This study examines how silence is used in psychotherapy and how the therapist and client treat silence in different contexts. We demonstrate that silence is treated as meaningful in that it sheds light on the client’s thought process and engagement with therapeutic interventions. We use conversation analysis (CA) as a data-driven approach to examine the utility and meaning of silence that participants attribute to it during therapeutic interactions.
Literature Review
Many writings on silence in psychotherapy are written from psychoanalytic perspectives in which the therapist is taught to allow silence to occur regardless of possible relevance of speaking (e.g., Prince, 1997). In psychoanalysis, silence is taken to be meaningful and attributable to the client with the exception of when the psychoanalyst decides the client is ready for an interpretation (Haim, 1990). Blos (1972) claims that attributions of the meaning of silences can be made based on the psychoanalyst’s feelings about the silence. Fliess (1949) even argues that silence is a form of sphincter closure to cause retention of words (substituting for excretory products). Such retention is supposedly intended to punish the psychoanalyst and is an act of transferring feelings about a dominant and cold parent to the psychoanalyst. The idea that silence can be equated with client aggression has been a particularly salient feature of psychoanalysis. In 1961, the Journal of the American Psychoanalytic Association published a special issue on silence as a form of acting out (cited in Ephratt, 2008). Gale and Sanchez (2005) challenge these traditional views of silence, operating from within a psychoanalytic framework, and arguing that taking silence out of the realm of language leads to mistakes in interpretation and denies clients “quiet areas” within the therapeutic community (cf. Davies, 2007, who reviews in detail the ways in which silence has been characterised in psychoanalytic theory).
Silence between turns does not always indicate a problem or forthcoming dispreferred response. Differences in silence can be contextual or cultural (Berger, 2011; Gardner et al., 2009), and in psychotherapy may be followed by heightened emotional expression (Hill et al., 2019; Soma et al., 2023). Nagaoka et al. (2013) found that in a Japanese sample, more silence during sessions was associated with higher satisfaction ratings. In some instances, silence is not only unproblematic but necessary for a preferred response to be treated as such. We examine this possibility in the context of psychotherapy. Most literature regarding silence in psychotherapy has focused on the meaning of silences and is based on clinical experience rather than empirical studies. Some explicitly advise clinicians to be especially tolerant of silence and to avoid filling silences. Most publications dealing with silence in psychotherapy are written from prescriptive, theoretical positions by practitioners who are writing from their own interpretations of their own practice.
Surprisingly, no research has looked in depth at silence in psychotherapy as a linguistic device beyond measuring “interactional silence” (e.g., Frankel et al., 2006). Most empirical research in this area has measured the talk time of each party or examined proposed intra-psychic meanings of silence through direct interpretation or interviews (e.g., Levitt, 1998, 2002). However, talk time and silence are not direct opposites, as talk can occur in overlap and each party is usually silent during another’s talk (Jefferson, 1984, 1986; Sacks et al., 1974). Mahl (1956) uses the formula: patient’s silence quotient = seconds of silence / seconds available to patient to talk, yet characterises such silences as dysfluent. Others measure the presence of silence or patterns of silence but with little reference to interactional function (e.g., Hill, 1978; Tomicic & Martínez Guzmán, 2011). Importantly, silence is not monolithic in its placement nor meanings (Hill et al., 2003; Levitt, 1998; Sacks et al., 1974). The current study considers how silence is used by clients as a way of conveying sincerity.
Although speaker transition regularly occurs with relatively little silence in everyday conversation (Sacks et al., 1974), longer silences following an initiating action are characteristic of dispreferred responses such as refusals and negatively valenced answers (Schegloff, 2007). However, this is not the only instance in which silence following an initiating action occurs. It can also be used as a sensitivity marker. For example, Silverman found that silence was one way in which people marked topics as sensitive in HIV/AIDS counselling (Silverman, 1996; Silverman & Peräkylä, 1990). Coupland and Coupland (1997) likewise observed silence as a feature of death-implicative talk (e.g., discussion of organ failure) in geriatric medicine. In all three cases, silence is used to show sensitivity to social expectations or norms, while also demonstrating sensitivity to the emotional state of recipients.
Cook (1964) presents one of the few empirical studies on silence in a strictly psychotherapeutic context. He used five cases rated by a therapist as “successful” and five cases rated as “unsuccessful” from a previous study (Tomlinson & Hart, 1962) that had examined an issue unrelated to silence–the Process Scale (A. Walker et al., 1960). From each session, Cook extracted nine 2-minute segments which were rated for successfulness using the Process Scale. He then calculated the client’s silence ratio (or total proportion of silence) for each segment. Only silences of 5 seconds or more that were terminated by the client were counted towards the silence ratio. As with most other authors that have explored silence in psychotherapy, Cook makes the assumption that silences are attributable to the client and that the “owner” of the silence will be the one to break it. Silence in the psychotherapy context occurs between two or more people. Silence in this context inherently comes about as a result of nobody speaking and therefore everybody engaging in silence–what McDermott and Tylbor (1983) term collusion. Despite this, silence in interaction is often attributable to a particular party or parties, making a verbal response relevantly absent. Nevertheless, Cook (1964) found that segments with the highest Process Scale ratings had 4%–20% silence. Interestingly, segments that had more than 20% silence were rated as more successful than segments with less than 4% silence, a finding echoed more recently by Nagaoka et al. (2013). The current study explores one reason why this might be the case, namely whether therapists treat quick responses as less sincere or otherwise problematic compared to delayed responses.
As observers, we are unable to definitively say whether a particular silence is due to intra-psychic factors. However, at least some silences will be due to various intra-psychic difficulties that may be exploited in interaction. Chafe (1980) discusses several factors that influence the presence of silence in conversation. Some of these include pragmatic boundaries and transitions, whilst at other times silence indicates a difficulty in transition, a change of state (cf. “oh” as a change-of-state marker, Heritage, 1984) or searching for a word (cf. Goodwin, 1983).
