegloff 1979), answerers in this “type” of call display no interest in personal identifications in caller’s first utterance. The foregoing discussion may seem to have introduced a confusion, for while the identification work asserted to be relevant in the overall structural organization of conversation is identification of the initiator or caller, it being his entitlement or his grounds for starting up that are at issue, the selfidentification offered in the answerer’s initial utterance is identification of the called. In this regard, it is, then, critical to note that an identification of either party makes relevant correlative identifications for the other. That this is so turns on the organization of identifications briefly described earlier in connection with the relevance of the identification “acquainted” as first order identification for members (pp. 82–85 above, and Sacks 1972a). Identification terms being organized into classes, the use of a term from a class may make the class or collection from which it is selected relevant for other identifications in the setting. Thus one rule Sacks proposes for the organization of identifications, the “consistency rule,” holds that if a first identification of a member from a population to be categorized or identified is selected from some collection of identification terms, the subsequent identifications in that population may be selected from the same collection of terms. While in selecting analyses of identifications employed by others the consistency rule may have the form of a strong rule, requiring hearing two identifications as selected from a same collection of terms if possible (as is argued in Sacks 1972b), as a rule for speaker’s selection of identifications it is a “weak” rule (i.e., second identifications “may” be selected from a same collection, but need not be). As such, it may be seen to have the status of a minimal preference rule, holding in the absence of any supervening rule, and allowing non-consistent identifications thereby to display the operation of a supervening rule. Unless a supervening rule is used to se-





Emanuel A. Schegloff

lect another identification, then a consistent identification, one from the same collection as an initial identification, will be seen to be relevant. One consequence of this organization of identifications is that when a term of self-identification is used by answerer for his answer, a collection of identification terms is thereby made relevant for the identification of the caller as well. In the absence of claims to, and use of, other identifications which might be selected by supervening rules (e.g., that acquaintances should so identify themselves), the identification from that collection that would hold for caller is the identification of him that has been made relevant for the conversation, and in terms of which answerer may be seen to attend a next utterance. In the case of the primary data under examination here, “Police Desk” makes relevant a collection of identifications for callers related to the “professional” identification of the answerer; in the case of most of the calls in our corpus, the identification this makes relevant for caller is “citizen” or “member of the public” or “complainant.” This is, as was noted earlier, an “identification of sorts,” i.e., an orientation to a “sort of caller” rather than to personal or named identification. It was suggested in the discussion of the overall structural organization of designed conversation that initiators might offer identification even when, by reference to the collection “acquainted-unacquainted,” they were not acquainted and therefore did not find in the identification adequate warrant for starting up the conversation. This was because the legitimacy of reasons, the alternative basis of entitlement, was related to identifications, some reasons which were not legitimate for unidentified members, “anybodies,” being legitimate for members identified in some way. There it was argued that identifications might be offered as orientations for reasons, establishing by the identification the legitimacy of the reason that was to be offered, and thereby also suggesting by the identification the sort of reason that might be forthcoming. This tie between identifications and legitimate reasons should now be seen for its restrictive possibilities as well. While an identification may make legitimate a reason that might not be legitimate (i.e., supplying adequate grounds for starting the conversation, and for possibly continuing it) for an unidentified initiator, or an otherwise identified initiator, a given identification may bound a domain of legitimate reasons for starting a conversation, and presumptively restrict the reasons that may properly be offered for initiating a conversation to one analyzable as within that domain. “Police Desk” (and self-identifications generally, for in these respects “American Airlines,” “Shoe Department,” “Service Department,” etc. are no different), then, serves to presumptively make “identification of sorts” relevant to the conversation it begins, and to deprive “acquaintanceship” identification,

Answering the phone

which requires personal identification, of immediate relevance as a way for caller to establish adequate grounds.24 As between identifications and reasons as adequate grounds for initiating the conversations, reasons are thereby made the primary resource. “Police Desk” does not, however, merely make “identification of sorts” the relevant type of identification; it makes identification “of a particular sort” relevant, i.e., identification as “citizen,” “member-of-public,” or “complainant” vis-a-vis “police” the sort of identification relevant to the initiated conversation. The reasons possibly relevant to initiating the conversation are, then, to be reasons for a “citizen” or “member-of-the-public” to initiate a conversation with “the police” (or, as in the present data more specifically, for “a complainant” to initiate a conversation with the “police complaint desk”). “Police Desk” makes the relevant task of caller’s first utterance giving a legitimate reason for starting “this conversation;” and by introducing the identifications it does, it bounds a domain of reasons, or sorts of reasons, from which a legitimate reason for starting “this conversation,” with “these parties” (i.e., “these parties so identified”) should be offered.25 Self-identification answers of the form of “Police Desk” type the prospective conversations they initiate as “business calls,” proffering identifications of the parties by reference to the sort of business the prospective conversation might be directed to, and bounding the domain from which reasons for the initiation of the conversation are properly selected as ones relevant to the type of business for which the identification of the parties is relevant. As with “fore-known” type calls, the relevance of “type” is not confined to “topical restrictions,” but extends to the ways in which topically relevant talk is constructed. In “shared priority topic” calls, it was noted, such talk may be constructed specifically using indexicals with no explicated referents, as one feature. In “business” type calls, the talk on topics may be constructed so as to display that the talk is organized by reference to the type of conversation, with the types of co-participants who are involved. (“Co-participant orientation” has come up earlier26 without being labelled as such; it will be further discussed below). Thus, for example, callers’ reasons may not only be drawn from a domain relevant to “police,” but the talk on those reasons may have its features co-selected “for the police.” Consider then the formulation of reasons in terms of the criminal law, such as the following:





Emanuel A. Schegloff

→

→

D: C: D: C: D: C:

Police Desk. I’d like to report a vandalism. (IPD, #377) Police Desk. I’d like to call up and report a break-in. Whereabouts. In my on my car, I don’t know for sure whether they got in or not. The both of the doors was still locked, but uh the windshield is on the driver’s side of it is broken. (IPD, #452)

The features of the setting-as-reported and the talk about them are co-selected in forming up a reason “for the police.” Doing “police business” as a type for the conversation has a relevance extending past the initial selection of legitimate reason for initiating the conversation; it is relevant to how the talk about that reason is organized and produced, both in its initial presentation and subsequently.27 Answer forms involving self-identifications such as “police desk” have been discussed here as preemptive identifications, as using the answerer’s first turn to accomplish an initial identification of the parties. When conceived of as an identification issue, however, a basic problem concerning identification may he seen to be involved, namely how the relevance of some among the many possibly correct identifications is established. Given that on any occasion of identification, in principle more than one identification term, from more than one collection of such terms, is available, the selection of some identification can be depicted as an achievement, and the method whereby the outcome was achieved needs to be described. In earlier discussions, it was suggested that where non-first identifications were involved, the consistency rule might provide the relevance of an identification selected from the same collection as a first identification. Thereby, the relevance of the identification of callers of the police being identified as “citizens” or “complainants” given the initial self-identification by answerers as “police;” and thereby the relevance to targets of attempted initiations of conversation of “possible acquaintance,” given that as a possibly relevant basis for the initiators starting up a conversation. But no such basis is available for first identifications; the consistency rule is of no help when there is no initial identification for a subsequent one to be consistent with. This problem was touched on earlier in establishing the basis for the relevance to the possible initiator of a conversation (rather than his target) of the identification “acquainted”; there, the relevance of such an identification was established by reference to greeting obligations

Answering the phone

among acquaintances, and the tasks of monitoring an environment for that obligation that is laid on those who might, by reference to it, have to greet first. The problem arises now again in the case of “Police Desk.” How is the relevance of such an identification established for the ones who answer the phone? How is it selected from the indefinitely large set of collections which have as members of them “correct” identifications for these answerers? The problem posed in this way is a relevant problem if the production of “Police Desk” is posed as an identification problem. That is, if it is taken as given that the answer is to be a self-identification, then the problem is how to select a relevant self-identification, or how to locate a relevant identification, or how the relevance of some selected identification is established. The problem may, however, be otherwise conceived, and is otherwise to be conceived in the case of answers such as “Police Desk.” For rather than the above format, the problem may be seen to involve that the selection of a self-identification form of answer ipso facto selects the particular self-identification to be employed (a similar format was suggested and rejected with respect to first.speakership in Schegloff 2002 [1970]), the possibility having been entertained that the determination that conversation was relevant would thereby select who should begin it). Indeed, it is not that the relevance of a self-identification form requires the selection of a particular self-identification, but that the relevance of a particular identification, germane to answering the phone in the first place, makes relevant the selection of a self-identification form of answer.

Answerers The discussion of the preceding sections has referred to “the answerer,” implicitly thereby either treating the selection of an answerer as non-problematic, or temporarily locating the matters being discussed at a point after an answerer has been selected. The selection (or self-selection) of an answerer is not, however, automatic, and since some of the considerations relevant to that selection are relevant to the answerer’s selection of an answer as well, it is useful to consider the matter briefly at this point. A sense of the issues involved may be conveyed by considering some hypothetical procedures by which an answerer might be selected (or self-selected) upon the ringing of the phone. Aspects of the situation such as the following might supply criteria for the selection, but turn out to be inadequate. Proximity might be used, the person (or adult) nearest the ringing phone being the one to answer it. Some stratified set of formulations of the available personnel





Emanuel A. Schegloff

might be used – the oldest or youngest, highest status or lowest status, person being the one whom the phone’s ring selects. “Answerer” might be made an “official position,” there being a pre-selection of a single person to answer on any occasion of the phone’s ringing, or the rights or obligations of the position being passed around as “turns to answer.” None of these procedures seem correct as first-order solutions to the selection of an answerer. Indeed, in the simplest case, in which there is but one person present when the phone rings, it might be thought there is no problem to be solved, the one present being the only available answerer. Yet it appears that the sole person present may not answer (as in the cases of the guest in the house, or the janitor in a business after hours). The problem is not one of selecting which of the present parties should answer, for it may happen that none of them is found eligible. Instead, the initial problem is one of determining a first-order eligible population of potential answerers, the selection of an answerer being in the first place a selection from that population of candidates, with others answering derivatively – by delegated right, or upon finding that no eligible potential answerer is present or available to answer. The above discussion raises, in a rather different form, a question entertained at the beginning of this whole project (Schegloff 1967, 1968, 2002 [1970]): “who speaks first.” There, the domain of reference was the population of two presumed in the setting features of two-party conversation. It was proposed that, in that form, it was not a general question and had no general solution. For a given formulation of members of that population as “caller-called,” a general solution seemed to hold, formulated by the distribution rule. It should now be seen why that question is not general for unformulated parties, and that, indeed, it is an altogether equivocal question. No basis for bounding the population as a population of two members (or a population of some particular two members between whom a selection of first speaker is to be made) was established. The two members who turn out to be involved in a two.party conversation is not available as a discriminated, bounded population of reference when the beginning of the conversation is at issue. To so treat it would be to assume the accomplished state of affairs as prerequisite to its accomplishment. Under specified circumstances, such a formulation of the problem may be in point, as when a “state of talk” may be seen to have been established between two parties before a first utterance has been made. Then the problem of which of the two should go first may be warranted, the population as a possible population

Answering the phone

of reference having been established (as in the earlier mentioned example of “breaking the ice.”) The problem considered in Schegloff (1968, etc.) was “who speaks first.” The problem here, while it can be seen to be relevant to the production of the same datum, is differently formulated: not “who speaks first” but “who answers.” The utterance concerned is, then, a “formulated utterance,” and although the selected person will indeed “speak first” in the conversation, it is not in the first place relevant that he will “speak first” with respect to the population from which he is selected, for other members of that population may not speak at all (although they are not precluded from doing so, as when members of a family talk serially to some caller). The population of reference here is a bounded population. While the speaker selection problem that was posed in Schegloff (1968, etc.) is thus equivocal and misconceived, there is a relevant speaker selection problem implicated in the data; it involves the selection of an answerer from those present in the environment of the ringing phone, and that involves initially the determination of an eligible population of potential answerers. The determination of particular eligible populations of potential answerers in particular settings, the composition of those populations, and the bases for inclusion or exclusion of particular parties are matters for local determination in each setting. But such local determinations are shaped by a major general order of consideration, with a discussion of which the present chapter will close. The determination of eligible answerers, and the selection by answerers of appropriate answers, are both sensitive to interactional considerations. The relevant considerations are not ones pertaining to interaction ongoing in the setting in which the phone rings. Indeed, it is precisely to such considerations that the structures set off by the telephone’s summons are insensitive: the procedures whereby an answerer is selected can have the consequence that current speakers do not complete their utterances, that selected next speakers do not speak next in order to answer, etc., these being precisely the features to which a telephone summoner – a caller – cannot have attended in placing the summons, i.e., temporal and priority reviews. As they are not employed by caller, and cannot be assumed by answerer, so also they are of diminished relevance in selecting an answerer. Procedures for determining a population of eligible answerers, and of selecting an answerer from among them, might include systematic provision for exempting current speakers, or selected next speakers, etc. This is not the case. The interactional considerations relevant to determination of eligible answerers pertain to the prospective conversation that the telephone ring fore-





Emanuel A. Schegloff

shadows: they are considerations of prospective co-participant orientation. In particular, they are considerations about the identification or formulation of prospective interlocutors, under the auspices of which callers may be seen to have designed the incipient conversation. What is central here is that the telephone is treated as a territorial phenomenon, being socially placed in some territory (in the ethological sense) or itself defining a location. A phone is, in the first place, “somewhere.” Overwhelmingly, furthermore, the territory to which a phone is attached is a “someone’s territory;” that is, members, or classes of members, are affiliated to the territory, having special claims on it; its phone is then their phone. There are classes of territories, and classes of affiliated personnel, the types of classes of personnel being co-relevant with types of class of territory. For telephones, two classes of territory to which phones can be attached are most important: “domestic” and “business.” Each provides for its own relevantly affiliated types of personnel: for domestic, the basic set of classes of members that is co-relevant can be grouped under the rubric “family.” This formulation of the affiliated personnel holds presumptively; that is, callers unacquainted with the persons affiliated to some phone known only to be “domestic,” may treat an answering male voice as “husband,” and ask for a “wife”:

→ →

A: B: A: B: A:

Hello Hello, Mr. Smith? Yes? May I speak to Mrs. Smith, please? There is no Mrs. Smith.

For acquainteds, the presumptive formulations of personnel affiliated to “domestic”’ phones may be replaced by others:

→

A: B: A:

Hello ‘allo G’ your roommate talks forever? or is it you. (CF, 8)

“Business” phones have other sets of formulations of the affiliated personnel presumptively relevant. In either case, an orientation to the type of territory to which a telephone is attached, and its appropriately formulated affiliated personnel, provides callers encountering an unrecognized voice ways of hearing it as “the secretary,” “the wife,” “the roommate,” “the girl friend,” “the house guest,” etc.

Answering the phone

It is membership in the classes of affiliated personnel, co-relevant with the territory in its type to which the phone is attached, which principally shapes the first-order eligible population of potential answerers. It is by reference to its affiliated personnel that callers call, and inspect answerers’ initial utterances. And it is by reference to such orientations on callers’ parts that who might answer is initially determined. It is, therefore, by reference to co-participant considerations for the incipient or prospective interaction, by reference to callers’ orientations, that answering is shaped. Considerations of proximity, seniority, current involvements, etc. may be involved in the organization that results in not everyone making a move for the telephone when it rings, but these considerations are relevant in the first instance for the locally eligible population of potential answerers, however locally composed, by reference to which it is presumed a caller came to get the particular phone to ring. Arrangements for answering are built, together with auspices for calling, to allow the possibility that an initial turn to talk will supply the caller with a “confirmation,” with evidence of having reached what he wanted to reach. Such considerations underlying the bounding of a population of eligible potential answerers, and a selection if possible from among them, it also underlies the selection of an answer in the conversation’s initial turn, bearing on such selections as “hello” for domestic phones, and a self-identification form for “business” phones. When other than eligibles, other than potential answerers, for some phone come to answer it (e.g., having been asked to do so by absent or otherwise occupied eligibles), they may select an answer form to display in first turn that other than an eligible has answered, providing a caller immediately with an account for the absence of the confirmation he might otherwise have been waiting for (letting him know thereby, for example, that he has not reached a wrong number). One such form, for domestic phones is: A:

Hello, Levy residence. (EAS, FN)

The telephone’s ring, then, is the occasion for reviews by all persons present of the place they are in, and the type of place in its phone-relevance, of their affiliation to that place, and thereby of their possible membership in the class of eligible potential answerers. Such reviews will be relevant for any present party, not only on his own behalf, but on behalf of other parties with whom he may be in current interaction, for even if some given party is not a potential answerer, his current interlocutor may be, and if that person is to deal with the summons, there are organizational requirements on the other participants in his ongoing interaction. The telephone’s ring, thus, invokes a set of con-



 Emanuel A. Schegloff

siderations on the part of members co-present with it about their relations to the place they are in, about other co-present persons’ relations to that place, about their current relations with one another, and about the orientations of an at-that-time anonymous caller who is about to be co-participant with one of them. It is out of such considerations, which adumbrate and realize aspects of the upcoming conversation before its first utterance is produced, that answerers get selected to answer the phone, and which strongly condition their selection of an answer form to constitute the conversation’s first turn.

Notes * This chapter is a revision, completed in 1969–1970, of chapter three of my doctoral dissertation, “The First Five Seconds: The Order of Conversational Openings” (Department of Sociology, University of California, Berkeley, 1967). However, bibliographical citations have been updated. The revision of chapter two, completed at the same time, is now available as Schegloff (2002 [1970]), where interested readers may pursue references in this chapter to prior parts of the argument. The data for this chapter and for the dissertation of which it was a part were composed of a corpus of some five hundred telephone calls to the police of a midwestern American city in the immediate aftermath of a disaster. Calls were routed from a central switchboard to a so-called “complaint desk,” at which point the recording of them began. I am indebted to the Disaster Research Center, at that time located at The Ohio State University, which collected the materials and made them available to me when this research was begun. I am also indebted to the Advance Research Projects Agency, DOD, for support through the Air Force Office of Scientific Research under Contract number AF 49(638)-1761. Finally, I am indebted to David Sudnow, whose encouragement and help came at a most strategic time. . In the revised version of the (preceding) chapter two of the work from which this text is taken. Now published as Schegloff (2002 [1970]). . We will have occasion below to justify this formulation of a class of settings, and the formulation “telephone conversation,” both of which have been used throughout the preceding discussion with no warrant. . There is, of course, a considerable range of such activities, and summoned may engage in further identificatory work in selecting an answer to the summons. He may, for example, search the setting, e.g., the appearance of the summoner, to decide, for example, whether the pre-sequenced activity is likely to be a solicitation or direction-asking, whether and/or what to answer turning on the outcome of such further identification. Errors are, of course, possible, as when one turns away from a “lower-class beggar” only to have him follow and reproachfully ask for a match. For a discussion of what “anybody owes anybody” in public places, see Goffman 1963a: 130. . For a discussion of the dimensions relevant to selection of first name as compared to title plus last name see Brown and Ford (1964). The whole corpus of literature on kinship

Answering the phone 

terminology presents materials on what might be called “address entitlements,” which from a summoned’s perspective may serve as resources in recovering from the address term used the position of one who is entitled to use it. For a review of the literature on terms of address, see Hymes (1964); for a recent discussion, see Ervin-Tripp (1969). This discussion is couched in terms of “information about the summoner,” i.e., the sort or category of summoner, for that is the more general case. There is, of course, the possibility, when summoned by voice in face-to-face interaction, of recognizing the voice, and thus recognizing “who the summoner is,” making a personal rather than social identification (see Goffman 1963b: 2, 43, 56–62). Voice recognition, and interest in personal identification and recognition versus category identification, are discussed below [and in Schegloff 1979]. . I do not mean thereby to imply that it is by reference to the differentness that members proceed; they are not engaged, as far as I can see, in comparing face-to-face with mediated interaction, and using the comparison or the respects in which they differ as the basis for their actions. In doing either form of interaction, they deal with the features of that form. Although my discussion appears to deal with telephone interaction as a special case of faceto-face, it is not clear that members deal with it that way, and it is thus here essentially a rhetorical device and part of the format of presentation. The methodological import of Goffman’s suggestion that “. . . telephone talk . . . must first be seen as a departure from the norm, else its structure and significance will be lost” (1964: 135–136) is thus unclear, though his point does not appear aimed at presentational formats. I will argue below, however, that the distinction between the forms is a members’ distinction, and they are attentive to the relevant features of each form. . For example, though persons at a ringing phone may exchange assurances as one goes to answer that “It’s Jim,” the answerer does not regularly answer “Hi, Jim.” . But “just a moment” is; it, however, has different properties when used in telephone conversation; while in unmediated interaction a summoner may go on, the fact of availability of a hearer (in a narrow sense) having been established by the occurrence of an answer, this is not the case in telephone interaction, where “just a moment, please” may be followed by a break in the acoustic channel, or unavailability of an ear at the earpiece. . One way of dealing with the use of variable answer forms is to see it as a matter of convention. But this does not go very far, and merely displaces the problem. For unless the convention is organized such that individuals adopt by convention some one form and use it invariably, which does not appear to be the case, the question remains what the convention is: i.e., what members are required to know, and on what basis selection is made (if selection is involved) which issues in the conventional patterns of use. It will not do, for example, simply to propose that some business establishment adopted the convention of having its telephone answered with the name of the business, and so instructed its employees. For the practice and realization of such a “convention” would still depend on employees finding themselves relevantly to be “at the business” (and thereby also for “persons” to find themselves to be, relevantly at the moment, “employees”) for the convention to be relevant, i.e., it requires on the part of such an employee-answerer certain orientations to relevant formulations of their context. What these are may be general, and relevant far beyond some local set of conventional practices, and may involve the use of analyses and usages of a general

 Emanuel A. Schegloff

character specified to this particular problem. Whether or not convention is involved, then, it is in point to consider what is involved in the selection between answer formats. . The consequences for conversational openings are, of course, not the only ones. What is needed is a description of “telephone conversation” as a technical object, a description that might imply a technical account of “face-to-face” interaction as well. Some elements of such a description are available, but would lead too far afield in the present context. [Cf. the later Schegloff 1993.] . How there came to be a selection issue for answers on the telephone, or how there came to be a selection issue for clearance cue answers but not between clearance cue and problematic answers on the telephone, as compared to face-to-face interaction, is a matter of the historical development of the conversational system, and its adaptation to the contingencies presented by the telephone. That a conversational system may develop to exploit the peculiar features of its setting is suggested by the exploitation of the resources for voice recognition in specially pure form in telephone interaction. (The last point was suggested by Harvey Sacks.) Voice recognition is discussed below. In any case, for current members, the selection issue for answers on the telephone is an established resource whose availability is not their responsibility or concern. . The relevance of the last observation will become clear in the later discussion; the contrast is with an answer like “Macy’s” which may also invoke a relevant “agenda of sorts,” but displays the basis of the agenda in the answer. . A methodological point may be in order here. This discussion of “yeah” may be compared to “yeah” as an answer to a summons in face-to-face interaction. They involve utterly different usages in many respects, by virtue of a) what they follow, b) the kind of selection issues that result in them and which they display, and c) the way those two features, combined with others, are fitted to – and partially constitute – different settings. An attempt to deal with an utterance such as “yeah” semantically, therefore, seems doomed. Interactional analysis is required. That much seems clear even before dealing with “yeah” in a second slot where the first slot is filled with a remark (other than a summons), let alone “yeah” elsewhere in conversation. . Discussed in Schegloff (2002 [1970]). . It should be noted that although “hi” may be the appropriate greeting form for “acquainteds” or “intimates,” and thus from the point of view of the caller, knowing he is calling an intimate, the greeting form it is appropriate for him to receive, if the answerer’s initial utterance is “hi,” it will not be heard as an appropriate greeting but as a mistaken opening of an expected call. Insofar as an initial “hello” is treated as a greeting, and one which turns out to be too “formal” for the caller, and which is then transformed to “hi,” it appears that the whole process is required: starting out with the “hi” that might turn out to be an appropriate form would not short-cut the process, but would be heard as a different action entirely. . As shown in Schegloff (2002 [1970]). . What the parameters of “close pacing” are remains to be determined. One possibility is the relationship of “next action” discussed in Schegloff (2002 [1970]). . Much of what follows is discussed in Goffman, 1963a: 112–148, although the present discussion differs in several respects from Goffman’s, most centrally in that Goffman is con-

Answering the phone 

cerned with the conditions for different sorts of “face engagements,” while I am concerned with establishing the basis for sorts of work that need to get done as constituent parts of the unit “a single conversation,” that supply features of the overall structural organization of that unit. . The remainder of this paragraph draws on Sacks, 1972a and 1972b. On recognition, see also Goffman 1963: 112–114; 57–68). . It is the occurrence of “absent first greetings” that forces the specifications about entitlement to reciprocal recognition, status orderings on who greets first, etc. For it might have been expected that if A does not greet B, B will greet A, and then either A will do a return greeting or he will not. In the former case, an exchange will have occurred; in the latter case, it will be a second, or return, greeting that will be absent. The fact that first greetings are sometimes found absent suggests that there may be grounds for B to not initiate the exchange if A has failed to do so; status relations, and a linked asymmetry of rights to be recognized by the other, may supply such grounds. . To be sure, settings may vary, and may be attended by members as varying, in their possible-acquaintance-richness. Where a setting, a place-time, is treated as possibleacquaintance-poor, monitoring for possible greeting occasions may not be relevant, and persons may then be found to fail to recognize. The relevance of “possible recognition” will be discussed below with respect to voice recognition on the telephone. . It has “no interest” insofar as it does not entail greeting obligations, although it may entail others, e.g., avoidance obligations. That its relevance to the possible legitimacy of initiating a conversation is oriented to can be seen in members’ use of opening utterances such as “You don’t know me, but. . .” . Where recognition is not available to visual inspection, as when persons who have corresponded and/or spoken on the telephone are to meet in a public place, they may undertake identification work to establish acquaintanceship as the adequate grounds for starting a conversation before greetings: A: B: A: B: A:

Paul Smith? Yes? I’m Al Jones. Oh, hi. Hi.

Note: l) the introductions at the beginning are specially notable in that introductions are “historically sensitive.” In contrast to greetings which are properly done whatever the history of contact between two persons, being relevant at first meetings and after fifty years of marriage (though for repeated encounters within a short period of time, their relevance may gradually fade, e.g., over the course of a business day in an office), introductions should not be done, should not have to be done, after a first, or possibly a second, meeting (cf. Sacks 1975). For ones who turn out to be acquainteds, introductions are therefore specially marked activities. Routinely, they should start with greetings. 2) There are circumstances in which one may find greetings and self-introductions in that order, but those are regularly features of by-product interactions, or ones designed to appear by-product, as in initiation of conversation at parties with ones who “happen to be” next to one. 3) A sequence of the

 Emanuel A. Schegloff

gross sort as the one cited may occur when it is not “identification as acquainteds” that is being accomplished, as when a piece of business is to be executed with “a Mr. Smith,” and one starts by enquiring or confirming that it is the relevant person one has located, and then identifying oneself. But in the data cited: a) the self-introduction employs the “recognition-relevant” frame (“I am...” rather than “My name is...”); b) the response is a recognition sign (see discussion of voice recognition below); and c) although it is not indicated in the citation, no (further) account for initiating the conversation is given, whereas in the contrasting case, such a sequence ought be followed by “the business” that might warrant having undertaken such an opening. In the last respect, considerations of a known shared agenda, known and shared by virtue of prior acquaintanceship or prior contact, such as was discussed in connection with “fore-known” calls, appears relevant, and marks the opening as establishing acquaintanceship as adequate grounds for starting the conversation, rather than self-introductions as preliminaries to announcing the adequate grounds for initiating the conversation. In telephone conversation, where acquaintanceship recognition is also not available to visual inspection, identification work will also be found relevant (see below). . The ways in which the alternative bases of warrant for the conversation are related to its overall structural organization remain to be described in detail. Such a full discussion cannot be entertained here, but one kind of linkage can be outlined. I have referred to one of the alternatives as “reasons”, that is a reference to the domain which is drawn upon in warranting the initiation of a conversation; it is a “sort” of warrant. For any particular conversation, an initiator does not require reasons; as a general matter, it would appear, that using this sort of warrant requires of an initiator “a reason.” Conversations initiated with “a reason” as their warrant (i.e., as their sole warrant, when acquaintanceship is not employed) may have their developing overall structure attended by the parties as one based on the topic the reason serves to constitute. They may find where they are in the conversation by finding where they are in that topic, the end of the “reason topic” serving as the occasion for starting the closing of the conversation (Schegloff & Sacks 1973), silences there being treated, for example, as closing-relevant rather than as occasions for transition to a next topic. Such conversations, i.e., designed conversations with a reason as their warrant, may be treated prospectively as presumptively mono-topical. It is such a joint orientation to the prospective structure of the conversation that may serve as the basis for an initiator of such a contact “prefacing” an initial topic with a marker such as “two things,” indicating thereby that the overall structure of the conversation should not be constituted progressively using the initial topic as model. “Monotopicality” is an attribute of overall structural organization. (It should be noted, however, that it is not the first in a series of attributes which might continue “bi-topicality, tri-topicality. . .” The critical distinction, for the purposes of overall structural organization, is between one and more than one, because of the way that sets different bases for coordinating the closing of the conversation. Accordingly, the “two” in “two things” is not to be, and is not, treated literally, constraining the conversation from moving to a third topic. What is critical is the “more-than-one.” The markers used to indicate prospective poly-topicality use a variety of means to accomplish that; I cite several for illustrative purposes, taken from a radio talk show, omitting the opening sequence in each:

Answering the phone  B:

one quick quesh’n before y’ get into my topic. (BC, tan, p. 62) _____________ B: First of all, I wanna c’ngrachulachu – c’ngeatchulatche on your progr’m. (BC, tan, p. 83) _____________ B: well numbuh one, the woman you just finished talking to, (BC, tan, p. 104) _______________ B: First of all, uh, I understand yer going intuh the hospital. A: Yes ma’am. B: And uh–let me jus’ wish you lots of good luck. A: Thank you. B: Uhm, the second thing, about the mayor an’ the governor, I’ve listen’ to all the things thet chu’ve said, an’ I agree with you so much. Now, I wanna ask you something . . . (BC, red, pp. 189–190) ______________ B: Two quick comments, A: Yes sir. B: D’you get the feelin thet we’ve been outflanked by the Pueblo an’ the garbage affayuhs, A: ehh heh heh heyeh heh heh heh heh heh ha ha heh B: Right/ An’ also the other fast comment . . . (BC, red, p. 89) The use of “list-ing” as a technique relies on the treatment of lists as having more than one member, a “first” being a “first” only if there is a “second.” These techniques may, of course, accomplish other outcomes in addition to indicating non-monotopicality; they may, for example, serve to indicate that the thing first to be talked about is not to be treated as “the reason,” an accomplishment which may itself be arrived at by different techniques; they may serve as indications of the degree to which topics ought to be “worked up” or elaborated, the projection of an “agenda” possibly affecting time allocation; or the degree to which “interactionally generated topics” will be accommodated, i.e., topics which “come up” by way of topics under way. In all these respects, as well, overall structural organization is constituted. In any case, these techniques seem addressed to heading off, or transforming, an otherwise presumptive type of overall structural organization for the initiated conversation, i.e., mono-topical. Here, then, is one instance of an operation referred to earlier in the text without examples – a transformation of conversational type after the opening. . This does not mean that callers cannot establish acquaintanceship; only that it, together with other forms of possible personal identification, are not features in terms of which a next utterance, or next utterances are inspected. Thus, callers’ first utterances which, when placed after “hello,” serve as occasions for possible voice recognition and are seen to invite possible voice recognition, are not so seen and are not so treated when placed after “police desk,” or self-identifications of that form. Callers may, for example, invite voice recognition after an answerer’s “hello” by using “hello” alone (or some variant) as their first utterance.

 Emanuel A. Schegloff

→ →

#197 Police make call. A(woman): hello D: Yeah. A: You mean yor’re not busy/ D: Oh yeah, but I thought I’d call and let you know what we know so far.

After “police desk” such utterances are heard not as voice-recognition-relevant, but as confusions of the summons-answer structure, and warrant a repeat of the answer.

→ →

#365 D: Police desk. C: Hello. D: Police desk.

. Caller’s utterances may be inspected in order to find how they might be ones which formulate reasons which can warrant the conversation with the parties as identified. Such inspection may lead to direct inquiries, e.g.: #115 D: C: D: C: D: C: →

D:

Police desk. Hello/ Hello. I come down here to see my wife jill, I’m from Missouri? Yes. And uh they ain’t down here now, they moved. (Pause) Well, what do you want me to do?

. In Schegloff (2002 [1970]). . The failure to find the “police relevant reason” in #115 cited in the preceding note may be related to caller’s failure there to co-select the parts of his talk so as to formulate “businessfor-the-police” as he might have, perhaps, by talk of “missing persons.”

References Birdwhistell, Ray L. (1970). Kinesics and Context: Essays on Body Motion Communication. New York: Ballentine Books. Brown, Roger, & Ford, M. (1964). “Address in American English”. In D. Hymes (Ed.), Language in Culture and Society (pp. 234–244). New York: Harper and Row. Ervin-Tripp, Susan (1969). “Sociolinguistic Rules of Address”. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology (pp. 93–107). Goffman, Erving (1953). Communication Conduct in an Island Community. Unpublished Ph.D. Dissertation. Department of Sociology, University of Chicago.

Answering the phone 

Goffman, Erving (1963a). Behavior in Public Places: Notes on the Social Organization of Gatherings. New York: Free Press. Goffman, Erving (1963b). Stigma: Notes on the Management of Spoiled Identity. New York: Simon and Schuster. Goffman, Erving (1964). “The Neglected Situation”. In J. J. Gumperz & D. Hymes, (Eds.), The Ethnography of Communication American Anthropologist, 66 (6 pt. II), 133–136. Hall, Edward T. (1959). The Silent Language. New York: Doubleday. Hymes, Dell (1964). Language in Culture and Society. New York: Harper & Row. Kendon, Adam (1967). “Some Functions of Gaze-Direction in Social Interaction”. Acta Psychologica, 26, 22–63. Sacks, Harvey (1972a). “An Initial Investigation of the Usability of Conversational Materials for Doing Sociology”. In D. N. Sudnow (Ed.), Studies in Social Interaction (pp. 31–74). New York: Free Press. Sacks, Harvey (1972b). “On the Analyzability of Stories by Children”. In J. J. Gumperz & D. Hymes (Eds.) Directions in Sociolinguistics: The Ethnography of Communication (pp. 325–345). New York: Holt, Rinehart and Winston. Sacks, Harvey (1975). “Everyone Has to Lie”. In M. Sanches & B. G. Blount (Eds.), Sociocultural Dimensions of Language Use (pp. 57–80). New York: Academic Press. Sacks, Harvey (1992a). Lectures on Conversation. Volume 1. Edited by Gail Jefferson, with an Introduction by Emanuel A. Schegloff. Oxford: Blackwell. Sacks, Harvey (1992b). Lectures on Conversation. Volume 2. Edited by Gail Jefferson, with an Introduction by Emanuel A. Schegloff. Oxford: Blackwell. Scheflen, Albert E. (1963). “Communication and Regulation in Psychotherapy”. Psychiatry, 26, 26–136. Scheflen, Albert E. (1964). “The Significance of Posture in Communication Systems”. Psychiatry, 27, 316–331. Schegloff, Emanuel A. (1967). The First Five Seconds: The Order of Conversational Openings. Unpublished Ph.D. Dissertation, Department of Sociology, University of California, Berkeley. Schegloff, Emanuel A. (1968). “Sequencing in Conversational Openings”. American Anthropologist, 70, 1075–1095. Schegloff, Emanuel A. (1979). “Identification and Recognition in Telephone Openings”. In G. Psathas (Ed.), Everyday Language: Studies in Ethnomethodology (pp. 23–78). New York: Erlbaum. Schegloff, Emanuel A. (1986). “The Routine as Achievement”. Human Studies, 9, 111–151. Schegloff, Emanuel A. (1993). “Telephone Conversation”. In R. E. Asher (Ed.), Encyclopedia of Language and Linguistic, Volume 9 (pp. 4547–4549). Oxford, England: Pergammon Press. Schegloff, Emanuel A. (2002 [1970]). “Opening Sequencing”. In J. E. Katz & M. Aakhus (Eds.), Perpetual Contact: Mobile communication, private talk, public performance (pp. 326–385). Cambridge: Cambridge University Press. Sommer, Robert (1959). “Studies in Personal Space”. Sociometry, 23, 247–260. Sommer, Robert (1969). Personal Space: The Behavioral Basis of Design. Englewood Cliffs, N.J.: Prentice-Hall.

Investigating reported absences ‘Neutrally’ catching the truants* Anita Pomerantz

In a fair number of jobs, at least some of the work that employees perform involves interacting with others. In these cases, the talk is not incidental to the work; rather it is the way the work gets done. One job in which the employees get their tasks done by talking is that of a clerk in a high school attendance office. As part of their job, clerks call the homes of students who are, or have been, absent to investigate the causes of the absences. I have an audiotape of calls in which a clerk called the homes of absent students to investigate the absences. In the series of calls, the clerk identified herself and talked as a representative of the attendance office. In exploring how the clerk talked, I identify some concerns that were part of the clerk’s investigative work and describe some devices or methods that she used in carrying out the investigations. I will argue that one of the clerk’s concerns, as exhibited in the talk, was conducting the investigation in a neutral way. I will show what neutrality meant in terms of conduct and discuss some methods she used in accomplishing neutrality. Before I discuss features of the clerk’s talk, I will provide some background on how high school attendance offices functioned. In high school, the school day was divided into periods; students went to different classes each period. During each period, teachers recorded the names of absent students and sent the names to the attendance office. In the attendance office, a clerk entered the absences on the records of the individual students. The attendance office had two categories of absences: “excused” or legitimate and “unexcused” or illegitimate. Parents and guardians were informed of school procedures for absences. They were to call the attendance office in the morning if their child was going to be absent. For a number of reported absences, however, the attendance

 Anita Pomerantz

office received no calls from the parents/guardians. An attendance office clerk investigated some of these unaccounted for absences by telephoning the home. An absence officially was classified as “excused” if and when the returning student brought in a written affirmation signed by the parent or guardian. That affirmation described a legitimate reason for the absence, for example, that the student was ill or had a family emergency. The participants referred to a signed affirmation as “a note.” The thesis is this: Throughout the interactions, the clerk dealt with absences that she viewed as probably legitimate differently than those that she viewed as possibly or probably illegitimate. I make inferences about when the clerk and the parent suspected truancy from their conduct. I will attempt to show that when the clerk suspected truancy, she engaged in more work to be ‘neutral.’ Her discourse displayed a tension between speaking from an assumption of probable truancy and speaking as if she made no assumption of guilt. Likewise, when the parent suspected truancy, she displayed a similar tension. I will review three cases which involve contrasts with respect to whether the interactants suspected truancy. In the first case (Call 1), I see no evidence of the clerk’s suspecting truancy at call beginning; the mother did not suspect truancy throughout the call. In the second case (Call 5), I do not believe the clerk suspected truancy at call beginning; the mother did suspect truancy. The mother subsequently remembered information which dismissed their suspicions. In the third case, I see evidence of the clerk’s suspicions at call beginning. As the call evolved, both parties suspected truancy. Prior to discussing the three cases, I will make some comments on how the clerk oriented to the records she used in her work. The records were fundamental to her work. Based on reports of absences, the clerk selected homes to call and asked the parents to provide information about the students’ whereabouts during the times that they were reported absent. Yet it seems that the clerk oriented to the records as at best accurate but possibly inaccurate. In making the phone calls, all the clerk had to go on were the records of reported absences. Inaccurate records could be problematic because she needed information concerning actual absences and, given that she represented the school, she might not have wanted to expose the inaccurate records. The clerk used a variety of solutions to deal with the problem of ‘possibly inaccurate records’. Consider her common opening information-seeking question: “Was _____ home from school ill today?” The parent warrantably could infer that the attendance office had reports indicating that the child was not in school that day. When a parent confirmed that the child was legitimately absent, no further investigation was needed. When the parent did not confirm

Investigating reported absences

it, the clerk itemized the extent of the absence by periods. In the instances in which the parent did not confirm the opening inquiry and the clerk enumerated the reported absences, often there was a mismatch between the duration in the opening inquiry and the extent of the reported absences. Call 5 illustrates this phenomenon. In her initial inquiry, the clerk implied a whole day’s absence (“Was Bryan home from school ill today?”). In enumerating the absences, she named only two periods (“Well he was reported absent from his thir:d and his fifth period cla:sses today.”). In the original inquiry, the clerk did not constrain herself to reference only reported absences and ask something like, “Well, was Bryan home between eleven and one o’clock?” or “Was Bryan home for two hours today?”. Rather she initially formulated the extent of the absence (a whole day) by what may have been the likely actual absence given the recorded absences. On the basis of this instance and others, I suggest that the clerk used reported absences as indicators of actual absences. She seemed to operate with the assumption that reported absences likely indicated real absences but that real absences sometimes went unreported. So she could read the record of reported absences in such a way as to perhaps guess what was likely as to the extent of a student’s actual absence. She selected formulations to cover not only the reported absences but the ‘more’ that possibly were not reported. I now turn to discussion of the three contrastive cases. Recall that the cases differ with respect to whether the participants suspect truancy. I will show how the parties oriented to the business of the call, that is, gathering/reporting the facts rather than assessing the status of the absence, in the varying circumstances.

Instance 1: Call 11 In Call 1, neither the clerk nor the mother exhibited signs of suspecting truancy. [Call 1] Clerk: ...this is Miss Devalo from Alvera High School calling? Mother: Mmhm:, Clerk: Uh was Mar:k home from school i:ll t’da:[y? Mother: [Ah::: yes he was in fact, I’m sorry I I didn’t ca:ll because uh: I slept in late I (.) haven’t been feeling well either, .hhhh and uh: .hh (0.4) he have a myihkno:w a:, fever:: (.) this mornin[g Clerk: [Uh huh, .hhh





Anita Pomerantz

Mother: A::nd uh I don’know y’know if h’l be:,h in tihmorrow fer sure er no:t he’s kine’v js been layin arou:-ndj(h)ih know[: Clerk: [O:kay well I’ll go’head’n:: uh::m I won’t call you t’morrow ni:ght if: we don’t see im t’morrow wi’ll jst assume he wz home i:ll.[.hhhh Mother: [.tch ih-Ri:gh[t c’z eeClerk: [A n : - send a note with him when he does return.

It seems that the clerk and mother understood that the purpose of the phone call was merely investigative: that it was appropriate to exchange information but not to draw the conclusion explicitly as to whether the absence was legitimate or illegitimate. The clerk’s job was just to collect the facts; the mother’s job was just to report them. The clerk’s orientation to the appropriateness of just collecting facts can be seen in the way she began the investigation. While the question “Was Mark home from school ill today?” was intelligible as “Was Mark’s absence legitimate?”, these two versions are neither equivalent nor interchangeable. The clerk did not ask the mother “Was Mark’s absence excused?” and the mother did not assert “Mark’s absence is excused.” So while both the clerk and mother understood the import of reporting that Mark was home ill, they did not say in so many words that the absence was excused. In short, the clerk asked the mother for a report regarding a circumstance to which she had direct access; the mother gave the report without adding a conclusion. When I discuss Call 4, I will make an argument that the clerk’s choice of a way to begin seeking information reflected whether or not she suspected truancy. In this call, the clerk began the investigation with a question that incorporated an instance of an excused absence. With the question, “Was Mark home from school ill toda:y?”, the clerk sought information by incorporating a common or typical reason for a legitimate absence. She offered “home ill” as the sort of information that she was trying to obtain, as a sample of the kind information that she was seeking. The clerk and the parents recognized “home ill” as an item in a set of items, as one reason among others for legitimate absences. In addition to “home ill,” they knew that “doctor’s appointment,” “family emergency,” and others provided for legitimate absences. When the clerk incorporated an item like “home ill” in a question, she was not asking the parent to merely confirm or disconfirm that item. Rather, she was asking whether that item, or another one from the same set, was true.2 Upon hearing the question, a parent had several options: to confirm the legitimate reason, to affirm an alternative legitimate reason for the absence, or to claim no knowledge of a legitimate reason.

Investigating reported absences

Later in the analysis, I hope to show that the clerk’s choice of incorporating a legitimate reason within a question to open the investigation probably was related to her not particularly suspecting truancy. When the mother confirmed that Mark was home ill, she supplied details that supported or substantiated the confirmation. She reported that “he have a myihkno:w a:, fever:: (.) this morning.” Embedded in the confirmation/elaboration of Mark’s being home ill, the mother apologized and accounted for not having called the attendance office to report his absence. Clerk: Uh was Mar:k home from school i:ll t’da:[y? Mother: [Ah::: yes he was in fact, I’m sorry I I didn’t ca:ll because uh: I slept in late I (.) haven’t been feeling well either, . . .

In apologizing and accounting for her failure to call, she displayed her knowledge that calling was the proper protocol for excused absences. Perhaps more interesting is the mother’s inclusion of “in fact” when she confirmed that her son was home ill. Including a claim of facticity is an attempt to bolster the facticity of what one is asserting. We bolster the facticity when doubt has been displayed or is anticipated. Given her ‘failure’ to call, it is plausible that the mother considered the call from the attendance office an indication that they suspected truancy; thus she bolstered the facticity of her assertion that her son was home ill. The organization of the work talk in this phone call was fairly straightforward. The clerk sought information that would account for the student’s absence by offering a common or typical account for legitimate absences. The mother confirmed with supporting details. The mother raised the possibility of the absence continuing through next day; the clerk informed the mother how the office would handle that contingency. Finally the clerk reminded the mother to send a note upon Mark’s return. Based on a small sample, I propose that the organization of work talk in the calls in which the parties displayed no particular suspicion fit this template: the clerk inquired, offering an instance of a legitimate absence, the parent confirmed a legitimate absence and indicated when the child would return to school, and the clerk reminded the parent to send a note. Once the clerk received confirmation of a legitimate reason, she treated the investigation as essentially closed. This sharply contrasts with calls in which the clerk received no testimony from the parents as to the legitimacy of the absences.





