Surprise as An Interactional Achievement by Wilkinson and Kitzinger

Surprise As an Interactional Achievement: Reaction Tokens in Conversation*

Loughborough University
University of York

The expression of surprise—at something unexpected—is a key form of emotional dis- play. Focusing on displays of surprise performed by means of reaction tokens (akin to Goffman’s “response cries”), such as wow, gosh, oh my god, ooh!, phew, we use an eth- nomethodological, conversation-analytic approach to analyze surprise in talk-in-inter- action. Our key contribution is to detach the psychology of surprise from its social expression by showing how co-conversationalists collaborate to bring off an interac- tionally achieved performance of surprise. Far from being a visceral eruption of emo- tion, the production of a surprise token is often prepared for several turns in advance. We also show how surprise can be recycled on an occasion subsequent to its initial pro- duction, and how surprise displays may be delayed by silence, ritualized disbelief, and other repair initiations. Finally, we consider some of the uses of surprise as an interac- tional resource, including its role in the reflection and reproduction of culture.

Surprise—the emotion experienced when encountering “unexpectedness” (Reizenzein 2000) or “expectancy violations” (Scherer, Zentner, and Stern 2004)—is com- monly regarded as a fundamental human emotion (Ortony and Turner 1990; Strongman 2003). Like other “basic” or “pri- mary” emotions—happiness, sadness, fear, anger, and disgust—it is typically inferred from characteristic facial expressions, bodily postures, and vocalizations (Tomkins 1962). Raised eyebrows, open mouth, upflung hands, and gasps and exclamations are com-
* Some of the material included here was present- ed at the 53rd annual conference of the International Communication Association, held in San Diego in May 2003. This paper was written while the first author was at Simon Fraser University, Burnaby, British Columbia. It developed from a broader pro- ject on reaction tokens carried out by the first author under the guidance of Emanuel A. Schegloff, to whom both authors are very grateful for his support, encour- agement, and exemplary teaching. We would like to thank Paul Drew, Gene Lerner, Richard Ogden, Emanuel A. Schegloff, and Anthony J. Wootton for their helpful comments on earlier versions. Cor- respondence should be sent to Sue Wilkinson, Department of Social Sciences, Loughborough University, Loughborough, Leicestershire LE11 3TU, UK;
mon expressions or displays of surprise (Plutchik 1980).
Social psychologists typically have theo- rized such displays with reference to one or the other of two distinctive frameworks, sometimes called “organismic” and “interac- tional” (Hochschild 1979). In organismic approaches, rooted in Darwin’s ([1892]/1998) classic evolutionary theory, emotion is under- stood as fundamentally biological, and dis- plays of emotion are seen as inadvertent manifestations of individual bodily processes. For Darwin, emotions are innate and instinc- tual physiological responses with functional significance; emotion displays are designed in the first instance not to communicate emo- tions but to serve adaptive purposes. According to Darwin, expressions of sur- prise, graduating into “astonishment” and thence to “stupefied amazement” (p. 178), originate in the biology of the “startle” response (although some subsequent researchers have argued that the startle response is more reflex than emotion: see Ekman, Friesen, and Simons 1985). In Darwin’s classic account, in the face of an unexpected event or a violated expectation, raised eyebrows enable “the eyes [to] be opened quickly and widely” (p. 278) and the
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
“open mouth of a man stupefied with amaze- ment” (p. 284) is “so as to draw a deep and rapid inspiration” (p. 97). The vocalizations characteristic of surprise (“a deep oh” or “a blowing, hissing, or whistling noise” such as “whew” (pp. 285–86) are a consequence of position of the mouth in interaction with the timing of expiration (p. 97). Organismic approaches, then, conceptualize emotion as “unbidden and uncontrollable” (Hochschild 1979:551); emotion displays are viewed as “visceral” eruptions, the involuntary over- flowing of internal states. Emotional expres- sion is not considered to be intentional or primarily communicative (although a “signal- ing” function is clearly adaptive in evolution- ary terms). Rather, it is a by-product of physiological responses within the individual.
By contrast, interactional approaches, such as social constructionism (Averill 1982; Harré 1986), ethnomethodology (Garfinkel 1967), and discursive psychology (Edwards 1997; Potter 1996), conceptualize emotion as fundamentally social or socially constructed rather than biological, and dis- plays of emotion are viewed as intentional communications rather than as involuntary exusions. From an interactional perspective, “[e]motions are not something which just happen to an individual; rather, they are acts which a person performs” (Averill 1974:182). Interactional approaches, while not denying individual experience, focus on the intrinsi- cally communicative function of emotion, particularly the interactional uses and inter- personal “management” of emotion displays (Cahill and Eggleston 1994; Hochschild 1983; Parkinson 1991). Work in this tradition focus- es on the way in which the expression of emo- tion is interactionally organized and attuned to others in the social world: for example, how facial expressions of elation or disap- pointment are directed to an audience rather than occurring in direct response to a precip- itating event (Kraut and Johnston 1979), or the way in which displays of empathic pain are inserted into an interaction sequence (Heath 1989). The emotional expression of surprise may be used as a “social signal” in play between mothers and infants: mothers were found to exclaim at a higher pitch when their children did not show a surprise facial expression at a jack-in-the-box than when
they did so (Reissland, Shepherd, and Cowie 2002). Interactional approaches, then, con- ceptualize emotion as “social through and through” (Parkinson 1996:672), and emotion displays are regarded as primarily commu- nicative.
Whereas the organismic approach seeks to uncover the physiological underpinnings of emotional expression, the interactional approach—most markedly from the social constructionist perspective—focuses instead on why and how (Western, modern) people conceptualize emotion in organismic terms, and perform it accordingly. Averill (1974) suggests that we understand emotions as vis- ceral because of centuries of “psychophysio- logical symbolism,” associating emotions with primitive, animal-like parts of the ner- vous system: the “visceral brain” and the guts. This, in turn, has informed the prevailing dis- course of emotion: “[w]e are ‘gripped,’ ‘seized,’ and ‘torn’ by emotion; we act ‘uncon- trollably’” (p. 152). As Harré (1986:5) puts it, the notion of an emotion as “a response suffered by a passive participant in some emotive event” is itself one of “the social strategies by which emotions and emotion declarations are used by people in certain interactions.”
The research reported here contributes to the interactional tradition, building specif- ically upon the work of Goffman (1978). Goffman theorized the interactional uses of the visceral understanding of emotion in his important paper on “response cries”: “exclamatory imprecations” (p. 798), which function as “exuded expressions, not inten- tionally sent messages” (p. 800). Akin to (some kinds of) emotional display, response cries are “a form of behavior whose very meaning is that it is something blurted out, something that has escaped control” (p. 799): they include pain cries (ow, ouch), “spill cries” (oops, whoops), revulsion sounds (eeuw), surprise sounds (eek, yipe), and lexi- calized items drawn from religion (hell, heav- ens) and taboo domains of bodily function (shit, fuck).
According to Goffman, social members understand response cries as visceral erup- tions of spontaneous emotions: “a natural overflowing, a flooding up of previously con- tained feeling, a 
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
straints, a case of being caught off-guard” (p. 800). Regardless of their lexical (or para- lexical) identities, what these response cries have in common is that they “externalize a presumed inward state” (Goffman 1978:794) and convey the sense of having been blurted out spontaneously, the involuntary exuding of a psychological state rather than an inten- tional piece of communication. Goffman claimed, but did not provide systematic data analysis to demonstrate, that these apparent- ly “blurted out” imprecations are fundamen- tally interactional events, “creatures of social situations” (p. 814). He made various pro- vocative suggestions as to how response cries might work socially, but without analyzing any actual instances of their use.
In this article we use a conversation- analytic approach to investigate a subset of response cries—we call these “reaction tokens”1—used analyzably to perform sur- prise, as they occur in naturalistically collect- ed data. We will refer to these as “surprise tokens,” or, when making analytic points that locate them in the field of reaction tokens more generally, as “(surprise) reaction tokens.”
The research presented here is part of a larger study in progress, in which we are exploring the interactional use of reaction tokens based on a recorded corpus of (so far) around 600 instances in English-language talk.2 We build on previous conversation-
analytic work (in particular, Freese and Maynard 1998; C. Goodwin 1996; Goodwin and Goodwin 1987, 2000): instead of treating reaction tokens as eruptions or leakages of internal, individual, physiological, or psycho- logical states, we explore them as interaction- al strategies in social contexts. Our aim is to rescue reaction tokens from the “implacable familiarity” (Schegloff, cited in Heritage 2003:17) of mundane everyday understand- ings, to situate them firmly within social action, and to analyze their deployment as specifically sociological phenomena.
The key contribution of this article is to detach the psychology of surprise (the emo- tional experience of encountering the unex- pected) from the social expression of surprise (the public display of finding something counter to expectation). We show empirically that expressions of surprise conveyed through surprise tokens (o:::h!, wow, golly, and so on) are not involuntary spontaneous emotional eruptions but interactionally orga- nized performances: that surprise is an inter- actional achievement. We also examine, particularly in the final section, what is accomplished socially by displays of surprise: that surprise is also an interactional resource. We show that actions accomplished by the performance (or withholding) of surprise include the reflection and reproduction of culture; the production and reinscription of membership categories; affiliation and disaf- filiation; and management of the local moral order. In these ways, then, we contribute to the interactional account of emotion, specifi- cally surprise, as a fundamentally social phe- nomenon.
People display surprise in many ways other than producing surprise tokens, includ- ing prosodic marking on questions and repeats of prior turns (Jefferson 1972; Selting 1996), facial expression (Ekman 1992), and
1 The term reaction tokens is intended to capture their reactive nature (some reaction or response is indelibly intricated into them); their bounded format (they are units that are not infiltrated into a turn in the way that laughter or other prosodic features may be, and are not grammatically part of it); and their minimal quality in that (unlike, for example, other dis- plays of surprise such as “Are you serious?!” or “He didn’t!”) they do not function as initiating actions, which make relevant further talk from the producer of the surprise source turn. The choice of term also differentiates our collection from Goffman’s (1978) somewhat differently defined version of the phenom- enon. Goffman also included, for example, filled paus- es (such as “uh,” “um”) in his collection of “response cries”; these are not included in our collection of “reaction tokens.”
2 The data from which these instances are drawn are Kitzinger’s Birth Crisis Calls (BCC) and Home Birth (HB) corpora; Wilkinson’s Breast Cancer Patients (BCP) corpus; some ad hoc personal record- ings; and the immense corpus of British and American audio and video recordings now available to
researchers in conversation analysis, transcribed pri- marily by Gail Jefferson. When the audio data have been available to us, however, we have undertaken some retranscription. We are particularly grateful to Doug Maynard and to Victoria Land for making avail- able to us the audio data for fragments 2/31; and 1, 19 and 20, respectively.

Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
gesture and body deployment (C. Goodwin 2000). Nonetheless, people can and do dis- play surprise, and are understood by co-inter- actants to be doing so, even when they are not visible to one another (as in the tele- phone conversations that constitute the majority of our data), using only surprise tokens. Although we are not claiming, then, to be documenting the range and variety of ways in which human beings display surprise to one another, we provide detailed analysis of one such means of display.
We did not stipulate in advance any defi- nitional features of surprise tokens except that they should be used analyzably to per- form surprise: that is, to display that some prior talk or event in the world3 is unexpect- ed or counter to expectation. The warrant that this is so is based on the interactional data analyses that constitute most of this arti- cle. It so happens that a wide array of differ- ent reaction tokens was used in our data set to display surprise, including wow, gee, gosh, jesus christ, my goodness, oh my word, oo:h!, oh:!, good gracious, oh my god, oh shit, blimey, and nonlexical tokens such as whis- tles and gasps. The lexical (and paralexical) items used to perform surprise also can be, and are, used to perform very different reac- tions. In our data set, for example, (oh) (my) god is analyzably deployed to perform, in addition to surprise (see fragments 12, 18, 19, and 24), disgust (at bags of rotten potatoes)
3 In this article we discuss only reaction tokens responsive to preceding talk, and not those responsive to events in the world, because the former constitute the majority of instances in our data. Two examples of the latter are shown below. In the first, Barbara’s my goodness is a surprise token responding to the tactile sensations of picking up a prosthetic breast for the
[RT193: Wilkinson: BCP2: 7:6-8]
and sympathetic dismay (about a dying moth- er). Another way of putting this is to say that reaction tokens are both context-free and context-dependent:4 that is, a single reaction token can be used flexibly across many differ- ent contexts. The particular reaction it per- forms on any given occasion depends upon the deployment and calibration of prosodic features and upon its local sequential context. Specific reaction tokens are selected in part with reference to vernacular poetics (Jefferson 1996; Schegloff 2002) and in part to display both the extent of the surprise (from mild puzzlement to deep shock) and its valence (from very positive to very negative), such that particular combinations of lexical and prosodic usage convey anything from awed amazement to horrified disbelief, from delighted astonishment to carefully neutral surprise (Bollinger 1985: 48).
In English, intonation makes an indepen- dent contribution to the meaning of an utter- ance. Early work focusing on the prosody of surprise suggested that pitch register, pitch movement, and relative volume were charac- teristic of “surprised” utterances, especially a rise-fall (Roach 1983:119) or rise-fall-rise (Bollinger 1989:286) intonational contour. Evidence in support of these early (and many subsequent) claims, however, has relied on analysts’ intuitions, remembered overhear- ings, and informants’ judgments about script- ed dialogues in laboratory settings (for
first time (she has recently undergone a mastectomy). In the second, Loretta’s oh my gosh responds to events she is watching on television, which are not accessible to her recipient. The latter, as established in the immediately preceding sequence, is not watching Daktari on TV.) For the key to the symbols used in transcription, see the appendix.
01 Bar: My goodness it feels so ni:ce. °huh-huh 02 huh huh huh° .hhh $ It even feels wa:rm. .h-huh!
[RT367 Trio III]
01 Lor: Oh my gosh Officer Henry is (.) ul-locked in the 02 ca:ge wi- (0.3) with a lion.
Surprise tokens, and other forms of reaction token, are also very common in reported speech. These are the focus of a forthcoming analysis and are not dis- cussed here.
4See Sacks, Schegloff, and Jefferson (1974:699, Note 8) for a discussion of this phenomenon in rela- tion to the turn-taking organization for conversation.
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
example, Levis 2002). We know of only two phonetic analyses of surprise based on empirical analysis of actual talk-in-interac- tion (Local 1996; Selting 1996; also see Freese and Maynard 1998, whose analysis of the prosodic features of news delivery includes surprise).
In Selting’s analysis of German conver- sation, repair initiations (such as was [what] and bitte [pardon]) were found to be differen- tiated prosodically between those under- stood by recipients to convey problems of hearing or understanding (as displayed through speakers’ repetition of prior turn), and those understood by recipients to convey “surprise” or “astonishment” (as displayed, for example, through speakers’ accounting for what is thereby treated as an unusual cir- cumstance that runs counter to expecta- tions). The latter are characterized by increased pitch and extra loudness in com- parison to surrounding units (Selting 1996).
In Local’s analysis of British and American English conversation, oh-tokens likewise were found to be differentiated prosodically between those understood as straightforward “news receipts” (thus “not necessarily associated with the degree to which an answer is unexpected,” Heritage 1984:309) and those which functioned as “surprised” receipts (Local 1996:202), signal- ing the unexpectedness of the news imparted. On the basis of his analysis of surprise ohs, Local identifies particular pitch configura- tions (high, wide-range, rising-falling) charac- teristic of displays of surprise, but cautions against “a simplistic assignment of meaning to pitch contours independent of the interac- tional, lexical and grammatical environments in which they occur” (Local 1996:206).
The contribution of these two studies to linguistics is to show that prosody does not operate independently of its local interac- tional, lexical, and sequential environment. Their contribution to conversation analysis (CA) is to describe some of the features of
prosodic marking without which some of the ways in which speakers perform surprise would not be hearable as such, even with their local environment otherwise fully accounted for. Nonlexical reaction tokens (such as oh ), are constituted as surprise reac- tion tokens largely through being “punched up” prosodically (represented in convention- al CA transcription notations as, for example, O↑:::h!).5
Our collection of surprise tokens, then, consists of that subset of items, from our larg- er collection of reaction tokens, which regis- ter the unexpectedness of information conveyed in a prior turn at talk (whether the unexpectedness is valenced positively, neu- trally, or negatively). As such, surprise tokens are differentiated from:
Reaction tokens performing emotions other than surprise (such as disgust, pleasure, or sympathy; see Wilkinson and Kitzinger forthcoming a);
“Oh” as a simple news receipt (see Heritage 1984; Local 1996);
Other vocalized ways of conveying sur- prise that are not “tokens”: for example, newsmarks and news receipts initiating new sequences (“Did he really?!,” “You’re kid- ding,” Heritage 1984; Jefferson 1981; Maynard 1997; also discussed in the section on “ritualized disbelief” below); claims to be surprised (“I am surprised at that”); and assessments of events as surprising (“That’s amazing,” see fragment 3 below). As we will see, these other methods of conveying sur- prise often occur in the same environment as surprise tokens, and to some extent shade
5 Participants’ orientation to the prosodic dis- tinction between news receipts and surprise tokens can be seen in Lesley’s report of reactions to her announcement that a burglary suspect had “looked suspiciously at [her] dining room.” She reports the classic facial expression of surprise (their “mouths all dropped open”), repairing the prosody on “oh” (line 3) so as to transform it from a news receipt to the surprise token more fitted to the open-mouthed reaction.
[RT202N: Holt X(C)2-1-6]
01 Les:
[When ↑↑I walked in that post office ‘n-:-: said oh ‘e looked suspiciously at my dining room
05 Les: An’ they said oh-
↑Oh: were ↑you bur:gled, too:?hh
their mouths all dropped open.
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
into them at the edges of our collection. Nonetheless, as we will show, these tokens constitute a distinctive set of practices that perform surprise (instead of merely claiming it) and do so in an as-if-visceral way.
Surprise tokens are commonly produced after some talk (typically a news telling,
Fragment 1: Pregnant Again
announcement, or informing) designed to elicit surprise from the recipient. Not all news is designed by its teller to elicit surprise. There is a clear distinction between talking so as to be heard as simply imparting new infor- mation (as in fragment 1 below) and talking so as to be heard as imparting surprising information: that is, new information counter to expectation (as in fragment 2 below).
[Land: YU9: 30:48]
01  Pau:  So:: .hh But- oh Alison’s pregnant.  Again.
02  Chl:  Oh ri:ght.  Yeah.
03  Pau:  Which I heard today as well when I popped into (   )
05  Chl:
07  Pau:  Four. (.) Four yeah four.
08  Chl:  yeah
09  Pau:  So: uh four year[s.]
10  Chl:                  [Oh] right.  Not as big as I thought.  Yeah.
11        [Yeah.]
12  Pau:  [And  ] um Sally’s just had a baby girl. [And ]
an’ she’s like (going) [(    )]
                       [.hh So]there’ll be quite a big gap
13 Chl: 14

betwee:n (.) them.
                                         [.hhh] They’re
sproutin’ everywhere. My colleague at work her sister’s
just had one.
In fragment 1, the teller (Paul) imparts two pieces of information marked as news (something only “heard today,” line 3; some- thing that has “just” happened, line 12), but not as surprising in any way. Surprise involves finding some object or event contrary to expectation. The turn-terminal “again” (line 1) treats the news of Alison’s pregnancy as rather to be expected (even though, it seems, it is four years since her last pregnancy); Chloe receipts the news with an “oh right” (a simple news receipt) rather than with any expression of surprise. Even more markedly, Chloe treats the further news that “Sally’s just had a baby girl” (line 12) as utterly unsur- prising; she invokes a context in which every-
Fragment 2: Speaking of Bottoms
one apparently is having babies, rendering Sally’s (and Alison’s) doing so entirely to be expected. Paul and Chloe are aligned in not being surprised, thereby producing them- selves as co-inhabitants of a world in which pregnancies and births are commonplace (even though the only further instance Chloe offers is at some remove: the sister of a col- league, line 14).
By contrast, fragment 2 (also see the ear- lier analysis by Maynard 1997) represents an interaction in which a speaker designs her talk to elicit surprise, and her recipient duly produces a surprise token: oh my goodness (line 9, arrowed).
[RT513: PND3:18 (from Maynard 1997)]
01  Andi:   .hhh well: (.) speaking of bottoms are you sitting
02          dow:n.
03  Betty:  Ye:ah.
04  Andi:   Well (.) we have some news for you:.
05  Betty:  What.
06  Andi:   .hhh th’t (.) may come as $ a bit of a surpri:se. ehhh!
07  Betty:  I see- $what are you telling me.$=
08  Andi:   =hhhh! Bob and I are gunna have a baby.
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
09 Betty: <°O:h my: go:od↑ness°> hho- (0.5) did you have a 10 reverse- he have a reversal?
11 Andi: Yeah.

Here, Andi clearly designs her telling as a surprise source: first via the elaborate pre- announcement heralding “news” (line 4) that might possibly (it is implied) cause her recipi- ent to faint, or at least go weak at the knees, with the shock of it (“are you sitting down,” line 1). After a go-ahead (“what,” line 5), which specifically invites the projected telling, she postpones telling the news by producing an increment to her prior turn, specifically labeling the news as “a bit of a surprise.” Immediately after the news announcement (“Bob and I are gunna have a baby,” line 8) Betty produces her surprise token, oh my goodness, and immediately inquires about precisely that aspect of the announcement which generates the “surprise”: Bob, it seems, is known by Betty to have had a vasectomy (lines 9–10). Andi’s surprise source turn is designed precisely for a recipient with some prior knowledge (that Bob has had a vasecto- my) but without other knowledge (that he subsequently has had it reversed). In the design of surprise source turns, speakers dis- play what they expect their recipient already to know and take for granted, as well as what they do not expect them to know and hence what will come as a surprise. This may be knowledge specific to an individual (as in fragment 2), or more general cultural or sub- cultural knowledge (as in many of the exam- ples that follow).
Fragment 2 is typical of the surprise exchanges in our data in the following ways. First, as in the majority of cases in our collec- tion, the surprise token is responsive to an interactant’s prior talk, and not to direct sen- sory impressions of the world (see notes 3 and 7 for instances of the latter). Second, two peo- ple collaborate to bring off the production of surprise: one produces something surprising; the other produces the surprise reaction. A more formal way of putting this is to say that the turn containing the surprise source and the turn in which the surprise token is pro- duced are in a structurally “preferred” rela- tionship (Sacks [1973]/1987, 1995) to one another. That is, the surprise source turn is analyzably designed precisely to elicit the sur-
prise duly performed in next turn, and the action of a speaker producing a surprise token aligns with the action of the prior speaker whose talk was designed to elicit it.
This, at least, is the normative form of these interactions and also numerically the most common in our data. It contrasts with instances in which speakers react with sur- prise tokens to prior turns at talk not analyz- ably designed to elicit surprise (see fragment 9, below) and with instances in which speak- ers withhold surprise tokens to prior talk ana- lyzably designed to elicit them (see fragment 10 below). Both of these latter instances are “dispreferred” in the structural sense of dis- playing a misalignment between speakers (Pomerantz 1984), and thus are nonnormative (as well as relatively rare in our collection).
Surprise Source Turns
In each of fragments 3 through 8 (below), speakers produce surprise tokens in response to prior talk analyzably designed to elicit sur- prise. In each instance, a surprise source turn informs a recipient about some event or expe- rience, designing that telling so as to highlight the contrast between what might be expected and what actually turned out to be the case. In each instance, the recipient responds with a surprise token that embodies, and claims to share, the prior speaker’s conveyed under- standing that something unusual or counter to expectation has been described.
As would be expected on the basis of the well-established finding that turns in a struc- turally preferred relationship are contiguous with one another (Sacks [1973]/1987), these surprise tokens are produced without delay, usually immediately following the projectable end of the surprise source turn (as in frag- ments 3–6) or with slightly early onset, in overlap with it (as in fragment 7). (We have marked the surprise token line with an arrow in each case.)
In fragment 3 (from a call to a birth crisis helpline), the call-taker’s ooh! (line 3) is well fitted to Eve’s account of her labor at this point in her telling.
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
Fragment 3: No Pain
[RT114N: Kitzinger BCC 103:11]
01  Eve:  I w- got to six centimeters and I hadn’t had no
03  Clt:
04  Eve:
05  Clt:
06  Eve:
pain at a::ll.
Oo::[h! ]

    [U:m] [a:nd ]
          [That’]s amazin[g.]
[ I] know.
Eve has designed this part of her telling to embody a stance towards having been six centimeters dilated and pain-free. She is not simply recounting a fact, but treating it as sur- prising. As a negative observation, her report treats the absence of pain as a noticeable departure from what might otherwise be expected (Schegloff 1988). “At all” is an extreme case formulation” (Pomerantz 1986), produced with exaggerated emphasis which further underscores the absence of (even mild) pain. The call-taker treats the sit- uation Eve describes as “amazing,” across two separate units: her surprise reaction (Oo::h!, line 3) and her assessment (“That’s amazing,” line 5). At line 6 Eve confirms that
Fragment 4: All-Night Party
indeed she had intended to convey how “amazing” her experience was, and “knows” ittobeso.
Negative observations, and extreme case formulations are common components of turns treated as surprise sources. Other examples of extreme case formulations in turns treated as surprise sources by recipients (via their production of surprise tokens) include “extreme case proportional formula- tions” (Pomerantz 1986:228), such as “all” (fragment 4) and “only” (fragments 5 and 6) as well as intensifiers such as “just” (fragment 24) and “very” (fragment 27); also see “even” (fragment 22).
[RT62: Holt: Sept-Oct 88:1:11]
01 Les: [Oh:] I I say Gordon’s jus’
02 come home he’s bin itta party all night.
03 Mum: O[h: good gracious=Has he got a fat head

