10 1 1 453 9317 PDF

Conversation Analysis
Pragmatics & Beyond New Series

Editor
Andreas H. Jucker
University of Zurich, English Department
Plattenstrasse 47, CH-8032 Zurich, Switzerland
e-mail: ahjucker@es.unizh.ch
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 Catherine Kerbrat-Orecchioni
Hebrew University of Jerusalem University of Lyon 2
Jean Caron Claudia de Lemos
Universit de Poitiers University of Campinas, Brazil
Robyn Carston Marina Sbis
University College London University of Trieste
Bruce Fraser Emanuel Scheglo
Boston University University of California at Los Angeles
Thorstein Fretheim Deborah Schirin
University of Trondheim Georgetown University
John Heritage Paul O. Takahara
University of California at Los Angeles Kansai Gaidai University
Susan Herring Sandra Thompson
University of Texas at Arlington University of California at Santa Barbara
Masako K. Hiraga Teun A. Van Dijk
St.Pauls (Rikkyo) University Pompeu Fabra, Barcelona
David Holdcroft Richard J. Watts
University of Leeds University of Berne
Sachiko Ide
Japan Womens University
Volume 125
Conversation Analysis: Studies from the rst generation
by Gene H. Lerner
Conversation Analysis
Studies from the rst generation
Gene H. Lerner
University of California, Santa Barbara
John Benjamins Publishing Company

Amsterdam/Philadelphia
TM The paper used in this publication meets the minimum requirements
8
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 rst 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

302.346-dc22 2004050204
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)
2004 John Benjamins B.V.

No part of this book may be reproduced in any form, by print, photoprint, microlm, 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 dont 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 19351975
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 orga-
nization of action that underpins social life from the obvious, mundane details
of conversation and other human conduct. Yet, beyond Sacks seminal contri-
butions 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 ma-
terials ways of working that other researchers could adopt and adapt. It is this
clarity of purpose and method embodied in Sacks investigations of conver-
sation 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, Har-
vey 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 Sche-
gloff 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
Jeffersons entry onto the scene was unique. Having enrolled in Sacks
course at UCLA in 1965 to fulfill one remaining requirement for her B.A. de-
gree 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 sus-
tained and concentrated association with the details of the recorded data by
coming to terms with those details and revealing some of the fine grained or-
derliness 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 in-
teraction. 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 Jeffersons, by then, proven approach to transcrip-
tion, he advised me to try to figure out a way to transcribe the data and then
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 ac-
tion sequencing. I first describe Jeffersons 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 Ana-
lytic 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) Jeffer-
son 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 fea-
tures 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 to-
kens (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 understand-
ing of three forms of organization generic to conversation: turn taking, action
formation and action sequencing. I now take up each of these forms of or-
ganization 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. Un-
derstanding 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 po-
sition 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 Har-
vey 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 establish-
ing 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 men-
tion 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 prob-
lematic participation, when more than one party is speaking at once. Her
contribution on overlapping speech in conversation derives from and com-
plements Sacks work on turn taking.8 It is only once we understand the orga-
nization of turn taking, that simultaneous speech or overlap can be properly
understood. Jefferson lays out the systematic forms of overlap competition
and introduces the important conceptual distinction between marked and

unmarked overlap competition. Almost half of the paper is devoted to a sys-
tematic 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 chap-
ter deepens our understanding of sequence organization, since it describes, in
detail, members interpretive work sequential interpretive work that un-
derpins 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 cru-
cial 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 war-
rants their inclusion by showing how identities become relevant and circum-
stances 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 investi-
gation 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. More-
over, 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-for-

granted 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.
Jeffersons chapter in this section, At First I Thought explores a conversa-
tional device that demonstrates how reported thoughts can be subject to social
organization in this instance sustaining the social world as an altogether or-
dinary place by showing ones 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 conver-
sational data, Jefferson shows how people routinely select first thoughts to
report, that are appropriate to local circumstances and/or category member-
ship. 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 anal-
ysis 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 mo-
ment. 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.
For almost 30 years, Terasakis 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 infor-
mation 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 treat-
able 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 ori-
entation to determining whether possible news is news for a deliverers current
recipient. She provides a detailed account of the linguistic composition and se-
quential organization of pre-announcement sequences and their connection to
the composition of subsequent news delivery turns.
Lerners contribution demonstrates another connection between syntax
and social action in this case between turn construction and action sequenc-
ing. 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 participants 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 comple-
tion. 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 environ-
ment. 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 de-
scribing 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 Masters degree in psychology, when I began
to look for a doctoral program that would allow me to extend my thesis re-
search on parent-child interaction. A fellow graduate student, who had gone
on to UC Irvines 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 mod-
eling. 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 deco-
rated 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 par-
ticular 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 im-
portant 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
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 en-
tirely 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 ar-
gument, 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 todays
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 Jeffersons 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 com-
ment was made, Sacks makes the unqualified statement, The structures for particulars
direction doesnt 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 Jeffersons work on overlap, as direct evidence for partys
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 Goffmans 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 Garfinkels 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. 199218). New York: Academic Press.
Jefferson, Gail (1973). A case of precision timing in ordinary conversation: overlapped tag-
positioned address terms in closing sequences. Semiotica, 9, 4796.
Joseph, Brian D. (2003) The Editors Department: Reviewing Our Contents. Language, 79
(3), 461463.
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. 5469).
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), 696735.
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. 757). New York: Academic Press.
Schegloff, Emanuel A. (1968). Sequencing in conversational openings. American

Anthropologist, 70, 10751079.
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 Id probably rather transcribe than any do any other part of the work
(analyzing, theorizing, lecturing, teaching, etc.), the one thing Id rather not do
is talk about transcribing. Its not a topic. You might as well talk about typewrit-
ing. 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 raise-
T. 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 you-
A. I got lost in one job, yeah.
T. Got lost in it, and your father-
C. He know your father? Yeah.
[ A. Sure he knows my father, but my fathers got nothing to do with it.
[ C. Ok he gave you the job.
A. No, hes got nothing to do with it. Huh uh, My fathers 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: ene sid that if I work fer a month: yih getta buck,h h thi[rty fi:ve=
(Dan): [((sniff))
Ken: =n hour en () evry month he uh ( ) he rai[ses you ] ( ) ]
Dan: [Howdju]g e t th]e jo:b,
(1.0)
Ken: I js wen down theren 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 (yr 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 [ bt my f a t h e rs gt] nothina ] do with it.
Al: [( ) they gave you] th jo:b,]
(0.7)
Ken: No: hes got nothin ddo wth it. Huh-uh, my fa(h)athers 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 read-
able transcript from Stanley I. Kutlers Abuse of Power (1997:253),1
(3) [AbPow:253]
((re the seamy things Hunt says he did for the White House Plumbers))
DEAN: I dont know the full extent of it,
PRESIDENT NIXON: I dont know about anything else.
DEAN: I dont 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 dont kno:w the () full extent v it.
(0.7)

Dean: Uh:::eh
(0.9)
Nixon: I donnoo bout anything else exchh e[pt
Dean: [I dont either in I: wd (h)als(h)o
hhate tuh learn [some a] these thi]ngs. hhhhhhhh
Nixon: [W e l l ] y a : h ]
(0.2)
Dean: So Thats,hhhh thats that situation.
Why put all that stuff in? Well, as they say, because its there. Of course theres
a whole lot of stuff there, i.e., in the tapes, and it doesnt all show up in my
transcripts; so its because its there, plus I think its 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 Ill 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), Ill 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 father-
C. 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 (yr 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 As father, and the mention of As father triggers a question by
patient C:
T: ...and your father-
C. He know your father?
In fragment (2) the issue is drastically obscured, first of all by the sheer dif-
ficulty 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 Kens having been bothered, and it just so happens that at that mo-
ment Al decides to cut in with a tangential question about Kens 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 irre-
sponsible to leave it out even though it not only messes up the transcript but
its so improbable that Im embarrassed by it.
But on thinking about it, yes, its more than likely that the father alterna-
tive is whats happening with Al, but that doesnt 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. Its 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 dont see any bugs. It might have, but I cant see any
Its a phenomenon. And it very likely happens more often than we have access
to in our materials, because people dont always, or even routinely, mention
the fact that they did a mishearing (sometimes they dont realize it, sometimes
they do realize it but it doesnt 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
bothered, which Al hears as his saying something about Kens 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
Ill just mention one other difference between the two fragments. One of
them shows an and while the other shows an or (pronounced er).
A. I got lost in one job, yeah.
T. Got lost in it, and your father
versus
Ken: I got ul u.-got () lost in one jub=Yea:h.
(0.5)
(it bo:thered)
Dan: Got lo:st innit e[r ( yr fa:ther ) ( )
Ken: [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 its or, then while repeating the proffered ver-
sion 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 ther-
apist accepting, the other shows him rejecting, the patients 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 dont know about anything else.
DEAN: I dont 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 donnoo bout anything else exchh e[pt
Dean: [I dont either in I: wd
(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:rsd stopped outsi:de.=
Ada: =eeYe::[s?
Jessie: [h An that whether he thow:t thet I edn
ac[cidnt] [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 wont co[st her anything, she starts up with her acceptance of
his offer:
(7) [SCC:DCD:37:R]
Bryant: ...but hh if youve gohtta credit neote ih weont
co[st you anything anyw]ay.
Sokol: [ Wul owright the:n, ]
Sokol: Thats 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, knowl-
edge of which Dean doesnt 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 specif-
ically ignores its projection of further things (i.e., his I dont either targets
Nixons initial proposal, I dont 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 pro-
duces the word except with some breathiness and an unvocalized mid-vowel:
exchh ept. Again, here is a possible case of a recurrent phenomenon: A recip-
ient of someones utterance may treat something in that utterance as laughter,
and respond with laughter of his own. Heres a simple case:5
(8) [Schenkein:II]
Bill: Dju watch by any chance Miss International Showcase las night?
Ellen: n:No I didn[I wz reading my-
Bill: [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 Bills 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. Hes not, then, to be heard as slavishly
laughing just because the President is laughing. As well, the work hes do-
ing 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 Nixons 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 frag-
ment (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: wd (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, Deans 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....
Heres another sort of thing that can be noticed in a comparison of frag-
ments (3) and (4). In the former, Dean, shown of course in standard or-
thography, is saying ...I [laughs] almost hate to learn some of these things.
In the latter, among other pronunciational particulars, hes 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 Califor-
nia motorcycle gang, the Hells Angels, produces them followed by dere
(their):
(9) [KPFK:GJ]
Joe: I tell them right tdere face . . .
For example, a Bronx janitor produces several versions of there (and theyre):
theyuh, theyre,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 theyre uh (0.2) the tu- () u. () t-two guys uh
deyuh, n me n James Walkuhs 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 aint 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, Whos that. Whos that punk right there., (and also
producing a vowel-begun version of thats, Ahz Davey More.), B.J., saying
things like yall see these pictichiz..., yknow who dat is, and Thats Davey
More.
(12) [Labov:Jets:3-5]
Mez: Hey lookih dat.
.
B.J.: Jo man, we gonna talk about fights man yall see these pictchiz
up here man yknow who dat is,
Mez: Whos that. Who that punk right there.
Mayall: Mothuhfuck[im!!
B.J.: [A punk, [Thats
Mayall: [He stinks!
Alex: [Shuddup Mayall!!
Leonard: [Man- eh heh! [khhkhkhkhkhkhkh[khkhkh
B.J.: [Davey More. [
Mez: [Ahz Davey More
B.J.: Th-thats Davey More en thats- Benny Kid Paret.
I then go on to locate some possible systematicities and interactional phenom-
ena 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) var-
ious 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 couldnt 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 dont 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 couldnt set up ay perfectly legitmate (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 Caufield who: wz I don know if you rmember
Ja:ck er not he wz...
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 were 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 janitors pronunciations of there, or John
Deans pronunciations of to. But now and then, something in one of our more
detailed transcripts would catch his attention, and wed be treated to a little
something I find enormously amusing, having to do with someones 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, wed be told
of a fascination he had with [t]hings like didje and wanche, and his hav-
ing 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 mun-
dane stuff he was known to work with. These little tiny things were perhaps
beginning to be just a bit too tiny for comfort.)
But theyre 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 elim-
inating 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 Als
question to Ken about father.
. For a discussion, and further cases, of this phenomenon, see Jefferson (1984: esp. 2829).
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: 8283).
. Transcribing the Watergate materials, Ive been struck by the delicacy of some of Deans
interactional work. I didnt 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 acceptance-
declination. In G. Psathas (Ed.), Everyday Language: Studies in Ethnomethodology
(pp. 7996). 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, 112. Tilburg: Tilburg University.
Jefferson, G. (1984). Notes on some orderlinesses of overlap onset. In V. DUrso & P.
Leonardi (Eds.), Discourse Analysis and Natural Rhetorics (pp. 1138). 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. 2434).
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.
// Double obliques indicate the point at which a current speakers talk is

overlapped by the talk of another. ((No longer in use.1 ))
Louise: N how t//all are you, Al,
Roger: How tallr 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 speakers
talk is broken up in the transcript, but is actually through-produced by
its speaker.
Pammy: Yeah well okeedoe=
Myra: =[Yeah.
Pammy: [I js thought Id ask
1. The asterisk (*) was used by some transcribers to indicate termination of simultaneous
speech in conjunction with double obliques (//). (Ed.)
In this case, Pammys utterance is produced as Yeah well okeedoe I js

thought Id ask, with Myras Yeah starting up immediately upon com-
pletion of okeedoe, and simultaneously with I js. . ..
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 imagine.=Now
(0.4) p since then Ive retained counsel.
(0.0) Numbers in parentheses indicate elapsed time by tenths of seconds.
Al: . . .js be a lotv (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: Im 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: |
_____ [Its 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 cooprated
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 its all in the chair=

Mike: =all th[at junks 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: =Howr 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 prolonga-
tion 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 syl-
lables thereafter are punched up.
Ehrlichman: He said e -I came dih you:,hh frm 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: Thets a good question.
(0.6)
Louise: Thank you.
.,?? Punctuation markers are used to indicate the usual intonation. (The ital-
icized 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 appro-
priate syntactical points, but occasionally there are such displays as the
following.
Maggie: Oh Id say hes about what.=five three enna ha:lf?=arent
chu Ronald,
Sometimes, at a point where a punctuation marker would be appropriate,
there isnt 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 () tihda: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: Bt the point is Jessie dont fehget no:w. h (0.3) eh:m
() e w so: close t Martn wa:snt 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 IM NO(h)T I(h)NTUH THA(h)*T! * *
In non-computer transcripts, the asterisk indicates creaky voice. (In com-
puter transcripts, Ive 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: [it-
Emma: Oh:: God ih wz 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* 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 wjist () understa:nd thet* uh: you en I are deh-
absooly 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.
t, d In more recent computer transcripts, a boldface consonant replaces the
single sub- or superimposed dot which serves as a hardener in my non-
computer transcripts. The above fragment would now be shown as:
Kalmbach: I wjist () understa:nd thet uh: you en I are deh-
absooly dihgether on tha:t,
Ehrlichman: No question about it?=uh hHerb
a.,e, When a single dot is not available, two dots over a vowel replace the sin-
gle 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 -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 dou-
ble dots over the i-, confirm that those sounds are short. I dont 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 or-
thography. 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 con-
centration, 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 cnditioner, 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.
.
.
(b) Kalmbach: . . .he said the rea:son thet wz: u.-fer the ca:ll wz. . .
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
its 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 evn putcher hand ou:ts:I:de the CAR
ih jiz(b)bu:rn.
...
Ehrlichman: But they- () the(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 theyve 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 frm
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 ef-
fect, 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) <I mean we go to ceramics,

...
Polly: Ysee its diff rent f me:. <eh f () the othuh boy:s,
word< A post-positioned left carat indicates that while a word is fully completed,
it seems to stop suddenly.
Meier: Uh well I fel like my lef side of my () chest I cd () mah
had a k- cramp<
A dash indicates a cut-off.
Vic: He said- yihknow, I get- I get sick behind it.
>< 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 hs indicates an inbreath. Without the dot, the hs
indicate an outbreath.
wohhrd A row of hs within a word indicates breathiness. In some transcripts the
hs are italicized, in some not.
Colson: . . .a ghhuy wh(h)os olso totally loyal<
And in computer transcripts, the hs are sometimes also superscribed.
Colson: Th:e thing thet worries me is the(p)() is the
phh ossibilityv. . .
(h) Parenthesized h indicates plosiveness. This can be associated with laugh-
ter, crying, breathlessness, etc.
Jim: Dont sound so (h)amp(h)itious fer Ch(h)risesake
(h)ih suh hh sou l(h)i yuh k(h)uh g(h)o tuh sleep
n the pho(h)one.
...
Maggie: I jst ran up th stai(hh)rs thats wh(h)y Im
huffing en puffing.
The pound-sterling sign indicates a certain quality of voice which con-
veys suppressed laughter (various transcripts have other symbols, e.g.,
the Dutch guilder-sign which no longer exists).
Ken: ahh ha I don know whos 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
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 un-
gotten talk.
Mike: No.
(0.4)
Mike: ( ),
In the speaker-designation column, the empty parentheses indicate tran-
scribers inability to identify a speaker.
Roger: Pazm zm 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 shd talk to Witnaw. (0.5) And uh: () jis brace
eem n tell im tih ()(offih sebbatikiss). . .
() 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 ahll protect chu but* uh)
|
( ): (0.7) (Okay.)
|
Nixon: __|__ (Thet uh) thets thats why:. (0.9) I (0.5) cant
let chu go (0.2) go dow:n.
(( )) Doubled parentheses contain transcribers 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 talks 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 features
preservation must take work, the taking of turns must be organizationally
achieved. Here, on the basis of audio recordings collected from naturally oc-
curring 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-takings organization in conversation and also
Harvey Sacks
because it is plainly not correct for other2 than conversation. The generaliza-
tion 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, meet-
ings 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 sug-
gest 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 ar-
rangements, 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 com-
pared 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 par-
ties 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 func-
tion that position permits maximization of, characterization of the organiza-
tion 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 gen-
eralization 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 speci-
fiable 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 corre-
lated 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 organization-

ally 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 inte-
grational problems. Except when a simultaneous no gap and no overlap tran-
sition between speakers is achieved, minimization of either one is not directly
compatible with minimization of the other. Means for reducing gaps encour-
age 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 de-
fines 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 tech-
nique 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
in a turn, the chance to allocate next turn is not assured. For at that point,
Rule 3 becomes operational, and current speakers turn may be terminated by
a self-selecting speaker.
Rule 3 makes self-selection optional. Rule 4 permits current-speaker con-
tinuation 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 con-
strain 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 self-
selection 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 comple-
tion 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 speakers 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-speaker-
selection if it is to be assured, and by providing that the possibility of current-
speaker-selection be a recurring first option that thereby constrains the place-
ment 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 self-
selection 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 in-
tegration 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-at-
a-time occurs and overlap and interruption these constituting the two
ways that more-than-one-speaker-at-a-time occurs. The distinctions are fur-
ther 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 minimiza-
tion 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
speakers selection of a next gaps may be avoided or reduced by newly start-
ing 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 anothers turn, then interruption differs from overlap, if we under-
stand by the latter such a start as is projected to occur on anothers 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 pos-
sible 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.
Current speaker might extend the length of a correctly appreciated last