Anita Pomerantz

Instance 2: Call 5 In Call 5, the clerk seemed not to be suspicious of truancy when she initiated the inquiry. Upon hearing the clerk’s initial inquiry, the mother did suspect truancy. I will examine the mother’s conduct while she apparently suspected truancy and the clerk’s responses to her conduct. When the mother was asked whether her son was home ill, she disconfirmed that possibility. However, she did not disconfirm it immediately. Rather, she turned to someone in the household and checked it out with him. Clerk:

Was Bryan home from school ill today? (0.4) Mother: .hhhh (0.7) Bryan wasn’ home ill today, was he. ((Off phone)) (0.5) Male: Not at all Mother: No. Clerk: M[m hmm Mother: [No he wasn’t

In a way it is curious that the mother turned to check with the other person before answering the clerk’s question. It is curious because the mother already had a answer to the question: Even while checking with the other person, she implicitly took the position that Bryan had not been home ill. Yet by turning to the other person, she treated her position as ‘uncertain.’ She increased the certainty of her position by double-checking with the other person. She reported her position to the clerk only after getting it confirmed by the other. In the course of two turns, there was a transformation from ‘thinking that something was the case’ to ‘knowing and asserting that it was the case.’ In double-checking, the mother displayed a tentativeness or lack of certainty about her position. After the strong confirmation, she asserted the position with certainty. At least some of the time, double-checking allows one to assert with confidence the facticity of a proposal. The power of a second opinion, or verification, can be quite startling. In double-checking, the mother displayed an awareness that her report constituted damaging or incriminating evidence. That the mother was disconfirming accounts for a legitimate absence was relevant to her doubled-checking before answering. When members of society recognize a matter as consequential, they are more likely to double-check on it. In double-checking, they treat the matter as consequential. In this call, the mother was in the process of reporting that her son wasn’t home that day. Inasmuch as she knew the import

Investigating reported absences

and consequentiality of that report, she elected to make certain of it before speaking out. It is conceivable that the clerk, based on the mother’s report that Bryan had not been home ill, might have concluded something like “It sounds like Bryan was truant.” Whatever her private conclusions might have been, in her conduct she treated the call as an occasion for exchanging information, not for making judgments. In response to the mother’s report that Bryan had not been home ill, she reported on the record of Bryan’s absences. In describing the record of Bryan’s absences, she described the events (absences) for which she sought an account. Clerk:

Was Bryan home from school ill today? ((lines deleted)) Mother: No he wasn’t Clerk: .hhh Well he was reported absent from his thir:d and his fifth period cla:sses today.

In using the term “reported,” the clerk framed the information on Bryan’s absences as reports given to the office. Had she said, “Well he was absent from his third and his fifth period classes today,” she implicitly would have been endorsing the reported absences as actual absences. The clerk sometimes talked about “absences” and sometimes talked about “reported absences.” “Reported absences” is a more careful formulation in that the speaker of it does not endorse the accuracy of the record. Let me sum up this initial exchange. In response to the clerk’s fact-seeking inquiry, the mother reported facts that were damaging to her son. Prior to giving the damaging report, she double-checked to ‘make certain’ of the facts. The clerk responded to the damaging report not by concluding anything about the absences but by detailing the reported absences. In doing so, she reported the events for which she sought an account. Upon hearing about Bryan’s reported absences, the mother apparently drew the conclusion that her son had been truant. Both the mother and the clerk dealt with a tension between the mother’s clearly having a conclusion and their understanding that they were to merely exchange information, not judge the son’s absence at this time. I will track the conversational moves that seem to exhibit this tension.





Anita Pomerantz

Clerk: Mother: Clerk:

Mother: Clerk: Mother: Clerk: Mother:

.hhh Well he was reported absent from his thir:d and his fifth period cla:sses today. Uh huh .hh A:n we need ’im to come inna the office in the morning to clear this up. (1.0) .hh O:kay I’ll tell’im that, [(uhm/and) [O:ka::y He has no excuse! Uh huh= =as far as I know,

The marked intonation of the mother’s acknowledgment “Uh huh” conveyed a message something like: ‘I fully understand.’ Rather than explicitly addressing the mother’s understanding regarding the guilt of the son, the clerk informed the mother of the attendance office procedures in handling the case. In telling the mother what the attendance office needed, the clerk framed the context for the resolution of the problematic absence to be between the student and the attendance office. The mother acknowledged the clerk’s instructions on handling the problematic absence (“Okay I’ll tell him that”) but went on to more explicitly judge the absence. In saying “He has no excuse!” the mother moved further away from reporting just the facts, just what she had access to. Here she asserted with certainly and emphasis not only the non-existence of “home ill” but of any legitimate excuse. If Bryan had no legitimate excuse for his absence, by definition he was truant. So while the mother did not literally say that Bryan was truant, she essentially said it. The clerk’s response to the mother’s condemnation was a very bland acknowledgement. Blandly acknowledging the prior was a way of not engaging further with it. With no gap after the clerk’s bland acknowledgment, the mother appended “as far as I know,”. Both the words and the tone of voice sound like the mother backed off, changed from strongly condemning her son to taking a more uncertain stance. When the mother declared “He has no excuse,” she asserted a condition that defined the son as guilty. In adding “as far as I know,” the mother displayed her awareness that her assertion and the implicated judgment of guilt were beyond what she could and/or should warrantably assert. To append the qualification, the mother apparently reviewed her own assertion and found it overstated.

Investigating reported absences

The mother’s engaging in this kind of review is understandable, almost expected, in a context in which she was being asked for facts and in which judgments were inappropriate. These types of qualifications or disclaimers are used in contexts in which people testify and are held accountable for the truth of their assertions. Even though the mother qualified her condemnation, she continued to speak from a framework of assuming her son’s guilt. Her recommendation regarding just punishment for her son presupposed his guilt: Mother: Clerk: Mother: Clerk: Mother:

He has no excuse! Uh huh= =as far as I know, Okay, A:nd uh (1.0) Mother: Maybe it’ll be a good idea tih have ’im sit in for the hours thet he missed. Clerk: Okay fine. Well, if he can’ bring in a note (he c’n)

The mother presented her idea for just punishment packaged as a tentative suggestion (“A:nd uh, (1.0) Maybe it’ll be a good idea...) Consistent with the way the clerk dealt with the mother’s previous judgment, the clerk again refrained from engaging in the discussion on the mother’s terms. The clerk responded to the mother’s suggestion of punishment with an acknowledgment (“Okay fine”) followed by a condition. Using the conditional form “if ”, the clerk transformed the mother’s assumption of guilt into a matter not yet established. Moreover, rather than referring directly to Bryan’s “guilt” as the matter not yet established, she referred to his not bringing in a note as the matter not yet established (“if he can’t bring in a note from you”). In this way, the clerk maintained a focus on the attendance office’s procedures, framing the judgment as something the office would do in the future. By casting punishment as contingent on whether or not the student would return to school with a note, the clerk implied that the mother’s discussion of punishment was premature. Punishment would be relevant only at a later stage and only if certain contingencies occurred. The clerk’s maintaining the stance that Bryan’s guilt or innocence would be determined through proper procedures at a later stage was part of achieving neutrality. In the call, the clerk had the job of acknowledging the mother’s contributions, moving forward in the interaction, while not validating or endorsing the mother’s assumption that her son had been truant. She managed this by bland





Anita Pomerantz

acknowledgments and invocations of routinized office procedures for handling unresolved absences with the student.

Instance 3: Call 2 In some of the calls on tape, the clerk seemed to suspect truancy from the outset of the call. I cannot say whether the clerk’s apparent suspicions were based on the way she read the pattern of recorded absences, the history of the child, or on other information. In Call 2, the clerk’s conduct in opening the investigation suggests that she suspected truancy. [Call 2] ((Standard Orthography)) Mother: Hello:, (.) Clerk: .hhh e-Hello:. Is this Missiz Auerbach? Mother: Yes it i:[s, Clerk: [.hhhh Uh this is Miss Devalo from Alvera High s:School calling? Mother: M-[hm::? Clerk: [.hhhh Ah I was calling about Michelle she has a couple of absences sin:ce: u-oh:: las:t Thursday, .hhh She’s been reported absent (0.2) .t all day last Thursda:y, (.) Mother: Uh hu:h, well she hasn’t been home i:ll,

The clerk began seeking information in a way that is strikingly different from the previous two calls examined. In the previous calls, the clerk began seeking information by asking a question which incorporated a common or typical account for an excused absence: “Was ______ home from school ill today?”. While the clerk explicitly sought information from the parent via the inquiry, she implicitly gave information to the parent as well. In asking whether the student was home ill that day, she implicitly informed the parent that the student was not in school that day. In starting her investigation with a question, the clerk explicitly asked for information while she implicitly informed the parent about a potential problem or circumstance upon which the call was predicated. In Call 2, the clerk began her investigation by informing the mother about her daughter’s absences.

Investigating reported absences

Clerk: .hhhh Uh this is Miss Devalo from Alvera High s:School calling? Mother: M-[hm::? Clerk: [.hhhh Ah I was calling about Michelle she has a couple of absences sin:ce: u-oh:: las:t Thursday, .hhh She’s been reported absent (0.2) .t all day last Thursda:y

The clerk began by naming the subject of the call “I was calling about Michelle...” She revealed whom the call was about but not what it was about. This format is sometimes used in delivering bad news. The clerk continued with an overview of the problematic situation: “she has a couple of absences sin:ce: u-oh:: las:t Thursday,”. Parenthetically, the clerk’s formulation “couple of absences” was a solution to possibly inaccurate attendance records. “A couple” is a proper gloss for different configurations of actual absences. It allowed the clerk to indicate the number of absences without having to be held accountable for the accuracy of a specified number. After giving the overview of absences, the clerk started to enumerate the reported absences (“She’s been reported absent (0.2) .t all day last Thursda:y,”). In describing the absences, the clerk implicitly sought information from the mother that would be relevant to the absences. When an interactant offers a version of an event based on limited access to a co-interactant who has (or should have) more access to the event, the co-interactant’s version becomes relevant (Pomerantz 1980). Instead of directly asking for information, interactants sometimes give ‘my side’ tellings in hopes that co-interactants will volunteer the information. What are the differences between the two methods the clerk used to initiate her investigations? Recall that the question format used by the clerk involved incorporating a legitimate reason for an absence. The clerk could appear neutral, as simply gathering information, if a legitimate excuse was incorporated into the question. The clerk’s second method consisted of informing the mother of the absences and letting her account for them. With this method, the clerk elicited information without having to formulate a question. The formulation of the question seems problematic in circumstances of suspicion. It is likely that the position incorporated within the question is taken as the speaker’s ‘best guess’ or expectation. If the clerk suspected truancy, her best guess would be that the student was not home ill for the duration of the absence. In this circumstance, the clerk might be reluctant to incorporate a legitimate excuse in her inquiry. Yet if she incorporated the negative case into her question, the clerk’s neutral-



 Anita Pomerantz

ity would appear compromised. She would not be seen to be simply gathering information but perhaps judging or pre-judging the child. With the informing method of information-seeking, the clerk avoided building into her inquiry any ‘guess’ as to the reason for the absences (Pomerantz 1988). The clerk produced no query instructing the parent on the type of information she sought or expected. If she had reason to keep her guesses or expectations to herself, the second method was well suited to this need. While both the mother and the clerk apparently suspected truancy during much of the call, both maintained the stance of neutrality. The clerk elicited only information (not judgments) and the mother reported only information. Whatever their private suspicions, they treated the status of Michelle’s absence as a matter not yet determined and in need further investigation. In starting her inquiry by informing the mother of her daughter’s absences, the clerk merely reported the facts as were available to her. However when one listens to the opening, one gets the sense that the clerk probably was suspicious. If she were, her conduct did not blatantly reveal it. She stuck to reporting only facts and did so carefully. It is perhaps that careful reporting of evidence that provides the sense of her suspicion. In both words and tone of voice, the mother’s response displayed her own suspicions about the absences. Clerk:

.hhhh Ah I was calling about Michelle she has a couple of absences sin:ce: u-oh:: las:t Thursday, .hhh She’s been reported absent (0.2) .t all day last Thursda:y, (.) Mother: Uh hu:h, well she hasn’t been home i:ll,

While the mother may have believed that her daughter had been truant, on record she did not offer a judgment. Recall the mother in Call 5 went beyond what she knew for sure and essentially issued a judgment when she asserted “He has no excuse.” In this call, the mother did not go that far. She asserted that the circumstance that commonly legitimizes absences, home ill, did not exist. While we can hear her reference to the one legitimate excuse as standing for all legitimate excuses, on record she denied only “home ill”, a circumstance to which she was expected to have had access. To reiterate, the clerk’s and the mother’s conduct suggests that they probably suspected truancy. Assuming this to be the case, the clerk avoided revealing her suspicions or expectations by inquiring without use of a question format. She presented the office’s evidence and let the mother account for the absences. The mother avoided explicitly judging her daughter as truant, as having no le-

Investigating reported absences

gitimate excuse for her absences, by asserting that the one circumstance that typically legitimizes absences, “home ill,” did not exist. There is a funny quality to these materials. In the context of the parties’ suspecting truancy, they seemed very careful in their claims. They worked to report ‘just the facts’ and they marked when they went beyond the facts. For example, the clerk was careful to mark when her assertion when beyond the evidence. When the mother asked whether Michelle’s absences were all day absences, the clerk did a very careful confirmation. She enumerated the absences shown in the record and then indicated that “all day absences” was a reasonable conclusion or assumption. Clerk:

.tlk We:ll (.) she: was absent Thursda:y? Frida:y? (0.8) .t.hhh and again toda:y. Mother: Are these all day absence:s or are they: (.) just (.) certain periods. Clerk: Uh:::: hhhhmhh.t.hhhh (0.7) Well let’s see it looks like fir:st second (0.3) third and fourth period for last Thursday and Friday .hhh and here’s sixth period so it’s (.) e- (.) we’d have to assume that it’s an all day a:bsence ye:s.

The clerk’s enumeration of reported absences in response to the mother’s inquiry can be seen as a careful way of speaking. It displays her apparent concern to make claims for only what she could support. By including “We’d have to assume...”, she claimed that “all day absences” was an assumption rather than a fact, though a very reasonable assumption. With respect to a carefulness about one’s claims, for me the weirdest line in the call was the clerk’s “And uh...You don’t know that she’s been home ill, huh?” Clerk: so it’s (.) e- (.) we’d have to assume that it’s an all day a:bsence ye:s.= Mother: =M-h[m::, Clerk: [.hhhh And uh:::, .hhhh.t You don’t kno:w that she’s been home ill hu:h, Mother: n:No:: not to my knowledge I (.) I’m (.) stay at ho:me so:,h (0.3) Clerk: Mm hm, .t[.hhh Mother: [I would know if she was ou:t,

Recall that upon hearing the clerk report on her daughter’s absences, the mother addressed the reason for the absences: “Well she hasn’t been home i:ll.” The clerk’s response was to overview the extent of the absences and they had several turns on that topic. To move on with the call’s business, the clerk needed to return to the mother’s addressing the reason for the absence. Inasmuch as



 Anita Pomerantz

the mother had asserted, “Well she hasn’t been home ill,”, the clerk might have returned to the topic with a request for confirmation: “And uh she hasn’t been home ill huh?” The clerk did request a confirmation, but the form she used was peculiar: “And uh:::, .hhhh.t You don’t kno:w that she’s been home ill hu:h,”. Rather than asking the mother to confirm what she has previously asserted, the clerk used a more cautious formulation. Rather than ask for confirmation that Michelle was not at home, the clerk asked for confirmation that the mother did not have the knowledge to confirm that Michelle was home ill. The clerk’s question exhibited a carefulness with respect to claims in that she asked only about the mother’s knowledge, not about the actual circumstance. The clerk made a very careful, perhaps overly careful, formulation. These types of careful formulations are produced in environments in which parties are held accountable for their claims. Inasmuch as the clerk oriented to the mother’s testimony as incriminating, she moved into a cautious mode of asking the mother to report ‘just what she knew for sure.’ Given the framing of the clerk’s question, the mother was put in a position of responding in those terms. The clerk asked the mother to confirm her lack of knowledge of her daughter’s having been home ill; the mother confirmed in those terms but added an argument that her knowledge as a good indicator of the actual circumstance. Clerk: You don’t kno:w that she’s been home ill hu:h, Mother: n:No:: not to my knowledge I (.) I’m (.) stay at ho:me so:,h (0.3) Clerk: Mm hm, .t[.hhh Mother: [I would know if she was ou:t,

In the context created by the clerk’s question, the mother gave a qualified report about her state of knowledge. However she added information with which she claimed that she was well positioned as an observer to give a valid report. Having gotten the mother’s confirmation that Michelle had not been home ill, she moved on to the remaining business of the call. She offered a conclusion based on the mother’s report and then informed the mother of the appropriate steps to take.

Investigating reported absences 

Clerk:

.hhhh And uh:::, .hhhh.t You don’t kno:w that she’s been home ill hu:h, Mother: n:No:: not to my knowledge I (.) I’m (.) stay at ho:me so:,h (0.3) Clerk: Mm hm, .t[.hhh Mother: [I would know if she was ou:t, Clerk: Yea:h. Well, u::h the- obviously she’s not going to her cla:sse:s an’ we need tuh find out where she is: going would you: talk t’her about it tonight’n .hhh send her into the attendance office tomorrow morning so that we c’n straighten these out. Mother: Mmmkay:: I’ll see what she says about it.

With the mother’s confirmation that her daughter had not been home ill, the clerk might have concluded that the absences were illegitimate. However the work of the call was to investigate absences, not judge them. At this point in the interaction, the clerk did draw a conclusion, but it was different from the obvious one. Rather than concluding truancy, the clerk offered a substitute conclusion: “Well obviously she’s not going to her classes.” In her conclusion, the clerk restated the problematic circumstance (the absences), formulating Michelle as the actor/agent. The clerk then offered the ’necessary’ response to this problematic circumstance: to investigate where Michelle was going. Rather than concluding that the absence was illegitimate, the clerk concluded that there was a problematic circumstance and that further investigation was required. Concluding that a problem existed that needed further investigation was consistent with her investigative stance. The response that the clerk advocated to deal with the problem was shaped as a collaborative effort between the mother and the attendance office. Well, u::h the- obviously she’s not going to her cla:sse:s an’ we need tuh find out where she is: going would you: talk t’her about it tonight ’n .hhh send her into the attendance office tomorrow morning so that we c’n straighten these out. (emphasis added)

The clerk cast the investigative activity as teamwork, specifying a part of the investigation that the mother should orchestrate. The tension over suspecting truancy on the one hand and not yet judging the absence (remaining neutral) on the other was reflected in the clerk’s talk of punishment. In raising the topic of the detention system, the clerk implied that truancy was a possibility.

 Anita Pomerantz

Clerk: Mother: Clerk: Clerk: Clerk: Clerk:

...send her into the attendance office tomorrow morning so that we c’n straighten these out. Mmmkay[:: I’ll see what she says about it. [Uh::: we have a new.hhh Okay:. We have a new uh:: detention system now (.) that if they don’ clear the:se, they’ll become truants. (.) .hh A:nd she will need to come in en clear them up.

While the clerk raised the topic of truancy, she did not explicitly talk about it as a possibility for Michelle. She did not say, “If Michelle is truant, she’ll go to detention.” Rather the clerk announced a new system involving detention. She described the policy as it applied to all students: If students do not clear up their absences, they become truancies. The clerk spoke about truancy in terms of the general policy, not specifically with reference to Michelle’s absences. The clerk then returned to speak about Michelle’s absences, informing the mother that the daughter needed to handle the absences (“clear them up”) with the attendance office upon returning. While the clerk raised the issues of truancy and detention, she did so by citing the general policy, the conditions under which students become truant and go to detention. When speaking about Michelle’s absences, she referred only to the necessity of dealing with the unresolved problem, of “straightening out” or “clearing up” the absences. Discussing punishment in the context of informing the mother of a new school policy was a solution to the clerk’s suspecting truancy while she officially remained neutral.

Postscript As a postscript to this talk, I would like to comment on two issues. The first issue is methodological, the second involves the substance of the findings or claims that I have offered. As I mentioned earlier, the corpus I worked with was a series of phone calls made by a clerk in a high school attendance office while she was engaged in checking whether or not absences were legitimate. In other words, I had a series of interactions with the same party doing the same job across different circumstances and with different recipients. Having that corpus allowed me to notice similarities and differences and make claims that I would not have been able to make had I had just one call or had the calls been made by

Investigating reported absences

different clerks. Across the corpus, the clerk used a formulaic invariant way of identifying herself; she used one of two different practices when she initiated her investigatory work. Since the speaker was constant, I was able to look to other aspects of the situation to account for the differences. Regarding the findings or claims I made in this presentation, I hope two phenomena receive attention from scholars in the future. The first is that when representatives of organizations rely on records that they view as quite possibly inaccurate, do they engage in practices that gloss the inaccuracies and minimize exposing the records as faulty? If so, what are those practices? If not, do they display concerns about possible consequences? The second is to ascertain whether, in different institutional contexts, personnel who are institutional mandated to be neutral use hyper-careful formulations when they are suspicious regarding the case in question?

Notes * In form, this chapter is a hybrid between a lecture and a paper. I originally gave the lecture at the University of Konstanz (Germany) in 1980. Gail Jefferson produced both a verbatim and an edited transcript; I tightened up the analysis while leaving the argument intact. . See Appendix for transcript of entire call for all three cases sited in the text. . Sacks named this type of question a Correction Invitation Device because in proffering one item, a speaker invites a correction (a different item) if applicable. As Sacks described it, this device works because a recipient of such a question determines/infers the project or interest behind the question and often will supply the information that satisfies it. See Sacks’ (1992, Vol. 1, pp. 21–23, 380–381) discussion of the Correction Invitation Device.

References Pomerantz, Anita (1980). “Telling my side: ‘Limited access’ as a fishing device”. Sociological Inquiry, 501 (3–4), 186–198. Pomerantz, Anita (1988). “Offering a Candidate: An information seeking strategy”. Communication Monographs, 55, 360–373. Sacks, Harvey (1992). Lectures on Conversation. Vol. I. Oxford, UK: Basil Blackwell.



 Anita Pomerantz

Appendix Call 1 Mother: Hello. Clerk: Mother: Clerk: Mother: Clerk: Mother:

Clerk: Mother: Clerk:

Mother: Clerk: Mother: Clerk: Mother: Clerk: Mother: Clerk:

(0.5) Hello Mister Johnson? (0.8) Uh: this is Missus Johnson, Uh Missus Johnson I’m sorry.=This is Miss Devalo from Alvera High School calling? Mmhm:, Uh was Mar:k home from school i:ll t’da:[y? [Ah::: yes he was in fact, I’m sorry I I didn’t ca:ll because uh: I slept in late I (.) haven’t been feeling well either, .hhhh and uh: .hh (0.4) he have a myihkno:w a:, fever:: (.) this mornin[g [Uh huh, .hhh A::nd uh I don’know y’know if h’l be:,h in tihmorrow fer sure er no:t he’s kine’v js been layin arou:-ndj(h)ih know[: [O:kay well I’ll go’head’n:: uh::m I won’t call you t’morrow ni:ght if: we don’t see im t’morrow wi’ll jst assume he wz home i:ll.[.hhhh [.tch ih-Ri:gh[t c’z ee[A n : - send a note with him when he does return. Mm [hm [O:kay. Okay= =Thank you Uh huh Bye [bye [B’bye

Call 2 Mother: Hello:, (.) Clerk: .hhh e-Hello:. Is this Missiz Auerbach? Mother: Yes it i:[s, Clerk: [.hhhh Uh this is Miss Devalo from Alvera High s:School calling?

Investigating reported absences 

Mother: M-[hm::? Clerk: [.hhhh Ah I was calling about Michelle she has a couple of absences sin:ce: u-oh:: las:t Thursday, .hhh She’s been reported absent (0.2) .t all day last Thursda:y, (.) Mother: Uh hu:h, well she hasn’t been home i:ll, (0.5) Clerk: .tlk We:ll (.) she: was absent Thursda:y? Frida:y? (0.8) .t.hhh and again toda:y. Mother: Are these all day absence:s or are they: (.) just (.) certain periods. Clerk: Uh:::: hhhhmhh.t.hhhh (0.7) Well let’s see it looks like fir:st second (0.3) third and fourth period for last Thursday and Friday .hhh and here’s sixth period so it’s (.) e- (.) we’d have to assume that it’s an all day a:bsence ye:s.= Mother: =M-h[m::, Clerk: [.hhhh And uh:::, .hhhh.t You don’t kno:w that she’s been home ill hu:h, Mother: n:No:: not to my knowledge I (.) I’m (.) stay at ho:me so:,h (0.3) Clerk: Mm hm, .t[.hhh Mother: [I would know if she was ou:t, Clerk: Yea:h. Well, u::h the- obviously she’s not going to her cla:sse:s an’ we need tuh find out where she is: going would you: talk t’her about it tonight’n .hhh send her into the attendance office tomorrow morning so that we c’n straighten these out. Mother: Mmmkay[:: I’ll see what she says about it. Clerk: [Uh::: we have a newClerk: .hhh Okay:. We have a new uh:: detention system now (.) Clerk: that if they don’ clear the:se, they’ll become truants. (.) Clerk: .hh A:nd she will need to come in en clear them up. (.) Mother: Nnk[ay Clerk: [Okay? Mother: Do: I have tuh get back t’you ’r (.) jus’ sending her is that enough. Clerk: .hhh Well if you c’n excu:se any of these with a note saying yes she’s been home ill er at the doctor’s or whatever .hhhh uh:: (.) just send a note but othe[rwise you don’ need tuh come in Mother: [(Yeah)

 Anita Pomerantz

Mother: Clerk: Mother: Clerk: Mother: Clerk: Mother: Clerk:

Okay then. Mka[y [I’ll talk tuh her about it en (.) we’ll get back tuh you. O:kay A[lright [Thank you Mb’bye Bye bye

Call 5 Male: Clerk: Male: Mother: Clerk: Mother: Clerk: Mother: Male: Mother: Clerk: Mother: Clerk: Mother: Clerk:

Mother: Clerk: Mother: Clerk: Mother: Clerk: Mother: Mother:

Hello? Hello, may I speak to Misses: Forde? eYeah hol’ on. (10.0) Hi:: Hello this is Miss Devalo from Alvera High School calling, Uh huh, Was Bryan home from school ill today? (0.4) .hhhh (0.7) Bryan wasn’ home ill today, was he. ((Off phone)) (0.5) Not at all No. M[m hmm [No he wasn’t .hhh Well he was reported absent from his thir:d and his fifth period cla:sses today. Uh huh .hh A:n we need ’im to come inna the office in the morning to clear this up. (1.0) .hh O:kay I’ll tell’im that, [(uhm/and) [O:ka::y He has no excuse! Uh huh= =as far as I know, Okay, A:nd uh (1.0) Maybe it’ll be a good idea tih have ’im sit in for the hours thet he missed.

Investigating reported absences 

Clerk: Okay fine. Well, if he can’ bring in a note [(he c’n) Mother: [Oh waitta minute e- waitta minute Clerk: Okay, Mother: .hh ah he hadda a do:ctor’s appointment. Clerk: [[Ah hah Mother: [[That’s right (.) Clerk: he did. heh h[eh .hhhhh Mother: [huh huh Mother: See I was gunna get ma:d right away. Clerk: Hhhuh huh huh .ehhhh[huh Mother: [Cause he is capable of doing that. Clerk: Mm hm[m(hh)m Mother: [Okay I’ll getta- give’m th’excuse. Clerk: O:ka[:y: tha:nk you Mother: [Okay Mother: Byebye

“At first I thought” A normalizing device for extraordinary events* Gail Jefferson

The Phenomenon Recurrently, in their talk about various sorts of events, people include an item that can be roughly formatted as ‘At First I Thought X, Then I Realized Y’.

A Brief History Among the scores of phenomena that the late Harvey Sacks collected (from ‘dreams’ to ‘shopping lists’, from ‘intonation’ to ‘mock facts’ to “How are you?”, from Caryl Chessman to ‘cigarettes’ to ‘symmetry’), some folders containing no more than a single instance and a brief preliminary consideration, others bursting at the seams, was ‘Joke/Serious as an Oriented-to Contrast Class’. It is a hefty folder with dozens of instances and a variety of considerations. One of those considerations was included in a lecture he gave in Fall 1967, in which he used ‘Joke/Serious’ as a way to approach the issue of Ambiguity. Here is an excerpt from that lecture (Sacks 1992, Vol. I: 671–672). One tends to think about ‘ambiguity’ that, for example, a word could mean this or that, or a sentence could mean this or that, or it could mean this, or that, or God only knows what else. Now, the sort of ‘ambiguity’ that I’m interested in specifically, is sequentialized ambiguity, where the issue is what sort of thing should go next, turning on what this thing might have been. For example, on the occurrence of some first pair-members (such things as questions, offers, requests, etc.), there can be a particular sort of sequential ambiguity present, the alternatives: Is this serious or is it a joke. And we can find such next utterances as “Are you kidding?”, “Are you serious?”, “You’re joking!”. Now, what

 Gail Jefferson

such an utterance is specifically attending is the issue of what sequence the first utterance should generate. Are you doing an ‘insult’ to which I should do a ‘return insult’? Are you doing a command’ which I should accept or reject? Are you making an ‘offer’ which I should accept or reject? Do you want me to marry you? Or, for any of these, are you just kidding. We are not, then, talking abut the issue of an ambiguity of meaning in the sense of is it this term or is it that term, but the issue of does it have this sequence appropriate after it or that sequence appropriate after it; a possible acceptance of the proposal, or laughter. Thus, the decision that someone is kidding, that something is a joke, means effectively: Whatever sequence this thing might generate if it were ‘serious’ does not apply.

That is the excerpt. In that lecture there is no particular piece of data mentioned, but after Sacks’ death in 1975 as I began to go through his research notes, in the folder with the ‘sequential ambiguity’ consideration I found newspaper clippings, excerpts from books, etc., all lumped together under the general topic ‘Joke/Serious’. Some of them are clearly the sort of thing Sacks must have been referring to in his lecture. For example: (1) [SPC Materials:1964] ((Woman talking about her husband who has threatened suicide)) I just acted like I thought he was just kidding. I didn’t want him to think I was taking him seriously. He said “Well Joey run down to the police station before I do something I don’t want to do.” . . . I says “Joey run outside. Daddy’s only kidding.” (2) [New York Times, November 20, 1964] Two women ran into a Bronx drugstore yesterday morning and one remarked: “Imagine! We just saw three men go into the bank with masks and a gun.” Andrew Mack, owner of the store, which is directly across the street from the bank, looked up and said: “Aw, you’re joking.” But the women were right, and Mr. Mack phoned the police . . . (3) [New York Times, July 19, 1965] ((Two boys walking down the street; one is killed by a sniper’s bullet)) He staggered several steps after the bullet hit him and collapsed on the concrete. His companion, 17-year-old Thomas Wilson, said later that he had heard “a light pop” but thought nothing of it. When Peter fell, the police said, his friend turned to him and said, “Cut out the kidding – don’t kid me.” Thomas then saw blood trickling from his companion’s arm and ran to a nearby tavern to phone the police. (4) [New York Times, July 13, 1968] Police commissioner Howard R. Leary apologized yesterday to a Bronx

“At first I thought”

clothing store owner who called the police on Thursday evening to tell them that a man was shooting at people on East 138th Street outside his shop. His call was met with disbelief. Three men were killed in the shooting. Leo Llonch, the store owner, said that when he called the police on the new citywide 911 emergency number, the policeman he spoke to asked, “Are you pulling my leg?”1 (5) [Excerpt from Kafka’s The Trial, pages 7–8] Who could these men be? What were they talking about? What authority could they represent? . . . one could certainly regard the whole thing as a joke, a rude joke which his colleagues in the Bank had concocted for some unknown reason, perhaps because this was his thirtieth birthday, that was of course possible, perhaps he had only to laugh knowingly in these men’s faces and they would laugh with him . . . (6) [New York Times, 1968] News of the invasion of his homeland fell like a crushing weight on Jan Kavan, a principal student leader in the Czechoslovak reform movement. . . . The 22-year-old student said he thought the first reports of the invasion were a joke. When they were confirmed, he said in an interview yesterday, he went into a state of shock. (7) [From The Witnesses. Testimony of Abraham Zapruder] I heard the first shot and I saw the President lean over and grab himself like this (holding left chest area). . . . For a moment I thought it was, you know, like you say “Oh he got me” . . . you’ve heard those expressions, and then I saw – I don’t believe the President is going to make jokes like this, but before I had a chance to organize my mind I heard a second shot and then I saw his head open up and the blood and everything came out and I started – I can hardly talk about it. (The witness is crying.)

Now, several of these fragments happen to have, not only the ‘Joke/Serious’ alternation, but another feature: Reports of ‘first thoughts’. In fragments (6) and (7) the reported ‘first thought’ is that it’s a joke. But, for example, in fragment (3) we have built into the report that the youngster “heard a light pop but thought nothing of it”. This, in a context where readers already know, and the boy already knew when he made this statement, that the “light pop” was the sniper’s gunshot. And it appears that Sacks was beginning to attend this feature in its own right, independent of the ‘Joke/Serious’ alternation. Specifically, in the folder marked ‘Joke/Serious’ are a couple of items which have nothing to do with that, but which are instances of these reported ‘first thoughts’. For example:



 Gail Jefferson

(8) [From The Witnesses.] ((Testimony of William R. Greer, the secret service agent driving the Presidential limousine.)) Well, when we were going down Elm Street, I heard a noise that I thought was a backfire of one of the motorcycle policemen. . . . And then I heard it again. And I glanced over my shoulder. And I saw Governor Connally like he was starting to fall. Then I realized there was something wrong.

This, and fragments like it, although they were stuck into the ‘Joke/Serious’ folder, have no mention of joking. And in my experience this is a fact, and a pleasure, of collecting instances: Inevitably one comes across materials that don’t fit under the heading one has set oneself to collect on some particular data-run, but which seem to be related; ‘ballpark phenomena’ that might cast some light on the focal phenomenon and/or point to independently interesting issues. Again, then, the material I show as fragment (8) is simply stuck into the ‘Joke/Serious’ folder although there is no reference to joking in it. And it appears that while the phenomenon was beginning to emerge, it had not yet ‘surfaced’. So, in a lecture in May, 1968 dealing with the workings of verbs, Sacks focuses on the ‘Thought/Realized’ alternation, using a piece of ‘first thought’ data which he refers to as his only case. Here is an excerpt from that lecture, the first time he presented this material (Sacks 1992, Vol. I: 787–788). Verbs seem to be one routine area for doing such a thing as ‘showing an intention of the truth of some statement’. I have a case in mind, and although it’s the only case I happen to have, I don’t think it’s peculiar. It’s from the New York Times, November second, nineteen sixty seven, headed “New Auto Fines System in Effect; First Public Reaction is Sour”.

(9) [New York Times, November 2, 1967] At about ten thirty a.m. yesterday an Adelphi College student parked his car at a meter on 78th Street between 5th and Madison Avenues, and went to pick up his girlfriend. A half hour later the student, David Searles, returned to the street with his girl and found the car was missing. At first he thought it had been stolen. Then he realized it had been towed away by the police, and still later he realized that he was one of the first victims of the new, higher parking fines that went into effect yesterday.2 Focusing on “. . . then he realized it had been towed away by the police”. In characterizing what he did as ‘realized’, what’s being said is that it turned out to have been correct. That is, in the use of ‘realized’ the correctness of his thoughts is proposed. Were the report to be delivered at the time

“At first I thought”

that he did his considerations about where the car is, we wouldn’t get “I realize the car has been taken by the police.” What we would likely get is, “I (guess, bet, wonder if) the car’s been taken by the police”, or “Maybe the car’s been taken by the police”, and things like that. So what we have is something like: ‘Realize’ stands in opposition to ‘thought’ by reference to the fact that ‘thought’ is used when it turns out to be wrong. “At first he thought it had been stolen.”.

That is the excerpt; the first time Sacks presented this sort of material. One thing that often happens is that someone presents a phenomenon, and thereafter others begin to come up with cases. After this lecture in late May of 1968, I found myself noticing and clipping materials in which the word ‘thought’, and things like it, are used when what was ‘thought’ turns out to be wrong. For example: (10) [Los Angeles Times, February 22, 1969]3 ((The R. F. Kennedy assassination inquest; testimony of a bystander who was shot)) “I felt someone kick me”, said Stroll, adding that he didn’t know at first that he had been shot. “Then I noticed – because I had on blue pants – that one of my legs was red.”. (11) [Los Angeles Times, April 11, 1969] A Pepperdine College security guard fatally shot Larry Kimmons, 16, without warning, four companions of the youth testified at a Los Angeles Coroner’s inquest Thursday. . . . [One companion] said his first reaction was that the guard was kidding and that he had only fired a blank shell as he yelled at Kimmons, “Come on Larry, get up, get up.” (12) [Los Angeles Times, January 19, 1970] ((Interview with a last-minute substitute for the Pro Bowl football game)) Caught in traffic, Larsen reached the Coliseum five minutes before the kickoff. Changing into uniform, he charged out of the tunnel at 1:10 p.m. A loud roar greeted his appearance. “For a second”, he said, “I thought the cheer was for me. Then I realized that the West had just gotten the ball for the first time.

And for this latter fragment, I had a companion piece: Some years earlier, a friend of mine, a novice actress, had described her reaction to a traditional occurrence at an Opening Night party in New York City: Every member of the cast, no matter how minor a role they have, is applauded when they enter the restaurant. This was her first experience of it.



 Gail Jefferson

(13) [Verbatim Report; a novice actress, ca 1962] When they started applauding I thought Mimsy [the star of the show] was behind me or something. I did one of these [she turns and looks behind her]. Nobody there. They were applauding for me!

These materials were simply marked as instances of the ‘Thought/Realized’ alternation that Sacks had talked of in his lecture; specifically, as materials which could be roughly formatted as a sequence: First I thought X, then I realized Y. And I began to develop some preliminary notes on reported first thoughts, as having an obscure relationship with – and not necessarily giving access to – what people are actually thinking. Reported first thoughts as assertions, as subject to social organization, i.e., as ‘selected appropriate first thoughts’; thoughts appropriate to some situation and/or Membership Category, for example, the striking fittedness of a football superstar’s asserted reaction to the roar of a crowd (as something he was due and accustomed to) and a novice actress’s asserted response to applause (‘surely not for little me’). Each, then, achieving the ‘arrogance’ or ‘modesty’ appropriate to their respective Membership Category. Other issues started coming to mind. For example, that while people don’t go around reporting each and every thought, on whatsoever, and especially not volunteering their wrong thoughts; and while the mass media do all sorts of editing-out of ramblings and irrelevancies in interviews, here were these objects, again and again, not only reported by people but preserved in the media. And I began to look at those reported ‘first thoughts’ in detail. One thing I noticed was that the offered ‘first thought’ in fragment (11) seemed far-fetched. Consider what sort of pre-planning and coordination it would require for this unknown guard and this boy, Larry Kimmons, to bring off a joke in which the guard suddenly fires a blank shell, and the boy, with appropriate timing, falls to the ground. Nevertheless, as a whole it seems perfectly acceptable, plausible, does not stimulate inquiry into its constituent features. I wondered if it might not be that this plausibility has to do with the context, i.e., has to do with what the reality turned out to be. By contrast to the extraordinary facts of the matter, the reported ‘first thought’ in fragment (11) and others, stand as unremarkable, usual, etc. I began to get a sense that these reported ‘first thoughts’ were products of a search/selection procedure for a formulation of some problematic event, where the search was geared to finding a likely, i.e., non-extraordinary formulation of the event.

“At first I thought” 

And counter-cases came to mind. Such things as ‘crank calls’ to the police, the burglar-under-the-bed phenomenon, etc. That is, there is a known set of things one should not make of a situation. As it happened, at that time and place (1969–1970, Irvine, California), there was a relevant recurrence: Again and again there would be disparaging mentions in the newspapers and on the radio, of people who would call the police asking if we were being invaded every time a nearby missile base sent up an experiment – some of which were spectacular to see. So there seemed to be a business here. Roughly, a reported ‘first thought’ recurrently constituted an innocuous, ordinary alternative to an extraordinary actuality. It turned out that Sacks had been struck by that aspect of the phenomenon. Here is an excerpt from a lecture about the achieved status of ‘being ordinary’ that he gave in Spring, 1970 (Sacks 1992, Vol. II: 220). It’s really remarkable to see people’s efforts to achieve the ‘nothing happened’ sense of really catastrophic events. I’ve been collecting fragments out of newspapers, of hijackings, and what the airplane passengers think when a hijacking takes place. The latest one I happened to find goes something like this: “I was walking up towards the front of the airplane and I saw by the cabin, the stewardess standing facing the cabin, and a fellow standing with a gun in her back. And my first thought was he’s showing her the gun, and then I realized that couldn’t be, and then it turned out he was hijacking the plane. And another goes (this was a Polish plane hijacking), a passenger reports: “I thought to myself, we just had a Polish hijacking a month ago and they’re already making a movie of it.” And a classically dramatic instance is, almost universally the initial report of the J. F. Kennedy assassination was of firecrackers. Just imagine the Old Testament in its monumental events, with ordinary people having gone through it. What would they have heard and seen, e.g., when voices called out to them, when it started to rain, etc. There is at least one place in the Old Testament where that happens. Lot was warned of the burning of Sodom and Gomorrah, and was permitted to bring his daughters and sons-in- law out:

(14) [Genesis, Chapter 19] And Lot went out, and spake unto his sons in law, which married his daughters, and said, Up, get you out of this place; for the Lord will destroy this city. But he seemed as one that mocked unto his sons in law. And they stayed behind.

That is the excerpt. A few years ago, Paul Drew came up with a case from the Book of Samuel. As it happens, this case fits a bit more closely to Sacks’ de-

 Gail Jefferson

scription of “ordinary people” reacting to such things as “voices calling out” than does Genesis:19. In Samuel:I:3 the “nothing happened” reaction is not, as between Lot and his sons-in-law, to the voice of a fellow human, but (albeit by a child) to the voice of God. (15) [Samuel:I:3] And the child Samuel ministered unto the Lord before Eli. And the word of the Lord was precious in those days; there was no open vision. And it came to pass at that time, when Eli was laid down in his place, and his eyes began to wax dim, that he could not see; And ere the lamp of God went out in the temple of the Lord, where the ark of God was, and Samuel was laid down to sleep; → That the Lord called Samuel: and he answered, Here am I. And he ran unto Eli, and said, Here am I; for thou calledst me. And he said, I called not; lie down again. And he went and lay down. → And the Lord called yet again, Samuel. And Samuel arose and went to Eli, and said, Here am I; for thou didst call me. And he answered, I called not, my son; lie down again. Now Samuel did not yet know the Lord, neither was the word of the Lord yet revealed unto him. → And the Lord called Samuel again the third time. And he arose and went to Eli and said, Here am I; for thou didst call me. And Eli perceived that the Lord had called the child.

In a sense, this is a double instance, i.e., not only does Samuel make the most ordinary sense of this voice in the night, figuring that Eli wants some service of him, but Eli, by his laconic treatment of Samuel, seems to figure that the boy is just dreaming – it takes God three attempts before Eli ‘perceives’ the extraordinary facts of the matter.4 Reviewing the foregoing excerpts, we find that over the years Sacks used the same sort of material in three different sorts of presentations: – – –

Fall 1967: As a basis for considering the ‘Joke/Serious’ alternation in terms of sequential ambiguity. Spring 1968: As a basis for considering the ‘Thought/Realized’ alternation, where “...‘thought’ is used when it turns out to be wrong.” Spring 1970: As a basis for considering ordinary perception of catastrophic events.

It was the Spring, 1968 lecture that had started me collecting cases of ‘at first I thought X, then I realized Y’, from which I’d begun to develop some consid-

“At first I thought” 

erations. But it was the Spring, 1970 lecture that clicked the phenomenon into place for me. The assertion, and preservation/transmission, of these ‘wrong’, sometimes really far-fetched, ‘first thoughts’ about terrible events was a device; an incantation; a ritual used to manage, to put into normal perspective, something that might otherwise be disruptive. As Sacks has it, “to achieve the ‘nothing happened’ sense of really catastrophic events”. In the sequential terms posed in the Fall, 1967 lecture, if these things could not be put into normal perspective, some action would have to be taken about this new reality. In the couple of weeks following the April 2, 1970 lecture, I found two more instances: (16) [Los Angeles Times, April 8, 1970] Mrs. Martha Harmon will never forget the sound of her children’s voices screaming in the night. “At first it sounded like they were just fussing,” she recalled with a shudder Tuesday. “But then I heard the oldest one yell fire. That woke me.” (17) [Television interview. Witness to a shooting at a student demonstration, Isla Vista, California, April 18, 1970] He said “I think I’ve been hit”, or “I think I’ve been shot.” Whatever he said, he wasn’t sure. . . . I thought he’d been hit by a rock or something. . .

And I’d begun to work up a consideration of the phenomenon. Following is a rough sketch of the phenomenon, the result of work done partly by Harvey Sacks, and partly by me.

Notes on “At first I thought” In the aftermath of some problematic events we recurrently find reports in the mass media which include people’s ‘first thoughts’ about the event – ‘first thoughts’ which turned out to have been wrong. Given what the event turned out to be, the wrong ‘first thoughts’ can be seen as plausible (although scrutiny might reveal them to be odd, far-fetched, etc.). Their plausibility resides in that they stand as innocuous, ordinary, Anybody’s Alternatives to what turned out, on some particular occasion, to be the actuality, i.e., they assert what any one of us would, could, should make of such an event.

 Gail Jefferson

Asserting the wrong ‘first thought’ reaffirms, in the face of some actuality, the in-principle correctness of the ordinary alternative. In effect it is proposing that the wrong ‘first thought’ should have been right. Now, the phenomenon of Reported First Thoughts is very much a matter of extraordinary events, witnesses, and mass media; people speaking for the record, their words being preserved and transmitted, etc. But to get a sense of what I’m proposing to be the in-principle-correctness work of these media objects, I’ll turn to the realm of utterly ordinary conversation about altogether trivial events. For here, in ordinary conversation, is the principle and wellspring of the resources being deployed on those more rarified occasions. Consider, for example, the way “I thought” is recurrently used in ordinary conversation. I’ll start off with a fragment in which someone produces the same “thought” for two consecutive participants; a little girl, and then the little girl’s father. Here, the little girl, Sharon, has phoned to invite her friend Stephanie to the beach, where Sharon and her family are spending their vacation. Her friend Stephanie isn’t home, and Sharon finds herself speaking to Stephanie’s mother, Fran, and eventually handing Fran over to her father (and Fran’s friend), Ted. (18) [NB:III:1:2-4:Standard Orthography] Fran: ↑Well when did you guys go::. Sharon: Ah: Saturday?hh Fran: → ↑Oh: for crying out loud. I thought it was the e:nd of the mo:nth you were go::i:[ng. Sharon: [Mm-mm,hh Fran: ·hhhh ↑Oh:::::::::::. . . ((ca 50 lines omitted; Sharon’s father is now on the line)) . Fran: → Well I thought you weren’t going down til next seh-u-the weekeh:-I mean the end of the mo:nth. (0.4) Ted: ↑No:, we’re down here for: two weeks, Fran: ↑Oh::: well you lucky gu::ys.