Fragment 5: Horse Hair Factory
[RT327: Holt:X(C)1:1:1:7]
Les: He wz a (0.2) .p a ↑buyer for the hoh- i- the ↑only horse hair fact’ry left in England.
06  Les:  And he wz their buyer,
Mum: Good gracious. (0.3)
Fragment 6: Twins
[RT584: Kitzinger BCC 217]
01 Clt: I only learnt six weeks before when I was having twins.
02 Jen: °°Oh my goodness!hhh°°
03 Clt: And d’you kno:w (.) what my m- what my doctor
04 said.
05 Jen: Wha(h)(h)t.
06 Clt: A- She was taking an awfully long time examining me

Extreme case formulations also are constructed recurrently with numerical values, displaying, for example, duration
(fragment 6), age (fragment 7), and weight and size (fragment 8). Notice, too, in frag- ment 7, the pre-announcement that Pat
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
uses in line 1: it seems to rely on Meg’s presumed inability to deduce Pat’s age by looking at her, and hence adumbrates a
Fragment 7: Seventy-Five
[RT194: Wilkinson BCP8:32:46-33:08]
01  Pat:  =D’you know old I a:m.
02        (.)
03  Meg:  No:: I’ve no id[ea. ]
contrast between appearance and reality which constitutes part of the surprise source.
04  Pat:
05  Meg:
07  Pat:
08  Wyn:
09  Meg:  m[m!]
10  Pat:   [ B]ut I don’t fee:l it.
Fragment 8: Ten-Pound Baby
[$I’m] seventy fi:[ve.$]
[G o:]shh!

huhhuh[°huh huh huh°]
[$Going on seventy ] SI:[X!$]

[ ha]h-hah-hah
[RT30: TG: 19:08]
01  Bee:  She had it yestihday.  Ten:: pou:nds.
02  Ava:  °Je:sus Christ.°
03  Bee:  She ha[dda ho:(hh)rse hh .hh]

In producing designed surprise source turns, speakers commonly rely on recipients’ being co-cultural members with shared knowledge that (for example) the pain-free early labor (3) and the ten-pound baby (8) are not normal, ordinary, taken-for-granted events in their culture. In producing a sur- prise reaction to a designed surprise source, a recipient endorses a speaker’s stance that some normative expectation has been breached. Conversely, when a turn is not designed as a surprise source, and a recipient does not treat it as such (as in fragment 1: “Pregnant again”), co-interactants likewise confirm for each other a world held in com- mon with no expectations breached. In align- ing about what is and what is not surprising, speakers and recipients invoke and repro- duce mundane understandings of what is normative for their culture.
Two Deviant Cases
In the instances presented so far, surprise sources and surprise tokens have been in a preferred, or aligning, relationship: that is,
surprise tokens have been produced in response to surprise source turns designed to elicit them. We end this section with some counterexamples: instances in which surprise sources and surprise tokens stand in a dispre- ferred, or non-aligning, relationship.
In fragment 9 below, a recipient produces a surprise token even though the speaker has not designed a prior turn to elicit it. In frag- ment 10, a recipient withholds surprise fol- lowing a surprise source turn designed to elicit it. In producing a surprise reaction to a turn not designed to elicit it (as in 9), or in withholding surprise as a reaction to a turn designed to elicit it (as in 10), recipients dis- play some alternative stance towards what can be expected in the normal course of things.
In fragment 9, Ted is doing nothing sur- prising in giving Fran (who has asked for it, line 1) his telephone number, nor does he treat his own telephone number as in any way unusual or counter to expectation. Her surprise reaction (jesus christ, line 11) is clearly not something he is seeking to elicit.
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
Fragment 9: Telephone Number
[RT378: NB:III.1]
01 Fra: [...] you(h)’d be:tter give me yer nu:mber.=
02 Ted: =↓O:kay,
03 (0.3)
04 Ted: Sev’n one four,
05 Fra: Se::ven o:ne fou[:r.
06 Ted: [.t It’s six sev’n three:,
07 Fra: Si::x sev’n three.
08 Ted: Six two,
09 Fra: Six two.=
10 Ted: =Sev’n sev’n.
11 Fra: Se:v’n sev’n.=Jesus Chr(h)i(h)st? wh(h)at (h)a lo:ng

13  Ted:  =mWe:ll, (.) You wouldn’haftuh dial the sev’n one
15  Fra:
17  Ted:  Et the end’v Harbor Boulevhh(h)a(h)rd?=
four if yih git down, .hhh[hh et least ruh the eh-] [But even so::.]
As this interaction illustrates, a surprise reaction can be produced in inauspicious cir- cumstances, the interactional equivalent of being cast upon stony ground in place of the fertile soil prepared for it by a designed sur- prise source.6 One distinctive feature of the surprise token turn is particularly worth not- ing: Fran’s surprise token, jesus christ (line 11), is immediately followed by talk locating its source (the “long number,” lines 11–12). This is unusual in our data corpus: compare, for example, Ava’s unelaborated jesus christ in fragment 8: “ten-pound baby” (and indeed the unelaborated surprise tokens in all of fragments 3–8). Ava’s jesus christ (like the other surprise tokens in 3–8) is produced in response to an immediately prior turn ana- lyzably designed to elicit it; surprise tokens in such environments are not normally followed by any additional talk designed to locate their source. Rather, interactants are assumed to understand the source of the sur- prise by virtue of the proximity of the sur- prise token to the prior turn produced in order to elicit precisely the kind of reaction its recipient has provided.
In our deviant case instance (“telephone number”), by contrast, Fran displays her analysis of the prior turn as not having been designed as a surprise source by following her surprise token jesus christ with informa-
tion locating its source. She thereby conveys her expectation that Ted might have difficul- ty making sense of her reaction without some indication of what could have occasioned it. This analysis is confirmed by Ted’s subse- quent turn, in which he accounts for the fea- ture of his prior talk which Fran has located as the surprise source (the length of his tele- phone number). Thus our academic analysis that Bee’s talk in fragment 8 is designed as a surprise source, whereas Ted’s talk in frag- ment 9 is not, is also the analysis displayed by the co-interactants.
The surprise token turn in fragment 9 bears a strong family resemblance to a phe- nomenon previously described by Charles Goodwin (1996:394–5) as an “elaborating sentence,” which “explicates the reaction that prefaced it.”7 Relatively few of the reaction
7 Goodwin (1996) cites three examples of this phe- nomenon (one produced in response to an event in the world; the others produced as part of reported speech). In his key exemplar, the speaker, working in an airline operations room, responds to images on a monitor array: her reaction token (“Uhoo:::eh:::”) alerts the other workers to some reactable feature of the environment, and her elaborating sentence (“It’s covering half of the ai(h)rpl(h)ane”) formulates its source. Some additional instances are displayed in Goodwin and Goodwin (2000), also in co-present interaction in which the recipients’ access to the observable feature of the world targeted by the reac- tion token has not yet been established. We suggest, then, that Goodwin’s “elaborating sentence” is specif- ic to contexts in which the surprise source is treated as not accessible to the recipient, and that the pattern we identify here is otherwise normative.
6 Fragments 19 (“pregnant sister”) and 24 (“banana leaf”) provide further examples of surprise reactions to turns not designed to elicit them.
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
tokens in our collection overall are followed by “elaborating sentences”: the exceptions are either non-aligning (as in Fragment 9; and Fragment 19 below), or are responsive not to a prior utterance but to an extra-linguistic event accessible only to the speaker (see the examples in note 3).
In our second deviant case, fragment 10
Fragment 10: Big Babies
(taken from a call to a birth crisis helpline), Gill reports surprise sources her babies’ birth weights. Here, surprise reactions (from the call taker) are relevantly missing. As the frag- ment opens, Gill, who is pregnant for the third time, is explaining her concern about her forthcoming labor with reference to the length of her previous one.
[Kitzinger BCC7: 5:17-7:13]
01  Gil:  um (.) VEry long: (.) well co(h)mpa(h)ra(h)tively
05  Clt:
07  Gil:
09  Clt:  mm hm
10  Gil:  ((swallows))  Christian was nine pounds twe:lve.
11        (.)
12  Gil:  So he was quite big. <But the second stage I felt
  1. 13         was the bit that DIDn’t go brilliantly well but
  2. 14         .hhhh I felt quite compromised on what position
  3. 15  I was in.
//        ((about 1 minute later))
90  Gil:  And he was born.
91        (.)
92  Gil:  And um (0.2) he was nine pounds twelve so(h) he
93        was quite big.
94  Clt:  $ We:ll you have very healthy big [bab]ies.
>anyway< (.) It was about thirteen hours first
stage..hhh Two hours second stage. .hh Um:
  [Th]at is long isn’t it for a second
bab[y. mm.]
   [Yeah. ] Well my first-
Marilyn was nine pounds.
95  Gil:
U:m (0.5) I had a physiological third stage
which I think (.) didn’t go brilliantly well
as well ((continues))
At line 8, Gill provides the information that her first baby weighed “nine pounds.” The average full-term birth weight (in the U.K.) is around eight pounds. This cultural knowledge, here presumed to be shared, in part constitutes the weights in lines 8 and 10 as surprise sources. The call taker, however, merely offers a continuer (“mm hm,” line 9). Gill then “ups the ante,” pointing out that her second baby was even larger (“nine pounds twelve,” line 10). In the absence of any reac- tion from the call taker at line 11, Gill offers her own assessment of the baby’s weight (“he was quite big,” line 12), before continuing with her story. About a minute later, Gill’s narrative reaches the birth itself (line 90), and she reiterates both the baby’s birth weight (line 92) and her assessment of it (“he was quite big,” lines 92–93). This time the
repeated information does elicit a turn from the call taker, but she does not align with Gill on the issue of size. The call taker (again) refuses the opportunity to marvel at the babies’ birth weights, substituting an empha- sis on their health (line 94). A surprise reac- tion is therefore still hearable as relevantly missing, and after a short delay (line 96), Gill continues with her narrative. It is likely that the reason for the call taker’s refusal to align is that Gill is offering size of baby as an account for length of labor. Displaying sur- prise at Gill’s unusually heavy babies might be heard here as endorsing this claim of causality, a claim explicitly challenged by the call taker later in the call.
In this section, then, we examined the design of surprise source turns, and showed that participants themselves orient to such
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
turns in accordance with their design. In all of the data we presented (with the exception of fragment 9), the turns at talk preceding the sur- prise token were analyzably designed to elicit surprise, and (with the exception of fragment 10) succeeded in doing so. Recipients’ surprise tokens in these auspicious interactional envi- ronments align with speakers’ own surprise in treating some feature of the world as unexpect- ed.
We outlined some design features of sur- prise source turns (such as pre-announcements, negative observations, extreme case formula- tions). We also began to consider not just how people perform surprise, but also why they do so: what these surprise performances achieve in their local interactional environments. By per- forming as-if-visceral surprise reactions to talk designed to elicit it, people confirm for each other a shared, taken-for-granted world defined by a set of norms, values, and expecta- tions of which the “surprising” behavior, event, or whatever constitutes a breach.
In the final section of this paper we offer a more substantial discussion of the interactional uses of surprise. First, however, we provide empirical evidence to support our claim, fol- lowing Goffman (1978), that reaction tokens are not spontaneous, visceral eruptions, but rather are designed to appear as-if-visceral; that they are little performances of viscerality. In the next two sections we show that people can pro- duce a second as-if-visceral surprise reaction to the same surprise source (recycled surprise); and that some other action, adumbrating or performing surprise in its own right, can inter- vene between the surprise source and the as-if- visceral surprise token (delayed surprise). The
Fragment 11: A Hundred and Fifteen (I)
[RT384: NBII:3:R:1]
01  Emm:  .hhh We BEEN tuh PA:LM SPRINGS.
02        (0.2)
03  Lot:  Oh: God ah be’t it’s
data analyses presented below show how speakers actively extend surprise reactions over multiple turns at talk and create slots for the subsequent production of surprise tokens in the ongoing talk. A visceral eruption (whatever its interactional significance turns out to be) cannot, by definition, have been designed by its producer to achieve some interactional effect. In showing that reaction tokens are constructed to be as-if-visceral, we lay the groundwork for exploring further the interactional work that surprise tokens are analyzably designed to per- form.
Surprise is not necessarily a one-shot per- formance: it can be revived and recycled on subsequent occasions. An as-if-visceral surprise reaction can be “blurted out” multiple times to the same surprise source without apparently thereby losing its spontaneous and impulsive character. In this section we display some examples of “recycled surprise”: instances in which a surprise source is produced and a sur- prise reaction to it is performed appropriately, only for the whole sequence to be recycled a lit- tle later in the interaction.
Fragments 11 and 12 are taken from a tele- phone call between Lottie and Emma, and are separated by about 50 seconds. In the first, very near the beginning of the conversation, Emma topicalizes her recent trip (line 1). She answers Lottie’s question about the heat (lines 3–4) with a surprise source turn giving the tempera- ture (line 5), to which Lottie, in appreciation of its extremity, reacts with a surprise token, oh gosh (line 7).8
surprise source surprise token
05  Emm:
07  Lot:
  1. 09  Emm: ↓Gu:ys th’guys played go:lf over there about a
  2. 10         hunderd’n fifty of’m ((continues))
[.hh[hunderd’n fiftee:n.h
8 Note that a short silence intervenes here between surprise source and surprise token. We discuss this feature in the next section.
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
In the second fragment, Emma checks 98) and again eliciting from Lottie a surprise out with Lottie the temperature back home, reaction (oh god, line 99).
before recycling the surprise source turn (line