word, and next speaker, intendedly projecting his start to occur on that words
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 cur-
rent speaker attempting to avoid a gap while intending next speaker is attempt-
ing to avoid overlap, their occurrence is irremediable. But interruption, as it
involves a start that is projected to occur within anothers 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 speakers 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 speakers 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
speakers 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 gaps 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 situa-
tion 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 ac-
quire 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 ver-
sion, beginning with this paragraph, the remainder of the manuscript has been retyped in-
corporating 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 incor-
porates 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 sys-
tem 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 peoples 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 per-
form specifiable activities. Within-overlap talk can involve procedures for re-
solving 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: Hes not gunnuh li:sten [tuh tha::t,
Jim: [Im 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: Hes not gunnuh li:sten [tuh tha: :t, ]
Jim: [Im 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 in-
quiry 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 proce-
dures. Specifically, a party can precision-place his talk in the course of anothers,
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: So he come[s home one nightn the sonofa] bitch [bit him.
Carol: [ 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: Fire Department, out at the Fairview Food[mart theres a-
Desk: [Yes.
Desk: Weve already got the uh call on that maam,
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: Hows everything look.=
Bud: =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 thm 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: Just like tha:[:t.
Mike: [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 co-
herent next utterance component while a next speaker is starting up in latch
position, overlap occurs.
Bert: Uh you been down here before [havenche.
Fred: [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 fea-
ture holds for relatively long pauses, for example, seven tenths of a second.
Timing is shown in parentheses between the two utterances.
Ava: Hen Jo were like on the outs, yihknow?
(0.7)
Ava: [[So uh,
Bee: [[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: (Ill) set it dehr own the sidewalk.
(0.15)
Vic: [[No.
James: [[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 pro-
ducing 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 over-
lapping 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: [[T-
Lady: [[Ye:s.
First starter can be found to drop out as second starter begins.
Essie: I think Cookie [ta-
Janet: [I didn even knowe was i::ll.
Gail Jefferson
And a second starter can be found to drop out immediately after an attempted
start.
Dan: May [be yer brother is . . .
Louise: [( )-
Multiple serial starts and drop outs can be found in association with an on-
going utterance which has several points of possible completion plus contin-
uation 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: I jus thought it was so kind of stupid=
Janet: =[[Y-
Polly: [[I didn even say anything=
Janet: =[[Eh-
Polly: [[when I came ho:me.
(0.3)
Janet: 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 pro-
duces an utterance with a first possible completion and a continuation (Well
Essie jus called + n I aftuh caller 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 ut-
terance, 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.
Johnson: [[I-
Roberts: [[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 mi-
cropause, both parties drop again, and subsequently one takes and is given the
turn space.
Edna: [[Well-
Bud: [[Un-
()
Edna: [[uh-
Bud: [[less-
()
Bud: If evrything 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: But [had-
Lady: [But
()
Lady: [[this-
Tracy: [[hh Ha-
()
Lady: [[Ni:neteen s e]v e n]ty I,
Tracy: [[Hadju finished] ( )]
()
Lady: was:: had n invitation to, (0.3) Buckingm 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, perse-
verance 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 ad-
justments; 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: [[Im glad to hear it.
Roberts: [[But- uh- uh, uh understand that um Franklin . . .
...................................................
Gladys: [[En-
Edna: [[En you need [som:e ]uh,]
Gladys: [S- s- ] sh:]redded lettuce?
(b) Stretches. In the following fragments a word is prolonged, and the pro-
longation 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 lotta-
Roger: [Les:::::::try it!
...................................................
Carol: Tha:[ts 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 utter-
ance components to be produced clear of overlap.
2. Segmental adjustments
A similar provision for some part of an overlapped utterances occurrence in

the clear can be observed for these devices. While the pronunciational adjust-
ments 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
an intuitive sense that one party is Turn Occupant, the other Turn Claimant,
with Turn Occupant simply producing an utterance while Turn Claimant pro-
duces serial segmental adjustments.
Mike: [[Th-
Vic: [[Yknow I cut [m y s e l f [o n y o[u : r f r e a k i n [gla:ss,]
Mike: [Th least e [ cdo- [Th least e coulda [do::ne,]
()
James: [[ Y e : h , ]
Mike: [[Least e c]dv done wz come dow:n en letchu know what
happen
Of the two overlapping utterances, one has a phrase recycled over the con-
tinuous course of another, permitting some projectable part of the recy-
cled utterance to occur clear of overlap. The same feature holds for the
following fragment.
Ken: No, theyre women whov devo[ded their l-
Roger: [Theyre women that hadda=
Roger: =[[bad love [lifen became nuns.hh [heh hh !
Ken: [[their [their life- [their life, to uh
(0.6)
Ken: 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 components actual content is known in advance, is formed
up for speech-delivery, etc., it can, in principle, as not yet more than a pro-
jected component type, be reserved for overlap clearance by a within-overlap
recycling of its prefatory segment(s).)
In these fragments, while Turn Claimants work is obvious, Turn Occu-
pants talk, as work, is not transparent, i.e., it is not obvious that Turn Occupant
is doing something vis--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 Occupants talk is subjected to.
More to the point, we might consider each recycle by Turn Claimant as a re-
quest 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 re-
cycled utterance segments) and one neither recognizes nor attempts to remedy
Gail Jefferson
an observable trouble. We might call the former procedure Marked Compe-

tition 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 con-
sequence 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 talks
consequence, and Other-Retrieval, via which a party to an overlap provides for
someone elses, not his own, talks consequence.
Further, these two groups each have Marked and Unmarked forms; forms
which announce trouble and explicitly retrieve talk out of the prior over-
lap, 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 partys 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: [[Hy -
Olive: [[Yeah.
()
Edna: Hide it.
At first starters dropping out as second speaker starts:
Rich: I think if [ you-
Carol: [Am I right?
()
Rich: If you bring it intuh them.
And at second starters dropping out after an attempted start:

Louise: I wnd if [rilly is someone back the[re,
Roger: [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 consider-
ation 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: [thats not ba:d.]
(0.2)
Bee: That class is suh:: yihknow this is the Indian class. . .
...................................................
Fran: Hes not gunnuh li:sten [tuh tha: :t, ]
Jim: [Im not say]in- Im 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 partys 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: [[ The-
Bert: [[Yah-
()
Fred: waves r about tuh wash us away.
...................................................
Roger: I happen tuh wear buloo jeans constantly.
(0.3)
Ken: Well,
()
Gail Jefferson
Roger: Even [in-

Ken: [so do I now,
()
Roger: formal occasions, yknow? 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 un-
marked self-retrieval; i.e., continuation, one constructing a continuous Well
uh while the other constructs a continuous Unless.
Edna: [[ Well-
Bud: [[Un-
()
Edna: [[ uh-
Bud: [[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: But [had
Lady: [But-
()
Lady: [[ this-
Tracy: [[ hh Ha
()
Lady: [[ Ni:neteen seventy I,
Tracy: [[ Hadju finished ( )
()
Lady: 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 initi-
ates repair procedures by requesting a repeat of his co-competitors overlapped
utterance.
In the following fragment, each party displays Turn Occupancy and the
continuously produced utterances reach completion simultaneously. There-
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 initia-
tion of a new topic; the explanation Before he gets home yielding to You
goin upn getcher hair fixed tuhday,. (Whether this is a generalizable feature of
overlap management remains to be seen.)
Gladys: En then you could return it uhb, oh along about noon.
()
Edna: You goin up n get [cher hair fixed tuhday, ]
Gladys: [Before he gets home. ]
(0.5)
Gladys: What deah?
()
Edna: Yer goin up tday en getcher hair fi [xed.
Gladys: [Oh no.
In the following, each party displays Turn Occupancy, with one partys utter-
ance extending beyond overlap resolution. Perhaps a way to characterize the
conditions under which it is decided who should yield is to find that First Stop-
per has yielded to Last Stopper. (Again, whether this is a generalizable feature
of overlap management remains to be seen.)
Lil: A:n Im hoping a lo:t,=
Lil: =hh thet [youll] [do a few a them,]
Tony: [(Yeh)] I [ bet yooer t e r r] ibly hoping,
(0.2)
Lil: t! Hu:h?
(0.2)
Tony: t! I bet yer terribly [hopi [ng.
Lil: [hhh [Im terribly terribly hoping.
In the following two fragments one party displays himself to be Turn Occu-
pant, 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 Occu-
pant/Unmarked Competitor/First Stopper produces an utterance plus contin-
uous, extended laughter while Turn Claimant/Marked Competitor/Last Stop-
per produces a complete question repeated three times over the course of the
continuous display.
Gail Jefferson
Mary: [[ I s i d I get mo]re the n h(h)alf=

Sue: [[Wuh year was it?]
Mary: =hh(h)unhh heh [hehh hnnnn hnnnn ]
Sue: [ Wht year was i:t? ]
()
Mary: u.-hnnn huh hu[h hnnnn!]
Sue: [Wuh year ] was it?
()
Mary: Ha:h?
()
Sue: What w-year wa[ s it? ]
Mary: [Ni:ne] teen fifty.
And in the second fragment Turn Occupant. . . etc produces a single coherent
utterance: Okay duh soopuh. Freak it. Hes a bitch he didnpud in duh light
own dih sekkin flaw, hh man tell im. Ykno:w? while Turn Claimant... etc.
produces revisions of a same question: Y didnt getta holda-, Listen man.,
Y couldnt gitta hol-, Jim wasn home uh what.
Vic: Its, the attitude of people!
(1.0)
Vic: Okay[
Mike: [Y[didnt getta holda-]
Vic: [d u h s o o p u h ]
()
Mike: Listen [man.
Vic: [Freak it. Hes a bitch he didn pud in
duh light own dih sekkin flaw, hh=
Mike: =Y couldnt gitta ho[l-
Vic: [Man tell im.
()
Mike: Jim[wasn home] uh what.
Vic: [ Y kno:w? ]
(0.5)
Vic: Hah?
()
Mike: Jim wasn home [ (Whenyih wenovuh deh) ]
Vic: [I d i d n g o b y t h e h. ]
(This particular case is transparent for the constructedness of the competi-
tive continuous coherent utterance, with the bulk of its components starting
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 Occupants Y kno:w? followed by Turn
Claimants 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 ac-
knowledgment token (Yeah, Uh huh, etc.), one party treats the others
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 people-
Bea: [So dont
()
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-competitors utterance, but not
directed to its subsequent consequence. So, for example, in the following frag-
ment an acknowledgment token turns out to preface a restart i.e., a marked
self-retrieval.
Helen: Specially when they [don know what theyre doin]=
Bill: [ N o t b e i n g b- ]
Helen: =in the [f i r s t ] place.]
Bill: [Yeh no]t being] bulatnt.
And in the following, an acknowledgment token turns out to preface a revised
a restart.
Gail Jefferson
Olive: I didn know what time you were gonna get down, so I
wen out [shopping yester[day,
Edna: [Oh::: we got- [Yeah=
Edna: =We didnt get down till about- (0.6) Oh I dont know,
six uhclock . . .
The acknowledgement token as accessory to competition may import some
function from its operation as a recognitional in recognition-placed over-
lap (see page 44, point 2, and Jefferson 1973); i.e., proposes that hearing-
understanding 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 anothers over-
lapped talk by incorporating the others possibly-unheard materials into his
own next utterance, resulting in an undisrupted flow of talk. For example:
Clara: Isnt that place [something.
May: [I tell yuh.
()
May: It is really something.
...................................................
Lottie: How come yih didn stay. Oh ih wzis [too hot huh,
Emma: [OH-:: there-
()
Emma: Jus too hot Lottie, an it was. . .
And the combination of acknowledgment and repeat provides for an unequiv-
ocal retrieval of another partys talk.
Ken: I-I made pretty good time, but its [tiresome.
Louise: [But it was one pers-=
Louise: =Yeah its tiresome.
...................................................
Carol: [[He-
Denise: [[Hes ser[vin Jody.]
Carol: . [ serves? ]
Carol: [[ Jo:dy?
Denise: [[Yeh!
(.)
Denise: He serves Jody.
(In this case both parties display Turn Occupancy, one with a serially con-
structed single coherent utterance: he + ser:ves? + Jo:dy?, the other with a
through-constructed single coherent utterance: Hes servin Jody. Note that
the party who provides the acknowledgement token thereafter produces the
other partys version of that same statement; i.e., shifts from Hes servin Jody
to He serves Jody.)
In the following it appears that each party is attempting to acknowledgment-
token retrieve the others version of a same statement (drops vs. dump).
The matter is resolved with a combined acknowledgment token/repeat.
Loren: . . .people thet jus come en 1eave their little kids
et th skating rinks appalling.
(pause)
Loren: Yknow[drops them ]
Kate: [dump them ]=
Loren: =Yeah.
()
Kate: Yeah.
()
Loren: 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 anothers talk, simultaneously
stand as an offer to delete ones 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 possi-
ble 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 Stoppers utter-
ance (Oh God). Subsequently the initially withdrawn, now repeat-request-
declined materials (that they had rescued a lot of people) reappear.
Gail Jefferson
Kitty: They said theyve been, huge wa:ves.=

Kitty: =en they [jus rescued-
Frank: [Oh::::: God. Over the pie:r,
()
Frank: Huh?
()
Kitty: Oh Go::d.
(0.2)
Frank: Excep tuhday its t-calm down, I guess ih was a storm
out et [sea,
Kitty: [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.
()
Frank: 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 pre-
sented 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 pro-
cedures constituting means of performing specifiable activities; specifically, a
party can precision-place his talk in the course of anothers, 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 ob-
ject 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
point is the moment of completion of an ongoing utterance, this last result-

ing in overlap should the currently speaking party continue talking beyond
that point.
(2) Within-Overlap Talk can involve systematic procedures for resolving over-
lap and/or attempted solutions to a problem which arises as to who should
drop out. It can also involve systematic procedures for competing within over-
lap, 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 over-
lapping talk, and can involve systematic procedures for retrieving talk poten-
tially not heard due to its occurrence in overlap; i.e., for providing for its con-
sequence in subsequent talk. And these procedures seem to be distributed into
types: Self-Retrieval and Other-Retrieval, each type with its marked and un-
marked forms; restarts constituting marked self-retrieval, continuations con-
stituting unmarked self-retrieval, repeat-requests constituting marked other-
retrieval, 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, 4796.
Schegloff, Emanuel A. (1987) [1973]. Recycled Turn Beginnings: A Precise Repair Mech-
anism in Conversations Turn-taking Organization. In G. Button & J. R. E. Lee (Eds.),
Talk and Social Organization (pp. 7085). 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*
It has been suggested1 that the production of an answer to a summons is a se-

lectional issue. It is selectional not only in the interactionally interesting sense,
whereby that an answerer can select a problematic answer can serve as a con-
straint 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 selectional-
ity 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 pro-
duction of a summons, he may, as was suggested above, see the outcome of
relevant pre-assessments by summoner of the comparative priorities of any on-
going 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 summoners analysis that it has such claims of
priority and temporal fit that would warrant its introduction into an ongo-
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 dis-
play 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 sum-
moner, 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 be-
ing 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 ad-
dress, 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 anal-
ysis how they were selected, given that that they are selected is something
summoned can orient to. They display features relevant to such identifica-
tions 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 sum-
mons 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, inter-
actional 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 an-
swer 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, be-
tween 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 sub-
ordinate 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 utter-
ance, 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
summoners preassessments. It provides, so to speak, a carte blanche.
Answering the summons on the telephone
The features of the opening of telephone conversations are somewhat differ-

ent.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 de-
veloping 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), pend-
ing business, regular calling practices, whos 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 pre-
assessment 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 decid-
ing 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 in-
ference that are relevant, they are typically not treated as adequate grounds
for answering the telephone with a problematic answer. Indeed, in the ab-
sence 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, how-
ever, a carte blanche clearance cue as the sole possibility. While such a non-
constraining 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 conver-
sation, 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 conse-
quences of the preceding discussion. Hello and self-identifications (or self-
formulations) are not typically used answers to summonses in unmediated in-
teraction. 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 prob-
lematic 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 sub-
class 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 inter-
actions and telephonic ones is not an analysts 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 interac-
tion, 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.
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 rele-
vant) 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 anal-
ysis is available to the summoned, he may answer yeah, as in the following
materials selected from the corpus
#500
D: Yes.
C: Do you want your private ambulance up there? etc.
#391
D: Yeah
C: Listen, I got a hold of Colonel _______ . . . etc.
#350
D: Yes.
C: Youd better put emergency 5 in on that, weve got . . . etc.
Or the following transfer of call by intercom, drawn from another organization:
L7: No, listen, do you want to talk to Hal,
he can probably be more technical.
S7: O.K.
L8: O.K. Just a minute.
S8: Right.
(pause)
H9: 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: Well, call er can call me back.

J: O.K.
E: Guhbye
J: Yeah.
(Trio, I)
After a three minute 37 second conversation with P, J calls E back:
E: Yah?
J: Well, she doesnt know.
(Trio, III)
These circumstances, in which answers regularly used in face-to-face interac-
tion are used on the phone as well, are ones in which at least some of the infor-
mation 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 instru-
ment which may show the call to be intercom, or by the temporal organization
of the telephones 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 ana-
lyzed 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 rel-
evant 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 se-
lected 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, re-
sumed, 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 ini-
tiated, 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
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 puz-
zle). 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 orien-
tation 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, invok-
ing 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 anothers 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 whats on everyones mind;
not only everyone displaying what is on his mind, but everyone displaying
what they take it is on everybody elses 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, soli-
darity, 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 whats on your mind? once he knows who you are.
Thus, in a call from a reporter to a fire department:
A: City Fire Department, Livingston speaking.
B: This is Carrie Fortune.
A: It was a garbage rack.
B: Oh(hh)kay. Thank//you
A: heh heh heh Ri(hh)ght
B: Gbye.
(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 project-
ing the conversation as a particular type of conversation, some considerations
and constraints are introduced to which the co-participants are to be oriented
in constructing the conversation in its course, considerations and constraints

which in the case of any particular conversation require attention to the partic-
ulars 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 ear-
lier 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 doesnt know.) Where a shared orientation to priority topic
has been invoked by the calleds answer, other possible moves by caller in the
next slot may show that the expectation on which the answerer based his se-
lection 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 answerers 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 an-
swerers 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 con-
fronted 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 prior-
ity topic, co-orientation to which is presumptively displayed by the answer,
non-proper callers 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: Yeah.
S: Hi, Hal?
H: Yeah, what can I do for ya?
S: Marty Anderson here.
H: 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 out-
sider 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 ea-
gerness [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 war-
rantable 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 answerers 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 pro-
ceed 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 inter-
action 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.
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 greet-
ings. 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 co-
participant, 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: Hello/
B: Marty Rosenthal calling collect, will you accept the charge/
A: Yes, certainly, operator.
(CF, p. 22)
A: Hello
B: Are you awake/
A: Yeah I // dis got up
B: I Oh didjuh/
A: Yeah
B: hh Weh gooud. Im alone.
(NB: IV:3, p. 1)
A: Hello
B: Is Jessie there/
A: (No) Jessies over et er grammas fer a couple days.
B: Aright thankyou
A: Yer welcome/
B: Bye
A: 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 refer-
ence to showing recognition, or being consistent with informal or friendly
relations, i.e., hi.14 For example:
A: Hello.
B: Marty, Al.
A: Oh, hi.
(CF, p. 24)
A: Hello/
B: Martin/
A: Yeah/
B: Hi, this is Sophia