One issue with studying silence in psychotherapeutic contexts is consideration of the roles that the therapists and clients take. Baker (1955) found that when psychiatrists took a passive role, clients were more engaged and had a vocabulary expansion rate similar to ordinary conversation. However, when psychiatrists adopted a more active role, clients were defensive and had less variety in their speech. This is a similar phenomenon to machine gun questions described by Tannen (1985) where recipients may feel obliged to respond quickly, with precisely what was asked, and in the minimum language required to do so. Esterly (1979) reported on the work of Goodman, a psychologist, who described the effect that a fast-talking style such as this can have on those unaccustomed to it–they may become “upset, dissatisfied, incompetent, though they may not understand why” (p. 68). Goodman postulated that people adopt a fast-talking style due to “anxiety, domination, boredom, [or] the need to express freshly stimulated thoughts” (Esterly, 1979, p. 68). He described this style as crowding, and although the adoption of a fast-talking style is not necessarily due to intra-psychic problems, incongruence between institutional and personal communication styles can cause severe misunderstandings.
In a different situation that often involves considerable threat to face and potentially dire consequences, such as interrogation of witnesses in courtrooms, silence is often advised in order to communicate thought-through and consistent responses, with some manuals advising no less than a 5-second gap following the question (Summit, 1978). However, opposing lawyers frequently interpret silence as lying or resistance (A. G. Walker, 1985). Cross-culturally, interpretation of silence can have detrimental effects. Eades (2000, 2007) and Mushin and Gardner (2009) found that Indigenous Australians whose backgrounds include regularly leaving long silences before verbally answering questions were frequently misinterpreted in court as uncooperative. The lawyers did not recognise that many Indigenous Australians’ answers to questions actually include considerable silence, i.e., as preparation and not simply a gap. Gardner et al. (2009) propose that this situation engenders a “clash not of fundamental cultural turn-taking practices, but of tolerance for silence in a particular institutional setting” (p. 84). Berger’s (2011) work on co-present interactions from the United States and United Kingdom supports Gardner’s idea that a lack of orientation to “getting things done” contributes to silence in conversation and that rather than having culturally bound turn-taking practices, silence can vary situationally within a culture.
In the current study, we explore how the conversational style and therapeutic agenda sometimes clash such that conversationally appropriate immediacy is treated therapeutically as resistance or insincerity, whereas the practice of leaving gaps contributes to the performance of sincerity. In order to examine the nature of silence in psychotherapeutic contexts, we conducted two related analyses. In Analysis 1, we examine cases in which clients produce responses immediately following the therapist’s therapeutically relevant turn. In Analysis 2, we examine in detail 2 minutes of a therapy session in which silence (and the lack thereof) was key to the breakdown and realignment of the interaction.
Method
Participants and Data
Participants were recruited purposively from private-practice psychotherapists, who identified clients who may be interested in participating. Nine hours of spiritually oriented humanistic psychotherapy was recorded between a woman, Paula, who is social work-trained, and two separate clients, Sally (3 hours) and Leif (6 hours). The participants are middle-aged European-Americans from the Mid-Western and Mountain regions of the United States. Sally is a woman who was assigned male at birth and had been seeing Paula for assistance with her gender affirmation for two years. She exhibits dyskinesia from her Parkinson’s Disease medication rather than slowing from the disease itself. Leif is a man who had been seeing Paula for 3 years for recovery from alcoholism and has mobility problems that cause him considerable anxiety. The study was approved by the University of Roehampton Ethics Board, and names and locations have been pseudonymised.
Design
The data were recorded using a consumer-grade video camera that the participants were able to control themselves. We transcribed the data using Jeffersonian transcription and performed CA. A guide to this method is in the Appendix. Jeffersonian transcription involves demonstrating linguistic variation in prosody, volume, and silences textually (see transcription key in the Appendix), while CA is intended to identify how interactants treat each other’s talk and utilise conversational strategies.
It does not appear to be the case that the silence discussed here is part of a therapeutic approach but rather the interaction. This is evinced by silence that can be attributed to either therapist or client regardless of prior speaker or who has rights or obligations to speak next. Additionally, there are some instances of interactants doing being silent, doing thinking, or doing not responding yet, but most of the silences between these dyads are not of this nature. The silences examined in this report occur in sequential positions in which the silence is attributable to the client, who is tasked with providing a therapeutically relevant response to the therapist’s prior action.
Quantitative aspects of the study were undertaken by the authors who coded therapeutic questions in ELAN (version 6.6; Sloetjes & Wittenburg, 2008), a tool that allows for codes and audio to be stored together.
Results
Analysis 1–Immediacy is Dispreferred: Accounting for Immediate Responses
One approach to identifying preferred responses within a sequence type is to examine the frequency with which different practices occur in responsive positions. A gap[1] is one of the best indicators that a dispreferred response is forthcoming, so ordinarily one would expect gaps following initiating actions to be uncommon. This was, however, not the case for overtly therapeutic questions. Out of 130 therapeutic questions, 120 (92.3%) were followed by gaps. Ten instances (7.7%) across the nine sessions involved therapeutic questions that were responded to immediately. These will be examined here in light of the local sequential environment.
Extract 1 consists of a series of delayed responses despite an immediate utterance.
Extract 1 [PL 100109]
01 Pau: And how do you know it's depresh: depressed
02 Lei:>> I don:-
03 Pau: W'I [I mean
04 Lei:>> [I
05 Pau: Instead of pleasant
06 (1.6)
07 Pau: What's the quality of it
08 (.)