(I have kept the fragments as brief as possible. Let me just note that each of the “ I thought” interchanges terminates after Fran’s “Oh::::” response, each with a return to the ‘reason for the call’, each return initiated by Fran – that with Sharon by saying “Well goodness sakes it’s too bad she’s not home, ’cause

“At first I thought”

she’d sure love to come down,” and that with Ted by saying “Well Sharon said something about Stephanie coming down and...”). As laymen, just reacting to the materials, we might get a feeling that Fran’s (repeated) reference to her wrong notion is somehow searching for an account, an explanation, i.e., is proposing “Somehow I got that impression. Did I misunderstand? Or was it originally so and your plans were changed?” And we can make at least one, rather more technical, observation; that each recipient of Fran’s proffered wrong ‘thought’ denies it; Sharon with “Mmmm,” and Ted with a more elaborate “↑No:, we’re down here for: two weeks,”. That is, each recipient of the proffered wrong ‘thought’ treats it as here-andnow operating on an assertion of fact, rather than, e.g., an interesting, amusing, puzzling commentary on the recipient’s state of mind. This turns out to be a recurrent sequence: Someone proposes that X is the case, a recipient produces “I thought Y”, and the prior speaker denies the Y, sometimes supplying an account, sometimes not. Here is a series of instances of the [X is the case, “I thought Y”, denial of Y] sequence. (For easier access to the phenomenon, the transcripts have been simplified.) (18a) [Holt:2:12:1-2] Joyce: In that envelope, there’s an NHR program. Leslie: → Yes it’s for anybody who’s not got one. Joyce: → Oh. I thought perhaps you’d left yours in there inadvertently. Leslie: → No. She’s left one in, in case anybody got left out. (18b) [BAC5R:ms:33] Jessup: But my point is, that was the question to them. Course: → Alright.=I thought you meant the question ◦ here.=Ex[cuse me.]◦ Jessup: → [No. ] (0.2) Course: Go [ahead.] Jessup: → [ N o ]↓:.=So ↑therefore... (18c) [Frankel:TCI:1:26] Sheila: → Michael’s in the midst of moving this weekend. Geri: → Thought it was last weekend. Sheila: → No::, he had some: complications.h



 Gail Jefferson

(18d) Alice: Fran: Alice: Fran: Alice:

[TCI(b):16:8-9] → Well Steven’s hair’s the same color as Crai:g’s, Is it? ↑Yeh → I thought Craig’s was li:ghter.= → No I don’t think so Craig’s hair isn:’t

(18e) [Gold:3] Jane → But I couldn’t arrange that becau:se Thomas is coming again. Reva: → I thought Thomas was going away to Ohio. To schoo:l. Jane: → No: that’s not until after he graduates from high school. (18f) [Schenkein:II:226-228:R] Lori: Next time you go to Fedco, I think I got this at Fedco. Get me a bunch of them. Ben: → The ones at Fedco are different. Lori: Where did I get them. I’ll have to try to remember I → thought I got those at Fe[dco, Ben: → [No:. Lori: I guess I got ’em someplace e:lse. (18g) Alan: Nell: Alan: Nell: Alan:

[Frankel:GS:X] Still growing. It’s got buds ’n everything else on it. Oh has i:t? → Buds. No[t bugs. ] → [Oh bu:d]s. I thought you said bugs. → No. I don’t see any bugs. It might have, but I can’t see any

(18h) [Goodwin:DP:21] Beth: ’n one thing they said in the article that was really intriguing was, in the United States at this point, there are → over a hundred thousand people who are over a hundred years old. Jan: No::! Beth: M-hm? Jan: → I thought they kept track and there were only a few peo:ple. Beth: → No: there’s over a hundred thou:sand according to this article.

“At first I thought” 

(18i) [SBL:3:2:R:5] Claire: → Well they’re not comi:ng, unless they can fi:nd someb[ody. Sara: [Oh= Sara: → =I thought they were coming. Claire: → No they’re not coming unless they can get another couple. (18j) [Schenkein:II:177-8:R] Ben: Yeah but Bill they came in from Corona del Mar. Pat: No we didn’t come from Corona del Ma::r, Ben: You came in on MacArthur Boulevard Pat: Yeah MacArthur, but we hit uh: Ben: Coast Highway. Pat: Yeah. Ben: → Right. [That’s C[orona. Bill: → [That- [That’s Corona del Mar Pat: Oh is it? Bill: Right where it hits the Coast Highway Pat: → Oh I didn’t realize I thought that was already: Ben: → No. (18k) [NB:IV:11:R:1-2] Emma: I was over to see you yesterda:y but you must’v been taking a nap I rang the be:ll and then I ca:lled you: later in the eveni:ng, (0.7) Emma: I don’t know where you were m[ayGladys: [Oh: I’ll tell you I heard the → pho:ne I was watching television by the time I got out here it’d stopped ringing. Emma: → Yea:h well I let it ring about ten times I thought well now maybe you[’re in the ba:]thtu:b. Gladys: → [N o n o : , ] Gladys: → No: uh with the television o:n you know half the time you don’t hear it.

 Gail Jefferson

(18l) [Adato:2:4-5] Jay: He said the first chance::: he gets. Meaning a certain time period. Jim: → Oh no. The first of next month. Jay: → Oh. I thought you said he was going to put you in for a raise the first chance he gets. Sy: → [Uh-uh. Jim: → [No. First of the month. Jay: Oh the first of the month. Oh.

The foregoing dozen cases show a recurrent sequence: [assertion that X is the case, “I thought Y”, denial of Y]. In these sequences the “I thought” is treated as here-and-now operating on the prior assertion of fact, and not a report of some perceptual glitch by the recipient. As to the sense one might get that “I thought Y” is deployed in pursuit of an account, we can at least note that in a few of those dozen cases (18c, 18e, 18k) the denial is followed by an account. And in the following 3 cases, with or without a ‘denial’ component, there is confirmation that the materials presented as “I thought Y” are not merely reasonable, but were at one time correct, and that some unreported change had occurred. (18m) [TCII(b):38:3-4] Lasche: Did you go out there last weekend Seaton: No, (0.4) Lasche: → Oh I thought you we:re. Seaton: → I was going to I was gonna go out there this weekend too but uh: Lasche: ↑ hhuh huh-huh↑ (1.2) Seaton: I just c-c-can’t get going hhuh-hhuh (18n) [Frankel:TCI:1:8] Sheila: ·t·hhh Michael went to San Diego today, Geri: → Thought you were going. Sheila: → No, (0.3) Sheila: → I changed my mind, I have to work tonight,

“At first I thought” 

(18o) [SBL:3:R3] Claire: We’ll just have the two tables unless she’s fou:nd uh another cou:pl[e or] Sara: [Yeh.] WellSara: → Oh you I thought you said you were gonna have your neigh:↓bors. Claire: → ·tk·hhh u.-No they couldn’t co:me.

And in the following fragment, “I thought Y” is specifically being used to elicit a self-correction from a coparticipant. (18p) [SPC:10(a):14] Desk: but it’s at- on three o’clock and she might just be free or between interviews. (1.0) Mr. O: w-What time is it now sir? Desk: Three isn’t it? (0.7) Mr. O: → (We:ll?) I thought it was earlier than tha:t, (0.3) Desk: → It’s two o’clock I’m sorry. Mr. O: Yeah. Desk: I got the hour wrong.5

Given the foregoing materials, I would argue that such an assertion as “I thought” need not be, and routinely is not, just a factual report, but is operating to propose the relevance and in-principle correctness (and on occasion, as in fragment (18p), the actual correctness) of the matters being formed up as “I thought Y”. Such a possibility enriches the ‘Thought/Realized’ alternation proposed by Sacks in his Spring, 1968 lecture. Now, not merely that “...‘thought’ is used when it turns out to be wrong”, but that ‘thought’ is used when it turns out to be wrong but is being pursued as in-principle correct, reasonable, right. What, then, is going on when the ‘At First I Thought X, Then I Realized Y’ format, with its innocuous alternative to an extraordinary reality, is deployed? By asserting the in-principle correct, ordinary alternative, the whatactually-happened is shown to be odd, surprising, exceptional; to be inprinciple wrong. That is, although this thing did happen, it merely happened. It is an incidental occurrence. In principle, things like this do not happen.

 Gail Jefferson

In one of his earliest lectures, given in the Fall of 1964, Sacks touches on just that issue, by reference to how decisions are made about whether a death was suicide or not. He says (Sacks 1992, Vol. I: 62): My own feeling about such matters is that a range of decisions are made in terms of ‘odd events’ versus ‘normal events’. And odd events, by and large, are just not added together. So that if one has a notion that some X is a normal event, then the fact that occasionally or two or three times in a row something else happens, that doesn’t provide for a shift. One doesn’t now say “Well, maybe X isn’t the normal event.” But, in part perhaps by way of the fact that what is normal gets incorporated into things like proverbs and becomes very stable, odd events are just sloughed off.

And rummaging through my own ever-increasing horde of newspaper clippings and hastily scribbled notes from news broadcasts, I’ve come across two lovely instances of “At first I thought” being used explicitly to propose ‘things like this don’t happen’. In both of these cases the reported ‘first thought’ is disbelief.6 (19) [ITN 7:00 News: November 30, 1985, Prime Minister Thatcher] ((During a miner’s strike, a taxi driver in Wales is driving one of the working miners, when a concrete block is dropped onto the cab from an overpass and the driver is killed.)) Thatcher: At first I could scarcely believe it. It was murder. And I almost went to new depths of despair. That such things could happen in Britain. . . . It isn’t British. This calculated malice is alien to us. (20) [Leeuwarder Courant, November 13, 1991] A bomb attack early this morning caused heavy damage to the home of Minister of Justice Aad Kosto in the North Holland Grootschermer. “I am extraordinarily grateful to the Alkmar police for taking the bomb threat seriously. . . . In the first instance I thought: it will surely be a false alarm. I didn’t want to believe that such a thing can happen in The Netherlands”, said the Minister.

In these two cases we’re seeing the work of “at first I thought” at its most explicit – not to mention grandiose, if not just plain elephantine: Things like this don’t happen here. Now, in Sacks’ ‘Joke/Serious’ folder were some materials, again having nothing to do with joking, but a matter of ‘first thoughts’, in which people assert that their ‘first thought’ was the catastrophic actuality. What is wonderful about these is that they specifically orient to the impropriety of these as ‘first

“At first I thought” 

thoughts’, marking that Anybody would not/should not have made that of it. Here is the most explicit version, from Sacks’ folder. (21) [From The Witnesses, page 3, Yarborough] As the motorcade went down the side of Elm Street toward the railroad underpass, a rifle shot was heard by me; a loud blast, close by. I have handled firearms for fifty years, and thought immediately that it was a rifle shot.

Note that this witness marks his perception as specifically not Anybody’s by providing an account (“I have handled firearms for fifty years”). Further, he specifically marks his thought as not an eventually arrived at ‘realization’ by providing a contrast term (“immediately”). The immediacy marker conveys that what ordinarily is, and is to be arrived at as, a ‘realization’, i.e., under the burden of additional, convincing evidence of something out of the ordinary (see, e.g., fragments 3, 8, 10 and 16), was in this case not so arrived at. It happened “immediately”. And the offering of credentials explains why the ordinary, perhaps proper, procedure was not carried out. It turns out that the combination of credentials and immediacy markers is recurrently used. (22) [Notes on an item on BBC News, ca 1985] An English nightclub dancer was caught in the outbreak of the Iran-Iraq war, in a small town in Iraq where she experienced all sorts of war-related sights and sounds. She and her colleagues moved on to Baghdad, expecting it to be undisturbed, but at around 2:00 a.m., coming out of the hotel disco, they heard a huge explosion. She says, “I knew straightaway it was a bomb.”

So: In her story, she builds in credentials: She was in a place earlier where she experienced the sounds of bombs. And with “straightaway”, she marks the immediacy which contrasts with the ordinarily necessary and proper steps to knowledge of that sort. (23) [de Telegraaf, February 12, 1992, re the crash of an F16 in Henglo] Aviation hobbyist Laurens Rorink, who for the last 20 years has closely followed the comings and goings of all that flies, and witness to the air disaster during a demonstration in the German city of Ramstein, knew immediately what was going on. “I recognized the sound. Vroom, such a typical noise. Damn, I thought, that’s a crash.”

 Gail Jefferson

Again: Credentials, both of a general sort (“for the last 20 years...etc.”), and built into a description specific to this event (“...such a typical noise”.) And the immediacy marker (“immediately”) – here, supplied by the writer of the article. And although we don’t get such a marker in the witness’ statement, perhaps the exclamation “Damn” (“Verdorie”) with which he precedes his reported thought works in a similar way. So what happens if one has no credentials and is reporting a correct catastrophic first thought? You use what you do have available, i.e., the immediacy markers. Here is a dramatic and eloquent instance. (24) [From The Witnesses, page 14, Governor Connally] We had just made the turn, well, when I heard what I thought was a shot. I heard this noise, which I immediately took to be a rifle shot. . . . I immediately– the only thought that crossed my mind was that this is an assassination attempt.

Again, the “immediately”s and “the only thought that crossed my mind” convey that these are not Anybody’s Proper First Thoughts. You get a sense that they came unbidden into his mind; that Connally himself is surprised at having had these as first thoughts. By forming it up in this way, he preserves his status as a normal, reasonable, ordinary man. Indeed, it appears that routinely persons without any particular credentials to offer, will, when reporting a correct catastrophic first thought, accompany it with an immediacy marker, thus invoking the proper alternative – that this should have been a later ‘realization’. (25) [Algemeen Dagblad, April 14, 1992 re an earthquake – rare in The Netherlands] F. van Duijnhoven: “We were woken by an enormous din; everything around us was moving, it was as if something huge was rumbling by, under the house. It lasted fifteen seconds; short to be sure, but if you’re sitting in the middle of it, an eternity. I knew immediately: This is an earthquake.” (26) [de Telegraaf, October 5, 1992] ((re the crash of an El Al aircraft into a housing development)) “It must have been around a quarter to seven. We had just eaten. We heard a terrible roaring sound. Naturally we looked outside at once. I thought immediately: An airplane is crashing.”

Given the recurrence of, at least immediacy markers, if not credentials plus immediacy markers for things that are not Anybody’s Proper First Thoughts, we can return to fragment (7), the Zapruder testimony, in which he simply

“At first I thought” 

produces a ‘naked’ report of his experience of the first shot, “I heard the first shot”. Period. Interestingly, the matter of the shots is not what he is focusing upon. He is focused on the next event, the President’s reaction to the shot. And this he handles in what is emerging as the canonical fashion, i.e., with a version of “at first I thought”, in Zapruder’s testimony, “For a moment I thought...”. Canonical it seems to be. Here are a dozen or so cases of the cases gathered between the next-to last (ca 1982) and last rewrite (1992) of this paper. (27) [BBC News, September 3, 1985] ((A witness to the Glifada grenade incident)) There was a big bang and we thought – We were just having some fun round the poolside, and we thought for just a moment that it was a firecracker. (28) [ITN 9:00 News, October 8, 1985] ((Statement of a young policeman who was shot in the stomach during the Tottenham, London disturbances of Sunday, October 6th, the first time in mainland Britain that the police had encountered gunfire during a civil disturbance.)) Police Constable Patt: “First of all I thought it was a brick.” (29) [BBC News, the night Indira Ghandi was assassinated] ((Telephone interview with Peter Ustinov who, with some other people, had been waiting to meet Mrs Ghandi in her garden when the shooting started.)) Ustinov: “At first there was speculation about firecrackers.” (30) [de Volkskrant, March 1,1990] ((contributed by Martha Komter)) ((In Arnhem, February 18th, a discussion at an outdoor café got out of hand...)) ...and S. shot his former comrade dead right before the eyes of hundreds of terrace-visitors. They thought that it was part of a performance of two street artists and burst out in loud applause. Later it became evident that S. had discharged lethal shots. (31) [BBC News, May 22, 1991, rebroadcast of an item on India TV] ((re the bomb blast that killed Rajiv Ghandi)) “...Initially it was mistaken for the burst of a cracker, to welcome him.” (33) [Algemeen Dagblad, May 23, 1991] ((Bhagwan Singh, eyewitness to the bomb blast that killed Rajiv Ghandi)) “A bright flash of light and a deafening blast. Splinters from wooden crush-barriers flew all around. My first reaction was that followers of the Congress Party had ignited an oversized firework.”

 Gail Jefferson

(34) [de Telegraaf, February 12, 1992] ((the Henglo F16 crash, cf fragment 23)) “I thought first of a gas explosion”. (35) [Nederland 1, 4:00 news, February 12, 1992] ((the Henglo F16 crash)) “I didn’t believe it at first, but there you are.” (36) [de Volkskrant, February 12, 1992] ((the Henglo F16 crash)) “I didn’t believe what I had seen, it was like a film.” (37) [de Telegraaf, March 7, 1992] ((3 charred bodies were found in a field)) The bodies were found by a supervisor of the recreation facility VoornePutten. He was alerted by smoke, which he thought was produced by burning reeds. Upon closer investigation of the marshy ground, he came across the mutilated, still smoking bodies, whereupon he immediately brought in the police. (38) [de Telegraaf, October 5, 1992] ((the El Al crash)) Mrs Augustinus: “We were sitting watching Studio Sport when we saw a dark spot approaching from Diemen with a weak sound of a motor. A second later the windows on the opposite side of our block of flats were lit up orange. Originally we thought that the garage was hit. If only it were just the garage...” she said. (39) [Algemeen Dagblad, October 24, 1992] ((the El Al crash)) Ayesha Alhassan (28): “We were sitting watching TV when it happened. I, my cousin, her friend, and a sub-tenant. We were blown across the living room by the blast . I thought that there was an earthquake.”

(There is a nice contrast between fragment (25) above, with its immediacymarked correct ‘first thought’ by F. van Duijnhoven about an earthquake in his native Netherlands, a place where earthquakes are rare, and fragment (39) here, by the immigrant Ayesha Alhassan, also about an event in The Netherlands, in which “earthquake” is reported as the wrong ‘first thought’. Perhaps Ayesha Alhassan came to this land from one in which earthquakes are more or less normal occurrences.) (40) [de Telegraaf, December 1, 1992] ((A train was derailed at Hoofddorp)) When [the road-mender from Leiden, Izaak] Colpa arrived in the twilight at his workplace just across from the disaster area and heard an enormous racket, he thought at first that one of the cranes had toppled over. “When we went to take a look at what was going on, we could see the destruction. . . . Everything was wrecked. Normally, you only see such a thing on TV, don’t you?”

“At first I thought”

(41) [Leeuwarder Courant, December 22, 1992] ((the Faro air disaster)) Mrs G. Voorthuis from Augustinusga, eyewitness to the disaster: “I was on the telephone and suddenly heard a very loud bang. Because it was extremely bad weather, I thought it was a lightning strike. After that, we saw, about 200 meters away, an enormous flame, and a bit later we heard an explosion. Then you think that a tank-truck exploded.

Such materials give a strong sense of the notion that there are oriented-to and used, proper first thoughts; those which affirm the normal, Anybody’s procedures for perceiving what is going on. A sharp bang is a firecracker, smoke coming from a field is burning reeds, a painful thud against your body is a rock or brick, etc. etc. Then there are things which are oriented to as improper first thoughts; those which turn out to be the bizarre, catastrophic, extraordinary facts of the matter, i.e., those which comprise the proper ‘then I realized’ component. Thus, one is a ‘crank’ if one produces such a report as, “At first I thought it was a shot, then I realized it was a firecracker.” One is a crank – or someone who has been exposed to a drastically altered reality. Such a case was witnessed by my colleague Judy Davidson years ago in Hawaii. A small group of American soldiers on leave from the fighting in Vietnam were walking along a Honolulu street on a day that happened to be Chinese New Year, when a celebratory bunch of firecrackers went off. To a man, the soldiers hit the ground. But this is an extreme case. The reporting of such first thoughts – not to mention such overt behavior – is heavily constrained and negatively sanctioned. I witnessed something similar, yet significantly different, by a pair of ordinary New York City-dwellers. Ten or maybe fifteen years ago on a visit to New York, I was walking down West 86th Street with two old friends of mine, longtime residents of New York, whom I thought ought to get to know each other, since they lived only a block or two apart. There was a bang. I heard it as a backfire and kept walking. These two people, who had never met before, simultaneously ducked. Now, the fact that they simultaneously ducked is similar to the extreme case of the Vietnam G.I.s. The fact that they only ducked, that they checked themselves and did not hit the ground, is significantly different; testimony, perhaps, to the power of the constraint against being a ‘crank’.7 Earlier we glimpsed that sort of constraint in the credentials and immediacy markers of fragments (21)–(26). For another sort of glimpse we can start off with a news clipping sent to me by Anita Pomerantz.





Gail Jefferson

(42) [Oxford Times, March 19, 1982] UFO Reports Stream In Mysterious purple lights were seen moving across the sky last Friday evening to the amazement of witnesses. Mr Derek Mansell, of Crown Road, Wheatley, said he saw a large red light steadily moving across the sky above his home. The light suddenly shot upwards and disappeared. “I thought it was an aircraft at first,” said Mr Mansell who is UFO research officer for Contact International UK, “but an aircraft could never have shot upwards like that so quickly.” People from Oxford, Maidenhead and Cirencester saw the lights and contacted the police and Mr Mansell [who] has been cataloguing UFOs since 1964.

For nearly 20 years, this man’s work has been the sighting of UFOs. Surely he would be entitled to a credentials-plus-immediacy-marker report. But perhaps his was the optimal choice when it comes to the phenomenon of UFOs, i.e., 20 years of UFO experience might well point towards rather than away from the ‘crank’ possibility.8 Alright, then, what about that Secret Service Agent, in fragment (8), the driver of the Presidential limousine, William R. Greer, who reported that he thought the gunshot was “a backfire”? Well, maybe he’s more to be thought of as a chauffeur than a steely-eyed, professionally paranoid Secret Service Agent. I more or less held onto that image of Greer until I read Bonar Menninger’s Mortal Error (1992: Appendix A, pp. 297–375.). One of the appendices in the Menninger book is a series of spoken and written statements to the Warren Commission by the Secret Service Agents accompanying the President. Quite a few contain an “at first I thought”: (43) [Warren Report, Testimony of Clinton J. Hill, SS, p. 305] ((Hill was riding in the follow-car, behind the Presidential limousine. His assignment was to pay special attention to Mrs Kennedy.) Spector: Now, as the motorcade proceeded at that point, tell us what happened. Hill: Well, as we came out of the curve, and began to straighten up, I was viewing the area which looked to be a park. There were people scattered throughout the entire park. And I heard → a noise from my right rear, which to me seemed to be a firecracker. I immediately looked to my right, and, in doing so, my eyes had to cross the Presidential limousine and I saw President Kennedy grab at himself and lurch forward and to the left. . . . I jumped from the car, realizing that something → was wrong, ran to the Presidential limousine.

“At first I thought”

(44) [Warren Report, Written Statement, William R. Greer, SS, p. 320] ((Greer was driving the Presidential limousine)) I was looking at the overpass that we were about to pass under in case → someone was on top of it, when I heard what I thought was the backfire of a motorcycle behind the President’s automobile. After the second shot, I glanced over my right shoulder and saw Governor Connally start to fall. → I knew then that something was wrong and I immediately pushed the accelerator to the floor...

Here we have an indication that Greer was not merely a chauffeur, but a working SS-man, alert to possible trouble. Maybe Hill and Greer should be drummed out of the Secret Service as hopelessly naïve. Or maybe what they’ve done is to produce a powerful display of their retaining the normal perceptions and reactions despite the paranoiagenic nature of their work. Those guys in the Secret Service? They’re just like you and me. Another SS-man’s statement is produced with a bit more professional distance: (45) [Warren Report, Written Statement, Roy H. Kellerman, SS, p. 322] ((Kellerman was riding next to Agent Greer in the Presidential limousine)) We were still traveling at the normal rate of speed from 12 to 15 miles per → hour when I heard a noise, similar to a firecracker, exploding in the area to the rear of the car, about 12:30 p.m.

The statements in the following 3 fragments provide a sort of middle ground between ‘ordinary man’ and ‘professional’. They do not formulate the thing as a ‘first thought’, but on the other hand, in contrast to Agent Kellerman’s disengaged statement in fragment (45), they do give an experiential description – albeit with the self-reference elided: (46) [Warren Report, Written Statement, John D. Ready, SS, p. 343] ((Ready was standing on the right front-door running board of the followcar)) At about 12:30 p.m. we began the approach to the Thornton Freeway traveling about 20–25 mph in a slight incline. I was about 25–30 feet from President Kennedy who was located in the right rear seat. I heard what → appeared to be firecrackers going off from my position. I immediately turned to my rear trying to locate the source.



 Gail Jefferson

(47) [Warren Report, Written Statement, Glen A. Bennet, SS, p. 353] ((Bennett was riding in the right rear seat of the follow-up car)) [A]bout 12:25 P.M., the Motorcade entered an intersection and then proceeded down a grade. At this point the well-wishers numbered but a few; the motorcade continued down this grade en route to the Trade Mart. At this point I heard what sounded like a fire-cracker. I immediately looked [away]from the right/crowd/physical area/ and looked towards the President... (48) [Warren Report, Written Statement, George W. Hickey Jr., SS, p. 355] ((Hickey was riding in the left rear seat of the follow-car)) Just prior to the shooting the Presidential car turned left at the intersection and started down an incline toward an underpass followed by [the followcar]. After a very short distance I heard a loud report which sounded like a firecracker. It appeared to come from the right and rear and seemed to me to be at ground level. I stood up and looked to my right and rear in an attempt to identify it.

By ‘elided self-reference’, I’m pointing to such phrases as “what appeared to be...” (fragment 46), “(what/which) sounded like...” (fragments 47 and 48); the elision being the absence of “...to me”, i.e., “what appeared to me to be”...”, “which sounded to me like...” (cf. Hickey’s “...and seemed to me to be at ground level” in fragment 48). It’s this referring to what one made of an occurrence without explicitly identifying oneself as the interpreter that places those reports in a middle ground between Anybody’s spontaneous “I thought it was X”, and Kellerman’s studied “It was similar to an X”.9 Across these materials we are seeing, in stronger or weaker forms, the positing of ‘normal’ in the face of the extraordinary. In a 1971 lecture, Sacks did a consideration which bears on this issue – by reference to, of all things, teenage dating practices, and for a different sort of device; ‘We were going to [X], but [1, 2...etc.] so we [Y]’ (Sacks 1992, Vol. I: 455–457).10 About ‘We were going to [X]’, Sacks says, “I raise the question, why put in a rejected alternative?” In answering that question he uses a phrase which resonates with a phrase of Karl Menninger’s that Sacks refers to in one of his earliest lectures (Sacks 1992, Vol. I: 66): “loyalty to reality”. In his 1971 lecture, Sacks speaks of the naming of the rejected alternative as showing one’s “commitment to the normal”. It’s a lovely phrase, and it’s obviously relevant to the UFO-spotter’s and Secret Service Agents’ handling of their encounters with the extraordinary. It

“At first I thought”

may also be an underlying issue for whomsoever, since we are always in the business of exhibiting our “commitment to the normal”, and devices like ‘We were going to [X]’ and ‘At first I thought [X]’, are resources for doing that. An overview of the Sacks materials I’ve so far referred to reveals that across time and disparate ‘topics’ (suicide, verbs, hijackings, teenage sex), in bits and pieces, here and there, are the elements of a unified analysis. In his answer to his own question raised by a bit of data to do with teenage dating practices – Why put in a rejected alternative? – Sacks characterizes it as a matter of using what we know to be “normal” as a way to “specifically locate what happened here as distinctly unusual”. Forget about the specific topic. This is an abstract, generalizable notion. Take the 1964 discussion of determining which deaths are or are not suicide (page 146, above). Forget suicide, there is an abstract, general notion about ‘odd events’ versus ‘normal events’, where odd events don’t count, are “just sloughed off ”, do not “provide for a shift” in how things are to be perceived or managed. Now, put these two together, i.e., that naming a rejected alternative provides that what happened here is distinctly unusual, plus that unusual, odd events don’t count. With those two notions we can come to see, technically, what can be involved in Sacks’ 1970 proposal that one aspect of ‘being ordinary’ is the achieving of “the ‘nothing happened’ sense of really catastrophic events” (page 137, above). It appears that ordinary people, wheresoever we find them, will search for some normalizing alternative to the extraordinary actuality, whereby they can both exhibit their “commitment to the normal”, and provide that, in principle, things like this don’t happen; that what actually happened merely happened, is an incidental occurrence. No “shift” has occurred or need occur; we can go about our business as usual. And this normalizing device, reinvented time after time by all the Anybodys who have been involved in something extraordinary, is, time after time, preserved in the media and sent out to do its work among everybody else. The sheer tenacity of the device may be seen in a program note for a BBC 2 documentary, “The Day the Sun Blowed Up”, sent to me by Pomerantz back in 1982. The ‘first thought’ in question occurred 37 years earlier:11 (49) [TV guide program note, BBC 2, August 6, 1982] Yesterday’s Witness in America The Day the Sun Blowed Up Narrated by James Cameron



 Gail Jefferson

→

On 16 July 1945 at 5:30 in the morning, the world’s first atom bomb was experimentally exploded in the desert of New Mexico. This is the story of the fateful days of secret preparation for the test. It is told by some of the scientists and soldiers who were intimately concerned . . . The story is also told by some of the local inhabitants for whom the birth of the nuclear age came as a total and alarming surprise. “I heard the explosion and thought something had blowed up in the yard out in front of the store. I went out and there was a man just standing there looking kind of dumbfounded and I asked him, ‘What blowed up out here?’ And he said, ‘Look over yonder,’ (looking to the east of us) ‘the sun blowed up!”’

Postscript I stopped adding new data to my presentations of the “At first I thought...” phenomenon in 1992, but couldn’t resist sticking clippings and hastily jotted notes into the “At first I thought...” folder as further cases cropped up. Here is a chronologically-ordered sampling of the next decade’s materials. (50) [caught in passing, BBC News, late October, 1993(?)] In troubles-torn northern Ireland, men in black clothing with Balaklava masks break into a house, run past a young girl, and shoot her brother. She reports that her first thought was that they were friends of her brother’s, dressed for a Halloween party. (51) [Leeuwarder Courant, December 8, 1994] ((A young woman passenger on the cruise ship Achille Lauro that burned and sank, kept a diary)) While the two young women were dancing in the discothek, the Achille Lauro caught fire. Shirley van Haaster wrote: “Suddenly one of the pursers began to run to the exit. I thought that there was a fight. Until I got near the door and a cloud of smoke drifted in.”

The following fragment resembles fragment (30), in which the reported first thoughts of witnesses to a shooting is that it is part of a performance. Here, the reported first thoughts are those of people not merely witnessing, but overtaken by, the event. (52) [Leeuwarder Courant, March 21, 1994] Earthquake didn’t belong in the Oscar show

“At first I thought”

→

Los Angeles (AP) – The earthquake yesterday in Los Angeles was taken as part of the show during a rehearsal for the Oscar-presentations. Just as a mock-up of a dinosaur handed over the envelope for the winner of the Oscar for ‘special effects’, the lamps began to swing and the seats in the hall to shake. The public treated it as a successful part of the celebration until several technicians raced to the podium and screamed “earthquake!”. There was no damage, and the rehearsal continued after a short break.

Earlier, noting a contrast between fragment (25) with its immediacy-marked correct ‘first thought’ about an earthquake in The Netherlands, a place where earthquakes are rare, and fragment (39), also about an event in The Netherlands, in which earthquake is reported as the wrong ‘first thought’, I remarked that perhaps Ayesha Alhassan, the woman reporting the latter, came to this land from one in which earthquakes are more or less normal occurrences. And indeed, we can find people accounting for their wrong ‘first thought’ having been of an earthquake by announcing that they’re from California (see, for example, fragment 57 below). But in fragment (52) above, and in the following fragment, people involved in a California earthquake are reported to have had wrong ‘first thoughts’. In (52), it appears that the standard and correct account (earthquake) was superceded by the possibility of Hollywood special effects made relevant by the fact of the rehearsal in progress. In the following fragment it appears that the standard and correct account was superceded by possibilities made relevant by the fact that the young couple involved were on a freeway overpass in a moving car when the earthquake struck. (53) [Caught in passing: CNN January 18, 1994] ((A young man, passenger in the car driven by his girlfriend when the L.A. earthquake of January 17th struck. They were on the Highway 14 overpass onto Highway 5, as it started bucking prior to its collapse.)) “...she thought it was a high wind, I thought initially it was a blow-out...”

The following fragment, and several others in this series (59, 60, 63, 66, and 68), have a similar character to the anecdotes of the Vietnam G.I.s in Hawaii and the pair of New York City-dwellers walking down West 86th Street, reacting to firecrackers and backfire as gunshots (page 151). (54) [Leeuwarder Courant, December 10, 1994] Peres fall causes consternation in Norway



 Gail Jefferson

→

Oslo (Reuter) – The Israeli minister of foreign affairs Shimon Peres stumbled over a trolley track yesterday evening during a walk through the inner city of the Norwegian capitol Oslo. The minister took a hard fall and suffered a wound near his eye. The fall led to great consternation among the massed security agents. Peres and the Israeli Prime Minister Jitzak Rabin – together in Oslo with PLO leader Jasser Arafat to receive the Nobel Peace Prize – were walking to their hotel after attending a service in the capitol’s synagogue. The Jewish sabbath forbade them from making use of the secret service automobiles. At the moment the Israeli minister fell, members of the eight-hundred man security service shouted “Stand still, don’t move”. The area was immediately searched for possible attackers. Bodyguards helped Peres to his feet. The minister had nothing worse than a bleeding wound near his right eye.

The following fragment is nice for the issue of ‘commitment to the normal’. Of all people, the Dutch painter Rob Scholte would not consider himself an ordinary man. He’s a creative, even a maverick. But when it comes to telling about the day four months earlier, when he lost his legs to a bomb planted in his car, he’s an Anybody. (55) [de Volkskrant, February 11, 1995] ((An interview with the painter Rob Scholte, four months after a bomb planted in his car exploded.)) “When I drove out of the parking spot, I heard three short ticks.” He ignored them. He’d hardly turned the corner when something happened that he describes as “a gray, yellow light, it was an implosion”. ... → “It didn’t occur to me that it was a bomb. I thought that the garage had done something wrong. The car hadn’t been driven for five days.” (56) [de Telegraaf, April 20, 1995] Bomb attack Oklahoma Bewildered office personnel were able to save dear life by ducking under a desk or a table, which protected them from flying glass and collapsing ceilings. → “I thought that it was an earthquake”, said an official of the Department of Agriculture. “I had just ducked under the desk when glass from the windows clattered all around me.” (57) [NRC Handlesblad, April 20, 1995] ((re the Oklahoma auto bomb))

“At first I thought” 

“I thought that an earthquake had taken place. I’m used to that because I come from California”, said another survivor. (58) [de Telegraaf, September 25, 1995] Boy (16) murders eleven people in France Toulon (AFP, DPA) – A sixteen year old boy murdered eleven people and wounded eight others...in the village of Solliès-Pont and the nearby town of Cuers, north of the southern French city of Toulon. ... [During the night, he killed several of his family members in Soliès-Pont by beating their heads in with a hammer and a cudgel.] After that he went 6 kilometers further to Cuers, where, early in the morning he began to randomly shoot at people on the street. ... → A resident of Cuers said that he thought at first that the boy was igniting fireworks, then that he was shooting at pigeons. “Then we saw a man with a wound in his leg.”

I would add the following case of an El Al commander whose aircraft began to fall apart, to that of the Vietnam G.I.s in Hawaii (page 151), who hit the ground when firecrackers went off. (59) [de Volkskrant, January 27, 1996] ((A review, including a moment-by-moment account taken from cockpit recordings, of the crash of an El Al Boeing 747 into an apartment building in the Bijlmer on October 4, 1992.)) ...[I]n the cockpit of the Boeing the explosions on the right hand side of the aircraft were heard. At the same moment, the craft made several wild, swerving movements. I’m taking over!” cried Fuchs to his co-pilot. While he struggled to regain control of his aircraft, the captain wondered what → could have happened. His first thought was that his craft was hit by a rocket, fired by Palestinian terrorists.

And how about residents of Sarajevo, not so long ago a war-torn shambles, now a site for making films about that war. The following is a polar opposite to the case cited by Harvey Sacks, where the wrong ‘first thought’ was that a filming was in progress (page 137). Here, that’s what is in fact going on. (60) [Friesch Dagblad, June 29, 1996] “They’re only shooting a film...”

 Gail Jefferson

→

Sarajevo – What is burning in the city center? Why is there a tank with Serbian markings next to the Presidential building? Who put up the barricades again near the former front line? Cars were being stopped by police agents. Residents of Sarajevo look around them, concerned and wondering. “They’re starting to shoot again”, asks a passerby of an agent. “Yes”, answers the policeman, “a film”. Sarajevo has become a popular location for making films. Bosnian and French directors have already filmed two movies there this year. This time [it’s] an American crew...

On the other hand, “some” residents of south-east Drenthe in the Netherlands, who “thought that a bomb had fallen” when in fact an earthquake had occurred, come off as perhaps just a bit crankish in the newspaper article which reports those ‘thoughts’. In this particular article, the ‘bomb’ possibility (arrow a) is succeeded by one which, in comparison not only to the reality of the earthquake, but now to the ‘first thought’ of “some” people, is normalcy itself (arrow b). (61) [Algemeen Dagblad, February 21, 1997] In the Drenthe village Roswinkel fear was everywhere after Wednesday evening’s earthquake. The KNMI registered 3.4 on the Richter Scale. Never before was a quake of that magnitude measured. The blame is being placed on the Dutch Oil Company. The extraction of gas in that area is considered to be the cause of the quake. ... Many concerned telephone calls were received by the police on Wednesday (a) → evening. “About a hundred people phoned. Some thought that a bomb had fallen.”, according to Bert Peters of the police in south-east Drenthe. ... (b) → Eddy Venema from Ter Apel, a village five kilometers further away, didn’t know what had happened to him. “My first thought was that the back part of my building had collapsed.”

And the following fragment can be added to those cases in which credentials are not cited, but an ‘immediacy marker’ is produced (e.g., fragments 24, 25, and 26).

“At first I thought”

(62) [de Volkskrant, March 22, 1997] Tel Aviv – On the spacious terrace of the popular café A Propos on Ben Gurian Street, a bomb exploded. The huge parasols slammed to the ground, glass flew though the air. ... [A woman helping clear the damage is interviewed] Because she lived nearby, → she heard the explosion and at once ran outside. “I knew immediately that it was an attack”, she said angrily.

And here is yet another case similar to the Vietnam G.I.s (page 151), the Oslo security agents, the El Al commander and the Sarajevo residents of fragments (54), (59) and (60), above. (63) [Leeuwarder Courant, February 6, 1997] Children with firecrackers cause panic Jerusalem (AFP) – A small group of children setting off firecrackers was the cause of great consternation yesterday in a shopping center in the northern Israel harbor city of Haifa. The shopping public thought at first → that it was an attack, and became panic stricken. A large contingent of police immediately closed off the shopping center until the source of the ‘explosions’ became clear.

And the following fragment joins ranks with the shooting which was applauded as street theater (fragment 30), and the earthquake in the midst of an academy awards rehearsal which was initially taken to be part of the show (fragment 52). (64) [Leeuwarder Courant, March 16, 1998] Singer dies during performance London (ANP/DPA) – The Reggae star Judge Dread died Saturday as the result of a heart attack that he suffered during a performance in the British Canterbury. The approximately fifty year old singer, whose real name is → Alex Hughes, collapsed at the end of the show. Many in the audience thought that his collapse was part of the performance.

Note, by the way, that in fragment (61) it is reported that “some” people thought an earthquake was a bomb, while in fragment (64) it is reported that “many” thought a singer’s collapse was part of the performance, while in fragment (63), the report has an entire population responding in identical fashion, “The shopping public thought at first that it was an attack...”. The point being that how any of these formulations measure the actual segment of each of the relevant populations is utterly obscure; their work seems to be directed to con-



 Gail Jefferson

veying the character of the response; crankish, as conveyed by fragment (61)’s “some”, or, say, reasonable under the circumstances as conveyed by fragment (64)’s “many”, not to mention fragment (63)’s global characterization, “the shopping public” – to which we might add the characterizations “to a man” in the VietNam G.I.s anecdote and the reference to the New York City-dwellers ducking as “simultaneous” (page 151). (65) [de Telegraaf, December 8, 1999] Dismay among students after shooting tragedy in Veghel schoolroom At the Veghel Leigraaf College feelings are dominated by incomprehension. After all, according to the school director R. Martinoh, the suspected shooter, the seventeen-year old Ali D., was not known as an aggressive boy. The director was sitting upstairs in a meeting at around two o’clock in the afternoon when he heard a series of loud reports. → “I thought at first that someone was throwing firecrackers”, declared a visibly shaken Martinoh. “I immediately ran downstairs. But at the bottom of the staircase, I came across the first victim. In the hall and in the computer area I found the other casualties bleeding on the ground. It was terrible, what I saw . . . ”.

The following fragment, involving yet another earthquake, might also be added to the growing sub-corpus of ‘first thoughts’ that under different circumstances would be crankish, not to mention paranoid, i.e., the Vietnam G.I.s. and the New York City-dwellers of page 151, and the Oslo security agents, El Al commander, and Sarajevo and Jerusalem residents of fragments (54), (59), (60) and (63); as well, perhaps, as the Tel Aviv resident of fragment (62) with her ‘immediacy’-marked report. (66) [Algemeen Dagblad, January 29, 2001] Victims first thought of a Pakistani bomb Mahendra Thakker accompanied politicians who visited the disaster area. → He told Kinhsuk Nag of The Times of India: “When the earthquake started, we took it for an enormous bomb explosion. We thought that Pakistan was making trouble on the Day of the Republic that we were celebrating.” Bhu lies on the border with Pakistan.

Having arrived at the year 2001 in my heap of cases, I found one striking absence. I had nothing at all from September 11th. I mentioned that to Gene Lerner, wondering if it was a feature of the event itself, or perhaps how it was covered. Whatever the reason for this gap, Lerner consulted internet, and

“At first I thought” 

emailed back some cases he’d found. Here is just one – I include this particular one because it resembles my own experience with the event.12 (67) [Marietta Times, September 15, 2001] I was in the shower that morning and my wife told me about what hap→ pened,” Hugh Hopper, 66, of Marietta, said. ”At first I thought it was just a small plane that accidentally hit the building and I told my wife, ’Don’t let it bother you. Those buildings can’t come down.”’

A half a year later, Pomerantz sent me an e-mail with yet another member of the Vietnam G.I.s in Hawaii, et al., sub-set. This, now, involving a New Yorker. (68) [E-mail from Anita Pomerantz, March 11, 2002] TV coverage of a fire at a tire dump – big flames in the background. Some → young guy was interviewed. They only had one line of his: “I thought it was a terrorist.” What does that say about normal these days?

And finally, the two most recent additions; in 2002, the Washington, D.C. snipers (at the time that the article excerpted in fragment (69) appeared, still being treated as a single “shooter”), and in 2003, the space shuttle Columbia disaster. (69) [Algemeen Dagblad, October 21, 2002] Shooter for the first time active in the weekend The uncapturable sharpshooter who has terrorized the Washington, D.C. area for almost three weeks, struck for the first time on a weekend. A 37year old man was hit in the stomach on Saturday evening on the parking lot of a highway restaurant, 140 kilometers south of the American capitol. He was taken to a hospital in critical condition. The attack took place in front of the Ponderosa Steakhouse, where the → victim and his wife had been eating. “I heard a bang like an automobile exhaust sometimes makes. My husband was able to take three more steps before he collapsed”, according to the deeply shocked wife. (70) [CNN, February 1, 2003] ((Texas witness to the space shuttle Columbia disaster)) “I live by the railroad tracks, and at first I thought a train had blown up.”

Just one more. Although I hadn’t planned on adding any further cases after the book went into the final stages of preparation, this one was just too rich to ignore:

 Gail Jefferson

(71) [Algemeen Dagblad, January 14, 2004] ((At lunch break, a high school student walks into the school’s crowded cafeteria and shoots a teacher in the head)) ‘The master lay in a pool of blood’ → Den Haag – They heard a bang and thought that rowdies (rotjongens) were at it again with fireworks. When youngsters began screaming and running, they looked around. To their astonishment they saw [vice-principal and economics teacher] Hans van Wieren lying on the ground. Around his head, a pool of blood. “A teacher shouted ‘Call 112, call 112!’ But I was too nervous to grab my mobile phone”, says Fatoush Benkalid, a pupil at Terra College, who was witness to the tragedy in the school cantine. Van Wieren was shot in the head from close range. According to witnesses, by Murat, a 17-year old pupil in the vocational middle school. “Before the → bang I heard people behind me shouting ‘Joke, joke’. I think everyone thought that the pistol was fake”, says [Fatoush’s] friend Mimoush Handi.

And so on...