A: Oh, hi.
(CF, p. 25)
A: Hello
B: Eddy
A: Yeh
B: Guy Huston.
A: 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 ini-
tial hello of some telephone conversations is treated by co-participants, or is
to be treated by analysts, as straightforwardly a greeting. The grounds for treat-
ing the initial hello as an answer to a summons have been developed in Sche-
gloff (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 rings 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, callers first utterance may begin with a greeting term which then
seems to operate as the second part of a pair, as in:
A: Hello/
B: Hello. Im trying to locate Professor ___________, is he there/
A: No, hes not.
B: Thank you
A: Mmm
(EAS, FN)
On the other hand, while the occurrence of the answerers initial hello does
not entail any subsequent greeting by the answerer, it is also the case that
callers 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: Hello.
B: Allo.
A: Gyouyour roomate talks forever or is it you?
B: 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 ini-
tial 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 first-
order activity that the utterance is accomplishing. Above all, the. initial hello
is an answer to a summons. Whatever else it accomplishes, it accomplishes (un-
less 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 non-
occurring, 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 accom-
plishes 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 callers 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 calleds 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.
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 ini-
tial 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 re-
turn greeting it need not be found absent. Are there then no constraints or no
specially relevant actions to be done in the callers 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 orga-
nized 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 ini-
tial 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 be-
ginnings; 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 re-
turn 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 sev-
eral 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. Em-
pirically, 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 inter-
ests 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 con-
versations 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 en-
counter 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 salu-
tation 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 possi-
ble); when there is a planned or intended encounter, more than a minimal
pair is done. The notion of accidental and intentional encounters implic-
itly 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 acquainteds version of by-product and designed.
That members may treat as relevant whether an action has been a pro-
duced designedly or as by-product, as a general distinction, can be seen as well
in Sacks observation (1992a: 792793) 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 invitations
achievement, the conversation having been initiated by the other; such an invi-
tation may be seen to be interactionally generated, while callers invitations may
(though they need not invariably) be seen as designed. While interactionally

generated invitations may be treated by offering counter-invitation (why
dont 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 argu-
ment here is that conversation as a whole may be similarly analyzed, or more
particularly, conversation beginnings may be, and different relevancies be as-
sociated with different findings (The point about specifying the discussion to
conversational beginnings is that once initiated, transformations of conversa-
tional 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. 104105, 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 ac-
complished as by-product when persons find each other to be co-present,
do co-presence acknowledgments (e.g., see Goffman 1963a: 8388 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 conver-
sation 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, oc-
cupying 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 ar-
rival, 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 con-
frontation (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 pre-
liminaries 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, conversa-
tion. (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 engi-
neered 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, sys-
tematic, or methodic account of the doing and analyzing of conversation
initiations as designed or by-product, it may suffice for the present oc-
casion to suggest the types involved and some basis for seeing that they are
relevantly discriminated by members. (Where the parties are entitled to mu-
tual 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 invari-
ably 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 con-
versation, 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 le-
gitimate reason for initiating the conversation. While I cannot here provide an
adequate account of what constitutes a legitimate reason, that being in any
case a matter that may be contested by the co-participants, several sorts of rea-
sons which may claim a sort of prima facie legitimacy have been suggested by
others; for example, Goffmans (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 anyones com-
ment (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, hav-
ing them assumes a different status, as will be discussed below). In face-to-face
interaction, of course, acquaintanceship may be established by visual inspec-
tion. 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 indef-
initely 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 pop-
ulation 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 ac-
quainted 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 col-
lection 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 ac-
knowledgment 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, mem-
bers 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 didnt even say hello, He (you) didnt
(dont) recognize me), or, if those are not supportable, Hes angry, etc. Mem-
bers 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
an orientation to meeting it, seem to entail for members monitoring the en-
vironment 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 un-
acquainted.21 It may therefore be more correct to see the initial adequate
grounds for starting a conversation to be the identification acquainted, rea-
sons 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 identifi-
cation, 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 ac-
quainted 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 unac-
quainteds.
While I have urged that acquainted is an identification of first-order rel-
evance to the beginning of conversation (although it should be recalled that
discussion is intendedly limited to designed conversation), other identifica-
tions 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 partic-
ular, 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 con-
versation 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 there-
fore, for achieving co-presence), acquaintances may have reasons too. How-
ever, 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 start-
ing 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 rel-
evance or adequacy of the acquaintanceship per se as a basis for conversation.
As Sacks observed (1992b: 163166), 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 havent 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 dont
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 particu-
lar now that they are starting a conversation. Indeed, the character of some
members acquaintanceship may be defined by the regular no reason initi-
ation 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 ev-
ery 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 collec-
tion 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 conversa-
tions 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
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 ba-
sis for the conversations 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 ac-
tion, 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 inter-
nal development of conversations that may be ongoing. Occupants of adjacent
seats on an airplane, those waiting in waiting rooms, at bus-stops, for eleva-
tors, at ski lifts, etc., are in such settings. When by-product conversations are
started in such settings, initial utterances regularly display attention to the co-
membership 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 ana-
lyzing 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 recog-
nized acquaintanceship, or by legitimate reasons (where legitimacy may be re-
lated 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 in-
variably 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 conversations
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 be-
ginning 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 ac-
quaintanceship 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 organiza-
tion 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 answerers initial hello does not prospectively invoke identifi-
cations or reasons by way of utterance-to-utterance organization, but rather
occasions their relevance through the overall structural organization of de-
signed conversation, a caller doing identification work or reasons must ac-
complish 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 answerers 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 callers 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 va-
riety 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 accom-
plish and how they are selected, to which callers first utterance is fitted (as
utterance and as type). In doing so, we will have occasion to examine some in-
teractional resources employed by answerer to select an answer (not to select
between clearance cue answers and problematic answers, for, as was suggested
above, the interactional resources for that selection appear unavailable in tele-
phone conversation beginnings; but to select between possible clearance cue
answers), and whose use for the selection is displayed in the answer, and is em-
ployed 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 prop-
erly 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 per-
sonal 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 set-
ting 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 sug-
gest 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 fea-
tures, 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: Hello
B: Hello::?
A: Yeah. Hello.
B: Wuhis this 293-4673
A: No, its 293-4637..
Answering the phone
B: Oh, Im awfully sorry.

...
A: Am I supposed to be a business firm?
B: Yes. Thats right. Thats exactly right. Im calling my office. They
never answer with hello::.
(GZ, p. 27)
Hello, then, in allowing possible business and possible personal con-
versations, 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 sum-

monses, as the answerers initial utterance, is that they are preemptive iden-
tifications. 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 initiators work in the case of telephone conversation, the callers
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 initiators (callers) first utterance after the possible start of a conver-
sation has been established; where an initiators 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 callers first utterance that that work is rele-
vant. Answering the telephone with a self-identification is preemptive because
it does the work of identification before the turn-taking organization has pro-
vided caller his first opportunity for doing so. At callers 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.
An answer such as Police Desk accomplishes an identification of a par-

ticular kind, and thereby makes the identification issue the provision of identi-
fications 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 of-
fered as part of the work of warranting a conversation, it identifies a particular
initiator or caller, it addresses the issue whos calling, and may, for example,
use a personal name to accomplish the identification. Self-identification an-
swers of the form of Police Desk make the relevant sort of identification not
personal but categorical. Identificational interest is then not in who in par-
ticular 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 initi-
ated), 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: Police Desk.
C: Uh, I have a large uh Pontiac station wagon. Do you think you could
use it?
D: 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.
C: Ill be right out there.
D: Thank you.
(IPD, #19)
D: Police Desk
C: Uh uh a car accident at 4700 East Lincoln, Taylors Lane and Lincoln.
D: 4700 East Lincoln?
C: Yes.
D: Anybody hurt there?
C: I dont believe so.
D: O.K. . . . Ill have somebody check.
C: 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: Police Desk, what is it please?

C: Un its Watson.
D: Huh?
C: Watson?
D: 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 callers first utterance may be inspected for pos-
sible acquaintanceship identification, for example, through voice recognition
(Schegloff 1979), answerers in this type of call display no interest in personal
identifications in callers 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 be-
ing his entitlement or his grounds for starting up that are at issue, the self-
identification offered in the answerers initial utterance is identification of the
called. In this regard, it is, then, critical to note that an identification of ei-
ther 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. 8285 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 identifi-
cations, the consistency rule, holds that if a first identification of a member
from a population to be categorized or identified is selected from some collec-
tion 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 col-
lection of terms if possible (as is argued in Sacks 1972b), as a rule for speakers
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 su-
pervening rule, and allowing non-consistent identifications thereby to display
the operation of a supervening rule. Unless a supervening rule is used to se-
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 identifi-
cation 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 them-
selves), 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 col-
lection of identifications for callers related to the professional identification
of the answerer; in the case of most of the calls in our corpus, the identifica-
tion 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 ac-
quainted 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 rea-
sons which were not legitimate for unidentified members, anybodies, being
legitimate for members identified in some way. There it was argued that iden-
tifications 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 forthcom-
ing. This tie between identifications and legitimate reasons should now be seen
for its restrictive possibilities as well. While an identification may make legit-
imate 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 identi-
fication 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 conver-
sation 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 specif-
ically, for a complainant to initiate a conversation with the police complaint
desk). Police Desk makes the relevant task of callers first utterance giving a
legitimate reason for starting this conversation; and by introducing the iden-
tifications 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 prospec-
tive conversations they initiate as business calls, proffering identifications
of the parties by reference to the sort of business the prospective conversa-
tion 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 top-
ical 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 orienta-
tion 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:
D: Police Desk.
C: Id like to report a vandalism.
(IPD, #377)
D: Police Desk.
C: Id like to call up and report a break-in.
D: Whereabouts.
C: In my on my car, I dont know for sure whether they got in or not.
The both of the doors was still locked, but uh the windshield is on
the drivers 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 le-
gitimate 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 answerers 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 avail-
able, 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 rele-
vance 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 rele-
vance 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 de-
termination that conversation was relevant would thereby select who should
begin it). Indeed, it is not that the relevance of a self-identification form re-
quires 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, implic-
itly thereby either treating the selection of an answerer as non-problematic,
or temporarily locating the matters being discussed at a point after an an-
swerer has been selected. The selection (or self-selection) of an answerer is not,
however, automatic, and since some of the considerations relevant to that se-
lection are relevant to the answerers 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 hypo-
thetical 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. Proxim-
ity 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
might be used the oldest or youngest, highest status or lowest status, person
being the one whom the phones 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 phones ringing, or the rights or obligations of the posi-
tion 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 ini-
tial 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 dele-
gated 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 en-
tertained 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 for-
mulation 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 mem-
bers 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 par-
ties 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 al-
though 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 popu-
lation 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 answer-
ers in particular settings, the composition of those populations, and the bases
for inclusion or exclusion of particular parties are matters for local determi-
nation 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 answer-
ers 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 consider-
ations that the structures set off by the telephones 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 an-
swerers pertain to the prospective conversation that the telephone ring fore-
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 phe-
nomenon, being socially placed in some territory (in the ethological sense)
or itself defining a location. A phone is, in the first place, somewhere. Over-
whelmingly, furthermore, the territory to which a phone is attached is a some-
ones 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: do-
mestic 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: Hello
B: Hello, Mr. Smith?
A: Yes?
B: May I speak to Mrs. Smith, please?
A: There is no Mrs. Smith.
For acquainteds, the presumptive formulations of personnel affiliated to do-
mestic phones may be replaced by others:
A: Hello
B: allo
A: 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 per-
sonnel, 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 af-
filiated 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 consid-
erations 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 con-
siderations 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 conversations initial turn, bearing on such selec-
tions as hello for domestic phones, and a self-identification form for busi-
ness 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 immedi-
ately 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 telephones 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 poten-
tial 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 telephones ring, thus, invokes a set of con-
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 an-
swerers get selected to answer the phone, and which strongly condition their
selection of an answer form to constitute the conversations first turn.
Notes
* This chapter is a revision, completed in 19691970, of chapter three of my doctoral dis-

sertation, 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 for-
mulation 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 summoneds 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, 5662). 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 face-
to-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
Goffmans suggestion that . . . telephone talk . . . must first be seen as a departure from the
norm, else its structure and significance will be lost (1964: 135136) 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 Its Jim, the answerer does not regularly answer Hi, Jim.
. But just a moment is; it, however, has different properties when used in telephone con-
versation; 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 con-
vention. 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 them-
selves 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 formu-
lations 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
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 prob-
lematic 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 con-
trast is with an answer like Macys 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 com-
pared 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, com-
bined 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 ac-
quainteds 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 answerers initial ut-
terance 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: 112148, although the present
discussion differs in several respects from Goffmans, 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: 112114; 5768).
. It is the occurrence of absent first greetings that forces the specifications about enti-
tlement 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 possible-
acquaintance-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 dont know me, but. . .
. Where recognition is not available to visual inspection, as when persons who have cor-
responded 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: Paul Smith?
B: Yes?
A: Im Al Jones.
B: Oh, hi.
A: 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 his-
tory 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
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 indi-
cated 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 con-
versation, 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 em-
ployed) 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 conver-
sation 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 orga-
nization. (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 indi-
cate 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 queshn before y get into my topic.

(BC, tan, p. 62)
_____________
B: First of all, I wanna cngrachulachu cngeatchulatche on your progrm.
(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 maam.
B: And uhlet me jus wish you lots of good luck.
A: Thank you.
B: Uhm, the second thing, about the mayor an the governor, Ive listen to all
the things thet chuve said, an I agree with you so much. Now, I wanna ask
you something . . .
(BC, red, pp. 189190)
______________
B: Two quick comments,
A: Yes sir.
B: Dyou get the feelin thet weve 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 answerers hello by using hello alone (or some variant) as their first utterance.
#197
Police make call.
A(woman): hello
D: Yeah.
A: You mean yorre not busy/
D: Oh yeah, but I thought Id call and let you know what we know
so far.
After police desk such utterances are heard not as voice-recognition-relevant, but as con-
fusions of the summons-answer structure, and warrant a repeat of the answer.
#365
D: Police desk.
C: Hello.
D: Police desk.
. Callers 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: Police desk.
C: Hello/
D: Hello.
C: I come down here to see my wife jill, Im from Missouri?
D: Yes.
C: And uh they aint down here now, they moved.
(Pause)
D: 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 callers failure there to co-select the parts of his talk so as to formulate business-
for-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. 234244). New York: Harper and Row.
Ervin-Tripp, Susan (1969). Sociolinguistic Rules of Address. In L. Berkowitz (Ed.),
Advances in Experimental Social Psychology (pp. 93107).
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), 133136.
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, 2263.
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. 3174).
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. 325345). 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. 5780). 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, 26136.
Scheflen, Albert E. (1964). The Significance of Posture in Communication Systems.
Psychiatry, 27, 316331.
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
Schegloff, Emanuel A. (1979). Identification and Recognition in Telephone Openings. In
G. Psathas (Ed.), Everyday Language: Studies in Ethnomethodology (pp. 2378). New
York: Erlbaum.
Schegloff, Emanuel A. (1986). The Routine as Achievement. Human Studies, 9, 111151.
Schegloff, Emanuel A. (1993). Telephone Conversation. In R. E. Asher (Ed.), Encyclopedia
of Language and Linguistic, Volume 9 (pp. 45474549). 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. 326385). Cambridge: Cambridge University Press.
Sommer, Robert (1959). Studies in Personal Space. Sociometry, 23, 247260.
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 clerks investiga-
tive work and describe some devices or methods that she used in carrying out
the investigations. I will argue that one of the clerks 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 clerks 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 go-
ing 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 ab-
sences 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 assump-
tion 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
clerks 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 sus-
pected 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 clerks 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 fundamen-
tal 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 ori-
ented 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 con-
cerning 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 ques-
tion: 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
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 enumer-
ated 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 days absence (Was Bryan home from school ill today?). In enumer-
ating 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 oclock? 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 re-
ported 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 re-
ported absences in such a way as to perhaps guess what was likely as to the
extent of a students 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 cir-
cumstances.
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 tda:[y?
Mother: [Ah::: yes he was in fact, Im
sorry I I didnt ca:ll because uh: I slept in late I (.) havent been feel-
ing 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 donknow yknow if hl be:,h in tihmorrow fer sure er no:t

hes kinev js been layin arou:-ndj(h)ih know[:
Clerk: [O:kay well Ill goheadn::
uh::m I wont call you tmorrow ni:ght if: we dont see im tmorrow will
jst assume he wz home i:ll.[.hhhh
Mother: [.tch ih-Ri:gh[t cz ee-
Clerk: [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 legit-
imate or illegitimate. The clerks job was just to collect the facts; the mothers
job was just to report them. The clerks 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
Marks absence legitimate?, these two versions are neither equivalent nor inter-
changeable. The clerk did not ask the mother Was Marks absence excused?
and the mother did not assert Marks 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 clerks 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 in-
corporated 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 doctors 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 legit-
imate reason, to affirm an alternative legitimate reason for the absence, or to
claim no knowledge of a legitimate reason.
Later in the analysis, I hope to show that the clerks 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 Marks being home ill, the
mother apologized and accounted for not having called the attendance office
to report his absence.
Mother: [Ah::: yes he was in fact,
Im sorry I I didnt ca:ll because uh: I slept in late I (.) havent been
feeling well either, . . .
In apologizing and accounting for her failure to call, she displayed her knowl-
edge that calling was the proper protocol for excused absences. Perhaps more
interesting is the mothers inclusion of in fact when she confirmed that her
son was home ill. Including a claim of facticity is an attempt to bolster the fac-
ticity 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 straight-
forward. The clerk sought information that would account for the students
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 Marks 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 tem-
plate: 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 clerks initial inquiry, the mother did suspect
truancy. I will examine the mothers conduct while she apparently suspected
truancy and the clerks responses to her conduct.
When the mother was asked whether her son was home ill, she discon-
firmed 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 wasnt
In a way it is curious that the mother turned to check with the other person
before answering the clerks 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 cer-
tainty 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 discon-
firming accounts for a legitimate absence was relevant to her doubled-checking
before answering. When members of society recognize a matter as consequen-
tial, 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 re-
porting that her son wasnt home that day. Inasmuch as she knew the import
and consequentiality of that report, she elected to make certain of it before