09 Lei:>> .I do(h)n't feel good about it
10 (2.4)
11 Pau: What are you thinking
In Extract 1, Paula’s self-repair intervenes during Leif’s production of an answer. In order to maintain the floor, Paula is engaging in a rather quick uptake of turns. Leif finally produces his full response in line 9 after two attempts to begin what could be the same answer (“I don-,” “I,” and finally “I don’t feel good about it”). Following Leif’s immediate response (line 9) to Paula’s final question (line 7, “What’s the quality of it?”), Paula indicates that his answer is inadequately specific by asking for more information (line 11). In Extract 1, the delays are caused by Paula’s self-repair, rather than Leif doing sincerity or thoughtfulness, and Paula treats Leif’s otherwise immediate response as inadequate. This leaves a further nine instances of immediate responses to therapeutic questions for us to examine to help in understanding how participants orient to the interactional requirements of the therapeutic agenda.
Structurally Provided-For Responses
The initiating action in Extract 2 is actually the first part of an insert sequence which pursues a response following a long gap of 10.4 seconds (line 12). Paula has given an interpretation (lines 1–11) to which Leif has not yet responded. He responds immediately (line 14) to Paula’s pursuit in line 13. In this instance, Leif has already exhibited delay, and further delay might demonstrate lack of engagement, rather than thoughtfulness. Recall that Cook (1964) found there was an optimal proportion of silence in his client-centred psychotherapy data.
Extract 2 [PL 010810]
01 Pau: You're the same person three weeks ago
02 (1.2)
03 Pau: And yet (2.4) your ability to (0.1) see the situation
04 differently (1.8) each time you encounter it (2.8) is
05 greater than it used to be
06 (4.4)
07 Lei: ((coug[h))
08 Pau: [It's like you're not (.) thinking habitually
09 (1.2)
10 Lei: ((cough))
11 Pau: Your approach to things is not so: (2.8) stereotypic °anymore°
12 (10.4) ((facial & postural gestures, "doing thinking"))
13 Pau:>> How do you like that
14 Lei:>> Well I like it fine
15 (4.3)
16 Lei: ((cough)) Of course naturally (2.1) I've tried to
17 figure out why
Although a response is not conditionally relevant following a statement, Leif makes moves to engage by placing coughs between Paula’s turns and using body movements typical of taking the floor (not shown). Following Paula’s interpretation, he abandons these practices and does thinking (line 12). Paula’s question, “How do you like that?” builds on an implicit acceptance of her interpretation and she presumes that her interpretation is accurate. Her question also makes a response conditionally relevant and directs Leif by providing a question-answer framework that may be easier for him to utilise due to its specificity. It also focuses him on the therapeutic agenda by directing him to the kind of information she is seeking.
Also following an interpretation, Paula asks Leif a question about one aspect of the problem that is directly related to her interpretation (Extract 3). Paula’s turn ends with a yes/no tag question, thereby increasing the likelihood of a response (Stivers & Rossano, 2010), to which Leif responds affirmatively and immediately.
Extract 3 [PL 100109]
01 Pau: hh so what is it (.) and if it passes (0.4) is there a
02 way for you to relate to it (1.6) differently than you
03 >> do now when it comes because there's kind of an "it"
04 >> quality to it isn't there
05 Lei:>> Yeah
06 Pau: You know you're not saying to me (1.6) I didn't get my
07 bike out and I meant to all weekend and it was a little
08 windy so I didn't do it and then I got depressed
Extracts 1–3 involve interpretations and confirmation-seeking in the form of questions. In ordinary conversation, the preferred responses to confirmation-seeking are yes responses, which are routinely without delay. However, without elaboration on the part of the client, the therapist treats these seemingly preferred responses as inadequate. In Extract 1, Paula explicitly requests elaboration, whereas in Extract 2, she remains silent for 4.3 seconds before Leif self-selects to elaborate. In Extract 3, Paula elaborates on her own prior turn, accounting for the interpretation by citing what Leif has (not) said, thereby treating Leif’s “yeah” as a dispreferred response to her interpretation.
Immediate Responses to Obvious Questions
Again, Paula presents an interpretation (Extract 4). This time, she presents Sally with an interpretation using rising intonation (line 4), marking it as a question to which a response is conditionally relevant. Sally confirms this interpretation (line 5) with an oh-prefaced response (Heritage, 1998), marking it as an obvious answer for which the question should not have been asked. Agreement is the preferred response in this position due to Paula’s turn design and the nature of the sequence. In this case, Sally treats Paula’s prior action as inappropriate through her oh-prefaced response. Following this sequence, a 1.4-second lapse develops before Sally makes an “um” vocalisation, pauses for 2.3 seconds, and changes the topic.
Extract 4 [PS 122309]
01 Sal: And we (2.3) had uh oh shaggy bells instead of silver silver
02 bells heh
03 (2.4)
04 Pau:>> So there is play once in a great while
05 Sal:>> Oh yeah
06 (1.4)
07 Sal: Um (2.3) somebody e-mailed me an (1.2) engineering drawing on
08 how to (0.9) uh () how to erect a Christmas tree
In Extract 5, Paula produces a preliminary turn to a therapeutic question (lines 1–4), in which she rephrases her query-in-progress as a reflexive wondering. Leif does not respond to this, and following a 4.4-second gap (line 5), she accounts for her earlier turn with “I mean I don’t know that just occurred,” followed by an overtly therapeutic question, “So do you have any theory on why you don’t get out on the bike if it’s a little too hot?” Leif responds immediately to this with “yes” in line 9. Following a 1.9-second gap (line 10), he reiterates and expands on his “yes” response and produces a piece of paper (lines 11–12). This paper contains the weekly newsletter from Alcoholics Anonymous and is a regular feature of their sessions, to which Paula orients by laughing in line 13. In line 14, Leif indicates that Paula has forgotten something from their last session, “We were just talking about that the other day weren’t we?” with emphasis on “weren’t we,” thereby marking her question as having an obvious answer (line 14). This ultimately results in Paula having to check her notes (line 17). Kitzinger (2011) has documented interactional trouble when interactants are caught having pretended to remember. Paula asked a question that made visible her not remembering, which is then treated by Leif as having an obvious answer and receives a dispreferred response.
