Read stokoe Chapter 3, Pages 63-86 ‘You are the turns you take’ ask 3 questions.
Talk: The Science of
Conversation
Elizabeth Stokoe
1
ROBINSON
First published in Great Britain in 2018 by Robinson
Copyright © Elizabeth Stokoe, 2018
The moral right of the author has been asserted.
All rights reserved.
No part of this publication may be reproduced, stored in a retrieval system, or
transmitted, in any form, or by any means, without the prior permission in writing of
the publisher, nor be otherwise circulated in any form of binding or cover other than
that in which it is published and without a similar condition including this condition
being imposed on the subsequent purchaser.
A CIP catalogue record for this book
is available from the British Library.
ISBN: 978-1-47214-082-1
Robinson
An imprint of
Little, Brown Book Group
Carmelite House
50 Victoria Embankment
London EC4Y 0DZ
An Hachette UK Company
www.hachette.co.uk
www.littlebrown.co.uk
2
For Eunice
3
To hear some of the recorded conversations that feature in the book, and
other bonus content, visit www.carmtraining.org/talk/extras
4
CONTENTS
1: The conversational racetrack
‘Hello!’
‘Um, hello . . . ?’
The conversational racetrack
2: Here comes the science
Scriptwriters are conversation analysts
The building blocks of talk
How to build a successful invitation
Interruptions, dating, and a roll-on deodorant
The science of conversation
3: You are the turns you take
First moves and ‘first movers’
How turns build people
You are the words you say you are
Conversation Analytic Personality Diagnostic
5
Talk builds people
4: Sticks, stones, and . . . talk
Do actions speak louder than words?
Is communication 93 per cent body language?
Do women and men talk differently?
Do different ‘cultures’ talk differently?
Are ‘ums’ errors?
Why are there so many myths about talk?
5: Every word matters
Word selection matters
The trouble with ‘women’
How words change outcomes
‘Some’ or ‘any’?
Getting people to donate to charity
‘Talk’ or ‘speak’?
‘Are you willing?’
Talk matters
6: Are you being served?
How to get help
Service in seconds
Service burden
6
Recruiters and the recruited
Thank you for not helping
7: How to have better conversations
The problem with communication skills
The future of talk
Five ways to have better conversations
Appendix
Acknowledgements
References
Index
7
1
THE CONVERSATIONAL
RACETRACK
This book will change the way you think about talk. It will show you how
to lift the lid on the engine that drives our everyday lives. It will show you
why there is a big pay-off to understanding talk scientifically.
Of course, when we try to understand something like talk
scientifically, it is not the same as trying to understand something like a
black hole. Talk, as a phenomenon of social life, exists only to be
understood.1 It is designed by humans for humans to get every facet of
life accomplished. We build, maintain and end our personal and
professional relationships through talk. We buy and sell. We get and give
help. We are excited, persuaded, irritated, embarrassed and consoled in
response to things others say to us. Talk is the tool we have to do all of
these things.
Talk is also our resource for fixing things that go wrong. When we
characterise some aspect of our talk as ‘communication breakdown’, we
are probably referring to feeling trapped in a conversation, or that we are
8
struggling to get heard, or as though we are on the back foot. This book
will show you how to unpack and understand talk and what to say to
anticipate, resist, avoid – or further aggravate – these problems.
Scientists strive to understand black holes, even though they do not
exist in the first place to be understood by humans. Talk, on the other
hand, exists only to be understood by humans. In that sense, a book about
the science of talk should be an easy read; the conceptual gap between the
audience and the phenomenon is small. On the other hand, it might be a
challenge to convince you that we need a scientific approach to the study
of something that seems, on the surface, so easy to understand. Can we
really gain anything by analysing something that we ‘just do’? The
chapters in this book will show you that the answer is ‘yes’. And it will
show you that the science of conversation is not just interesting, or fun. It
is crucial.
Along with many other academic colleagues around the world, I have
spent the last twenty years working as a conversation analyst, studying
recordings of real talk from real people talking to each other in real time.
While the linguist Noam Chomsky once described conversation as a
‘disorderly phenomenon’, it is, in fact, highly organised. And we are quite
unaware of how systematic our talk is, and how different words lead to
different outcomes. So while we all keep talking, we are not good at
understanding precisely what went wrong in an encounter, or what went
right. Analysing real talk in the wild – and in slow motion – shows us the
incredible power talk has to shape our daily lives.
Think about the last time you went on a rollercoaster.2 You might have
seen – even if you did not purchase it – the snapshot of yourself as you
zoomed down the steepest incline. Rollercoaster snapshots reveal how
you looked at a particular moment, even though you are unable to
remember or reproduce it later. Similarly, we are not capable of recalling
and reproducing – with the exact words, the exact intonation and the
exact facial expression – what we said at, say, thirty-three seconds into a
particular conversation. Studying talk scientifically allows us to freezeframe that moment and scrutinise it to see how it worked, what worked
and what did not work.
In this opening chapter, we investigate what happens at the start of
encounters. Talk can run smoothly or awkwardly from the very first
9
‘hello’. We can predict what is likely to happen next from the first words
uttered. This is because all conversations have a landscape; a
conversational racetrack. We start at the beginning of an encounter with
another person or people and, along the way, complete projects. We
anticipate and avoid or crash into hurdles. Conversations become tense,
difficult or all-out war. An opening ‘hello’ can even be, quite simply, a
matter of life and death.
‘Hello!’
The discipline of conversation analysis was invented by three academics
in the 1960s: Harvey Sacks, Emanuel Schegloff and Gail Jefferson.3
Sacks was killed in a car crash in 1975 and, like celebrities who die
young or in mysterious circumstances, there is something of a rock ’n’
roll mystique about him. There are also many arguments over his legacy
and competitive in-groups and sub-camps typical of any academic field.
This book is not the place for lengthy accounts of the discipline, but it
is worth knowing that each of these three figures holds a special position
in the hearts and minds of conversation analysts around the world. And,
as an advocate for conversation analysis to a likely non-specialist
audience, and for context, you might be surprised to learn that
conversation analytic research is among the most cited in academia, even
compared to ‘hard science’ disciplines.4
Before we see our first example of two people starting to talk, let me
explain briefly how their talk will be presented. As with all the other data
analysed in this book, recordings of real (generally anonymised)
interactions are transcribed using a technical system developed by
Jefferson.5 The system is the first stage of analysis and cannot be
reproduced by a machine – even though speech-to-text software will
produce reasonably accurate verbatim transcripts (see ‘The future of talk’,
in Chapter 7).
The ‘Jefferson system’ is designed to represent not just what is said
but how it is produced, placed, timed and so on. It includes information
about intonation, lengths of pauses and gaps within and between turns at
talk, the onset and end of overlapping talk, and the precise moment when
10
who is speaking changes. It also tells us how speakers build actions
through talk – the component activities that comprise complete
encounters. At first, the transcripts might seem overly technical. But this
is how people talk – keeping in the detail of how talk works, rather than
stripping it away to the words alone. It’s what actually happened.
For the first few extracts in this chapter, though, I will present a
verbatim version of the transcript alongside the technical version, to
enable you to become familiar with the system. A key to all of the
symbols appears at the end of the book. You can also listen to some of the
examples I present online,6 where they feature as part of a lecture or
presentation.
Our first example is an ordinary domestic telephone call between two
American friends, Hyla and Nancy. Hyla is calling Nancy; this is a
telephone line without caller ID.7 Her dialling is the summons that starts
the conversational ball rolling. It is the first action that must be completed
for any further talk to happen.
Example 1: Hyla and Nancy
Take a look at Nancy’s answering turn, ‘H’llo:?’ The punctuation marks
help researchers, and you, understand exactly how she said it. For
instance, the colon indicates that the ‘o’ sound is slightly elongated. The
question mark indicates ‘questioning intonation’. You might be surprised
to learn that ‘questioning intonation’ does not always accompany the
spoken production of a question. In fact, questions are more often
delivered with falling or ‘closing’ intonation, indicated by a full stop.
This is one of many communication myths that we will bust in this book.
11
Systematic patterns exist in the way conversation starts, and not just
on the telephone. This is because although the means of communication
might be different (e.g., language choice, modality, gesture, sign
language, written, spoken), and the exact words used also vary (e.g.,
‘hello’, ‘hiya’, ‘yo!’), the actions being done remain the same.
Watch how Hyla and Nancy’s conversation unfolds.
When Hyla says, ‘Hi:,’ she delivers a greeting but also communicates
recognition of her friend’s voice. The comma indicates what conversation
analysts call ‘continuing intonation’ – think how it sounds to read items
on a shopping list or the digits in a telephone number.
At line 04, Nancy gives a second greeting, but it sounds very different
to her answering-the-phone ‘H’llo:?’. This time, the ‘HI::.’ is louder,
brighter, and more animated. It conveys recognition of Hyla’s voice.
Next, Hyla and Nancy exchange ‘how-are-yous’.