Fragment 12: A Hundred and Fifteen (II)
[RT385: NBII:3:R:2]
91  Emm:  Oh it fe:lt ni:ce tih come down
  1. 92         guess ih wz ho:t heere yesterday
  2. 93         wasn’t it..h=
94  Lot:  =Teah it was ri:l ril ni- yesterday
95        morning it wz ri:l fo:ggy.
96  Emm:  Wa: it. I thought it wa:s becuz ih
wz kahna ha:zy out ther-=.hh
‘mA::gine a hunnerd’n fifteen
Oh:: Go:d.=
99  Lot:
100 Emm:  =En ar air conditioner went out. ((continues))
recycled surprise source recycled surprise token
Between these two interactions, Emma, who apparently was launching a telling about her trip, interrupts herself and offers to hang up because Lottie has “got company” (back- ground voices are audible in this call). It is only after Lottie says goodbye to her com- panions, and reassures Emma that she is not “busy,” that Emma relaunches her telling, selecting her (already successfully produced) surprise source turn to do so. In fragment 12, then, the surprise source is recycled as a first part of a telling; in producing a second sur-
Fragment 13: Gold Bridge (I)
[RT386: NB:II:5:R]
01  Lot:  Wt’s new with you:.
02  Emm:  .hhhh Oh:: ah wen’tih th’dentis’n
03        [uh:: G]od’e wantuh pull=
04  Lot:  [Ye:ah?]
05  Emm:  =a tooth ‘n [make me a] new go:ld
prise token, Lottie aligns as recipient for that telling.
In a different conversation between these two women, it is Lottie (again the pro- ducer of the surprise token) who recycles the surprise source turn originally produced by Emma. In fragment 13, Emma replies to Lottie’s personal state enquiry (a version of “howareyou,” line 1) with bad news about her experience at the dentist (the surprise source turn, lines 2–9), and Lottie aligns to produce dismayed surprise (oh shit, line 10).
06  Lot:
07  Emm:
  1. 09  Emm:
  2. 10  Lot:
11 Emm: 12
[.hhh] uh: .hhhh (.) bridge fer
°Oh:: sh::i[:t.° ]
[ Shi]:t.
surprise source surprise token
13  Emm:  Is ri:ght.
14  Lot:  Tha:t’s a big (.) Tha:t’s a big uh:::=
15  Emm:  =.p.hhh=
16  Lot:  =[gimmick.-]
17 Emm: 18

=[He wo:n’t]ry: tuh sa↑:ve this tooth this’s a new de:ntis’he said it’s go:tta d-ih doesn’hu:rt me et A::LL ((continues))
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
Less than a minute later, Lottie (the original recipient) recycles the surprise source turn as a candidate understanding (line 52).9 On receiving confirmation (lines 53–54) she produces another sur-
Fragment 14: Gold Bridge (II)
[RT387: NB:II:5:R]
51  Lot: =°’N tha:t’s a bunch’v money-°
53  Emm:
56  Lot:
58  Emm:
61  Lot:  °O[h::: °<]
62  Emm:    [En I’m] not gonna have this:
prise token (oh shoot, line 56). The down- graded version of the original shit perhaps was selected to avoid having Emma repeat quite so emphatically an obscenity that Lottie herself had delivered sotto voce.
recycled surprise source
recycled surprise token
Ei:ght hundred do:ll[a r s ?] [↑Ihn th]at
°Oh:: sh:::oot° I:’d go tuh somebuddy e[: l s e.]

           [I’m goan] ba:ck tih
my other dentist,h
tooth pulled? iz not ↓b:o:thern me↑:?
cheaper dentist, but one she can trust not to remove a tooth that isn’t bothering her.) The recycled surprise sequence, launched by the initial recipient of the surprise source, serves here to extricate Lottie from her position as recipient of talk with which she is unwilling to express alignment, and to return the con- versation to something on which they can (again) align: the high cost of the dentistry.
Fragments 15 and 16 are taken from classroom interaction and are separated by 20 minutes. In the first of these, a student is in the early stages of her in-class presentation of her project (part of the assessed work for this course). She reports having a collection of “about two hundred” instances (line 1); the professor produces an immediate reaction token (Ph:::::::::ew!, line 3).
Between these two interactions, Emma continues her complaint about her dentist’s insistence on extraction, saying that she “was- n’t ready to have [her] tooth pulled”; that he said “Well I don’t wanna even fix it”; and that she doesn’t “trust him.” Lottie does not align with Emma’s position that a tooth which “doesn’t hurt” should be fixed rather than pulled, and does not offer even minimal responses until she speaks again at line 51 above. Here she returns the complaint more specifically to the cost of the treatment. This concern is additional to those Emma has just been listing (note the “n” [and] with which she launches her turn), as opposed to whether it constitutes good dentistry. (Emma orients to this distinction too, and reinforces, after the recycled surprise sequence, her own position [lines 62–63]: she wants not just a
Fragment 15: Two Hundred Instances (I)
[RT94N: SW: class (I)]
01  Std:  I’ve got a collection of about
02        two hundred=
03  Prf:  =[Phe:::::::::w!]
04  Std:  =[of these (.)  ] things.  Taken
surprise source surprise token
9 Fragment 19 (“pregnant sister”) offers another example of a recipient recycling the surprise source turn and producing another surprise token (lines 19–20). Here, Chloe reproduces the surprising ele- ment of Paul’s turn, not designed as such (“your sis-
ter’s pregnant”), and places her second surprise token (oh my god!) contiguously. Thereby she makes clear that it is the pregnancy itself, and not simply the upcoming reconfiguration of Paul’s family (line 17) to which she is reacting.
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
  1. 05         from about twenty different data
  2. 06         sources.  So I guess I’m working
  3. 07         with a rather bigger collection than
  4. 08         (.) than most of you.  At times I’ve
  5. 09         felt like I was wading in a morass of-
  6. 10  of data.
Although the student offers the large number of items in her collection as a possi- ble surprise source, she does not design the turn to be heard as deliberately eliciting sur- prise. (Compare Mike’s laconic introduction of the vintage Cord automobiles in fragment 27 below.) The professor’s surprise reaction is produced in overlap with the student’s con- tinuing talk. The speech-exchange system here is a “presentation,” and he is a member of her “audience.” She continues with talk that orients to his intervention (with the modest acknowledgment of having “a rather bigger collection,” line 7). This talk, however,
Fragment 16: Two Hundred Instances (II)
[RT217: SW: class (I)]
46  Prf:  So you’ve got two hundred
47        of these do you.
48  Std:  Uh yes.
48         (.)
49  Prf:  GOod Lo:rd!
We have seen three instances of recy- cled surprise. In each case, the recycled sur- prise token performs an aligning action: reestablishing alignment after an interrup- tion (“a hundred and fifteen”) and a non- aligning sequence (“gold bridge”); and reissuing praise (“two hundred instances”), although presumably not only for the bene- fit of the student concerned but also with the pedagogic goal of inspiring other stu- dents in the class to make similarly exten- sive collections. In these dramatic “revivals,” then, as-if-visceral surprise reac- tions to the same surprise source are pro- duced on more than one occasion. They are recycled in new sequences designed expressly to prepare for the apparently spontaneous “blurting out” of an impulsive emotional reaction, which achieves (align- ing) interactional goals.
In producing immediate surprise tokens, and particularly in producing them
is addressed not to him, but to her fellow stu- dents (via the “you” at line 8), and it deflects his possible praise by proceeding immediate- ly to treat the size of her collection as prob- lematic (“wading in a morass of data,” lines 9–10). Twenty minutes later, at the end of the student’s presentation, the professor, an orig- inal recipient of the surprise source turn, recycles it as a candidate understanding (lines 46–47), seeking confirmation that the student has 200 instances of the phenome- non. He then performs a post-confirmation surprise reaction (good lord!, line 49).
recycled surprise source
recycled surprise token
without elaboration, recipients display their understanding that the prior turn was designed precisely to elicit the surprise they have so promptly produced. The sur- prise tokens that follow such turns produce the effect of being visceral eruptions in part because they appear to be “blurted out” in direct and immediate response to the event they (thereby) target as their source (even though, as we have seen, some of these sur- prise sources are recycled).
In this section we focus on surprise tokens that do not immediately follow (are not contiguous with) the surprise source turn. Our data here consist of cases in which the person who subsequently pro- duces a surprise token does something else first, after the surprise source and before producing the surprise token. We have found that a very limited set of practices occupies this slot: silence; displays of ritual- ized disbelief and other repair initiations; and multiple inserts, consisting of combina- tions of these practices, which considerably
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
delay the production of the surprise token.10 We will show that these practices are also methods for performing surprise: that silences, ritualized disbelief, and other repair initiations typically are understood by recipients (at least retrospectively, after the surprise token has been enacted) as constituting little performances of surprise in their own right. In these instances, then, the surprise token is not an immediate vis- ceral reaction but rather is an extension, an upgrade, or a confirmation of surprise already produced or adumbrated through other practices.
Fragment 17: Four Thousand Youngsters
[RT312: Holt:x(c)1:2:7:9]
It is not unusual for silence to intervene between a surprise source turn (clearly designed as such) and a surprise token. We have already shown one such instance (frag- ment 11, line 6); some further examples are displayed below. In fragments 17 and 18, the speakers design surprise source turns using (emphatically produced) extreme case for- mulations (like those in fragments 3–8) to highlight the element of their telling that they treat as counter to expectation. In both cases the recipients produce an aligning surprise token, but only after a short gap of silence.
01 Les:
02 03
  1. 04  Mum:
  2. 05  Les:
‘N there w’r four h: thousan::d youngsters← surprise source in the ↓ha:ll
>My wo:r[d< ]
[.hh] An’ they were ↑a:ll wearing thei:r (0.2) outdoor clothes of course ((continues))
surprise token
surprise source silence
surprise token
Fragment 18: Four Bathrooms
[RT136: NBIV10R]
01  Lot:  Yeah let’s see she’s [go:t
02  Emm:
03  () :  ((sniff))
05  Lot:
07  Emm:
10  Lot:  e-En:: (.) tshe- eehhh She’s a great pers’n nuh
11        run arou:n’ n:aygid yihkno:w ((continues))
[( )
Four ba:throoms.
°°Oh my Gohht.°°
Research on preference organization shows that when a response is “preferred” (aligns with the course of action initiated by the prior talk), it is characteristically pro- duced immediately and without delay (Pomerantz 1984). We have found, however, that a substantial proportion of surprise reac- tions produced in alignment with surprise sources designed to elicit them (as in 11, 17,
18) are delayed by silence (of up to 0.6 sec- ond) without participants treating this as problematic. Rather than understanding such silences as necessarily implying disagreement or dispreference, participants can treat them as little performances of “doing being sur- prised,” which support the eventual produc- tion of the surprise token.
The following observations suggest that silence is not necessarily dispreferred in sur- prise sequences:
First, short gaps are fairly common before the production of surprise tokens fol- lowing designed surprise sources (that is, in sequences that otherwise bear all the hall- marks of being preferred): they occur in
10 The only other practice we have found between a surprise source and a surprise token is requesting additional information pertinent to the surprise source; in every case in our collection, such requests occur in conjunction with silences, displays of ritual- ized disbelief, and repair initiations. We show one instance of this in fragment 27: “two Cords.”
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
around one quarter of the instances in our collection.
In addition, speakers of designedly sur- prising turns that are followed by silence do not generally modify or preemptively refor- mulate those turns (Pomerantz 1984; Schegloff 1995:72–75), as speakers typically do when they hear silence as signaling an upcoming dispreferred response.
Further, there is a common cultural understanding of extreme surprise as a vis- ceral emotion that leaves one temporarily speechless such that silence can—and here often does—embody the reaction of being too shocked or amazed to speak: of demon- strating that “the breath has been knocked out of me” or that one is “at a loss for words,” left “open mouthed” (see note 4), “struck dumb,” “dumbfounded,” “stunned,” or “voiceless” with surprise (compare Darwin’s [1892]/1998: 78] “stupefied amazement”).
Fragment 19: Pregnant Sister
[RT262N+RT590+RT591: Land: YU9: 5:14]
01  Chl:  .hhhh Uhm so d’you know what Asraf’s
  1. 02         doin’ for Christmas.
  2. 03  (.)
04  Pau:  He’s stayin’ at ours.
05  Chl:  Right.  [ Cool. ]
06  Pau:          [He’s um] (.) comin’ over [fer Christ]mas=
07  Chl:                                    [ ‘Cause-  ]
08  Pau:  =E:ve ‘cause my sister’s havin’ a party (at ours)=
09  Chl:  =>Righ<=
10  Pau:  =not a big- .hhh ‘Cau:se uh my sister’s pregnant.
Finally, even when a surprise source turn is not designed as such, a silence following talk that a recipient subsequently treats as surprising, via production of a surprise token, can be heard retrospectively as an initial per- formance of surprise (see fragment 19 below).
In the following fragment, Chloe, one of the more dramatic speakers in our data cor- pora, apparently is so much stunned by Paul’s casually mentioned news of his sister’s preg- nancy that her immediate reaction (line 11) is silence of more than a second (a very long silence in conversation). This fragment comes from earlier in the same interaction as frag- ment 1 (“pregnant again”), in which Paul presents Chloe with news about Alison’s pregnancy and Sally’s delivery of a baby girl; neither of these (as we have seen) is treated as surprising.
12  Chl:  hhhh!
13  Pau:  Hello::?.
surprise source silence
surprise token
HUH HUH ↑↑HELLO::[: : : !!] [huhhuh][huhhuh ]
  1. 14  Chl:
  2. 15  Pau:
  3. 16  Chl:
  4. 17  Chl:
18 Pau: $ I kno:[w. ‘N that’s- ]
19 Chl: [↑↑YOUR SISTER’S] PREGNANT OH MY 20 GOD! Couldn’t they wait til the wedding.