Extract 5 [PL 012210]
01 Pau: Have you ever wondered if it just occurred to me (0.2)
02 I wonder how often our excuses (0.8) reflect our fears
03 you know we make the excuse before we're even conscious
04 of what's the (.) fear that going and doing this
05 (4.4)
06 Pau: I mean I don't know that just occurred
07 Pau:>> So do you have any theory on why (1.0) you don't get out on
08 the bike if it's a little too hot
09 Lei:>> Ye:::s.
10 (1.9)
11 Lei: Yes I think it has a lot to do with it I uh as a matter
12 of fact ((takes out paper))
13 Pau: heheheh
14 Lei:>> We were just talking about that the other day weren't we::
15 Pau: °Yea:h°
16 Lei: I think we were
17 Pau:>> I think you were what were we talking about ((checks notes))
Sometimes participants treat questions as having obvious answers that are not evident specifically from previous turns or interactions but from broader contexts. Extract 6 involves Sally describing a neurosurgical procedure that she does not want.
Extract 6 [PS 110409]
01 Sal: But you know they have to (1.8) stimulate things in
02 (1.6) you know what happened (1.1) do you get back
03 memories from the past er (0.8) do you smell
04 something er (1.6) °does it affect motor function°
05 Pau:>> And that really unnerves you
06 Sal:>> °Oh yeah°
07 Pau: Do you suppose that's something that you and I can
08 prepare you for
She is talking about being awake during surgery as one of the top reasons she does not want it (lines 1–4). Paula then asks, “And that really unnerves you?” (line 5), which Sally treats as obvious by using an immediate, oh-prefaced response (Heritage, 1998) in line 6.
Disaffiliative Immediate Responses
Disaffiliation is one possible feature of dispreferred responses. Disaffiliation is when a party displays a stance that opposes that of another party (cf. Stivers, 2008). This may be explicit or implicit, as in Extract 7. Sally has been discussing how her neurologist has been pressuring her to enter a clinical trial for an invasive procedure that, according to Sally, only temporarily “resets the clock” and involves significant risks that she is not willing to take.
Extract 7 [PS 110409]
01 Pau:>> Have you ever talked to someone who's had it
02 Sal:>> Yes
03 Pau: Okay more than a few people
04 (2.7)
05 Sal: UM (3.5) I guess that depends on how close you're (6.0) h(h)how
06 how one to one the talk has to be
07 (1.2)
08 Sal: Y'know I've as far as presentations in front of groups=
09 Pau: Mm
10 Sal: =um (2.3) I've seen y'know probly a dozen or so
Paula asks Sally whether she has ever talked to someone who has had the procedure (line 1), to which Sally responds immediately “yes” (line 2). Although Paula’s turn in line 1 is designed for a yes response, which Sally provides, Paula treats “yes” as reflecting a small number and thus inadequate consideration (line 3). Following Paula’s clarification request, Sally produces longer turns with gaps and pauses and accounts for her “yes” response by expanding on her sources of information. Paula now treats this response (lines 5–10) as adequate with the minimal response token “mm” (see Gardner, 1997, 2002).
Extract 8 involves Paula asking Leif a follow-up question to how Leif feels about making excuses. Leif has said that he wonders what his grandmother would say if he was making excuses around her, and Paula and Leif have just completed a repair sequence related to the unexpectedness of Leif’s comment (data not shown). This extract is on the same general topic as Extract 5 and follows it. Paula’s question (lines 1–2), “What would your granma if she was here right now what would she be saying?” is a therapeutic question designed for Leif to articulate his thoughts on excuses. Leif responds immediately and ironically with “She probably wouldn’t say anything she’s probably whack me up the side of the head with a frying pan and tell me to get out and start pickin potatoes or something” (lines 5–6). Paula does not, however, treat Leif’s turn as a joke. She instead responds with an interpretation, “So her response would be don’t over don’t overanalyse it.”
Extract 8 [PL 100109]
01 Pau:>> [What would your granma (0.7) if she was here right now what
02 would she be saying
03 Lei:>> Sh[e probably w]ouldn' say anything she's probably=
04 Pau: [( )]
05 Lei:>> =whack me up the side of the head with a frying pan and tell
06 >> me to- (1.2) get out and start pickin potatoes or something
07 (0.7)
08 Lei: Pulli°n' potatoes°
09 (0.9)
10 Pau: S:o (0.3) her (0.3) response (1.2) would be (1.1) don't over
11 don't overanalyse it
Soon after, in Extract 9, Paula attempts a therapeutic intervention to help Leif be less judgemental on himself (lines 1–6), in which she invokes the image of his grandmother hitting him with a frying pan. Leif responds immediately in line 7 with a laxly pronounced affirmative/negative token “nyeah,” which can be produced to do sensitive interactional work (Jefferson, 1978). In this instance, Leif is agreeing with the intervention but disagreeing with the image through which his self-judgement has been conveyed. Because Paula did not treat his previous turn as a joke, there is a chance that she has understood the image as realistic. As we see in his next turn at lines 8 and 10, Leif takes back the image of his grandmother (Extract 8, lines 5–6). As Leif is beginning this turn, Paula begins simultaneously “It’s just you know,” orienting to the negative element of “nyeah” and displaying accountability for the interpretation. Paula drops out, however, and does not articulate a full account. Instead, she produces acknowledgement tokens “mkay” (line 11) following possible completion and “mmhm” (line 12) following a second possible completion of Leif’s turn. Ultimately, Paula implicitly accounts for the intervention in lines 15–17 by responding that although Leif’s grandmother might not be violent, he might be (figuratively) violent to himself.