Notes * Between 1970 and 1995 I presented versions of this paper any number of times. I’ve never wanted to publish it; for one, the matter of authorship is problematic – Harvey Sacks had so much to do with it, And secondly, it became a sort of pet. I didn’t want to let go of it; just kept piling up instances over the years. But this collection is the obvious place for it to appear – if anything is a First Generation paper, this one, with its materials spanning almost four decades, is. . Note that the account of the policeman’s response to the call is built into the report, i.e., the emergency number is “new” – perhaps some people have been treating it as a new toy, placing joking calls for help. Furthermore, the account is pre-positioned and thus comes off as a ‘description’, as ‘setting the scene’, rather than, e.g., a ‘defense’. . While in fragment (4), that a facility is “new” serves as a possible account of a wrong ‘first thought’, here it may be that this particular wrong ‘first thought’, about a car towed away by the police, would not have been reported in the paper at all had it not been tied to something newsworthy (about which there is no reported ‘first thought’), i.e., the introduction of “new” higher fines. . A noticeable difference between Sacks’ clippings and mine is that, although he now lived and worked in Los Angeles, he subscribed to the New York Times (see fragments 2, 3, 4, 6, and 9), while I contented myself with the local rag, the Los Angeles Times. . The wrong ‘first thought’ in this case is, in a sense, doubly accounted for. In the first place, Samuel, the recipient of the voice of the Lord is a child, who couldn’t be expected to know any better. Secondly, we get a similar sort of pre-positioned account as that of fragment

“At first I thought” 

(4). Here, we are specifically advised that this was a time when such occurrences were rare (and thus not even the adult, Eli, would be prepared for it): “And the word of the Lord was precious in those days; there was no open vision.” (Had my first encounter with the text been the version which appears above, I would not have known what to make of it . As it happens, I first encountered it in a Dutch bible; roughly translated: “In those days the word of the Lord was scarce; there were not many visions.”, which I took to be a reference to the rarity of such apparitions. I recently checked with the American Standard Version of the Holy Bible, and got “...there was no frequent vision.”, and the Artscroll’s Stone Edition of the Tanach sent to me by Jenny Mandelbaum, has it as “vision (i.e., prophecy) was not widespread.” And incidentally, the Artscroll version has, not that “Eli perceived...”, but the verb discussed by Sacks (page 6), i.e., that “Eli realized...”). . These materials begin to resonate with the ‘fishing device’ considered by Anita Pomerantz (1980:186–198). Roughly, such things as accounts can be elicited from a coparticipant, by reporting one’s own experience of the coparticipant’s circumstances. For example, “Your line was busy” routinely gets such responses as “Oh I was talking to so-and-so”. This, in contrast to, e.g., the ‘correction invitation device’ (Sacks 1992, Vol. I: 21–23, 380–381), where such information is elicited from a coparticipant by making a guess, e.g., “Were you talking to Larry?”, “No, I was talking to...”. . Fragments taken from the newspapers de Volkskrant, Algemeen Dagblad, de Telegraaf, and Leeuwarder Courant, and from the television broadcast of Nederland 1, are roughly translated from the Dutch. . One of these two friends of mine was the late Jim Schenkein, in many ways a maverick, but his instantaneous response on that occasion was identical to that of my other friend, a far more conventionally-oriented person. . In his book The UFO Experience: A Scientific Inquiry (1972: 15), J. Allen Hynek reports “...my work with UFO reporters of high caliber indicate (sic) that they wish to see or to explain their observations in terms of the familiar. A typical statement is: ‘At first I thought it might be an accident up ahead on the road – the lights looked something like flasher beacons on squad cars. Then I realized that the lights were too high, and then I thought maybe it was an airplane in trouble coming in for a crash landing with power off, since I didn’t hear any sound. Then I realized it was no aircraft.”’ For Hynek, “this process of going from the simple, quick description and explanation, step by step, to the realization that no conventional description would suffice (escalation of hypotheses)” argues powerfully that the many UFO reporters who employ it are not cranks (p. 15). As it happens, 17 of the 70 relevant fragments in this paper resonate with Hynek’s “process of going from the simple, quick description and explanation, step by step, to the realization that no conventional description would suffice”, to which he gives the elegant title, “escalation of hypotheses”. Specifically, we find as a recurrent – although far from defining feature of the ‘at first I thought’ phenomenon, reference to what it was that occasioned the shift from ‘ordinary first thought’ to ‘extraordinary realization’ (See fragments 3, 7, 8, 10, 13, 16, 37, 40, 41, 42, 43, 44, 51, 52, 58, 63, and 65).Only two of those cases, however, converge with Hynek’s “typical statement”, with its series of escalating possibilities – fragment 9 (a multi-stepped report by a witness of an air disaster; step 1 “I thought it was a lightning strike”, step 2, “Then you think that a tank-truck exploded.”, with no mention of how she came to realize that an airliner had crashed) and

 Gail Jefferson

fragment 15 (a multi-stepped report by a witness to a series of random shootings; step 1, “[he] said that he thought at first that the boy was igniting fireworks”, step 2, “then that he was shooting at pigeons.”, and step 3, “ ‘Then we saw a man with a wound in his leg.’ ”) In the rest, we find reference to a single, decisive feature of the situation; that feature – perhaps recognizably, relevantly for production purposes – being adequate for realization. Even Mr. Derek Mansell, UFO research officer for Contact International UK, who might well qualify as one of Hynek’s “UFO reporters of high caliber”, offers, with his “I thought it was an aircraft at first...but an aircraft could never have shot upwards like that so quickly.”, a one-step, single-feature-adequate/decisive move from the ordinary ‘first thought’ to the extraordinary ‘realization’ (see fragment 42). (Of course it’s possible that in the various newspaper articles, the serial character of the arrival at realization has been edited out – which itself would be interesting; that would mean that across time and in various cultures news personnel take it that one step is adequate/decisive. But then, that would mean that their readership accepts one step as adequate, which would suggest that the persons making the quoted statements – presumably members of the population for which the articles are written – did offer the reported one-step adequate/decisive accounts.) . I just want to note, without knowing what to make of it, that in fragments (44), (46) and (47), Agents Greer, Ready and Bennett provide immediacy markers – in each case, not for a perception, but for an action: “I immediately pushed the accelerator to the floor”, “I immediately turned to my rear trying to locate the source”, and “I immediately looked [away] from the right/crowd/physical area/ and looked towards the President”, respectively. . From the Kennedy assassination to teenage dating practices! From the monumental to the miniscule. But, given teenagers’ attentiveness to normalcy, it’s not surprising that in the teenage-talk Sacks was examining, such a device showed up. (In this case, the device was used for a change from the usual smooching venue, the guy’s car, to a guest-house in back of the girl’s family home.) . As with the case of Samuel and Eli in fragment (15), what we have here is a double; an ‘ordinary’ first thought by one person, followed by an extraordinary next by someone else. However, in this case the extraordinary next goes beyond the facts of the matter, and resonates with the anecdote about the Vietnam G.I.s on leave in Hawaii (page 151) in the sense that the man who provides the extraordinary next has just been confronted with an altered reality, the Atomic Age. . My reaction to the earliest computer graphics on CNN – at about 10:20 a.m. New York time, 3:20 p.m. over here – was that the scale was altogether wrong. They were showing some huge thing that would have to have been maybe a Boeing 747, but the early descriptions of the aircraft were of a “private plane”, which I took to be something like a little Cessna.

References Hynek, Allen J. (1972). The UFO Experience: A Scientific Inquiry. New York: Ballentine. Menninger, Bonar (1992). Mortal Error. New York: St. Martin’s.

“At first I thought” 

Pomerantz, Anita (1980). “Telling my side: ‘limited access’ as a ‘fishing’ device”. In Sociological Inquiry, 50 (3–4). Sacks, Harvey (1992). Lectures on Conversation, Volumes I and II, G. Jefferson (Ed.). Cambridge, MA: Blackwell.

P III

Sequencing actions

Pre-announcement sequences in conversation Alene Kiku Terasaki

It has been widely observed that, in their design, languages show a concern for marking the distinction between information in an utterance which can be considered ‘new’ for this discourse and information which can be treated as ‘given’ (Chafe 1970; Halliday 1967; Haviland & Clark 1975). Information marked as ‘given’ in an utterance is information presented by the speaker as available to the recipient situationally or in prior talk while ‘new’ information is marked as not so available. Thus one of the sources of the designation of some information as either given or new resides in its prior history in this particular conversation.1 References to a previously introduced item will be constrained to be done as references to ‘given’ information – i.e., information which has already been mentioned in the discourse. Conversely, the presentation of some piece of information as ‘new’ for the discourse marks it as not having been previously mentioned.2 The initial realm of the operation of the given/new distinction, its area of applicability, then, is the single conversation or discourse. Although the general observation has been made that the information structure of an individual utterance organizes its parts in terms of their relation to the preceding discourse, the question of whether the distinction organizes segments of discourse larger than the single utterance has not been investigated in detail. That is, on the cross-utterance level, there is the question of what will count as ‘new-for-the-discourse’ information and subsequently there is the question of whether the given/new distinction can be found to organize series of utterances across speakers into unitary segments. Over the past several years, work in the Sequential Analysis of Conversation by Sacks, Schegloff and others has investigated various aspects of the organization of speaker utterances across serial turns at talk. A major focus of the

 Alene Kiku Terasaki

enterprise has been the organization of speaker turns and the utterances they contain into sequentially-ordered units or ‘sequences’ of discourse. The unitary character of such sequences has been demonstrated in a series of papers, most notably, Sacks et al. (1974) and Schegloff (1968, 1972). It is argued there that the serial adjacency of specifiable utterances such as Question/Answer to one another across speaker turns reflects an independent organization of the adjacent items into a unit discriminable from surrounding turns and internally organized as consecutively occasioned utterances. A property of the relation between consecutive utterances across turns was formulated to specify that ‘sequential’ organization: In order to use the term ‘sequence’ in strong fashion – to refer not merely to ‘subsequent occurrence’ in the sense of successive positions of the hands of a clock, but rather to a specifically sequential organization – a property called conditional relevance was proposed to hold between the parts of the sequence unit. When one utterance (A) is conditionally relevant on another (S), then the occurrence of S provides for the relevance of the occurrence of A. If A occurs, it occurs (i.e., is produced and heard) as ‘responsive to’ S, i.e., in a serial or sequenced relation to it; and, if it does not occur, its non-occurrence is an event, i.e., it is not only non-occurring (as is each member of an indefinitely extendable list of possible occurrences), it is absent, or ‘officially’ or ‘notably’ absent. That it is an event can be seen not only from its ‘noticeability,’ but from its use as legitimate and recognizable grounds for a set of inferences (e.g., about the participant who failed to produce it). (Schegloff 1972: 76)

As an illustration we cite the following fragment of conversation in which the absence of a response to the utterance in line l occasions a remark directed to its absence: (1) [CN II-11:6]3 l A: Well they, ((silent whisper)) had gotten married. (2.5) 2 A: Yuh heard me. (0.5) 3 B: Hitched?!?

The parts of a sequence, its consecutive turns, are thus in a relationship of conditional relevance to one another. The minimal conversational sequence is comprised then of two such related turns at talk in which the occurrence of the first makes the occurrence of the second ‘expectable.’ Members of the culture recognize and use a range of such sequences in their achievement of everyday

Pre-announcement sequences

activities. Greetings, farewells, informings, corrections, etc. are all sequenced activities. A major class of utterance sequences which has been investigated is a set of two-utterance sequences termed Adjacency Pairs.4 The name describes twoturn pairs in which the turns are adjacently placed and produced by alternate speakers. The occurrence of an Adjacency Pair ‘first pair part’ makes the occurrence of one of a set of discriminatively related ‘second pair parts’ expectable. For example, an offer and its acceptance or rejection; a greeting and a return greeting; a request and a grant or denial; a question and its answer. The nonoccurrence of one of these related second pair parts can warrant a finding of their absence and thus a set of possible complaints; to wit, that one is ‘evading,’ had not heard, is not present to hear, etc. Relatedly, the turn adjacent to a first pair part may be taken up with items concerned to account for its nonproduction in the adjacent turn (Schegloff 1972: 76). Thus the orientation of speakers to the unitary character of such sequences (as well as those comprised of more extended series of turns) is discernible in a variety of ways separate from their otherwise routine appearance as adjacently placed turns. Our question earlier was whether the given/new distinction could be found to operate in the organization of a series of utterances across speakers. Sequence organization does make this distinction. Reports on sequences concerned specifically with the transmission of new information and its reception (by Goldberg (1975) on the transmission of instructions and by Sacks (1974) on story telling in conversation) indicate that sequence organization reflects a sensitivity to marking information to be conveyed or in its conveyance as given or new in the serial presentation of the utterances. Pomerantz’ (1975) work on Assessments as adjacency pair second pair parts to Informings in conversation has suggested some additional integration of given/new marking and sequence structures. She argues that recipient Assessments are the conditionally relevant second pair parts to deliveries of news and notes that the Assessment utterance regularly preserves the referent of the prior item in a pronominalization of it (1975: 33ff.). For example (indicated by brackets): (2) [JG AB:2] l A: .. and there’s a good chance that they might be hiring her. 2 B: Oh! [That’s] good.

The use of the pronominal form marks the news as now ‘given’ for the discourse and coincides with the sequence-unit boundary:



 Alene Kiku Terasaki

l 2

A: B:

Informing Assessment

(Adjacency Pair First Pair Part) (Adjacency Pair Second Pair Part)

It appears then that the introduction of information as ‘new’ for the conversation is followed by its subsequent treatment as now ‘given’ and that this marking can intersect with sequential structures in the talk. We propose to undertake an investigation of this intersection with sequential structures through an analysis of a conversational sequence specific to the transmission of new information: Announcements of news in conversation. Clearly we are not attempting to capture in our description all instances of the introduction of new information in conversation. Rather, we focus on instances of speech events5 characterizable as Announcements of news. These are treated as specifically sequential phenomena, embedded in the organization of conversation.

A sequential characterization of announcements of news Informings in conversation occur in a variety of sequential units: as Announcements, in inquiries, in requests for news, as stories, etc. With the term ‘Announcements’ we have in mind those conversational events in which occurrences are reported as ‘announceable news’ – frequently under the auspices of ‘good’ or ‘bad’ news. These can range from births, deaths, and marriages to new jobs, or victories at the golf course. (Please see Appendix A for illustrative examples.) While there may be externally specifiable criteria for the selection of some event as announceable, we avoid an entirely content-based definition of Announcements because it appears that features of the design and placement of the item in the overall structure of the conversation contribute as much to its recognition as do content considerations. Items which appear to otherwise qualify as – ‘announceable’ in a content-based definition are regularly done and not treated as Announcements. For example: (3) [NB:–2] l B: 2 3 4 5 A:

So, Elizabeth’n Will were s’poze tuh come down las’night but [there was death ‘n the fam’ly] so they couldn’ come so Guy’s asked Dan tuh play with the comp’ny deal, so I guess he c’n play with’im. So, Oh good.

Pre-announcement sequences

In this example the news of the death (indicated by brackets) is not remarked on, while the news that the golf game will take place is received as assessable news. Our suggestion is that a major factor in the recognition of Announcements by speakers is to some degree independent of the content of the events they report and resides, instead, in the organization of their presentation in the talk. Thus, Announcements should be differentiated from talk about occurrences which might otherwise appear to be announceable but can be shown to have been ‘buried’ in their presentation. There are additionally instances of talk in which a recipient treats some talk as news-to-them which were not marked by the deliverer as Announcements. It is not criterial to the corpus of Announcements that speakers know in fact that their intended recipients have not heard the impending news. Parties to the talk as well as analysts can locate the utterance as a possible Announcement independent of whether it is known that the recipients have not heard. Our argument is that the recognition of some utterance as a possible Announcement resides in a dual set of features it displays: (1) its syntactic format, and (2) its sequential features. It is well known in linguistics that the same sentential form can participate in a wide variety of distinct speech acts. Conversely, speech acts select a syntactic format for their production. Announcements regularly occur in a highly attenuated simplex (declarative) sentence form in conjunction with features of design and placement in the conversation which mark them as instances of a news delivery. We have characterized Announcements as ‘reports’ done in the course of particular conversations. The occasion of the Announcement provides for a set of complementary identities [Deliverer (D) and Recipient (R)]6 which organizes the parties to the talk and provides orderly sets of procedures for the delivery and reception of the news across a series of turns at talk.7 The turn series have regular components that occupy characterizable turns in which the minimal turn set is composed of two turns: 1) The Announcement of the news by Deliverer. 2) The Assessment of the news by Recipient. For example: (4) [JG NC:1] l D: Guess what=I haven’t had a drink for eight days now. 2 R: Fan-tas-tic!



 Alene Kiku Terasaki

Just as the intending Deliverer has specifiable delivery procedures, the intended Recipient has reception procedures which provide for showing that the ‘news’ is or is not ‘news-for-them’ and that it has been received and appreciated. In the case of Announcements, Assessments of the news’ relevance for the Deliverer and for the Recipient regularly occur in the turn directly following the Announcement. They operate there to display the Recipient’s ratification of the Announcement as news-for-them and to display their understanding of it as ‘good’ or ‘bad’ news. Assessments are thus one regularly occurring ‘second pair part’ to Announcements which, together with Announcements, comprise a minimal Announcement Sequence. Announcements regularly occur over turn series larger than the two-turn pair in the serial presentation of the news. These ‘expansions’ of the pair also contain specifiable components. Expanded Announcement Sequences retain the initial pair components and include two optional Adjacency Pairs: 1. Turn sets prior to the Announcement turn: Pre-Announcement Sequences 2. Turn sets between the Announcement turn and its Assessment: Insertion Sequences8 The turn series are thus organized into a specifiable Announcement Sequence which has as its minimal form the Adjacency Pair: Announcement/Assessment and which can include regular expansions in the form of Pre-Sequences and Insertion Sequences which occur prior to and between the main components. Following is an instance of such an Announcement Sequence with the components labeled: (5) [JG 2:1] Pre-announcement 1st pair pt.

D:

Pre-announcement 2nd pair pt. Announcement #1 Insert Seq. 1st pr. pt. Insert Seq. 2nd pr. pt. Assessment of #1 Announcement #2 Insert Seq. 1st pr. pt. Insert Seq. 2nd pr. pt. Assessment of #2

R: D: R: D: R: D: R: D: R:

I forgot t’tell y’the two best things thet happen’tuh me t’day. Oh super=What were they. I gotta B plus on my math test. On yer final? Uh huh? Oh that’s wonderful! hh And I got athletic award. REALLY?!? Uh huh. From Sports Club. Oh that’s terrific Ronald.

Pre-announcement sequences 

These findings have been formulated from an initial examination of a sample of 85 extended Announcement Sequences taken from transcripts of actual audio- and video-recorded conversations. They are abstractly summarized by the following list of features: 1. 2. 3. 4.

Announcements in conversation occupy turn series. The series of turns are adjacently placed. The series are organized into adjacency pairs. The adjacency pairs, moreover, are related to one another as Pre-Sequences and as Insertion Sequences to the Announcement Sequence. 5. The Announcements regularly take assessments as their conditionally relevant second pair parts. We are arguing that sentences occurring in natural conversation which are otherwise linguistically characterizable as informationally ‘new’ are not isolated occurrences but occur in the environment of other sequentially-related utterances as part of a discourse unit that organizes serial utterances across speakers.

News deliveries as conversational artifacts An operative principle in the construction of utterances in conversation is the principle of recipient design.9 Utterances are designed for just this recipient, taking account of the relevancies that obtain between just this speaker and this recipient. As a particularizing operation in such things as word selection, topic selection, sequence ordering, etc., it provides for the variability of design in conversation which is generally glossed as ‘context sensitivity.’ In the instance of ‘informings’ in conversation, recipient design is manifested in an overriding preference not to report things already known to one’s recipients. That requires any intending Deliverer of news to make some determination of the character of their information as news-for-this-scene. While the beginning of a possible news delivery may be done, the news itself may not be ultimately produced as ‘news’ if the potential Recipients are found to have ‘already heard.’ The ‘news-ness’ of an item appears to be a matter collaboratively arrived at across a news delivery.10 Its status as ‘new’ information is not predetermined but is proposed and either ratified or rejected in the interaction. We find that the determination of whether some piece of information has already been heard is central to its assertability as news. Its character as news is negotiated and achieved across the telling by the parties to the talk. That places the phenomenon of news introduction into the scene of its telling as the prod-

 Alene Kiku Terasaki

uct of local search procedures. Viewed as a locally-managed determination, the presentation of news conversationally is thus locked into the scenes of its use. Our interest here is to examine how the organization of the delivery and reception of Announcements as a particular type of Informing Sequence reflects such concerns.11 We can now specify our assertion that the preference not to tell things already known is implicated in news deliveries in terms of the organization of Announcement Sequences. Turns surrounding the Announcement turn give major design emphasis to marking the upcoming turn as containing news and thereafter to ratifying its proposed news-ness. The preference not to retell organizes large segments of the sequence structure as well as the design of individual turns within the sequence. Using the feature list after example (5) as an overall framework, this section presents findings regarding two major components of these Sequences: (1) the organization of Pre- sequences to Announcements and (2) the design of the Announcement turn. How the preference not to re-tell is manifested in the design of these two components is outlined and suggestions are made regarding issues of linguistic interest which turn up in the course of the sequential analysis. We propose that a consideration of Announcement Sequence structures will have a payoff for both linguistics and conversation analysis. Where syntactic and semantic concepts may be illuminating in a consideration of the design of utterances in the Announcement turn, the selection of Announcement Sequences as a bounded domain of inquiry into the operation of information structures12 may permit linguistic findings of a more general nature in regard to the operation of given/new structures in actual conversation. Our description is done under the auspices of an interest in the sequential organization of that structure and in it we make use of such terms as ‘First,’ ‘Second,’ ‘prior’ and ‘next.’ In encountering these terms we ask the reader to keep in mind that our frame of reference is the serial ordering of speaker turns and that it organizes their use. Our use of the terms ‘First’ and ‘Second’ refers to the parts of the Adjacency Pair organization outlined above. Note that any particular Adjacency Pair First Pair Part may additionally comprise the initial turn of a Sequence and, concomitantly, that any particular Second Pair Part may comprise the final turn of a Sequence. (Later in our discussion there is a further treatment of the issue of sequence boundaries.) In the next section we describe the internal structure of Pre-Announcement Sequences and its relationship to Announcement turn design.

Pre-announcement sequences 

Pre- sequences Schegloff ’s (1968) characterization of Summons/Answer sequences has detailed the operation of two-utterance/turn sequences which have the character of prefaces or preludes to some projected next action or sequence. Rather than being internally complete units, these Pre- Sequences are integrally tied to the ensuing actions as their preliminaries. Summons/Answer sequences such as the following are produced as anticipatory of some next activity, here, the initial exchange of a conversation they preface. (6) [BS 2,1:80] A: Jim? B: Yeah? (7) [KR: l] A: Mo::m. B: What. (8) [TRIO 1:1] A: ((phone rings)) B: Hello. (9) [SN4:l] A: ((knock knock knock)) B: Come in:::.

The recipients are summoned into availability for further talk and their answer establishes their availability. As Pre- sequence types, Summons/Answer sequences are concerned with providing for coordinated entry into the talk they preface. In establishing the availability of the parties for further talk and thereby warranting the occurrence of that talk, Summons/Answer sequences share with other Pre- sequence types the feature of contingently pre-cursing some next action. The occurrence of the projected sequence is made contingent on the outcome of the Pre- sequence.13 If one party is found not available on the occurrence of the Summons, the proposed further talk is delayed or averted. Similarly, Pre-offers like the following make the production of the offer dependent on the response to the Pre- sequence First:

 Alene Kiku Terasaki

(10) [BS 2,1:107] B: PRE-OFFER 1st PRE-OFFER 2nd

A: B: A:

=I’m gunna buy a thermometer though because I think she’s // (got a temperature). We have a thermometer. Yih do? Wanna use it?

Responses that solicit the next action will produce that next action while other responses can prevent its production.14 In the earlier discussion of Adjacency Pair types, it was noted that the selection of the Second Pair Part of the sequence could be made from among the members of more than one category of type-related Seconds. For instance, invitations can take acceptances or refusals, compliments may be accepted or declined, etc. The selection of one alternative over another has marked consequences for extensions of the sequence into the ensuing talk. A preferential ordering of one alternative over the other appears to operate quite generally in the selection of such Seconds. Acceptances of offers are largely preferred to refusals, for instance, while compliments are preferably declined rather than accepted (Pomerantz 1975: 112ff.). Both the initiator of the sequence and their recipient have procedures available to them for averting the production of the non-preferred alternative. The procedures themselves are sequentially organized and built to operate as expansions either prior to or inside the projected sequence.15 Pre- sequence initiations are a major first-speaker technique for discovering in advance whether the production of the projected sequence’s First Pair Part is warranted. If the response to the Pre-invitation First, for example, indicates that the projected invitation will be refused, the invitation will not be produced: (11) [HS ST1:1] PRE-INVITATION 1st PRE-INVITATION 2nd REPORT OF INTENDED INVITATION16

A: B: A:

Say what’r you doing. Well, we’re going out. Why. Oh, I was just gonna say come out and come over here an’ talk this evening, but if you’re going out you can’t very well do that.

In Announcement sequences, a category of Seconds alternative to Assessments involves the marking of the Announcement’s news as ‘already heard’ or as ‘not surprising.’ For example:

Pre-announcement sequences

(12) [AT ST1:1] R: How are ya?17 D: Oh I don’ know. I don’t feel good. I had trouble with my stomach. I had pains all day. R: We all do. (13) [J&G:24] D: Jim’s, very sick. R: Yea:h I went over there’n talked to him today.

Between the two alternatives, Assessments (or ‘News Marks’ such as ‘Really?’) occur much more frequently than ‘not news’ responses. As with other Adjacency Pair types, there is a preferential ordering in Announcement sequence Seconds which locates ‘I-know’ responses as dispreferred alternatives. Various aspects of the design of the Announcement turn and of Pre-Announcement sequences attend that dispreference.

A Pre-Announcement type A major class of Pre-Announcement objects involves the presentation of some news to come without therein providing that news.18 We refer here to instances of items such as: (14) [BS 2,1:73] D: hh Oh guess what. R: What. (15) [KC 4:2] D: Hey we got good news. R: What’s the good news. (16) [JSLR:38] D: Hey I got sump’n thet’s wi::1d.. R: What.

As Pre-Announcement types they display features in common with other Presequences which have projective uses for the Sequence they preface. In particular they share components with another Pre- sequence type: Prefaces to Stories told in conversation. Sacks (1974: 340) has discussed the sequentiallyrelevant features of these components in his consideration of story prefaces. Pre-Announcement sequences of the type under consideration typically dis-



 Alene Kiku Terasaki

play some or all of these components In the design of the Pre-Announcement First. Briefly, they are: 1. A naming of the projected sequence, e.g.,: “Hey we got good [news]. 2. A characterization or evaluation of it, e.g.: “Hey we got [good] news. 3. Some reference to the recency of the event reported, e.g.: I forgot t’tell you the two best things thet happen’t’me [t’day]” 4. An offer or request to tell, e.g.: “[Y’wanna know] who I got stoned with a few w(h)eeks ago?hh!” As components directed to the production of the projected sequence, each feature of the Pre- sequence item can be found to provide an antecedent for features of that projected sequence, either in the form of its presentation or in its elicitation. The offer to tell and the naming of the projected sequence establish the relevance of the occurrence of not just a single next turn (e.g., an acceptance or rejection of the offer) but project a next-next action, the Announcement of the news. The item named and offered in the Pre-Announcement is made projectively available in the next-next turn contingent on an acceptance of the offer to tell. The characterization of the news-to-come has projective uses in the upcoming sequence in that it marks the Deliverer’s own evaluation of the news. Not all announceables are equivalently good or bad news for the local parties and the determination of the news import is a matter much attended to in the turns surrounding the Assessment. Deliverer’s evaluation of the news thereby provides the Recipient with materials to locate the sort of Assessment the Deliverer seeks at the culmination of the Announcement. Thus Assessments are regularly matched to the characterization provided in the Pre- sequence and can be found to be the basis of agreement and disagreement structures subsequent to the Announcement (Pomerantz 1975: 66ff.). Finally, though a formulation of the issue is more complex than this would suggest, an apparent constraint on the delivery of announceable news is that it be told on the first opportunity to tell.19 Thus the time reference may serve to warrant the tellabillty of the news as reporting events which have happened since the last interaction of the Deliverer and Recipient. We can thus characterize the components of the Pre- sequence as ‘sequentially relevant’ in that they are directed to some management of projected expectable next actions in the proposed Announcement sequence. We argued earlier that the occurrence of the Announcement as a ‘fact’ of any particular conversation’s history is a locally-accomplished production made up out of known-in-common procedures for the delivery of possible news and its ratification as ‘in fact’ news. We argued further that the finding that the proposed

Pre-announcement sequences 

news is already known to the intended Recipients is a dispreferred alternative and that that dispreference organizes various aspects of the Announcement sequence design. Our task here is to attempt to demonstrate the ways in which that dispreference is manifested in sequentially-relevant features of the PreAnnouncement sequence. The next section of the paper is directed to that issue.

Design of the Pre-Announcement First Pair Part In this section we examine three aspects of the design of Pre-Announcement sequence components which attend the preference not to re-tell known news. We consider in turn: (1) Deliverer’s methods for preserving the preference not to re-tell, (2) Already Informed parties’ methods, and (3) Recipient’s methods for the recognition of possibly known news.

(1) The design of Pre-Announcement Firsts is regularly aimed at determining in advance of an Announcement whether its news has been heard. Sudnow (1967: 37) provides a particularly fitted instance of this concern in his discussion of death announcements among hospital staff: The announcement of a death from one shift member to another can and does occur in the course of an ordinary greeting conversation, and on these wards, where deaths are not so much announced as they are mentioned, their mention does not noticeably inhibit ordinary conversation. When a death occurs in an unexpected place within the hospital, or when deaths occur in rather unusual circumstances, news spreads quickly and the conversation about death is much more dramatically attenuated. On one occasion, a diabetic woman died in childbirth, a relatively infrequent happening, and by the time a nurse arrived on the OB ward for the evening shift, she had already heard of the morning’s death. She was greeted by a daytime nurse as she approached the station with, “Have you heard?” and answered, “Yes, Mrs. B. stopped me in the hall downstairs and told me,” whereupon a conversation was entered about “what happened”....

Another such instance comes from the opening of a taped conversation: (17) [KC 4:2] D: How are you all? R: Oh very very well. D: Good. D: Hey we got good news! R: I know.

 Alene Kiku Terasaki

Additionally: (18) [HS ST:l] D: Didju hear the terrible news? R: No, what. (19) [JSLR:2l6) D: Didju hear about that guy who got-tho-tho- that family in New Mex//ico ( ) R: hhh Oh::: Oh::: that mercury ( ) poisoning?

The Pre-Announcement First is produced in the form of a question which asks whether the intended Recipient has heard while not naming the news. The doing of the Pre- as a question raises the possibility that the news has already been heard in a way which allows the intended Recipient to preserve it as a ‘correct’ possibility in their next turn, as in example (19) and in Sudnow’s example. We contrast this instance to the following in which the intended Recipient produces the dispreferred ‘I know’ which elicits surprise from the intending Deliverer: (20) [KC 4:2) D: Hey we got good news! R: I know. D: Oh ya do?

Where the Pre-Announcement did not directly incorporate into its design a concern for the possible known-ness of the news, the intending Deliverer was ‘caught out,’ found to have been ‘in error.’ Deliverer’s concern not to be found to have told a piece of known news may also manifest itself in an expansion of the Pre-Announcement First’s turn components to include some formulation of its possible known-ness as in the following example (indicated by brackets): (21) [JSLR:41) D: Toni en Bill I have something tuh tell you. [You prob’ly heard about it. -already but just in case you haven’t.]

In this instance, Deliverer’s proprietary concern for her Recipients is produced as the warrant for delivering information she otherwise marks as probably already known.20 The following is another instance of a Pre-Announcement First directed to locating the intended Recipient’s state of knowledge about the news. In it, a

Pre-announcement sequences

caller has made a request for a rather routine police inquiry in the midst of a major emergency situation. As a preliminary to his refusal to grant the request on the grounds that an emergency exists and despite the fact that the caller has made no indication that he is aware of the emergency, the dispatcher produces the following Pre-Announcement First: (22) [PD:l6 #33) D: You know what’s happening at the Fairgrounds then. R: Uh, what? ((as in ‘didn’t hear’)) O: D’you know what’s happening at the Fairgrounds then? R: No

The preference not to tell known news manifests itself in an apparent preference to, as Sacks has called it, ‘under-tell and over-suppose.’ 21 The initial item is constructed to prefer an agreement to its presupposition that the caller knows. The ‘non-hearing’ subsequently elicits a revision of the initial item into one which does not suppose the caller knows but nonetheless preserves the possibility that he may, while not producing the news itself. In those instances where the Pre-announced news is found to be known, the production of the projected Announcement is suspended and further talk on the matter is done in ways which treat the item as ‘given:’ (23) [JSLR:217] PRE- 1st D: Didju hear about that guy who got-tho-tho that family in New Mex//ico ( ) PRE- 2nd R1: .hhh Oh::: Oh:::that mercury ( ) poisoning? PRE- 2nd R2: Oh thee, you mean in thee “heat? // in thee grain? R1: ((gasp)) PRE- 2nd R3: The mercury poison//ing? Assessment R2: Oh isn’t that//terrible? Ri: ((gasp)) Agreement D: Wasn’t that tragic?

Pomerantz (1975: 13) notes that “speaker’s source and extent of knowledge of the relevant referents generally bear on assessment productions.” In the prior example, the precursed news is marked as known by several intended Recipients in the Pre-Announcement Second position. The Assessment which is done is constructed as one which solicits agreement to it as a mutually-known-about and assessable item. The character of the talk about the item is rapidly altered to reflect a common state of knowledge about the proposed news.



 Alene Kiku Terasaki

The determination of an item’s status as ‘in fact’ news is thus apparently a constraint on the design of its earliest presentation in the talk. Responses which indicate that the proposed news is known avert the production of the Announcement structure we have outlined and markedly shift the character of the talk done under the auspices of ‘news’ to talk done about a known-in-common event. The character of the Pre-Announcement First as projective of talk well beyond its immediate occurrence is used and oriented to by the parties.

(2) A second feature of the organization of Pre-Announcement sequences which attends their projective aspects involves the operation of Pre- sequences as entry devices. In Summons/Answer sequences, the occurrence of the Summons provides for the relevance of some business to come, business which will be presented by the initiator of the Summons. Jefferson has argued that Summones using the format [Name] (often with rising intonation, e.g., “Jim?”) propose in their occurrence that the Summoner has some business and is not simply calling the other into ‘casual’ talk.22 Instances of its use in call openings regularly do not precede such things as “How are you?” but instead preface mono-topical calls whose single business is such things as invitations, requests, inquiries, orders, proposals, etc. Thus in such Opening sequences, the Summoner is provided with an identity for the projected sequence as the party who has possible business and the Answerer with an identity as the intended recipient of that business. This permits us to talk in some more detail about terms we have up to now used in an unexamined way to refer to the parties to Announcement sequences. We have spoken of intending Deliverer and proposed Recipient in Pre-Announcements as a set of identities local to and generated by the occurrence of the Pre-Announcement First. Another category of local identity is possible: that of parties Already Informed (AI) of the news. In a multiparty conversation, the presence of parties who have already heard the news or who have been told prior to the arrival of some uninformed parties presents a difficulty in preserving the preference not to re-tell. Actual announcement deliveries often display this as a concern by marking the news as a retelling for some of those present, as in the following example: (24) [CC 2:8] D: [All of you are not aware of the fact] thet we have engaged a consultant.=[Some of you d-kno:w, ’n others don’t.]

Pre-announcement sequences 

The Pre-Announcement turn may be utilized to achieve similar things. It is used as a place where the proper identities are ‘discovered’ for the projected sequence. The character of the Pre-Announcement First as containing a set of projective identities is attended to and utilized by Already Informed parties to mark their status vis-à-vis the projected sequence. How they can be expected to talk to the impending news is apparently a matter of concern. The production of the Pre-Announcement First is often taken over by the Already Informed party or collaboratively produced in conjunction with them. (While still reserving the actual delivery of the news to the Deliverer.) They thereby mark the relevancy of the Announcement’s occurrence for their next actions as not the same as for Recipients. For example: (25) [J&G:4] PRE- 1st

D:

PRE- 2nd COLLABORATIVE ANNOUNCEMENT

R: AI: D:

You’d prob’ly turn green with envy if I tolja about what I got ( ) first quarter. Whadju get first //quarter? Tell ’em about it. Four A’s in two B’s.

R:

What’s new Jim,

AI: D: R:

Oh? – He’s a little richer. Yeh I’m a li’l richer, tss! Richer. (2.0) Whaddiyuh mean. Think we’re gonna getta rai:se, first a ‘nex’ month.

(26) [Adato:2,l5] INQUIRY COLLABORATIVE PRE- 1st AGREEMENT BY D PRE- 2nd

ANNOUNCEMENT

R: D:

Collaborative work by Already Informed’s provides for the tell-ability of the news even though for certain of those present it will be a re-telling. In jointly producing the Pre-announced news, Already Informed’s permit the preservation of the preference not to re-tell.

(3) Finally, we find that intended Recipients appear to display attention to producing a recognition of the news at the earliest possible point. That is, they appear to utilize features of the Pre-Announcement First to recognize the impending news as known to them and to thereby avert its delivery. We contrast this to the possibility that they may have waited until the news is actually delivered to, say, ‘make certain’ that this is the same news they have already heard.

 Alene Kiku Terasaki

Of the eight instances we have of news found to have been known, the three which are recognized after the Announcement turn are not Pre-announced and the remaining five are recognized in the Pre-Announcement Sequence. They are listed here: (27) [JSLF:2l6] D: Didju hear about thee, pottery en lead poisoning // ( ) R: Yeah, Evie wz just telling us // ( ) D: I read en article en I ca- in one a’the- I d’know whether it it wz Newseek’r Time’r what. I think ih wz Time’r Health. (28) [JSLR:2l7] D: Didju hear about that guy who got-tho-tho that family in New Mex//ico ( ) R1: .hhh Oh::: Oh::: that mercury ( ) posoning? : : R2: Oh thee, you mean in thee wheat? // in thee grain? R1: ((gasp)) R3: The mercury poison//ing? R2: Oh isn’t that // terrible? ((whispered)) R1: ((gasp)) D: Wasn’ that tragic? [[ ( ): ( ) R4: Yeah R1: ◦ Oh, God,◦ R2: No y- I think you oughta- maybe chill that a li’l bit (honey)? (29) [AT FT:l]23 D: Oh. You know, ((lowers voice)) Yuri did a terrible thing. R: hhh! I know. D: You know? She committed suicide. R: That thing was staged. D: What- if she died then it won’t have been a put on. (1.5) R: She died? (30) [Sudnow, cited instance] D: Have you heard? R: Yes, Mrs. B. stopped me in the hall downstairs and told

Pre-announcement sequences 

(31) [KC 4:2] ((Note there are two D’s and two R’s)) D: How are you all? R: Oh very- very well. D: Good. D: Hey we got good news. R: I know. [[ R2: What’s the good news. D: Oh ya do? [[ D2: Ya heard it? D: Oh good. R2: Oh yeah, mm hmm. R: Except I don’ know what a giant fullicullar lympho-blastoma is. D: Who the hell does, exc//ept a doctor. R: Well R2: Mm R: (I d’n) D2: This is nice did you make this? ((accomplishes shift in topic))

In each instance the production of the projected Announcement is suspended and further talk is done in ways which treat the items as ‘given’ (except in (29) where the possible distinction between a staged suicide and an actual death is raised). That brings us, however, to a puzzling matter. The proposed Recipients of the news apparently call on design features of the Pre-Announcement First to determine whether they ‘in fact’ know or can guess the proposed news to come. Yet the features of the Pre-Announcement First appear to provide few materials in themselves for locating what among the possible pieces of news they have heard is here being referred to, e.g., “Have you heard?” It appears that Recipients utilize ethnographic/biographic aspects of the scene in conjunction with the Pre-Announcement components we have outlined above to come to that determination. Sudnow suggests that news’ status as a report of a ‘non-ordinary’ event for any particular ethnographic setting is marked in its presentation via the use of a Pre-Announcement First (what he refers to as a ‘presentational format’).24 The delivery of some item with a Pre-Announcement narrows its possible referents to some category of events which, for that setting, are ‘non-ordinary.’ Thus, in a hospital setting where deaths are a routine feature of the organization’s

 Alene Kiku Terasaki

events, not all deaths are announceable matters. If the intended Recipient has recently heard some news which is a candidate member of such a category, ‘announceable event,’ the occurrence of a Pre-Announcement First provides them with the means for locating the news it prefaces as possibly the same news they have heard.25 We argued earlier that Pre-Announcement Firsts are designed to withhold their news while precursing it. The extent to which the news is specified appears to cover a range going from the most unspecified, as in: “Have you heard?” to quite specified, as in: “Y’wanna know who I got stoned with a couple weeks ago?” And, of course, the more specified the news-to-come has been in the First, the more recognizable it is to the intended Recipients. In instances (27) and (28) a good deal of identificatory information is present in the Pre-Announcement. The ‘subject’ of the news, its locale and participants are presented. In example (29) the case relations26 of the projected Announcement are preserved in the pro-termed verb phrase of the Pre- such that along with the characterization of the news as ‘terrible’ the intended Recipients can find what the Pre- projects. The use of the characterizing term ‘terrible’ appears, furthermore, to have uses particular to death Announcements. Of the three instances of its use which we have found, each prefaces news of a death. Its recognizability as premonitoring death news as a matter known in common to speakers is evidenced more strongly by the fact that in the two additional instances we have, both result in guesses that the impending news is of a death:27 (32) [HS ST:l] D: Didju hear the [terrible] news? R: No. What. D: Y’know your Grandpa Bill’s brother Dan? R: He died. D: Yeah. (33) [DA:4] D: I-I-I had something- [terrible] t’tell you. So // uh R: How terrible is it. D: Oh, th- as worse it could be. (0.8) R: W- y’mean Edna? D: Uh yah. R: What she do, die? D: Mm:hm,

Pre-announcement sequences

Also in example (31), “Hey we got good news”, the characterization appears to locate the news which is prefaced. Certain ethnographic features are brought to bear on its usability here. The news concerns the finding that a mutual friend does not, after all, have cancer. Deliverer has just arrived at the home of Recipient. It was known that on the way over, Deliverer would stop at the hospital to visit the mutual friend and that news of the biopsy results would most probably be available. In the interim, Recipient had found out for herself what the results were and could locate in the characterization of the impending delivery as ‘good’ that it referenced the same news she knows and otherwise expects her guest to deliver. Various components of Pre-Announcement First design such as naming and characterization of the news are thus utilized by intended Recipients to recognize the news as known. There are, however, concomitant features of PreAnnouncement Firsts which provide for Recipient’s recognition that are not located in their design features but in their placement in the overall organization of the conversation. Parties to a conversation use and orient to the history of its topical progression. New topics emerge from prior topics and are presented as taking account of their subsequence to their priors, sometimes as having been warranted by the occurrence of that prior topic. Announcement sequences regularly occur in topic-initial position in a conversation – that is, as the initial segments of new topic talk. They thereby inherit the constraints of that position. Instance (28) of the recognized Announcements occurs in a position subsequent to the prior attempted Announcement of instance (27) regarding the lead poisoning. The talk is uniformly on-topic to the lead poisoning up to the point where the Pre-Announcement First of (28) is done. While it is clearly possible for parties to shift the talk at such a point to matters entirely separate from those of the prior talk (as is the case after (28) and (31)), it is regularly the case that new topics locate the just prior talk as their assertable antecedent. In this case, then, the recognition of the Pre-Announcement in (28) as referencing the news of an instance of poisoning (here, by mercury) is locatable to the parties in part in terms of its possible ‘topicality’ for the conversation-so-far. There is an additional feature of placement shared by some of these instances. In example (30), the arriving nurse is described as being ‘greeted with’ the Pre-Announcement First. The sequence is presented as the initial exchange of the conversation.28 Examples (29) and (31) occur quite early in the conversation, on the completion of a single ‘How are you?’ sequence. In each instance the Pre-announced item emerges as the ‘first topic’ of the conversation.



 Alene Kiku Terasaki

Sacks (1992) has detailed the operation of ‘first topic’ as used by parties to a conversation. He argues that the opportunity to introduce a ‘first topic’ is a matter routinely provided for in the organization of openings to conversation.29 The first routine opportunity to introduce a first topic for the talk for some opening types is in the answer position to the question, “How are you?” and the second opportunity occurs on the completion of a ‘pass’ answer to that question, such as “Fine.” First topic position is a generic locale of many Announcement types. Certain classes of announceables are treated as things which can warrant a call directed solely to their delivery or require that they be told ‘on a first opportunity to tell.’ Their delivery in first topic position marks them as important news. In our data, instance (30) is deported as preempting even an exchange of greetings. Instances (29) and (31) occur in the second routine position for first topic introduction. Thus the placement of the items can be found to locate the sort of news they preface in ways which aid in their recognition as Pre-Announcements to news known by the Recipients. Finally, the placement of such Pre-Announcement Firsts early on in the conversation provides another source of their recognizability as references to known news. Goffman (1967: 71) notes that the early parts of occasions are taken up with talk which reports news-which-has-happened-since-we-lastmet. For those who see one another regularly, the presentation of a Presequenced piece of news contains a built-in time reference. Locating the news which the Pre-Announcement First references is thus bounded by a knownin-common unit of time which permits the proposed Recipient of the news to find the thing referred to as having occurred, say, ‘since yesterday.’ Our treatment of this problem here should be viewed as only suggestive of the directions a subsequent account might take. We currently lack enough instances in the data collection we have to make any strong claims. Furthermore, our argument in regard to first topic position is related to the problem of the placement of Announcements in the overall organization of a conversation. In order to substantiate our claim that the Pre-Announcement First’s location early on is indeed relevant to its recognizability we need a much more detailed specification of the placement issue. Pre-Announcement Firsts, then, provide for the recognizability of the news they preface in their placement in the conversation and in the design of their presentation without requiring a presentation of the news itself. If the news is not recognizable to the intended Recipient, its delivery will follow; if it is recognizable, the preference not to re-tell will have been preserved. The strength of the Pre-Announcement First as projective of the upcoming talk and thus

Pre-announcement sequences 

organizing of central features of that talk appears to necessitate moves which will shift the basis of the talk from ‘new’ to ‘given’ as early as possible. The preference not to tell known news emerges then as not simply a matter of avoiding ‘informational redundancy’ but as having highly localized import for the immediate exigencies of the speakers’ problem, ‘how to talk next.’ To this point our concern has been to examine the operation of the preference not to re-tell as it was manifested in the design and use of sequentiallyrelevant components of the Pre-Announcement First Pair Part. We have not given attention to the question of how any particular utterance is recognized as an instance of a Pre-Announcement First, nor have we considered how its design organizes the shape of the projected sequence. Aspects of the structural organization of the Pre-Announcement Sequence and their relationship to the design of the Announcement turn itself are considered next.

The first three turns The following is a listing of instances of Pre-Announcement Sequences of the type we will characterize. (Speakers are designated ‘D’ for Deliverer, ‘R’ for Recipient, and ‘AI’ for Already Informed.) Again they are taken from transcripts of actual audio- or video-recorded conversations: (34) [KC 4:2] D: Hey we got good news. R: What’s the good news. (35) [JSLR: 38] D: Hey I got sump’n thet’s wi::ld. R: What. (36) [BS 2,1:73] D: Oh guess what. R: What. (37) [J&G:2] D: Y’wanna know who I got stoned with a few w(h)eeks ago? hh! R: Who. (38) [JG 1:1] D: I forgot t’tell you the two best things thet happen’t’me t’day. R: Oh super, what were they.