Hyla and Nancy exchange ‘how-are-yous’ rapidly, and reciprocally. The
speed of exchange is represented by the equals signs (‘=’), which indicate
that the two turns are ‘latched’ together. The quick pace of their turntaking is also indicated implicitly, because when gaps occur between turns
– as we will see in later examples – they are measured and included in the
transcript.
On the face of it, this is utterly mundane. There’s no science here!
Well, actually there is. Conversation analysts have shown that, across
12
settings including phone calls, face-to-face encounters, Skype calls and
even instant messaging, conversations recurrently open with three rapid,
reciprocal, component pairs of actions:
1. The summons and answer (the opening at lines 01–02).
2. The greetings and identification (for Hyla and Nancy, just the sound of the
voice is enough for identification at lines 03–04).
3. The initial enquiries (the ‘how-are-yous’ at lines 05–06).
This sequence of actions is pretty robust across opening sequences,
whether on the telephone or face to face. In Example 2, Dad and Liz are
Skyping. The pound signs indicate ‘smile-voice’ – how people sound
when they smile as they talk. You can watch this clip online.8
Example 2: Dad and Liz
In Skype conversations, the ‘summons and answer’ sequence still occurs
because someone still has to ‘go first’ in any interaction. However, the
identification part may be redundant, because we generally know who we
will talk to when we initiate a video chat. Of course, I can use my
partner’s mobile phone or Skype account to make a call, and the recipient
may not expect to see or hear me when they answer the summons.
Example 3 comes from a written exchange between two friends on a
social media messaging app.9
Example 3: Isla and Joe
13
The same components are in place: ‘summons and answer’; ‘greetings
and identification’, and the reciprocal ‘how-are-yous’. They take a bit
longer, because writing is slower than talking. But otherwise, the same
actions happen.
Example 4 is from a conversation between two office workers from
different departments of the local council.
Example 4: Katy and Darcy
In this recording, the square brackets at lines 05 and 06 represent
overlapping talk. So as Darcy says ‘m’duck’ – one of many terms of
14
endearment used in the UK, such as ‘love’, ‘hen’ or ‘pet’ – Katy begins to
respond to Darcy’s ‘how-are-you’.
There are several features that you may not initially notice in this
example, but that are important when it comes to understanding the more
complex conversational openings we come to later. So, first, note that the
way Katy answers the phone tells us that this is not a domestic setting, but
a workplace. Second, Katy’s next turn, at line 04, is delivered with ‘smilevoice’, after she knows who she is speaking to. She does not smile when
she answers the phone.
Third, Darcy and Katy do not know each other well enough for Darcy
to simply respond ‘Hi’ at line 03, as we have seen in previous examples:
she returns Katy’s greeting fully and identifies herself explicitly. But
Darcy and Katy do know each other well enough to now move into a
reciprocal ‘how-are-you’ sequence.
The equals signs at lines 03, 04 and 05 indicate when speakers rapidly
add another part to a turn that is potentially already complete.
Conversation analysts refer to the ‘point of possible completion’ that
happens when a speaker has done something that can be responded to by
the next speaker, such as asking a question. These are usually
grammatically complete, too. Points of completion are also indicated
strongly by intonation. As you read this book, read out loud, if you can.
Each time a full stop happens, your intonation should fall. In the Jefferson
system, full stops indicate ‘closing intonation’ and that a ‘unit’ of a turn
has been completed.
It is worth pausing here, to point out that in everyday written text, full
stops (‘periods’ in the USA) are not, however, intonation markers! They
are grammatical markers, placed at the end of grammatical sentences.
That is something that all readers of this book will know, whether
intuitively or explicitly. Jefferson borrowed and reassigned the full stop
and other conventions of written text (colons, commas, question marks,
etc.) to talk’s intonational patterns. Of course, both transcription and
everyday written punctuation are technical and precise – we are just more
used to the conventions for written text.
Returning to Darcy and Katy’s conversation, at line 03, Darcy’s turn is
complete as she says ‘good morning t’you.’ but she rapidly adds a second
‘unit’ to the end of her turn: ‘.=it’s Darcy:.’ What a simplified verbatim
15
transcript would not show is Darcy’s tacit analysis of her response to
Katy. The rapid addition of a second unit reveals Darcy’s ongoing
monitoring of the unfolding interaction. She needs to identify herself.
This is not my analysis of what Katy and Darcy are doing, it is Darcy’s.
My job is to reveal their analysis to you.
Think about when a friend invites you for dinner. They may say
something like, ‘Do you fancy dinner on Friday? Or Saturday?’ The
addition of ‘Or Saturday’ shows their ongoing monitoring of your lack of
immediate response. The idea that taking a turn in a conversation requires
‘processing time’, producing pauses, is another myth about talk that we
will bust in this book. In fact, we can and do respond very quickly within
milliseconds. Indeed, speakers are actually monitoring reactions while
their own turn is in progress, which is what enables such rapid
interchange to take place. And the fact that we (can) respond quickly
provides us with the evidence that delays, gaps, silences indicate an
upcoming problem. Of course, we can respond quickly with turns that
start with built-in delay-tactics (e.g., ‘Um:::::’). But a delay in responding
to an invitation indicates an upcoming turndown. Because they are fast to
recognise a possible turndown, your friend quickly adds another option,
to which you can hopefully say ‘yes’ . . .!
This is the tacit knowledge that people have for interacting. We reveal
it as we construct our turns, word by word, turn by turn, although we
cannot articulate what we are doing. But, without knowing it, we
anticipate hurdles on the racetrack of conversation, and try to avoid them.
Finally, look closely at the overlapping talk at lines 05 and 06. The
overlap shows that Katy has heard enough of the second unit of Darcy’s
turn to know what action Darcy is doing even though it is not complete:
she is doing a reciprocal ‘how-are-you’. This sort of overlap is very
common in interaction: people can sufficiently anticipate the action that is
coming to begin to respond without leaving any gap between turns.
You might be thinking that everything you have read so far is worthy
of an entry in a Private Eye ‘Pseud’s Corner’ jibe – and, in fact, an early
observation from Sacks, Schegloff and Jefferson, that ‘one speaker speaks
at a time and speaker change recurs’, actually was! It is certainly common
to hear people describe or reflect on conversational components like
‘how-are-you’ sequences critically, as a bit of pointless ‘filler’, said ‘just
16
to be polite’, in a meaningless and ‘non-genuine’ way. I will show that
these sorts of comments misunderstand the importance of starting to
unpack the basic machinery of talk as a precursor to persuading you of
some crucial insights into how talk works.
You might also be thinking that there are numerous instances where
conversation openings are quite different. And, of course, they are. Here
are four quick examples.
Example 5: Calling the doctor’s
Example 6: Salesperson calling a business
In Example 5, a patient phones her local surgery. The parties greet each
other (‘>Good< mornin:g,’ and ‘Hello’), but, in this opening moment,
17
only the receptionist provides identification of both the surgery and her
first name. Who the patient is becomes relevant later; for now, the main
action is requesting an appointment. There are no ‘how-are-yous’ –
potentially tricky in this setting! Example 6 has two people at work
calling each other, but, unlike Katy and Darcy, they do not know each
other.
Example 7 is a domestic phone call, but Lesley, who answers the
ringing phone, is being called by someone who wants to make a delivery
to Lesley’s house. Example 8 is another of Katy’s calls, but this one is to
her daughter at home. As with all our examples, use the transcription key
presented in the Appendix to unpack the technical aspects about which
you want to learn.
Example 7: Lesley and Mr Harris (simplified transcript)
Example 8: Katy and her daughter
18
There are reciprocal greetings – spot Lesley’s second one at line 06 in
which she also shows she recognises Mr Harris and has expected him to
call. See also that Mr Harris recognises Lesley from her answering
‘hello’. As the caller, he is in a position to know who is likely to answer
the phone. However, he identifies himself using his own name and the
name of the business. He does not expect Lesley to recognise him from
his voice alone; they do not have the kind of relationship where Mr Harris
could simply say ‘Hi!’ next.
Example 8 neatly demonstrates how we calibrate the purpose of our
call by including or dispensing with things that would ordinarily happen
in the conversation. While mother and daughter exchange greetings, there
are no ‘how-are-yous’. Why not? Dispensing with ‘how-are-yous’ helps
us know the kind of conversation we are in. This is going to be a quick
call to check on the important activity of getting into the house and
disabling alarms. So by asking a question about the house’s alarm in the
slot usually filled with a ‘how-are-you’, Katy indicates to her daughter
what type of call this will be.
Call openings vary, then, depending on contingencies like who is
calling whom; whether or not speakers know each other; the urgency of
the situation; and so on. These are not random, messy variations, but
systematic ones, by which we construct and recognise the particular
nature of each type of call. It is the fact that we do this – making our
actions recognisable for each other – that makes those same actions and
methods recognisable for conversation analysts.