Extract 9 [PL 100109]
01 Pau: >We enjOY IT but hhh so I guess what I'm e-askin is (0.3) could
02 >> you take the grandmother:: (0.2) image (1.2) and instead of it
03 >> being (1.2) someone whacking you with a ruler or a frying pan 04
>> and saying >eh get back to work< (0.8) judging you (1.3) what
05 >> if she's an aspect of yourself that knows: (1.1) this too shall
06 >> pa:ss
07 Lei:>> Nyeah
08 Lei:>> [Actually y']know she she was she really wasn't the type=
09 Pau: [It's just you know]
10 Lei:>> =that'd take a frying pan or[ a] roller to somebody
11 Pau: [°°mkay°°]
12 Pau: Mmhm
13 Lei: But uh
14 (1.4)
15 Pau: But you might be
16 (1.3)
17 Pau: To yourself
Analysis 2–Silence is Preferred: An In-Depth Analysis
The final case of immediate response requires a lengthier discussion and benefits from greater background information. Rather than providing this final extract here, we present an analysis of the entire first 2 minutes of the session, which involves an interaction that has gone awry and is eventually realigned. Features include a characteristic delayed sincere or thoughtful response, a quick response that minimises, and a quick response that challenges. We focus here on the use of silence and immediacy by Leif and how the associated utterances are treated by Paula, the therapist.
Characteristic Delayed Sincere or Thoughtful Responses
We begin with an example that exemplifies a sincere or thoughtful response to the therapist’s talk. Extract 10 involves Paula and Leif discussing Leif’s former drinking.
Extract 10 [PL 101609]
1. Pau: [What about it was so seductive
2. (0.9)
3. Lei: .hhhh Wl I've been thinking about that,
4. (0.8)
5. Lei: ↑I really have
6. (1.6)
7. Lei: Um
8. (1.8)
9. Lei: MOSTly what I didn- (3.3) mostly what I don't miss
10. (2.1) is disappointing myself
11. (0.7)
12. Pau: Hm
13. (2.0)
14. Pau: Hm
15. (1.0)
16. Lei: In other words (0.5) spending ti::me (6.6)=
17. Pau: hmhehhu[h ((cough))
18. Lei: =[basically getting wa:sted (2.3) when I probably
19. could have been doing s(h)omething else
20. (.)
21. Pau: °Yeah°
In line 1, Paula produces a request for information, “What about it was so seductive?” Leif immediately displays thoughtfulness by remaining silent briefly (line 2) and reporting that he has been thinking about Paula’s very question on his own (lines 3 and 5). Line 3 is followed by another brief silence, and line 5 is followed by a lengthier one of 1.6 seconds (line 6), “um” (line 7), and another lengthy silence of 1.8 seconds (line 8). Leif begins to formulate his actual response in line 9 “mostly what I didn-,” followed by a 3.3-second pause and continuing with a replacement repair “mostly what I don’t miss,” a 2.1-second pause, and “is disappointing myself.” Leif’s silences before he begins his actual response demonstrate consideration of his response and are broken up to emphasise that a response is being prepared and thus that Paula should not initiate pursuit. Leif is then silent for another 3.7 seconds (lines 11–15), during which Paula displays affiliation with his response by saying “hm” twice (lines 12 and 14). Leif has not yet provided an in-depth answer, although he has provided what appears to be a sincere answer. In refraining from talk at this point, Paula recognises that there is more to come or that more is expected, while her “hm” vocalisations show that Leif is on the right track in terms of what kind of answer is required. In line 16, Leif continues by expanding his prior turn with “in other words (0.5) spending time” followed by a 6.6-second pause. Paula coughs (line 17), and Leif continues his turn with “basically getting wasted (2.3) when I probably could have been doing something else” (lines 18–19). Paula’s “yeah” (line 21) explicitly affiliates with Leif’s response and serves as an acknowledgement token.
Extract 10 involves an apparently sincere response, in which the client uses lengthy silences, to a therapeutic question by the therapist. Although relatively long silences are common with these participants, this extract contains very long silences of up to 6.6 seconds. In the subsequent sections, we show how the presence of lengthy silences of this sort are not mere happenstance but are one way in which clients can demonstrate sincerity. We contrast the smooth flow and silences of Extract 10 with that of the immediately preceding discussion to show how responses with an inadequate gap following interventions are treated as incorrect or insincere responses by the therapist and how quick responses can be used by the client to challenge or minimise the therapist’s prior action.
Uses of Quick Responses
Quick Responses that Minimise
Extract 11 is the beginning of the exchange that includes Extract 10. It involves an instance of a quick response that minimises the therapist’s prior talk. Immediately prior to Extract 11, Paula has finished setting up the camera in order to record their session for the day, and the pair have exchanged a brief, unrelated sequence. Paula’s topic initiator, “Do you miss it?” (line 1) does not have a clear referent. Leif’s “What?” (line 3) is an other-initiated repair on “it”, and Paula’s “drinkin” (line 5) is the repair solution. Lines 2, 4, and 6 contain silences, but the typical micropause or interturn silence for this dyad is approximately 0.4 seconds. As a result, these silences do not on their own constitute “silence” for the participants. Thus Leif’s “no” (line 7) comes quite quickly. He follows this up at line 9 with “Not really.” Paula does not attempt a turn during this time, nor for another 0.7 seconds (line 10). Thus, she has treated Leif’s “no” as an as yet inadequate response.
