The phenomenon of voice-hearing involves an individual’s perception of speech that others cannot hear—a personal ‘non-silence’ in the context of an interpersonal ‘silence’. It is, in fact, among the key diagnostic criteria for psychosis. Discussions of voice-hearing therefore naturally focus on the characteristics of that private non-silence: what the voices say, how often, and to whom; whether and how the voice-hearer responds; and so on. Yet, even in a context that is defined by perceptions of speech, silences continue to exist. In this chapter, we focus on references to silence in first-person accounts of voice-hearing drawn from interviews with forty voice-hearers with a diagnosis of psychosis—the Voices in Psychosis (VIP) interviews. All interviewees were using the Early Intervention in Psychosis services provided by the UK’s National Health Service in the North East of England in 2017–2018. The interviews contain frequent references to speech not occurring, as in: ‘I didn’t want to tell mum and dad’ and ‘I can’t talk back to the computer [voice]’ (Dan). These references are part of the broader phenomenon of negation (e.g. didn’t and can’t), which, as we show below, is a distinctive linguistic characteristic of interviewees’ responses.
In the study of communication, both silence and negation matter. Silence among potential interlocutors is always meaningful. It is not a mere absence of sound or speech, but rather a multifaceted and interpretable communicative act, potentially indicating, for example, compliance or resistance, the inability to express oneself, the decision and freedom to refrain from speaking, and so on (Jaworski, 1993). Therefore, silence has implications for social relationships and potential asymmetries of power and control. As such, references to silence in the interviews can inform our understanding of the experience of voice-hearing, since as Bell (2013, p. 1) asserts, ‘hallucinated voices have a social identity with clear interpersonal relevance’ and ‘are primarily experienced as social actors the hearers can relate to and interact with’.
Negation—the ability of language to express what is not—is both one of the universal and one of the most complex phenomena in language(s) (Horn, 2010). The forms and functions of negation can vary greatly, but often negation will expose a rupture with what is expected: out of the infinite number of things that are not the case, we usually choose to explicitly mention only those that need to be ruled out in a specific context, often because they may otherwise be believed to be the case. For this reason, negation is always a marked choice; it stands out and invites questions (Roitman, 2017). Like silence, negation also has implications for social power dynamics: in a particular context of communication, interlocutors may vary in terms of their ability and power to refuse, reject, or refute (Roitman, 2017, drawing on others).
In this chapter, we are particularly concerned with the implications of references to silence for the phenomenology of voice-hearing and for the portrayal of the relationships that the voice-hearer has with friends and family, and with the voices themselves.
A Linguistic Approach to the VIP interviews: Negation, Verbs of Speech, and Silence
As linguists, we began our analysis of the interviews by asking the question: what words are particularly distinctive of voice-hearers’ contributions to the interviews? To answer this, we employed a technique from the field of corpus linguistics known as a ‘keyness analysis’ (McEnery and Hardie, 2011). This involved using specialized software to compare the word frequencies in the interviews (excluding the interviewer’s questions) with the word frequencies in a set of oral history interviews collected as part of the British National Corpus (Aston and Burnard, 1998).1 This data set was similar enough to our interviews to make the comparison meaningful (as opposed to, for example, using a collection of news reports), but different enough to allow us to consider what is linguistically ‘special’ about the participants’ reporting of their voice-hearing experiences. The analysis generated a rank-ordered list of words that are ‘overused’ to a statistically significant extent in the VIP interviews, as compared with the oral history interviews.2 In corpus linguistics, these words are known as ‘keywords’.
As is often the case with this kind of analysis, the top keywords in the VIP interviews include some predictable findings (e.g. voice and voices), as well as some less predictable ones. Among the latter were the words not and n’t, i.e. the two main written realizations of the adverb for negation in English. In combination, they occur 4653 times in the interviews, and both are among the top 25 keywords.
In the VIP interviews, not and n’t (henceforth referred to simply as not) are used to negate a wide range of states of affairs. While some instances of negation occur in answer to yes/no questions from the interviewer (e.g. ‘Can you always understand what’s being said?’), most do not. Not was more often used to negate certain (expected) characteristics of the voices (e.g. ‘at first it wasn’t so much aggressive’), beliefs of voice-hearers (‘I don’t believe in ghosts and spirits’), and their emotional reactions (‘I didn’t get angry’). In 248 cases, however, not was followed by a verb related to speech. Given the focus on communication from, about, and with voices in the interviews, we decided to concentrate on these. Depending on the choice of verb and surrounding expressions, different aspects of communication were negated, such as manner of speaking (e.g. ‘They don’t talk as loudly as they did then’), topic (‘They haven’t been talking about us at all’), and addressee (‘They don’t talk to me’). We also identified a substantial minority of cases where not was used to negate the occurrence of speech itself, resulting in different kinds of ‘silence’.3
Four Types of Silences in the Experience of Voice-Hearing
By examining all cases where not was used by interviewees to negate the occurrence of speech, we identified four main types of ‘silences’, which differ depending on who is silent and how/why: (1) silent voices; (2) voice-hearer silent about the voices; (3) voice-hearer silenced by the voices; and (4) voice-hearer silent with the voices. These types of silences are all, in different ways, highly meaningful in the experiences of interviewees, and tell us something about how the participants’ social relationships are affected by their voice-hearing experiences. More specifically, what we call silent voices are a particular aspect of the phenomenology of voice-hearing, whereas the other three types are all related to aspects of voice-hearers’ interactions with the voices and with other people.
Silent Voices
Studies of voice-hearing have shown that ‘hallucinated voices are usually experienced as having identities and making coherent communicative speech acts’ (Bell, 2013, p. 1). However, the data show that, in some cases, not is used to negate the occurrence of speech on the part of the entities that are referred to throughout the interviews as ‘voices’. This results in what we oxymoronically call Silent voices.