 Alene Kiku Terasaki

(39) [JG 2:1] D: I got two good things tuhday. R: You di::d. (40) [BS2,1:83] D: Y’know what he wannid me tuh do? R: What. (41) [JSLR:l31] D: Hey you’ll never guess whacher Dad is lookih- is lookin’ at. R: What’re you lookin at. (42) [J&G:3] D: Y’d prob’ly turn green with envy if I tolja all about what I got ( ) first quarter. R: Whadju get first quarter. (43) [GTS 5:43] AI: Well I know that Jim, for example, is facing some things this week which are uh ((pause)) a little more serious. ((pause)) -aren’t they, regarding school. D: Yeh. R: What’s happening. (44) [DA:4] D: I- Uh:: I did wanna tell you, en’ I didn’ wanna tell you, uh:: uh last night. Uh because you had entert- uh, company. I-I-I had somethingterrible t’tell you, So// uh R: How terrible is it. (45) [HS ST:l] D: Didju hear the terrible news? R: No. What.

We propose to examine features of the placement and linguistic design of the Pre-Announcement type in terms of their sequential relevance for the Announcement they preface. Our interest is in how the Pre-Announcement First sets up a solicitation of the news as its conditionally relevant Second Pair Part. In the course of our analysis we will detail the relationship of the PreAnnouncement sequence to its elicited Announcement and touch on aspects of how this relationship affects the design of the eventual Announcement itself. We have argued that the design of Pre- sequence Firsts is directed to the management of some projected, expectable next actions in the prefaced Se-

Pre-announcement sequences 

quence. The response to the Pre- sequence First, i.e., the Pre-Announcement Second Pair Part, can operate to either halt or forward the sequence. Pre- sequence Firsts articulate with the turn-taking structure as contingent ‘bids’ by the speaker for the next-next turn. In Pre- sequences such as the following, the response to the Pre- sequence First warrants the production of the prefaced action but does not directly seek that projected next: (46) [JG CN:1] PRE-INVITATION A: Whatcha doin’ B: Nothin’. A: Wanna drink? (47) [BS 2,1:561] PRE-OFFER A: We have a thermometer. B: Yih do? A: Wanna use it?

By contrast, the response to the Pre-Announcement type under examination solicits the projected Announcement through the organization of the first three turns of the expanded Announcement sequence. The Pre-Announcement sequence (Turns l & 2) stands in an integral relationship to the Announcement (Turn 3) as the setting up of a solicit of the Announcement. In the following instances, that organization is outlined to the side of the data: (48) [J&G:2] PRE-ANN. 1st

D:

SOLICIT AS 2nd ANNOUNCEMENT

R: D:

(49) [BS 2,1:73] PRE-ANN. 1st SOLICIT ANNOUNCEMENT

D: R: D:

Y’wanna know who I got stoned with last wee(hh)k? Who. Mary Carter ’n her boy(hh)frie(hh)nd! .hh Oh guess what. What. Professor Deelies came in, ’n he- put another book on iz order.

If we examine the design of the Pre-Announcement Second we find that it displays interrogative features in its use of the WH-marked form. The interrogative form of the Pre-Announcement Second operates to strongly forward the sequence via its independent characterizability as an Adjacency Pair First in its own right: WH-Question/Answer. In it, an answer is made conditionally relevant on the occurrence of the Question as a response to the Pre-Announcement First. The occurrence of the Question as a ‘first’ however, is itself occasioned

 Alene Kiku Terasaki

by the prior Pre-Announcement First. We have argued that the occurrence of a specific Pre- sequence First makes the occurrence of matched Seconds expectable. The Pre-Announcement First, here, can be found to set up the Q/A pair which in turn produces the Announcement in its answer position. Our question is how the turn bid for in the Pre-Announcement First was elicited by that First. What features of the Pre-Announcement initial item provide for its recognizability by speaker as a Pre-Announcement First and for the return of a Solicit question as its preferred and conditionally relevant next? Pre- sequences are heard as Sequence-initial items; they are heard, not as discrete units, but as part of some unit of turns larger than those which immediately comprise the Pre- sequence. Adjacency Pairs establish the relatedness of one turn to its adjacent turn via the property of conditional relevance. Our task here is to establish the relatedness of turn sets to adjacent turn sets. We need, then, to consider the notion sequence ‘boundaries’ and how they are comprised. In some Adjacency Pairs, the occurrence of the First Pair Part establishes the beginning edge of the Sequence while the occurrence of the Second Pair Part establishes its end. Not all Second Pair Parts constitute possible Sequence boundaries, however. In his discussion of Summons/Answer Sequences, Schegloff (1968: 1081) has formulated the notion of non-terminality to describe a sequence property which relates turn sets to one another. Pre- sequences appear to be a regular locus of the occurrence of Adjacency Pairs in which the Second Pair Part regularly constitutes in some of its features an orientation to the non-terminality of the pair. It was argued there that the occurrence of a Summons generically announces more to come. Thus an Answer to a Summons such as “Joe?” “Yeah?” was noted to contain some expectation of a further turn. The return of a Solicit question to a Pre-Announcement First relies on a recognition of its prior as a Pre-Announcement First. The next section is directed to establishing those features of Pre-Announcement Firsts which make them recognizable to parties to the talk as Pre-Announcement Firsts.30

Linguistic features of the Pre-Announcement sequence We have argued so far that as the preferred alternative Second Pair Part to a Pre-Announcement First, WH-questions constitute a Solicit of the Announcement. What linguistic features of the Pre-Announcement First can be shown to be sequentially relevant to the elicitation of the Announcement Solicit as

Pre-announcement sequences 

its proper and preferred - second pair part? And, how does the Recipient’s response (the Solicit) display their recognition of the prior utterance/turn as a Pre-Announcement First? An examination of the Pre-Announcement Firsts in examples (34) to (45) reveals two lexical items which recur in their design: (1) a WH-word (e.g., ‘Guess [what],’ ‘Ywanna know [who],’ ‘[what] I got first quarter.’) OR (2) the noun, ‘thing’ or ‘somethin’ (e.g., ‘I got [sumpn] thet’s wi::ld,’ ‘I got two good [things]’, ‘I had [something] terible t’tell you.’) Where the items occur in conjunction with head sentences we can locate a restricted set of main verbs for which the items operate in object position. They are: ‘to know,’ ‘to guess,’ ‘to tell,’ and ‘to hear.’ Notice that they are each verbs which can take sentential complements. For example: “Did he tell you [that] he went on a binge?”; “I hear [that] X has been around at the parties lately.”, etc. Linguistic arguments have been advanced to the effect that WH-marked and ‘something’ items stand in a complementary syntactic distribution. It is argued that there is an underlying declarative presupposition attending WHquestions such as ‘Y’know what he wannid me t’do?’ which takes the form: ‘He wannid me t’do something’.31 In such instances, the pre-suppositional ‘something’ is proposed to underlie its surface representation as the WH-marked item. Instances of items which occur in the imperative mood such as ‘Guess what’ are proposed to contain an ‘embedded’ WH-question which is similarly derived from an underlying declarative presupposition so that linguistically it can be derived from the expanded form ‘Guess what happened’ i.e., ‘[something] happened.’ Our interest here is to relate these shared features of the Pre-Announcement Firsts to regularities in the Solicit or Second Pair Part turn of the Sequence. A major regularity of that turn is that it is almost exclusively occupied by a WH-Question. It seems to be typical that the syntactic relations (and often the case relations) set up in the Pre-Announcement Firsts for the WHand ‘thing’ words are maintained in the WH-Question which makes up the Pre-Announcement Second. For example: (50) [BS 2,1:38] PRE 1st D: SOLICIT R:

Y’know what he wannid me t’do? What.

In this example, ‘what’ in both turns of the sequence is linguistically describable as the object of the verb ‘to do’ in the initially embedded sentence: he wanted me to do something. The Solicit turn is only interpretable if the same features are assumed to apply to ‘what’ in both turns.

 Alene Kiku Terasaki

If we now consider similarities in the design of the Pre-Announcement First and the Solicit, some regularities are apparent for the instances in which the WH-word occurs in the First:

a. Agreement of number and person is preserved across the adjacent turns.32 E.g.: (51) [JG 1:1] D: I forgot t’tell you the [two best things] thet happen t’me t’day. R: Oh super, what were [they]. (52) [J&G:3] D: Y’d prob’ly turn green with envy if I tolja all about what [I] got ( ) first quarter. R: Whad[ju] get first quarter.

b. Not just any WH-question, but the same interrogative WH- as occurred in the Pre-Announcement First turn is produced in the Solicit, e.g.: (53) [J&G:2] D: Y’wanna know [who] I got stoned with a few weeks ago? R: [Who]

c. And, the Solicit turn is frequently solely occupied with the single WHinterrogative but when more than that single item occurs, the additional utterance is a replication (with appropriate transformations for speaker) of the WH-marked constituents of the prior turn into Question form, e.g.: (54) [JSLR:l3l] D: Hey you’ll never guess [whatcher Dad ((the speaker)) is lookin’ at.] R: [What’re you looking at.] (55) [J&G:3] D: You’d prob’ly turn green with envy if I tolja all about [what I got ( ) first quarter] R: [Whadju get first quarter.]

Implicit in our description of this Pre-Announcement Sequence type as one in which the delivery of the news is ‘withheld’ was the notion that it permits a ‘staged’ delivery of the news. By ‘staged’ we intend to point not to the notion of artifice but to that of serial revelation. We want to argue that the use of these WH-marked items in the Pre-Announcement First constitutes a pronom-

Pre-announcement sequences 

inalization operation which refers forward (or ‘cataphorically’) to a thereby promised replacement or referent in the next-next turn of the sequence. That is, their occurrence in the Pre-Announcement First proposes that their referents will be specified in the upcoming proposed Announcement turn. We are arguing, then, that the Pre-Announcement First Pair Part sets up a serial delivery of the news in which semantically-unspecified lexical items are initially presented and then subsequently replaced or specified in the Announcement turn. Furthermore, what can stand as a replacement or specification of the pro-termed items is constrained by the design of the PreAnnouncement First in syntactic and possibly case terms. We have so far considered the operation of WH-marked items in the PreAnnouncement. The nominal forms ‘thing’ and ‘something’ have not been dealt with. Although the generic noun ‘thing’ is not generally considered to be grammatically equivalent to WH-words, it appears to operate here similarly to the WH-words as a pro-form for a non-specified utterance that is projected as replaceable in the Announcement turn. See the following utterance: (56) [JG 1:1] D: I forgot t’tell you the [two best things] thet happen’t’me t’day. R: Oh super, what were they. D: I got [a B+ on my math test,] : [And] I got [athletic award]

The projective character of the item [thing] marked with the quantifier ‘two’ is preserved in the Announcement turns via the use of the conjoined sentence structure and the repetition of the frame “I got...” Notice also that in the Solicit turn, the WH-question form parallels the shape of sequences in which there is a WH-marked item in the First. ‘Someone’ and ‘something’ are linguistically characterized as indefinite pronouns. In their operation as pronouns they appear to provide for the same types of forward reference we have described for instances of WHmarked items. The occurrence of ‘something’ in the Pre-Announcement: “I got sump’n thet’s wi::ld,” is replaceable in the Announcement turn by a single NP: “a big red fire alarm box.” Other occurrences of ‘something’ are not so specified: (57) [DA:4] D: I had something- terrible t’tell you. : ((Edna died))

 Alene Kiku Terasaki

(58) [JSLR:41] D: Toni en Bob I have something tuh tell you... : You must not use- any’v the pottery you picked up der any-tuh cook in. (59) [JSLR:153] D: Say Tom, I wanna tell you sump’n : ...he thoughtchu were my son!

We notice, however, that in each of these instances ‘something’ occurs as the proform for the sentential complement of the verb ‘to tell.’ It appears that here, the verb specifies the referent for ‘something’ as a full sentence replacement.33 This conjunction of ‘to tell’ with the pro-term operates correspondingly in example (56) which uses ‘things’. That suggests a possibly more general operation of ‘thing’ and ‘something’ in recognizable Pre- sequences as pro-forms for a ‘tellable’ or ‘sayable’ in the next-next turn of the sequences they preface. And, insofar as these two items underlie WH-marked items, the WH-items can be argued to operate in the same way. The notion of a pro-form as a regular component of this Pre-Announcement type which has sequential implications allows us to expand our analysis to include the nominal form: “news” as a candidate Pre-Announcement pro-form. For example: “Hey we got good news.” As a naming of the projected action, it duplicates in its operation the syntactic and case relations in the lexical item ‘thing’. The list of admissible pro-forms we have discussed is thus restricted to a set of specifiable items not necessarily characterized as linguistically equivalent but which operate as sequential equivalents. A sequential characterization provides for their locatable similarity as components of Pre-Announcement First Pair Parts. Similarly, while the Pre-Announcement Firsts are realized by a range of sentence types (imperatives: “Guess what”; questions: “Didju hear”; declaratives: “I got two things”), the co-occurrence of a specifiable set of linguistic and sequential features identifies them as instances of the ‘same thing’ conversationally. Recipient’s recognition of an utterance as a Pre-Announcement First is then in part provided for in the presence of what we can now call ‘sequential’ or ‘sequentially implicative’ pro-forms and verbs.34 In summary, linguistic features of the design of Pre-Announcement Firsts have been argued to provide for the operation of the WH- marked and ‘thing’

Pre-announcement sequences 

items as pro-forms for forwardly referenced ‘tellables’ or ‘sayables’ in the contingently prefaced Announcement turn of the sequence. Further it was argued that the linguistic environment of these pro-forms specifies in advance of their appearance the shape of their replacements in the Announcement. The notion of a staged delivery of the news across the serial turns of the sequence allows us, consequently, to treat instances such as, “I got two things t’day,” (which appears to have announcement /declarative characteristics) not as counter-examples to our formulation, but as further instances of the same serial operation.

The solicit turn We noted earlier in our discussion of features of the Solicit turn that the WHquestion which occupies that turn selects the same WH- item of the Pre- or its syntactically corresponding Pro-form. It is a member’s problem to determine which sequence type any prior utterance is an instance of in order to locate their appropriate next action. The realization of a Pre-Announcement First by a range of sentence types creates such a problem insofar as it is otherwise the case that inquiries, for example, properly take answers, commands to guess take guesses, WH-questions take replacements of the WH-marked items as their answers, etc. In Pre-Announcement sequences, however, each of these items takes a Solicit (WH-question) as their next action. For example, the occurrence of a question such as: “Y’know what he wannid me t’do?” could comprise for its recipient the occasion for a selection of their appropriate next as the production of a replacement for the WH-marked item: “He wannid you to do X.” On the other hand, in our instance, the occurrence of the question appears to invoke the operation of a ‘selection rule’ of an alternative sort, one which directs the recipient to select a WH-interrogative matched to the one in the prior turn as their next action. We have argued to this point that it is the presence of sequentiallyimplicative pro-forms and verbs which in part account for recipient’s recognition of an item as a Pre-Announcement First. While there are other features which attend its recognition (such as placement), that formulation still appears to be inadequate in the face of the alternative possible second we cited above. What we want to propose is that the inadequacy of our analyst’s formulation is as well a member’s formulational problem. That is, the recognition, in this instance, of whether the item is ‘in fact’ a Question or ‘in fact’ a PreAnnouncement First is problematic to its recipients and can constitute the basis of recognitional ‘mistakes.’ We cite the following instance of such an error

 Alene Kiku Terasaki

occurring in a structure closely allied to Pre-Announcements, that of Riddle questions: (60) [KR:2] 1 Kid: 2 Mom: 3 Kid: 4 Mom: 5 6 Kid: 7 Mom: 8 Kid:

I know where you’re going. Where. To: that (meeting ). Right. Yah! Do you know who’s going to that meeting? Who I don’t know! Ouh:: prob’ly: Mr. Murphy an’ Dad said prob’ly Mrs. Timpte en some a’ the teachers.

The utterance in turn 1 of the example is correctly recognized and treated by its recipient as an instance of a riddle question/the first turn of a Riddle Sequence. The utterance in turn 5, however, is mistakenly treated by its recipient as a ‘return’ riddle, while its deliverer treats it as to be heard as an Inquiry. Notice that the return of the solicit does not imply that the recipient could not otherwise provide an ‘answer’ since one is done in turn 8. What the solicit evidences instead is that the utterance was recognized as an instance of a Riddle question, something appropriately responded to by a Solicit, regardless of the state of one’s knowledge on the matter. That recognitional errors can be made by actual speakers provides for the possibility that on the occurrence of such utterances there are alternative possible interpretations available. We propose that the interpretation of an utterance as a Pre-Announcement First of this type constitutes one of a set of orderly and regular alternatives to an interpretation of the utterance as one of the sentence types which are conventionally thought of in terms of their surface syntactic features. Thus, utterances such as: “y’know what he wannid me t’do?” may be treated by speakers as an instance of the linguistically conventional ‘Yes/No’ question with an embedded WH-question or it may be treated as an instance of a Pre-Announcement First. That alternative interpretations are available is not unique to Pre-Announcements. Riddles, riddle jokes, and rhetorical questions as well as a range of other Pre- sequence types appear to exhibit this same feature. As such they comprise a class of allied conversational structures not fully encompassed by a strictly grammatical description of their features. On their occurrence and a recognition of them as a possible instance of this class, a Solicit, regardless of its morphological realization, is their proper and preferred Second Pair Part.

Pre-announcement sequences 

It has been argued elsewhere35 that such instances as the following: (61) [BS 2,1:63] D: Guess what. R: What. (62) [GTS 5:1] D: You know where I went last night? R: Where.

Represent cases of ‘elided’ sequences in which a finding that one cannot guess or does not know the answer results in a deletion of the otherwise syntacticallyfitted Second Pair Part and a ‘skipping to’ the solicit questions. So, for example, (61) is actually: (63) [BS 2,1:63] D: Guess what. R: DELETE: (I can’t guess). R: What.

That analysis relies on a conception of conversational sequences (or Adjacency Pairs) as mentalistic structures, detached from the actual scenes of their use. While it is certainly the case that such otherwise matched seconds do occur, as in the following instances, they regularly do not.36 (64) [GTS 1:23] D: What’s black ’n white ’n hides in caves, (0.4) R: [A’right I give up,] What’s black ’n white ’n hides in(65) [HS ST:1] D: Didju hear the terrible news? R: [No,] What.

Our analysis of the linguistic and sequentially relevant features of the PreAnnouncement First’s design argues that the proper next to any one of these items is not located in the syntactic form of the utterance alone but is determinable by speakers only through an examination of its occasioned appearance as an instance of a possible Pre-Announcement First or Pre- sequence First. That determination and not its otherwise characterizability as an Adjacency Pair First is what locates its conditionally relevant Next as the Solicit. A second

 Alene Kiku Terasaki

pair part matched to the sentence type of the Pre-Announcement is not, then, ‘missing’ or ‘elided’, it is not called for. The relevant analytic unit is that of: D: Pre-Announcement First Pair Part R: Pre-Announcement Second Pair Part/Solicit of the Announcement The existence of other utterance organizations displaying some common grammatical features which would otherwise require a matched second is simply not a consideration in recipients’ selection of the Solicit once they have opted to treat the item as an instance of a Pre-Announcement First. Pre-Announcement sequences stand as differentiable alternatives to, for example, WH-question sequences, not as deviant instances of them. One is not more ‘fundamental’ than the other; they are alternative Adjacency Pair types. We have been concerned in this section to specify features of the PreAnnouncement First directed to the elicitation of the Announcement Solicit as its proper and preferred Second Pair Part. The responsiveness of the Announcement Solicit to its prior was proposed to be visible in its selection of the WH-marked item or syntactic equivalent in the Pre-First. Earlier it was argued that the Solicit constitutes an Adjacency Pair (WH-Q/Answer) in its own right which makes the appearance of the Announcement conditionally relevant. It was further argued that while the Solicit comprises an Adjacency Pair First, it is not the initial turn of the Sequence but displays its subsequence to the PreAnnouncement First. We suggested that the form of the Solicit is present in the design of the Pre-Announcement First. The next section considers variations in the form of the Solicit turn which provides us with one further indication that these utterances are part of a larger Announcement Sequence. That discussion will also take us to a consideration of the Pre-Announcement sequence’s relationship to the design of the Announcement turn itself.

The form of the Solicit question It was noted earlier that the WH-question of the Solicit takes two shapes: (1) as a replication of just the single WH-marked item in the Pre-Announcement, e.g., Guess [what] . [What.] or (2) as a replication of the WH-marked or ‘pro’d’ constituent of the Pre- in question form, e.g., Hey you’ll never guess [whatcher Dad is lookin’ at] [What’r you lookin’ at.] In this respect, the forms overlap with three forms of WH-questions which occur regularly in conversation:

Pre-announcement sequences 

1.

Full sentential WH-questions:

“What’r you lookin’ at.” “Where are you going?”

2.

One-word WH-questions:

“Who.” “What.”

3.

Partial WH-questions or Appendor questions:37

“On what.” “From whom.”

Our discussion to this point has utilized the notion of ‘sequentially relevant’ features to describe those aspects of the design of an utterance which attend its projective uses in the upcoming talk and display a sensitivity to its sequential placement vis-à-vis prior talk. That notion provides us with a way of organizing the forms of WH-questions we have listed. WH-questions which occur as complete sentences can occur as either First or as subsequent to some prior turn sets. Their positioning in relation to prior turns is not necessarily marked in their design. By contrast, One-word questions and Appendor questions display a relatedness to their prior turns which marks them as clearly subsequent to an antecedent turn which contains ‘what the question is about.’ Appendor questions such as the following are syntactically-fitted to their prior turns as extensions of the syntax of the utterance of that turn: (66) [J&G:4] APPENDOR Q Answer

A: B: A:

He had an operation. On what. On his hand.

One word questions also display their subsequence to prior turns. One major category of their occurrence is as what Sacks (1992) has called ‘next-turn repair initiators’ (NTRI’s): (67) [TG:26] NTRI Repair

A: B: A:

Sibbie’s sistuh hadda baby bowa:::y. Who? Sibbie’s sister.

On the occurrence of a problematic item, devices which solicit a repair of that item in next turn can be instituted. The attenuated form of the question relies on its immediate adjacency to the just prior utterance for its sense. Recipients locate what the question is about strictly in terms of its adjacency to an utterance which contains matched constituents, here the NP, “Sibbie’s sistuh.” As we argued earlier in regard to the WH-Solicit, the WH- item is only in-

 Alene Kiku Terasaki

terpretable if its features are assumed to be matched to those of constituents present in the prior turn. While it is entirely possible for the repair initiator to be done in full sentential form, i.e., “who hadda baby boy?”, its adjacent placement makes that unnecessary. Furthermore it may be that full repeats operate to convey quite different orders of information – such as surprise rather than non-hearing. (See Jefferson’s (1972) paper on “Side Sequences”.) Thus the use of the one-word question may not simply attend considerations of ‘economy.’ It may differentiate a type of sequence. Our interest in its use here is that its occurrence in a turn series is constrained to be one which is subsequent to a prior turn. We suggested earlier that Pre-Announcement Firsts are Sequence-initial items and that Adjacency Pairs can occur inside larger Sequences. Our problem was to be able to relate sets of adjacent turns to one another as Sequence parts – in this case the occurrence of Solicits as second to Pre-Announcement Firsts, but independently characterizable as Adjacency Pair First Pair Parts of a WHquestion/Answer sequence. One-word WH-questions show up as Solicits in response to Pre-Announcement Firsts of the type we have been examining, particularly in the form “What.” The general operation of One-word WH-questions as items which in their form mark their internal and subsequent relationship to prior turns suggests how we are able to argue for the unitary character of the serial turns of the Pre-Announcement Sequence. The form of the Solicits as One-word questions makes visible to us an orientation by the recipients of Pre-Announcement Firsts to the subsequence of the Solicit. Earlier we noted that a major way that speakers may display their understanding of some prior utterance is by producing its sequentially relevant next action. When that next action coincides with a return to otherwise intuitively available items such as ‘Question/Answer’ the form of return seems non-problematic. When that next action overlaps with one of those intuitively available items such as Question/Question (“Y’wanna know who. ..” “Who.”), however, the specification of how the return is antecedently related to its prior is troublesome. The designing of a return item to attend features of its subsequence to a prior turn would be an indication of its sequential relationship to that item. Here, the use of the One-word question after a recognizable PreAnnouncement First marks its subsequence to the Pre- while still providing for the ‘strong forwarding’ of the larger Sequence through its appearance as a Question.

Pre-announcement sequences 

Note that we are not here remarking on an analyst’s finding only. If a concern of speakers is to display their grasp of a prior utterance, then designing their own utterance to reflect a syntactic analysis of the prior would be a major means at their disposal. For the case of One-word Solicits, then, we have one further feature of the Pre-Announcement Sequence which argues for the serial occasioning of the two utterances as part of a larger, unitary Sequence.

The design of the Announcement turn The operation of Pre- sequences as sequence-initial items provides us with a way to locate the initially noted features of Pre-Announcement Firsts as related serially to their subsequent turns. As components of the Pre-Announcement First, the sequential pro-forms we have suggested establish the range of their possible specifications and direct the selection of the return Solicit from a set of grammatically matched options. The design of the Pre-Announcement First can thus be seen to strongly control its projectable outcomes in the Solicit turn. If we now track the pro-forms into the Announcement turn, we find convergent similarities. For example: (68) [J&G:2] D: Y’wanna know who I got stoned with a few(hh)eeks ago? hh! R: Who. D: Mary Carter ’n her boy(hh)frie(hh)nd. hh. (69) [BS 2,1:73] D: .hh Oh guess what. R: What D: Professor Deelles came in. – An’ he put another book on iz order. (70) [JG 1:1] D: I forgot t’tell yih the two best things then happen’tuh me t’day. R: Oh super, what were they. D: I got a B+ on my math test.....and I got athletic award. (71) [JSLR:13l] D: Hey you’ll never guess whatcher Dad is lookih- is lookin’ at. : R: What’re you looking at. D: A Radar Range.

 Alene Kiku Terasaki

(72) [JSLR:38] D: Hey I got sump’n thet’s wi::ld. R: What. D: Y’know one a’ these great big red fire alarm boxes thet’r on the corners? I got one. (73) [J&G:3] D: Y’d probl’y turn green with envy if I tolja all about what I got ( ) first quarter. R: Whaddju get first quarter. D: Four A’s ’n two B’s. (74) [GTS 5:43] AI: Well I know that Jim, for example, is facing some things this week which are uh ((pause)) a little more serious ((pause)) -aren’t they, regarding school D: Yeh. R: What’s hap’ning. D: I got kicked outta the University.

The occurrence of the Announcement utterance as a response to the WHSolicit marks it as a replacement of the WH- or other pro-form items originally presented in the Pre-Announcement First. It does this by way of the Solicit turn’s preservation and forwarding of the proform as a WH-question. Several of the Announcement First Pair Parts appear to operate in line with the workings of WH-question/Answer pairs. That is, the rule for answering WH-questions directs the answerer to provide only that sentence constituent which is WH-marked in the question. The syntactic distribution of the WHmarked constituent is preserved in the replacement. For example: (75) [J&G:3] D: Y’d prob’ly turn green with envy if I tolja all about what I got ( ) first quarter. R: Whadju get first quarter. D: Four A’s ’n two B’s.

In several of the instances, the Pre-Announcement First provides the entire sentence frame for the replacement, singling out the constituent type which can admissibly replace it, as in the example above. In others, however, that replacement relationship is not so well defined. Items generally regarded as pro-forms for noun phrases (‘what’, ‘something’, ‘thing’) can be replaced by whole sentences. For example:

Pre-announcement sequences 

(76) [BS 2,1:73] D: Oh guess [what]. R: [What.] D: [Professor Deelies came in, ’n he- put another book on iz order]

If we separate those Pre-Announcement Firsts which get phrasal constituent replacements from those which get full sentences we find that generally the interrogative WH- pro-forms are replaced by phrasal constituents while the items pro’d as ‘thing,’ ‘something,’ ‘news,’ and ‘what’ are more regularly replaced by full sentences. As we noted earlier in our discussion of ‘thing’ and ‘something’, the full sentence replacements are sentential complements of the verbs ‘to tell,’ ‘to know,’ ‘to hear,’ and ‘to guess’ which occur in the Pre-Announcement First. The phrasal constituents, on the other hand, are more usually noun phrases specified in the Pre-Announcement. As a function of the syntactic design of the Pre-Announcement First, then, the occurrence of the pro-form types sets up the sort of replacement which will stand as their Announcements. In this respect, they parallel the operation of the relationship we outlined between the Pre-First and its Solicit. In his discussion of the distinction between ‘marked’ and ‘unmarked’ information focus, Halliday (1967–1968: 207–208) suggests that where information focus is not marked, the whole of an utterance may come under its domain as ‘new’ information. He proposes as a device for locating such instances that one consider “the implied questions to which the information unit could stand as an answer.” It may be that his analysis implicitly attends the sequential organization of such utterances as they occur in informing sequences such as the one we have been considering. His treatment of these utterances as marked across their entirety as ‘new’ may arise from some recognition of them as utterances which occur in Announcement turns as responses to WHSolicits.38,39

Phrasal constituent replacements as Announcements That perspective brings our attention to the fact that where Pre-Announcement Firsts have specified a phrasal constituent replacement, the replacement is the sole occupant of the Announcement. (See examples (68), (71), (72) & (73).) For these instances, a further articulation of the sequential analysis with informational structures of the utterance is suggested. In contrast to the case of unmarked focus in which the domain of ‘new’ is the entire utterance, ‘parked focus’ selects a single constituent as bearing “the

 Alene Kiku Terasaki

main burden of the message” (Halliday 1967–1968: 204). Sequential features of the structure of expanded Announcement Sequences permit the isolation of the phrasal constituent – as the ‘information focus’ of the utterance. By marking one constituent of the Pre-Announcement First as not specified and to be replaced in the next-next (Announcement) turn, the selection and designation of the ‘marked focus’ (that is, the ‘news’) of the utterance is accomplished sequentially. The isolation of the WH-marked constituent from the remainder of the utterance proposes in its design that the WH-marked item is the news of the utterance. It concomitantly preserves the interactional preference not to re-tell by producing virtually the entire utterance without its news. At the outset of this paper we suggested that sentences otherwise characterizable as informings or as informationally ‘new’ are not isolated occurrences but appear in the environment of conversational sequences directed to their delivery and receipt. The production of ‘new’ information in the utterance is most dramatically embedded in its sequential presentation in these instances of phrasal constituent replacements in the Announcement turn. We notice also that while full sentential replacements are not so embedded, that their appearance following the occurrence of a pro-form in the Pre-Announcement First and its forwarding in the Second make them also visibly the result of a delivery sequence. Our concern has been to demonstrate how these utterances are observably parts of a sequence rather than de-contexted utterances.

Summary We have considered the linguistic and sequential organization of a particular Pre-Announcement sequence type. The sequential description characterized the successive turns of the sequence as each serially occasioned by the occurrence and design of its prior. The recognizability of the Pre-Announcement First was argued to reside in its placement across the topical organization of the conversation as well as in its component features. Its identification as a Pre-Announcement then provides a basis for the finding that the turns which ensue are occasioned on its occurrence and organized into a three turn series which, along with an Assessment as a fourth turn, constitute an expanded Announcement Sequence. Finally, the organization of the parts of the sequence was shown to attend a preference not to re-tell which articulates with information structures suggested by analysts of the given/new distinction. The informational organization of the parts of the sequence was argued to make the achievement of the news

Pre-announcement sequences

delivery an interactionally-accomplished matter, assembled across the occurrence of the sequence as a production local to this telling. Syntactic structures were found to be fitted to the exigencies of the delivery, and the recognizability of utterances as Announcements was argued to reside in part in their placement in the Announcement turn of the sequence. This suggested that, rather than being an independent matter of ‘content,’ the ‘news-ness’ of an utterance in conversation is developed in and through its treatment and reception in some actual scene of reportage.

Notes on methodology A brief introduction to the methodology of Conversation Analysis seems appropriate here since we depart from the more traditional experimental and analytic techniques employed in language studies. A thematic concern of work which has been done in Conversation Analysis has been to discover the role of the sequential organization of conversation in an understanding of language in use. In speaking of the sequential organization of conversation we are locating the focus of our interest on the level of contexted speech events rather than on that of single utterances of speakers. Central to this interest is a concern to attend those features of speakers’ utterances which are directed to forwarding a major mode of organizing spoken interactions: the turn-taking system. As detailed by Sacks et al. (1974), the analysis of conversational materials reveals the existence of a system which organizes speaker turns and the design of utterances contained in those turns. Utterances in conversation occupy speaker turns which are organized to allow one party to speak at a time while providing for recurrent speaker change. A fundamental aspect of the understanding of an utterance in conversation appears to involve any hearer’s determination of whether a current speaker’s utterance selects her/him to speak next. Thus, any current speaker, in designing their utterance, may employ a variety of current-speaker selects next-speaker techniques which allocate the upcoming turn to some one of their hearers. A major set of such techniques is the set of type-characterized utterances called ‘adjacency pair first pair parts’ such as offer, complaint, or greeting which take, as their appropriate ‘second pair parts,’ matched utterances such as acceptance, apology, or return greeting. Thus a hearer’s analysis of a prior utterance directed at him as an instance of an adjacency pair first pair part will provide for the return of a matched second pair part to that item (e.g., requests take grants of the request or denials of it). Hearers must analyze current utterances to find



 Alene Kiku Terasaki

if they are selected to speak next and to find what they should say. This type of account suggests that the understanding and design of utterances is centrally affected by the presence of the turn-taking system as something which must be attended to. We find that the kinds of phenomena which reveal the existence of this sequential organization are frequently to be found in what would otherwise seem to be ‘minor’ features of utterances such as ‘errors,’ false starts, appositional particles, tag questions and the like. Consequently, our work has been directed to the development of an observational methodology designed to capture the particulars of situated speech. Our use of audio- and video-recordings of naturally-occurring conversation in conjunction with detailed transcriptions is directed at preserving those particulars in a way which will permit close and repeated scrutiny for orders of organization hitherto not available to analysts. Work in Conversation Analysis on objects such as appositional particles (Sacks et al. 1974; Pomerantz 1975), errors (Schegloff 1987 [1973]; Jefferson 1972, 1975), and idiomatic or ‘sentence fragment’ utterances (Davidson 1975; Schegloff 1968; Goldberg 1975) demonstrate the import of these procedures for securing the sorts of findings about the sequential organization of conversation to which we refer.40 Given the recognized importance of ‘context’ in establishing the understandability of utterances (Firth 1935; Hymes 1974; Fillmore 1968; Goffman 1964) we have been interested in developing a method of examining ‘contexted’ as against ‘isolated’ utterances of speakers. Methodologies which rely on intuitive reconstruction of situated speech may omit phenomena which turn out to be central to speakers’ understanding of an utterance in terms of what it reveals to them about prior actions as well as about their own next actions. Austin (1962), Searle (1970), Labov (1972) and others have pointed out the central import of distinguishing linguistic characterizations of utterances, ‘what is said,’ from those concerned with the actions that the utterances accomplish, ‘what is done.’ Similarly, as Gumperz (1972: 24) has suggested, interview and elicitation techniques which ask speakers to reproduce ‘de-contexted’ speech activities may result in the delivery of idealized versions of that activity which remove significant features of the data from the possibility of investigation. Rather than glossing variations in the range of situated activities and identities we encounter as matters of context we attempt to address those variations directly for the deeper generalizations which will encompass them.41 The organization of the turn-taking system also provides us with an approach to the question of the verifiability of our findings. In the production of

Pre-announcement sequences

a matched second pair part to a prior turn, speakers show their understanding of that turn. Sacks et al. (1974: 728–729) have argued the methodologic significance of this in the following way: It is a systematic consequence of the turn-taking organization of conversation that it obliges its participants to display to each other, in a turn’s talk, their understanding of other turns’ talk. More generally, a turn’s talk will be heard as directed to a prior turn’s talk, unless special techniques are used to locate some other talk to which it is directed. . . . But while understandings of other turns’ talk are displayed to co-participants, they are available as well to professional analysts, who are thereby afforded a proof criterion (and a search procedure) for the analysis of what a turn’s talk is occupied with. Since it is the parties’ understandings of prior turns’ talk that is relevant to their construction of next turns, it is their understandings that are wanted for analysis. The display of those understandings in the talk of subsequent turns affords both a resource for the analysis of prior turns and a proof procedure for professional analyses of prior turns – resources intrinsic to the data themselves.

Speaker’s display of understanding, then, is criterial to an evaluation of the validity of our findings as analysts and that criterion anchors our analysis in the data corpus such that repeated examinations of that data or any similar data would yield equivalent findings. Our interest is to build a methodology which allows for the treatment of our observational findings as systematic achievements of the parties to any conversation.

Notes . An additional criterion used by speakers to determine the informational status of an utterance is previous mention of the item in earlier contacts with these or other coparticipants. Interactants develop a common historical biography with one another across their serial encounters which they can be shown to take account of in successive contacts. . Previous mention in this particular conversation is not the only criterion which determines an item’s informational status. Chafe (1974: 119–122) discusses the role of environmental features. We have a further comment on this. Interactants orient to the production of conversational occurrences we are calling ‘deliveries of news.’ ‘News’ is for conversation some report produced by its deliverer as not known to its recipient and subsequently interactionally ratified by the recipient as news-to-them. The factual character of the item’s status as not having been previously mentioned between these two parties or previously known to the recipient is not an issue since interactants can be shown to have ‘in fact’ known some item and yet treat it as not-known. (Recipients may have a variety of good interactional



 Alene Kiku Terasaki

grounds for not displaying that they have heard an item previously. For example, not revealing they have been privy to a ‘secret.’) In contrast to an interest in the factual character of the recipient’s knowledge or ignorance, our concern here is with the treatment of an item as not-known. . Citations refer to transcriptions of actual video- or audio- taped conversations. Items before the colon refer to the title of the transcript, those following it refer to the page or line of the transcript. Names are altered. The use of “...” within an utterance or “:” between utterances indicates omitted material. . A full discussion of Adjacency Pairs can be found in Sacks (1992) and in Sacks et al. (1974). A brief description is given in Schegloff and Sacks (1973) . We use the term ‘speech event’ here to refer to occurrences in conversation which interactants treat as prescriptive of next actions. It is not, then, coterminous with Hymes’ (1974) use of ‘speech event’ nor with Searle’s (1970) treatment of it. For a discussion of the equivocal nature of events in talk, see Jefferson (1979). . This set of terms was first used by Pomerantz (1975: 33). . The assignment of these identities as operative for the interaction can be a matter which is locally negotiated as the sequence proceeds. See the example which follows in which the intending Deliverer now becomes a Recipient of news: [AT STI:1] A: B: A: B:

How are ya. Oh I don’ know. I don’t feel good. I had trouble with my stomach. I’ve had pains all day. We all do. Is that right?

‘News marks’ such as ‘Really?’ and ‘Is that right?’ appear to operate to signal the receipt of a piece of information not known to its recipient. . For a full discussion of Insertion Sequences see Schegloff (1972). . For a description of the import of Recipient Design see Sacks et al. (1974: 727ff.). . The significance of the notion of collaborative production is detailed in Schegloff and Sacks (1973: 246). . Massive numbers of things which would appear to qualify as Announcements via their appearance in declarative sentence format are not in our technical sense Announcements. One of our concerns is to at least initially restrict ourselves to instances of clear deliveries of ‘good’ or ‘bad’ news and their attendant Assessments. . This term is used by Halliday (1967: 206). . Sacks (1992) gives an initial discussion of Pre-Sequences. . Gordon and Lakoff (1971) are treating an instance of a PRE-REQUEST in their example, “It’s cold in here.” . For additional instances of such internal expansions, see Schegloff (1972) and Jefferson (1972).

Pre-announcement sequences

. The utterance’s visibility as projective of some next is attended by the recipient in their production of ‘Why’ which is then answered with a report of the intended invitation. It also attends the possibility that now that she has been found to be ‘busy’ for the evening, the invitation will not be voluntarily produced. That intercepted invitations can be nonetheless solicited then provides for the possibility that arrangements can be me which will permit its ultimate acceptance. . A second major environment of Announcement items is in the Answer turn following News Solicit Questions such as “How are you.” and “What’s new.” as well as Topic-Initial Questions like “How’s Mr. White.” which request ‘up-dates.’ These instances are treated in the current report as equivalent to Announcements which do not occur in Answer position since they otherwise display many structural features in common. Their alternicity with non-questioned Announcements is a matter which will be taken up in subsequent reports. . It should be kept in mind that not all, nor even the majority of Announcement Sequences contain Pre-announcements. Our interest here is to examine one type of Preannouncement which does occur. A second type of Pre-announcement appears to be differentiated from this type according to the manner of their integration with the prior talk. While the type under consideration displays features of design and placement which mark it as ‘topic initial,’ the other type is produced as topically ‘hooked’ to the on-going talk. For example, the following instance is done as an agreement to the foregoing talk, not as disjunctive from it: [GTS 3:10] ((D=Deliverer, R=Recipient)) K: Gunna be a good morning, isn’t it. R: Yea:h. One a’ those mornings, K: (Oh yeah.) D: Yeah it’s gonna be one a’ those mornings. R: What’s the matter. D: ’m not coming back any more. R: Oh? How come? . That is, news appears to have a ‘decay rate.’ If it is not told on the first opportunity to tell, it is not necessarily equivalently tellable on a next opportunity and may not then be told at all. . For this example and others we have just cited, the source of Deliverer’s determination that the news may be known is perhaps located in the fact that the intended news comes out of the realm of ‘public news,’ i.e., television, radio, newspapers, magazines, etc., which the Recipients can be thought of as having had equivalent access to for a first hearing. Deliverer’s concern to mark news as possibly already heard certainly attends other concerns as well. Among these is the desire to mark one’s place in delivery networks vis-à-vis intended Recipient’s. The taking or making of an opportunity to personally deliver news of significant events in one’s life is used by parties to mark the sort of relationship they have with the Recipient. The failure to have delivered such news can warrant complaints directed to that issue such as, “Why didn’t you tell me,” or “The wife is always the last to know.” The delivery of some otherwise personally deliverable news by a third party as ‘possibly already known’ permits that third party to preserve the relationship of Recipient to the ‘appropriate Deliv-



 Alene Kiku Terasaki

erer’ of the news in the face of the possibility that they, and not the appropriate deliverer, are the first Deliverer. The ‘possibly heard’ expansion of the Pre-announcement First may mark the news as something which the current Deliverer ‘just happened’ to have delivered before the appropriate Deliverer did. . Harvey Sacks, personal communication, UCI Seminar, 1974. . Gail Jefferson, personal communication, UCI Seminar, 1970. . This fragment is taken from the tape of the Japanese film, “As a Woman Ascends the Stairs,” directed by Oshima. The translation was done with the help of a native speaker and professional translator under the proviso that the translation be as ‘literal’ as possible. The naturalness of film dialogue is of course a difficult problem. However, native speakers of Japanese whom we have asked to hear the tape report that it is not ‘contrived’ or ‘Westernized’ for them. . Sudnow (1967: fn. 117): “I intend to restrict attention to those events which have a clearly perceived announcement- type structure, events with presentational formats such as, “I have something to tell you.” . They can of course be mistaken in their recognition although we have no instances of this in our data so far. . We refer here to the semantic device proposed by Fillmore and others to encompass semantic relations of utterances. We cite one of Fillmore’s (1968: 24) definitions: The case notions comprise a set of universal, presumably innate, concepts which identify certain types of judgments human beings are capable of making about the events that are going on around them, judgments about such matters as who did it, who it happened to, and what got changed.

. There may be a preference structure attached to such Announcements which motivates the attempted guesses of the news in preference to having it announced. See Tesser & Rosen (1975) and Rosen et al. (1974) for psychological discussion of this which taps into the preference long attended in fables and adages about ‘the bearers of bad tidings’. . There are other instances of Deliverers arriving on the scene with their Announcements, as in the following example [BS 2,2:53] D: Good news! ((D approaches from a distance)) Got good news for ya. (1.0) I’m the bringer of- good tidings. ((as D arrives)) . We are distinguishing here between ‘routine’ and ‘pre-empting’ moves. Speakers may preempt the regular organization of openings to do Announcements. When they do not preempt, they can be found to utilize that regular organization to produce their news. . For a discussion of the methodological import of this as an analytic constraint – that the distinctions be made by parties to the scene – please see the section on Methods. . See particularly Quirk (1972: 396). . “What” does not preserve these features. We will discuss its operation subsequently. . This was pointed out by Gerald Delahunty, personal communication.

Pre-announcement sequences  . For a discussion of children’s use of this sequence item as a floor seeker, see Sacks (1972: 342–244) and Garvey & Hogan (1973). . See especially Churchill (1972–1973). Linguistic treatments of elision in answers to questions employ a similar strategy. . A regular exception is the instance of ‘joking guesses’ as in the following instance: [GTS 1:23] D: What’s black’n white’n hides in caves, (0.4) R1: A’right I give up, What’s black’n white // ’n hides in R2: A newspaper. RI: Hhh D: No, (0.4) Pregnant nun. Such items are regularly done interruptively and ‘pushed up’ as Joke insertions often marked with laugh tokens which signal their non-serious guess nature. This instance relies further on the availability of the riddle start as like the one: “What’s black’n white’n red (read) all over.” which D2 clearly knows is not the same riddle being done here since it has already been fully presented once. The joking/put-down aspect of the guess relies on the suggestion that D can be found to have produced an ‘old’, ‘everyone knows it’ joke riddle. . Discussed in Sacks, et al. (1974: 718). . A major aspect of deliveries which has not been treated here is the issue of intonation as marking information focus in Announcements. . Note that for most of the Announcement utterances there is an identity of Halliday’s three analytic concepts: They each represent cases of unmarked focus and are thus equivalently representations of theme and given. . Sequential organization is not the only order of utterance organization studied in Conversation Analysis but it is a major focus of interest. . We do not intend this as a criticism of existing methodologies – since they are designed to meet the needs and purposes of their research problems – but as a defense of our own observational approach. . Although A is apparently speaking about “Mittie” here, she mis-uses the name of her recipient B ( “Jeannette”).

References Austin, John (1970). Speech Acts: An essay in the philosophy of language. Cambridge: Cambridge University Press. Chafe, Wallace (1970). Meaning and the Structure of Language. Chicago: University of Chicago Press.