Extract 11a [PL 101609]
1. Pau: hh do you miss it
2. (0.6)
3. Lei: What
4. (0.5)
5. Pau: Drinkin
6. (0.5)
7. Lei: No
8. (0.4)
9. Lei: Not really
10. (0.7)
When Paula does speak again (line 11, below), she challenges Leif’s assertion that he does not miss drinking. Leif does not respond during the 1.1 seconds (line 12) of silence following Paula’s “Really?” (line 11). Paula produces an account for her disbelief (lines 13–14) comparing drinking to smoking (she is an ex-smoker). After a 0.8-second gap (line 15), she expands this turn with an emotive account of how one can miss smoking (lines 16–17), followed by a 0.9-second gap. She further continues, bringing her talk back to Leif, “You don’t miss (1.9) you don’t miss drinking?” (line 19) and “You don’t miss tying one on?” (line 20). This receives no response from Leif over a 3.8-second gap (line 21). Paula then produces a direct suggestion of what Leif does not miss about drinking “You don’t miss Patty [his wife] getting mad at you?” After a brief gap, Leif agrees (line 24), and they both laugh (lines 24–25).
Extract 11b [PL 101609]
11. Pau: Really?
12. (1.1)
13. Pau: Cause I thought for most people it's sorta like
14. (0.9) s:moking (.) you miss tha:t
15. (0.8)
16. Pau: There are moments when you just (0.8) boy it'd be
17. a perfect moment for a cigarette you know
18. (0.9)
19. Pau: You don't miss: (1.9) °you don't miss drinking°
20. Pau: You don't miss tying one on
21. (3.8)
22. Pau: You don't miss Patty getting mad at you
23. (0.7)
24. Lei: No I don't miss [that
25. Pau: [.hhuh
Extract 11 represents a lengthy and problematic portion of the session, with Paula resorting to reporting her assumptions and experiences (13–17) to account for having asked Leif whether he misses drinking. This troubled interaction is brought on by Leif’s quick, minimising response to Paula’s therapeutic question in line 1. Paula produces six turns during the course of Extract 11b before Leif responds vocally, indicating his strong disaffiliation. Paula’s turn at line 11 (“Really?”) challenges Leif’s “no” response (lines 7–9), and he shows no uptake until Paula attempts humour at line 22, when Leif then agrees (line 24).
Quick Responses That Challenge
A quick response is also a feature of client responsive actions that challenge the assumptions of the therapist’s initiating action. Extract 12 follows directly from Extract 11 and contains three instances of Leif challenging Paula’s assumptions (lines 3, 11, and 29). Once Paula has received confirmation from Leif about something that he does not miss (Extract 2b, line 24), she adds to the list of things that she asserts that Leif does not miss “You don’t miss (1.1) falling down because of drinking?” Leif immediately replies, “Well I didn’t usually do that” (line 3). This is significant, because, as we have noted, micropauses and interturn silences tend to be approximately 0.4 seconds between Leif and Paula. Leif’s immediate response is thus very quick in this context. Paula produces an open class repair initiator (Drew, 1997) in line 4, which is associated with inapposite or disaffiliative prior turns (i.e., Leif’s disagreement in line 3). Leif specifies the referent of “do that” as “fall down” in line 5, thus providing the repair solution.
Paula now must deal with a declined interpretation. She has a number of options to regain the alignment and affiliation demonstrated in Extract 11b, lines 22–25. She attempts to reformulate her perspective to appear as though she meant something in particular all along with which Leif can agree, i.e., that his falling down was not because he was drunk but because he was in withdrawal (line 7). Her hesitation and reformulation invokes Leif’s doctor’s authority (lines 8–10). Leif refuses this interpretation with “that was really” in terminal overlap with Paula (line 11). This turn challenges Paula’s assertion and comes even more quickly than his previous challenge. Paula offers a collaborative completion (line 12), reasserting her version of events, “That was kind of a time limited thing,” effectively interrupting Leif’s turn constructional unit (TCU) as this too is refused. Leif continues his turn in overlap with Paula, “That was really ataxia” (line 13). After a 2.2-second gap (line 14), Leif produces an increment “so it was kind of a clugeon[2] of ((gesture))” (lines 15–17). Paula produces “yeah” at line 18, claiming affiliation and alignment with Leif’s version.
Extract 12a [PL 101609]
1. Pau: You don't miss: (1.1) falling down becau::se of
2. drinking
3. Lei: W' I didn't usually do that
4. Pau: Pardon
5. Lei: I didn't usually fall down
6. (0.8)
7. Pau: Even with the PAWS[^3]
8. (.)
9. Pau: Even when um (0.8) your doctor was talking
10. about the post withdrawa[l
11. Lei: [That that really .hh ['s
12. Pau: [That kind of was a time limi[ted thing
13. Lei: [That was really ataxia
14. (2.2)
15. Lei: An so it was kind of a *clugeon* of ((*brings
16. hands together interlocking fingers slightly
17. during this word))
18. Pau: ↓Yeah
Leif continues his prior turn after a 1.2-second gap (lines 20–22) with “two things but I y’know I didn’t fall down yeah.” After a 3.2-second gap, Leif reformulates his turn at line 22 with “Yeah I didn’t (2.3) you know puke all over myself or anything like that uh” (lines 24–25). After a brief gap, Paula says, “Well no you don’t miss the disgusting things I guess at all but.” This turn claims both affiliation and alignment but demonstrates instead misalignment and affiliation with something Leif did not say–a lack of intersubjectivity with Leif’s course of action (i.e., Leif said this did not happen; Paula implied that it did and that he would not miss these “disgusting things”). Leif rejects Paula’s interpretation “Y’I didn’t- really didn’t-.”