19
We conclude this section on ‘Hello!’ by focusing again on the three
actions done in openings: the summons and answer; greeting and
identification, and ‘how-are-yous’. Example 9 is another call in which we
encounter Hyla, this time talking to her boyfriend, Richard. Their opening
shows the irresistibility of completing these three actions, which further
undermines the notion that they are pointless ‘filler’. They are doing
something important in an interaction: framing it, establishing its footing
and the relationships involved.
Harvey Sacks made a classic observation about ‘how-are-yous’ in a
paper called ‘Everyone Has to Lie’.10 He was referring to the kind of
social situation that requires a ‘fine, thanks, how are you’ response, not a
long (happy or sad) answer to the question. Not everyone is the right
person to receive the ‘true’ answer (e.g., ‘I’m feeling lousy’; ‘I’m so
excited!’), but neither is every slot in a conversation the right place to say
it. In Hyla and Nancy’s call, it later turns out that Hyla is not fine – after
the opening ‘how-are-yous’ she goes on to recount to Nancy her problems
with Richard. But everybody has to lie.
The call between Hyla and Richard has an extended opening. There is
a lot of laughter and breathiness in the call; Richard arrived home to hear
the phone ringing inside the house and has raced in to answer it. The
transcript represents the laughter and breathiness. I hope that, by now, you
can see how impoverished a verbatim transcript would be, when
compared to this technical version. Imagine if you were not permitted to
use emoticons or emojis when writing text messages or using instant
messaging services. Imagine I presented an analysis of text-based
interaction and deliberately removed these features. Representing talk
with this technical system is not just good science. It’s an ethical decision.
This is how people actually talk, and I want to show you what really
happened when Hyla called Richard.
Our
target
three
actions
–
the
summons/answer;
greetings/identification; and ‘how-are-yous’ – are highlighted.
Example 9: Hyla and Richard
20
21
All three actions are present in the call, with the summons and answer and
greeting and identification done rapidly and sequentially. However, the
greetings are done a second and third time, punctuated with Richard’s
breathiness and Hyla’s laughter.
Hyla asks Richard ‘how are you’ for the first time at line 10, in the
expected slot. His response is more heavy breathing, to which Hyla
responds with more laughter. Her laughter is, in fact, helping Richard out.
22
She fills in the slot for his turn with laughter, as well as providing an
account for why he cannot talk – he is ‘out of brea(h) th(h).’, she says.
Eventually, Richard produces the standard ‘fine’ response at line 14.
Hyla and Richard’s opening sequence is extended with more talk;
Hyla reports that she nearly hung up the call; Richard accounts for why
he took so long to answer. It is interesting that Hyla asks a second ‘howare-you’ at line 23. This time, Richard gives an ironic response: ‘Oh, I
love driving up to the door and hearing the ph-’. But see what he does
next! He cuts off this response, which is starting to sound like a complaint
about Hyla making him run to the house. He then says ‘no’, explicitly
verbalising the fact that he ‘will not go there’; he will not keep
complaining. In so doing, Richard halts the development of a possible
argument before it has started. Instead, he asks a reciprocal ‘how-areyou’. He uses his tacit knowledge of what should happen next to avoid
potential conflict, and Hyla gives the standard response.
Hyla and Richard show us that the ordinary words and phrases, like
greetings and ‘how-are-yous’, are a useful resource. They start an
encounter. They can be extended to help someone who is out of breath.
More importantly, though, the fact that Hyla and Richard do and redo
these actions shows that they are an interactional imperative. They simply
have to be done. People show us this all the time. Speakers monitor each
other’s turns for their appropriateness of action and position.
The interactional imperative holds for written talk too, as we saw
earlier in Isla and Joe’s messaging app conversation. Here is Isla talking
to another friend. When Jane opens the conversation in a non-standard
way (‘urgh’), Isla responds a minute or so later with a comment about
Jane’s inapposite conversation opener.
Example 10: Isla and Jane11
23
These early conversational actions, whose patterns are robust across
settings of all kinds, can also be usurped as resources for conflict and
argument, as we will see in the next section.
‘Um, hello . . . ?’
So far, the conversations we have encountered have been friendly, or
affiliative. They have started smoothly, with the speakers more or less in
alignment. In this section, we move on to examine some rather more
problematic – and even life-changing – conversational openings.
We start with a call between boyfriend and girlfriend, Dana and
Gordon. They are students who are home for their holidays and living
with their respective parents.
Example 11: Dana and Gordon
Let us stop here, before Dana has even spoken. Lines 01–03 are all we
need to see to know that there is trouble ahead. This is because of what
happens at line 03; or rather, what does not happen. Dana does not issue a
reciprocal greeting. Instead, there is a silence of 0.7 seconds. This might
not seem like long, but it is. And it is just this kind of detail that allows us
to zoom in on parts of the conversational racetrack where there might be
either smooth progress – alignment – or trouble – disaffiliation.
It is worth restating that the silence of 0.7 seconds is not required for
Dana to think about what her response should be. Go back to Hyla and
Nancy’s call – Example 1. Between their turns at talk there was no gap,
and no overlap. Their turns were produced in rapid succession – but not
automatically.
24
Conversation analysis is not a branch of behaviourism. Saying ‘hello’
does not guarantee a ‘hello’ in response. In fact, a missing ‘hello’ can
indicate mishearing, rudeness, or some other potential communication
problem. But the take-home message is that because conversations do
regularly open like Hyla and Nancy’s, something interesting is happening
when the pattern breaks.
Returning to Gordon and Dana’s conversation, here is Dana’s delayed
response.
So, Dana returns the greeting at the start of her turn, but immediately
opens up a new action: a question. And it is a challenging question. It is
not just an information-seeking question; it has got some bite in it. It is a
complaint. She is Gordon’s girlfriend and so is somewhat entitled – her
question presupposes this – to know where he is. She should not have had
to try to get him ‘all morning’. Of course, ‘all morning’ is not meant
literally. But Dana uses this rather extreme – but very ordinary and
recognisable – way of describing her sustained attempts to speak to
Gordon. Including – or not including – ‘all morning’ makes a difference.
Every word matters, for the things we are doing with our turns at talk.
It turns out later that Gordon had phoned Dana’s house, very late and
drunk, the previous evening. Now Dana is in trouble with her mum. What
possibilities does Gordon have for responding to Dana’s question? He
could say, ‘What do you mean, where have I been all morning?’, or
‘We’re not joined at the hip’, or something else that meets the challenge
head on, attacking Dana or defending himself. And then the couple might
head straight into an argument.
Another option is to push back on the trajectory started by Dana’s
question, which is what Gordon does.
25
Gordon responds with a bright-sounding ‘Hello!’, very much like Nancy’s
‘Hi!’ of recognition in Example 1. In so doing, Gordon recruits his tacit
knowledge of what kind of turn typically fills this slot in an encounter. In
other words, Gordon recruits the conversational racetrack.
This is not the end of Gordon’s turn, however. It unfolds in a way that
further marks Dana’s question as, literally, out of turn.
Note two tiny details that only a Jefferson transcript can magnify: an
‘Uhm’ and a 0.6 second pause. Sounds like ‘Uh’, ‘um’, ‘uh:::::’ can be
subtly altered in myriad ways with intonation. They are not random, or
meaningless, or speech production errors. They can be systematic,
cropping up in exactly the same types of conversational environments.
For instance, a little ‘uh’ sound frequently crops up when speakers are
confronted with an inapposite or unexpected prior turn – which Dana’s is.
And this analysis is supported by the fact that what Gordon did first was
what Dana could expect – he gave a ‘hello’ of recognition.
These three items at the start of Gordon’s turn, the bright ‘hello!’, the
‘uh’ and the 0.6 second pause, all push back on Dana’s first turn – her
first move. The idea that Dana is a ‘first mover’ – the kind of person who
opens an encounter with such a challenging first turn – is something that
we will return to in Chapter 3. We will see how we use talk to assess
everything about a person from their personality and disposition to their
motivation, emotions and attitude. Talk is often the only evidence we
have to make such assessments. We are the turns we take.
Only after these three items have been produced does Gordon answer
Dana’s question. For those who are enjoying getting into the detail, there
are other tiny perturbations in his turn: an in-breath right at the start (.hh)
– this is what ‘take-a-deep-breath’ looks like! And he makes a lip-smack
sound (.pt).
26
And after answering the question, Gordon once more reveals his tacit
understanding of what should happen next.
So Gordon initiates the third part of a typical opening sequence: the ‘howare-you’. Now, if Dana was satisfied with Gordon’s answer at line 05, and
ready to get on with the conversation, we would expect to see her respond
with ‘Fine’ (or a variant) and a reciprocal ‘how-are-you’. However, this is
what happens next.
By now, you should know enough about conversation analysis to interpret
a delay of 0.5 seconds as an indication that Dana is ‘not fine’. We can
now predict that she is not likely to move into a reciprocal pair of ‘howare-you’ turns with Gordon. Are we right?
Yes! We can predict the future when we study talk scientifically. Dana’s
intonation makes her response sound anything but ‘okay,’ (but everyone
has to lie . . .). There is no smile-voice. The ‘continuing intonation’ on
‘okay,’ suggests there is more (about Dana) to come. And she does not
ask Gordon how he is.