speaking out.
It is conceivable that the clerk, based on the mothers 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 mothers report that Bryan had not been home
ill, she reported on the record of Bryans absences. In describing the record
of Bryans absences, she described the events (absences) for which she sought
an account.
((lines deleted))
Mother: No he wasnt
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 Bryans ab-
sences 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 endors-
ing the reported absences as actual absences. The clerk sometimes talked about
absences and sometimes talked about reported absences. Reported ab-
sences 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 clerks fact-seeking
inquiry, the mother reported facts that were damaging to her son. Prior to giv-
ing 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 Bryans 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 mothers clearly having a conclusion and
their understanding that they were to merely exchange information, not judge
the sons absence at this time. I will track the conversational moves that seem
to exhibit this tension.
Anita Pomerantz
Clerk: .hhh Well he was reported absent from his thir:d and his fifth
period cla:sses today.
Mother: Uh huh
Clerk: .hh A:n we need im to come inna the office in the morning to clear
this up.
(1.0)
Mother: .hh O:kay Ill tellim that, [(uhm/and)
Clerk: [O:ka::y
Mother: He has no excuse!
Clerk: Uh huh=
Mother: =as far as I know,
The marked intonation of the mothers acknowledgment Uh huh conveyed a
message something like: I fully understand. Rather than explicitly addressing
the mothers 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 clerks instructions on handling the prob-
lematic absence (Okay Ill 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 def-
inition he was truant. So while the mother did not literally say that Bryan was
truant, she essentially said it.
The clerks response to the mothers condemnation was a very bland ac-
knowledgement. Blandly acknowledging the prior was a way of not engaging
further with it.
With no gap after the clerks 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 un-
certain 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.
The mothers 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 sons guilt. Her recommendation
regarding just punishment for her son presupposed his guilt:
Clerk: Uh huh=
Clerk: Okay,
Mother: A:nd uh
(1.0)
Mother: Maybe itll 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 cn)
The mother presented her idea for just punishment packaged as a tentative
suggestion (A:nd uh, (1.0) Maybe itll be a good idea...)
Consistent with the way the clerk dealt with the mothers previous judg-
ment, the clerk again refrained from engaging in the discussion on the mothers
terms. The clerk responded to the mothers suggestion of punishment with an
acknowledgment (Okay fine) followed by a condition. Using the conditional
form if , the clerk transformed the mothers assumption of guilt into a matter
not yet established. Moreover, rather than referring directly to Bryans guilt
as the matter not yet established, she referred to his not bringing in a note as the
matter not yet established (if he cant bring in a note from you). In this way,
the clerk maintained a focus on the attendance offices 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 mothers discussion
of punishment was premature. Punishment would be relevant only at a later
stage and only if certain contingencies occurred. The clerks maintaining the
stance that Bryans 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 mothers contribu-
tions, moving forward in the interaction, while not validating or endorsing the
mothers 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 clerks 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 clerks 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 ab-
sences sin:ce: u-oh:: las:t Thursday, .hhh Shes been reported absent
(0.2) .t all day last Thursda:y,
(.)
Mother: Uh hu:h, well she hasnt 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 seek-
ing 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 stu-
dent 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 daughters 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 ab-
sences sin:ce: u-oh:: las:t Thursday, .hhh Shes 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 clerks formulation couple
of absences was a solution to possibly inaccurate attendance records. A cou-
ple 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 ac-
countable for the accuracy of a specified number. After giving the overview
of absences, the clerk started to enumerate the reported absences (Shes 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 of-
fers 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-interactants version be-
comes 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 neu-
tral, as simply gathering information, if a legitimate excuse was incorporated
into the question.
The clerks second method consisted of informing the mother of the ab-
sences 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 speakers 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 clerks 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 (Pomer-
antz 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 Michelles absence
as a matter not yet determined and in need further investigation.
In starting her inquiry by informing the mother of her daughters 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 mothers 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 Shes been reported absent (0.2)
.t all day last Thursda:y,
(.)
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 clerks and the mothers conduct suggests that they proba-
bly 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 offices evidence and let the mother account for the absences.
The mother avoided explicitly judging her daughter as truant, as having no le-
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 Michelles absences were all day ab-
sences, 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 lets see it looks like fir:st second
(0.3) third and fourth period for last Thursday and Friday .hhh and
heres sixth period so its (.) e- (.) wed have to assume that its an
all day a:bsence ye:s.
The clerks enumeration of reported absences in response to the mothers in-
quiry 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 Wed 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 ones claims, for me the weirdest line in
the call was the clerks And uh...You dont know that shes been home ill, huh?
Clerk: so its (.) e- (.) wed have to assume that its an all day a:bsence ye:s.=
Mother: =M-h[m::,
Clerk: [.hhhh And uh:::, .hhhh.t You dont kno:w that shes been
home ill hu:h,
Mother: n:No:: not to my knowledge I (.) Im (.) 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 daughters absences, the
mother addressed the reason for the absences: Well she hasnt been home i:ll.
The clerks response was to overview the extent of the absences and they had
several turns on that topic. To move on with the calls business, the clerk needed
to return to the mothers addressing the reason for the absence. Inasmuch as
Anita Pomerantz
the mother had asserted, Well she hasnt been home ill,, the clerk might have
returned to the topic with a request for confirmation: And uh she hasnt been
home ill huh? The clerk did request a confirmation, but the form she used was
peculiar: And uh:::, .hhhh.t You dont kno:w that shes 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 clerks
question exhibited a carefulness with respect to claims in that she asked only
about the mothers 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
mothers 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 clerks 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 daughters 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 dont kno:w that shes been home ill hu:h,
(0.3)
In the context created by the clerks 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 mothers 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 mothers report and then informed the mother of the appropriate
steps to take.
Clerk: .hhhh And uh:::, .hhhh.t You dont kno:w that shes been home ill
hu:h,
(0.3)
Clerk: Yea:h. Well, u::h the- obviously shes not going to her cla:sse:s an
we need tuh find out where she is: going would you: talk ther about
it tonightn .hhh send her into the attendance office tomorrow
morning so that we cn straighten these out.
Mother: Mmmkay:: Ill see what she says about it.
With the mothers 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 shes 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 con-
cluded that there was a problematic circumstance and that further investigation
was required. Concluding that a problem existed that needed further investi-
gation 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 shes not going to her cla:sse:s an we need tuh find
out where she is: going would you: talk ther about it tonight n .hhh send her
into the attendance office tomorrow morning so that we cn 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 clerks talk of
punishment. In raising the topic of the detention system, the clerk implied that
truancy was a possibility.
Anita Pomerantz
Clerk: ...send her into the attendance office tomorrow morning so that
we cn straighten these out.
Mother: Mmmkay[:: Ill see what she says about it.
Clerk: [Uh::: we have a new-
Clerk: .hhh Okay:. We have a new uh:: detention system now
(.)
Clerk: that if they don clear the:se, theyll become truants.
(.)
Clerk: .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, shell 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 Michelles absences. The
clerk then returned to speak about Michelles 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 cit-
ing the general policy, the conditions under which students become truant and
go to detention. When speaking about Michelles 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 inform-
ing the mother of a new school policy was a solution to the clerks 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
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 pos-
sibly 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 insti-
tutional 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. 2123, 380381) discussion of the Correction Invitation Device.
References
Pomerantz, Anita (1980). Telling my side: Limited access as a fishing device. Sociological
Inquiry, 501 (34), 186198.
Pomerantz, Anita (1988). Offering a Candidate: An information seeking strategy.
Communication Monographs, 55, 360373.
Sacks, Harvey (1992). Lectures on Conversation. Vol. I. Oxford, UK: Basil Blackwell.
Anita Pomerantz
Appendix
Call 1
Mother: Hello.
(0.5)
Clerk: Hello Mister Johnson?
(0.8)
Mother: Uh: this is Missus Johnson,
Clerk: Uh Missus Johnson Im sorry.=This is Miss Devalo from Alvera High
School calling?
Mother: Mmhm:,
Mother: [Ah::: yes he was in fact, Im
sorry I I didnt ca:ll because uh: I slept in late I (.) havent been feel-
ing well either, .hhhh and uh: .hh (0.4) he have a myihkno:w a:, fever::
(.) this mornin[g
Clerk: [Uh huh, .hhh
Mother: A::nd uh I donknow yknow if hl be:,h in tihmorrow fer sure er no:t
hes kinev js been layin arou:-ndj(h)ih know[:
Clerk: [O:kay well Ill goheadn::
uh::m I wont call you tmorrow ni:ght if: we dont see im tmorrow will
jst assume he wz home i:ll.[.hhhh
Mother: [.tch ih-Ri:gh[t cz ee-
Clerk: [A n : - send a note with him
when he does return.
Mother: Mm [hm
Clerk: [O:kay.
Mother: Okay=
Clerk: =Thank you
Mother: Uh huh Bye [bye
Clerk: [Bbye
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?
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 Shes been reported absent (0.2) .t all
day last Thursda:y,
(.)
(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 lets see it looks like fir:st second (0.3)
third and fourth period for last Thursday and Friday .hhh and heres
sixth period so its (.) e- (.) wed have to assume that its an all day
a:bsence ye:s.=
Mother: =M-h[m::,
Clerk: [.hhhh And uh:::, .hhhh.t You dont kno:w that shes been home
ill hu:h,
(0.3)
Clerk: Yea:h. Well, u::h the- obviously shes not going to her cla:sse:s an we
need tuh find out where she is: going would you: talk ther about it
tonightn .hhh send her into the attendance office tomorrow morning
so that we cn straighten these out.
Mother: Mmmkay[:: Ill see what she says about it.
Clerk: [Uh::: we have a new-
Clerk: .hhh Okay:. We have a new uh:: detention system now
(.)
Clerk: that if they don clear the:se, theyll 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 tyou r (.) jus sending her is that enough.
Clerk: .hhh Well if you cn excu:se any of these with a note saying yes shes been
home ill er at the doctors or whatever .hhhh uh:: (.) just send a note but
othe[rwise you don need tuh come in
Mother: [(Yeah)
Anita Pomerantz
Mother: Okay then.

Clerk: Mka[y
Mother: [Ill talk tuh her about it en (.) well get back tuh you.
Clerk: O:kay
Mother: A[lright
Clerk: [Thank you
Mother: Mbbye
Clerk: Bye bye
Call 5
Male: Hello?
Clerk: Hello, may I speak to Misses: Forde?
Male: eYeah hol on.
(10.0)
Mother: Hi::
Clerk: Hello this is Miss Devalo from Alvera High School calling,
Mother: Uh huh,
(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 wasnt
Clerk: .hhh Well he was reported absent from his thir:d and his fifth period
cla:sses today.
Mother: Uh huh
Clerk: .hh A:n we need im to come inna the office in the morning to clear this
up.
(1.0)
Mother: .hh O:kay Ill tellim that, [(uhm/and)
Clerk: [O:ka::y
Clerk: Uh huh=
Clerk: Okay,
Mother: A:nd uh
(1.0)
Mother: Maybe itll 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 cn)

Mother: [Oh waitta minute e- waitta
minute
Clerk: Okay,
Mother: .hh ah he hadda a do:ctors appointment.
Clerk: [[Ah hah
Mother: [[Thats 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 Ill getta- givem thexcuse.
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: 671672).
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 Im 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 ut-
terances as Are you kidding?, Are you serious?, Youre 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, talk-
ing 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 men-
tioned, 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 news-
paper clippings, excerpts from books, etc., all lumped together under the gen-
eral 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 didnt want him to think
I was taking him seriously. He said Well Joey run down to the police
station before I do something I dont want to do. . . . I says Joey run
outside. Daddys only kidding.
(2) [New York Times, November 20, 1964]
Two women ran into a Bronx drugstore yesterday morning and one re-
marked: 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, youre 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 snipers 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 dont
kid me. Thomas then saw blood trickling from his companions 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 shoot-
ing. 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 Kafkas The Trial, pages 78]
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 mens
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 Ka-
van, 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 . . . youve heard those expressions, and
then I saw I dont 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 its 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
snipers 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 Pres-
idential 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
dont fit under the heading one has set oneself to collect on some particu-
lar data-run, but which seem to be related; ballpark phenomena that might
cast some light on the focal phenomenon and/or point to independently in-
teresting 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 ap-
pears 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: 787788).
Verbs seem to be one routine area for doing such a thing as showing an in-
tention of the truth of some statement. I have a case in mind, and although
its the only case I happen to have, I dont think its peculiar. Its 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, re-
turned 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, whats 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 wouldnt get I
realize the car has been taken by the police. What we would likely get is,
I (guess, bet, wonder if) the cars been taken by the police, or Maybe
the cars 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 there-
after 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 didnt 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
Coroners 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 al-

ternation 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 sub-
ject 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 superstars asserted reaction to the roar of a
crowd (as something he was due and accustomed to) and a novice actresss as-
serted 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 dont
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 ob-
jects, 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 (19691970, 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).
Its really remarkable to see peoples efforts to achieve the nothing happened
sense of really catastrophic events. Ive been collecting fragments out of news-
papers, 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 stew-
ardess standing facing the cabin, and a fellow standing with a gun in her back.
And my first thought was hes showing her the gun, and then I realized that
couldnt 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 my-
self, we just had a Polish hijacking a month ago and theyre 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 imag-
ine 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 de-
stroy 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 Id 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 incan-
tation; a ritual used to manage, to put into normal perspective, something that
might otherwise be disruptive. As Sacks has it, to achieve the nothing hap-
pened 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 childrens 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 Ive been hit, or I think Ive been shot. Whatever he
said, he wasnt sure. . . . I thought hed been hit by a rock or something. . .
And Id 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 peoples 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, Any-
bodys 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 Im proposing to be the in-principle-correctness work of these media
objects, Ill 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.
Ill start off with a fragment in which someone produces the same
thought for two consecutive participants; a little girl, and then the little girls
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 isnt home, and Sharon finds herself speaking to Stephanies mother,
Fran, and eventually handing Fran over to her father (and Frans 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; Sharons father is now on the line))
.
Fran: Well I thought you werent going down til next seh-u-the
weekeh:-I mean the end of the mo:nth.
(0.4)
Ted: No:, were 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 Frans 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 its too bad shes not home, cause
At first I thought
shed 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 Frans
(repeated) reference to her wrong notion is somehow searching for an ac-
count, 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 Frans proffered wrong thought denies it; Sharon with Mm-
mm, and Ted with a more elaborate No:, were down here for: two weeks,.
That is, each recipient of the proffered wrong thought treats it as here-and-
now operating on an assertion of fact, rather than, e.g., an interesting, amusing,
puzzling commentary on the recipients 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, theres an NHR program.
Leslie: Yes its for anybody whos not got one.
Joyce: Oh. I thought perhaps youd left yours in there inadvertently.
Leslie: No. Shes 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: Michaels in the midst of moving this weekend.
Geri: Thought it was last weekend.
Sheila: No::, he had some: complications.h
Gail Jefferson
(18d) [TCI(b):16:8-9]
Alice: Well Stevens hairs the same color as Crai:gs,
Fran: Is it?
Alice: Yeh
Fran: I thought Craigs was li:ghter.=
Alice: No I dont think so Craigs hair isn:t
(18e) [Gold:3]
Jane But I couldnt arrange that becau:se Thomas is coming again.
Reva: I thought Thomas was going away to Ohio. To schoo:l.
Jane: No: thats 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. Ill 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) [Frankel:GS:X]
Alan: Still growing. Its 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 dont see any bugs. It might have, but I cant 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: theres over a hundred thou:sand according to this article.
At first I thought
(18i) [SBL:3:2:R:5]
Claire: Well theyre not comi:ng, unless they can fi:nd someb[ody.
Sara: [Oh=
Sara: =I thought they were coming.
Claire: No theyre 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 didnt 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. [Thats C[orona.
Bill: [That- [Thats Corona del Mar
Pat: Oh is it?
Bill: Right where it hits the Coast Highway
Pat: Oh I didnt 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 mustv been taking a
nap I rang the be:ll and then I ca:lled you: later in the eveni:ng,
(0.7)
Emma: I dont know where you were m[ay-
Gladys: [Oh: Ill tell you I heard the
pho:ne I was watching television by the time I got out here itd
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 dont
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-cant get going hhuh-hhuh
(18n) [Frankel:TCI:1:8]
Sheila: thhh 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: Well just have the two tables unless shes fou:nd uh
another cou:pl[e or]
Sara: [Yeh.] Well-
Sara: Oh you I thought you said you were gonna have your
neigh:bors.
Claire: tkhhh u.-No they couldnt 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 its at- on three oclock and she might just be free or
between interviews.
(1.0)
Mr. O: w-What time is it now sir?
Desk: Three isnt it?
(0.7)
Mr. O: (We:ll?) I thought it was earlier than tha:t,
(0.3)
Desk: Its two oclock Im 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 oper-
ating 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 what-
actually-happened is shown to be odd, surprising, exceptional; to be in-
principle 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 some-
thing else happens, that doesnt provide for a shift. One doesnt now say Well,
maybe X isnt 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 clip-

pings and hastily scribbled notes from news broadcasts, Ive come across two
lovely instances of At first I thought being used explicitly to propose things
like this dont 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 miners 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 isnt 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 didnt want to believe that such a thing can happen in The Netherlands,
said the Minister.
In these two cases were seeing the work of at first I thought at its most ex-
plicit not to mention grandiose, if not just plain elephantine: Things like this
dont 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 as-
sert 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 han-
dled firearms for fifty years, and thought immediately that it was a rifle
shot.
Note that this witness marks his perception as specifically not Anybodys 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 ordi-
nary (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, expect-
ing 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, thats 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 dont 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 catas-
trophic 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 immediatelys and the only thought that crossed my mind con-
vey that these are not Anybodys Proper First Thoughts. You get a sense that
they came unbidden into his mind; that Connally himself is surprised at hav-
ing 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 Nether-
lands]
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 youre 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 Anybodys 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 Presidents 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 Zapruders 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 inci-
dent))
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 caf 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 didnt believe it at first, but there you are.
(36) [de Volkskrant, February 12, 1992] ((the Henglo F16 crash))
I didnt 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 Voorne-
Putten. 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 immediacy-
marked 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,
dont 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, Anybodys pro-
cedures 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, extraordi-
nary 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 or-
dinary 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 them-
selves 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 immedi-
acy 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 mans 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 hes 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 Menningers
Mortal Error (1992: Appendix A, pp. 297375.). 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 do-
ing 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 Presidents 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 nave. Or maybe what theyve
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? Theyre just like you and me.
Another SS-mans statement is produced with a bit more professional dis-
tance:
(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 Kellermans dis-
engaged 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 follow-
car))
At about 12:30 p.m. we began the approach to the Thornton Freeway trav-
eling about 2025 mph in a slight incline. I was about 2530 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 pro-
ceeded 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 Presi-
dent...
(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 follow-
car]. 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, Im 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. Hickeys ...and seemed to me to be at ground
level in fragment 48). Its 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 Anybodys spontaneous I thought it was X, and
Kellermans 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: 455457).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 Menningers that Sacks refers to in one of his
earliest lectures (Sacks 1992, Vol. I: 66): loyalty to reality. In his 1971 lec-
ture, Sacks speaks of the naming of the rejected alternative as showing ones
commitment to the normal.
Its a lovely phrase, and its obviously relevant to the UFO-spotters 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 Ive 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 dont 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 pro-
vides that what happened here is distinctly unusual, plus that unusual, odd
events dont count. With those two notions we can come to see, technically,
what can be involved in Sacks 1970 proposal that one aspect of being ordi-
nary 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 dont 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 Any-
bodys 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]
Yesterdays 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 worlds 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 couldnt 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 decades 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 brothers,
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 didnt 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 Nether-
lands, 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 in-
deed, we can find people accounting for their wrong first thought having been
of an earthquake by announcing that theyre 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 su-
perceded 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 stum-
bled 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 suf-
fered 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 capitols 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, dont move. The area was im-
mediately 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. Hes 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, hes 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 ig-
nored them. Hed hardly turned the corner when something happened
that he describes as a gray, yellow light, it was an implosion.
...
It didnt occur to me that it was a bomb. I thought that the garage had
done something wrong. The car hadnt 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 win-
dows 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. Im 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 Sollis-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 Solis-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. Im 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, thats what is in fact going on.
(60) [Friesch Dagblad, June 29, 1996]
Theyre 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 barri-
cades again near the former front line? Cars were being stopped by police
agents.
Residents of Sarajevo look around them, concerned and wondering.
Theyre 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
[its] 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 oc-
curred, come off as perhaps just a bit crankish in the newspaper article which
reports those thoughts. In this particular article, the bomb possibility (ar-
row 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
evenings 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, didnt
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 creden-
tials 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 caf 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 singers collapse was part of the performance, while in frag-
ment (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 incomprehen-
sion. 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 oclock in the
afternoon when he heard a series of loud reports.
I thought at first that someone was throwing firecrackers, declared a visi-
bly 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 com-
mander, 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 ab-
sence. 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 hed 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, Dont
let it bother you. Those buildings cant 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 37-
year 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 hadnt 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 schools 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 [Fatoushs] friend Mimoush Handi.
And so on...
Notes
* Between 1970 and 1995 I presented versions of this paper any number of times. Ive 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 didnt 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 policemans 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 couldnt 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 Artscrolls 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
(1980:186198). Roughly, such things as accounts can be elicited from a coparticipant, by
reporting ones own experience of the coparticipants 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: 2123, 380381), 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 ex-
plain 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 didnt hear any
sound. Then I realized it was no aircraft. For Hynek, this process of going from the sim-
ple, 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 Hyneks 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 recur-
rent 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 real-
ization (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 Hyneks typical statement, with its series of
escalating possibilities fragment 9 (a multi-stepped report by a witness of an air disas-
ter; 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 Hyneks UFO reporters of high caliber, offers, with his I thought it was an air-
craft 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 its 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 im-
mediately 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, its 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 guys car, to a guest-house in back
of the girls 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. Martins.
At first I thought
Pomerantz, Anita (1980). Telling my side: limited access as a fishing device. In Sociol-
ogical Inquiry, 50 (34).
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 presenta-
tion 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 investi-
gated 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 organi-
zation 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 se-
ries 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 prop-
erty 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
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 two-
turn pairs in which the turns are adjacently placed and produced by alternate
speakers. The occurrence of an Adjacency Pair first pair part makes the occur-
rence 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 non-
occurrence 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 evad-
ing, 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 non-
production 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. Se-
quence organization does make this distinction. Reports on sequences con-
cerned 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 As-
sessments 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 theres a good chance that they might be hiring her.
2 B: Oh! [Thats] 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 A: Informing (Adjacency Pair First Pair Part)