Unlike Andi’s pre-announcement buildup to the news of her pregnancy in fragment 2 (“speaking of bottoms”), Paul’s news (line 10) is not designed as a surprise source. Rather, it is touched off incidentally by his report, in response to a question from Chloe about his boyfriend’s plans for Christmas (lines 1–2), that Asraf will be visiting his (Paul’s) parental home (line 10). Chloe’s silence (line 11) is so extensive that Paul eventually pursues a
response (“hello?” line 13). This elicits an exag- geratedly loud, high-pitched response from Chloe (line 14), before she finally produces a (long-delayed) very loud, very high-pitched surprise token (“↑BLIMEY:!” line 16).11 This
11 Chloe’s enactment of surprise is prosodically very marked, both in volume and in pitch. Her mean pitch during this episode overall is 315 Hertz, com- pared with her normal pitch of around 200 Hertz, and it rises dramatically to 684 Hertz (close to the physio-
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
The use of silence in surprise sequences is also consequential for the production of the surprise token itself. Many are much lower in volume than the talk that surrounds them,12 even whispered (see fragments 6, 8, 13, 14, 18, 22, 31:9). At least in some cases, this may reflect the social mandate to suppress profanities. Some embody “breathlessness” or “voicelessness” in their manner of deliv- ery: sharp intakes of breath (see fragment 20; compare Darwin’s [[1892]/1998:97] “deep and rapid inspiration”) and sharp exhalations or whistles (see fragment 15; compare Darwin’s [[1892]/1998:285–6] “blowing, hiss- ing or whistling noise.” The nonlexical sur- prise token in fragment 20 is a dramatic embodiment of being struck not-quite-dumb with amazement—of being barely able to speak. Here, Chloe’s surprise token (lines 5–6) is provoked by her father’s (dispre- ferred) response (line 4) to her question (line 3) displaying the presumption that the family will, as is apparently their custom, attend the annual public firework display in the grounds of the local stately home (“Windhall”). The surprise token consists of a dramatic perfor- mance of breathlessness with three succes- sive gasping inbreaths, each more pronounced than the one before.
is followed by an elaborating sentence (line 17), that locates the source of surprise, fore- grounds the pregnancy as a topic in its own right, and identifies it as directly relevant to Paul and his reconstituted kinship network. Given that Paul has not designed his news to be heard as surprising, Chloe offers an account for her treatment of the pregnancy as unexpected: Paul’s sister is not yet married (line 20), and therefore, as Chloe later puts it, runs the risk of being “preggers in her dress” (the wedding dress for which fittings have already begun). In addition, as also becomes apparent later in the conversation, Chloe knows that Paul’s sister has a fertility prob- lem (polycystic ovarian syndrome). In this fragment, the surprise token comes nearly three seconds after the surprise source.
In sum, whether or not a turn is analyz- ably designed as a surprise source, a recipient who subsequently produces a surprise token may remain silent for some time before doing so, in which case the surprise token itself can- not be characterized as an immediate viscer- al reaction. And although silence is generally characterized as dispreference-implicative, it may be understood here, at least retrospec- tively, as constituting part of the surprise reaction itself.
Fragment 20: Fighting for Breath
[RT257N: Land:YU8:1:34]
01  Dad:  We’ve got firework night with friends
03  Chl:
04  Dad:
05  Chl:
07  Dad:
on Wednes[da:y]we’ve got [.hh (      )
Uh- No, I d- [well we’re
             [  .h h h !
.u::HHHH! [hhh hhh
[At Windha:ll?
not quite s[ure]
surprise source surprise token
          [I know I know y- I:’m fighting for
breath too at the thought we might not be at
Chloe’s performance is treated unprob- lematically by her recipient, who claims understanding of and alignment with it (“I know I know,” line 7) and reformulates it as “fighting for breath” at the possibility that the family might break with tradition and not engage in their expected behavior that
Wednesday. The “too” (line 8) produces “fighting for breath” as a characterization of Chloe’s performance of surprise by propos- ing Dad’s claim to be “fighting for breath” as identical with (rather than additional to) the action Chloe demonstrated in the prior turn.
In this section we have shown that, counter to what may be expected from classic
logical maximum for many female speakers) on the surprise token blimey. We are grateful to Richard Ogden for these pitch measurements and discussion of them.
12 Occasionally, however, they are notably louder than the surrounding talk (see fragments 19:19–20; 31:18).
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
CA work on preference organization, silence intervening between surprise source and sur- prise token is not necessarily dispreferred. Nor does it apparently detract from the as-if- visceral surprise token. Rather, silence (as well as displays of inability to speak, or diffi- culty in speaking) itself may constitute a mini-performance of surprise, adumbrating or enhancing the surprise embodied in a sub- sequent surprise token.
Displays of Ritualized Disbelief and Other Repair Initiations
In this section we build on previous work (in particular Heritage 1984; Jefferson 1981; Schegloff, 1995, 1997; Selting 1996) to show how a variety of different turn-types that can be used as repair initiations also can be used to do surprise. Like silence, these turns inter- vene between surprise source and surprise token, so that the surprise token is not an immediate visceral eruption.As we will show, the intervention of a repair-initiating sequence means that the surprise token is prepared for (by its eventual producer) at least two turns in advance. We also explore some of the interactional uses of surprise, lay- ing the ground for a more extended discus- sion in the final section of this article.
Repair refers to practices for dealing with ostensible problems in speaking, hear- ing, or understanding talk (Schegloff, Jefferson, and Sacks 1977). Repair initiated by the recipient of a turn at talk (other-initi- ated repair) delays the production of a response fitted to the prior turn by claiming some problem of hearing or understanding which must be addressed so that an appropri- ate response can be produced. One impor- tant commonality across the various types of other-initiated repair is that they single out some or all of the prior talk for special atten- tion, as not meeting the recipient’s expecta-
Fragment 21: Five Home Births
[RT319: Kitzinger HB35]
01  Clt:  I’ve had five at home
02  Ros:  Fi::ve,hh
03  Clt:  mm
tions in some way (Drew 1997; Jefferson 1972; Schegloff 2000). As is now a well-estab- lished finding in the CA literature, these fea- tures make of repair initiations and the practices through which they are implement- ed a resource for adumbrating or displaying dispreferred responses such as disagree- ments, rejections, or declinations (Schegloff 1995:101). We show here that they are also a resource for implementing surprise: in our data set they are commonly used as such to adumbrate or display surprise following turns that were not designed as surprise sources.
In each of the following fragments (21–23), a sequence is inserted between sur- prise source turn and surprise token, in which a recipient of a designedly surprising prior turn ostensibly checks on its accuracy (in each case confirmed). In each case, the first speaker presents a surprise source: an unex- pectedly high number (of home births, 21), or cost (of a cake, 23); the extension of a catego- ry-bound activity (breastfeeding) beyond the category to which it is normatively bound (birth mothers), introduced with an intensifi- er (“even”) to mark its unexpectedness (22). In each case, the recipient then marks the prior turn as surprising, using one or another of the turn-types which can be used for other- initiation of repair, but which also can be used more generally, as here, to mark some utterance or utterance-part as of special interest. These include “repeats or partial repeats, ‘pro-repeats’ (such as ‘he is?’[.|.|.]), and ‘really,’ all with or without a preceding ‘oh’” Schegloff 1995:159). Following Heritage (1984:339), we would add “yer kidding.”13 Finally, in each case, the first speaker pro- duces a minimal confirmation of the surprise source turn (mm, yup, yah); its recipient then produces a surprise token (goodness, my goodness, gosh).
ritualized disbelief confirmation
13 Two instances of “yer kidding” formulations appear in fragment 27 (lines 15 and 26); also see Wilkinson and Kitzinger (forthcoming b).
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
04 Ros: Goodness! surprise token
Fragment 22: Breastfeeding
[RT: 471: Kitzinger BCC483] (They are talking about breastfeeding)
01  Clt:  Even adoptive mothers can do it you kno:w.
  1. 03  Clr:
  2. 04  Clt:
  3. 05  Clr:
06  Clt:  .hhh I:f the:y’ve (.) I mean it’s much easier
07        if they’ve already had a ba:by ((continues))
Fragment 23: Wedding Cake
[RT82: Holt: M88:2:4:42]
01  Dee:  [I] mean hyou’re talkin’ a hundred pound’v a
°↑Can they°. Yup.hh
°My goodness!°

ritualized disbelief confirmation
surprise token
ritualized disbelief confirmation
↑wedding cake today Mar[k
[.tlk Are you really?

Ya:h. (0.2)
Go:sh ↑.huhh ↑huh huh huh↑ .hhh
(Why don’t you all joi:n.) [((sniff)) huhhhh

  1. 03  Mar:
  2. 04  Dee:
06  Mar:
08  Dee:
Various analysts have drawn attention to turns such as the second turns (the first arrowed turns) in the fragments above, refer- ring to them as “news marks” and “news receipts” (Jefferson 1981; Maynard 1997)14 or as “ritualized disbelief” (Heritage 1984:339), the term we adopt here. Unlike other sur- prise tokens, these items initiate sequences: that is, they are regularly understood as mak- ing relevant a response (as at 21:3, 22:4, and 23:4, the second arrowed turns above). They also serve thereby to invite the prior speaker to expand the on-topic talk, as previously pointed out by Jefferson (1981), and as Schegloff (1995: 151) suggests is the case with oh reallys in third position. In each of the cases cited above, that invitation is declined (even though a gap is left for it at 23:5), and the surprise token is produced after the prior
surprise token [We:ll this is it dear ((continues))
speaker has offered nothing more than mere confirmation. In such instances, the surprise token is produced by a speaker put in the position (through the minimal response by the other) of not yet having aligned suffi- ciently in relation to the surprise source: the surprise token upgrades an earlier display of surprise. We often find a peremptory charac- ter to these (oh) reallys, repeats, and yer kid- dings, in that their speakers may not wait for the confirmation they are ostensibly eliciting before producing their surprise reaction. These displays of ritualized disbelief do not so much “ask questions” as convey a stance: that news in the prior turn is unexpected in some way and needs confirmation before it can be otherwise receipted and reacted to. In doing so, they thereby constitute a surprise response in their own right.
Other types of repair initiators (such as interrogatives of various types and candidate understandings) also can be used to launch repair sequences that intervene between sur- prise source and surprise token, thus delaying the production of the latter. Space con- straints permit just one example here: for more, see Wilkinson and Kitzinger (forth- coming b).
In fragment 24 below, the surprise token—the radio program presenter’s oh my
14 Our account bears some resemblance to Maynard’s (1997) and Freese and Maynard’s (1998) analyses (following Jefferson 1981) of the “news delivery sequence,” with the important difference that these previous analyses do not differentiate systemat- ically between items such as oh god and oh my gosh (what we call reaction tokens) and items such as oh how lovely and oh well that’s good (assessments). Crucially, as we discuss in the final section, surprise need not involve assessment of an informing as any- thing other than unexpected.
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
ed that the leaves were purchased. The guest corrects this presupposition, responding that she grows them (line 2). Because this response violates the presupposition the pre- senter has just displayed in her immediately prior turn, she treats it as a surprise source. The presenter initiates repair: she interrupts the guest’s response to first repeat, and then reformulate, the surprise source, presenting it as a candidate understanding (“Oh you just cut them off the tree,” lines 3–4). After the guest confirms this understanding as correct (“Exactly, yes,” line 5), the presenter pro- duces her postconfirmation surprise reaction (line 7).
god! (line 7)—is delayed by an inserted repair sequence (lines 3–6), in which the sur- prise source is reformulated as a candidate understanding.15 This fragment is taken from a British radio cooking program in which guest celebrities are asked to prepare dishes of personal significance. Just before the frag- ment begins, the week’s guest, a Balinese dancer/singer living in London, has explained, as she starts to cook, that her dish is typical of an offering to the god of the Gamelan orchestra (playing in the back- ground). The presenter asks (at line 1) where she bought the banana leaves she is using in preparing the dish, thereby taking for grant-
Fragment 24: Banana Leaf
[RT191: Banana Leaf: BBC Radio 4: 9/06/02]
01  Pre:  Where did you buy tho:se.
02  Gue:
03  Pre:
05  Gue:
07  Pre:
08  Gue:  °huh huh huh° .hhhh In Bali for GOD we do
surprise source
cand understanding
confirmation surprise token
I gro:w- I [gro:w-]
           [Oh you] grow- Oh you just cut
them off the tree[:: ! ]
                 [Ex(h)]actly Ye(huh)s(huh).
[huh huh huh ]
[Oh my go:d! ]
10 Pre: 11