 Alene Kiku Terasaki

Churchill, Lindsey (1972–1973). The grammar of questioning. In R. W. Shuy (Ed.), 23rd Georgetown University Roundtable on Language and Linguistics. Washington, DC: Georgetown University Press. Davidson, Judy (1975). Ending Structures in Conversation. Unpublished Ph.D. dissertation, University of California, Irvine. Fillmore, Charles (1968). “The Case for Case.” In E. Back & R. Harms (Eds.), Universals in Linguistic Theory. New York: Holt, Rinehart & Winston. Firth, J. R. (1935). “The technique of semantics.” Transactions of the Philogical Society, 36–72. London. Garvey, Catherine, & Hogan, R. (1973). “Social Speech and Social Interaction: Egocentrism Revisited”. Child Development, 44 (3), 562–568. Goffman, Erving (1964). “The neglected situation.” American Anthropologist, 66 (2), 133– 136. Goffman, Erving (1967). Interaction Ritual. New York: Doubleday Anchor. Goldberg, Jo Ann (1975). “A System for the transfer of instructions in natural settings”. Semiotica, 14, 3. Gordon, D., & Lakoff, George (1971). “Conversational Postulates”. Proc. of the Seventh Regional Meeting of the Chicago Linguistic Society. Chicago. Gumperz, John (1972). “Introduction.” In J. Gumperz & D. Hymes (Eds.), New Directions in Sociolinguistics. New York: Holt, Rinehart & Winston. Halliday, M. A. K. (1967–1968). “Notes on transitivity and theme in English”, Parts I–III. Journal of Linguistics, 3, 4. Haviland, S. E., & Clark, Herb. H. (1975). “Comprehension and the Given-New Contract.” Social Sciences Working Paper #67. University of California, Irvine. Hymes, Dell (1974). Foundations in Sociolinguistics: An Ethnographic Approach. Philadelphia: University of Pennsylvania Press. Jefferson, Gail (1972). “Side Sequences.” In Sudnow, D. (Ed.), Studies in Social Interactions. New York: Free Press. Jefferson, Gail (1975). “Error Correction as an Interactional Resource.” Language in Society, 2, 181–199. Jefferson, Gail (1979) “A technique for inviting laughter and its subsequent acceptance/ declination”. In G. Psathas (Ed.), Everyday language: studies in ethnomethodology (pp. 79–96). New York: Irvington. Labov, William (1972). Sociolinguistic Patterns. Philadelphia: University of Pennsylvania. Pomerantz, Anita (1975). Second Assessments: A study of some features of Agreement/ Disagreement. Unpublished Ph.D. Dissertation, University of California, Irvine. Quirk, Randolph et al. (1972). A Grammar of Contemporary English. (London: Longham Group Ltd.). Rosen, S. et al. (1974). “Discriminatory buck passing: delegating transmission of bad news.” Organizational Behavior and Human Performance, 12, 249–263. Sacks, Harvey (1974). “An Analysis of the Course of a Joke’s Telling in Conversation.” In J. Sherzer & R. Bauman (Eds.), Explorations in the Ethnography of Speaking. Cambridge: Cambridge University Press. Sacks, Harvey (1992). Lectures on Conversation. Edited by Gail Jefferson, with an Introduction by Emanuel A. Schegloff. Oxford: Blackwell.

Pre-announcement sequences 

Sacks, Harvey, Schegloff, Emanuel A., & Jefferson, Gail (1974). “A Simplest Systematics for the Organization of Turn-Taking for Conversation”. Language, 50 (4), 696–735. Schegloff, Emanuel A. (1968). “Sequencing in Conversational Openings.” American Anthropologist, 70, 1075–1095. Schegloff, Emanuel A. (1972). “Notes on a Conversational Practice: Formulating Place”. In D. Sudnow (Ed.), Studies in Social Interaction. New York: Free Press. Schegloff, Emanuel A. (1987) [1973]. “Recycled turn beginnings: a precise repair mechanism in conversation’s turn taking organization”. In G. Button & J. R. E. Lee (Eds.), Talk and social organization (pp. 70–85). Clevedon: Multilingual Matters. Schegloff, Emanuel A., & Sacks, Harvey (1973). “Opening up Closings.” Semiotica, 8, 289– 327. Searle, John R. (1970). Speech Acts: An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Sudnow, David (1967). Passing On: The Social Organization of Dying. New Jersey: PrenticeHall. Sudnow, David (Ed.) (1972). Studies in Social Interaction. New York: Free Press. Tesser, A., & S. Rosen (1975). “The Reluctance to Transmit Bad News.” In L. Berkowitz (Ed.), Advances in Experimental Social Psychology, Vol. 8. New York: Academic Press.

 Alene Kiku Terasaki

Appendix A Instances of Announcement Sequences (1) [TG:26] A: Tch! I’ll get some advance birthday cards, hhm hmh! (0.6) A: hhh A::nd uh, (0.5) A: Me:h, (0.2) A: Oh Sylvie’s sistuh hadda ba:by bo:way. B: Who? A: Sylvie’s sister. B: Oh really? A: Myeah,= B: (That’s) =[[ A: She had it yestihday.=Ten:: pou:nds. B: Je:sus Christ. A: She had//da ho:(hh)rse hh hh B: (That’s a ba:by.) (2) [JSLR:48-49] A: Oh you haftuh tell’m about y’r typewriter honey, B: Oh yes. C: Yeah didju hear from them? B: Yes, A: We had m- more // trouble, B: (Olympia’s gonna) put in a; A: Oh you told ’em I forgot. B: (They’re g’nna) put inna new, – keyboard. D: Oh they are?= C: =Fer nothing!? D: (Yeh) C: W’l good fer you::::! (3) [Trio:l8] A: Oh you know, Mittie- Gordon, eh- Gordon, Mittie’s husban’ died. (0.3) B: Oh whe::n. A: Well it was in the paper this morning.

Pre-announcement sequences 

B: It wa::s, A: Yeah, (1.2) B: You mean, it was in the obit colum? // or // -an article about ’im. A: Yeh. Y’know N:no, no, just in the, B: Oh, hell. I took my paper tuh- (0.6) -tuh work an’ I put’t in the wasteba//sket I didn’tA: Yeh, B: I[[ A: They jus’ said Gordon Fremont, y’know, husband a’ Jeanette,1 n all, ’n– (0.2) B: Oh::::. (4) [HS ST:2:4] A: Oh you didn’- You didn’ hear the news didju. We were out there before Thanksgiving. B: Oh. You were. A: Yeah. ((pause)) A: Weren’t we? B: Oh. Out here? A: Yeah. B: Yeah. Yeah. Right. Right. A: Yeah. Angie’s gonna have a baby. B: Oh really! A: Yeah. B: Well, congratulations A: Isn’t that sump’n // I didn’t think sheB: ((aside, off phone)) Angie’s gonna have a baby. (5) [CN II-11:4] A: Y’know I toldjuh thet uh they’d only been there a day’r so when he got sick. B: Mm hm, A: Well they, ((silent whisper)) had gotten married. (2.5) A: Yuh heard me, (0.5) B: Hitched? A: Yuh.

 Alene Kiku Terasaki

B: OH no::. A: Yeh. [[ B: Ohhhhhh my go(hh)d heh heh (6) [JSLR:38] A: Hey I got sump’n thet’s wi::ld. ((up amplitude)) B: What. (1.0) A: Y’know one’a these great big red fire alarm boxes thet’r on the corners? I got one. B: Really? A: With all the works The inside // en everything. C: How (grea:t)! [[ B: O::h (that’ll // be good.) (7) [NB:202] A: I’ve quit smokin y’know en ev’rything. (1.0) B: Well whenjeh stop that. A: The day you left. (0.6) B: Left where. A: From here in September. B: How may cigarettes yih had. (0.8) A: NO:NE. B: Oh really? A: No:. B: Very good. A: Very good. (8) [NB 4: 4-9] A: But uh I didn’t get home til, .hhh two las’night I met a very, very, nice guy. B: Didju:::. A: I really did // through these friends of mine, B: Goo::d. A: .hhhh // En it was ril cute I’ll ‘aftuh tell yih’bout it, B: Oh::::.

Pre-announcement sequences 

(9) [J&G:2] A: How ‘bout you Danny. How you doing school-wise?= B: =He hasn’ changed his hair color. ((ref. to earlier topic)) C: Nhh huh huh // huh D: huh huh [[ ( ): huh huh .hh E: Y’d prob’ly turn green with envy if I tolja all about what I got ( ) first // quarter. D: Tell’em abou//t it. E: Four A’s ‘n two B’s. (1.0) A: You were takin’ all those courses? E: Yeah. D: C’n you believe that? A: That’s outta sight Danny. (10) [FDII:88] A: Yeah, We got- we gotta little bit of it out here. ((an earthquake)) B: Yeh not too much though huh, A: We::ll, – Oh I got hurt a little bit last night, B: You did. A: Yeah, B: What happena you. A: Aw::: I- like tuh lost muh little finger, (0.2) A: They had me in surgery g’bout two an’ a half hours. – gettin’ // (it fixed). B: Aw:: Jeez. A: This mornin’. B: Doin’ alright now though huh, A: We::ll, it hurts a little, but I’m alright. B: Ye(hh)ah I c’n imagine how it hurts,

Collaborative turn sequences* Gene H. Lerner

In conversation, the pre-emptive completion of one speaker’s turn-constructional unit (Sacks, Schegloff, & Jefferson 1974) by a subsequent speaker can operate on that unit in a way that transforms it’s production into a sequence – a collaborative turn sequence. By completing the prior speaker’s turn-constructional unit (or TCU) – that is, by producing a version of what had been projected as a part of the prior speaker’s turn – a sequence can be initiated. Here a recipient responds to a prior speaker, not by waiting until completion to act, but by pre-empting that completion as a method of responding. When the completion is addressed to the original speaker in this way, the acceptability of the proffered completion can be implicated as a next action. The original speaker thereby maintains authority over the turn’s construction even when completed by another. Not all proffered completions result in such two-turn sequences. A TCU completion that is addressed to the original recipient of the turn-so-far (and thereby concludes the action implemented through that turn for its original recipient) will ordinarily not be treated as confirmable, and thus will not launch a sequence.1 In this circumstance, the shared authority inherent in the coconstruction of an action is left undisturbed by confirmation/disconfirmation, as when two participants co-construct an explanation (Lerner & Takagi 1999) or co-tell a story (Lerner 1992) for the same recipient.2 In contrast, the present chapter focuses on those proffered completions that do seem designed to launch a sequence – a collaborative turn sequence – in which the original speaker ordinarily reasserts authority over the turn’s talk by responding to the proffered completion (or by producing an alternative to it).

 Gene H. Lerner

The Affiliating utterance as TCU completion The affiliating utterance that launches a collaborative turn sequence can be characterized by the following observations.

1. Affiliating utterances are used in a wide range of interactional environments and with various types of compound TCUs. What is common to all of them is their use of the constructional format of the ongoing turn as a method to propose a version of the current speaker’s projected talk prior to its occurrence (or in the case of teasing or heckling, the affiliating utterance will use the constructional format to pointedly mis-project the content of the turn’s completion ). 2. Affiliating utterances are built to be contiguous with the preliminary component of the TCU-in-progress, and are placed in the ongoing turn in a way that displays this contiguity (i.e., through placement at an opportunity space). Thus, affiliating utterances maintain the progressivity of the talk (or at least display an orientation to maintaining progressivity) across a change in speakers.3 It is this feature that distinguishes a bid for speakership – in the course of another speaker’s TCU – that is only a bid for conditional access to that ongoing turn (for pre-emptive completion) from a bid for next speakership at the same point which is a bid to begin a new turn. With pre-emptive completion, the projected turn-constructional format remains unchanged, achieving a syntactically unmarked speaker transition. In contrast, other next speaker startups (in the course of an ongoing turn) often display an orientation to their out-of-turn character, i.e., their character as a bid for turn transition at other than a transition-relevance place. This can be done, for example, by using the turn initial position to account for a midutterance startup. (1) [JJ:Invitation] B: I was just gonna say come out and come over here and talk this evening. But if you’re going out // you can’t very well do that. C: Talk you mean get drunk, don’t you. B: what?

Here the turn-initial repeat of “talk” is used to tie C’s response to an object in an earlier TCU – once it becomes clear that the next TCU begun in the turn is moving away from the invitation report. Locating a response target that occurred prior to the most recent transition-relevance place provides a warrant

Collaborative turn sequences 

for C’s mid-utterance startup as a very late attempt at a response. That is, it locates the prior transition-relevance place as a transition space for a late-starting utterance.4 Progressivity is maintained by the contiguous placement of the preempting utterance. However, when contiguity is lost (e.g. when the affiliating utterance begins after the current speaker has begun the projected final component) immediate lexical progressivity is relaxed in order to achieve component contiguity. This suggests a relative ordering of component contiguity and sheer lexical progressivity in the sequential structure of the opportunity space. Component contiguity is maintained at the expense of progressivity. (cf. Lerner 1996a: 245–251).

3. Affiliating utterances are produced to bring the turn to completion, and to bring it to completion at the next possible completion. Further, affiliating utterances are oriented-to by their recipients as taking the already-projected turn unit to its next possible completion, and no farther. Massively, affiliating utterances go only to next possible completion.5 They are built as a continuation of the turn-in-progress and as a completion to that turn. The initiation of an affiliating utterance is not a bid for continued talk; it is a conditional entry device. What is projectable from the preliminary component of a turn-in-progress is a component type that will bring the turn to the next possible completion. However, when an affiliating utterance is initiated, the projected next possible completion is ordinarily as far as that speaker goes, and as far as the prior-as-next speaker allows them to go – the next speaker after the affiliating utterance treating the affiliating utterance as standing in for the completion of the turn it is tied to. In the following case, this orientation to the affiliating utterance being one turn component long (bringing the current compound TCU to only the next possible completion) is especially visible because the location of the next possible completion is itself unclear. (2) [BC:III:Green] Caller: ... you feel like you c’n handle evrything. BC: Mm:://hm::, Caller: and anything. BC: except a cross wind at the short end a’ that runway. . Caller: hhh Oh d’crosswind didn’ you know wid a J-T’ree, well BC: jus’ drop one wing an’ // slide in, Caller: That’s right,

 Gene H. Lerner

The speaker of the affiliating utterance (“jus’ drop one wing an’ slide in,”) produces only the completion of the current turn unit. However, this is overlapped by the next utterance, which is aimed at what could be taken to be an earlier possible completion. The overlap does not occur because the speaker of the affiliating utterance is attempting to produce more than the completion. It occurs because the location of the first possible completion after the onset of the pre-emptive completion is problematic for the participants (cf. Jefferson 1973).

4. Once an affiliating utterance is begun, it is treated by the original speaker of the TCU as a candidate completion. When a second speaker begins in the course of a turn that contains an opportunity space for such completions, and begins in the vicinity of that space, then that speaker’s utterance is examined for its within-turn sequential import. That is, it is taken as continuing the current TCU, and as a possible instance of that unit’s completion. This can be seen in the following instance. (3) [GTS] Ken: Louise: Ken:

insteada my grandmother offering him a drink, of beer she’ll say [ wouldju[wanna glassa milk? [ hehhh [no. wouldju like a little bitta he’ing? ((i.e. “herring”))

Here, the placement of the affiliating utterance (“wanna glassa milk”) is central to it not being taken for a new turn beginning. If the target turn is removed from the fragment as in (4), the sequence is transformed into something entirely different. (4) [GTS] Louise: Ken:

wanna glassa milk? [ hehhh [no. wouldju like a little bitta he’ing?

However, a quote is in force and it is projected to be an offer. By beginning the offer after the authorship attribution (“She’ll say”), Louise produces an utterance that is not taken to be an actual offer, but a report of an offer by Ken’s grandmother. The “no. wouldju like a little bitta he’ing?” is readily understood, not as a rejection of an offer (or even the answer to a question) followed by a counter-offer, but as a rejection of “wanna glassa milk” as a candidate instance of what Ken was about to say. Affiliating utterances require no special tying devices, while other sorts of mid-utterance startups systematically use such devices as misplacement mark-

Collaborative turn sequences 

ers to suspend the “nextness” for which the utterance will otherwise be inspected. Objects placed in the opportunity space are inspected for their turn continuation features. This is the case even when the affiliating utterance could otherwise be heard to be the beginning of a new turn as in (3). In summary, affiliating utterances 1. use the format of the TCU-inprogress, 2. maintain the progressivity of the utterance from an opportunity space, 3. bring the turn unit-in-progress to completion and 4. are treated as candidate versions of what was about to have been said. That is, affiliating utterances are built as and treated as a turn-completing action.6

The collaborative turn sequence In conversation, turn size is locally managed and interactionally achieved through a turn-taking system – that is, turn size is not prearranged. Therefore, as far as the turn-taking system is concerned, any next possible completion may or may not turn out to be the actual completion of a TCU, and may or may not turn out to be the place at which transition to next speaker is accomplished. However, sequence organization influences turn-taking outcomes. When an adjacency pair first pair-part occupies a TCU, that unit is likely to be the last TCU a speaker produces prior to speaker transition (Schegloff & Sacks 1973). This constraint on turn organization is in the interest of preserving the contiguity of the first and second pair-parts of the sequence (Sacks 1987). Sequence contiguity provides one systematic basis for the empirical finding that pre-empting utterances (at least those that are designed to launch a sequence) are limited to a single component. Pre-emptive completions are ordinarily produced as a rendition of ‘what the other was going to say’ but are not composed as a guess (e.g. with a try-marker) that would explicitly inviting acceptance or rejection. However, pre-emptive completions are taken by the original speaker of the TCU they complete as candidate completions (implementing an action) that can be accepted or rejected. The production of 1) a TCU pre-emptive completion by 2) an addressed recipient of an ongoing turn and 3) addressed to that turn’s original speaker selects that last speaker as next speaker, and sequentially implicates as a next action, the acceptability of the pre-empting utterance as a completion for the turn. This is the collaborative turn sequence. A collaborative turn sequence is a collaboration of two speakers producing a single syntactic unit not only in that a next speaker produces the comple-

 Gene H. Lerner

tion to a TCU begun by a prior speaker, and that prior speaker does not continue once the pre-emptive completion begins, but also in that the first speaker ratifies the completion after its occurrence as an adequate rendition of the completion of the TCU they were about to voice. The acceptance/rejection of the pre-emptive completion – or more precisely, the receipt of the completion in terms germane to the action implemented by the proposed completion – is the relevant next action. 7 However, outright rejection rarely occurs in the receipt slot. For one thing, a pre-emptive completion need not be done, since the action it implements does not ordinarily become conditionally relevant on the completion of the preliminary component of the ongoing turn. Responding actions, for the most part, are made conditionally relevant on completion of the sequence-initiating action – that is, on completion of the turn that carries the action. The completion is occasioned by the production of a preliminary component of a compound TCU, and it can be responsive to the action the ongoing turn is implementing, but in most circumstances that responding action is not called for until possible completion of the full TCU. Responding after the preliminary component, then, is in a sense optional, while responding upon completion may not be. Thus a recipient is free to offer a completion or not under most circumstances.8 The receipt slot is the place for the original speaker of the TCU to acknowledge the other speaker’s completion as a continuation of their turn-in-progress by accepting it as in (5) or rejecting it as in (6).9 (5) [Theodore] A: if you start watering, it [will get greeB: [it will come back A: y- yes uh huh (6) [H and M] Hal: ... the answer is perhaps, though I don’t really know, that it isn’t a substitution, Max: it’s a transformation. Hal: No it’s not even a transformation

Other forms of completion-acknowledging response are possible – that fall somewhere between acceptance and outright rejection as in (7).

Collaborative turn sequences

(7) [SEWING] (The participants are making a pillow.) Daughter: Oh here dad (0.2) a good way to get tho:se corners out (0.2) Dad: is to stick yer finger insi:de. Daughter: well, that’s one way.

In (7) Daughter seems to reluctantly accept Dad’s method, but in a way that makes it clear that his proffered method was not the advice she was about to give – though it does acknowledge his completion as the proper type of completion, if not the exact one she had in mind. Another way to acknowledge the pre-emptive completion is for the original speaker to repeat the just-produced completion as in (8). (8) [DTA: simplified] B: you don’t go primarily because alcohol is obtainable there. You go there cz its a whole social interaction. Your gonna be doing other things, your gonna hustle ladies, your gonna see stuff [ yer yer C: [( ) gonna meet people B: yer gonna meet people you know

When this second completion can be understood as substantially repeating the pre-emptive completion, then the original speaker can be seen as acknowledging and accepting it by now incorporating it into their own completion.10 The collaborative turn sequence occurs, not across two distinct turns separated by a transition-relevance place, but within the purview of one participant’s turn at talk – within that participant’s projected turn space. The collaborative turn sequence’s sequence-initiating utterance occurs within the turn space of another participant, thereby colliding with the standard turn-taking practice of speaker change taking place at possible completion of TCUs. It is the onset of this second utterance – in this manner – that initiates and shapes the sequence, and informs what action is accomplished through it.11 The placement and form of pre-emptive completion re-produces the features of ordinary turn-taking in which one participant speaks at a time and turns are comprised of complete TCUs, while relaxing only the entitlement of a speaker to produce a complete TCU on their own. An orientation to a participant’s right to complete one TCU, having been allocated a turn at talk, is nevertheless shown both by the initiation of a small sequence (the collaborative turn sequence) that allocates the next turn to the original speaker of the TCU in order to ratify the proffered completion as in fact ‘what was about to have



 Gene H. Lerner

been said’, and by the use of the receipt slot alternative action of delayed completion, a turn-taking system repair device (described in the next section) that allows the original speaker of the TCU to regain speakership of their original turn and complete the original TCU themselves. One way to further explicate the organization of collaborative turn sequences is through a comparison with repair. There is a striking parallel between pre-emptive completion and other-initiated repair (Schegloff, Jefferson, & Sacks 1977). Both are launched by sequence-initiating actions that locate the just prior speaker’s talk as the object on which the sequence operates.12 Furthermore, both pre-emptive completions (by an addressed recipient) and next-turn repair initiators systematically select that just prior speaker as next speaker. Speaker selection occurs without the use of an address term in both cases. Not only is an address term unnecessary, but will not properly occur. This displays an orientation by participants to the right a speaker has to maintain control or authority over a turn to its completion. That is, the speaker who begins a turn maintains a right to determine what the utterance in that turn space will come to be. For next-turn repair initiators, this means having a right to repair any trouble even though it might be located by another participant. For collaborative turn sequences, this means maintaining control over what the completion will look like even in the face of a second participant making a bid for speakership within their turn space. Both the speaker of the original turn-in-progress and the speaker of the pre-emptive completion orient to the turn after the pre-emptive completion as a turn allocated to the original speaker. This can be seen clearly in instance (9) in which the addressed recipient of the turn-in-progress (Daughter) produces the pre-emptive completion, but addresses it to Dad, rather than the TCU’s original speaker (Mom), thus, in a sense, markedly “shifting” recipients. (9) [MOTHER’S DAY] Mom: if you should decide to live with a fella (.) Daughter: Mom will still talk to me ((laugh)) Mom: I’ll still talk ta ya // but please take- please take the pill Dad: I d- I didn’ s+

Here, Mom (the original speaker) does a receipt in the receipt slot position, while Dad, who is the addressed recipient of the completion, holds off his denial until sequence completion. The following instance also suggests that participants other than those involved in producing the collaborative turn sequence take the pre-emptive

Collaborative turn sequences 

completion to be within the current turn space, and treat it as initiating a two-turn sequence with prior speaker selected as next speaker. (10) [GTS] Dan: Roger: Dan: Ken:

but it seemed to be, to Ken at least the wrong kind the wrong kind // of distinction well you don’t wanna- I mean

A participant referred to (as opposed to addressed) in the course of a turn can be topically in a position to speak in the turn after next turn – next turn regularly going to a selected recipient of the current turn. (The mentioned participant is excised – in and through the reference – from those present who could be considered a selected recipient.) In the above instance, Ken is referred to by name, but does not begin to speak until a possible completion is reached in the course of Dan’s receipt. The mentioned participant does not begin after the next speaker (Roger), but waits until sequence completion. Thus, there seems to be an orientation to the receipt slot as a place for the original speaker to address the adequacy of the proffered completion. However, someone other than the original speaker may speak after the proffered completion (i.e. in the receipt slot), or some action other than a receipt may be produced in this slot by the original speaker. First I examine cases in which the speaker of the pre-emptive completion (and not the original speaker) speaks in the receipt slot, and then in later sections I examine other actions the original speaker can take in the receipt slot.

Claiming authoritative knowledge over a pre-empted completion Though the speaker of an affiliating utterance regularly produces just the completion for the TCU-in-progress and then stops (allowing the original speaker to acknowledge the completion and assess its acceptability), a pre-empting speaker occasionally appends an agreement token to the pre-emptive completion – thus, in effect, pre-empting the receipt slot, as well as the turn’s completion. In (11), which includes two instances of this [completion + agreement token] format, Ann begins by saying that she is not sure of what she is about to propose. Since these assertions are ‘produced with uncertainty,’ confirmation seems like a particularly relevant next action for Jenny, as a way of revealing or

 Gene H. Lerner

claiming more authoritative knowledge about a matter presented to her with uncertainty. (11) [Rahman:C:2 (simplified)] Ann: Now we’ve been told but I don’t know whether this is true:. Jenny: Mm::, Ann: that (0.9) no wa:y can you get out, Jenny:
In both cases, Jenny not only claims her authority by confirming Ann’s understanding, but she demonstrates her authoritative knowledge by producing the key elements herself. It is just when a speaker is uncertain about their own utterance (or when a recipient can otherwise claim superior personal or membership categorical authority) that the receipt slot may be pre-empted by the speaker of the pre-emptive completion. The uncertainty (or other reduced entitlement to the utterance they are voicing) weakens the original speaker’s entitlement to confirm the appropriateness of the completion.13 In addition, this format can be used to confirm a pointedly unfinished (and possibly trailed-off) delicate assertion as in (12). (Note that speaker C is actually referring to “grampa” at the beginning of line 1, though she could seem to be addressing him.) (12) [PV] C: well grampa it’s bad enough when ’e when he uhm::=tells you how much t’make, but when ’e tells you what t’ co:ok, (0.3) M: then it’s rilly bad=yeah. yeah // didju- uh C: He tol me t’ be sure’n check iz minu up there on the wall

Here, M offering the strongly projected, but unspoken, completion as an independently arrived at negative appraisal by quickly appending agreement tokens to it. In this way, she might be said to take equal responsibility for the negative appraisal and thereby for the complaint it helps constitute.14 In the next sections, I turn to a discussion of two receipt slot alternatives available to the recipient of the pre-emptive completion (i.e. the original speaker of the TCU): delayed completion and list construction.

Collaborative turn sequences 

Delayed completion as a receipt slot alternative to acceptance/rejection15 A recipient of a turn-in-progress can produce a pre-emptive completion which is offered as an assertedly correct completion (i.e., not as a candidate or heckling response ), but which is not accepted – but which is also not rejected.16 One device which provides a receipt slot alternative to pre-emptive completion acceptance and rejection is delayed completion. It was noted that pre-emptive completions are rarely rejected. An examination of delayed completion will provide one systematic basis for this finding. Delayed completion is a device used to link a speaker’s current utterance, across the talk of another participant, to their prior syntactically unfinished utterance, by constructing the current utterance as a syntactically fitted continuation of their own prior utterance which completes the turn unit begun in that earlier utterance.17 Delayed completion can provide a way to bring a current TCU to completion in the clear after another speaker initiates an utterance in the course of that turn unit.18 (13) [GTS] Ken: Seems like every week somebody- somebody in this group gets stepped on royal. Somebody gets- gets com Roger: Why doncha all step on me. Ken: pletely cremated.

Its use provides a way to resolve an overlap of utterances by stopping talking, while not losing the competition for the turn space. (14) [GTS] Dan: as a matter of fact we may not have a group going after // the uh Roger: maybe you’re screening ’em too hard Dan: next couple of weeks

Delayed completion is a current-speaker device for handling onset of speech by another participant within a TCU – with or without simultaneous speech, as in (14) and (13), respectively. For a next-speaker device see Schegloff ’s (1987) description of “recycled turn beginnings.” In addition, delayed completion can be used to turn a possibly complete TCU into merely the first part of a TCU the delayed completion now finishes.

 Gene H. Lerner

(15) [GTS] Roger: I don’ wanna accept that responsibility. (0.7) cause I’m not trained along those lines (.) Dan: Mh= Roger: =so I wanna bu- I don’ wanna raise an underachiever. (0.2) an an i(f)- ’n further the problem. (0.4) perpetuate the=uh (0.3) underachiever, (0.6) so ah’ll just leave it (.) to somebody who is (0.2) Dan: you’re not going to have children? (.) Roger: so trained, (.) t’have children.

Here the delayed completion (“so trained, (.) t’have children.”) locates the intervening utterance as interruptive of the TCU-in-progress. As a continuation of the speaker’s prior utterance it interdicts the sequential implicativeness of Dan’s intervening question. Though the question is in effect answered, it is not built as an answer to a question. Delayed completion can also provide a warrant for the initiation of overlap, through its claim of the unfinishedness of a speaker’s just prior utterance. (16) [GTS] Ken: My opinion of the school system, the Los Angeles school district, dis- district, is the most fucked over, Roger: Yeah well we // all got that opinion. Ken: school syst’m WAIT is the most fucked over school system in the world. (17) [Frankel: House Burning] Pen: I don’ wanna make yih ta:lk cuz I don’t wantche tuh: (.) Pat: No: I f- I really do feel a lot // bettuh (I feel like) Pen: upset chiself a*ll over agai:n,

By producing an utterance which is a syntactic continuation of their own prior utterance a speaker claims to be merely continuing a TCU-in-progress. The continuation (i.e., the delayed completion) is thereby asserted to be part of the same turn space occupied by that prior utterance, making the intervening utterance out to be interruptive of that turn space whether or not there is any overlap.

Collaborative turn sequences 

So, delayed completion can 1. provide a means to produce a complete TCU across intervening talk, 2. make out an intervening utterance to have been interruptive of a turn at talking, 3. provide a warrant for the initiation of overlap, and 4. interdict the sequential implicativeness of the intervening talk. With these features of delayed completion in hand, we can now look at the use of this device as a receipt slot alternative in the construction of collaborative turn sequences. Pre-emptive completion is a bid to enter the ongoing turn of another speaker. Its success hinges on the way the sequence runs off. A delayed completion can be used to interdict the (collaborative turn) sequential relevance of the pre-empting completion. In (18) a delayed completion (“then you think he’s gonna fall asleep”) realizes a competing completion. It doesn’t not acknowledge or ratify the pre-emptive completion. (18) [GTS] Ken: Louise: Ken: Louise:

no its when he turns a bright red that everybody has to start worrying no when he gets his eyes like this an’ he starts thinkin, you know then you get to worry then you think he’s gonna fall asleep.

And in (19), an interrupting (though not disagreeing) delayed completion interdicts the pre-empting utterance and its sequentially implicated receipt by not allowing it to come to completion in the clear and by outlasting it. (19) [Labov:TA] C: Fact I said tuh Larry yuh don’t think it’s- thet- y’know thet the kids thet’r skinny, (0.7) are gonnuh yihknow haftuh worry about it. They c’n eat twice iz much iz you, D: en it doesn’t mean // anything C: en not gain wei//::ght.

When the receipt slot alternative of delayed completion is produced, allowing the original speaker of the TCU to complete the syntactic unit and interdict the sequential relevance of the pre-emptive completion, the understanding, agreement or other responding action (first attempted through the pre-emptive completion) can remain relevant for next turn. Pre-emptive completion provides a way to display understanding of or agreement with an ongoing turn. It can be, for example, a same-turn alternative to producing an agreement token in next turn to show understanding or agreement.

 Gene H. Lerner

However, once a delayed completion is produced, its receipt – and through it the receipt of the original speaker’s turn as a whole – then becomes a relevant action for next turn. Next turn is now, again, a sequential slot for the original recipient (i.e., the speaker of the pre-emptive completion) to respond. This can be seen in (20). (20) [HIC] Sparky: Kerry: Sparky: Kerry:

it sounds like what you’re saying is that let them make the decisions an let us know wh//at it is and let us know what it is yeah

Here delayed completion is used to reclaim the speakership of a TCU, thus reversing again the earlier reversal of speakership/recipientship that was accomplished by the pre-emptive completion.19 In this way, the collaborative turn sequence is transformed from: preliminary component (completion source) pre-emptive completion (sequence-initiating action) receipt (sequence-responding action)

into: preliminary component (completion source) pre-emptive completion (sequence-initiating action) delayed completion (original turn/action completion) receipt (response to completed turn)

An attempted pre-emptive display of understanding is transformed into a next turn token assertion of understanding. The expanded form is, however, still a two-turn sequence. Delayed completion and receipt of the pre-emptive completion are slot alternatives. One can find both occurring after the production of pre-empting utterances. However, since pre-emptive completion by an addressed recipient of an ongoing turn selects a particular party – last speaker – as next, one would not expect to get both an acknowledgement and a delayed completion occurring in the same instance – produced by different participants. There is, however, at least one environment where this can occur. When the party addressed by the pre-emptive completion is a multipleparticipant party (cf. Lerner 1993; Schegloff 1995a), then any member of the party can be a proper next speaker, since the party, and not a particular member

Collaborative turn sequences 

of that party, has been selected as next speaker. Thus, more than one member may select to speak as Dad and Mom do in (21). (21) [HIC] Dad: Kerry: Dad: Mom:

he’s guaranteed, but (.) the rest of the members cn control that member from him: goin any higher [ than tha[t [goin hi [gher than that [ye::ah

Both co-members of the party (here Dad and Mom as co-explainers) may talk in the receipt slot and as in (21) one produces a delayed completion in the course of the pre-emptive completion, while the other produces a receipt of the pre-emptive completion on its completion. Two analytically distinct aspects of receipt slot work are discernable from the above discussion. (That is, there are systematically available practices to underwrite this distinction.) One job of a speaker selected to speak in the receipt position is the acknowledgement of the pre-emptive completion as the de facto completion of the turn. That is, a preliminary component speaker is in a position to either acknowledge or disregard the pre-empting utterance. A second job is the explicit acceptance or rejection of the completion as an instance of what was about to be said. It is possible to acknowledge the pre-empting utterance as a TCU completion without agreeing with it. This can be seen in (22). The daughter’s utterance (“well, that’s one way.”) acknowledges the completion (“is to stick yer finger inside”), but in so doing shows that it was not what was about to be said – though it may well be an acceptable alternative to the pre-empted advice. (22) [SEWING] ((The participants are making a pillow.)) Daughter: Oh here dad (0.2) a good way to get tho:se corners out (0.2) Dad: is to stick yer finger insi:de. Daughter: well, that’s one way.

The following data array illustrates the separation of the two tasks. 1. Acknowledgement of the completion through agreement with it

 Gene H. Lerner

(23) [CDHQ:II:3] Marty: Now most machines don’t record that slow. So I’d wanna- when I make a tape, Josh: be able tuh speed it up. Marty: Yeah.

2. Acknowledgement of the completion through disagreement (24) [HYLA:simplified] Hyla: I wz deciding if if I sh’d write im the thankyou no:te // fer the birthday gi:ft, Nancy: Yea:h Hyla: hh.hh I decided no:t to // though Nancy: How co:me, . Hyla: t hhhhh (.) Becuz I figure, hhhh//hhh Nancy: If ’e hasn’ written ye:t, (0.4) then ’e doesn’ want to. (0.2) Hyla: Oh:: don’t say thahhh//a(h)t Nancy: NO is tha’whatcher think//ing? Hyla: No::,

3. Disregard of the completion through delayed completion (25) [Labov:TA] C: Fact I said tuh Larry yuh don’t think it’s- thet- y’know thet the kids thet’r skinny, (0.7) are gonnuh yihknow haftuh worry about it. They c’n eat twice iz much iz you, D: en it doesn’t mean // anything C: en not gain wei//::ght. D: Right.

The form for acknowledgement of the completion (as the de facto turn unit completion) is the acceptance (23) or rejection (24) of the pre-emptive completion as an instance of what was about to be said, while the form used to leave the completion unacknowledged is primarily delayed completion (25) of the TCU by the turn unit’s original speaker. This interdicts the sequential relevance of the pre-emptive completion. The acceptance or rejection of the pre-emptive completion as ‘what was about to be said’ ratifies the pre-emptive completion as an acknowledged completion qua completion. Rejection of the pre-emptive completion (as an instance of what was about to be said) can be circumventing. This can be done by producing a delayed completion, rather than ratifying the pre-emptive comple-

Collaborative turn sequences 

tion’s status as a continuation of the turn unit-in-progress by disagreeing with it. This is another way that a preference for agreement is sustained in the practices of talk-in-interaction. As such, pre-emptive completion of a compound TCU provides a systematic locus for delayed completion. Before moving on to another type of receipt slot practice (list construction), the relationship of collaborative turn sequences to overlap management can be sketch out. The following section describes how pre-emptive completion and actions ensuing from it provide systematic sites for overlap.

A place for overlap Collaborative turn sequences constitute a systematic site for the occurrence of overlapping talk. Pre-emptive completions are built to continue a turn from a projected opportunity space. However, not all attempts at collaborative completion are successful. The speaker of the compound TCU-in-progress can talk through the onset of a developing pre-emptive completion and thereby retain control of the turn space. (26) [HIC] ((Sparky is addressing Kerry)) Sparky: it sounds like what you’re saying is that // let them make the decisions Kerry: (if this is)

Here Sparky continues talking after Kerry begins; Kerry then stops without finishing the pre-emptive completion. The attempted collaboration fails, since the original speaker continues (cf. Jefferson 1983). This can also be seen in (27), but in this case the original speaker retrieves the overlapped pre-emptive completion after finishing the TCU involved in the overlap. (cf. Jefferson, this volume). (27) [ADATO] J: Well it’s a, it’s a mideastern yihknow it’s- they make it in Greece, Turke::y, //right around there. B: Armenia, J: Yeah, Armenia

So, overlapping talk can occur when the target speaker continues talking beyond the opportunity space, as in (26) and (27). Overlap can also occur when the speaker of the pre-emptive completion initiates the utterance outside of the slight pause between components.

 Gene H. Lerner

(28) [HIC] Sparky: Dad:

when it doesn’ invol:ve thee basic agreement, it is b://y stock its by major*ity

Here Sparky has already begun the final component of his compound TCU, when Dad begins his “late” pre-emptive completion. In addition, delayed completion also provides a systematic place within collaborative turn sequence organization for simultaneous talk. The turn after the pre-emptive completion is allocated to the prior speaker and is allocated to that participant to produce a receipt. To interdict the pre-emptive completion, the utterance done in the receipt slot must be heard as a delayed completion of the preliminary component, and not as a receipt of the pre-emptive completion. Some delayed completions can be seen to be interdictive of the prior utterance through their formulation as, for example, a competing completion as in (29). (29) [GTS] Ken: Louise: Ken: Louise:

no its when he turns a bright red that everybody has to start worrying no when he gets his eyes like this an’ he starts thinkin, you know then you get to worry then you think he’s gonna fall asleep.

The discordance of a competing completion constitutes it as a rival completion for the turn unit. However, a delayed completion that repeats, extends, or perhaps even slightly revises a successfully completed pre-emptive completion may be taken to be an acknowledging receipt of the pre-emptive completion – unless the delayed completion is somehow made to be seen as a replacement for it. Position can achieve this, when composition does not – that is, overlapping the ongoing pre-emptive completion can interdiction it’s relevance for subsequent action, when the content of the delayed completion does not do so. Starting the delayed completion in the course of the pre-emptive completion as in (30) can accomplish this. (30) [HIC] Sparky: Kerry: Sparky: Kerry:

it sounds like what you’re saying is that let them make the decisions an let us know wh//at it is and let us know what it is yeah

Collaborative turn sequences 

Pre-emptive completion provides a systematic place for delayed completion – and delayed (especially non-discordant) completion provides a systematic place for overlap. One practice associated with overlap management in conversation is the post-overlap retrieval of a previously overlapped utterance (Jefferson, this volume). For instance: (31) [GTS] Dan: Louise: Dan:

Like Ken is describing a guy who c’n sort of (1.0) get [ in with the group= [be himself. =but still be himself,

When overlap of the pre-emptive completion with a continuing (or delayed) final component occurs then both systematics (collaborative turn sequence and overlap management) are concurrently relevant. Since overlap and pre-emptive completion intersect here, the next position after an overlapping completion may be both a receipt slot position and an overlap retrieval position. In this case, the two interactional tasks – the ratification/interdiction of a pre-emptive completion and the retrieval of an overlapped utterance – can be accomplished concurrently. The post overlap/receipt position can be used to retrieve the overlapped pre-emptive completion as Dan (the original speaker of the TCU) does in (32). (32) [GTS:5] Dan: And as you said there’s a who:le segment of our society, (.) that is [dropping out.] Roger: [are failing. ] (Right). Dan: are failing. Sure.

Or this position can be used to reassert the original speaker’s own formulation of the events as Ken does in (33) by adding an increment to his own completion. (33) [GTS] Ken: Louise: Ken:

And you think I really got pleasure out of getting uh well Igetti//ng in that debate? stomped on. Cause that’s what it ended up to be, a big debate.

In addition, when overlap occurs, the speaker who initiated the pre-empting utterance may also engage in post-overlap utterance retrieval as D does in (34).

 Gene H. Lerner

(34) [Gerald] R: if you don’t put things on yer calendar (.) [(f o r g e t I t)] D: [yer outta luck.] yeah(p). fo:getit

In this way, D ratifies R’s completion, thereby explicitly marking the preemptive completion as unsuccessful. The use of the receipt position to accept a pre-emptive completion or reclaim the turn space by producing a delayed completion, and the post-overlap retrieval of an utterance are both aimed at the same result: they are ways of reasserting a claim over a turn after another participant has either made a claim for the turn space (overlap competition), or has made a conditional entry into the turn space (pre-emptive completion), or both.

List construction as a receipt slot alternative to acceptance/rejection20 Jefferson (1990) has shown that “less-than-three-item three-part lists are recurrently constructed by occupying a third slot with a generalized list completer.” She gives the following instance as an example. (35) [JG:II(a):3] Heather: And they had like a concession stand like at a fair where you can buy coke and popcorn and that type of thing.

Here the list, “coke and popcorn and that type of thing,” contains two items and is followed by a generalized list completer which locates the first two items as members of a class. This type of list construction seems to be a way to formulate a reference to a class of items. The reference includes the listed items but is not limited to them. The construction of a list in the receipt slot, incorporating the just produced pre-emptive completion as an item, can both propose an initial acceptance of the affiliating utterance and then transform that acceptance into something else as in (36). (36) [GTS] Ken:

Louise: Ken:

He said all the colored people uh walk- walk down the street and they may be all dressed up or somethin and these guys eh white- white guys’ll come by with . hh mud. mud, ink or anything and throw it at ’em

Collaborative turn sequences 

The (unprojected opportunity space) terminal item completion, “mud,” is followed in receipt position by a repeat of the completion by Ken. This alone could constitute an acceptance of the completion, but “mud” becomes the first item in a list. Incorporation of another speaker’s utterance into a list proposes that that item is one among others, rather than the single, correct (i.e., acceptable) item. The availability of this analysis to participants can be seen in instance (37). Here Jay uses a list structure to propose a series of clues in pursuit of a recognition by Sy of the object referenced as “my box.” In the course of the recognition pursuit, Jim proposes “Forms.” This utterance is not built as a continuation of the list. It is not linked to the prior with a conjunction and is not produced with an upward intonation. In fact, it is produced with a distinct downward intonation. (37) [ADATO] Jay: I- I told Jim thet uh, I wz going to:: leave you my um, my box of, –thet I use? Sy: What box, Jay: fer the SLIPS? ’n PAPERS? Jim: Forms. (pause) Jay: ’N FORMS? Jim: Forms.

The form of the repeat is what is of interest here. The receipt of Jim’s utterance is constructed as the third item of the list of clues. This is constructed by Jay and understood by Jim to be another clue in a series of clues. In the last line, Jim displays this analysis of the receipt by repeating his utterance in its original form, reasserting that it is a substitution for the prior list items, and not simply another clue. The incorporation of an utterance into a list is a device for accepting a candidate in a way which also displays that it is not the exclusive acceptable candidate. Whereas, a simple repeat in the receipt slot can be used to indicate acceptance of the pre-emptive completion. In (36), Ken repeats the terminal item completion. However, the utterance that continues on from that item is constructed as a list, and therefore the repeat of “mud” comes to be available as having been the first item of a list. And since it is a list of the [item + item + generalized list completer] form, the sort of object being referenced is transformed from the items themselves into the class to which the first two items belong. (One might think of the “generalized list completer” as a generalizing list completer.) In this way, the grounds on

 Gene H. Lerner

which the pre-emptive completion is accepted is changed from an acceptance of “mud,” to an acceptance of the class of objects (perhaps throwable objects) to which “mud” belongs. It accomplishes a move away from acceptance, without outright rejection of the candidate. List construction can also be used in receipt position to retrieve an utterance from overlap as in (38). (38) [GTS] Dan: Roger: Dan:

well I do know last week that uh Al was certainly very (0.6) pi//ssed off upset, ’n pissed off, ’n angry en w’z bout ready tuh fi:ght+uh with Ken

In this instance, “upset” as a characterization of how Al was feeling last week could be a possible completion of the TCU-in-progress. One feature of list organization shown here is that a single, in itself sufficient, person characterization term can be turned into having been a first term rather than the only term. In this way, list construction can be used as an overlap resolution device. In this instance, the list is produced using the item (“upset”) + item (“pissed off ”) + item (“angry”) format. The list construction format seems to be a way to acknowledge the overlapped pre-emptive completion, while not endorsing it. Though the second and third items seem to be different formulations of the same reference, both of which are upgrades of the original characterization (“upset”), they are produced in a way (placing conjunctions between the items) that formulates them as, in fact, somewhat different states. This list format allows a shift to “angry” rather than a substitution of “angry” for “pissed off.” In this way, the pre-emptive completion is not rejected outright, but it is also not acknowledged as the sole acceptable item projected to follow “very.” As in instance (36), the construction of a list in (38) provides a way to acknowledge another speaker’s terminal item completion. The turn is extended by turning the final word of the TCU into the first item in a list. Dan incorporates the overlapped utterance as the next item in the list. By producing the acknowledgement as a part of a list, the speaker can move away from the acceptance of the candidate as the sole completion of the source turn-in-progress. These technically-described practices underwrite a solution to a local problem. In (38), Dan is producing a ‘delicate’ characterization. It is marked as such through the break in progressivity just prior to the terminal item of his turn. The search for a “just right” word can reveal the searched-for word as delicate

Collaborative turn sequences 

(Lerner 1999). Also, he chooses the careful term “upset.” One issue seems to be how to refer to the way one of the participants was feeling at the group’s last meeting. Roger is proposing what might be seen by this group to be a teenager/hotrodder version (cf. Sacks 1979), while Dan seems to be proposing, with “upset,” an adult/therapist version. Now given that Dan is doing therapist’s talk, and also taking it that one part of therapist’s talk seems to be to acknowledge the talk of the participants, a problem occurs. How can Dan continue to talk as a therapist, while at the same time displaying ‘empathy’ with his clients? The use of a list format in the receipt slot provides a solution. Its use allows a shift from, but not abandonment of “upset.” After acknowledging “pissed off,” Dan produces the third item (“angry”). This retains the upgrade of “pissed off,” while returning to the adult/therapist register of “upset.”