2 B: Assessment (Adjacency Pair Second Pair Part)
It appears then that the introduction of information as new for the conver-
sation is followed by its subsequent treatment as now given and that this
marking can intersect with sequential structures in the talk. We propose to un-
dertake 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 Announce-

ments, in inquiries, in requests for news, as stories, etc. With the term An-
nouncements we have in mind those conversational events in which occur-
rences 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 An-
nouncements 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: So, Elizabethn Will were spoze tuh come down lasnight
2 but [there was death n the famly] so they couldn
3 come so Guys asked Dan tuh play with the compny deal,
4 so I guess he cn play withim. So,
5 A: Oh good.
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 assess-
able news. Our suggestion is that a major factor in the recognition of An-
nouncements 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 occur-
rences 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 Announce-
ment independent of whether it is known that the recipients have not heard.
Our argument is that the recognition of some utterance as a possible An-
nouncement 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 sen-
tential 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 or-
ganizes 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 havent 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 De-
liverer and for the Recipient regularly occur in the turn directly following the
Announcement. They operate there to display the Recipients 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 compo-
nents. Following is an instance of such an Announcement Sequence with the
components labeled:
(5) [JG 2:1]
Pre-announcement 1st pair pt. D: I forgot ttell ythe two best things thet
happentuh me tday.
Pre-announcement 2nd pair pt. R: Oh super=What were they.
Announcement #1 D: I gotta B plus on my math test.
Insert Seq. 1st pr. pt. R: On yer final?
Insert Seq. 2nd pr. pt. D: Uh huh?
Assessment of #1 R: Oh thats wonderful!
Announcement #2 D: hh And I got athletic award.
Insert Seq. 1st pr. pt. R: REALLY?!?
Insert Seq. 2nd pr. pt. D: Uh huh. From Sports Club.
Assessment of #2 R: Oh thats terrific Ronald.
These findings have been formulated from an initial examination of a sam-

ple 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. Announcements in conversation occupy turn series.
2. The series of turns are adjacently placed.
3. The series are organized into adjacency pairs.
4. 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 rel-
evant second pair parts.
We are arguing that sentences occurring in natural conversation which are oth-
erwise linguistically characterizable as informationally new are not isolated
occurrences but occur in the environment of other sequentially-related utter-
ances 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 ones 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 re-
ception 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 al-
ready 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 re-
tell 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 ma-
jor components of these Sequences: (1) the organization of Pre- sequences to
Announcements and (2) the design of the Announcement turn. How the pref-
erence 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 syntac-
tic and semantic concepts may be illuminating in a consideration of the design
of utterances in the Announcement turn, the selection of Announcement Se-
quences 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- sequences
Schegloff s (1968) characterization of Summons/Answer sequences has de-

tailed 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: =Im gunna buy a thermometer though
because I think shes // (got a temperature).
PRE-OFFER 1st A: We have a thermometer.
PRE-OFFER 2nd B: Yih do?
A: 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 se-
lection 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 con-
sequences 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 orga-
nized 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 discover-
ing in advance whether the production of the projected sequences First Pair
Part is warranted. If the response to the Pre-invitation First, for example, in-
dicates that the projected invitation will be refused, the invitation will not
be produced:
(11) [HS ST1:1]
PRE-INVITATION 1st A: Say whatr you doing.
PRE-INVITATION 2nd B: Well, were going out. Why.
REPORT OF INTENDED A: Oh, I was just gonna say
INVITATION16 come out and come over here
an talk this evening, but if youre
going out you cant very well do that.
In Announcement sequences, a category of Seconds alternative to Assessments
involves the marking of the Announcements news as already heard or as not
surprising. For example:
(12) [AT ST1:1]

R: How are ya?17
D: Oh I don know. I dont feel good. I had trouble with my stomach. I
had pains all day.
R: We all do.
(13) [J&G:24]
D: Jims, very sick.
R: Yea:h I went over theren 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 Adja-
cency 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: Whats the good news.
(16) [JSLR:38]
D: Hey I got sumpn thets wi::1d..
R: What.
As Pre-Announcement types they display features in common with other Pre-

sequences which have projective uses for the Sequence they preface. In par-
ticular they share components with another Pre- sequence type: Prefaces to
Stories told in conversation. Sacks (1974: 340) has discussed the sequentially-
relevant 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 ttell you
the two best things thet happentme [tday]
4. An offer or request to tell, e.g.: [Ywanna 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 fea-
tures 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 Deliverers 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. Deliverers 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 sequen-
tially 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
conversations history is a locally-accomplished production made up out of
known-in-common procedures for the delivery of possible news and its rati-
fication as in fact news. We argued further that the finding that the proposed
news is already known to the intended Recipients is a dispreferred alternative

and that that dispreference organizes various aspects of the Announcement se-
quence design. Our task here is to attempt to demonstrate the ways in which
that dispreference is manifested in sequentially-relevant features of the Pre-
Announcement 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) Deliverers methods for preserving the preference not
to re-tell, (2) Already Informed parties methods, and (3) Recipients methods
for the recognition of possibly known news.
(1) The design of Pre-Announcement Firsts is regularly aimed at determin-

ing in advance of an Announcement whether its news has been heard. Sudnow
(1967: 37) provides a particularly fitted instance of this concern in his discus-
sion 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 men-
tion does not noticeably inhibit ordinary conversation. When a death occurs
in an unexpected place within the hospital, or when deaths occur in rather un-
usual 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
mornings 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 do-
ing 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 Sudnows 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.
Deliverers concern not to be found to have told a piece of known news
may also manifest itself in an expansion of the Pre-Announcement Firsts 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 probly heard about
it. -already but just in case you havent.]
In this instance, Deliverers 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 Recipients state of knowledge about the news. In it, a
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 whats happening at the Fairgrounds then.
R: Uh, what? ((as in didnt hear))
O: Dyou know whats happening at the Fairgrounds then?
R: No
The preference not to tell known news manifests itself in an apparent pref-
erence 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 isnt that//terrible?
Ri: ((gasp))
Agreement D: Wasnt that tragic?
Pomerantz (1975: 13) notes that speakers 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 Recipi-
ents 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 items 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 An-
nouncement 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 Sum-
mons provides for the relevance of some business to come, business which will
be presented by the initiator of the Summons. Jefferson has argued that Sum-
mones 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 Sum-
moner 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 se-
quences. We have spoken of intending Deliverer and proposed Recipient in
Pre-Announcements as a set of identities local to and generated by the oc-
currence of the Pre-Announcement First. Another category of local identity
is possible: that of parties Already Informed (AI) of the news. In a multi-
party 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 consul-
tant.=[Some of you d-kno:w, n others dont.]
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--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 re-
serving the actual delivery of the news to the Deliverer.) They thereby mark the
relevancy of the Announcements occurrence for their next actions as not the
same as for Recipients. For example:
(25) [J&G:4]
PRE- 1st D: Youd probly turn green with envy if I tolja about
what I got ( ) first quarter.
PRE- 2nd R: Whadju get first //quarter?
COLLABORATIVE AI: Tell em about it.
ANNOUNCEMENT D: Four As in two Bs.
(26) [Adato:2,l5]
INQUIRY R: Whats new Jim,
COLLABORATIVE
PRE- 1st AI: Oh? Hes a little richer.
AGREEMENT BY D D: Yeh Im a lil richer, tss!
PRE- 2nd R: Richer.
(2.0)
R: Whaddiyuh mean.
ANNOUNCEMENT D: Think were gonna getta rai:se, first a nex month.
Collaborative work by Already Informeds 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 Informeds permit the preserva-
tion 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 athe- I dknow whether it it wz
Newseekr Timer what. I think ih wz Timer 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 isnt 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 lil 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 wont 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
(31) [KC 4:2] ((Note there are two Ds and two Rs))
D: How are you all?
R: Oh very- very well.
D: Good.
R: I know.
[[
R2: Whats 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 dn)
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 organizations
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: Ywanna 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 pre-
monitoring 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: Yknow your Grandpa Bills brother Dan?
R: He died.
D: Yeah.
(33) [DA:4]
D: I-I-I had something- [terrible] ttell you. So // uh
R: How terrible is it.
D: Oh, th- as worse it could be.
(0.8)
R: W- ymean Edna?
D: Uh yah.
R: What she do, die?
D: Mm:hm,
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 Recip-
ient. 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 prob-
ably 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 Pre-
Announcement Firsts which provide for Recipients recognition that are not
located in their design features but in their placement in the overall organiza-
tion of the conversation.
Parties to a conversation use and orient to the history of its topical progres-
sion. 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 in-
stances. 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 conver-
sation, 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 conversa-
tion.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. Cer-
tain 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 oppor-
tunity 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 lo-
cate 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-last-
met. For those who see one another regularly, the presentation of a Pre-
sequenced piece of news contains a built-in time reference. Locating the news
which the Pre-Announcement First references is thus bounded by a known-
in-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 in-
stances 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 Firsts 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 rec-
ognizable, 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
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 avoid-
ing 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 pref-
erence not to re-tell as it was manifested in the design and use of sequentially-
relevant 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 Re-
cipient, and AI for Already Informed.) Again they are taken from transcripts
of actual audio- or video-recorded conversations:
(34) [KC 4:2]
R: Whats the good news.
(35) [JSLR: 38]
D: Hey I got sumpn thets wi::ld.
R: What.
(36) [BS 2,1:73]
D: Oh guess what.
R: What.
(37) [J&G:2]
D: Ywanna know who I got stoned with a few w(h)eeks ago? hh!
R: Who.
(38) [JG 1:1]
D: I forgot ttell you the two best things thet happentme tday.
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: Yknow what he wannid me tuh do?
R: What.
(41) [JSLR:l31]
D: Hey youll never guess whacher Dad is lookih- is lookin at.
R: Whatre you lookin at.
(42) [J&G:3]
D: Yd probly 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)) -arent they,
regarding school.
D: Yeh.
R: Whats 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 something-
terrible ttell you, So// uh
R: How terrible is it.
(45) [HS ST:l]
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 An-
nouncement 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 Pre-
Announcement 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-
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- se-
quence 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 se-
quence (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: Ywanna know who I got stoned with last
wee(hh)k?
SOLICIT AS 2nd R: Who.
ANNOUNCEMENT D: Mary Carter n her boy(hh)frie(hh)nd!
(49) [BS 2,1:73]
PRE-ANN. 1st D: .hh Oh guess what.
SOLICIT R: What.
ANNOUNCEMENT D: 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 interrog-
ative 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 rele-
vant 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 ex-
pectable. 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 pro-
vide 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 imme-
diately 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, Sche-
gloff (1968: 1081) has formulated the notion of non-terminality to describe a
sequence property which relates turn sets to one another. Pre- sequences ap-
pear 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 di-
rected 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 Announce-
ment. What linguistic features of the Pre-Announcement First can be shown
to be sequentially relevant to the elicitation of the Announcement Solicit as
its proper and preferred - second pair part? And, how does the Recipients re-
sponse (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] thets wi::ld, I got two good
[things], I had [something] terible ttell 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 WH-
questions such as Yknow what he wannid me tdo? which takes the form: He
wannid me tdo something.31 In such instances, the pre-suppositional some-
thing 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., [some-
thing] 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 of-
ten the case relations) set up in the Pre-Announcement Firsts for the WH-
and 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: Yknow what he wannid me tdo?
SOLICIT R: 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 ttell you the [two best things] thet happen tme tday.
R: Oh super, what were [they].
(52) [J&G:3]
D: Yd probly 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: Ywanna know [who] I got stoned with a few weeks ago?
R: [Who]
c. And, the Solicit turn is frequently solely occupied with the single WH-
interrogative but when more than that single item occurs, the additional ut-
terance 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 youll never guess [whatcher Dad ((the speaker)) is lookin at.]
R: [Whatre you looking at.]
(55) [J&G:3]
D: Youd probly 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 no-
tion 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-
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 An-
nouncement turn. Furthermore, what can stand as a replacement or spec-
ification of the pro-termed items is constrained by the design of the Pre-
Announcement First in syntactic and possibly case terms.
We have so far considered the operation of WH-marked items in the Pre-
Announcement. 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 ttell you the [two best things] thet happentme tday.
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 WH-
marked items.
The occurrence of something in the Pre-Announcement: I got sumpn
thets 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 ttell you.
:
((Edna died))
Alene Kiku Terasaki
(58) [JSLR:41]
D: Toni en Bob I have something tuh tell you...
:
You must not use- anyv the pottery you picked up der any-tuh cook
in.
(59) [JSLR:153]
D: Say Tom, I wanna tell you sumpn
:
...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 some-
thing 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-Announce-
ment type which has sequential implications allows us to expand our anal-
ysis 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 pro-
jected 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 sequen-
tial 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; declara-
tives: I got two things), the co-occurrence of a specifiable set of linguistic and
sequential features identifies them as instances of the same thing conversa-
tionally. Recipients 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
items as pro-forms for forwardly referenced tellables or sayables in the con-

tingently 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 tday, (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 WH-
question which occupies that turn selects the same WH- item of the Pre- or its
syntactically corresponding Pro-form. It is a members 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: Yknow what he wan-
nid me tdo? 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 oc-
currence 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 sequentially-
implicative pro-forms and verbs which in part account for recipients recogni-
tion of an item as a Pre-Announcement First. While there are other features
which attend its recognition (such as placement), that formulation still ap-
pears 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 analysts for-
mulation is as well a members formulational problem. That is, the recognition,
in this instance, of whether the item is in fact a Question or in fact a Pre-
Announcement 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: I know where youre going.
2 Mom: Where.
3 Kid: To: that (meeting ).
4 Mom: Right. Yah!
5 Do you know whos going to that meeting?
6 Kid: Who
7 Mom: I dont know!
8 Kid: Ouh:: probly: Mr. Murphy an Dad said probly 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 re-
turn 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 in-
stead 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
ones 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 possi-
ble 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: yknow what he wannid me tdo? 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 ques-
tions 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.
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 syntactically-
fitted 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 cant 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: Whats black n white n hides in caves,
(0.4)
R: [Aright I give up,]
Whats black n white n hides in-
(65) [HS ST:1]
R: [No,]
What.
Our analysis of the linguistic and sequentially relevant features of the Pre-
Announcement Firsts 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 deter-
minable 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 gram-
matical 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 se-
quences, 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 Pre-
Announcement First directed to the elicitation of the Announcement Solicit
as its proper and preferred Second Pair Part. The responsiveness of the An-
nouncement 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 Pre-
Announcement 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 discus-
sion will also take us to a consideration of the Pre-Announcement sequences
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 prod
constituent of the Pre- in question form, e.g., Hey youll never guess [whatcher
Dad is lookin at] [Whatr you lookin at.] In this respect, the forms overlap
with three forms of WH-questions which occur regularly in conversation:
1. Full sentential WH-questions: Whatr you lookin at.

Where are you going?
2. One-word WH-questions: Who.
What.
3. Partial WH-questions or On what.
Appendor questions:37 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--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 ques-
tions 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]
A: He had an operation.
APPENDOR Q B: On what.
Answer A: 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 (NTRIs):
(67) [TG:26]
A: Sibbies sistuh hadda baby bowa:::y.
NTRI B: Who?
Repair A: Sibbies 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 ut-
terance which contains matched constituents, here the NP, Sibbies 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 place-
ment 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 Jeffersons (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 WH-
question/Answer sequence.
One-word WH-questions show up as Solicits in response to Pre-Announce-
ment 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 sug-
gests 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 ques-
tions 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 un-
derstanding of some prior utterance is by producing its sequentially relevant
next action. When that next action coincides with a return to otherwise in-
tuitively 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 (Ywanna 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 Pre-
Announcement First marks its subsequence to the Pre- while still providing
for the strong forwarding of the larger Sequence through its appearance as
a Question.
Note that we are not here remarking on an analysts finding only. If a con-
cern 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: Ywanna 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 ttell yih the two best things then happentuh me tday.
D: I got a B+ on my math test.....and I got athletic award.
(71) [JSLR:13l]
D: Hey youll never guess whatcher Dad is lookih- is lookin at.
:
R: Whatre you looking at.
D: A Radar Range.
Alene Kiku Terasaki
(72) [JSLR:38]
D: Hey I got sumpn thets wi::ld.
R: What.
D: Yknow one a these great big red fire alarm boxes thetr on the
corners? I got one.
(73) [J&G:3]
quarter.
R: Whaddju get first quarter.
D: Four As n two Bs.
(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)) -arent they,
regarding school
D: Yeh.
R: Whats hapning.
D: I got kicked outta the University.
The occurrence of the Announcement utterance as a response to the WH-

Solicit 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
turns 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 WH-
marked constituent is preserved in the replacement. For example:
(75) [J&G:3]
quarter.
R: Whadju get first quarter.
D: Four As n two Bs.
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:
(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 in-
terrogative WH- pro-forms are replaced by phrasal constituents while the items
prod 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 (19671968: 207208) suggests that where infor-
mation 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 in-
stances 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 se-
quential 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 WH-
Solicits.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 infor-
mational 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 19671968: 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 mark-
ing 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 se-
quentially. 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 charac-
terizable 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 appear-
ance 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 deliv-
ery 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 occur-
rence 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 se-
ries 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 sug-
gested 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
delivery an interactionally-accomplished matter, assembled across the occur-

rence 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 place-
ment 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 ap-

propriate 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 ut-
terances 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 hearers determination of whether a current speakers ut-
terance 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 hearers analysis of a prior utterance di-
rected 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, appo-
sitional particles, tag questions and the like. Consequently, our work has been
directed to the development of an observational methodology designed to cap-
ture the particulars of situated speech. Our use of audio- and video-recordings
of naturally-occurring conversation in conjunction with detailed transcrip-
tions 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]; Jef-
ferson 1972, 1975), and idiomatic or sentence fragment utterances (Davidson
1975; Schegloff 1968; Goldberg 1975) demonstrate the import of these proce-
dures 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 under-
standability 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 intu-
itive reconstruction of situated speech may omit phenomena which turn out
to be central to speakers understanding of an utterance in terms of what it re-
veals 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 im-
port 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 en-
counter 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 ap-
proach to the question of the verifiability of our findings. In the production of
a matched second pair part to a prior turn, speakers show their understand-
ing of that turn. Sacks et al. (1974: 728729) 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 turns talk, their
understanding of other turns talk. More generally, a turns talk will be heard
as directed to a prior turns 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 turns 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.
Speakers 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 co-
participants. 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 deter-
mines an items informational status. Chafe (1974: 119122) discusses the role of environ-
mental 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 inter-
actionally ratified by the recipient as news-to-them. The factual character of the items 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 reveal-
ing they have been privy to a secret.) In contrast to an interest in the factual character of
the recipients 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 interac-
tants treat as prescriptive of next actions. It is not, then, coterminous with Hymes (1974) use
of speech event nor with Searles (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: How are ya.
B: Oh I don know. I dont feel good. I had trouble with my stomach. Ive had pains
all day.
A: We all do.
B: 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,
Its cold in here.
. For additional instances of such internal expansions, see Schegloff (1972) and Jefferson
(1972).
. The utterances 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 Whats new. as well as Topic-Initial
Questions like Hows 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 Se-
quences contain Pre-announcements. Our interest here is to examine one type of Pre-
announcement which does occur. A second type of Pre-announcement appears to be dif-
ferentiated 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, isnt it.
R: Yea:h. One a those mornings,
K: (Oh yeah.)
D: Yeah its gonna be one a those mornings.
R: Whats 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 Deliverers 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. De-
liverers concern to mark news as possibly already heard certainly attends other concerns as
well. Among these is the desire to mark ones place in delivery networks vis--vis intended
Recipients. The taking or making of an opportunity to personally deliver news of signifi-
cant events in ones 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 didnt 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 speak-
ers 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 Fillmores (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 moti-