Extract 12b [PL 101609]
19. (1.2)
20. Lei: .h two things but [I I=
21. Pau: [.hhhhh
22. Lei: =>Yknow<I didn't fall down yeah
23. (3.2)
24. Lei: Yeah I didn't (2.3) you know puke all over myself
25. or anything like that uh
26. (0.9)
27. Pau: h well no you don't miss the disgusting things I
28. guess °at all° but w-
29. Lei: Y'I didn't- really di[dn't-
Re-Establishing Alignment and Affiliation
Revisiting Extract 10, reprinted here as Extract 13 with the final line of Extract 12 included, one can identify striking differences between the sequential environments of Extracts 11 and 12 and Extract 10/13. In line 1, Leif denies Paula’s assertion that he “doesn’t miss the disgusting things” that he said he did not experience. Paula again interrupts him. This time, however, she does so with a therapeutic question, “What about it was so seductive?” (line 2), which does not presume any particular experiences other than being drawn to alcohol. There is an immediate, observable switch in the interaction, and Leif performs sincerity primarily through silence. Line 2 is Paula’s therapeutic question following misaligned talk about Leif’s experiences with alcohol. This is followed by a brief silence at line 3 and the beginning of Leif’s answer at line 4, which after many silences and much expansion are fully realised in lines 19–20. Although Leif’s answer could continue further, it is at this point that substantial information has been conveyed and at which Paula explicitly affiliates with his response (line 22, “Yeah”).
Extract 13 [PL 101609]
1. Lei: Y'I didn't- really di[dn't-
2. Pau:>> [What about it was so seductive
3. (0.9)
4. Lei:>> .hhhh Wl I've been thinking about that,
5. (0.8)
6. Lei: ↑I really have
7. (1.6)
8. Lei: Um
9. (1.8)
10. Lei: MOSTly what I didn- (3.3) mostly what I don't miss
11. (2.1) is disappointing myself
12. (0.7)
13. Pau: Hm
14. (2.0)
15. Pau: Hm
16. (1.0)
17. Lei: In other words (0.5) spending ti::me (6.6)=
18. Pau: hmhehhu[h ((cough))
19. Lei:>> =[basically getting wa:sted (2.3) when I probably
20. could have been doing s(h)omething else
21. (.)
22. Pau: °Yeah°
A similar situation occurs in Extract 14, prior to which Leif and Paula have been discussing what Leif is experiencing as depression. In line 1, Paula challenges Leif’s conceptualisation of sleeping late and not getting things done as depression, “And how do you know it’s depression?” Leif immediately responds (line 2) and is unable to complete his turn before Paula produces a self-repair (line 3) that is not necessary for comprehension. Instead, this self-repair functions to account for having asked the question. By producing it shortly after Leif begins his response, “I don-,” Paula orients to Leif’s response as dispreferred.
Extract 14 [PL 100109]
01 Pau: And how do you know it's depresh: depressed
02 Lei:>> I don:-
03 Pau: W'I [I mean
04 Lei:>> [I
05 Pau: Instead of pleasant
06 (1.6)
07 Pau: What's the quality of it
08 (.)
09 Lei:>> .I do(h)n't feel good about it
10 (2.4)
11 Pau: What are you thinking
12 (2.2)
13 Lei: I'm
14 Pau: =( ) don't feel good about it sounds like a thought.
15 Lei: I'm I'm pretty much (0.9) feeling like (2.5) you know there's
16 really no place for me here (0.5) you know I['m,
In line 4, Leif tries again but is in overlap with Paula and drops out. Paula ultimately reissues her question as “What’s the quality of it?” (line 7), to which Leif produces a complete turn, “I don’t feel good about it.” This turn is likely that which he had previously attempted, because it is prosodically similar and contains disturbances during both words that were present in his earlier partial turns. Disturbances such as aspiration can indicate trouble with a word (Potter & Hepburn, 2010), and in this case indexes prior interactional trouble rather than a problem with the words themselves. In line 11, Paula produces a follow-up question, “What are you thinking?” indicating that Leif’s response is inadequate. After a 2.2-second gap (line 12), both parties begin speaking almost simultaneously. Although Leif begins first (line 13), Paula’s turn is latched to his first syllable (line 14) and explicitly targets Leif’s response as inadequate, “( ) don’t feel good about it sounds like a thought.” Although Leif’s second try in responding to Paula’s last question comes immediately following her turn at line 14, he pauses multiple times, and his focused talk in lines 15–16 continues beyond this extract (data not shown). By engaging in silence, Leif makes moves toward realignment with Paula.
Increments and Reformulated Turns Versus Continuation
The therapist generally provides ample time for responding but does, on occasion, produce increments or redoes a turn. This is particularly evident in Extract 11, partially reproduced here as Extract 15.
Extract 15 [PL 101609]
1. Pau: hh do you miss it
2. ((lines deleted, repair sequence))
3. (0.5)
4. Lei: No
5. (0.4)
6. Lei: Not really
7. (0.7)
8. Pau: Really?
9. (1.1)
10. Pau: Cause I thought for most people it's sorta like
11. (0.9) s:moking (.) you miss tha:t
12. (0.8)
13. Pau: There are moments when you just (0.8) boy it'd be
14. a perfect moment for a cigarette you know
15. (0.9)
16. Pau: You don't miss: (1.9) °you don't miss drinking°
17. Pau: You don't miss tying one on
18. (3.8)
19. Pau: You don't miss Patty getting mad at you
In Extract 15, following a repair sequence (line 2), Leif responds “No” (line 4) to Paula’s yes/no interrogative, “Do you miss it [drinking]?” (line 1). He expands this answer in line 6 as “not really” which mitigates the dispreferred response. Although missing drinking is negatively valued, in the context of psychotherapy, a simple “no” to a therapeutically relevant question blocks the therapeutic agenda. After a brief silence (line 7), Paula expands this sequence with “Really?” with rising intonation, demonstrating surprise and an element of disbelief. Leif does not respond here, and a 1.1-second gap develops (line 9). Paula produces additional turns at lines 10–11, 13–14, and 16–17, whilst Leif still does not respond during moderate silences of 0.8–1.9 seconds. At line 18, a 3.8-second gap occurs before Paula takes another turn. Paula’s turns claim a degree of continuity between each one. Compare this to Extract 12b (partially reproduced here as Extract 16), in which Leif breaks the silences.