everything. [huh huh huh]
            [huh huh huh] .hhhh Yes I used
an apt exclamation.
In this episode, the guest’s turn “I grow-I grow” (line 2) is produced prosaically, with emphasis designed to perform correction. It is produced only retrospectively as a surprise source, through the presenter’s display of sur- prise. By reformulating the turn as an under- standing check, the presenter singles it out for special attention, thereby setting in motion a sequence that, subject to a confirm- ing response, provides her with a slot to dis- play surprise at the growing and harvesting of banana leaves in central London. This may be done here for the benefit of an overhearing audience, but the practice is not specific to such situations.
We have shown here that several of the turn-types that can be used to initiate repair on prior talk, such as repeats, (oh) reallys and candidate understandings, also can be used to
adumbrate or to do surprise: in fact, oh really may be at heart a surprise token and only derivatively a repair initiation. These perfor- mances of surprise often take the form of a display of ritualized disbelief; in each case, the subsequent production of a surprise token is delayed by the inserted repair sequence. In each of fragments 21 through 24, the incipient producer of a surprise token sin- gles out particular information not designed for special attention by the speaker of the prior turn, initiates repair on it, and on receiving the repair solution, produces an as- if-visceral surprise token. The later produc- tion of any surprise token is not an immediate visceral reaction: rather, it consti- tutes a second (or subsequent) expression of surprise, extending and upgrading the sur- prise already displayed through ritualized disbelief or other-initiation of repair.
Finally, just as the use of silence in sur- prise sequences is consequential for the pro-
15 For a further instance of repair initiated by means of a candidate understanding see fragment 27, line 9.
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
duction of the surprise token itself (as in fragment 22 and the other quiet, whispered, breathless, and breathy examples cited above), so too are displays of ritualized dis- belief and other repair initiations imbricated into the design of surprise tokens; sometimes they are treated as putative surprise tokens in their own right, as in fragments 25 and 26 below.
Just before the opening of fragment 25, Parvati, in an account too long to present here, has been describing her own surprise and disapproval at her Indian relatives’ ideas
Fragment 25: Caesarean
about how she should conduct her pregnancy and labor. Because she has just reported that her pregnancy is progressing normally, it is counter to (British) expectation that she would have a caesarean section. Thus her report of a (distant) relative’s presumption that she might elect to have a caesarean has been built as a source of surprise (and con- demnation). The call taker’s surprise token (an emphatic no, line 3, repeated in line 5) is an expression of ritualized disbelief at Parvati’s report. (We analyze this interaction in greater detail in our final section.)
[RT501+RT502: Kitzinger BCC344]
01  Par:  She said “Oh how are you doing Parvati
  1. 02         are you having a caesarean”
  2. 03  Clt: No::! ritualized disbelief/ surprise token
04 Par: Yes!
05 Clt: [No::. ]
ritualized disbelief/ surprise token 06 Par: [That’s] the way people ta:lk in India.
07 Clt: ((breathy)) Really!hhh

In fragment 26, taken from a “chicken dinner” involving two heterosexual student couples, the cook, Vivian, Shane’s partner, teasingly reports having given their guest,
Fragment 26: Biggest Piece of Chicken
[RT378+RT379: Chicken Dinner:8:2-14]
Michael, the biggest piece of chicken (line 1). She now offers him the one piece that remains.
[I gave Michael the bigges’ ↑piece- too:. (0.9)

01 VIV:
03  SHA:
05  MI?:  °( [ )°
06  NAN:     [Yeh I sa[w tha:]t.
07  SHA:              [Wha:t?]
08  MIC:  We know’oo[rates he:re]:.=
09  VIV:            [Of chicken,]
10  SHA:  =Is this true?
11  MIC:  .t ?hh-?hh (0.2) She gaym’the biggis’
  1. 12  b’↓ta:y/(potato) the biggis’ ↓chicken=
  2. 13  SHA: =nah ↑ha:h ↑O-kay (ul en w’) talk about that later.
surprise token
surprise token
According to Schegloff (1997:541), Shane’s whats (lines 3 and 7) “appear more to be expressions of shock and outrage at hav- ing been short changed by his companion (akin to what Goffman [1978] termed ‘response cries’) than other-initiated repair.” Like reaction tokens more generally, this
kind of what is not unambiguously sequence- launching (but see line 9), nor is it followed by a subsequent surprise token. Rather, this item, also frequently deployed in other-initi- ated repair (Schegloff 2000), is the surprise token.
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
lar part for a vintage car. At line 9, Curt launch- es the first of a whole string of post-expansion questions to convey a stance of awed amaze- ment in response to Mike’s report of a guy he knows who owns two original Cords (vintage cars). Between the surprise source (lines 6–8) and the surprise token (line 35), Curt produces one candidate understanding (line 9), four dis- plays of ritualized disbelief (one repeat, line 32; one oh really, line 12; and two versions of yer kidding, lines 15 and 26), and an information request (line 18). The surprise display is deferred until this supporting information has been elicited (and sometimes receipted and/or assessed), as if to establish the veracity of, and parameters to, the surprise source itself.
03  Gar:
05  Mik:  Well I can’t say they’re ol’ clunkers
Multiple Inserts
The production of a surprise token can be substantially delayed by multiple insert sequences, which themselves often adumbrate or perform surprise, as displayed in fragment 27. The interaction from which this fragment is taken has been examined extensively by other conversation analysts (notably, C. Goodwin 1996, 2002; M. Goodwin 1980; Schegloff 1987); we draw on their analyses in developing an account of the “shocked, elaborated amaze- ment” (C. Goodwin, 2002:157) displayed by one of the co-interactants.
In the fragment displayed here, Mike is responding to Curt’s earlier question (not shown) about where he might locate a particu-
Fragment 27: Two Cords
[RT117N: Auto Discussion, 15-16]
01  Mik:  [Lemme ask] a guy at work. He’s
gotta bunch a’old clu[nkers.
                     [Y’know Marlon Liddle?
06 07
  1. 08  Mik:
  2. 09  Cur:
11  Mik:  Oh yes. Very origi(h)nal
surprise source
upgraded surprise source
cand understanding
ritualized disbelief 1
ritualized disbelief 2
information request
assessment contd
ritualized disbelief 3
ritualized disbelief 4 surprise token
eez gotta Co:rd?
Two Co:rds, (1.0) [And
                  [Not original,
Oh:: reall[y?
          [Yah. Ve(h)ry
°Awhhh are you shittin  m[e?
[No I’m not.
12  Cur:
13  Mik:
15  Cur:
16  Mik:
What’s iz na:me.
18  Cur:
20  Mik:  Harry uh (1.0) Schirmer¿ <Schure¿
21  Cur:  [°Jeez I’d    [like-
22  Mik:  [°Schirmer. = [Schirmer.
23  Cur:  Ahhd like t’mee(hh)t da(h)t
25  Gar:  Has’e [gotta lo:t?
26  Cur:
27  Mik:
29  Gar:  Shu[re?
30  Mik:     [He’s from Milan Ohio.
31  Gar:  Shuur.
32 Cur: Two [Co:rds.
33 Mik: [°Shure.
34 Mik: Ya[h w’l one’s] en o:ld uh, 35 Cur: [Oh:: :]

gu(h)u(h)y .hhhh
[ No: kidd[ing?
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
As Goodwin’s (2002) analysis shows, Curt’s insistently surprised admiration of the two vintage Cords and their owner is not dis- affiliative with the stance taken by the delib- erately deadpan and understated Mike. Rather, social participants have methods of making something available as an assessable object (here a source of awed amazement) without being seen overtly to produce it as such: here, Curt plays knowledgeable enthu- siast to Mike’s knowledgeable insouciance. (We will return to this fragment in the fol- lowing, final section.)
In sum, we have focused here on delayed surprise tokens (those which do not immedi- ately follow the surprise source turn). We have exemplified the practices that typically occupy the slot between the surprise source and the surprise token—silence, ritualized disbelief and other forms of repair initiation, and multiple inserts—and have shown how these practices constitute little performances of surprise in their own right. In these instances, then, the surprise token is not an immediate visceral reaction, but is rather an extension, upgrade, or confirmation of sur- prise already produced or adumbrated through other practices.
So far our key argument has been that surprise is an interactional achievement. Using examples from real life talk-in-interac- tion, we have shown how surprise is socially organized and interactionally produced. Our analysis of surprise reactions has provided concrete evidence for Goffman’s (1978:814) claim that reaction tokens, generally heard as precultural visceral eruptions in fact are “creatures of social situations.” We have examined how co-interactants collaborate to bring off together a performance of surprise; and we have shown how these allegedly spontaneous, impulsive expressions of emo- tion can be recycled on subsequent occasions, without (apparently) thereby losing their spontaneous or impulsive character. We have shown how as-if-visceral surprise tokens often follow and upgrade more extensive sur- prise performances, characterized by silence, ritualized disbelief, and other repair
sequences (and sometimes extended combi- nations of these).
In this final section, we build on the analyses made above to demonstrate that the production of surprise is not only an interac- tional achievement, but also a powerful inter- actional resource for social members. In particular, we show how the expression of surprise is a resource for displaying cultural and category memberships; how it acts as a vehicle for performing other actions (such as apologies and justifications); and how it may be used to defer, or even to displace, other actions.
Surprise Displays (Sub)cultural and Category Memberships
First, “in talk about the world, speakers show whether or not they share one” (Moerman 1988:112). Surprise is one resource for doing just that: it provides a basis for displaying cultural, subcultural, and category memberships. As we noted in rela- tion to fragments 2 through 8, consensual sur- prise displays define a normative world (here breached), and thereby produce interactants as co-members (or not) of that world and co- category members (or not) within it. Our data analyses have shown that surprise is shaped with reference to social judgments about what is to be expected and what is not, and that these judgments are reflected in, and reproduced by, the production of consensual surprise. These displays of surprise show what co-members treat as unexpected, excep- tional, or unusual (all-night parties, fragment 4; ten-pound babies, fragment 8; houses with four bathrooms, fragment 18; and so on), and thereby what they take to be expected, unex- ceptional, or business as usual. In designing some informing so as to elicit surprise and in reacting to that informing with a surprise token (as in fragments 2–8; also see 17, 18, 21–23, 25), co-conversationalists collaborate to reflect and reproduce a shared culture.
Conversely, surprise that is not shared can work to partition co-interactants into dif- ferent (sub)cultural memberships. When a recipient fails to produce surprise following a turn clearly designed to elicit it, or does pro- duce surprise following a turn clearly not designed to elicit it, a discrepancy is revealed
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
between the interactants’ taken-for-granted understandings of, or orientations to, some feature of their world. The discrepancy may arise from, and reproduce, individual differ- ences in what is taken for granted, differ- ences between professional and lay knowledge (as in “big babies,” fragment 10); or cultural differences (as in “banana leaf,” fragment 24). Recipients of such surprise source turns could elect to “let them pass by” unremarked, without thereby having created a “noticeable absence.” In electing to pro- duce, after intervening talk that prepares for it, an as-if-visceral reaction, they choose instead to draw attention to this discrepancy of worldview, giving it particular interaction- al salience.
In “big babies,” the call taker’s noticeably absent surprise at a baby’s birth weight (pre- sented as a surprise source) produces and reinscribes her membership of the category “childbirth expert” and her co-interactant as a “lay” woman. In “banana leaf,” the English radio presenter’s display of surprise at the Balinese cook’s use of banana leaves (not presented as a surprise source) is treated by the latter as marking a cultural difference between them: she teasingly takes as literal the exclamation oh my god, commenting “In Bali for GOD we do everything” (lines 8–9), and thereby producing herself as Balinese and her co-interactant as someone outside that culture. In responding to the format of the surprise token (oh my god) and not to the surprise it enacts, she is also maintaining as normative her own cultural stance to the growing of banana leaves for culinary pur- poses, and declining to align with the presen- ter’s surprise (or, indeed, to engage with it at all).
In fragment 25 (“caesarean”), we saw how Parvati successfully elicited the call- taker’s surprise about her Indian relative’s assumption that she might choose an elective caesarean, thereby producing both interac- tants as co-cultural members (as “not Indian,” aligned in disapproving surprise about Indian values). Immediately subse- quently, however, Parvati produces herself as someone with expertise on “people.|.|.in India” (line 6), thereby making relevant both her identity as the British descendant of Indian immigrants (she has already informed
the call-taker that this is so) and the call- taker’s presumed identity as someone with- out direct knowledge of India. Cultural difference is deployed here (line 6) to nor- malize (for a culture of which the co-interac- tants both are produced again as non-members) an informing already treated as a surprise source.
Surprise, then, does not simply display the acquisition of unexpected new informa- tion; it also claims or displays preexisting knowledge. Surprise displays give rise to the recipient’s inferences both about what was not known by the surprised person and about what was known (providing the basis for whatever it was they treated as unexpected). This (sometimes sophisticated) taken-for- granted knowledge, which makes surprise relevant on any given occasion, may be dis- tributed differentially between co-interac- tants. When one interactant displays surprise that another does not or cannot, or can dis- play only ineptly, their respective category memberships are displayed and reinscribed. For example, by contrast with Curt and Mike, Gary (the third participant in fragment 27, “two Cords”) seems not to know enough about vintage cars to display surprise appro- priately. His question “Has ‘e gotta lot?” (line 29) targets quantity, rather than quality, as the possibly salient feature of the cars (and he gets no response from the other two). Through Gary’s misplaced question, the co- interactants are produced as “vintage car afi- cionados” and Other.
Speaker’s prior knowledge can be inferred either from the absence (or mitiga- tion) of surprise or from its production, depending on the nature of the surprise source turn. The issue of whether to display surprise, and with what degree of intensity, is a recurrent concern for participants, in part because of what it makes inferrable about their preexisting knowledge and (therefore) about their category and cultural member- ships.
Surprise As a Vehicle for Other Actions
Second, surprise acts as a vehicle for other actions; it may be displayed or with- held (and its intensity calibrated) with refer- ence to the other actions it can be
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
understood to convey in its local sequential context. Many of the actions that surprise is used to perform are affiliative, especially when a surprise token follows a surprise source turn designed to elicit it. For exam- ple, some of the exchanges we have shown celebrate the surprising achievements of the speaker of the surprise source turn: not looking one’s age (fragment 7), giving birth to five babies at home (fragment 21), build- ing a large collection (fragment 15). Recycled surprise sequences in our data set are likewise overwhelmingly in pursuit of affiliative goals: reestablishing alignment after an interruption (fragment 12: “a hun- dred and fifteen”) or a non-aligning sequence (fragment 14: “gold bridge”), or praising an industrious student (fragment 16: “two hundred instances”).
Even when people display surprise to
Fragment 28: Tupperware
[RT389: from McCarthy (2003): 50]
01  Amy:  What time are your other
02        people coming, Janet?
03  Jan:  Well the thing officially
  1. 04         starts at two.
  2. 05   Amy:  Oh right ooh gosh.
06  Jan:  The Tupperware lady said she’d
  1. 07         come about half one but as far
  2. 08         as I’m concerned if she comes
  3. 09         about half one-
10  Amy:  Yeah.  Oh sorry we’re early.
At line 5, Janet first receipts the informa- tion about the start time (with a change-of- state token claiming that something not previously known is now known; Heritage 1984). She then performs a surprise reaction to it (“ooh gosh” ), thereby claiming this new information to be unexpected, and hence that her arrival at Janet’s home more than half an
Fragment 29: Half Past Three
turns not designed to elicit them, or with- hold surprise to turns designed to elicit them, very often they do so in pursuit of affiliative goals, as we have seen already from Fran’s deployment of unelicited sur- prise to invite topic talk (about a long tele- phone number, fragment 9) and Curt’s insistently surprised admiration of the two vintage Cords (fragment 27). One common use of (unelicited) surprise displays is to register the breach of a local moral order, and to produce embedded accounts, excuses, and justifications for that breach, as in frag- ments 28 and 29 below.
In fragment 28, Amy is at Janet’s home before a Tupperware sales promotional party. She realizes (with Janet’s response at lines 3–4) that she has arrived too early, and produces a surprise reaction.