vates 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)
Im 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.
. For a discussion of childrens use of this sequence item as a floor seeker, see Sacks
(1972: 342244) and Garvey & Hogan (1973).
. See especially Churchill (19721973). 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: Whats blackn whiten hides in caves,
(0.4)
R1: Aright I give up, Whats blackn 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: Whats blackn whiten 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 Hallidays
three analytic concepts: They each represent cases of unmarked focus and are thus equiva-
lently representations of theme and given.
. Sequential organization is not the only order of utterance organization studied in Con-
versation 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 (19721973). 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, 3672.
London.
Garvey, Catherine, & Hogan, R. (1973). Social Speech and Social Interaction: Egocentrism
Revisited. Child Development, 44 (3), 562568.
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. (19671968). Notes on transitivity and theme in English, Parts IIII.
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. Phila-
delphia: 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, 181199.
Jefferson, Gail (1979) A technique for inviting laughter and its subsequent acceptance/
declination. In G. Psathas (Ed.), Everyday language: studies in ethnomethodology (pp.
7996). 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, 249263.
Sacks, Harvey (1974). An Analysis of the Course of a Jokes Telling in Conversation. In
J. Sherzer & R. Bauman (Eds.), Explorations in the Ethnography of Speaking. Cambridge:
Sacks, Harvey (1992). Lectures on Conversation. Edited by Gail Jefferson, with an
Introduction by Emanuel A. Schegloff. Oxford: Blackwell.
the Organization of Turn-Taking for Conversation. Language, 50 (4), 696735.
Schegloff, Emanuel A. (1968). Sequencing in Conversational Openings. American
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 conversations turn taking organization. In G. Button & J. R. E. Lee (Eds.), Talk and
social organization (pp. 7085). 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:
Sudnow, David (1967). Passing On: The Social Organization of Dying. New Jersey: Prentice-
Hall.
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! Ill get some advance birthday cards, hhm hmh!
(0.6)
A: hhh A::nd uh,
(0.5)
A: Me:h,
(0.2)
A: Oh Sylvies sistuh hadda ba:by bo:way.
B: Who?
A: Sylvies sister.
B: Oh really?
A: Myeah,=
B: (Thats)
=[[
A: She had it yestihday.=Ten:: pou:nds.
B: Je:sus Christ.
A: She had//da ho:(hh)rse hh hh
B: (Thats a ba:by.)
(2) [JSLR:48-49]
A: Oh you haftuh tellm about yr typewriter honey,
B: Oh yes.
C: Yeah didju hear from them?
B: Yes,
A: We had m- more // trouble,
B: (Olympias gonna) put in a;
A: Oh you told em I forgot.
B: (Theyre gnna) put inna new, keyboard.
D: Oh they are?=
C: =Fer nothing!?
D: (Yeh)
C: Wl good fer you::::!
(3) [Trio:l8]
A: Oh you know, Mittie- Gordon, eh- Gordon, Mitties husban died.
(0.3)
B: Oh whe::n.
A: Well it was in the paper this morning.
B: It wa::s,
A: Yeah,
(1.2)
B: You mean, it was in the obit colum? // or // -an article about im.
A: Yeh. Yknow N:no, no, just in the,
B: Oh, hell. I took my paper tuh- (0.6) -tuh work an I putt in the
wasteba//sket I didnt-
A: Yeh,
B: I-
[[
A: They jus said Gordon Fremont, yknow, 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: Werent we?
B: Oh. Out here?
A: Yeah.
B: Yeah. Yeah. Right. Right.
A: Yeah. Angies gonna have a baby.
B: Oh really!
A: Yeah.
B: Well, congratulations
A: Isnt that sumpn // I didnt think she-
B: ((aside, off phone)) Angies gonna have a baby.
(5) [CN II-11:4]
A: Yknow I toldjuh thet uh theyd only been there a dayr 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 sumpn thets wi::ld. ((up amplitude))
B: What.
(1.0)
A: Yknow onea these great big red fire alarm boxes thetr on the corners?
I got one.
B: Really?
A: With all the works The inside // en everything.
C: How (grea:t)!
[[
B: O::h (thatll // be good.)
(7) [NB:202]
A: Ive quit smokin yknow en evrything.
(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 didnt get home til, .hhh two lasnight 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 Ill aftuh tell yihbout it,
B: Oh::::.
(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: Yd probly turn green with envy if I tolja all about what I got ( ) first //
quarter.
D: Tellem abou//t it.
E: Four As n two Bs.
(1.0)
A: You were takin all those courses?
E: Yeah.
D: Cn you believe that?
A: Thats 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 gbout 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 Im alright.
B: Ye(hh)ah I cn imagine how it hurts,
Collaborative turn sequences*
Gene H. Lerner
In conversation, the pre-emptive completion of one speakers turn-construc-

tional unit (Sacks, Schegloff, & Jefferson 1974) by a subsequent speaker can
operate on that unit in a way that transforms its production into a se-
quence a collaborative turn sequence. By completing the prior speakers
turn-constructional unit (or TCU) that is, by producing a version of what
had been projected as a part of the prior speakers turn a sequence can be
initiated. Here a recipient responds to a prior speaker, not by waiting until com-
pletion 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 ac-
ceptability of the proffered completion can be implicated as a next action. The
original speaker thereby maintains authority over the turns 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 re-
cipient) will ordinarily not be treated as confirmable, and thus will not launch
a sequence.1 In this circumstance, the shared authority inherent in the co-
construction 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 turns 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 speakers projected talk prior to its occur-
rence (or in the case of teasing or heckling, the affiliating utterance will use
the constructional format to pointedly mis-project the content of the turns
completion ).
2. Affiliating utterances are built to be contiguous with the preliminary com-

ponent 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 speak-
ers.3 It is this feature that distinguishes a bid for speakership in the course of
another speakers TCU that is only a bid for conditional access to that on-
going 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 comple-
tion, 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 mid-
utterance startup.
(1) [JJ:Invitation]
B: I was just gonna say come out and come over here and talk this
evening. But if youre going out // you cant very well do that.
C: Talk you mean get drunk, dont you.
B: what?
Here the turn-initial repeat of talk is used to tie Cs 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 oc-
curred prior to the most recent transition-relevance place provides a warrant
Collaborative turn sequences
for Cs mid-utterance startup as a very late attempt at a response. That is, it lo-
cates the prior transition-relevance place as a transition space for a late-starting
utterance.4
Progressivity is maintained by the contiguous placement of the pre-
empting utterance. However, when contiguity is lost (e.g. when the affiliating
utterance begins after the current speaker has begun the projected final compo-
nent) 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. Com-
ponent contiguity is maintained at the expense of progressivity. (cf. Lerner
1996a: 245251).
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 compo-
nent 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 cn handle evrything.
BC: Mm:://hm::,
Caller: and anything.
BC: except a cross wind at the short end a that runway.
.
Caller: hhh Oh dcrosswind didn you know wid a J-Tree, well
BC: jus drop one wing an // slide in,
Caller: Thats right,
Gene H. Lerner
The speaker of the affiliating utterance (jus drop one wing an slide in,) pro-
duces 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 oc-
curs 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 speakers 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 units completion. This can be seen in
the following instance.
(3) [GTS]
Ken: insteada my grandmother offering him a drink, of beer shell
say [ wouldju-
Louise: [wanna glassa milk? [ hehhh
Ken: [no. wouldju like a little bitta heing?
((i.e. herring))
Here, the placement of the affiliating utterance (wanna glassa milk) is cen-
tral 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 en-
tirely different.
(4) [GTS]
Louise: wanna glassa milk? [ hehhh
Ken: [no. wouldju like a little bitta heing?
However, a quote is in force and it is projected to be an offer. By beginning the
offer after the authorship attribution (Shell say), Louise produces an utter-
ance that is not taken to be an actual offer, but a report of an offer by Kens
grandmother. The no. wouldju like a little bitta heing? 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-
ers to suspend the nextness for which the utterance will otherwise be in-
spected. 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-in-
progress, 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 ac-
complished. 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). Se-
quence 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 ac-
cepted 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 turns 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 pre-
cisely, the receipt of the completion in terms germane to the action imple-
mented 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 ordi-
narily 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 oc-
casioned 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 com-
pletion 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 acknowl-
edge the other speakers 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 gree-
B: [it will come back
A: y- yes uh huh
(6) [H and M]
Hal: ... the answer is perhaps, though I dont really know, that it isnt
a substitution,
Max: its a transformation.
Hal: No its not even a transformation
Other forms of completion-acknowledging response are possible that fall
somewhere between acceptance and outright rejection as in (7).
(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, thats one way.
In (7) Daughter seems to reluctantly accept Dads 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 com-
pletion, 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 dont 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 acknowledg-
ing and accepting it by now incorporating it into their own completion.10
The collaborative turn sequence occurs, not across two distinct turns sep-
arated by a transition-relevance place, but within the purview of one partic-
ipants turn at talk within that participants projected turn space. The col-
laborative turn sequences 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 fea-
tures 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 par-
ticipants 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 com-
pletion, 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 se-
quences is through a comparison with repair. There is a striking parallel be-
tween pre-emptive completion and other-initiated repair (Schegloff, Jefferson,
& Sacks 1977). Both are launched by sequence-initiating actions that locate
the just prior speakers 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 main-
tain 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 TCUs
original speaker (Mom), thus, in a sense, markedly shifting recipients.
(9) [MOTHERS DAY]
Mom: if you should decide to live with a fella
(.)
Daughter: Mom will still talk to me ((laugh))
Mom: Ill 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
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: but it seemed to be, to Ken at least
Roger: the wrong kind
Dan: the wrong kind // of distinction
Ken: well you dont 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 par-
ticipant 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 Dans 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 com-
pletion 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 com-
pletion thus, in effect, pre-empting the receipt slot, as well as the turns
completion.
In (11), which includes two instances of this [completion + agreement to-
ken] 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 weve been told but I dont know whether this is true:.
Jenny: Mm::,
Ann: that (0.9) no wa:y can you get out,
Jenny: <once n option is started. Ye:s thats right. Yes.
:
Ann: Ahn if you nohrnlly pay ahnnually. evn if you hahvent paid
ye:t,
Jenny: youve gotta pay up tih date. Yes.
In both cases, Jenny not only claims her authority by confirming Anns un-
derstanding, 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 speakers
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 actu-
ally referring to grampa at the beginning of line 1, though she could seem to
be addressing him.)
(12) [PV]
C: well grampa its bad enough when e when he uhm::=tells you how
much tmake, but when e tells you what t co:ok,
(0.3)
M: then its rilly bad=yeah. yeah // didju- uh
C: He tol me t be suren check iz minu up there on the wall
Here, M offering the strongly projected, but unspoken, completion as an inde-

pendently 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 alterna-
tives available to the recipient of the pre-emptive completion (i.e. the original
speaker of the TCU): delayed completion and list construction.
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 heck-
ling 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 speakers current utterance,
across the talk of another participant, to their prior syntactically unfinished
utterance, by constructing the current utterance as a syntactically fitted con-
tinuation 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 comple-
tion 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 youre 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 Im 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 ahll just leave it (.) to somebody who is
(0.2)
Dan: youre not going to have children?
(.)
Roger: so trained, (.) thave children.
Here the delayed completion (so trained, (.) thave children.) locates the in-
tervening utterance as interruptive of the TCU-in-progress. As a continuation
of the speakers prior utterance it interdicts the sequential implicativeness of
Dans 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 speakers just prior utterance.
(16) [GTS]
Ken: My opinion of the school system, the Los Angeles school dis-
trict, dis- district, is the most fucked over,
Roger: Yeah well we // all got that opinion.
Ken: school systm 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 dont 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 interven-
ing utterance out to be interruptive of that turn space whether or not there is
any overlap.
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 in-
terruptive 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 comple-
tion can be used to interdict the (collaborative turn) sequential relevance of the
pre-empting completion. In (18) a delayed completion (then you think hes
gonna fall asleep) realizes a competing completion. It doesnt not acknowledge
or ratify the pre-emptive completion.
(18) [GTS]
Ken: no its when he turns a bright red that everybody has to start
worrying
Louise: no when he gets his eyes like this an he starts thinkin, you know
Ken: then you get to worry
Louise: then you think hes gonna fall asleep.
And in (19), an interrupting (though not disagreeing) delayed completion in-
terdicts 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 dont think its- thet- yknow thet
the kids thetr skinny, (0.7) are gonnuh yihknow haftuh worry
about it. They cn eat twice iz much iz you,
D: en it doesnt 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, agree-
ment or other responding action (first attempted through the pre-emptive
completion) can remain relevant for next turn. Pre-emptive completion pro-
vides 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 speakers 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: it sounds like what youre saying is that let them make the
decisions
Kerry: an let us know wh//at it is
Sparky: and let us know what it is
Kerry: yeah
Here delayed completion is used to reclaim the speakership of a TCU, thus
reversing again the earlier reversal of speakership/recipientship that was ac-
complished 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 al-
ternatives. One can find both occurring after the production of pre-empting
utterances. However, since pre-emptive completion by an addressed recipi-
ent 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 multiple-
participant 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
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: hes guaranteed, but (.) the rest of the members cn control that
member from him:
Kerry: goin any higher [ than tha[t
Dad: [goin hi [gher than that
Mom: [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 re-
ceipt 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 daughters utterance (well, thats 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, thats 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 dont record that slow. So Id 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 shd 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:: dont say thahhh//a(h)t
Nancy: NO is thawhatcher think//ing?
Hyla: No::,
3. Disregard of the completion through delayed completion
(25) [Labov:TA]
C: Fact I said tuh Larry yuh dont think its- thet- yknow thet
the kids thetr skinny, (0.7) are gonnuh yihknow haftuh worry
about it. They cn eat twice iz much iz you,
D: en it doesnt 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 com-
pletion 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 units 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 com-
pletion qua completion. Rejection of the pre-emptive completion (as an in-
stance 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-
tions 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 prac-
tices 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 construc-
tion), the relationship of collaborative turn sequences to overlap management
can be sketch out. The following section describes how pre-emptive comple-
tion 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 com-
pletion 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 youre 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 its a, its a mideastern yihknow its- 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 be-
yond 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: when it doesn invol:ve thee basic agreement, it is b://y stock
Dad: 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 com-
pletion. 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: no its when he turns a bright red that everybody has to start
worrying
Louise: no when he gets his eyes like this an he starts thinkin, you know
Ken: then you get to worry
Louise: then you think hes 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 de-
layed 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 its relevance for subsequent action,
when the content of the delayed completion does not do so. Starting the de-
layed completion in the course of the pre-emptive completion as in (30) can
accomplish this.
(30) [HIC]
Sparky: it sounds like what youre saying is that let them make the
decisions
Kerry: an let us know wh//at it is
Sparky: and let us know what it is
Kerry: yeah
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 vol-
ume). For instance:
(31) [GTS]
Dan: Like Ken is describing a guy who cn sort of (1.0) get
[ in with the group=
Louise: [be himself.
Dan: =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 theres 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 speakers own formula-
tion of the events as Ken does in (33) by adding an increment to his own
completion.
(33) [GTS]
Ken: And you think I really got pleasure out of getting uh well I-
getti//ng in that debate?
Louise: stomped on.
Ken: Cause thats 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 dont 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 Rs completion, thereby explicitly marking the pre-
emptive completion as unsuccessful.
The use of the receipt position to accept a pre-emptive completion or re-
claim 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 re-
asserting 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 recur-
rently 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 pro-
duced pre-emptive completion as an item, can both propose an initial ac-
ceptance of the affiliating utterance and then transform that acceptance into
something else as in (36).
(36) [GTS]
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 guysll come by with . hh
Louise: mud.
Ken: mud, ink or anything and throw it at em
The (unprojected opportunity space) terminal item completion, mud, is fol-

lowed 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 speakers utterance into a list pro-
poses 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 recogni-
tion 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 Jims 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 sim-
ply 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 re-
peat 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 utter-
ance from overlap as in (38).
(38) [GTS]
Dan: well I do know last week that uh Al was certainly very
(0.6)
Roger: pi//ssed off
Dan: upset, n pissed off, n angry en wz 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 character-
ization 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 for-
mulations 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 out-
right, 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 ac-
knowledge another speakers terminal item completion. The turn is extended
by turning the final word of the TCU into the first item in a list. Dan incor-
porates 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 accep-
tance of the candidate as the sole completion of the source turn-in-progress.
These technically-described practices underwrite a solution to a local prob-
lem. 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
(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 groups
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 therapists talk, and also taking it that one
part of therapists 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 aban-
donment 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 as-
sertedly correct completions for the TCU-in-progress. The receipt slot provides
a place to acknowledge the pre-emptive completion as the proper contin-
uation 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) [MOTHERS DAY]
D: .h and they do thi:ngs
(1.2)
C: ta hurt thm (0.2) huh huh=
D: =no:no: Im sayin that ugh ugh (0.7) thats 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: you gotta get up. Gway. Ygotta get up. ((heavy whisper)) put
the dogs out. Feed the fish
Roger: hyuhh heh hehh t(hh)ake the fi(hh)sh fer a wa(h)lk hh hehh //
hheh hh //hehh
Ken: No
Ken: He takes the dogs, heaves em in the backyard,
In (41), the pre-empting speaker does append a laugh token to her comple-
tion, 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 Kens grandmother offers his
father (Shell say)
(41) [GTS]
Ken: shell say // wouldja-
Louise: wanna glassa milk? // hehhh
Ken: No. wouldju like a little bitta heing?
Louise: heh// ha ha
Ken: wouldja like some crekles?
Louise: ehh ha ha ha ha
Ken: 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.
(42) [GTS]
Ken: I was on a road on the way to Roswell New Mexico. An I was,
yknow, plonkin along at a regular speed,
Roger: an a grasshopper // jumped onto the road who you recognized.
Ken: An- wait minute
Ken: No. //No I was goin along, an I had
Jim: ehhhhha ha: hhhh hhh
Here, the not serious character of the utterance itself (invoking an earlier refer-
ence to grasshoppers in a rather nonsensical way) shows that Roger is putting
words into Kens 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 utter-
ance. 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 pro-
jected 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 (con-
tinuing) 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 interac-
tional import of the receipt position as the place for the original speaker to
maintain authority i.e., authorship over their turns 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 ac-
tion 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 two-
action 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 co-
constructs 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 expla-
nation or the events of a story.
. Contiguity refers to the placement of next speakers 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: The cops, over the hill. Theres a place up in Mulholland where theyve-
where theyre building those hous//ing projects?
Roger: 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-
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 co-
participant that attempts to extend the preliminary component of a compound TCU-in-
progress, 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: See this is little peanuts now, but if it ever got to be something
[people would be fi:]:ghting an ah: thats
Dad: [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
moms 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 co-
participants. 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 com-
pletion 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 com-
pletion 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 con-
cept. 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: 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
Roger: to begin with
Rather than completing a TCU-in-progress, Rogers 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 pre-
empted 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 speakers 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 speakers action,
but where that prior action did not make such a response conditionally relevant.
. This is one way that Goffmans (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 comple-
tion, 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 un-
said. 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
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) [Mothers 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)
B: . t . hhh Oh well thank you but you we ha- yihknow Thomas.
(16.4) [Agorio]
E: Didju request a copy of the: (0.3) instructions then?
R: Yeah.
(16.5) [Smith:Thanksgiving]
R: Wheres thah:
T: Its on the counter ((pointing))
R: Oh, okay.
Further, a completion need not be produced even when one seems to be encouraged by a
speaker as in Bs second utterance in (16.6).
(16.6) [GL:DS]
B: and uh but then she says she gets to thinking, oh well shes just not
gonna worry about it.
A: Mm hm.
B: you know, shes just gonna
A: 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 re-
sist 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 in-
spected 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 ini-
tial 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 accep-
tance 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 re-
jected 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 childs drink (milk) in contrast to the
adults 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 childs offer (a snack). Kens 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 Kens list become more juvenile with each item. In a way,
Kens third item does seem to capture the same flavor, if I can put it that way, of Louises
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. 102128). Cambridge, Cambridge University Press.
Drew, Paul (1987). Po-Faced Receipts of Teases. Linguistics, 25, 219253.
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), 127149.
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. 299345).
Cambridge: Cambridge University Press.
Heritage, John, & Raymond, Geoff (2003). The Terms of Agreement: Indexing Episte-
mic Authority and Subordination in Talk-in-Interaction. Paper presented at the
International Institute on Conversation Analysis, Copenhagen.
Jefferson, Gail (1973). A Case of Precision Timing in Ordinary Conversation: Overlapped

Tag-Positioned Address Terms in Closing Sequences. Semiotica, 9, 4796.
Jefferson, Gail (1978). Sequential Aspects of Storytelling in Conversation. In J. Schenkein
(Ed.), Studies in the Organization of Conversational Interaction (pp. 219248). New
York, Academic Press.
Jefferson, Gail (1983). Notes on some orderlinesses of overlap onset. In V. DUrso &
P. Leonardi (Eds.), Discourse Analysis and Natural Rhetoric (pp. 1138). Padua, Cleup
Editore.
Jefferson, Gail (1990). List Construction as a Task and Interactional Resource. In
G. Psathas (Ed.), Interaction Competence (pp. 6392). 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), 167177.
Lerner, Gene H. (1991). On the Syntax of Sentences in Progress. Language In Society, 20,
441458.
Lerner, Gene H. (1992). Assisted storytelling: Deploying shared knowledge as a practical
matter. Qualitative Sociology, 15 (13), 247271.
Lerner, Gene H. (1993). Collectivities in action: Establishing the relevance of conjoined
participation in conversation. Text, 13 (2), 213245.
Lerner, Gene H. (1994). Responsive list construction: A conversational resource for
accomplishing multifaceted social action. Journal of Language and Social Psychology,
13 (1), 2033.
Lerner, Gene H. (1995). Turn design and the organization of participation in instructional
activities. Discourse Processes, 19 (1), 111131.
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. 238276).
Cambridge: Cambridge University Press.
Lerner, Gene H. (1996b). Finding face in the preference structures of talk-in-interaction.
Social Psychology Quarterly, 59 (4), 303321.
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 talk-
in-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), 4975.
Gene H. Lerner
Sacks, Harvey (1979). Hotrodder: A Revolutionary Category. In G. Psathas (Ed.), Everyday