Dana has given Gordon enough evidence for him to pick up on the fact
that something is wrong. So now he has another option. Should he ask
Dana ‘what’s up?’ That could, of course, open up a can of worms. So
instead, he treats Dana’s ‘I’m okay,’ as a straightforward and positive
response to his question.
27
The ongoing saga of Dana and Gordon has grabbed the interest of online
commentators who have seen my talks with this clip presented with
audio. People have tweeted to ask, ‘Whatever DID happen to Gordon and
Dana?’, or commented that ‘Gordon can do better. IMHO’. And consider
the following comment from a YouTuber.
Example 12: YouTuber
Audience responses such as these provide evidence that these openings
are useful for explaining what conversation analysts do. And they show
how accessible, recognisable and reproducible talk is – after all, it is what
we all do . . .
Example 13: Lyrics from ‘Telephone Line’, 1976, Electric Light Orchestra
If a silence of 0.7 seconds is enough to indicate trouble for Gordon, a
silence of 2.5 seconds is enough for Jeff Lynne’s caller to know they are
in a very poor situation. The caller moves to ask ‘how-are-you’, when no
response is forthcoming. The ~tilde~ sign represents ‘wobbly voice’, a
feature added to the Jefferson system by Alexa Hepburn,12 who is an
expert in transcribing crying and other emotional sounds.
28
Dana and Gordon’s call is important because it explains some
fundamental things about talk. First, it shows how two people can have
quite different agendas, or projects, in their encounters. Dana’s project
was to get to talk to Gordon about the trouble she is in with her mother,
caused by him. Gordon’s project was to avoid having that conversation!
Next, the 0.7-second silence before Dana uttered a word shows that we
can pinpoint moments of trouble very precisely. This sort of precision has
big pay-offs when it comes to understanding what works and what is less
effective in, say, professional or workplace encounters. For example, we
can search for silences and find out what happened previously to produce
it; what particular word, question, description or phrase. Examples
throughout the book will show that a question, an explanation, an offer,
an assessment, an invitation – and lots of other actions – can be
pinpointed as failing by looking at what happens in the very next turn.
There are other ways that call openings tell us there is trouble ahead,
or they can misfire completely. In Example 14, Frannie has called her
friend Shirley at work. She has been transferred to Shirley’s line by a
receptionist. These are the first turns exchanged between the pair.
Example 14: Frannie and Shirley
Shirley answers the phone in ‘work mode’. Stating ‘District Attorney’s
office’ is a self-identification in a way that ‘hello’ is not.13 In the ensuing
29
turns, the punctuation marks and arrows indicate a level of emotional
delivery that is hard to represent on paper. But it should be clear that
dispensing with greetings and ‘how-are-yous’ are signals for turbulence
ahead. Shirley uses her friend’s name with a sharp tone in her first turn,
rather than ‘hello’. People often think that using names ‘builds rapport’
between speakers. This is another communication myth we will bust later
in the book.
If you are reading this book looking for communication tips, then one
option Shirley has to respond is to do what Gordon did: give a bright
‘Hello!’ She could do what typically happens next, at the start of the
conversational racetrack. But Shirley responds using the same format and
vexed intonation as Frannie. The two are in perfect alignment, with
conflict their joint project. And look at the next two turns: Frannie makes
a ‘first move’ like Dana’s – a challenging question. But rather than push
back on the challenge as Gordon did, Shirley responds defensively. This
is not communication breakdown. It is the flawlessly produced, precisiontimed, start of a fight.
Example 15 comes from a business-to-business cold call. The
salesman, Jack, is calling Max for the first time: they have never spoken
before. Jack’s project is to try and make an appointment to meet Max with
a view to selling a contract for telephone lines.
Example 15: Business-to-business cold-call sales
30
Like Katy and Darcy in Example 4, we can easily identify this as a
workplace call, not a domestic call. Max’s answer to the summons
contains exactly the same components as Katy’s: a greeting (Katy said,
‘Good morning’); an identification (Katy said, ‘Katy Green’), and a
question about help (Katy said, ‘c’n I he:lp you,’).
There is a micropause (line 03) after Max answers the phone. The first
part of Jack’s response is similar to Darcy’s: a return greeting and
identification (‘He:llo Katy good morning t’you.=it’s Darcy:.’). However,
things start to go wrong immediately afterwards. Note the full stop after
‘Ocom.’, and an immediately latched next unit of subsequent talk:
‘=£How’re you doing this morning.£’.
This is very similar to what happens in Katy and Darcy’s conversation
– except that Katy, not Darcy, does the first ‘how-are-you’. Not only that,
Jack’s ‘how-are-you’ is delivered with premature ‘smile-voice’ –
remember, Katy only used ‘smile-voice’ once she knew who she was
talking to – and when she realised she knew the person she was talking to!
Jack’s turn is problematic because it is said in the wrong place with the
wrong intonation. It is this kind of turn that people can quite rightly
criticise as ‘filler’ or ‘non-genuine’. But don’t take my word for it. This is
Max’s analysis too, as we can see in his response. If Jack and Max knew
each other, like Katy and Darcy, then we would expect Max to reciprocate
a ‘how-are-you’.
Max answers the question, but does not ask Jack how he is. This does not
matter for Jack, who simply takes the next turn as if he has been asked a
‘how-are-you’! This is one of the strangest snippets of conversation I
31
have ever analysed, in which a salesperson embodies the problems of
cold-calling and scripted interaction so transparently.
Things go from awkward to worse as Jack tries, unsuccessfully, to
engage Max in conversation. We will return to the cold-callers in Chapter
7 when we take up the issue of ‘rapport’ and examine how (not) to build
it. We will think about how communication training to ‘build rapport’ is
something of an oxymoron; if you have to train for it, you probably do
not have it – as Max and Jack neatly show us . . .
Sometimes what happens in call openings can be, quite simply, a
matter of life and death. Example 15 comes from a famous conversation
analytic paper called ‘When words fail’.14 The authors analysed a call to
emergency services. The case was an early example of what is
commonplace now – the intense public scrutiny of a telephone call
between a member of the public and a member of an emergency service,
in which communication failure results in dramatic service failure.
The case in question took place in Dallas, Texas, in which a caller
ended up arguing for seven minutes with an ambulance dispatcher while
his stepmother lay dying. Here is the very start of the call.
Example 16: Emergency call
32
I will not reproduce the whole three-minute call here. Suffice to say that
the caller’s first request does not sufficiently convey urgency to the
dispatcher, who asks for more detail at line 06. The fact that the caller
treats the dispatcher’s question as problematic, and even offensive,
enables us to see how quickly an encounter can escalate into conflict. It
shows us, as the authors say, ‘how this sort of thing can happen’.
In a follow-up call, the caller’s roommate makes a different request:
‘Hello uh: I need an ambulance (for) someone that appears tuh (.) have
almost stopped breathing?’ However, by the time the ambulance arrives,
the outcome is fatal.
Example 17 is similarly dramatic. It is the start of a call between a
person in crisis, Kevin, threatening suicide, and a police negotiator, Steve.
Example 17: Crisis negotiation
33
You have probably noticed that this call opening is a failure. Our evidence
is that the person in crisis says only an ‘answering’ hello before hanging
up. He does not do a second ‘greeting’ hello and there are no ‘how-areyous’. What becomes clear, in analysing conversations from negotiation
to selling, is that ‘fake’ rapport and scripted talk is easy to spot and
almost always fails to do the job it is designed for.
Several components of Steve’s turn do not work: thanking Kevin for
putting his phone on, saying what he would like to do, saying what is
important to Kevin, and saying that Kevin is important to Steve. There is
also very little space for Kevin to say anything. However, Steve makes a
second attempt at opening a conversation with Kevin, which is successful.
It looks very different.
34
Example 18: Crisis negotiation
Steve’s second attempt erases much of the ‘rapport-building’ content we
saw in his earlier conversation. He also asks a ‘closed’ question – never
recommended in communication guidance – replacing his open-ended
request to talk in the first attempt. Yet despite being asked a ‘yes/no’
question, Kevin starts talking. Steve also focuses his question on action,
and what can be done, rather than talking things through.
By comparing Steve’s failed and successful attempts to get Kevin to
talk, we can identify what works. As we will see, throughout the book,
what works is often not what we probably think will work. A closed
question, without an attempt at rapport, is effective in this negotiation
with a suicidal person in crisis.
This is what I do as a conversation analyst. I collect lots of instances
of, say, negotiation openings, analysing each word by word, turn by turn.
I can then map different types of turns and patterns to different outcomes.
Outcomes may be built into the very encounter being studied. They
happen inside the encounter. A person says yes or no; buys something or
does not; gets an appointment or does not – commits suicide, or does not .
. . Talk is not trivial, easy or mundane. It is crucial.
35
The conversational racetrack
Conversations are encounters with a landscape, with a start and an end
like a racetrack. We start at the beginning with our recipient or recipients
and, along the way, complete various projects, like we saw Hyla and
Nancy, and Dana and Gordon do. We design and build openings with
summons and answers, greetings and identifications, and ‘how-are-yous’.