Extract 16 [PL 101609]
1. Lei: =>Yknow<I didn't fall down yeah
2. (3.2)
3. Lei: Yeah I didn't (2.3) you know puke all over myself
4. or anything like that uh
Following Leif’s turn at line 1, which is neither conditionally relevant nor incomplete, a 3.2-second silence ensues (line 2). Leif then self-selects at lines 3–4 with a continuation of line 1, recycling “yeah” as the first word in line 3 and on the same topic. In both Extracts 15 and 16, Paula and Leif have employed the same practices of self-selection following a gap. This is a very different practice from Leif’s use of silence to perform sincerity in Extract 10.
Extract 10 [PL 101609]
1. Pau: [What about it was so seductive
2. (0.9)
3. Lei: .hhhh Wl I've been thinking about that,
4. (0.8)
5. Lei: ↑I really have
6. (1.6)
7. Lei: Um
8. (1.8)
9. Lei: MOSTly what I didn- (3.3) mostly what I don't miss
10. (2.1) is disappointing myself
11. (0.7)
12. Pau: Hm
13. (2.0)
14. Pau: Hm
15. (1.0)
16. Lei: In other words (0.5) spending ti::me (6.6)=
17. Pau: hmhehhu[h ((cough))
18. Lei: =[basically getting wa:sted (2.3) when I probably
19. could have been doing s(h)omething else
20. (.)
21. Pau: °Yeah°
In Extract 10, Paula asks Leif a yes/no interrogative (line 1), to which a response is conditionally relevant. Leif’s delay at line 2, rather than indicating a dispreferred response as in ordinary conversation, displays consideration of the question and participating in the therapeutic process. He produces a preliminary “Well I’ve been thinking about that … I really have” (lines 3–5). Following this action that is composed of two TCUs, a 1.6-second silence develops. This is only broken by Leif’s “um” at line 7 and is followed by another 1.8 seconds of silence. Leif has not yet produced his actual response to Paula’s question at line 1. However, he has shown orientation to and consideration of it. He begins his actual response at line 9 and initiates self-repair mid-TCU. The 3.3-second silence in line 9 is allowed to develop by both parties, as is Leif’s next 2.1-second pause in line 10.
Leif has the right to continue his turn and the obligation to provide an adequate response to Paula’s question. Leif’s “mostly what I don’t miss is disappointing myself” does not fully answer Paula’s question; it lacks the specificity of a fully developed idea in the therapeutic process. Paula, however, at this point acknowledges Leif’s response as a fully fledged response but not one that is ready for a therapeutic intervention such as an interpretation. Instead, she also shows that she is considering–not her own response–but Leif’s reported thoughts on the matter. Paula says “hm” twice during an otherwise silent 3.7 seconds (lines 11–15). Leif then offers further specification at line 16, “In other words spending time,” which could be a grammatically and syntactically complete turn but would be more complete with a gerund phrase (which is then added as “… getting wasted …”). The 6.6-second silence in line 16 is thus treated as a pause by Paula, who merely coughs (line 17). Leif continues his turn in lines 18–19 in overlap with Paula’s cough “basically getting wasted (2.3) when I probably could have been doing something else.” At this point, Paula produces “yeah” (line 21), which displays both alignment and affiliation with Leif’s response. It is a sequence-closing third, which does not make closure mandatory but instead can serve multiple functions that participants can exploit as they continue to progress the conversation. By carefully balancing her own talk and silence, Paula is able to encourage Leif to talk and relinquishes some control of the interaction, consistent with humanistic psychotherapy. By engaging in this balancing, Paula also maintains a contemplative demeanour that is consistent with her approach to spiritually oriented psychotherapy.
Implications
The current study sheds light on how silence manifests in psychotherapy. Coupled with findings that silence may be followed by heightened emotional expression (Hill et al., 2019; Soma et al., 2023) and that silence is associated with greater satisfaction with therapy sessions (Nagaoka et al., 2013), the current findings may also shed light on how this is accomplished. Understanding that silence in everyday interactions operates in the opposite way to how it functions in psychotherapy can assist trainee psychotherapists in adapting to the therapeutic environment. It can also aid psychotherapists in identifying when further work is needed on a topic in therapy. Implications for CA methodology include understanding silence as a contextual device rather than an all-encompassing indicator of dispreferred responses.
Conclusion
From a structural point of view, the silences that are attributable to Leif in Extract 1 occur as a consequence of not yet producing the implicated response but orienting to its requirement nonetheless. The silences are also simultaneously responsible in part for conveying the gravity of Leif’s response. Just as silences that are associated with dispreferred responses in ordinary conversation demonstrate consideration and reluctance, so do silences associated with preferred responses to therapeutic interventions. A prompt response in ordinary conversation is associated with preferred responses, and prompt dispreferred responses are considered accountable and/or rude. In ordinary conversation, immediacy and spontaneity are valued. But when a client responds to an intervention such as a therapeutic question promptly, the therapist often treats this as inappropriate to the sequential environment and is likely to challenge the client’s response.
This study is one of the few that considers silence empirically in a psychotherapeutic context. Previous work in this area has been largely conjectural and based on the memories of the practitioner, and those that have examined silence empirically have tended to neglect the local sequential environment. By studying silence that occurs in actual, recorded interactions from a CA perspective, we can move forward from the top-down approach of psychoanalytic theory that has so far dominated the literature on silence in psychotherapy. By recognising that silence is not necessarily problematic in psychotherapy, and especially that silence may be a sign that the client is in fact giving due consideration to a therapeutic intervention rather than resisting it, psychotherapists and other counselling professionals can achieve better understanding of and rapport with clients.