Language: Studies in Ethnomethodology (pp. 714). 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.
the Organization of Turn-Taking for Conversation. Language, 50, 696735.
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. 7193). Washington, D.C.: Georgetown University Press.
Schegloff, Emanuel A. (1987). Recycled Turn Beginnings: A Precise Repair Mechanism in
Conversations Turn-taking Organization. In G. Button & J. R. E. Lee (Eds.), Talk and
Social Organization (pp. 7085). 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. 3142). Washington, DC:
University Press of America.
Schegloff, Emanuel A. (1995b). Sequence organization. Unpublished Ms, UCLA Depart-
ment of Sociology.
Schegloff, Emanuel A. (1996). Confirming allusions: toward an empirical account of
action. American Journal of Sociology, 104 (1), 161216.
Schegloff Emanuel A., Jefferson, Gail, & Sacks, Harvey (1977). The Preference for Self-
Correction in the Organization of Repair in Conversation. Language, 53, 361382.
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 ut-
terances 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 inves-
tigation 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 consider-
able 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 fo-
cus 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 conver-
sational 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 extra-
sequence utterance by an upward amplitude shift of the utterance initiating
the sequence. I shall present materials below as they contribute to further in-
sight 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 sub-
components of the sequence under construction and, secondarily, to any auxil-
iary parts (local extensions) its sub-components may evidence. Specifically, the
joining of a speakers own successive utterances within a singly operative adja-
cency pair unit and the joining of a speakers 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 produc-
tion 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 conversa-
tion. 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 conversa-
tional 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 cir-
cumstances there can be... silence after a speakers 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 ex-
changes of greetings, and need not close segments with closing sections and
terminal exchanges. (Schegloff & Sacks 1973: 324325)
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 utter-

ances with regard to their productional properties and positioning within a
conversation. With regard to productional properties, I indicate a few com-
monly used close initiator types. One such common productional type em-
ploys 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). How-
ever, close initiators can be context-independent so that they do not necessarily
reflect aspects of the conversation of which theyre 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). How-
ever, clearly, a great many of the types discussed here have as their principal
feature that they display elements of the conversation theyre 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. Empiri-
cally, close initiators exhibit the analyzable property of selective positioning,
i.e., positioning with reference to another order of conversational organiza-
tion. 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 frag-
ment. Other party agreement will typically complete the close of the topic (cf.
Schegloff & Sacks 1973).
M: But uh I do feel frm indications that things are startin a ( ) sti:r a little
bit.
P: Well thats 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 yknow, theyre (r-) theyre bou:nd to.
M: Yeah.
P: In certain industry.
M: //Yeah
P: :hT.hh3
P: Different things ill pick up when it begins tbe spring a the yea:r an
evething.
M: Yeah.
P: :hT .hh But I think itll iron itself out.
M: I sure hope // so.
P: Ill 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 co-
conversationalists just prior utterance. A speakers selection of one particular
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 speakers 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 connected-
ness 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 po-
sitioning 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 implic-
itly 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 re-
source 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 Mecha-
nism 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 utter-
ance 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 terri-
tory 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 speakers immediately
prior extra-sequence utterances.
(1) JGP4
M: An so anyway tsneak out an uh call Miss M :hT.hh yknow/
F: Mhm.
M: :hT.hh So:
9.3 F: Hm
16. M: Thats all I know
31. F: Oka:y buddy ((Close intoned))
27. M: Okay ol buddy
F: Thank you Mi//ltie
M: Yeah.*
(2) NB
C: Bt Ill put it o:n the bottom of the drawer, nex to the meddle
( // )
B: Yeah.
B: Okay.
(0.4)
C: //( )-
B: Thatll be alright.
(1.0)
C: O//kay,
12.1 B: En Ill pick it up Saturday.

3.9 C: Okay. ((Close intoned))
6. C: Well see yuh (later)
17.1 B: O::kay Bill
(3) JCP3
R: I foun out whadi wanned f r // dinner
C: Okay.
R: .hhh 1 decided cause
.
.
.
R: An ken ya get some milk.
C: Milk/
R: Mhm.
R: Cuz theres none left.
7.7 C: Alright.
7 R: .hhh Okay.
22. R: Thats what l want f r dinner // hhh
8.4 C: Alri
R: Mkay
R: Goo//bye.
C: Bye.
R: 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: Hello Marty/
M: (h) hi Pete
P: How are you:=
M: Im great
P: Good
M: How are you
P: Pretty good.=
M: Hey,=
M: Were havin a meeting Tuesday night.
P: Tuesday night/
Jo Ann Goldberg
M: Yea:h a//n
.
.
.
M: So can you make it Tuesday night/
P: Yeah.
P: What time.
.
.
.
P: :hT.hh Different things ll pick up when it begins tbe spring a
the yea:r an evething.
M: Yeah.
P: :hT.hh But I think itll iron itself out.
M: I sure hope // so.
.- P: Ill see ya Tuesday.
9. M: Right.
P: O//kay Marty.
M: You-youre al*right. //You can get there.
P: Ye-
P: Yeah.
M: Okay.
P: Okay.
P: //Thank you.
7.7 M: See ya Pat.
5.5 P: Bye
(5) NB
C: Bt Ill put it o:n the bottom of the drawer, nex to the meddle
( // )
B: Yeah.
B: Okay.
(0.4)
C: // ( )-
B: Thatll be alright.
(1.0)
C: O//kay,
6. C: Well, see yuh (later)
B: O::kay Bill ((Close intoned)
4.2 C: Bye bye=
(6) JGP4
M: An so anyway tsneak out an uh call Miss M :hT.hh yknow/
F: Mhm.
M: :hT.hh So:
F: Hm
M: Thats all I know
31. F: Oka:y buddy ((Close intoned))
27. M: Okay ol buddy
F: Thank you Mi//ltie
M: Yeah.*
M: Yeah.
M: Oka//:y
F: Ill see ya.
M: Yeah.=
M: Okay Fr//aiik
F: Take it easy pa//rtnuh.
M: Yeah.
.- F: //Yeah.
.- M: Slon.
We observe here the same empirical regularities in the amplitude shifts in clos-
ing 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 ter-
minal exchanges are not necessarily adjacently positioned within the closing
construction. (See fragments 4, arrow l and 2; 6, arrow l and 2.) In fact, speak-
ers 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), re-
turn 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 tnight/ 1
P: Na:h.=I don think Im goin tnight Mel.=
M: Oh rilly/
P: No. H//uh uh
M: Ive got a happen//ing out at Mercer if y//ou wanna come.
.
.
.
M: Well // (okeydo)
P: .hhh
M: //I thought Id // ask ya, yknow jist // t (get) 2
P: Yeah
P: Yeah
P: Yeah
P: Yeah
P: Oka:y
M: //0kay P//ete.
P: .hhh
P: Ill see ya tamarra up the shop then. 3
M: Yeah,=
M: Okay // buddy.
P: .hhh
P: Okay Me//l
M: Okay.
P: Yeah.=
P: Th//anks. 4
M: Bye* bye.
P: Yeah.
(8) JG6:1
F: We:ll Ill give ya call maybe Thursday an see 1
whattu well,
F: Whadda you do on Saturday/
R: .hhh Nothine.=
R: Go tlaundryhh//hh .hhh hhh
F: Oh.
(0.5)
F: Uh huh.
F: .hh Well I might call ya Tliursday evening h 2
R: Okay hh
F: Have ya been, wuz ya home last Tuesday evening/
(0.5)
when I suppose tcall you // at fo:ur/
.
.
.
F: We:ll whadda ya got some lessons tget/
R: Yea:hhh
F: We:ll okay.
F: Be a good bo//::y.
R: .hh Okay // I
F: Ill try
F: Ill call ya Thursday even//ing. 3
R: Okay
R: gabye:
F: Well see whatjure do//in
R: .hh Alright
R: //Bye bye
F: Okay
F: Gbye Ro//bert
R: Bye
(9) JCP3
C: Hello:
R: :hT.hh Mother/
C: Ye:s.=
R: I foun out whadi wanned f r // dinner 1
C: Okay.
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.
C: Alright.
R: .hhh Okay,
R: Thats what I want f r dinner // hhh 2
C: Alri
R: Mkay
R: Goo//bye
C: Bye.
R: Bye.
(10) JGP4
M: Thats all I know.
F: Oka:y buddy ((Close intoned)) 1
M: Okay ol buddy
F: Thank you Mi//ltie. 2
M: Yeah.*
M: Yeah.
M: Oka//:y
F: Ill see ya 3
M: Yeah.=
M: Okay Fr//ank
F: Take it easy pa//rtnuh
M: Yeah
F: //Yeah
M: Slon
(11) JG 6:3
D: A:ndum hh mkay well then well look forward f- I hope he has a good
lesson on Friday.
S: //Yeah.
D: Hha-hm-hm-hm .hhhh
D: Okay honey.
D: .hhh An thank you so much f r calling. :h//hh 1
S: Ari
D: Bye bye
S: Bye
Given the above cited collection of close-constructional components, each ev-

idenced as an adjacency pair sub-component, we may ask whether the Ampli-
tude Shift Mechanism for affiliation of speakers own successive utterances will
remain operative over this lengthened course.6 We might reformulate this ques-
tion 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 clos-
ings 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: Well you wanna come out and have lunch with me/
B: No,
B: Let me take you to lunch, some//time.
A: OOOHHH! No. No.
A: Come on. Come out uh uhh say you get out here bout uh twelve
thirty. 1
B: Twelve thirty.
B: Dont fix very much, though, I just I uh eat very light anymore,
Im tryina get slim
.
.
.
A: Well Ill see you about twelve thirty 2
B: Twelve thirty Mon//day
Jo Ann Goldberg
A: Monday
B: Okay dear.
23. A: A:righ//t
25. B: Thank you
18.5 A: Bye
20. 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: Mkay well the:n well 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: Ari
10. D: Bye bye
4.8 S: Bye
(14) JGP6
W: Hello/
P: Is John there/
W: .hh No:.= Im expecting him anytime.=
W: //Is there a message/
P: Oh ah see.
P: No.
P: This is Pat from down at the drum corps down Santa Rosa Bay
Bay Shrine Club. .hh//h
W: Auhuh.
P: Is he goin tdrum practice tahmorrow mornin do you know/ 1
W: No: : . Hes no: t.
P: Oh I see.
.
.
.
W: .hh Well I know hes no:t
P: Oh I see.
25.1 P: Well okay jus longzi know. 2

W: Oh(h)ok//ay
23.2 P: Tha:nk // you
.- W: Thanku.
13. W: Bye bye.
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 weeks notice r something // maybe you cd
make it on a Friday night huh/
C: .hh Well that way I cd work ou:t yknow uh least a schedu:le en see
how itd uh works out with the rest of the people thet Im involved
wi://th.
D: ((Very quietly)) Yea:h.
(1.0)
D: Gee one a these Friday nights Ill give yuh a call if youd like to-
Id 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
yknow well see what we cn work out.
D: Great.
22.9 C: Awrighty/=
11. D: Okay Cathy,
.- D: //Ill see* ya then.=
20.3 C: Okay love
D: Yeah
C: Righ=
C: Byebye.
(17) JGP2
P: Oka:y ((Close intoned)) 1
M: //Okay P//ete.
P: .hhh
Jo Ann Goldberg
32. P: Ill see ya tamarra up the shop then.

M: Yeah,=
.- M: Okay // buddy.
P: .hhh
23.1 P: Okay Me//I
.- M: Okay.
P: Yeah.=
P: Th//anks.
M: Bye*bye.
P: 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-close-
initial 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 rela-
tive 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 clos-
ing sequence, the tactic has been to start with the terminal exchange and to
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 ampli-
tude level of speakers successive sub-component utterances were lowered over
the course of the sequence. It is by such means that larger sequential construc-
tions composed of multiple adjacency pair sub-components are cohered. (See
fragments 18, 19, and 20 below.)
(18) JGPP3
R: An ken ya get some milk.
C: Milk/
R: Mhm
C: Alright.
R: .hhh Okay
22. R: Thatss what I want f r dinner // hhh
8.4 C: Alri
10.9 R: Mkay pass
11.4 R: Goo//bye.
.- C: Bye.
10.9 R: Bye.
(19) NB
C: No hanky pan/ky.
B: No:: hanky panky.
(0.4)
15.1 C: Well have a good time.
23.1 B: Oh(hh) kay // (h) Ill-
14.5 C: N Ill, see yuh later.
17. B: Okay (Bill)
7.1 C: Yah. pass
.- B: //Bye.
.- C: Bye.
(20) SBI:6:l
A: Yea:h/
B: Mh//hm
Jo Ann Goldberg
28.5 A: Well Ill see you about twelve thirty

32.5 B: Twelve thirtyMon//day
27. A: Monday
22. B: Okay dear pass
23. A: Alrigh//ty
25. B: Thank you
18.5 A: Bye
20. B: Bye
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 speakers successive
utterances of a single adjacency pair exchange is differentially accomplished.
It is differentially accomplished by means of the Amplitude Shift Mechanisms
discriminative capacity by which it differentiates amongst primary and aux-
iliary 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 speakers own successive utter-
ances 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 Mechanisms affiliation/disaffiliation op-
eration evidenced in a speakers 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 pro-
ducing 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
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 adja-
cency 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 se-
quence 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 re-
cipients 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-conversationalists first.
(21) JGP3
22. R: Thats what I want f r dinner // hhh
C: Alri
10.9 R: Mkay pass
11.4 R: Goo//bye terminal
C: Bye
R: Bye
(22) JG6:l
R: Yea:hhh
F: We:ll okay
F: Be a good bo//:y
R: .hh Okay // I
F: Ill try
13.5 F: Ill call ya Thursday even//ing return to arrangements
R: Okay
R: Gabye:
10.4 F: Well see whatjure do//in return to arr.((cont))
R: .hh Alriglit
R: //bye bye
.- F: Okay. pass
5.6 F: Gbye Ro//bert terminal
R: Bye terminal
Jo Ann Goldberg
(23) JG6:8
L: I got a:ll the stuff. All I need is (ta) //see you.
M: Jst pick a time. //Ha ha ha ha
L: Okay clear heart
M: .hhh Thank you for calling an tell Dawn I said hello.
6.1 L: A wor, a word outta the past, huh/ signature 1
M: Ri:ght
3. L: Alright dear.
M: Gbye.
4.5 L: Gbye. 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 turns utterance is affixed. In the case of affil-
iation to a priors 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 priors prior by maintaining the
amplitude level of the present lower than that of the priors 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 affili-
ates. (See fragments 24, 25 and 26 below.)
(24) JG(R/M)
24. R: Iz it inna back of the park or in front, (0.6) or on the
sides. question one
M: .hh Its onna corner of .hhh the road that goes down by the park
and Bradshaw.
9.9 R: Oh.
(0.4)
21.2 R: What side right or left. question two
M: Its on Maple and Bradshaw.
.- R: Oh. pass
17.9 R: Its onna corner/ question three
M: Yeah.
6. R: .hhh Shoo pass
(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: Wheres e on twenny-first er
somethin/ question two
M: Ye.= Someting like that.=
12.6 P: Yeah pass
(26) JGP2
25.3 F: Whadiyou been doin. question one
M: (n) Oh I- I- We win out Hollywood tday.
17. F: Oh didja/ question extension
M: Yea//h
12.1 F: Yeah= pass
M: Yeah
2.7 F: :hh Hm. pass
23. F: Ja have a good ti:me/ question two
M: Well we looked over you know they had an open house.
17. F: Oh did they. question extension
M: Th-theres no u:h yknow mm races goin or anyth//een.
7.1 F: Mhm. pass
M: On Sundee.
.- F: M//hm. pass
M: .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 speakers successively positioned next first or sec-
ond pair-part was lower in peak amplitude than his just prior sub-components
extension. Such counter cases were not without their orderly aspects. Quite reg-
ularly, 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: HIlo:
Jo Ann Goldberg
R: Hlo.=
R: Thisiz Sills.=
R: When will you-Dyou need a lift/
.
.
.
C: Ill 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 didnt
(1.7)
C: Well then Ill haftuh stop an get im so//mething so, gbye
R: :hT.hh
R: Okay.=
28.9 R: So ya nea-so ya dont- so ya needa lift.= close initiator
C: ((Angrily)) I DONT.
27.2 D: You do/ repair initiator
C: ((Increased anger)) I DONT.
15.8 R: Okay. pass
C: Gbye. terminal
1.9 R: Gbye.
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 Rs Gbye 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 ut-
terances of a single adjacency pair exchange and its extension is marked by
successive lowering of peak amplitudes over its course. A similar operation ob-
tains for the joining of multiple adjacency first and second pair-parts where
they constitute sub-components of a larger sequential construction. The Am-
plitude Shift Mechanisms discriminative capacity differentiated between these
two operations.
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 de-
viation 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 ampli-
tudes 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 pair-
part 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 mkay
18.7 D: Well the:n well 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: Ari
D: Bye bye
S: Bye
Jo Ann Goldberg
(29) JGP2
M: //I thought Id // ask ya, yknow jist // t(get)
P: Yeah
P: Yeah
P: Yeah
P: Yeah
19. P: Oka:y ((Close intoned)) close initiator
M: //Okay P//ete. close return
P: .hhh
31.6 P: Ill see ya tamarra up the shop O.P./C2 /arrangement
M: Yeah,=
M: Okay // buddy.
P: .hhh
P: Okay Me//I
M: Okay
P: Yeah.=
P: Th//anks.
M: Bye* bye.
P: Yeah.
(30) JGP4
M: Thats all I know.
31. F: Oka:y buddy. ((Close intoned)) close initiator
M: Okay ol buddy. close return
35.2 F: Thank you Mi//ltie. O.P./C2 /thank you
M: Yeah.*
M: Yeah.
M: Oka//:y
F: Ill see ya.
M: Yeah.=
M: Okay Fr//ank.
F: Take it easy pa//rtnuh.
M: Yeah.
F: //Yeah.
M: Slon.
(31) JG6:2
C: Yknow I got that figured out too.
R: Mhm
7.2 C: Okay baby lemme letju go honey // (cuz I gotta go over to the
post office cuz I dont have anybody in the store right now n it//
wont matter if I close. close initiator
R: Mkay close return
R: Mhm
R: Mkay
C: Okay honey pass
R: Mkay
8.3 C: Okay well // Ill see ya probably ohh around
sixish. C2 /arrangement
R: Gbye
R: Mkay
R: Gbye
C: Bye dear
R: Bye
However, after the occurrence of this second raise in peak amplitude, suc-
cessively 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 D: Well the:n well 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
20. D: .hh An thank you so much f r calling. C2 /thank you
:h//hh
S: Ari
10. D: Bye bye terminal
S: Bye
Jo Ann Goldberg
(33) JGP2
19. P: Oka:y ((Close intoned)) close initiator
M: //Okay P//ete. close return
P: .hhh
31.6 P: Ill see ya tamarra up the shop O.P./C2 /arrangement
M: Yeah,=
M: Okay // buddy.
P: .hhh
25.1 P: Okay Me//It
M: Okay.
14.8 P: Yeah.= pass
16.8 P: Th//anks. thank you
M: Bye* bye
13. P: Yeah. terminal
(34) JGP4
31. F: Oka:y buddy ((Close intoned)) close initiator