Think about the encounters you have with friends, partners, the
checkout person at the supermarket, your children’s schoolteacher, the
doctor, a first date. Each of these has a landscape with projects, or actions,
that comprise the complete encounter. Some actions will be the same, like
greetings, openings and closings. Others will be particular to the setting,
like diagnoses, flirts, storytelling, complaints, requests or instructions. We
may move smoothly along the racetrack from one project to the next, like
Katy and Darcy, or bump along the sides of the racetrack, on the rumble
strips, like Max and Jack.
Years of popular psychology have taught us to think about life and
behaviour in terms of gender, culture, personality and other variables. We
tend to think we are individuals who behave according to our intentions,
in idiosyncratic ways. This book will show you that we are far more
interesting than this. It will also show you that we are pushed and pulled
around by words, phrases, intonation – by talk – far more than we realise.
Conversation analysts study conversational racetracks of all kinds,
from people on first dates to police interrogations of suspects; from
doctors and patients in hospital to aeroplane cockpit communication. By
zooming in on the projects that comprise complete encounters, and by
analysing many instances of, say, a patient asking for an appointment, or a
police officer asking a suspect what they did, we can understand how
different ways of asking questions lead to different outcomes – patients
get an appointment or do not; suspects give full accounts or do not. This
book will guide you along many racetracks. You will see how to avoid an
argument like Gordon did; how to avoid sounding scripted or fake like
Jack did: how to use the ordinary resources of words to get things done.
The next chapter will take you deeper into conversation analysis. It
will introduce you to some of the technicalities involved in analysing talk.
So sit back, relax . . . here comes the science.
36
__________
1
Edwards (2012).
2
3
4
Albert et al. (2018).
Sacks et al. (1974).
An article in Nature entitled ‘The top 100 papers’, listing the most-cited research of
all time, identified papers in the ‘hard’ sciences with citations ranging from 305K to
12.2K, at the time of publication in 2014. The foundational piece by Sacks,
Schegloff and Jefferson currently has over 16K cites in March 2018; Schegloff has
76K citations; Jefferson 45K and Sacks 50K. For a relatively unknown discipline,
its impact is actually rather massive!
5
6
Jefferson (2004).
Listen to Hyla and Nancy (and other examples from the book) at
www.carmtraining.org/talk/extras.
Comparing mobile calls to landlines without caller ID, ‘answers to the summons of
a mobile call treat the summons as being personalised in giving information on the
caller, and allowing the answer to be designed accordingly’ (Arminen and
Leinonen, 2006).
7
8
Listen to Dad and Liz (and other examples from the book) at
www.carmtraining.org/talk/extras.
9 Meredith (2014).
10 Sacks (1975).
11 Meredith (2014).
12 Hepburn (2004).
13 Schegloff (2002).
14 Whalen, Zimmerman, and Whalen (1988).
37
2
HERE COMES THE SCIENCE
As talkers, we know – tacitly at least – pretty much everything there is to
know about talking. We know how to build a turn at talk. We know when
to take a turn and when someone else is likely to finish their turn. We
know how to keep the floor and what it means to interrupt. We know how
to fix a misspoken turn. And we know how to use the gaze of our eyes,
the position of our bodies, and the material environment to augment or
replace speech.
When we participate in conversation, we analyse everything our
interlocutors say so that we can take a turn. We are doing a kind of
38
analysis all the time, without being conscious of it. We assess constantly
what people say to us, mulling our assessments over privately (‘I wonder
what he meant by that . . .?’). We talk to our friends about what people
say to us and to them (‘I think she was just being friendly’). And we
challenge people directly (‘What did you mean by that?’). We can spot
when others fish for compliments or are disingenuous; we can tell people
that they missed someone being hypocritical or flirting with them.
But we do not know how talk works scientifically. And some of what
we think we know about talk is wrong. We will examine common myths
about talk more fully later in the book. One example is that silence does
not do anything. Or that silence is just what happens when our brains take
time to process what someone has just said to us. Both of these assertions
are far too simple. As talkers, we already know this. Our knowledge of
silence is built into idiomatic phrases like, ‘your silence speaks volumes’.
We say this to tell the person we are talking to that we are interpreting
their lack of response as problematic in some way. But we do not treat
silences – generally – as a processing delay.
Indeed, the fact that silence can indicate trouble ahead can be found in
both real and scripted talk. In The Duchess, a film based on a biography
of aristocrat Georgiana Cavendish, the Duchess is talking to Charles Grey
(of Earl Grey tea fame, and who was Prime Minister of Great Britain in
the early nineteenth century), with whom she is having an affair.
Georgiana is played by Keira Knightley, and Charles Grey by Dominic
Cooper. I have transcribed the performance from the film’s audio track.
So the silence at line 02 is how the scene was actually delivered.
Example 1: Scene from The Duchess
39
Just before we analyse those two seconds of silence, this is what the audio
track looks like when it is represented in the software editor I use to
transcribe the recording.
The sound wave forms peaks and troughs as words are uttered. The long
silence is easy to spot, though there is some background noise in the
second half of the silence. We can also hear Charles take a breath and
pause before he delivers the words of his turn, though this is hard to spot
simply by looking at the wave pattern.
The sound wave gives an indication of a speaker’s loudness, but not
their intonation or pitch. Some analysts find it difficult to ‘hear’ the pitch
of talk accurately (it helps if you are musical!) – does the pitch of the
final word of a turn rise or fall? Other software can reveal how speakers
shift pitch up and down as they talk, producing images like the one below,
of Charles’s turn, ‘You ought to know I do:.’
There are no extremes of pitch movement in his turn, but you should see
that it rises slightly when he says ‘know’ and falls at the end –
40
represented in the transcript by underlining on ‘know’ and a full stop at
the end of ‘do:.’.
Returning to the conversation itself, Georgiana asks her lover a
question. It is a ‘yes/no’ question (sometimes referred to as a ‘closed’
question), designed to get ‘yes’ or ‘no’ (at least!) in response. These
options are not equivalent. In this context, ‘yes’ is clearly a better
response.
However, before Charles responds with either ‘yes’ or ‘no’, what
happens next is . . . silence. After a two-second delay, he says, ‘You ought
to know I do.’ His response implies that Georgiana did not need to ask
this question. She should take for granted the fact that he thinks of her
when they are apart. Yet Georgiana’s next response, ‘You hesitated before
you replied’, highlights the negative implications of delaying a response.
The conversation analyst Emily Hofstetter uses a similar example
from the blockbuster film Frozen to show the conversation analytic skills
of one of the film’s main characters, Olaf the snowman. In the scene
Hofstetter describes, Princess Anna is talking to an iceman, Kristoff,
about her hair changing colour from brown to white, becoming frozen.
Example 2: Scene from Frozen, from EM does CA15
In Example 1, ‘yes’ is what conversation analysts refer to as a ‘preferred’
response; in Example 2, the ‘preferred response’ is ‘no’. This is a
structural, not a psychological, kind of preference – to do with the
systematic ways in which people respond to questions, invitations,
requests, offers and other conversational actions. We will explore the
technicalities of ‘preference’ later in the chapter.
41
For now, though, notice that, like Georgiana’s question, Anna’s
question is also a ‘yes/no’ question. Here, Anna is probably hoping for a
‘no’ – which should also be said immediately! And notice that it is only
after a two-second silence that Kristoff replies. At this point in the film,
Olaf pops up into the scene, to say to Kristoff, ‘You hesitated.’ Olaf’s
interpretation is the same as Georgiana’s.
Hofstetter uses this example to explain a number of technical features
of conversation analysis. We will also examine some of these later in the
chapter. But it is clear that each of these silences are doing something
other than allowing Charles, and Kristoff, time for cognitive processing.
Their delays predict something about their upcoming responses. Pointing
out the hesitations implies that Kristoff’s response, while it is the ‘right’
one, is less genuine or sincere. He says ‘no’ because it is right or
expected, not because he means it! And this suggests that genuine
responses happen rapidly.
The very fact that scriptwriters write these scenes – to be enlivened by
actors and to be comprehensible for the audience – tells us something
fundamental about the systematic nature of talk. It is no coincidence that
our examples above contain two-second delays. It is unlikely that the
actors were told to ‘pause for two seconds’. But their tacit knowledge of
how long a hesitation is – a clear, impossible-to-miss hesitation – has
resulted in almost identical performances.
Let us have a look at some unscripted dialogue, now, from the
television show, Location, Location, Location.16 In the programme, the
hosts are two property experts, Phil Spencer and Kirsty Allsopp. They
help house-hunters find a property to buy. Phil and Kirsty show the
house-hunters several properties in different locations during the course
of each episode.
In Example 3, Phil is asking a couple, Suzie and Andy, if they know
the town where the next property to be viewed is located.