Rejected pre-emptive completions The initiation of a pre-emptive completion implicates a receipt for next turn. Pre-emptive completions are not built as candidates, but are produced as assertedly correct completions for the TCU-in-progress. The receipt slot provides a place to acknowledge the pre-emptive completion as the proper continuation of the turn unit-in-progress. However, the receipt slot alternative to acceptance is ordinarily not the explicit rejection of the completion. Delayed completion can be used by the original speaker to produce an alternative, possibly competing completion that interdicts the sequential relevance of the pre-emptive completion. Nevertheless receipt slot rejection does occasionally occur as in (39). (39) [MOTHER’S DAY] D: .h and they do thi:ngs (1.2) C: ta hurt th’m (0.2) huh huh= D: =no:no: I’m sayin that ugh ugh (0.7) that’s the compromise they have ta do with themselves

Overwhelmingly, collaborative turn sequences that include explicit rejection of the pre-emptive completion also contain laugh tokens produced by the speaker of the pre-emptive completion. This can be appended to the completion as in (39) or it can occur as in-speech laughter as in (40).

 Gene H. Lerner

(40) [GTS] Ken: Roger: Ken: Ken:

you gotta get up. G’way. Y’gotta get up. ((heavy whisper)) put the dogs out. Feed the fish hyuhh heh hehh t(hh)ake the fi(hh)sh fer a wa(h)lk hh hehh // hheh hh //hehh No He takes the dogs, heaves em in the backyard,

In (41), the pre-empting speaker does append a laugh token to her completion, however, the receipt slot rejection begins simultaneously with it and so cannot be said to be responsive to it. Yet, the laugh token does show this was not produced as a serious contender for what Ken’s grandmother offers his father (“She’ll say”) (41) [GTS] Ken: Louise: Ken: Louise: Ken: Louise: Ken:

she’ll say // wouldjawanna glassa milk? // hehhh No. wouldju like a little bitta he’ing? heh// ha ha wouldja like some crekles? ehh ha ha ha ha wouldja like a peanut butter an’ jelly sandwich?

It is these recognizably not serious completions that original speakers reject. The point is that the pre-emptive completion is produced to be rejectable. It is produced as not serious and is marked by the laugh tokens to display that it is not being produced as assertedly correct. It is done as a recognizably not serious version of what the other was about to say. But it can nonetheless be inaccurate in a way that displays by its very selection that it has resulted from accurate knowledge of what sort of utterance had been projected to complete the TCU. For example, in (41), “glass of milk” catches perfectly the complaint being issued by the original speaker – that his grandmother treats his father like a child.21 The following instance reveals that it is actually not necessary for the pre-empting speaker to employ laugh tokens at all to display the not serious character of the completion.

Collaborative turn sequences 

(42) [GTS] Ken: Roger: Ken: Ken: Jim:

I was on a road on the way to Roswell New Mexico. An’ I was, y’know, plonkin along at a regular speed, an’ a grasshopper // jumped onto the road who you recognized. An’- wait minute No. //No I was goin along, an’ I had ehhhhha ha: hhhh hhh

Here, the not serious character of the utterance itself (invoking an earlier reference to “grasshoppers” in a rather nonsensical way) shows that Roger is putting words into Ken’s mouth, and is therefore built to be rejected. One might even say that it requires a denial to be successful (cf. Drew 1987). Since Roger has an audience here, he need not initiate the laughter himself (Glenn 1989).

Concluding remarks Utterances which continue an in-progress TCU and are addressed to its speaker make relevant a next action and select a next speaker. They select last speaker as next and make relevant as a next action the acceptability of the affiliated utterance. Pre-emptive completions are noteworthy as sequence-initiating actions (or Firsts) not only in that their production makes relevant a characterizable set of second actions, but especially in that both their placement (at a projected or unprojected opportunity space) and composition (as a syntactic unit completion) can be characterized formally. The production of a pre-emptive completion proposes for the original speaker of the turn-in-progress the alternative possibilities of retrieving (continuing) the turn from the opportunity space or addressing the completion as a continuation of the ongoing turn. What a TCU-in-progress will come to be remains the province of the speaker that initiated the unit. Collaborative turn sequences represent evidence of an orientation to speaking turns in a series as including an orientation to turn spaces in a series, where entry of a speaker into the turn space of another party is interactionally organized as a conditional entry. This is the interactional import of the receipt position – as the place for the original speaker to maintain authority– i.e., authorship – over their turn’s talk. The use of delayed completion can be seen to, in a sense, push back the recipe slot to next turn, rather than displace it. Since pre-emptive completion implements a form of early response to a turn, delayed completion returns the

 Gene H. Lerner

relevance of the receipt to its original place, thus co-constructing as a four action sequence [preliminary component + pre-emptive completion + delayed completion + receipt] (issuing from the pre-emptive completion), what on other occasions can be accomplished in its unexpanded form through a twoaction sequence produced in two turns: a sequence-initiating action (in one turn by one speaker) followed by a sequence-responding receipt (in a next turn by another speaker).

Notes * This chapter is a somewhat revised version of Chapter 4 of my doctoral dissertation (Lerner 1987). For example, I have updated the references. For a description of the features of turn construction that furnish occasions for co-participant completion, see Lerner (1991, 1996a). For a description of some actions that can be accomplished through this practice, see Lerner (1996b). . The “directionality of address” is too complex a matter to develop here, but suffice it to say that when a pre-emptive completion is produced by an addressed recipient, then it will almost always be addressed to prior speaker, while completions produced by someone other than the addressed recipient will almost always be addressed to that addressed recipient. In other words, who the completion is addressed to is not independent of who the prior turn was addressed to and thus the footing of the participant who produces it as addressed recipient or non-addressed recipient is consequential. . If the original speaker does confirm or reject this type of completion – one that coconstructs an action for the original recipient of the TCU – they can be seen to be asserting special (unilateral) authority over the action – for example, over the elements of an explanation or the events of a story. . ‘Contiguity’ refers to the placement of next speaker’s utterance, while progressivity is the one-after-another placement of words within the turn. Progressivity then is maintained by the contiguous placement of the pre-emptive utterance. . Another device commonly used is the turn-initial disjunction marker, “Oh” (Jefferson 1978; Heritage 1984). Jefferson provides the following instance: [GTS:II] Ken: Roger:

The cops, over the hill. There’s a place up in Mulholland where they’vewhere they’re building those hous//ing projects? Oh have you ever taken them Mulhollan’ time trials? ...

In contrast with these marked turn beginnings, pre-emptive utterances are constructed without any display of misplacement. . On occasion the pre-emptive completion itself can be extended to include an agreement token. The addition of such a token can do distinct interactional work. This work can best be understood once the regular shape of the collaborative turn sequence (pre-emptive com-

Collaborative turn sequences

pletion followed by a receipt by prior speaker) is described, and so this form will not be examined until after the sequence itself has been presented. . For the record I have found only one instance of a TCU-continuing utterance by a coparticipant that attempts to extend the preliminary component of a compound TCU-inprogress, rather than bring the unit to a next possible completion. (It is not irrelevant in this case that the parenthetical expansion of the preliminary component – which does not bring the ongoing TCU to its next possible completion – is addressed to the recipient of the ongoing turn, and thus does not make its acceptability relevant as a next action.) [HIC] Mom: Dad:

See this is little peanuts now, but if it ever got to be something [people would be fi:]:ghting an ah: that’s [an an it could ]

Here Mom is admonishing one of her adult sons (during a family meeting) not to take lightly the preparation of a written “basic agreement” the family is preparing as the guiding document for a newly established family investment club. Dad chimes in to corroborate mom’s premise (“if it ever got to be something”). Here Dad uses the projected opportunity space to second the premise implemented in the preliminary component, rather than to produce a completion and thereby align with its consequences. . Sometimes it may be more suitable to use the terms “accept/reject” at other times “agree/disagree” or “confirm/reject.” In this chapter, I am primarily concerned with those turns at talk that make relevant some form of understanding and/or alignment by coparticipants. Pre-emptive completion of these actions constitutes one way to strongly demonstrate such understanding/alignment by voicing a part of the TCU which realizes that action. It is ordinarily then up to the original speaker to validate the proposed completion as an adequate completion for their turn. However, some types of action (e.g. a request or other-correction) are not primarily built for alignment. Here the use of pre-emptive completion can implement a complementary action (e.g. an offer or self-correction), and thus the receipt of that action may also be of a different character (cf. Lerner 1996b). . There may be some special activity contexts in which the absence of a pre-emptive completion by a recipient at the end of a preliminary component may become noticeably absent. For example, this may be the case in certain instructional or tutorial exchanges in which, for example, an instructor/tutor is leading a student toward the understanding of some concept. Instructors/tutors may elicit completion as a tutorial practice, and similarly students may elicit completion from their instructor/turor (cf. Fox 1993; Lerner 1995). . For another form of rejection receipt, [Disagreement token + substitute completion], see instance (3). . See Schegloff (1996) for a description of another environment in which repeating is used as a method for agreeing – in this case, to confirm an allusion. There is a further parallel here. In the case of confirming an allusion, the original speaker has made clear (but not in an explicit fashion) some state of affairs which is then made explicit by a recipient. It is this that can be confirmed by a repeat. In the present case of pre-emptive completion, the original speaker has made clear some not-yet-said component of their turn (and its attendant state of affairs) which is then voiced (i.e. made explicit) by a recipient. Here again, a repeat (of the



 Gene H. Lerner

proffered completion) can confirm what had not yet been said explicitly. However, as we will see in the next section, if the original speaker begins their own “delayed” completion before the pre-emptive completion has come to a possible completion itself or if they produce a delayed completion that disagrees with the pre-emptive completion or both occur, then their completion can count as a replacement for the pre-emptive completion. . Projected and unprojected opportunity spaces provide for the sequential possibility of producing a syntactically fitted TCU completion. For completeness, I should mention that there is, in a sense, an additional opportunity space. The transition space also constitutes an opportunity space for the production of a syntactically fitted TCU completion, so that a single syntactic unit is produced across speakers as in pre-emptive completion. [GTS] Ken: Roger:

Well there was a certain amount of uh as far as I could see anyway there was a certain amount of uh dis- discontent between uh you and Louise to begin with

Rather than completing a TCU-in-progress, Roger’s utterance (“to begin with”) is appended to an already possibly complete unit. By “re-completing” the prior TCU, he can demonstrate agreement, but one which attaches a distinct modification. The prior utterance is not preempted in its course; rather the second speaker begins in the transition space and adds an additional increment to the TCU. He therefore does not violate the prior speaker’s right to speak at least until a first possible completion is reached, but shows that the action carried by that TCU had not been adequately formulated (cf. Sacks 1992, Vol. 1: 657; Lerner 2004). . This type of sequence relationship has more recently been dubbed a “retro-sequence” (Schegloff 1995b) in which a sequence is initiated as a response to a prior speaker’s action, but where that prior action did not make such a response conditionally relevant. . This is one way that Goffman’s (1981) concept of “speaker footing” can be seen to operate in interaction. . For a recent discussion of the negotiation of the relative “epistemic authority” parties can claim toward agreement in assessment sequences, see Heritage and Raymond (2003). For a discussion of the negotiation of the “authorship of delicates” see Lerner (1999). . A revised version of this section appeared as Lerner (1989). . One straight forward reason pre-emptive completions are rarely rejected is that they need not be produced. For the most part, preliminary components of compound TCUs do not require – that is, do not make conditionally relevant – the production of a completion by another participant. Though they do furnish distinct opportunities for completion, that action is ordinarily optional. This can be seen in each of the following instances. In (16.1), no one takes the opportunity to pre-empt the completion and the TCU is finished by the original speaker. In (16.2), another speaker does enter the turn space at the completion of the preliminary component, but produces a continuer rather than pre-empt the completion, while the original speaker finishes the TCU. In (16.3). the original speaker produces a preliminary component and stops speaking – leaving an explicit offer projected, but unsaid. The completion is not produced by the recipient, who instead responds to the offer. And in (16.4) there is a break in the progressivity of the turn (thus creating an unprojected opportunity space), but no recipient speaks and the original speaker completes their own

Collaborative turn sequences 

turn. Finally, in (16.5) a recipient does speak after a break in progressivity, but responds to the question, rather than produce the completion. (16.1) [HIC] Sparky:

If dad and Sherrie got together, (.) they would have a quorum

(16.2) [Mother’s Day (standard orthography)] Dad: So, if you were strong in your feelings about (0.2) people (0.2) Daughter: Mm hm Dad: your thet you liked ((continues to completion)) (16.3) [F:TC (in Davidson 1984: 125)] A: So if you guys want a place tuh sta:y. (0.3) . t . hhh Oh well thank you but you we ha- yihknow Thomas. B: (16.4) [Agorio] E: Didju request a copy of the: (0.3) instructions then? R: Yeah. (16.5) [Smith:Thanksgiving] R: Where’s thah: ◦ It’s on the counter ◦ ((pointing)) T: R: Oh, okay. Further, a completion need not be produced even when one seems to be encouraged by a speaker as in B’s second utterance in (16.6). (16.6) [GL:DS] B: A: B: A:

and uh but then she says she gets to thinking, oh well she’s just not gonna worry about it. Mm hm. you know, she’s just gonna – yeah.

Uncompleted TCUs, whether ended at a projected or unprojected opportunity space, can be treated as a trail-off, whose final part was left unspoken, yet can be seen to have been understood when a recipient responds as in (16.3), (16.5) and (16.6). (Token responses resist repair. That is, responding to an action with an agreement/disagreement token, is not vulnerable to rejection in the same way as a pre-emptive completion, since it cannot be inspected for content correctness, but only action-type correctness.) In these ways, potentially rejectable completions need not be done, and they are not noticeably absent. . Delayed completion can be seen as one device for achieving what Sacks (1992) referred to as “skip-connecting.” . Instances (13) to (17) do not contain pre-emptive completions. They are included to illustrate the phenomenon of delayed completion only. The use of this devise is not limited to pre-emptive completion.

 Gene H. Lerner . Formally, the production of a pre-emptive completion switches the occupants of the speakership/recipientship positions and converts a turn into a sequence, selecting the initial speaker of the turn unit as next speaker. A second switch in positions can be achieved by countering the pre-emptive completion in the receipt slot position. This occurs when the original speaker of the TCU deletes the sequential implicativeness of the pre-emptive completer by producing a (delayed) completion of their own rather than producing an acceptance or rejection of the pre-emptive completer. . A revised version of this section appeared as Lerner (1994). . Consistent with the use of list construction as an alternative to straightforward acceptance and rejection of the candidate, list construction can also be used in the receipt slot as part of a pre-emptive completion rejection. In (41), the use of a list displays that what is rejected is not simply the particular quoted offer completion. A [rejection token + substitute] could well be seen to be quibbling. The pre-emptive completion “wanna glassa milk” does catch the point being made – his grandmother treats his father like a child – though it is built as a punch line. In this case, Louise proffers a child’s drink (“milk”) in contrast to the adult’s drink (“beer”) that has been already been mentions. However, the contrast turns out to have been – or one might say is made out to have been – an adult offer (a drink) versus a child’s offer (a snack). Ken’s contrast is accomplished through the construction of a list. The replacement of the pre-emptive completion with another offerable item would simply be another instance of what you can offer a child. But the construction of a list displays the class of items that is intended and thereby, warrants the rejection of “milk” since it is not in that class. Nevertheless, notice that Ken’s list become more juvenile with each item. In a way, Ken’s third item does seem to capture the same flavor, if I can put it that way, of Louise’s pre-emptive completion, thereby moving in the direction of agreement even while explicitly rejecting her completion at the outset of his receipt (cf. Sacks 1987).

References Davidson, Judy (1984) “Subsequent Versions of Invitations, Offers, Requests, and Proposals Dealing with Potential or Actual Rejection”. In J. M. Atkinson & J. Heritage (Eds.), Structures of Social Action (pp. 102–128). Cambridge, Cambridge University Press. Drew, Paul (1987). “Po-Faced Receipts of Teases”. Linguistics, 25, 219–253. Fox, Barbara (1993). The Human Tutorial Dialogue Project: Issues in the design of instructional systems. Lawrence Erlbaum: Hillsdale, New Jersey. Glenn, Phil (1989). “Initiating shared laughter in multi-party conversations”. Western Journal of Speech Communication, 53 (2), 127–149. Goffman, Erving (1981). Forms of Talk. Philadelphia: University of Pennsylvania Press. Heritage, John (1984). “A Change-of-State Token and Aspects of Its Sequential Placement”. In J. M. Atkinson & J. Heritage (Eds.), Structures of Social Action (pp. 299–345). Cambridge: Cambridge University Press. Heritage, John, & Raymond, Geoff (2003). The Terms of Agreement: Indexing Epistemic Authority and Subordination in Talk-in-Interaction. Paper presented at the International Institute on Conversation Analysis, Copenhagen.

Collaborative turn sequences 

Jefferson, Gail (1973). “A Case of Precision Timing in Ordinary Conversation: Overlapped Tag-Positioned Address Terms in Closing Sequences”. Semiotica, 9, 47–96. Jefferson, Gail (1978). “Sequential Aspects of Storytelling in Conversation”. In J. Schenkein (Ed.), Studies in the Organization of Conversational Interaction (pp. 219–248). New York, Academic Press. Jefferson, Gail (1983). “Notes on some orderlinesses of overlap onset”. In V. D’Urso & P. Leonardi (Eds.), Discourse Analysis and Natural Rhetoric (pp. 11–38). Padua, Cleup Editore. Jefferson, Gail (1990). “List Construction as a Task and Interactional Resource”. In G. Psathas (Ed.), Interaction Competence (pp. 63–92). Washington, International Institute for Ethnomethodology and Conversation Analysis and University Press of America. Jefferson, Gail (This volume). “Some Orderly Aspects of Overlap in Natural Conversation”. In G. H. Lerner (Ed.), Conversation Analysis: Studies from the first generation. Amsterdam: John Benjamins. Lerner, Gene H. (1987). Collaborative Turn Sequences: Sentence Construction and Social Action. Unpublished Ph.D. Dissertation, University of California at Irvine. Lerner, Gene H. (1989). “Notes on overlap management in conversation: The case of delayed completion”. Western Journal of Speech Communication, 53 (Spring), 167–177. Lerner, Gene H. (1991). “On the Syntax of Sentences in Progress”. Language In Society, 20, 441–458. Lerner, Gene H. (1992). “Assisted storytelling: Deploying shared knowledge as a practical matter”. Qualitative Sociology, 15 (13), 247–271. Lerner, Gene H. (1993). “Collectivities in action: Establishing the relevance of conjoined participation in conversation”. Text, 13 (2), 213–245. Lerner, Gene H. (1994). “Responsive list construction: A conversational resource for accomplishing multifaceted social action”. Journal of Language and Social Psychology, 13 (1), 20–33. Lerner, Gene H. (1995). “Turn design and the organization of participation in instructional activities”. Discourse Processes, 19 (1), 111–131. Lerner, Gene H. (1996a). “On the “semi-permeable” character of grammatical units in conversation: Conditional entry into the turn space of another speaker”. In E. Ochs, E. A. Schegloff, & S. Thompson (Eds.), Interaction and Grammar (pp. 238–276). Cambridge: Cambridge University Press. Lerner, Gene H. (1996b). “Finding ‘face’ in the preference structures of talk-in-interaction”. Social Psychology Quarterly, 59 (4), 303–321. Lerner, Gene H. (1999). The Collaborative Construction of Delicate Formulations: Negotiating the Authorship/Ownership of Unkind Words. Paper presented at the convention of the National Communication Association, Chicago. Lerner, Gene H. (2004). “On the place of linguistic resources in the organization of talkin-interaction: Grammar as action in prompting a speaker to elaborate”. Research on Language and Social Interaction, 37 (2). Lerner, Gene H., & Takagi, Tomoyo (1999). “On the place of linguistic resources in the organization of talk-in-interaction: A co-investigation of English and Japanese grammatical practices”. Journal of Pragmatics, 31 (1), 49–75.

 Gene H. Lerner

Sacks, Harvey (1979). “Hotrodder: A Revolutionary Category”. In G. Psathas (Ed.), Everyday Language: Studies in Ethnomethodology (pp. 7–14). New York: Irvington. Sacks, Harvey (1987). “On the Preferences for Agreement and Contiguity in Sequences in Conversation”. In G. Button & J. R. E. Lee (Eds.), Talk and Social Organization (pp. 54– 69). Clevedon, England: Multilingual Matter. Sacks, Harvey (1992). Lectures on Conversation (2 Vols.). Oxford: Basil Blackwell. Sacks, Harvey, Schegloff, Emanuel A., & Jefferson, Gail (1974). “A Simplest Systematics for the Organization of Turn-Taking for Conversation”. Language, 50, 696–735. Schegloff, Emanuel A. (1982). “Discourse as an interactional achievement: some uses of ‘uh huh’ and other things that come between sentences”. In D. Tannen (Ed.), Analyzing Discourse (Georgetown University Roundtable on Languages and Linguistics 1981) (pp. 71–93). Washington, D.C.: Georgetown University Press. Schegloff, Emanuel A. (1987). “Recycled Turn Beginnings: A Precise Repair Mechanism in Conversation’s Turn-taking Organization”. In G. Button & J. R. E. Lee (Eds.), Talk and Social Organization (pp. 70–85). Clevedon, England: Multilingual Matters. Schegloff, Emanuel A. (1995a). “Parties and talking together: two ways in which numbers are significant for talk-in-interaction”. In P. Ten Have & G. Psathas (Eds.), Situated Order: Studies in Social Organization and Embodied Activities (pp. 31–42). Washington, DC: University Press of America. Schegloff, Emanuel A. (1995b). Sequence organization. Unpublished Ms, UCLA Department of Sociology. Schegloff, Emanuel A. (1996). “Confirming allusions: toward an empirical account of action”. American Journal of Sociology, 104 (1), 161–216. Schegloff Emanuel A., Jefferson, Gail, & Sacks, Harvey (1977). “The Preference for SelfCorrection in the Organization of Repair in Conversation”. Language, 53, 361–382. Schegloff, Emanuel A., & Sacks, Harvey (1973). “Opening Up Closings”. Semiotica, 8, 289– 327.

The amplitude shift mechanism in conversational closing sequences Jo Ann Goldberg

In an earlier investigation I established the existence of the Amplitude Shift Mechanism for utterance affiliation in natural conversation (Goldberg 1978). Amplitude shift can be used to indicate the inter-turn relationship between utterances by a speaker, either affiliating one utterance to a prior in a sequence by a downward shift in amplitude, or disaffiliating with a prior to indicate the initiation of a new sequence through an upward shift. I now proceed to (1) further explore aspects of its operation and (2) initiate another line of investigation as to the generality of its scope of operation, viz., to examine another type of conversational sequence in which the Amplitude Shift Mechanism is operative. The question/answer sequence was found to be a format of considerable generality in terms of the conversational tasks it embodied, “summoning,” “requesting,” “inviting,” etc. The operational generality of the Amplitude Shift Mechanism for a range of conversational tasks was thereby implicit. Here I focus on one pervasive conversational task: ending a conversation by means of a “closing sequence,” a termination apparatus for the structural unit “a single conversation.”1 After describing the closing sequence, I will develop two observations.

1. The Amplitude Shift Mechanism as described within the question/answer sequence (Goldberg 1978) displays similar operational regularities within the closing sequences inspected. I will provide confirmation of the original findings in this sequence type. More importantly, the materials suggest another conversational sequence over which the Amplitude Shift Mechanism is a relevantly operative device.

 Jo Ann Goldberg

2. The closing data will allow us to provide a finer order specification of the kind of work the Amplitude Shift Mechanism does as an agent of the sequence whose utterances it coheres. I have previously indicated one such aspect of its operation as a boundary marker at sequence-initial position. It was there that the Mechanism differentiated a sequence-initial utterance from a prior extrasequence utterance by an upward amplitude shift of the utterance initiating the sequence. I shall present materials below as they contribute to further insight into the ways the Mechanism evidences itself to be an agent of sequence construction. Two points will be developed in this regard: 2.1 The Amplitude Shift Mechanism displays a ‘discriminative capacity’ in its response to different order turn-constructional components over the course of the sequence. The mechanism is directed foremost to the affiliation to one another of principal first and second pair-parts of the adjacency pair subcomponents of the sequence under construction and, secondarily, to any auxiliary parts (local extensions) its sub-components may evidence. Specifically, the joining of a speaker’s own successive utterances within a singly operative adjacency pair unit and the joining of a speaker’s own successive utterances within an adjacency pair unit which operates with other such units as sub-components of a larger sequential construction is differentially accomplished. In the case of the latter (affiliation of adjacency pair units operative as sub-components of a larger sequential construction), it is in the joining of first and second pair-part sub-components that affiliation by amplitude descent is evidenced. 2.2 However, over the course of those sub-components so joined, a systematic variance of this pattern is evidenced. Specifically, in about half of the closings inspected a particular environment – an ‘opportunity position’ – was found in which the descent operation was suspended. For these cases, after the production of a raised amplitude sequence-initial exchange, a next first pair-part (and at times its return) evidenced a further raise in peak amplitude. Orderly aspects of such occurrences will be explored. I shall first discuss the general features of the closing sequence in conversation. I will then discuss the operation of the Amplitude Shift mechanism within this sequence.

Amplitude shift in closing sequences 

The closing sequence By “closing sequence” I refer to one major system-provided-mechanism with which parties may co-ordinatedly terminate a conversation, a closing sequence not being an invariantly present element for all forms of talk in interaction. In light of this focus on closings, our data has been drawn from conversational events which evidence this structure. The corpus on which the analysis is based has been drawn exclusively from telephone interactions. For such cases the closing is more or less coterminous within the closing of the interactive episode in which it occurs. It is a closing to “a conversation” and not to some segment of other types of conversational activity. Again, a closing sequence is not an invariantly present component of all conversational activity. It does not hold for members of a household in their living room, employees who share an office, passengers together in an automobile, etc., that is, persons who could be said to be in a “continuing state of incipient talk.” In such circumstances there can be... silence after a speaker’s utterance which is neither an attributable silence nor a termination... These are adjournments, and seem to be done in a manner different from closings. Persons in such a continuing state of incipient talk need not begin new segments of conversation with exchanges of greetings, and need not close segments with closing sections and terminal exchanges. (Schegloff & Sacks 1973: 324–325)

First, I shall indicate some of the independently analyzable features of the closing sequence and then turn to an examination of amplitude shift as a mechanism for utterance affiliation.

Close initiators I begin by noting some standard aspects of speakers’ close-initiating utterances with regard to their productional properties and positioning within a conversation. With regard to productional properties, I indicate a few commonly used close initiator types. One such common productional type employs materials developed within the conversation itself. Instances include, ‘return to/formulation of the reason for call (see fragments 7, arrows 1 and 2; 9, arrows l and 2; 14, arrows l and 2) and ‘reference to prior arrangements’ (see fragments 4, arrows l, 2 and 3; 12, arrows l and 2; 16, arrows 1 and 2). However, close initiators can be context-independent so that they do not necessarily reflect aspects of the conversation of which they’re a feature. One such type is found in the production of a single lexical unit such as “Okay” or “Okay” + [name of co-participant], the “Okay” being produced with a characteristic

 Jo Ann Goldberg

downward intonation. (See fragments 2, arrow l; 10, arrow l; l7, arrow l). However, clearly, a great many of the types discussed here have as their principal feature that they display elements of the conversation they’re closing.2 By proposing the existence of typical close-initiator type utterances, I do not intend to suggest that the production of such type utterances anywhere within the conversation will be in itself sufficient to initiate closing. Empirically, close initiators exhibit the analyzable property of selective positioning, i.e., positioning with reference to another order of conversational organization. In this case, selective positioning refers to placement of close initiators just after the completion of an analyzably terminated “topic.” Although not all topics have analyzable ends, co-conversationalists may specifically attend to the production of a topic boundary by employment of bounding mechanisms. For instance, a party can offer a formulation of the ‘moral’ or ‘lesson’ to be drawn from the topic and thereby begin to close the topic, as in the following fragment. Other party agreement will typically complete the close of the topic (cf. Schegloff & Sacks 1973). M: But uh I do feel fr’m indications that things are startin a ( ) sti:r a little bit. P: Well that’s what I heard the other day thet they were gettin pick up a lille // bit,* M: Myeah P: But uh certain things will do that y’know, they’re (r-) they’re bou:nd to. M: Yeah. P: In certain industry. M: //Yeah P: :hT.hh3 P: Different things ’ill pick up when it begins t’be spring a the yea:r an eve’thing. M: Yeah. P: :hT .hh But I think it’ll iron itself out. ← M: I sure hope // so. P: I’ll see ya Tuesday.

Returns to close initiators There exists a vast range of conversation-based mechanisms with which a speaker may display the relatedness of his just produced utterance to a coconversationalist’s just prior utterance. A speaker’s selection of one particular

Amplitude shift in closing sequences 

type of mechanism on some occasion will turn on the type of relatedness it is then relevant to indicate. Some of these mechanisms operate at the level of grammar, as when a party producing a return to a prior speaker’s utterance will incorporate in his utterance an element, such as a phrase, used by the prior speaker in his utterance. An aspect of adjacency pair organization operates in the service of the display of inter-utterance connectedness at the level of turn series. In the case of a party producing a return to an adjacency pair sequence initiator (such as a close initiator), one prominent means by which connectedness is indicated is by ‘selective positioning’ of this utterance in relation to the close initiator. I noted that close initiators display the property of selective positioning with reference to topical organization of the conversation. In the case of closing-returns, their proper positioning is accomplished with reference to adjacency pair organization: an adjacency pair second pair-part or return to a first pair-part (such as a question and close initiator) is demonstrated to be such by speakers primarily by positioning their utterance ‘next’ – that is, just after the completion of the first pair-part. Of course within that position the party producing the return may mark closing as an enterprise with which he may or may not co-operate. Whatever the party producing the return may elect to do, he is also implicitly ‘showing’ that he understands the conversationally organized task selected for him by the other party. It may be the case that for the party producing a return to close initiation, the Amplitude Shift Mechanism provides one resource for showing understanding and co-operation with the sequential task at hand. The matter might be stated in the following fashion: as we shall see, a party initiating closing employs the resource of the Amplitude Shift Mechanism whereby he may mark the disengagement of his present sequence-initial utterance from his prior. We shall also see that the party producing the return disengages his utterance from his prior. It may be the case that for the party producing the return, a corresponding upward amplitude shift marks both a disengagement of present from prior utterance, and a demonstration of his understanding of, and his co-operation with, the action proposed by the initiator. This demonstration is achieved through the production of a parallel directional amplitude shift with that just produced by the initiator. That is, it may be that one element in the demonstration of the acceptance of the proposal to close is the production of the upward amplitude shift of his return utterance over his prior.4

 Jo Ann Goldberg

The closing sequence and the question/answer sequence Unlike question/answer exchanges which can achieve completion within the span of a single adjacency pair unit, closing sequences require at least two sets of adjacency pair exchanges as sub-components: the exchange to initiate closing and the terminal exchange. The greater sequentially-described territory of turns that closings routinely cover allows for inspection of aspects of the Amplitude Shift Mechanism not obviously present in the question/answer sequence. The amplitude shift mechanism as described within the question/answer sequence displays similar operational regularities within the closing sequences inspected. As I have found to be the case in question/answer sequences (Goldberg, 1978), the sequence-initial utterance and the return utterance of the closing sequence will be raised in peak amplitude relative to the speaker’s immediately prior extra-sequence utterances. (1) JGP4

9.3 16. 31. 27.

M: F: M: F: M: F: M:

An so anyway t’sneak out an uh call Miss M :hT.hh y’know/ Mhm. :hT.hh So: Hm That’s all I know Oka:y buddy ((Close intoned)) Okay ol buddy

F: M:

Thank you Mi//ltie Yeah.*

C:

B’t I’ll put it o:n the bottom of the drawer, nex’ to the meddle ( // ) Yeah. Okay. (0.4) //( )That’ll be alright. (1.0) O//kay,

(2) NB

B: B: C: B: C:

Amplitude shift in closing sequences 

12.1 3.9 6. 17.1

B: C: C: B:

(3) JCP3 R: C: R:

7.7 7 22. 8.4

En I’ll pick it up Saturday. Okay. ((Close intoned)) We’ll see yuh (later) O::kay Bill

R: C: R: R: C: R: R: C:

I foun out whadi wanned f ’r // dinner Okay. .hhh 1 decided cause . . . An ken ya get some milk. Milk/ Mhm. Cuz there’s none left. Alright. .hhh Okay. That’s what l want f ’r dinner // hhh Alri

R: R: C: R:

Mkay Goo//bye. Bye. Bye.

← ←

Speakers’ intended close-terminal utterances are lowered in peak amplitude relative to their respective initial and return utterances.5 (4) JG 3:8

P: M: P: M: P: M: P: M: M: P:

. . Hello Marty/ (h) hi Pete How are you:= I’m great Good How are you Pretty good.= Hey,= We’re havin a meeting Tuesday night. Tuesday night/

←

 Jo Ann Goldberg

M:

M: P: P:

P:

–.9.

7.7 5.5

M: P: M: P: M: P: M: P: P: M: P: P: M: P:

Yea:h a//n . . . So can you make it Tuesday night/ ← Yeah. What time. . . . :hT.hh Different things ’ll pick up when it begins t’be spring a the yea:r an eve’thing. Yeah. :hT.hh But I think it’ll iron itself out. I sure hope // so. ← I’ll see ya Tuesday. Right. O//kay Marty. You-you’re al*right. //You can get there. YeYeah. Okay. Okay. //Thank you. See ya Pat. ← Bye

(5) NB C: B: B: C: B:

6. 4.2

C: C: B: C:

B’t I’ll put it o:n the bottom of the drawer, nex’ to the meddle ( // ) Yeah. Okay. (0.4) // ( )That’ll be alright. (1.0) O//kay, We’ll, see yuh (later) O::kay Bill ((Close intoned) Bye bye=

Amplitude shift in closing sequences 

(6) JGP4 M: F: M: F: M: 31. F: 27. M: F: M: M: M: F: M: M: F: M: –.- F: –.- M:

An so anyway t’sneak out an uh call Miss M :hT.hh y’know/ Mhm. :hT.hh So: Hm That’s all I know Oka:y buddy ((Close intoned)) ← Okay ol buddy Thank you Mi//ltie Yeah.* Yeah. Oka//:y I’ll see ya. Yeah.= Okay Fr//aiik Take it easy pa//rtnuh. Yeah. //Yeah. S’lon. ←

We observe here the same empirical regularities in the amplitude shifts in closing as occurred in the question/answer sequence. The conversational task of the Amplitude Shift Mechanism in affiliation/disaffiliation applies here and provides additional evidence for the existence of this Mechanism. It is relevant to note that close-initiating exchanges and their intended terminal exchanges are not necessarily adjacently positioned within the closing construction. (See fragments 4, arrow l and 2; 6, arrow l and 2.) In fact, speakers routinely produce a range of within-closing adjacency pair sub-components such that the closing sequence can be rather extensively occupied. However, it is an occupation by a range of conversational activities whose proper position is within the closing section. Such activities may include one or more of the following: “first reference or re-reference to arrangements” made within the conversation (see fragments 7, arrows 2 and 3; 8, arrows l and 2; 3), “reference to an unspecified next encounter” (see fragments l0, arrow 3; 2, arrow 2), “return to or display of original reason for the call” (see fragments 7, arrows l and 2; 9, arrows l and 2), a “thank you” or “final thank you” for some thankable occasioned within the conversation (see fragments 7, arrow 4, 10, arrow 2; 11, arrow l) and “signature” to the conversation (see fragment 23, arrow 1).

 Jo Ann Goldberg

For this project I am counting as closing components all those things listed above and nothing else. However, it is clear that other things can happen in closing sequences. One other major thing that can and routinely does, happen is that parties re-open topical talk which thereby suspends closing only later to be reinitiated. (7) JGP2 P: Hello: M: Hi ol’ buddy. P: Yea:h. M: Hey ya got anything goin t’night/ P: Na:h.=I don’ think I’m goin t’night Mel.= M: Oh rilly/ P: No. H//uh uh M: I’ve got a happen//ing out at Mercer if y//ou wanna come. . . . M: Well // (okeydo) P: M: P: P: P: P: P: M: P: P: M: M: P: P: M: P: P: M: P:

.hhh //I thought I’d // ask ya, y’know jist // t’ (get) Yeah Yeah Yeah Yeah Oka:y //0kay P//ete. .hhh I’ll see ya tamarra up the shop then. Yeah,= Okay // buddy. .hhh Okay Me//l Okay. Yeah.= Th//anks. Bye* bye. Yeah.

←1

←2

←3

←4

Amplitude shift in closing sequences 

(8) JG6:1 F: We:ll I’ll give y’a call maybe Thursday an see whattu well, F: Whadda you do on Saturday/ R: .hhh Nothine.= R: Go t’laundryhh//hh .hhh hhh F: Oh. (0.5) F: Uh huh. F: .hh Well I might call ya Tliursday evening h R: Okay hh F: Have ya been, wuz ya home last Tuesday evening/ (0.5) when I suppose t’call you // at fo:ur/ . . . F: We:ll whadda ya got some lessons t’get/ R: Yea:hhh F: We:ll okay. F: Be a good bo//::y. R: .hh Okay // I F: I’ll try F: I’ll call ya Thursday even//ing. R: Okay R: ga’bye: F: We’ll see whatju’re do//in R: .hh Alright R: //Bye bye F: Okay F: G’bye Ro//bert R: Bye (9) JCP3 C: Hello: R: :hT.hh Mother/ C: Ye:s.= R: I foun out whadi wanned f ’r // dinner C: Okay.

←1

←2

←3

←1

 Jo Ann Goldberg

R: .hhh I decided cause Pete an I jist had a blow of pits poop .hhh split pea soup an a piece a toast hh: C: Oh iz tha wha ju wan/= R: .hhh Ah huh :hhh . . . R: An ken ya get some milk C: Milk/ R: Mhm. R: Cuz there’s none left. C: Alright. R: .hhh Okay, R: That’s what I want f ’r dinner // hhh ←2 C: Alri R: Mkay R: Goo//bye C: Bye. R: Bye. (10) JGP4 M: That’s all I know. F: Oka:y buddy ((Close intoned)) M: Okay ol buddy F: Thank you Mi//ltie. M: Yeah.* M: Yeah. M: Oka//:y F: I’ll see ya M: Yeah.= M: Okay Fr//ank F: Take it easy pa//rtnuh M: Yeah F: //Yeah M: S’lon

←1 ←2

←3

(11) JG 6:3 D: A:ndum hh m’kay well the’n we’ll look forward f- I hope he has a good lesson on Friday. S: //Yeah.

Amplitude shift in closing sequences 

D: D: D: S: D: S:

Hha-hm-hm-hm .hhhh Okay honey. .hhh An thank you so much f ’r calling. :h//hh A’ri Bye bye Bye

←1

Given the above cited collection of close-constructional components, each evidenced as an adjacency pair sub-component, we may ask whether the Amplitude Shift Mechanism for affiliation of speakers’ own successive utterances will remain operative over this lengthened course.6 We might reformulate this question to embody our initial observations. They were: the peak amplitude level of speakers’ utterances in the exchange initiating closing (Cil )7 are ordinarily raised over their immediately prior extra-sequence utterance (x), (I diagram this as x–⇑ →Cil ); the peak amplitude level of speakers’ terminal utterance (Ct ) are ordinarily lowered relative to their Cil utterances, (in diagram: x–⇑ →Cil – ⇓ →Ct ). This is so for cases where Ct is positioned directly after Ci or in its near proximity. Does this amplitude patterning obtain in the major portion of closings investigated where Cil is not positioned just prior to Ct , but where there occur intervening close-related operations? I look to relevant materials below. In the case that Ct component is preceded by a closing component (C) which is not Cil , the Ct utterance will be lower in amplitude relative to that immediately prior utterance (C—⇓ →Ct ). (12) SBI:6:I A: B: B: A: A:

Well you wanna come out and have lunch with me/ No, Let me take you to lunch, some//time. OOOHHH! No. No. Come on. Come out uh uhh say you get out here ’bout uh twelve thirty. ←1 B: Twelve thirty. B: Don’t fix very much, though, I just I uh eat very light anymore, I’m tryina get slim . . . A: Well I’ll see you about twelve thirty B: Twelve thirty– Mon//day

←2

 Jo Ann Goldberg

23. 25. 18.5 20.

A: Monday B: Okay dear. A: A:righ//t B: Thank you A: Bye B: Bye

(13) JG6:3 D: .hh An so the I: will uh u::mh keep it here by the pho:ne.= An I think everything ’ill bequite alright, you know. S: Yeah. D: A:ndum hh D: M’kay well the:n we’ll look forward f- I hope he has a good lesson on Friday. S: Yeah. D: Hha-hm-hm-hm .hhhh D: Okay honey. 20. D: .hh An thank you so much f ’ r calling. :h//hh 7. S: A’ri 10. D: Bye bye 4.8 S: Bye (14) JGP6 W: P: W: W: P: P: P: W: P: W: P:

W: P:

Hello/ Is John there/ .hh No:.= I’m expecting him anytime.= //Is there a message/ Oh ah see. No. This is Pat from down at the drum corps down Santa Rosa Bay Bay Shrine Club. .hh//h Auhuh. Is he goin t’drum practice tahmorrow mornin do you know/← 1 No: : . He’s no: t. Oh I see. . . . .hh Well I know he’s no:t Oh I see.

Amplitude shift in closing sequences 

25.1 23.2

P: W: P:

Well okay jus longzi know. Oh(h)ok//ay Tha:nk // you

–.13.

W: W:

Thanku. Bye bye.

←2

In the case that a speaker produces some non-terminal closing utterance (C) that is preceded by some non-initial closing utterance (C), it is itself lower in peak amplitude relative to the immediately prior utterance, C–⇓ →C–⇓ →Ct . (16) JGPl D: But uh ’f I gave you a week’s notice ’r something // maybe you c’d make it on a Friday night huh/ C: .hh Well that way I c’d work ou:t y’know uh least a schedu:le en see how itd uh works out with the rest of the people thet I’m involved wi://th. D: ((Very quietly)) Yea:h. (1.0) D: Gee one a’ these Friday nights I’ll give yuh a call if you’d like toI’d give you // ( ) ←1 C: Grea:t! . . . D: But I” give yih ca:ll/ e:nu:::h ←2 C: Okay Don C: Well then gimme a call in the next coupla weeks en then un .hh y’know we’ll see what we c’n work out. D: Great. 22.9 C: Awrighty/= 11. D: Okay Cathy, –.- D: //I’ll see* ya then.= 20.3 C: Okay love D: Yeah C: Righ= C: Byebye. (17) JGP2 P: Oka:y ((Close intoned)) M: //Okay P//ete. P: .hhh

←1

 Jo Ann Goldberg

32. –.23.1 –.-

P: M: M: P: P: M: P: P: M: P:

I’ll see ya tamarra up the shop then. Yeah,= Okay // buddy. .hhh Okay Me//I Okay. Yeah.= Th//anks. Bye*bye. Yeah.

Thus we see that where the adjacency pair exchange sub-component initiating closing is not positioned just prior to the terminal exchange, but where there occur intervening close-related sub-components, the same amplitude patterns for affiliation obtains. In sum, the following regularities in speakers’ shifts in the peak amplitude level of their own successive utterances over the course of the closing sequence were observed: 1. The peak amplitude levels of the close-initial and the return-to-the-closeinitial utterances were raised over that of speakers’ respective immediately prior utterance. x–⇑ →Cil 2. The peak amplitude levels of the close terminal exchange utterances were lower relative to those of speakers’ close-initial and return utterances. Cil –⇓ →Ct 3. The peak amplitude levels of the close terminal exchange utterances were lower relative to that of their respective immediately prior non-initial and return close utterances. C–⇓ →Ct 4. The peak amplitude levels of non-initial close utterances were lower relative to that of their own immediately prior non-initial close utterances. C–⇓ →C In our inspection of regularities in speakers’ shifts in peak amplitude level of their successive adjacency pair sub-components in the production of the closing sequence, the tactic has been to start with the terminal exchange and to

Amplitude shift in closing sequences 

move back up through the components, to just prior to the sequence-initial exchange. The finding has been that, routinely, sub-component utterances, so serially arrayed, display an amplitude shift patterning such that the peak amplitude level of speaker’s successive sub-component utterances were lowered over the course of the sequence. It is by such means that larger sequential constructions composed of multiple adjacency pair sub-components are cohered. (See fragments 18, 19, and 20 below.) (18) JGPP3 R: C: R: R: C: R: 22. R: 8.4 C: 10.9 R: 11.4 R: —.- C: 10.9 R:

An ken ya get some milk. Milk/ Mhm Cuz there’s none left. Alright. .hhh Okay That’ss what I want f ’r dinner // hhh Alri Mkay Goo//bye. Bye. Bye.

← pass

(19) NB

15.1 23.1 14.5 17. 7.1 –.–.-

C: No hanky pan/ky. B: No:: hanky panky. (0.4) C: Well have a good time. B: Oh(hh) kay // (h) I’llC: ’N I’ll, see yuh later. B: Okay (Bill) C: Yah. B: //Bye. C: Bye.

(20) SBI:6:l A: Yea:h/ B: Mh//hm

← pass

 Jo Ann Goldberg

28.5 32.5 27. 22. 23. 25. 18.5 20.