M: Okay ol buddy
35.2 F: Thank you Mi//ltie. O.P./C2 /thank you
M: Yeah.*
M: Yeah
M: Oka//:y
27.5 F: Ill see ya arrangement reference
M: Yeah.=
M: Okay Fr//ank
.- F: Take it easy pa//rtnuh well wish
M: Yeah.
18. F: //Yeah. terminal
M: Slon
(35) JG6:2
7.2 C: Okay baby lemme letju go honey // (cuz I gotta go over to the
post office cuz I dont have anybody in the store right now n it //
wont matter if I close. close initiator
R: Mkay
R: Mhm
R: Mkay
C: Okay honey pass
R: Mkay
8.3 C: Okay well // Ill see ya probably ohh around
sixish. O.P./C2 /arrangement
R: Gbye
R: Mkay
R: Gbye
3.8 C: Bye dear terminal
R: Bye
In the data presented above, two types of closing components were presented,
Thank yous and arrangements. Each evidenced a raised amplitude deviation
in the position just after the close-initiating exchange. But, when such com-
ponents 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: Oka:y ((Close intoned))
M: //Okay P//ete.
P: .hhh
31.6 P: Ill see ya tamarra up the shop then.
M: Yeah,=
M: Okay //buddy.
P: .hhh
25.1 P: Okay Me//It
M: Okay
14.7 P: Yeah.= pass
16.8 P: Th//anks. thank you
M: Bye* bye
P: Yeah.
Jo Ann Goldberg
(37) JG6:1
R: Yea:hhh
F: Well okay.
14.3 F: Be a good bo//:y
R: .hh Okay // I
F: Ill try
13. F: Ill call ya reference to arrangement
R: Okay
R: gabye:
10.4 F: Well see whatjure do//in
R: .hh Alright
R: //Bye bye
.- F: Okay pass
5.6 F: Gbye Ro//bert
R: Bye
(38) JGP4
F: Oka:y buddy.
M: Okay ol buddy.
35.2 F: Thank you Mi//ltie.
M: Yeah.*
M: Yeah.
M: Oka//:y.
27.5 F: Ill see ya. reference to arrangements
M: Yeah.=
M: Okay Fr//ank
.- F: Take it easy pa//rtnuh.
M: Yeah.
18. F: //Yeah.
M: Slon.
(39) SB2:2:1
B: Well I // wont keep you Bea, but ( )
A: Well
A: Yeah.
A: Okay
A: Well Im sorry thetchuhh
B: Well som I bt // ( )
A: Maybe its just as well you discovered it.
B: Well it tis, cause Id hate to be sliding backwards down that hill

A: Yeah ehhh heh heh uh rather
B: Ehheh heh heh
A: Okay
29. B: Well you give // ( )
A: See yuh-
A: And Ill see yuh Sunday
28.6 B: A::Iright, // Bea
A: Alrighty.
20.6 B: Thanks, // a million thank you
A: Bye
18.5 B: Bye
These data illustrate the placement of two proper closing components in the
position just after the close-initiating exchange: Thank yous (see fragments
32 and 34) and references to arrangement (see fragments 33 and 35). Al-
though 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 yous), 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 occu-
pied 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. Par-
enthetically, 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 po-
sitioned 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 arrange-
ment/rearrangement operations, but is elsewhere evidenced over a range of
other conversational activities which include second stories and return invita-
tions to name but two. These occurrences reflect conversationalists preference
for fittedness in the placement of such activities. The fitting of a conversa-
tional 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 arrange-
ments are such commonly employed close-initiated components shall not be
considered. What is of interest here is that such occurrences result in a close-
initiating 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 fit-
tedness 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)
11.5 A: Well then Ill see you tomorrow close initiation/

evening. reference to arrangement
B: Ill s-Ill see you tomorrow,
B: And you said six oclock/
A: No seven
B: Seven
A: Its just dessert you see.
B: Yeah.
B: Okay, then.
A: Seven oclock.
B: //Oh ( )
16. A: Uh however. However, uh you can rearrangement
come in earlier if you want to
B: Oh no, heh heh heh Ill be in a little earlier of // course.
A: Yes.
A: Uh huh
B: And uh an thats the-
B: An Im disappointed because I had- re-initiation
Id 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 dyou-
uh the reservation, or will you
B: Oh thats right.
(41) JG6:3
D: .hh Okay darling well thatll be jist fine.
D: An of course as I say I dont expect any extra time but jist tlike
for the finish of a thought hha
S: Oh sure.
Jo Ann Goldberg
19. D: :hhhh Okay then well see you- close initiator/

ala- hell see you then reference to arrangement
around about nine-thirty on friday morning.
S: Fi://ne.
D: Aha:.
D: .h//h Okay well see you then,
S: Bye bye
19.8 D: .hh And listen shall I leave you a check for seven
dollars/ rearrangement
S: Yeah :hh
D: Okay
S: Thats alright
D: Ahuh.
14. D: Okay honey.= ((Close intoned)) second close initiation
13.5 D: Thank // you so mu
S: Nobody ill, nobody else ill rearrangement
be home.= Thats good :hh // topic initiation
:hh :hh
23. D: We:ll now I dont know uh//h topic initiation
((Clears throat))
S: Thats alright.
S: I mean you know this doesnt 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 yknow for // .hh
S: Yeah.
S: Well I tell you something ease I like to either .hhh uhh always
feel free tcall me uphhh=
Second, on occasions that the rearrangement component is produced in
briefest form, its extra-close sequence potential may be documented by its con-
sequences for the subsequent course of the closing. Commonly, after comple-
tion of brief rearrangements to close-initiating arrangements, parties in their
return to the course of the closing produce yet another reference to arrange-
ments. (See fragments 42 and 43.)
(42) JG3:8
P: Ill see ya Tuesday. reference to arrangements
9. M: Right.
P: O//kay Marcus
19. M: You- youre al*right.
//You can get there. rearrangement
P: Ye
P: Yeah.
M: Okay. pass
P: Okay.=
P: //Thank you.
7.7 M: See ya Pete. reference to arrangements
P: Bye.
(43) JGP110
D: But Ill 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
cn work out.
D: Great-.
C: Awrighty/=
11. D: Okay Cathy,
.- D: // Ill 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 reini-
tiated 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: Well thanks f r calling okay/= close initiation
R: Yeah.
11.8 R: Okay.
26. R: Wyou gonna go tmorrow/ topic
M: Oh yeah.
R: T Okay.
M: Okay/=
R: The bus doesnt come Im not going.
.
.
.
12. R: Dr. Buckley come pick us all up.
M: Oh hhlhh wouldnt do it. Oh huh // I (ra-)
25.1 R: rrral see ya. second close initiation
M: Okay.
M: Bye.=
R: Bye.
(45) NB(A/C)
C: Oh I love tuh-
C: Gee I ride mine all // a time
A: Yeah.
C: I love it.
(0.2)
A: .hh Well, honey/ Ill pobly see yuh one a close initiation
these da : ys,
12. C: O::h God yeah,
A: //Uhh huh!
C: We
17. C: Bt I c- I jis // couldn get down //there topic
A: Oh-
A: Oh I know.
A: Im-not asking...
.
.
.
A: But Im fine,
(0.4)
C: //Yeah.
20.6 A: I think Ill make it.
C: Okay.
25.5 A: .hhhh Alright,
C: Ill see yuh next week then second close initiation
A: Bah bye
C: Bye bye,
(46) NB (GH)
G: ArigIit,
(2.4)
G: Ill see if I cn (1.0) -rout sumpn
up.
G: Ill call yuh back inna few minnits. close initiation
10. E: Alright,
15. E: San Clemente is always uh topic
I know darn well ycn get on
there.
G: Ken we/
E: //Yeah.
G: Oh-* Oii Saturdee/
E: Yeah.
.
.
.
10. E: So I dont know how ihd be tday.
Bt- ygivem a try.
13. E: Why donche try en nen, gimme second close initiation
call.
G: Alright.
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 reg-
ularly 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 ap-
paratus 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. Al-
though 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 at-
tended. 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.
It has been argued here that re-initiations of topical talk and rearrange-
ments stand in an extra-sequence (potential) relationship to the close-initiating
utterances they succeed. A party initiating such a sequentially recognizable al-
ternative 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 close-
initiating 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 lo-
cated: re-initiations of topical talk (see fragments 40 and 4l), rearrangements
(see fragments 40, 4l, 42 and 43), Thank yous (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 top-
ical 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 opera-
tions can serve as the occasion for their rearrangement, both inside and outside
of closing sequences. I argued that, since reference to arrangements is a promi-
nently 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 vari-
ous 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 am-

plitude, 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 fur-
ther 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 speakers 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 environ-
ment in which more than one sequence type can find systematic and position-
specific 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 ramifi-
cations for this pattern. That is, the Mechanisms patterned operations may
exhibit systematic variances according to the different sequences whose con-
struction it is in the service of and the complications that such sequences may
undergo. Such a possibility awaits further investigation.
Notes
. Discussion of the phenomenon of a closing structure for conversation will be in the ser-
vice 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 clos-
ing 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 proce-
dure 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 other-
wise 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 turn-
constructional component such as thank you (Ps last line in fragment 4) or See ya +
((name)) (Ms 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 distinc-
tion 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 speakers 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 be-
tween 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 compo-
nent 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 ad-
jacency 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 producers immediately prior utterance. The notation C2 will indicate the occur-
rence 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 Ill call ya Thursday evening is preceded by a close
component which is not itself the close-initiating utterance. Rather, the close-initiating ut-
terance 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 dis-
tinction 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.
Ill 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 sub-
components. 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, Ds Ill give you a call stands as an arrangement component to an in-
definite next encounter. After Cs 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 ampli-
tude initiator. That half was closely divided amongst the types of operations mentioned
here: Thank yous, arrangements, re-initiations of topical talk, and rearrangements. It is
obviously the case that a finer order investigation of the hypothesis I forward here awaits col-
lection 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. 199218). New York: Academic Press.
Schegloff, Emanuel A., & Sacks, Harvey (1973). Opening Up Closings. Semiotica, 8, 289327.
Index
A designed conversations vs. by-product

account-elicitation devices 165 conversations 79
acquainted-unacquainted 83 double-checking 114
acquainted-stranger relationship
65
E
already informed 186
ethnographic setting 189
ambiguity 131
answerer 95
announcements F
content-based definition of first topic 192
announcements 174
expanded announcement
sequences 176, 195 G
assessments 173 gap and overlap 40
authoritative knowledge 233 gap vs. pause 40
overlap onset 44, 58, 241
overlap resolution 45, 59, 243
B generalized list completer 244
being ordinary 137 generalizing list completer 245
by-product conversations see designed given/new distinction 171, 193
conversations vs. by-product see also marked and unmarked
conversations information focus
C H
clearance cue answers 67, 70 hello 74, 88
closing sequence 259 see also possible greeting
auxiliary components 274 Hynek, J. Allen 165
opportunity position 279
rearrangements 286
re-initiation of topic talk 290 I
comic book orthography 19 inaccurate records 110
commitment to the normal 154 information focus 209
see also odd events interruption 42
courtesy terms 64 investigative work 109
D J
damaging evidence 114 joke/serious 131
Index
L Q
laughter 18 question/answer sequence 262
local identity 186
loyalty to reality 154
R
recipient design 177
M receipt slot 230, 239, 243
marked forms 73 receipt slot alternative 235, 244
marked competition 50 receipt slot pre-emption 233
marked and unmarked receipt slot rejection 247
information focus 209 recognitional response 18, 44, 58
Menninger, Karl 154 reference a matter vs. revive a matter
minimal preference rule see preference 285
mishearing 16
mono-topical calls 186
S
Sacks, Harvey 22, 79, 82, 125, 131,
N 137, 146, 154, 165, 192
neutrality 117, 120 self-identification 89, 93
new information see given/new summons/answer sequences 63,
distinction 179, 186
O T
odd events 146 telephone answerer see answerer
see also commitment to the telephone as territorial 98
normal thank you 285
overall structural organization 79, terms of address 64
86, 174, 191 turn taking
overlap see gap and overlap turn allocational techniques 37
turn claimant 49, 59
turn occupant 49, 59
P turn-taking repair 232
pause see gap and overlap who speaks first 96
polar position of conversation 36 turn-taking system 229
Pomerantz, Anita 165, 173 type of conversation 69
possible greeting 77
preference
U
minimal preference rule 91
uncertainty 233
preference not to re-tell 177,
unmarked competition 50
183, 192
pro-form 199, 201
progressivity 226 V
break in progressivity 246 varying pronunciation 20
In the Pragmatics & Beyond New Series the following titles have been published thus far
or are scheduled for publication:
74 TROSBORG, Anna (ed.): Analysing Professional Genres. 2000. xvi, 256 pp.
75 PILKINGTON, Adrian: Poetic Eects. A relevance theory perspective. 2000. xiv, 214 pp.
76 MATSUI, Tomoko: Bridging and Relevance. 2000. xii, 251 pp.
77 VANDERVEKEN, Daniel and Susumu KUBO (eds.): Essays in Speech Act Theory. 2002. vi, 328 pp.
78 SELL, Roger D.: Literature as Communication. The foundations of mediating criticism. 2000.
xiv, 348 pp.
79 ANDERSEN, Gisle and Thorstein FRETHEIM (eds.): Pragmatic Markers and Propositional Attitude.
2000. viii, 273 pp.
80 UNGERER, Friedrich (ed.): English Media Texts Past and Present. Language and textual structure.
2000. xiv, 286 pp.
81 DI LUZIO, Aldo, Susanne GNTHNER and Franca ORLETTI (eds.): Culture in Communication.
Analyses of intercultural situations. 2001. xvi, 341 pp.
82 KHALIL, Esam N.: Grounding in English and Arabic News Discourse. 2000. x, 274 pp.
83 MRQUEZ REITER, Rosina: Linguistic Politeness in Britain and Uruguay. A contrastive study of
requests and apologies. 2000. xviii, 225 pp.
84 ANDERSEN, Gisle: Pragmatic Markers and Sociolinguistic Variation. A relevance-theoretic approach
to the language of adolescents. 2001. ix, 352 pp.
85 COLLINS, Daniel E.: Reanimated Voices. Speech reporting in a historical-pragmatic perspective. 2001.
xx, 384 pp.
86 IFANTIDOU, Elly: Evidentials and Relevance. 2001. xii, 225 pp.
87 MUSHIN, Ilana: Evidentiality and Epistemological Stance. Narrative Retelling. 2001. xviii, 244 pp.
88 BAYRAKTAROLU, Arn and Maria SIFIANOU (eds.): Linguistic Politeness Across Boundaries.
The case of Greek and Turkish. 2001. xiv, 439 pp.
89 ITAKURA, Hiroko: Conversational Dominance and Gender. A study of Japanese speakers in rst and
second language contexts. 2001. xviii, 231 pp.
90 KENESEI, Istvn and Robert M. HARNISH (eds.): Perspectives on Semantics, Pragmatics, and
Discourse. A Festschrift for Ferenc Kiefer. 2001. xxii, 352 pp.
91 GROSS, Joan: Speaking in Other Voices. An ethnography of Walloon puppet theaters. 2001.
xxviii, 341 pp.
92 GARDNER, Rod: When Listeners Talk. Response tokens and listener stance. 2001. xxii, 281 pp.
93 BARON, Bettina and Helga KOTTHOFF (eds.): Gender in Interaction. Perspectives on femininity
and masculinity in ethnography and discourse. 2002. xxiv, 357 pp.
94 McILVENNY, Paul (ed.): Talking Gender and Sexuality. 2002. x, 332 pp.
95 FITZMAURICE, Susan M.: The Familiar Letter in Early Modern English. A pragmatic approach. 2002.
viii, 263 pp.
96 HAVERKATE, Henk: The Syntax, Semantics and Pragmatics of Spanish Mood. 2002. vi, 241 pp.
97 MAYNARD, Senko K.: Linguistic Emotivity. Centrality of place, the topic-comment dynamic, and an
ideology of pathos in Japanese discourse. 2002. xiv, 481 pp.
98 DUSZAK, Anna (ed.): Us and Others. Social identities across languages, discourses and cultures. 2002.
viii, 522 pp.
99 JASZCZOLT, Katarzyna M. and Ken TURNER (eds.): Meaning Through Language Contrast. Volume
1. 2003. xii, 388 pp.
100 JASZCZOLT, Katarzyna M. and Ken TURNER (eds.): Meaning Through Language Contrast. Volume
2. 2003. viii, 496 pp.
101 LUKE, Kang Kwong and Theodossia-Soula PAVLIDOU (eds.): Telephone Calls. Unity and diversity
in conversational structure across languages and cultures. 2002. x, 295 pp.
102 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 KHNLEIN, Peter, Hannes RIESER and Henk ZEEVAT (eds.): Perspectives on Dialogue in the New
Millennium. 2003. xii, 400 pp.
115 KRKKINEN, 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 nal 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 STENSTRM (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 GONZLEZ, Montserrat: Pragmatic Markers in Oral Narrative. The case of English and Catalan.
2004. xvi, 409 pp.
123 MRQUEZ REITER, Rosina and Mara 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 rst generation. 2004. x, 300 pp.
126 WU, Yian: 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, Mara 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

10 1 1 453 9317 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

10 1 1 453 9317 PDF

Uploaded by

Copyright:

Available Formats

Conversation Analysis

Pragmatics & Beyond New Series

John Benjamins Publishing Company

Library of Congress Cataloging-in-Publication Data

P95.45. C6644 2004

2004 John Benjamins B.V.

I. Taking turns speaking

II. Implementing actions

III. Sequencing actions

In addition to these introductory remarks, this volume also includes a second

Taking turns speaking

Turn taking practices organize the allocation of opportunities to participate in

and introduces the important conceptual distinction between marked and

on the practices they employ. She is able to ground participants taken-for-

Activities carried out in conversation are organized into sequences of actions.

On meeting Harvey Sacks A concluding remark

I first met Harvey Sacks in the course of investigating doctoral programs. I

be an understatement to say that I was not prepared for what I encountered in

Schegloff, Emanuel A. (1968). Sequencing in conversational openings. American

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

Jefferson, G. (1985). An exercise in the transcription and analysis of laughter. In T. van

Glossary of transcript symbols

// Double obliques indicate the point at which a current speakers talk is

In this case, Pammys utterance is produced as Yeah well okeedoe I js

Mike: iYeh its all in the chair=

word Degree signs bracketing an utterance or utterance-part indicates that the

2. The inverted question-mark () is also used as a substitute for the question-mark/comma

t*, d* In early computer transcripts, an asterisk following a consonant replaces

Ruth: Monday nights we play, (0.3) <I mean we go to ceramics,

Taking turns speaking

. A positioning of conversation among the speech-exchange systems1

. A simplest systematics for the organization of turn-taking

For conversation, preservation of one party talking at a time is organization-

Achieving minimization of gap and overlap

Since Rule 2 provides that if current-speaker-selection of a next speaker is done,

Current speaker might extend the length of a correctly appreciated last

Overlapping talk initially appears to be messy a product of peoples not much

adjustments; i.e., manipulation of larger parts of the utterance a speaker is

(a) Stutters. Depending upon our transcripts detailedness, stutters are

A similar provision for some part of an overlapped utterances occurrence in

an observable trouble. We might call the former procedure Marked Compe-

1. Marked self-retrieval: Restart

A party to an overlap, having dropped out to resolve overlap and re-establish

And at second starters dropping out after an attempted start:

2. Unmarked self-retrieval: Continuation

A party to an overlap, having dropped out to resolve overlap and re-establish

Roger: Even [in-

3. Marked other-retrieval: Repeat request

This procedure appears to have some regularity of occurrence after competitive

Mary: [[ I s i d I get mo]re the n h(h)alf=

4. Unmarked other-retrieval: Acknowledgment and embedded repeat

Kitty: They said theyve been, huge wa:ves.=

point is the moment of completion of an ongoing utterance, this last result-

Jefferson, Gail (1973). A Case of Precision Timing in Ordinary Conversation: Overlapped

It has been suggested1 that the production of an answer to a summons is a se-

Answering a summons in face-to-face settings

In selecting an answer, one set of resources is the availability to the summoned

Answering the summons on the telephone

The features of the opening of telephone conversations are somewhat differ-

answers. Aside from yeah, telephone summonses may be answered by hello

They may be employed by answerers when that information may be treated as

E: Well, call er can call me back.

partially constitute it, or are consistent with its constitution as an intercom

t, d In early computer transcripts, an asterisk following a consonant replaces