Example 3: Location, Location, Location (simplified transcript)
42
Suzie and Andy respond rapidly and positively to Phil’s first question,
‘Do you know Hersham?’ with a series of ‘yeahs’. But a now-familiar
two-second delay follows Phil’s follow-up question, ‘Is that a good
thing?’
Now, this conversation represents just a tiny fragment of what it takes
to buy a house. But a ‘yes’ to Phil’s second question would indicate Suzie
and Andy’s positive disposition towards the town where the house is
located (never mind the house itself). In other words, a tiny ‘yes’ at this
moment would mean that the parties take one step further along the
conversational racetrack. Saying ‘no’ halts this progress.
So, like our earlier examples, a two-second delay tells us pretty much
all we need to know about the likely direction of a response – before it is
even uttered. Phil understands this too – he names the couple’s delay
(‘Silence!’) and makes explicit what saying nothing means (‘Not a good
thing!’). One word – ‘no’ – has big implications!
43
This chapter will examine the building blocks of talk using examples
from familiar television programmes and films, as well as from real-life
conversation in different settings, to explain how talk works. We will see
what it takes to make a conversation move forward, what the common
types of move in talk are, and how people do things with talk.
In the first chapter, we focused on how people do one type of thing –
greet each other. In this chapter, we will see how people do other things –
actions like inviting, requesting, complaining, interrupting – all using the
machinery of conversation.
Scriptwriters are conversation analysts
Olaf from Frozen might well be a conversation analyst but, of course, his
character is brought to life, not just by the magical powers of Queen Elsa,
but by those who write scripts for film and television. One of the best
sources of scripted interaction for helping us to understand the technology
of conversation analysis, and the science of talk, is the American sitcom
Friends.17 The programme followed the friendships, families, romantic
and career lives of six friends (including a brother and sister and two flatsharing pairs) living in Manhattan, who often hung out at their local
coffee shop, Central Perk. The programme ran for a decade between 1994
and 2004, with a huge audience and massive cultural impact (and plenty
of scholarly analysis).
Love or loathe the programme, the scripts from Friends work well to
help us understand the science of conversation. The scriptwriters
juxtapose spoken turns in ways that throw into sharp relief the
underpinning machinery of talk. Because a lot of the humour in Friends
depends on clever conversational devices – not set-piece jokes or oneliners – audience laughter can be used as another useful tool for pinning
down how talk works.18, 19, 20
Our first example comes from an episode in which two of the main
characters, Monica and Phoebe, are about to start a conversation. The
setting is the apartment Monica shares with another character, Rachel. In
the transcript, audience laughter is treated as a participant.
44
Example 4: Friends: ‘The One With Ross’s New Girlfriend’21
As the scene starts, Monica crosses the floor of her apartment to where
Phoebe is sitting. She says hello to Phoebe, who raises her head in
response. Monica asks Phoebe, ‘Y’know what I’m thinki:n’?’ What kind
of turn is this? Well, by dint of its grammatical design, as written, this is a
‘declarative’. Questioning intonation is used to turn declarative into an
‘interrogative’ – a ‘yes/no’ polar question.
In transformational grammar, there is a transformation called ‘dofronting’ that converts declaratives into interrogatives. However, in talk,
people are often elliptical; that is, they miss words out. So ‘Y’know what
I’m thinki:n’?’ is readily heard as an abbreviated version of ‘Do you
know what I’m thinking?’, restoring it as an interrogative.
In fact, this grammatical form is a vehicle for a particular function. It
is not solely after a ‘yes’ or a ‘no’ in response. Monica’s question is
designed to lay the foundation for something else to happen. It is what
conversation analysts call a ‘pre’ – a type of turn that sets up a future
activity. So a pre-announcement, or a pre-telling, such as Monica
produces, is designed to get a ‘go-ahead’ from Phoebe – something like,
‘no, what?’ If Phoebe gives her the go-ahead, Monica can report her
thoughts on some unspecified topic.
Setting the stage for a future conversational action happens regularly.
In a real-life telephone call below, Donny is calling Marsha.
Example 5: Donny and Marsha
45
Once they have said their hellos, and established that they know who each
other is, Donny says, ‘Guess what.’ This is a ‘pre’ action, like Monica’s
‘Y’know what I’m thinking?’ Donny and Monica both seek a go-ahead
from the person they are talking to.
Let us see what happens next, in both conversations.
In the ordinary telephone call, Donny gets a go-ahead from Marsha
(‘What.’). By doing this, she returns the floor to Donny, who then delivers
some bad news about his car.
But something different happens next in our Friends example.
46
If Phoebe had replied to Monica in the same way that Marsha replied to
Donny, with a ‘What?’, Monica would have her go-ahead. But Phoebe
responds to the form not the function of Monica’s turn, treating it literally
as an invitation to guess what she is thinking.
Evidence that this is an odd thing to do comes from what happens
next: the audience breaks into laughter.
If Phoebe had given Monica a go-ahead, most likely the audience would
not have laughed. So why does the audience laugh, at just this point?
There is no obvious sitcom joke. The humour is generated by what
happens at line 04; the audience recognises the correct function of
Monica’s ‘pre’ turn.
As the laughter subsides, Phoebe begins to respond to Monica.
The transcription symbols ‘>’ tell us that Phoebe slows down her
pace of talking, quite dramatically, as she searches for an answer to
Monica’s question, ‘You know what I’m thinking?’ The laughter that
follows Phoebe’s suggestion – that Monica is wondering if sex is still the
47
same – is about the ridiculousness of the hypothesis, and the obvious
tease: that it’s been so long since Monica had sex (the tease) that sex itself
might have changed in the interval (the ridiculous extreme)!
In Donny and Marsha’s conversation, Donny sets the scene for Marsha
to offer help, following his car breakdown. Even though the Friends
example is scripted, it turns out that Monica’s ‘pre’ turn similarly lays the
foundation for eliciting an offer from Phoebe to cut Monica’s hair.
Phoebe finally gives Monica the go-ahead, and now Monica can say what
was on her mind (‘Well, I was thinking . . .’). She tries to elicit an offer of
a haircut from Phoebe. At line 14, Phoebe shows that she knows where
Monica’s ‘pre’ turn was heading. While the audience does not laugh in
response to Phoebe’s smiley ‘Oh!!’, they do laugh after she delivers
something that is definitely not an offer!
Monica did not ask Phoebe to cut her hair directly (e.g., ‘Can you cut
my hair?’); she made an indirect, rather than direct, request. However,
Phoebe skips past this indirectness and turns Monica down baldly, with
no explanation. In another context, this might start an argument or be
treated as rude; here, the scriptwriters have produced a line designed for a
laugh. We will return to how people request and offer help in Chapter 6,
‘Are you being served?’
48
The building blocks of talk
In developing (realistic-ish22) dialogue for Friends, the scriptwriters show
us their implicit knowledge about how talk works. To start with, they
know that turns are organised into pairs. Monica’s first turn – ‘You know
what I’m thinking?’ – is the ‘first pair part’ of what conversation analysts
call an ‘adjacency pair’. An adjacency pair is a basic building block for
talking. It provides the foundation for constructing ‘sequences’ of activity.
After one speaker has produced a ‘first pair part’, their recipient is
expected to respond with a turn that delivers a second action. The second
action is paired with, and fitted to, the first one – for instance, ‘questionanswer’; ‘greeting-greeting’; ‘invitation-acceptance’; ‘request-offer’. The
second action is the ‘second pair part’.
Once the first pair part has been produced, almost anything produced
next comprises the second pair part. This includes silence and delay. The
second pair part may be inspected for its timeliness (‘your silence speaks
volumes!’), whether or not it is fitted properly to the first (‘I was asking
for a go-ahead, not to play a guessing game!’), and whether or not it helps
the overall sequence progress smoothly or stall and falter.
Think about the last time you invited someone to do something.
Without realising it explicitly, you analysed what your interlocutor said in
response to your invitation, how they said it, and when they said it. You
interpreted their response as an acceptance, a rejection, a stalling
manoeuvre, an account for non-acceptance, or whatever.
This way of understanding social interaction is rooted in a style of
sociology called ‘ethnomethodology’, from which conversation analysis
evolved. Ethnomethodology, invented by Harold Garfinkel,23 means the
‘study of people’s methods’. Garfinkel’s central concern was to make
common sense, or the ‘seen but unnoticed’ organisation of social life,
visible for academic scrutiny.
Think about how you walk around a museum. Your pace will be rather
different to that used in an airport. You notice if a small child runs in an
art gallery, and hope that their parents fulfil their obligations to socialise
their children. The instruction to ‘stop dawdling’ makes more sense on
the walk to school than in a museum.
49
One way to make these invisible but powerful ‘methods’ evident is to
examine what happens when tacit rules are breached. Garfinkel conducted
what he called ‘breaching experiments’. He instructed his students to talk
to their friends and ask them to ‘clarify the sense of commonplace
remarks’ (1967, p. 42). Example 6 is one student’s reported conversation.
Example 6: Garfinkel’s ‘breaching experiment’
In each turn, the student attempts to get their unwitting friend to make
explicit what would normally be understood without such explicitness.