hour before the start of the party was inad- vertent, not planned, and based on not having been informed properly about the start time. She subsequently produces an explicit apology (line 10).
In fragment 29, Mum is enquiring whether Lesley’s husband and children have left for their skiing trip.
[RT43: Holt:1:1,1:23]
01  Mum:  ‘Av your family gone o:ff?
02        (.)
03  Les:  Ye:s,
04  Mum:  Oh goo:d,=
05  Les:  =A:t um: half past three: this morning.
  1. 06  (0.3)
  2. 07  Mum: Oh my wo:rd.
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
In response to Mum’s enquiry (line 1), Lesley confirms the family’s departure (line 2), and Mum receipts this as good news (line 4). Lesley comes in quickly after Mum’s “good news” receipt to counterpose the fact of the departure’s unsocial hour (line 5), hearable as a mild complaint on Lesley’s part. By producing a surprise reaction to what is (thereby) treated as unexpected information, Mum excuses herself from hav- ing receipted as “good” an event that, it now turns out, had negative consequences for Lesley.
Fragment 30: Couple of Bottles
Conversely, the withholding of sur- prise—indeed, the explicit denial that any- thing surprising has happened—also can perform other actions. Again, what those actions are depends heaving on context. In fragment 30 below, Nan retorts that Gordon’s account of his previous evening’s activities is entirely as she would have expected. (“What a surprise” is heavily ironic.) This reply treats Gordon and his friends as people who typi- cally drink in the manner described.
[Holt: F:SO88:2:5:3]
01  Gor:  All the old, all the old lads (.)
  1. 02         .t .hhhh gettin’ together we uh .hhh
  2. 03         we polished off a couple a’ bottles.
04  Nan:  What a surprise
Across all of these interactions, then, sur- prise performances embody the assessment that something is unexpected or unusual, and such assessments can act as vehicles for actions including praise (fragment 16: “two hundred instances”), apology (fragment 28: “Tupperware”), and knowledgeable admira- tion (fragment 27: “two Cords”). In withhold- ing surprise, a speaker claims to have had prior expectations (for example, as regards the drinking behavior of Gordon and the “old lads,” fragment 30); such claims likewise act as vehicles for actions such as the claim to cultural knowledge (fragment 25: “caesare- an”) and the attribution of a characteristic trait to a co-interactant (fragment 30: “couple of bottles”). The interactional meaning car- ried by displays of being surprised (or not surprised) is both context-free, insofar as it hinges around the core notion of “expectabil- ity,” and context-dependent, insofar as the meaning of treating something as expected (or unexpected) in any given context depends on what else is being done, deferred, or supplanted by the expression of surprise.
Surprise As an Alternative to Other Actions
Third, surprise tokens are alternative to other possible actions that could have been performed in the interactional context in which they occur, so that the action per- formed by the surprise token depends to
some extent on how it stands in relation to such alternatives. For example, tellings, announcements, informings, and so on typi- cally elicit assessments (as good or bad). The appropriate valence of the assessment, how- ever, is not always immediately apparent, and the interactional costs of getting this wrong— of treating good news as bad news, or vice versa—can be high. In response to an announcement, a surprise token sometimes may be deployed precisely because it can be designed so as merely to mark some news as unexpected and nothing more. Thus surprise tokens can defer assessment until more infor- mation has been given (or can displace assessment altogether). As Sacks (1995:574) put it, “The surprise thing can be treated as reserving rights to future expression of emo- tion, saying, ‘I see that this is the thing that I will express emotion about. Let me give you some more room to tell me about it. Then you’ll hear me give a wail.’”
The interactional advantage of giving the prior speaker “more room to tell” is that this may enable a recipient to determine whether wailing or celebration is the more appropri- ate response. For example, in fragment 31 below (an expansion of fragment 2: “speak- ing of bottoms”), Betty seems to deploy the surprise token precisely to pass the turn at talk back to the prior speaker, in order to gather more information about the valence
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
of response required. As Maynard (1997:116) points out, Betty’s surprise token enables her to postpone any assessment of the news (of Andi’s pregnancy) as either good or bad until after establishing Andi’s own stance. Only after Betty has established that the pregnancy
is intended (the couple having elected to reverse the vasectomy) does she produce the assessment “oh I’m so happy” (line 21), treat- ing the pregnancy as “good news” (whereas her earlier surprise token merely registered its unexpectedness).
Fragment 31: Speaking of Bottoms (Expansion of Fragment 2)
[RT513+RT514: PND3:18 (from Maynard 1997)]
01  Andi:   .hhh well: (.) speaking of bottoms are you sitting
02          dow:n.
03  Betty:  Ye:ah.
04  Andi:   Well (.) we have some news for you:.
05  Betty:  What.
06  Andi:   .hhh th’t (.) may come as $ a bit of a surpri:se. ehhh!
07  Betty:  I see- $what are you telling me.$=
have a
08  Andi:
09  Betty:
11  Andi:   Yeah.
12          (0.5)
13  Andi:   .hhh[::::::::: ]
=hhhh! Bob and I are gunna have a baby. <°O:h my: go:od↑ness°> hho- (0.5) did you reverse- he have a reversal?
[ ↑↑whe::n.]
tch e-yup. Last ↑Ma:rch. (0.5)
14  Betty:
15  Andi:
17  Andi:   .mhhh  ((sniff))
18  Betty:
19  Andi:
20  Andi:
21  Betty:
O:H [ MY ↑GO:OD ]NESS: [And it was-]
It was (.)[very successful.][Very $quickly] hh::h .hhh [ O:H I’M SO: ][ ↑HA:PPY. ]

When responding to a prior turn that is producing complaint, surprise tokens are characteristically disaffiliative. That is, affilia- tion with a complaint calls for more than sim- ply an expression of surprise: for example, it requires confirming the speaker’s assessment of the situation, joining in the complaint, or expressing dismay or sympathy (see Curl, Drew, and Ogden forthcoming). In fragment 32 below, although Mum treats the matter of
Fragment 32: Letter
the first-class letter that hasn’t arrived as a surprise source, via her surprise token good gracious (line 7), she does not offer any (neg- ative) assessment of the postal service, nor does she join in Lesley’s complaint. Indeed, her subsequent question about the contents of the letter (line 9) is hearable as possibly implying that the delay in arrival may not matter.
[RT315: Holt:X:(c):1:2:7:3]
01 Les: Becuz I posted a letter (.) to Katharine last Monda:y 02 Mum: Hm:?
03 Les: First cla:ss,
05  Mum:  Y[es]
06  Les:
07  Mum:
09  Mum:  Is there anything important in it?
[An]d she hasn’ got it yet.
↑Good gracious. (0.7)
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
Here the surprise token provides a response that registers the unexpectedness of the complained-about behavior, but without engaging with the “complaint” element. In effect, it displaces an appropriate alternative action, such as co-complaining or sympathiz- ing.
In sum, because surprise tokens can be limited to simply marking the unexpected- ness of a telling or an announcement, they can displace or defer other actions made sequentially relevant. The action accom- plished by reacting with surprise to a prior turn depends very much on the action pur- sued by that prior turn (and indeed within the sequence more generally). Analysis of any particular surprise token, in its sequential context, requires analyzing it for what it is an alternative to, as well as for what surprise itself does in that position.
In this final section we have considered some of the interactional uses of surprise: the display of (sub)cultural and category mem- berships (through differential displays of sur- prise to some feature of the world, indexing differences in cultural, specialist, or individ- ual knowledge); the use of surprise as a vehi- cle for other actions (such as pursuit of affiliative goals, attending to breaches in a local moral order, or trait attribution); and the production of surprise as an alternative to some other action (such as an assessment, a co-complaint, or an expression of sympathy), or as a means of deferring it. In all of these ways, then, surprise is an interactional resource as well as an interactional achieve- ment.
We have offered a detailed empirical analysis of how surprise is constructed and deployed in everyday social interaction through the use of reaction tokens. Further, we have examined why this is done, in specif- ic interactional contexts.
We have built upon Goffman’s (1978) pioneering work on “response cries,” provid- ing empirical evidence in support of his claim that these apparently spontaneous, uncon- trolled “exclamatory imprecations” (p. 798), which we call reaction tokens, are fundamen- tally interactional events. We have shown
empirically that expressions of surprise con- veyed through surprise tokens are not invol- untary emotional eruptions but interactionally organized performances. First, surprise tokens typically are produced after talk analyzably designed to elicit sur- prise from its recipient: for example, by using devices such as pre-announcements, negative observations, and extreme case formulations to build a surprise source turn. Second, sur- prise is not necessarily produced on one occasion only, but can be recycled: surprise tokens can be produced on multiple occa- sions to the same surprise source without (apparently) thereby losing their as-if-viscer- al character. Third, surprise tokens also can be delayed by practices such as silence, dis- plays of ritualized disbelief, and other repair initiations. These practices themselves appear to constitute mini-performances of surprise, thus rendering the surprise token an exten- sion, upgrade, or confirmation of an earlier intimation of surprise rather than an immedi- ate visceral reaction to the unexpected. All of these observations support our analysis of surprise as an interactional achievement.
We have moved well beyond Goffman’s notion of “response cries” as “creatures of social situations” (p. 814), not only in our specification of how surprise tokens work, but especially in our examination of what they can be used to do. We have shown that surprise provides a potent interactional resource for social members, and that the expression (and withholding) of surprise is thick with culture. In performing as-if-viscer- al surprise in response to talk designed to elicit it, people confirm for each other a shared, taken-for-granted world, defined by a set of norms, values, and expectations, of which the surprising behavior or event con- stitutes a breach. The interactional uses of surprise that we have identified are the reflection and reproduction of a normative world; the production and reinscription of membership categories; affiliation and disaf- filiation; and the management of the local moral order.
The analyses presented above contribute substantially to the interactional tradition of studying emotion, specifically surprise. Our key contribution has been to separate the personal experience of encountering the
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
unexpected (the psychology of surprise) from the public display of finding something counter to expectation (the social expression of surprise). Further, in analyzing surprise displays in some detail, we have shown not only that these are interactionally organized performances, but also that these perfor-
Appendix. Transcription Key
[X] brackets
= equals sign
(0.5) time in parentheses
(.) period in parentheses ::: colons
. period , comma
? question mark
¿ inverted question mark
! exclamation mark - dash
here underlining
HERE capitals
HERE underlining and capitals °here° degree signs
°°here°° double degree signs
>here< “more than”/”less than” <here> “less than”/”more than” sym-
< “less than” symbol
> “more than” symbol
↑↓ up or down arrow
_: underlined letter before colon : underlined colon
$ dollar sign
* asterisk
((f)) letter f in double parentheses hah
empty parentheses
(bring) word(s) in parentheses (boy)/(buy) slash separating word(s) in
((sniff)) word(s) in double parentheses
[...] three periods in brackets // double slash
mances accomplish an array of social actions. In this way we also regard our work, specify- ing how one particular aspect of human behavior is produced and socially managed in situ, as a contribution to the broad tradi- tion of ethnomethodological and conversa- tion-analytic work in sociology.
overlapping talk
no space between turns
intervals within or between talk (measured in tenths of a