A: B: A: B: A: B: A: B:

Well I’ll see you about twelve thirty Twelve thirty–Mon//day Monday Okay dear Alrigh//ty Thank you Bye Bye

← pass

Amplitude shift and the sub-components of the closing sequence The closing data allow us to determine a finer order specification of the kind of work the amplitude shift mechanism does as an agent of the sequence whose sub-components it coheres. The joining of adjacency pair sub-components into a larger sequential construction and the joining of a speaker’s successive utterances of a single adjacency pair exchange is differentially accomplished. It is differentially accomplished by means of the Amplitude Shift Mechanism’s “discriminative capacity” by which it differentiates amongst primary and auxiliary components of the adjacency pair sub-components over the course of sequence construction. The Amplitude Shift Mechanism is directed foremost to the affiliation to one another of principal first and second pair-parts of the adjacency pair sub-components of the sequence under construction and, secondarily, to any auxiliary parts (local extensions) its sub-components may evidence. To note that the peak amplitude level of a speaker’s own successive utterances decreases over the course of the sequences inspected is to report on one of the most massively recurrent empirical findings. I have inferred, in part from this finding, that the Amplitude Shift Mechanism’s affiliation/disaffiliation operation evidenced in a speaker’s successive utterances over the course of the sequence is conducted on a turn-by-turn basis, i.e., is continuously operative. However, this is not to suggest that turn order is the primary factor in producing the directionality of the corresponding amplitude shift. What appears to be foremost in determining the directionality of the shift is whether or not the successive utterance in question is a “primary” or “auxiliary” component of the sequence under construction. The Amplitude Shift Mechanism displays a discriminative capacity by which it differentiates amongst such primary and auxiliary components. I shall present evidence below in support of the position that the Amplitude Shift Mechanism is directed foremost to the affiliation to one another of speakers’ own successive first and second pair-parts of adjacency

Amplitude shift in closing sequences 

pair sub-components of the sequence under construction and, secondarily, to any local extension its sub-components may evidence. (A prominent instance of a sub-component extension type is the “pass.” The pass, of course, may be terminal to the sub-component but not the larger organization, the closing sequence.) Investigation of relevant materials will be initiated below. To begin, I offer a sequential characterization of the aspects of the adjacency pair sub-components of interest. In each of the fragments below one of the parties introduces a close-related activity either in initiating the closing sequence or as some next sub-component over its course. (See the ‘return to the reason for the call’ of fragment 21, the ‘arrangement’ reference of fragment 22,8 and the conversational ‘signature’ of fragment 23.) For each, after the recipient’s return utterance (the second pair-part) is produced, the initiator then produces a pass in third position. The matter of focal interest occurs next. After the pass, the initiator of the prior adjacency pair either introduces a new first or produces a return or second pair-part to his co-conversationalist’s first. (21) JGP3 22. R: C: 10.9 R: 11.4 R: C: R:

That’s what I want f ’r dinner // hhh Alri Mkay Goo//bye Bye Bye

(22) JG6:l F: R: F: F: R: F: 13.5 F: R: R: 10.4 F: R: R: –.- F: 5.6 F: R:

We:ll whadda ya got some lessons t’get/ Yea:hhh We:ll okay Be a good bo//:y .hh Okay // I I’ll try I’ll call ya Thursday even//ing ← return to arrangements Okay Ga’bye: We’ll see whatju’re do//in ← return to arr.((cont)) .hh Alriglit //bye bye Okay. ← pass G’bye Ro//bert ← terminal Bye ← terminal

← pass ← terminal

 Jo Ann Goldberg

(23) JG6:8 L: M: L: M: 6.1 L: M: 3. L: M: 4.5 L:

I got a:ll the stuff. All I need is (ta–) //see you. J’st pick a time. //Ha ha ha ha Okay clear heart .hhh Thank you for calling an tell Dawn I said hello. A wor’, a word outta the past, huh/ ← signature ←1 Ri:ght Alright dear. G’bye. G’bye. ← terminal

It is such orders of occurrence that suggests the existence of a discriminative capacity of the Amplitude Shift Mechanism. It is a means by which a speaker may specify, within a range of at least the two just prior turns, to which prior utterance of that turn his present turn’s utterance is affixed. In the case of affiliation to a prior’s prior, a speaker specifically disaffiliates his present from his just prior by raising the peak amplitude of the present over that of the just prior, but simultaneously affiliates his present to the prior’s prior by maintaining the amplitude level of the present lower than that of the prior’s prior. In the sequential environment so constituted there was displayed a strong tendency to directly affiliate successive adjacency pair sub-component first or second pair-parts, bypassing intervening extensions. A similar orientation was found to occur in like-constituted environments of successive question affiliates. (See fragments 24, 25 and 26 below.) (24) JG(R/M) 24.

9.9 21.2 –.17.9 6.

R:

Iz it inna back of the park or in front, (0.6) or on the sides. ← question one M: .hh It’s onna corner of .hhh the road that goes down by the park and Bradshaw. R: Oh. (0.4) R: What side right or left. ← question two M: It’s on Maple and Bradshaw. R: Oh. ← pass R: It’s onna corner/ ← question three M: Yeah. R: .hhh Shoo ← pass

Amplitude shift in closing sequences 

(25) JG(P/M) M: I dropped him off frum a meeting the other night. 20. P: :h.hh In Sana Moiiica,=huh.= ← question one M: Yeah 6. P: Yeah ← pass 14. P: Where’s ’e on twenny-first er somethin/ ← question two M: Ye.= Somet’ing like that.= 12.6 P: Yeah ← pass (26) JGP2 25.3 F: M: 17. F: M: 12.1 F: M: 2.7 F: 23. F: M: 17. F: M: 7.1 F: M: –.- F: M:

Whadiyou been doin. ← question one (n) Oh I- I- We win out Hollywood t’day. Oh didja/ ← question extension Yea//h Yeah= ← pass Yeah :hh Hm. ← pass Ja have a good ti:me/ ← question two Well we looked over you know they had an open house. Oh did they. ← question extension Th-there’s no u:h y’know mm races goin or anyth//een. Mhm. ← pass On Sundee. M//hm. ← pass .hh

For such instances, the Amplitude Shift Mechanism does not just affiliate some successive first or second pair-part of an adjacency pair sub-component to just any last utterance, but to the first or second pair-part of the prior adjacency pair sub-component of the larger sequential construction.9 There were found a few instances which ran counter to the above reported cases. On such occasions, a speaker’s successively positioned next first or second pair-part was lower in peak amplitude than his just prior sub-component’s extension. Such counter cases were not without their orderly aspects. Quite regularly, the lowered successive sub-component was the first or second pair-part of the terminal exchange. One such instance is presented here: (27) JG3:12 C: H’Ilo:

 Jo Ann Goldberg

28.9 27.2 15.8 1.9

R: H’lo.= R: Thisiz Sills.= R: When will you-D’you need a lift/ . . . C: I’ll come home with Leslie cuz I gotta stop at the store an get something for kitty. (0.8) R: Oh. R: Waidami R: ((Off phone)) .hhh Didju get anything fer Smo:key/ G: No. R: ((On phone)) No he didn’t (1.7) C: Well then I’ll haftuh stop an get im so//mething so, g’bye R: :hT.hh R: Okay.= R: So ya nea-so ya don’t- so ya needa lift.= ← close initiator C: ((Angrily)) I DON’T. D: You do/ ← repair initiator C: ((Increased anger)) I DON’T. R: Okay. ← pass C: G’bye. ← terminal R: G’bye.

It must be noted of such instances that we are not witnessing lapses in affiliation per se but affiliation of a first or second pair-part (as R’s “G’bye” in the above case) to a prior first or second pair-part. However, in that there were so few of the above type lapses found in our corpus, a fine order specification of the grounds by which the discriminative capacity is activated or left inactive awaits accumulation of more materials. In sum, for the major portion of cases so constituted, the affiliation of utterances of a single adjacency pair exchange and its extension is marked by successive lowering of peak amplitudes over its course. A similar operation obtains for the joining of multiple adjacency first and second pair-parts where they constitute sub-components of a larger sequential construction. The Amplitude Shift Mechanism’s discriminative capacity differentiated between these two operations.

Amplitude shift in closing sequences 

Systematic variations to continuous amplitude descent There is a systematic variance in the amplitude descent pattern. The position in which this variance is found is constant and represents a temporary deviation from descent. This variance can be understood with reference to the sorts of operations parties engage in there. I shall designate the position as an “opportunity position.” There was found, in about half of the closings inspected, an environment in which the pattern of amplitude descent over the series of first and second pair-parts of the closing was suspended. For those cases, after the raised amplitudes of the exchange to initiate closing, there was evidenced in the utterance of the initiator of the second exchange (and, on occasion, the return utterance) a further rise in peak amplitude. I designate the position in which the variance occurs as an “opportunity position” (O.P.). A range of sequentially specifiable moves can occur in this position. These are displayed in the data below. The variance can be understood with reference to this opportunity position and the order of phenomena that regularly occur there. I shall first establish the orderly character of this deviation in the pattern of amplitude descent. I shall then investigate grounds for its occurrence. (The significance of the notation C2 in the data will be explained below.) I turn now to consideration of orderly aspects of this second amplitude raise in closing. Over the course of the closing, when an adjacency pair first or second pairpart component is raised in peak amplitude over that of the immediately prior exchange, its position is just after the initiating exchange (and its extension(s)). (28) JG6:3 D: .hh An so then I: will u:h u::mh keep it here by the pho:ne an I think everything ’ill be quite alright, you know. S: Yeah. D: A:ndum hh m’kay 18.7 D: Well the:n we’ll look forward f- I hope he has a good lesson on Friday. ← close initiator S: //Yeah. ← close return D: Hha - hm - hm - hm .hhhh D: Okay honey. ← pass 20. D: .hh An thank you so much f ’r calling. ← O.P./C2 /thank you :h//hh S: A’ri D: Bye bye S: Bye

 Jo Ann Goldberg

(29) JGP2 M: P: P: P: P: 19. P: M: P: 31.6 P: M: M: P: P: M: P: P: M: P: (30) JGP4 M: 31. F: M: 35.2 F: M: M: M: F: M: M: F: M: F: M:

//I thought I’d // ask ya, y’know jist // t’(get) Yeah Yeah Yeah Yeah Oka:y ((Close intoned)) ← close initiator //Okay P//ete. ← close return .hhh I’ll see ya tamarra up the shop ←O.P./C2 /arrangement Yeah,= Okay // buddy. .hhh Okay Me//I Okay Yeah.= Th//anks. Bye* bye. Yeah. That’s all I know. Oka:y buddy. ((Close intoned)) Okay ol buddy. Thank you Mi//ltie. Yeah.* Yeah. Oka//:y I’ll see ya. Yeah.= Okay Fr//ank. Take it easy pa//rtnuh. Yeah. //Yeah. S’lon.

(31) JG6:2 C: Y’know I got that figured out too. R: Mhm

← close initiator ← close return ←O.P./C2 /thank you

Amplitude shift in closing sequences 

7.2

8.3

C: Okay baby lemme letju go honey // (cuz I gotta go over to the post office cuz I don’t have anybody in the store right now ’n it// won’t matter if I close. ← close initiator R: Mkay ← close return R: Mhm R: Mkay C: Okay honey ← pass R: Mkay C: Okay well // I’ll see ya probably ohh around sixish. ←C2 /arrangement R: G’bye R: Mkay R: G’bye C: Bye dear R: Bye

However, after the occurrence of this second raise in peak amplitude, successively produced closing sub-components typically evidence a return to the pattern of descent over the remaining course of the sequence. This second raise as a suspension of the pattern of amplitude descent is temporary. (32) JG6:3 18.7

20.

10.

D: Well the:n we’ll look forward f- I hope he has a good lesson on Friday. ← close initiator S: //Yeah. ← close return D: Hha-hm-hm-hm .hhh D: Okay honey. ← pass D: .hh An thank you so much f ’r calling. ← C2 /thank you :h//hh S: A’ri D: Bye bye ← terminal S: Bye

 Jo Ann Goldberg

(33) JGP2 19.

31.6

25.1 14.8 16.8 13.

P: M: P: P: M: M: P: P: M: P: P: M: P:

Oka:y ((Close intoned)) //Okay P//ete. .hhh I’ll see ya tamarra up the shop Yeah,= Okay // buddy. .hhh Okay Me//It Okay. Yeah.= Th//anks. Bye* bye Yeah.

← close initiator ← close return

Oka:y buddy ((Close intoned)) Okay ol buddy Thank you Mi//ltie. Yeah.* Yeah Oka//:y I’ll see ya Yeah.= Okay Fr//ank Take it easy pa//rtnuh Yeah. //Yeah. S’lon

← close initiator

←O.P./C2 /arrangement

← pass ← thank you ← terminal

(34) JGP4 31. 35.2

27.5

–.18.

F: M: F: M: M: M: F: M: M: F: M: F: M:

←O.P./C2 /thank you

← arrangement reference

← well wish ← terminal

Amplitude shift in closing sequences 

(35) JG6:2 7.2 C: Okay baby lemme letju go honey // (cuz I gotta go over to the post office cuz I don’t have anybody in the store right now ’n it // won’t matter if I close. ← close initiator R: Mkay R: Mhm R: Mkay C: Okay honey ← pass R: Mkay 8.3 C: Okay well // I’ll see ya probably ohh around sixish. ←O.P./C2 /arrangement R: G’bye R: Mkay R: G’bye 3.8 C: Bye dear ← terminal R: Bye

In the data presented above, two types of closing components were presented, “Thank you”s and arrangements. Each evidenced a raised amplitude deviation in the position just after the close-initiating exchange. But, when such components were not positioned directly after the initiating exchange, they were found to be produced in accord with the pattern of amplitude descent in the close environments in which they appeared. (See fragments 36, 37, 38, and 39 below). Thus, not only is this suspension in the pattern of amplitude descent temporary, but it is as well position specific. (36) JGP2 19. P: M: P: 31.6 P: M: M: P: 25.1 P: M: 14.7 P: 16.8 P: M: P:

Oka:y ((Close intoned)) //Okay P//ete. .hhh I’ll see ya tamarra up the shop then. Yeah,= Okay //buddy. .hhh Okay Me//It Okay Yeah.= Th//anks. Bye* bye Yeah.

← pass ← thank you

 Jo Ann Goldberg

(37) JG6:1 F: R: F: 14.3 F: R: F: 13. F: R: R: 10.4 F: R: R: –.- F: 5.6 F: R:

We:ll whadda ya got some lessons t’get/ Yea:hhh Well okay. Be a good bo//:y .hh Okay // I I’ll try I’ll call ya ←reference to arrangement Okay ga’bye: We’ll see whatju’re do//in .hh Alright //Bye bye Okay ← pass G’bye Ro//bert Bye

(38) JGP4

35.2

27.5

–.18.

F: M: F: M: M: M: F: M: M: F: M: F: M:

(39) SB2:2:1 B: A: A: A: A: B: A:

Oka:y buddy. Okay ol buddy. Thank you Mi//ltie. Yeah.* Yeah. Oka//:y. I’ll see ya. Yeah.= Okay Fr//ank Take it easy pa//rtnuh. Yeah. //Yeah. S’lon.

←reference to arrangements

Well I // won’t keep you Bea, but ( ) Well Yeah. Okay Well I’m sorry thetchu–hh Well so’m I b’t // ( ) Maybe it’s just as well you discovered it.

Amplitude shift in closing sequences 

29.

28.6 20.6 18.5

B: A: B: A: B: A: A: B: A: B: A: B:

Well it tis, cause I’d hate to be sliding backwards down that hill Yeah ehhh heh heh uh rather Ehheh heh heh Okay Well you give // ( ) See yuhAnd I’ll see yuh Sunday A::Iright, // Bea Alrighty. Thanks, // a million ← thank you Bye Bye

These data illustrate the placement of two proper closing components in the position just after the close-initiating exchange: “Thank you”s (see fragments 32 and 34) and references to arrangement (see fragments 33 and 35). Although a frequent productional feature of such proper closing components in this specific sequential locus (just after the close-initiating exchange) is their production in raised amplitude over the exchange they succeed (elsewhere a disaffiliation operation), they are properly considered closing components. The production of components such as these (e.g. “Thank you”s), whether they occur just after the close-initiating exchange or later in the closing sequence are built to briefly reference in the closing and not revive the matters they attend. In both of these sites, they do not suspend the closing sequence by inviting their development beyond the position in which they are produced. When, as is the case here, they are essentially closing components, I designate them as C2 (non-initial closing component exhibiting a second rise in peak amplitude). This position just after the close-initiating exchange is not exclusively occupied by proper closing moves. Two types of extra-close sequence moves will be investigated. First, I shall consider “rearrangements,” i.e., those operations on “references to arrangements” which act as modifications or amendments. Parenthetically, it is not the case that arrangement and rearrangement operations are limited in placement to the closing sequence. Here, I consider only those that appear in closings. Although rearrangements in closings are typically positioned with reference to operations within the same sequence they sometimes trigger a return to topical talk and thereby exhibit what shall be referred to as “extra-sequence potential.” Second, I shall consider “re-initiations of topical talk” which suspend closing operations and are thereby extra-close sequence moves. I shall not in these cases designate them as proper closing sequence components (C2 ). In our consideration of rearrangements and re-initiations

 Jo Ann Goldberg

of topical talk after the initiation of closing, focus will be upon the impact of their extra sequence potential upon the course of closing. I first consider rearrangements.

Rearrangements When a rearrangement appears in the closing sequence it typically exhibits the property of selective positioning, viz., a rearrangement is positioned not just anywhere over the course of closing, but just after the closing component it seeks to amend, the reference to arrangements. Such orderliness in positioning of subsequent instances of a same action is not an exclusive feature of arrangement/rearrangement operations, but is elsewhere evidenced over a range of other conversational activities which include second stories and return invitations to name but two. These occurrences reflect conversationalist’s “preference for fittedness” in the placement of such activities. The fitting of a conversational activity in specifiable proximity to a prior related activity provides the structural auspices for the production of the latter by reference to the former. The positioning of a rearrangement component just after the arrangement it seeks to amend has consequences for the placement of the former in the closing by virtue of the fact that a great many closings are initiated by reference to arrangement components. The organization by reference to which arrangements are such commonly employed close-initiated components shall not be considered. What is of interest here is that such occurrences result in a closeinitiating exchange, reference to arrangements, which just after its production is met with the initiation of an exchange (rearrangements) with potential to disrupt the course of the sequence under construction. That is, despite the fittedness to the prior close-initiating exchange and despite the dispatch with which the rearrangements may be executed, a rearrangement stands in two of its aspects in a relationship of potential rupture to the closing. For one, a rearrangement may be used as the occasion for its producer or its recipient to re-engage in topical talk. Here the talk attending the rearrangement may be used as the auspices to bring up a different but related matter. On such occasions, closing is temporarily abandoned. (See fragments 40 and 41.) (40) SB2:6:8 A: Oh for goodness sake B: (Look) ninety five or ninety six B: Well anyway, uhm (2.0)

Amplitude shift in closing sequences 

11.5

16.

A: Well then I’ll see you tomorrow ← close initiation/ evening. reference to arrangement B: I’ll s-I’ll see you tomorrow, B: And you said six o’clock/ A: No seven B: Seven A: It’s just dessert you see. B: Yeah. B: Okay, then. A: Seven o’clock. B: //Oh ( ) A: Uh however. However, uh you can ← rearrangement come in earlier if you want to B: Oh no, heh heh heh I’ll be in a little earlier of // course. A: Yes. A: Uh huh B: And uh an that’s theB: An I’m disappointed because I had← re-initiation I’d been looking forward tuh doing of topical talk // something for them, // an’A: Yes. A: Uh huh, A: Well now do you want me to cancel the uh the order, or d’youuh the reservation, or will you B: Oh that’s right.

(41) JG6:3 D: .hh Okay darling well that’ll be jist fine. D: An of course as I say I don’t expect any extra time but jist t’like for the finish of a thought hha S: Oh sure.

 Jo Ann Goldberg

19.

19.8

14. 13.5

23.

D: :hhhh Okay then we’ll see you← close initiator/ ala- he’ll see you then reference to arrangement around about nine-thirty on friday morning. S: Fi://ne. D: Aha:. D: .h//h Okay we’ll see you then, S: Bye bye D: .hh And listen shall I leave you a check for seven dollars/ ← rearrangement S: Yeah :hh D: Okay S: That’s alright D: Ahuh. D: Okay honey.= ((Close intoned)) ←second close initiation D: Thank // you so mu S: Nobody ’ill, nobody else ’ill ← rearrangement be home.= That’s good :hh // ← topic initiation :hh :hh D: We:ll now I don’t know uh//h– ← topic initiation ((Clears throat)) S: That’s alright. S: I mean you know this doesn’t bo//ther me. D: I have to be go:ne uh butuh .hh I think that perhaps see he would be here by himself. //-at least y’know for // .hh S: Yeah. S: Well I’ tell you something ease I like to either .hhh uhh always feel free t’call me uphhh=

Second, on occasions that the rearrangement component is produced in briefest form, its extra-close sequence potential may be documented by its consequences for the subsequent course of the closing. Commonly, after completion of brief rearrangements to close-initiating arrangements, parties in their return to the course of the closing produce yet another reference to arrangements. (See fragments 42 and 43.)

Amplitude shift in closing sequences 

(42) JG3:8 P: 9. M: P: 19. M:

7.7

I’ll see ya Tuesday. Right. O//kay Marcus You- you’re al*right. //You can get there. P: Ye P: Yeah. M: Okay. P: Okay.= P: //Thank you. M: See ya Pete. P: Bye.

← reference to arrangements

← rearrangement

← pass

←reference to arrangements

(43) JGP110 D: But I’ll give yih a ca:ll/ ← reference to arrangements 24. C: Okay Don 27.9 C: Well then gimme a call in the ← rearrangement next coupla weeks en then h .hh y’ know we’ Il see what we c’n work out. D: Great-. C: Awrighty/= 11. D: Okay Cathy, –.- D: // I’ll see* ya then.= ← reference to arrangements C: Okay love D: Yeah. C: Righ= C: Byebye.

It is suggested that the subsequent reference to arrangements may operate to relate the present closing move to the prior closing move, the prior reference to arrangements. That the relating is done may suggest that parties treat the intervening rearrangement talk as disruptive to the advancement of closing. However, such subsequent references to arrangements do not share, with reinitiated closings after reinitiated topical talk, the feature of being produced in raised amplitude. That is, they are not marked as closing sequence initiators in their relation to prior talk. The rupture effect of such brief rearrangements may be thereby seen as treated by parties as limited in effect on the course of the closing sequence. (See fragments 42 and 43.)

 Jo Ann Goldberg

Re-initiations of topical talk I turn now to re-initiations of topical talk. Such occurrences are extra-close sequence moves. The topic reinitiating utterance is produced by its speaker in raised amplitude over his just prior utterance of the close-initiating exchange. Topical re-initiations suspend the advancement of the closing sequence. (44) JGR/M M: R: 11.8 R: 26. R: M: R: M: R:

12. 25.1

R: M: R: M: M: R:

Well thanks f ’r calling okay/= ← close initiation Yeah. Okay. W’you gonna go t’morrow/ ← topic Oh yeah. T Okay. Okay/= The bus doesn’t come I’m not going. . . . Dr. Buckley come pick us all up. Oh hhlhh wouldn’t do it. Oh huh // I (ra-) rrral see ya. ← second close initiation Okay. Bye.= Bye.

(45) NB(A/C) C: Oh I love tuhC: Gee I ride mine all // a’ time A: Yeah. C: I love it. (0.2) A: .hh Well, honey/ I’ll pob’ly see yuh one a these da : ys, 12. C: O::h God yeah, A: //Uhh huh! C: We 17. C: B’t I c- I jis // couldn’ get down //there A: OhA: Oh I know.

← close initiation

← topic

Amplitude shift in closing sequences 

A: I’m-not asking... . . . A: But I’m fine, (0.4) C: //Yeah. 20.6 25.5

A: C: A: C: A: C:

I think I’ll make it. Okay. .hhhh Alright, I’ll see yuh next week then Bah bye Bye bye,

(46) NB (GH) G: A’rigIit, (2.4) G: I’ll see if I c’n (1.0) -rout sump’n up. G: I’ll call yuh back inna few minnits. 10. E: Alright, 15. E: San Clemente is always uh I know darn well y’c’n get on there. G: Ken we/ E: //Yeah. G: Oh-* Oii Saturdee/ E: Yeah. . . . 10. 13.

E: So I don’t know how ihd be t’day. B’t- y’give’m a try. E: Why donche try en nen, gimme call. G: Alright.

← second close initiation

← close initiation ← topic

← second close initiation

Re-initiation of topical talk (in a sense returning parties to the “body” of the conversation) is not an uncommon occurrence and speaks to a feature of closing: the closing structure is open to its possible suspension through the

 Jo Ann Goldberg

introduction of new topical materials. That such occurrences can and do regularly happen does not speak to a deficiency of the closing apparatus. One sense of deficiency may be that the apparatus is easily counteracted and the conversation never evidences termination by the closing structure. However, empirically, this is not the case. Rather, that a conversation does not achieve closure upon first initiation means only that parties have disengaged the apparatus on the occasion of the re-initiation of topical talk and that they can and do reinitiate closing at some subsequent point. It is not uncommon for a conversation to evidence multiple moves to close, the last being the one that achieves completion.11 Re-initiations of topical talk occur in a range of positions over the course of the closing sequence even (however rare) after the terminal exchange. Although they evidence such a freedom in positioning, the position just after the close-initiating exchange remains as the one most recurrently used by a party to indicate to a co-participant that there is still conversational business to be attended. That this position is one so prominently used may be a consequence of sequentially derived pressures based on the fact that the sequence just initiated is built to bring about conversational closure. I elaborate on this matter below. In the body of the conversation, that a party in the course of some ongoing talk finds that he has some new topic to bring up, does not provide a warrant for introducing the matter then and there. In the body of the conversation there is a range of evidence that supports the suggestion that a speaker may not only await completion of any current matter but await placement of his new matter in a conversationally designated locus as would give evidence of the relatedness of the present matter to just prior matters. I spoke of this as “preference for fittedness” in the discussion of rearrangements. But, that a speaker may await later placement can only be done by virtue of his reliance on there being further talk. Here, the issue for him is where in the course of such further talk this new matter may best be brought forth. However, in the case of initiation of closing, a speaker is precluded from such a reliance for closing is specifically built to lead to a condition of no more talk. This is most prominently the case in the corpus from which our data has been drawn, telephone interactions, where parties unlike those in face-to-face interactions do not remain, if only briefly, in visual and auditory contact after closing their conversation. As a consequence, once closing is initiated while there still remains something yet to be discussed, there may be pressure to initiate that matter as early into the closing sequence as possible. Of course, by definition, that earliest possible position is the one just after the initiation of the closing.

Amplitude shift in closing sequences 

It has been argued here that re-initiations of topical talk and rearrangements stand in an extra-sequence (potential) relationship to the close-initiating utterances they succeed. A party initiating such a sequentially recognizable alternative to the designated task at hand, closing, thereby properly produces his utterance in raised amplitude over his immediately prior indicating the new sequence, disaffiliated character of his present utterance relative to his own prior.

Conclusion I have considered four types of operations positioned just after the closeinitiating exchange. In that position, a speaker would commonly produce such utterances in raised peak amplitude relative to his own just prior utterance of the close-initiating exchange. In that position a range of operations were located: re-initiations of topical talk (see fragments 40 and 4l), rearrangements (see fragments 40, 4l, 42 and 43), “Thank you”s (see fragments 28 and 30), and references to arrangements (see fragments 29 and 3l). The last two were distinguished from the first two by being proper closing components. It was suggested that re-initiations of topical talk and rearrangements have a feature in common as operations that occur within the closing sequence: each stands in a relationship of “rupture” to the advancement of the closing sequence by virtue of their (potential) extra sequence consequences. It was proposed that the arrival of each, just after the close-initiating exchange, could be understood by reference to different sequential issues. In the case of the re-initiation of topical talk I argued that given that the closing sequence is building to a state of no more talk and given the presence of something yet to be discussed, there may be a pressure for parties to announce the existence of that discussable as early in the course of the closing as possible, that earliest possible position being just after the close-initiating exchange. In the case of rearrangements in the closing, other matters were taken into consideration. Reference to arrangements operations can serve as the occasion for their rearrangement, both inside and outside of closing sequences. I argued that, since reference to arrangements is a prominently used type of closing initiator, the position just after the close-initiating exchange seems ripe for occupation by a rearrangement move. This position just after the close-initiating exchange has been designated as an opportunity position. By this label I refer to a place in the closing, the position just after the close-initiating exchange, where a party under various structurally designated auspices can initiate an extra-close sequence move.

 Jo Ann Goldberg

When I spoke of C2 I referred to a type of action, a second raise in peak amplitude, that appears as a productional feature of proper closing components when positioned immediately after the close-initiating exchange, otherwise a feature of such disaffiliating moves. Such proper closing components appear to advance the closing and not invite its suspension. The question, then, is why such proper closing components are not here produced in compliance with the pattern of amplitude descent over the course of closing elsewhere and otherwise exhibited. The hypothesis that I forward is that they may deviate with reference to the types of extra-sequence moves that otherwise might be similarly positioned. That is, given the “ripeness” of this position for sequence-suspension-implicative activities, it may be the case that occupation of this position by proper closing moves in raised amplitude is a means by which a party may mark as recognized the alternative sequence moves for the position he occupies, but one he will not elect to take. He thereby doubly asserts his commitment to the advancement of the sequence under construction.12 Investigation of what I have here designated as the opportunity position and the raised amplitude components featured there have implications for further research on the Amplitude Shift Mechanism, a matter to which I shall briefly turn. The Amplitude Shift Mechanism as an agent of sequence construction has as its most prominent feature the joining of a speaker’s own utterances in the sequence under construction by first raising the amplitude at sequence initiation and then by joining within-sequence first and second pair-parts by successive amplitude descent over its course. Investigation of what I have here designated as the opportunity position specified a within-sequence environment in which more than one sequence type can find systematic and positionspecific expression in a single sequential locus. For closing, after the initiating exchange, alternative sequential operations can be and regularly are initiated. Such occurrences are properly investigated without reference to the amplitude shift pattern they may evidence. However, such occurrences do have ramifications for this pattern. That is, the Mechanism’s patterned operations may exhibit systematic variances according to the different sequences whose construction it is in the service of and the complications that such sequences may undergo. Such a possibility awaits further investigation.

Amplitude shift in closing sequences 

Notes . Discussion of the phenomenon of a closing structure for conversation will be in the service of goals for the investigation of the Amplitude Shift Mechanism and thereby limited. The reader is urged to consult Schegloff and Sacks (1973) for a systematic treatment of closing as an aspect of the structural unit ’a single conversation.’ Our discussion of the closing apparatus is indebted to this work. . Strictly speaking, it is not the case that there is always only one occurrence of closing per conversation. It is not uncommon for conversationalists to initiate and then suspend closing in a return to topical talk. For such occasions, the conversation will evidence more than one initiation of a closing sequence, only one of which, at least, will reach completion. However, in pursuit of the task at hand, our corpus has been drawn primarily from conversations exhibiting a single close initiation that reached completion. . Goldberg uses the symbol “:hT” to indicate an audible nasal in-breath. When it appears with an audible oral in-breath (“:hT.hh”) it indicates a nasal in-breath that is transformed into an oral in-breath. (Ed.) . The directionality of amplitude shifts of both speakers, the party initiating close and the party producing the return-to-close, will be indicated for the data fragments. This procedure is of course in order, since I am investigating an amplitude affiliation mechanism for conversationalists regardless of whether their utterances are in initial or in return position within the sequences investigated. . I am using the expression ‘intended close terminal utterance’ to stand for two otherwise distinguishable elements. In the first case, I refer to a turn-constructional component that announces its lastness both for the closing sequence and the conversation per se. The instance type is “bye bye” or some variant of it. In the second case, I refer to a turnconstructional component such as “thank you” (P’s last line in fragment 4) or “See ya + ((name))” (M’s last line in fragment 4). Such components are de facto terminal but, unlike the component “bye bye,” have no systematic claim to the position they occupy. A distinction between these two types of terminals may be highly relevant for a conversationalist in coming to understand utterances positioned after each type. For example, where a “thanks” positioned after a “see ya” may be hearable as just some next proper closing move, a “thanks” positioned after a same speaker’s “bye bye” may have a systematic basis for being heard as an otherwise forgotten but just remembered expression of appreciation. However, with regard to investigation of amplitude shift patterning and the work of affiliation, the distinction between these two types of terminal utterances is not relevant. For the task of affiliation, the relevant issue is whether or not that component (which happens to be last) is a component of the closing under construction. In this spirit I shall put aside this elsewhere-relevant distinction and refer to both as close terminals. . With regard to our characterization of the closing sequence as being composed of adjacency pair sub-components I reiterate that each sub-component in its turn may achieve extension. Initially, I can mention two common extension types: passes and repairs. Where relevant to the investigation of the data fragments I shall so mark those extensions. For the moment, the reader is asked to disregard them in determining the amplitude shift regularities of the larger sequence. The orderliness they evidence will be attended shortly.

 Jo Ann Goldberg . By the notation Cil I indicate l) an utterance construction for closing (C), 2) that is sequence initial (Ci ), and 3) as I have indicated, routinely raised in peak amplitude (Cil ) over the producer’s immediately prior utterance. The notation C2 will indicate the occurrence of a second raise in peak amplitude on a subsequent proper closing component. The positioning of that C2 raise will be the focus of investigation below. . It is suggested that the arrangement “I’ll call ya Thursday evening” is preceded by a close component which is not itself the close-initiating utterance. Rather, the close-initiating utterance is the one just prior to it, “Well whaddaya got some lessons to get?” This latter is not the most self-evident instance of a close initiator. The auspices under which it might be so considered needs elaboration. It will be argued that “...ya got some lessons to get?” stands as a possible variant of a type of closing initiator hitherto not mentioned. What is that type? One type of strategy conversationalists employ in initiating closing is based upon on a distinction between caller and called. The closing type of interest here is one reserved for callers. That is, a caller may use as grounds for initiating closing that it is done on behalf of called. “I’ll let you go” is a prototype. It is possibly the case that “ya have some lessons to get?” is a variant of this strategy. So formulated, it might be said to constitute a “search” for grounds whereby the conversation could be warrantably discontinued. It may be characterized as a variant strategy insofar as it seeks, rather than announces, a warrant for termination of the conversation. . One of the grounds for investigating the closing structure was that it covered a greater sequentially described territory. By constitution, closing is regularly composed of at least two adjacency pair subcomponents: the exchange initiating close and the terminal exchange. Even when closing was found to occupy a greater quantity of turns, it was an occupation by more such first and second pair-part sub-components and not by extensions of subcomponents. As to why such extensions were not produced is not within the domain of this investigation. I only note that parties, having extension as a possible next move, more often elected instead to initiate some next adjacency pair sub-component. However, it was on the occasions of such extensions that affiliation of subsequent firsts and seconds to prior firsts and seconds by means of this discriminative capacity was evidenced. . In this fragment, D’s “I’ll give you a call” stands as an arrangement component to an indefinite next encounter. After C’s return to this indefinite next, she proposes a rearrangement to it in order to make the timing of the next call more definite. . The capacity of this conversational sequence to be repositioned is not a feature of all conversational sequences. For example, ‘greetings’ when not exchanged initially will not be repositioned at a later point in the conversation. Similarly, the repair apparatus is built to be initiated and achieve completion just after the occasion of the production of the repair object. After that repair position, under the condition that the repairable itself is not later reintroduced or in some way consequential for subsequent talk, I find no such structurally provided locus for the return of the repair apparatus. . I must reiterate that this portion of the research is based on investigation of forty-seven closing sequences. Half of those closings evidenced occupation of the position just after the close-initiating exchange with an utterance in raised amplitude over the raised amplitude initiator. That half was closely divided amongst the types of operations mentioned here: “Thank you”s, arrangements, re-initiations of topical talk, and rearrangements. It is

Amplitude shift in closing sequences 

obviously the case that a finer order investigation of the hypothesis I forward here awaits collection of more materials. The hypothesis I forward to account for this systematic variance appears to be a fruitful line for investigation.

References Goldberg, Jo Ann (1978). “Amplitude shift: a mechanism for the affiliation of utterances in conversational interaction”. In J. N. Schenkein (Ed.), Studies in the Organization of Conversational Interaction (pp. 199–218). New York: Academic Press. Schegloff, Emanuel A., & Sacks, Harvey (1973). Opening Up Closings. Semiotica, 8, 289–327.

Index

A account-elicitation devices 165 acquainted-unacquainted 83 acquainted-stranger relationship 65 already informed 186 ambiguity 131 answerer 95 announcements content-based definition of announcements 174 expanded announcement sequences 176, 195 assessments 173 authoritative knowledge 233 B being ordinary 137 by-product conversations see designed conversations vs. by-product conversations

designed conversations vs. by-product conversations 79 double-checking 114 E ethnographic setting 189 F first topic 192 G gap and overlap 40 gap vs. pause 40 overlap onset 44, 58, 241 overlap resolution 45, 59, 243 generalized list completer 244 generalizing list completer 245 given/new distinction 171, 193 see also marked and unmarked information focus

C clearance cue answers 67, 70 closing sequence 259 auxiliary components 274 opportunity position 279 rearrangements 286 re-initiation of topic talk 290 comic book orthography 19 commitment to the normal 154 see also odd events courtesy terms 64

H hello

D damaging evidence

J joke/serious 131

114

74, 88 see also possible greeting Hynek, J. Allen 165 I inaccurate records 110 information focus 209 interruption 42 investigative work 109

 Index

L laughter 18 local identity 186 loyalty to reality 154 M marked forms 73 marked competition 50 marked and unmarked information focus 209 Menninger, Karl 154 minimal preference rule see preference mishearing 16 mono-topical calls 186 N neutrality 117, 120 new information see given/new distinction O odd events 146 see also commitment to the normal overall structural organization 79, 86, 174, 191 overlap see gap and overlap P pause see gap and overlap polar position of conversation 36 Pomerantz, Anita 165, 173 possible greeting 77 preference minimal preference rule 91 preference not to re-tell 177, 183, 192 pro-form 199, 201 progressivity 226 break in progressivity 246

Q question/answer sequence

262

R recipient design 177 receipt slot 230, 239, 243 receipt slot alternative 235, 244 receipt slot pre-emption 233 receipt slot rejection 247 recognitional response 18, 44, 58 reference a matter vs. revive a matter 285 S Sacks, Harvey 22, 79, 82, 125, 131, 137, 146, 154, 165, 192 self-identification 89, 93 summons/answer sequences 63, 179, 186 T telephone answerer see answerer telephone as territorial 98 thank you 285 terms of address 64 turn taking turn allocational techniques turn claimant 49, 59 turn occupant 49, 59 turn-taking repair 232 “who speaks first” 96 turn-taking system 229 type of conversation 69 U uncertainty 233 unmarked competition

50

V varying pronunciation

20

37

In the Pragmatics & Beyond New Series the following titles have been published thus far or are scheduled for publication: 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102

TROSBORG, Anna (ed.): Analysing Professional Genres. 2000. xvi, 256 pp. PILKINGTON, Adrian: Poetic Effects. A relevance theory perspective. 2000. xiv, 214 pp. MATSUI, Tomoko: Bridging and Relevance. 2000. xii, 251 pp. VANDERVEKEN, Daniel and Susumu KUBO (eds.): Essays in Speech Act Theory. 2002. vi, 328 pp. SELL, Roger D.: Literature as Communication. The foundations of mediating criticism. 2000. xiv, 348 pp. ANDERSEN, Gisle and Thorstein FRETHEIM (eds.): Pragmatic Markers and Propositional Attitude. 2000. viii, 273 pp. UNGERER, Friedrich (ed.): English Media Texts – Past and Present. Language and textual structure. 2000. xiv, 286 pp. DI LUZIO, Aldo, Susanne GÜNTHNER and Franca ORLETTI (eds.): Culture in Communication. Analyses of intercultural situations. 2001. xvi, 341 pp. KHALIL, Esam N.: Grounding in English and Arabic News Discourse. 2000. x, 274 pp. MÁRQUEZ REITER, Rosina: Linguistic Politeness in Britain and Uruguay. A contrastive study of requests and apologies. 2000. xviii, 225 pp. ANDERSEN, Gisle: Pragmatic Markers and Sociolinguistic Variation. A relevance-theoretic approach to the language of adolescents. 2001. ix, 352 pp. COLLINS, Daniel E.: Reanimated Voices. Speech reporting in a historical-pragmatic perspective. 2001. xx, 384 pp. IFANTIDOU, Elly: Evidentials and Relevance. 2001. xii, 225 pp. MUSHIN, Ilana: Evidentiality and Epistemological Stance. Narrative Retelling. 2001. xviii, 244 pp. BAYRAKTAROĞLU, Arın and Maria SIFIANOU (eds.): Linguistic Politeness Across Boundaries. The case of Greek and Turkish. 2001. xiv, 439 pp. ITAKURA, Hiroko: Conversational Dominance and Gender. A study of Japanese speakers in first and second language contexts. 2001. xviii, 231 pp. KENESEI, István and Robert M. HARNISH (eds.): Perspectives on Semantics, Pragmatics, and Discourse. A Festschrift for Ferenc Kiefer. 2001. xxii, 352 pp. GROSS, Joan: Speaking in Other Voices. An ethnography of Walloon puppet theaters. 2001. xxviii, 341 pp. GARDNER, Rod: When Listeners Talk. Response tokens and listener stance. 2001. xxii, 281 pp. BARON, Bettina and Helga KOTTHOFF (eds.): Gender in Interaction. Perspectives on femininity and masculinity in ethnography and discourse. 2002. xxiv, 357 pp. McILVENNY, Paul (ed.): Talking Gender and Sexuality. 2002. x, 332 pp. FITZMAURICE, Susan M.: The Familiar Letter in Early Modern English. A pragmatic approach. 2002. viii, 263 pp. HAVERKATE, Henk: The Syntax, Semantics and Pragmatics of Spanish Mood. 2002. vi, 241 pp. MAYNARD, Senko K.: Linguistic Emotivity. Centrality of place, the topic-comment dynamic, and an ideology of pathos in Japanese discourse. 2002. xiv, 481 pp. DUSZAK, Anna (ed.): Us and Others. Social identities across languages, discourses and cultures. 2002. viii, 522 pp. JASZCZOLT, Katarzyna M. and Ken TURNER (eds.): Meaning Through Language Contrast. Volume 1. 2003. xii, 388 pp. JASZCZOLT, Katarzyna M. and Ken TURNER (eds.): Meaning Through Language Contrast. Volume 2. 2003. viii, 496 pp. LUKE, Kang Kwong and Theodossia-Soula PAVLIDOU (eds.): Telephone Calls. Unity and diversity in conversational structure across languages and cultures. 2002. x, 295 pp. LEAFGREN, John: Degrees of Explicitness. Information structure and the packaging of Bulgarian subjects and objects. 2002. xii, 252 pp.

103 FETZER, Anita and Christiane MEIERKORD (eds.): Rethinking Sequentiality. Linguistics meets conversational interaction. 2002. vi, 300 pp. 104 BEECHING, Kate: Gender, Politeness and Pragmatic Particles in French. 2002. x, 251 pp. 105 BLACKWELL, Sarah E.: Implicatures in Discourse. The case of Spanish NP anaphora. 2003. xvi, 303 pp. 106 BUSSE, Ulrich: Linguistic Variation in the Shakespeare Corpus. Morpho-syntactic variability of second person pronouns. 2002. xiv, 344 pp. 107 TAAVITSAINEN, Irma and Andreas H. JUCKER (eds.): Diachronic Perspectives on Address Term Systems. 2003. viii, 446 pp. 108 BARRON, Anne: Acquisition in Interlanguage Pragmatics. Learning how to do things with words in a study abroad context. 2003. xviii, 403 pp. 109 MAYES, Patricia: Language, Social Structure, and Culture. A genre analysis of cooking classes in Japan and America. 2003. xiv, 228 pp. 110 ANDROUTSOPOULOS, Jannis K. and Alexandra GEORGAKOPOULOU (eds.): Discourse Constructions of Youth Identities. 2003. viii, 343 pp. 111 ENSINK, Titus and Christoph SAUER (eds.): Framing and Perspectivising in Discourse. 2003. viii, 227 pp. 112 LENZ, Friedrich (ed.): Deictic Conceptualisation of Space, Time and Person. 2003. xiv, 279 pp. 113 PANTHER, Klaus-Uwe and Linda L. THORNBURG (eds.): Metonymy and Pragmatic Inferencing. 2003. xii, 285 pp. 114 KÜHNLEIN, Peter, Hannes RIESER and Henk ZEEVAT (eds.): Perspectives on Dialogue in the New Millennium. 2003. xii, 400 pp. 115 KÄRKKÄINEN, Elise: Epistemic Stance in English Conversation. A description of its interactional functions, with a focus on I think. 2003. xii, 213 pp. 116 GRANT, Colin B. (ed.): Rethinking Communicative Interaction. New interdisciplinary horizons. 2003. viii, 330 pp. 117 WU, Ruey-Jiuan Regina: Stance in Talk. A conversation analysis of Mandarin final particles. 2004. xiv, 263 pp. + index. 118 CHENG, Winnie: Intercultural Conversation. 2003. xii, 279 pp. 119 HILTUNEN, Risto and Janne SKAFFARI (eds.): Discourse Perspectives on English. Medieval to modern. 2003. viii, 243 pp. 120 AIJMER, Karin and Anna-Brita STENSTRÖM (eds.): Discourse Patterns in Spoken and Written Corpora. 2004. viii, 279 pp. 121 FETZER, Anita: Recontextualizing Context. Grammaticality meets appropriateness. 2004. x, 272 pp. 122 GONZÁLEZ, Montserrat: Pragmatic Markers in Oral Narrative. The case of English and Catalan. 2004. xvi, 409 pp. 123 MÁRQUEZ REITER, Rosina and María Elena PLACENCIA (eds.): Current Trends in the Pragmatics of Spanish. 2004. xvi, 383 pp. 124 VINE, Bernadette: Getting Things Done at Work. The discourse of power in workplace interaction. 2004. x, 278 pp. 125 LERNER, Gene H. (ed.): Conversation Analysis. Studies from the first generation. 2004. x, 300 pp. 126 WU, Yi’an: Spatial Demonstratives in English and Chinese. Text and cognition. xviii, 226 pp. + index. Expected Fall 2004 127 BRISARD, Frank, Michael MEEUWIS and Bart VANDENABEELE (eds.): Seduction, Community, Speech. A Festschrift for Herman Parret. vi, 196 pp. + index. Expected Fall 2004 128 CORDELLA, Marisa: The Dynamic Consultation. A discourse analytical study of doctor-patient communication. xii, 247 pp. + index. Expected Fall 2004 129 TABOADA, María Teresa: Building Coherence and Cohesion. Task-oriented dialogue in English and Spanish. xvii, 244 pp. + index. Expected Fall 2004 130 HALMARI, Helena and Tuija VIRTANEN (eds.): Persuasion Across Genres. A linguistic approach. Expected Fall 2004

A complete list of titles in this series can be found on the publishers website, www.benjamins.com