Responses to students were often hostile, although they sometimes
thought the student was actually trying to be funny. Like our example
50
from Friends, then, when speakers do not respond in the way one expects,
a possible outcome is laughter.
Garfinkel’s breaching experiments allowed him to identify what it is
that gives ordinary scenes their ‘familiar, life-as-usual character’ (p. 37).
Taking this one step further, we would not be able to understand a
‘breach’ without understanding how systematic our everyday actions
actually are. This rule applies to how we communicate, too.
How to build a successful invitation
Almost every social action we might conceive of requires communication
– talk, writing, gesture, signing – to make it happen. Actions are
accomplished in turns, one after another. Turns are built from individual
words, formed into units. These units may be sentences, in the way we
traditionally understand grammar and composition – though they may not
be. And sequences of turns build conversational racetracks.
In conversation, people have to monitor each other’s turns for when it
might be relevant for them to take a turn. They have to figure out who
should talk next, and when – and how – to take a longer turn than one unit
of talk. Speakers have to understand ongoing talk and make a ‘fitted’ –
that is, apposite – response. They must deal with trouble in speaking,
hearing and understanding. Finally, and crucially, speakers build turns
from a massive range of lexical and grammatical possibilities. One word
can make a difference to the outcome of an encounter, as we will see later
in the book.
The idea that people just take turns, one after the other, may sound
simple. It is not. Think about what is required, word for word, turn by
turn, to build a successful invitation. It is straightforward, right?
Example 7: A basic invitation
51
To invite someone to go to the cinema, we need to build a turn of talk that
does the action of inviting. There are many ways to do this. In the
example above, Speaker A asks, ‘Want to go to the cinema?’ But other
combinations of words and phrases, with varying intonation and other
paralinguistic subtleties, will also deliver an invitation.
The design of an invitation – and any action – will depend on who you
are asking, what kind of relationship you have with your recipient, what is
at stake, how important an acceptance (or rejection) is, and so on.
Conversation analysts have shown that these sorts of contingencies
inform the design of the turn and are embedded in those designs.
If the person is your partner, for instance, just saying ‘Cinema . . .?’
may do the job. From ‘fancy a movie?’ to ‘I wondered if you’d like to
come with me to the cinema’ – there are multiple options.
In Example 7, B responds positively to A. You may have spotted,
however, that only lines 03 and 06 are included in the transcript – at least
four turns are missing. The invitation and the acceptance are present – the
first pair part and the second pair part. Together, two turns build a ‘base’
sequence – a base ‘adjacency pair’. The base completes the core activity –
the invitation – being done.
But much more can happen in what is still a simple action. Let us fill
in lines 01–02.
The meaning of ‘preference’, introduced at the start of this chapter, comes
alive in the first two lines of this encounter. We are not talking about
psychological or personal preference, though. Although we might think it
is obvious that the ‘preferred’ response to an invitation is acceptance, it is
not so obvious what the preferred response to a compliment is.
52
‘Preference’ refers to something structural about paired turns of talk.
They keep the action moving forward. Put simply, ‘preferred’ responses
take less conversational effort; ‘dispreferred’ responses take more.
If the preferred response to an invitation is acceptance, what are the
best ways to ensure a positive outcome? The answer is to do a ‘pre’. We
encountered ‘pre’ turns earlier in the chapter, when Monica and Donny
used them to get go-aheads from Phoebe and Marsha. Asking ‘what are
you up to this evening?’ is a pre-invitation. If B says they are busy, then A
knows issuing an invitation is likely to receive a rejection – a dispreferred
response.
People design their talk to help those they are talking to avoid having
to give such responses. Imagine how strange this conversation would be:
Returning to Example 7, A has established some grounds for going ahead
with an invitation, with a ‘pre-sequence’ comprised of another adjacency
pair of turns. But there are still two lines missing before B accepts the
invitation.
53
B is available to go to the cinema. But B does not accept A’s invitation
before establishing that there is a film they would actually like to watch.
B establishes this by initiating another adjacency pair of turns which
expands the base sequence. In technical terms, this pair of turns is
inserted between the base pair (lines 03–06).
Now we can see why B does not accept A’s invitation until line 06.
The sequence is nearly complete.
54
A and B take two final turns, expanding past the base sequence into a
‘post’ sequence, finalising arrangements for their cinema trip (lines 07–
08). Now the core activity, with its pre-invitation sequence, its two
expansions (inserted and post-sequences), is complete, across eight turns
at talk.
A neat illustration of the concepts we have been introduced to,
including ‘adjacency pairs’ and ‘preference’, is found in scripted dialogue
too. This next example from Friends is also about invitations. Rachel has
arrived in New York seeking her old friend Monica, after abandoning her
wedding and leaving her fiancé at the altar. Monica’s brother, Ross, also
newly single, is moving into a new apartment. Ross asks Rachel if she
wants to come to his apartment with two of the other friends, Joey and
Chandler, to help him assemble his new furniture. Ross’s ‘pre-invitation’
starts the sequence.
Example 8: Friends: ‘Pilot’24, 25
The core activity in this scene is Ross’s invitation to Rachel to help put
together his furniture. When Ross asks, ‘So Rachel what’re you uh:
what’re you up to tonight.’, we can recognise it easily as a ‘preinvitation’. It is laying the ground for a possible invitation, depending on
Rachel’s answer. If she is busy, Ross can avoid later rejection by not
issuing the invitation in the first place. Or, more technically, he can ensure
that the conversation avoids a failed base ‘first pair part’. So, what is
Rachel doing tonight?
55
What are Rachel’s options for responding to Ross’s ‘pre’ turn? She could
report that she has plans for the evening, blocking any forthcoming
invitation. In the script, Rachel gives a ‘go-ahead’ response (she is doing
‘hno:thin(h)g.’), but it is said ironically. She is doing nothing because she
is not on her planned honeymoon. Ross’s question now seems clumsy; he
has asked her what she is ‘up to’ on the night of her abandoned wedding,
while she sits in front of him in a wedding dress.
Having received Rachel’s, albeit ironic, go-ahead, Ross goes ahead
with his (pretty indirect) invitation.
Ross’s invitation takes into account Rachel’s earlier reply that she is
‘doing nothing’. In the first part of his invitation, he shows some
consideration for her circumstances: ‘If you don’t feel like being alone
tonight’.
Invitations can, of course, be accepted or declined. But these are not
symmetrical, equally valued alternatives. To keep talk flowing smoothly,
the preferred response to an invitation is to accept it. If Rachel accepted,
the course of action initiated by Ross would be successfully accomplished
– she would go to his apartment to help with the furniture.
But Rachel produces a dispreferred response – she says ‘no’. Look at
how many words it takes to say ‘no’. If you do not understand that
56
Rachel’s words add up to ‘no’, remedial work is needed on your
conversational skills. It is a good job you are reading this book!
Conversation analysts have shown that preferred responses are
generally short and unelaborated. Dispreferred responses are more
intricate. Dispreferred turns often occur after a delay (silence speaks
volumes . . .!). They may start with words like ‘well’. They may contain
an appreciation (e.g., for the invitation). They may contain pauses and
other signs of perturbation. They may also contain explanations, accounts,
or excuses for why the speaker cannot produce the preferred response.
Rachel’s response to Ross’s invitation, as a declination, contains all
the right features for saying ‘no’. She starts with ‘well’; her turn contains
an appreciation of the invitation (‘actually thanks:’), and it includes an
account for saying ‘no’ (‘it’s been a long day.’). Because it contains all
these features, Rachel has said ‘no’ in the regular way. There is no breach.
It is not funny. The audience does not laugh.
As the scene moves on, Joey and Chandler ask Phoebe if she would
like to help with the furniture. This time, when Phoebe says ‘no’, the
audience laughs. Why?
57
Joey makes an ‘Okay’ gesture
Joey asks Phoebe if she would like to help – it is a ‘yes/no’ question.
Saying ‘yes’ would mean giving the preferred response. It would most
likely be done immediately, without elaboration, and it probably would
not be funny. We can tell that Phoebe is going to say ‘no’ from the first
part of her response ‘↑Oh:: ↑I wish I could’. This looks like it will be the
start of a regular dispreferred turn, just like Rachel’s, showing an
appreciation of the invitation before turning it down.
It is the second part of Phoebe’s response that gets the audience
laughing. She says, ‘I don’t want to.’ Although this does provide an
explanation for saying ‘no’, it is not the sort of reason generally found in
declinations. The standard way to say ‘no’ is to say that you cannot accept
the invitation, not that you do not want to accept it. In ordinary
conversation, such a turn might start an argument (‘What’s the matter
with you?’).
Reasons for declining invitations are generally situational or
circumstantial, set against matters of volition or desire. But more simply,
note that ‘I wish I could’ is contradicted by ‘I don’t want to’. So instead
of providing a psychological disposition to accept, being thwarted by a
circumstantial reason for having to refuse, Phoebe states a desire (‘wish’)
to accept and contradicts it with a desire (‘want’) to reject. It is this bald
58
contrast that the audience finds funny, as well as the unadorned nature of
the rejection. Phoebe is flouting the whole normative business of
providing circumstantial reasons for rejecting invitations that one would
otherwise want to accept.