discernable pause or gap, too short to measure
extension of preceding sound (the more colons, the greater the

closing intonation (not necessarily the end of a sentence) continuing intonation (not necessarily between clauses of

rising intonation (not necessarily a question)
rising intonation weaker than that indicated by a question

animated tone (not necessarily an exclamation) abrupt cutoff of sound
loud, relative to surrounding talk
very loud and emphatic, relative to surrounding talk soft, relative to surrounding talk
very soft, or whispered, relative to surrounding talk speeded up, relative to surrounding talk symbols slowed down, relative to surrounding talk

rapid start to following talk
slow ending to preceding talk
marked rise or fall in pitch, immediately following the arrow downward pitch turn (falling intonation contour on the vowel) upward pitch turn (rising intonation contour on the vowel) smile voice
“creaky” voice
high-pitched (“falsetto”) voice
laughter particle inserted into talk
audible outbreath (number of
hs indicates length)
audible inbreath (number of
h s indicates length)
transcriber unable to hear words
transcriber uncertain of hearing
alternate hearings by transcriber

sounds or other material hard to transcribe; other comments by transcriber
words omitted for presentational purposes lines omitted for presentational purposes
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
Averill, James R. 1974. “An Analysis of Psychophysiological Symbolism and Its Influence on Theories of Emotion.” Journal for the Theory of Social Behaviour 4:147–90.
———. 1982. Anger and Aggression: An Essay on Emotion. New York: Springer-Verlag.
Bollinger, Dwight. 1985. Intonation and Its Parts. London: Edward Arnold.
———. 1989. Intonation and Its Uses. London: Edward Arnold.
Cahill, Spencer E. and Robin Eggleston. 1994. “Managing Emotions in Public: The Case of Wheelchair Users.” Social Psychology Quarterly 57:300–12.
Curl, Traci, Paul Drew, and Richard Ogden. Forthcoming. Linguistic Resources for Social Action. Cambridge, UK: Cambridge University Press.
Darwin, Charles. [1872]/1998. The Expression of the Emotions in Man and Animals. London: HarperCollins.
Drew, Paul. 1997. “’Open’ Class Repair Initiators in Response to Sequential Sources of Troubles in Conversation.” Journal of Pragmatics 28:69–101.
Edwards, Derek. 1997. Discourse and Cognition. London: Sage.
Ekman, Paul. 1992. “An Argument for Basic Emotions,” Cognition and Emotion 6:169–200.
Ekman, Paul, Wallace V. Friesen, and Ronald C. Simons. 1985. “Is the Startle Reaction an Emotion?” Journal of Personality and Social Psychology 49:1416–26.
Freese, Jeremy and Douglas W. Maynard. 1998. “Prosodic Features of Bad News and Good News in Conversation.” Language in Society 27:195–219.
Garfinkel, Harold. 1967. Studies in Ethnomethodology. Englewood Cliffs, NJ: Prentice-Hall.
Goffman, Erving. 1978. “Response Cries.” Language 54:787–815.
Goodwin, Charles. 1996. “Transparent Vision.” Pp. 370–404 in Interaction and Grammar, edited by Elinor Ochs, Emanuel A. Schegloff, and Sandra A. Thompson. Cambridge, UK: Cambridge University Press.
———. 2000. “Action and Embodiment Within Situated Human Interaction.” Journal of Pragmatics 32:1489–1522.
———. 2002. “Recognizing Assessable Names.” Pp. 151–61 in Studies in Language and Social Interaction: A Festschrift in Honor of Robert Hopper, edited by Philip Glenn, Curt LeBaron, and Jennifer Mandelbaum. Mahwah, NJ: Erlbaum.
Goodwin, Charles and Marjorie H. Goodwin.
1987. “Concurrent Operations on Talk: Notes on the Interactive Management of Assessments.” IPRA Papers in Pragmatics 1:1–54.
Goodwin, Marjorie H. 1980. “Processes of Mutual Monitoring Implicated in the Production of Description Sequences.” Sociological Inquiry 50:303–17.
Goodwin, Marjorie H. and Charles Goodwin. 2000. “Emotion Within Situated Activity.” Pp. 239–57 in Linguistic Anthropology: A Reader, edited by Alessandro Duranti. Oxford: Blackwell.
Harré, Rom. 1986. “An Outline of the Social Constructionist Viewpoint.” Pp. 2–14 in The Social Construction of Emotions, edited by Rom Harré. Oxford: Blackwell.
Heath, Christian. 1989. “Pain Talk: The Expression of Suffering in the Medical Consultation.” Social Psychology Quarterly 52:113–25.
Heritage, John. 1984. “A Change of State Token and Aspects of Its Sequential Placement.” Pp. 299–345 in Structures of Social Action, edited by J. Max Atkinson and John Heritage. Cambridge, UK: Cambridge University Press.
———. 2003. “Presenting Emanuel Schegloff.” Pp. 1–10 in Discussing Conversation Analysis: The Work of Emanuel A. Schegloff, edited by Carlo L. Prevignan and Paul J. Thibault. Amsterdam: Benjamins.
Hochschild, Arlie Russell. 1979. “Emotion Work, Feeling Rules and Social Structure.” American Journal of Sociology 85:551–75.
———. 1983. The Managed Heart: The Commercialization of Human Feeling. Berkeley: University of California Press.
Jefferson, Gail. 1972. “Side Sequences.” Pp. 294–338 in Studies in Social Interaction, edit- ed by David Sudnow. New York: Free Press.
———. 1981. “The Abominable ‘Ne’?: A Working Paper Exploring the Phenomenon of Post- Response Pursuit of Response.” Occasional Paper 6, Department of Sociology, University of Manchester.
———. 1996. “On the Poetics of Ordinary Talk.” Text and Performance Quarterly 16:1–61.
Kraut, R.E. and R.E. Johnston. 1979. “Social and Emotional Messages of Smiling: An Ethological Approach.” Journal of Personality and Social Psychology 37:1539–53.
Levis, John M. 2002. “Reconsidering Low-Rising Intonation in American English.” Applied Linguistics 23:56–82.
Local, John. 1996. “Conversational Phonetics: Some Aspects of News Receipts in Everyday Talk.” Pp. 177–230 in Prosody in Conversation, edited by Elizabeth Couper-
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
Kuhlen and Margret Selting. Cambridge,
UK: Cambridge University Press.
Maynard, Douglas W. 1997. “The News Delivery Sequence: Bad News and Good News in Conversational Interaction.”
Research on
Language and Social Interaction 30:93–130. Moerman, Michael. 1988. Talking Culture: Ethnography and Conversation Analysis. Philadelphia: University of Pennsylvania
Ortony, Andrew and Terence J. Turner. 1990.

“What’s Basic About Basic Emotions?”
Psychological Review 97:315–31. Parkinson, Brian. 1991. “Emotional Stylists: Strategies of Expressive Management Among Trainee Hairdressers.” Cognition
and Emotion 5:419–34.
———. 1996. “Emotions Are Social.”
Journal of Psychology 87:663–83.
Plutchik, Robert. 1980.
Emotion: A Psychoevolutionary Synthesis. New York:
Harper & Row.
Pomerantz, Anita. 1984. “Agreeing and

Disagreeing With Assessments: Some Features of Preferred/Dispreferred Turn Shapes.” Pp. 57–101 in Structures of Social Action, edited by J. Max Atkinson and John Heritage. Cambridge, UK: Cambridge University Press.
———. 1986. “Extreme Case Formulations: A Way of Legitimizing Claims.” Human Studies 9:219–29.
Potter, Jonathan A. 1996. Representing Reality. London: Sage.
Reisenzein, R. 2000. “Exploring the Strength of Association Between the Components of Emotion Syndromes: The Case of Surprise.” Cognition and Emotion 14:1–38.
Reissland, N., J. Shepherd, and L. Cowie. 2002. “The Melody of Surprise: Maternal Surprise Vocalizations During Play With Her Infant.” Infant and Child Development 11:271–78.
Roach, P. 1983. English Phonetics and Phonology: A Practical Course. Cambridge, UK: Cambridge University Press.
Sacks, Harvey. [1973]/1987. “On the Preference for Agreement and Contiguity in Sequences in Conversation.” Pp. 54–69 in Talk and Social Organization, edited by Graham Button and John R.E. Lee. Clevedon, UK: Multilingual Matters.
———. 1995. Lectures on Conversation, vols. 1 and 2. Oxford: Blackwell.
Sacks, Harvey, Emanuel A. Schegloff, and Gail Jefferson. 1974. “A Simplest Systematics for the Organization of Turn-Taking for Conversation.” Language 50:696–735.
Schegloff, Emanuel A. 1987. “Analyzing Single Episodes of Interaction: An Exercise in Conversation Analysis.” Social Psychology Quarterly 50:101–14.
———. 1988. “Goffman and the Analysis of Conversation.” Pp. 89–135 in Erving Goffman: Exploring the Interaction Order, edited by Paul Drew and Anthony Wootton. Cambridge, UK: Polity.
———. 1995. “Sequence Organization.” Department of Sociology, UCLA. Unpublished manuscript.
———. 1997. “Practices and Actions: Boundary Cases of Other-Initiated Repair.” Discourse Processes 23:499–545.
———. 2000. “When ‘Others’ Initiate Repair.” Applied Linguistics 21:205–43.
———. 2002. “The Surfacing of the Supressed.” Pp.241–62 in Studies in Language and Social Interaction: A Festschrift in Honor of Robert Hopper, edited by Philip Glenn, Curt LeBaron, and Jennifer Mandelbaum. Mahwah, NJ: Erlbaum.
Schegloff, Emanuel A., Gail Jefferson, and Harvey Sacks. 1977. “The Preference for Self- Correction in the Organization of Repair in Conversation.” Language 53:361–82.
Scherer, K.R., M.R. Zentner, and D. Stern. 2004. “Beyond Surprise: The Puzzle of Infants’ Expressive Reactions to Expectancy Violation.” Emotion 4:389–402.
Selting, Margret. 1996. “Prosody As an Activity- Type Distinctive Cue in Conversation: The Case of So-Called ‘Astonished’ Questions in Repair Initiation.” Pp. 231–70 in Prosody in Conversation, edited by Elizabeth Couper- Coulen and Margret Selting. Cambridge, UK: Cambridge University Press.
Strongman, K.T. 2003. The Psychology of Emotion: From Everyday Life to Theory. 5th ed. New York: Wiley.
Tomkins, S.S. 1962. Affect, Imagery and Consciousness. New York: Springer.
Wilkinson, Sue and Celia Kitzinger. Forthcoming a. “Sympathy Tokens in Talk-in-Interaction.” ———. Forthcoming b. “Repair As an Interactional Resource: Other-Initiated
Repair in the Performance of Surprise.”
Downloaded from by Wayne Mellinger on October 22, 2011
#2843—Social Psychology Quarterly—VOL. 69 NO. 2—69203–Wilkinson
Sue Wilkinson is a professor of feminist and health studies in the Department of Social Sciences, Loughborough University, and editor of Feminism & Psychology: An International Journal. Her current research interests include reaction tokens in conversation, the experience of breast cancer, health care interactions, and same-sex marriage.
Celia Kitzinger is a professor of conversation analysis, gender, and sexuality in the Department of Sociology, University of York. She is completing a book on feminism and conversation analy- sis. Her current research interests include the experience of trauma following childbirth, the social construction of the emotions, and same-sex marriage.