Alongside the audience’s laughter is Joey’s bemused, embodied ‘okay’
gesture. His response ends the sequence and is a light-hearted assessment
of ‘typical’ Phoebe – typical Phoebe who baldly said ‘no’ earlier, when
Monica asked her for a haircut. The scriptwriters often have Phoebe
making these kinds of conversational breaches.
By writing this scene, and by laughing at the place designed for
laughter, the scriptwriters, actors, audience and viewers show that they
know a lot about talk already. We might not use technical terms like ‘first
pair part’ or ‘preference organisation’ to describe talk, but we sense the
way it is organised. And we know that talk gets things done.
Interruptions, dating, and a roll-on deodorant
Dawn, Ella, Marie and Kate are friends (in real life; this is not television).
They are getting ready to go out to a nightclub.
Example 9: Real friends talking
Let us return to the start of this book, and this chapter, and to the idea that
we all talk. So we already know a lot about talk. But the purpose of this
book is to inform its readers about how talk works scientifically. A basic
observation from the inventors of conversation analysis was that, in order
to talk, one speaker speaks at a time and speakers change. We noted that
59
this statement was reproduced in Private Eye as an example of
pretentious academic nonsense.
In Example 9, one speaker speaks at a time and speakers change.
Dawn is watching the clock; Ella shows that she is on board with leaving
soon. This is a basic, two-turn sequence. They take turns; one at a time.
After two turns, the activity is over. No one else responds. Dawn does not
pursue a response from anyone else.
Next, Marie starts a new sequence. There is only one problem with
going out; she is already feeling sweaty.
Something happens at line 06 that we have not paid much attention to so
far. Two speakers talk at the same time. How does this happen? Is it a
mistake? An interruption? In fact, it is an example of collaboration
between friends.
Marie is searching for a word to complete her turn, that she is
‘sweating like a . . .’ Marie has not finished her turn: it is not complete in
any sense – grammatically (it stops mid-air); intonationally (she does not
sound like she has completed the turn) and in terms of the action being
done (making an analogy is not complete). The pause of four-tenths of a
second – which is still mid-turn – tells us that she is searching for a word.
We have some grounds to say that this kind of silence is a processing
silence. The right thing to do is wait for processing to happen.
So why does Kate start talking, when Marie is clearly not finished?
Note the point where the overlapping talk starts. Kate starts talking at the
point that she judges is enough time for Marie to find the right word to
complete her analogy – in this case, four-tenths of a second. She sees her
friend failing to come up with a word and so makes a suggestion: Marie is
‘sweating like a . . . man’. But just as Kate says ‘man’, Marie finds the
word she was looking for . . . ‘rapist’.
60
You are probably thinking that this is a rather offensive way to
complete the analogy. ‘Sweating like a pig’ is the common idiom.
Marie’s turn – offensive or not; in regular use or not – is a first pair
part. What kind of action is it doing? It could be a ‘pre’ – it could set the
scene for help with her sweatiness.
However, her friends’ response – each of them doing a second pair part in
overlap with one another – combines shock, horror and laughter. Their
responses show disapproval, but in a light-hearted way. Notice – while we
examine this conversation – that all four speakers talk at once. But this is
not hostile interruption. Speakers can and do regularly talk at the same
time, particularly when they are all working on the same action at the
same time.
In the next first pair part, which keeps the talk moving forward, Marie
says that she is ‘really hot’. At line 12, Ella closes this particular topic by
telling Marie (and probably Dawn) that they have to stop using ‘that
phrase’. We discover, here, that ‘sweating like a rapist’ is not a one-off
expression but has evolved in the friendship group. We have some insight
into the culture of the group.
Marie moves the talk on again, with another ‘first pair part’.
61
Marie’s turn is comprised of one unit of talk (called a ‘turn construction
unit’). It starts with, ‘↑Has ↑anyone-’, which we can see from the rest of
her turn is the start of a ‘yes/no’ question. No one can respond yet,
though. No action has been completed. Marie pauses briefly, then starts
again. These kinds of false starts and little glitches litter our talk.
Marie recycles the original start of the turn and then completes it. We
know Marie’s turn is complete for the following reasons. First, it ends
with falling intonation (indicated with a full stop; it is a myth that
questions always end with rising intonation). It completes an action
(asking for deodorant). It is grammatically complete. It is treated as
complete by Dawn, who takes the next turn. Finally, the end of the word
‘stuff’ is the first point that the turn is ‘possibly’ complete. This means
that changing speaker is now apposite; normatively permitted and
expected.
Dawn’s turn is the second pair part to the adjacency pair set up by Marie.
It shows us that Dawn has understood Marie’s first pair part correctly, as
both a question and a request for deodorant.
Dawn’s response is built from two parts. The first part, ‘Dave has’,
answers the question. It is a preferred turn – delivered without elaboration
or delay. However, Dawn adds a second part to her turn, ‘but you’ll smell
like a ma:n’. This qualifies her response and implicit offer of Dave’s
deodorant – which might not be what Marie is looking for. Each
component of Dawn’s turn, then, does a different, separate, action.
Let us see what happens next:
62
As Kate starts to laugh, Marie makes her request a second time, with
some subtle changes. The word ‘really’ (‘has anyone got any really non:
sweaty stuff’) is replaced with the word ‘fe:minine’ – with exaggerated
intonation. Marie’s request is responded to with an offer – which is the
‘preferred’ response.
This conversation gives us some technical insight and know-how into
the organisation of turn-taking. We can identify points where it is possible
for speakers to take turns, but also talk at the same time.
However, overlapping talk can also be treated as an interruption and
thus a violation of one speaker’s rights to take a turn. In the clip below, a
police officer is interviewing a teenage suspect who has been arrested on
suspicion of criminal damage. Our interest is that, although there are
actually no moments of overlapping talk, the police officer treats the
suspect as though she had interrupted her.
Example 10: Police interview
63
The transcription of the police officer’s opening question is important. It
ends with a comma. This represents ‘continuing’ intonation which, you
may remember, sounds like reading a shopping list of items, or the chunks
of a telephone number. Halfway through, our intonation is higher than
when we come to the end. By the end of line 02, the police officer’s turn
is not complete in terms of grammar, intonation, or in terms of finishing
an action that the suspect could respond to.
The four-tenths-of-a-second gap on line 03 is, coincidentally, the same
length of time that Kate waited for Marie to come up with the word
‘rapist’. However, in the friends’ talk, Kate tried to help Marie out to
complete her turn. Here, the suspect jumps in with an answer to an
incomplete question. She does not help the officer out by completing her
question. The suspect’s next turn is also not complete, and the police
officer comes back into the conversation when the suspect runs out of
something to say.
Any notion you might have that power is attributed straightforwardly
to police officers and not suspects is undermined by the suspect’s
response to the officer’s ‘yes/no’ question, ‘Are y’gonna let me finish
64
what I’m sayin’.’. The suspect does not say ‘yes’ or ‘no’. She challenges
the very asking of the question – defending the fact that she started to
talk. The little circles around her turn indicate that she is speaking quietly,
suggesting that the suspect will not risk a more aggressive rebuttal.
This episode tells us that it doesn’t take an overlap, or clash of voices,
to constitute an interruption. The suspect interrupts the police officer’s
building of an action but does not talk over her.
Sometimes overlapping talk is the conversational problem. The next
scene is again from Friends. The storyline focuses on Monica, who wants
a baby but has no partner. She has decided to visit a sperm bank. Her
brother, Ross, is trying to discourage her.
Example 11: Friends: ‘The One with the Jam’27
65
Monica gestures that her lips are still moving
Monica has not finished her turn when Ross starts talking in overlap. This
is treated as a violation of turn-taking by Monica, who stops what she is
saying to chastise Ross: ‘↑lips moving, ↑still talking.’ This is a sitcom,
and so the audience laughs. In ordinary talk, the parties might laugh – or
might start arguing.
To generate this scene, Ross must begin speaking at a point where
Monica has obviously not finished her turn. What is the tacit knowledge
needed to write and act the scene? The scriptwriters, and the actors, need
to know several things about turn-taking and how to do an interruption.
When Ross starts talking, Monica is in the midst of an incomplete turn –
she is saying ‘to do . . .’ This is grammatically incomplete. It is
incomplete in terms of its action. And it is incomplete in terms of its
intonation. The first point of possible completion arrives on the word
‘something’, when the sentence is complete, Monica’s pitch drops, and an
action is accomplished.
Conversation analysts have found that speakers are generally entitled
(by each other) to one unit in a turn. For Monica, one unit would end with
the word ‘something’. However, she tries to produce a multi-unit turn,
which means keeping the floor. If you have ever wondered what not
66
‘getting a word in edgeways’ looks like, it is the moment in the transcript
where Monica latches a second unit immediately on to the first (‘=