For example, while talking about his experience of hearing the voices of absent relatives, Hugh mentions an uncle in particular:4
Extract 1
Hugh: [M]e Uncle’s just, just there . . .
Interviewer: Right.
Hugh: . . . not moving, he’s just there, in a prayer stance . . . Ehm, and that’s comfort. So like me neighbour said he’d be freaking out, and I said, no, I could see him . . .
Interviewer: Yeah.
Hugh: . . . but he didn’t say anything.
Another interviewee, Xander, says that he hears three voices and, as the interview progresses, describes one of them as ‘the quiet one’. When asked directly about whether he would ‘ever have a conversation with the quiet voice’, he replies:
Extract 2
Xander: I would, but at times because he doesn’t speak, it’s hard to know if he’s there.
Both cases exemplify an aspect of the phenomenology of voice-hearing that undermines the use of the term ‘voice’ itself, namely the experience of the presence of a personified agent who could potentially speak but does not, or not always (see Alderson-Day and Fernyhough, 2016). The silence that results from non-speaking voices can be experienced in different ways. In Extract 1, Hugh describes the silent presence of his uncle as ‘comfort’ and reports himself as contradicting his neighbour’s imagined reaction of ‘freaking out’ in that situation. In Extract 2, Xander seems to present the fact that one of the voices does not always speak as a missed opportunity for social interaction with that voice. In both cases, however, the silence is meaningful, both as an experience in the moment and to reiterate that the entities labelled as ‘voices’ are not always perceived to ‘speak’. In contrast, in the three other types of silence, it is the voice-hearer’s speech that is being negated.
Voice-Hearer Silent about the Voices
Some of the instances of negated speech occurred as part of interview questions on whether voice-hearers talk about their voices with people in their lives. Ryan, for example, says:
Extract 3
Ryan: [W]hen I do, they do get upset. So it is, it’s quite a difficult thing to do, unless I sugar coat it quite a lot and then they can, they can sort of comprehend it but I find it easier just not to talk about it.
In response to similar questions, Xander mentions that he has talked about the voices with his mother and sister, but also adds:
Extract 4
Xander: [W]ith my dad, I don’t really talk to him about it, just like, it’s always been like that with my dad. . . . with my gran and my granddad, I don’t talk to them about it, because I don’t want them to worry.
Here the negation of speech activities does not result in prototypical silence in the sense of a total absence of the sound of speaking, but rather in a form of self-censorship that constitutes a social kind of silence: the absence of communication about the voices in interactional contexts where such communication is, in principle, possible, and arguably even desirable.
The reasons for this kind of silence highlight both the sensitivity of voice-hearing as a topic and how closely intertwined the voice-hearing experience is with the person’s most intimate social relationships. As a topic, the experience of voice-hearing is, to some extent, ‘untellable’ (Jaworski, 1993): it is hard to express (‘I sugar coat it quite a lot and then they can, they can sort of comprehend it’); it is private and subjective, and therefore not suitable for people with whom communication is not usually particularly intimate (‘it’s always been like that with my dad’); and it is potentially upsetting for people with whom mutual relationships are more intimate and caring (‘they do get upset and I don’t want them to worry’).
This kind of silence therefore reflects differences in the voice-hearer’s existing relationships with other people, and in the ways in which these relationships are, in turn, affected by the voice-hearer’s decision to share the experience of voice-hearing or not. While the underlying self-censorship is driven by the wish to protect others, as well as by one’s own public image, it also obviously deprives the voice-hearer of the empathy and support that they could otherwise potentially receive from people in their lives. The detrimental effects of ‘self-silencing’ have been reported as affecting interpersonal relationships with respect to various health complaints (Jack, 1991).
Voice-Hearer Silenced by the Voices
In other cases, it is the voices that are presented as preventing hearers from engaging with others in their social environment. Ian, for example, says that his voice ‘tells us not to talk to people’ and that he ‘hardly see[s] friends now’. Similarly, Will quotes his own voice as telling him:
Extract 5
Will: Don’t talk to that person because they don’t like you, or don’t . . . ehm . . . speak to them because you don’t, you can’t trust them.
In the previous types of silence, we saw how hearers are isolated through their own reluctance to discuss their voices, as a difficult or ‘upsetting’ topic. Here, the voice-hearers are discouraged from engaging with others altogether, despite their desire to do so. In the example above, Will both receives explicit instruction not to talk to someone and reports a more implicit form of oppression in which his confidence is undermined and the voice creates distrust (‘you can’t trust them’). In addition, in Extract 6, although the voice does not explicitly forbid talking to others, the name-calling that Carl experiences from the voices affects his participation in conversations with other people:
Extract 6
Carl: When I’m talking to people and that, and . . . I don’t want to talk because I feel embarrassed.
Whether or not voice-hearing itself is experienced as distressing, this particular type of oppressive relationship with voices—where the voices prevent people from leading their lives as they would wish to—is known to cause distress (Varese et al., 2017).
Voice-Hearer Silent with the Voices
The final type of silence consists of cases where the voice-hearer does not ‘respond’ or ‘talk back’ in the voice-hearing situation, even though the voice is speaking. This group includes examples such as:
Extract 7
Dan: I can’t talk back to the computer one. . . . That’s never happened.
Extract 8
Hugh: And during the day I don’t respond . . . to it, well I try not to respond to it anyway. . . . I’m just hearing it and doing that thing of putting it in the box . . . to deal with later. Because that’s the only way I can deal with it.
In these cases, not responding to the voices is potentially a strategy that people have adopted, a kind of coping mechanism, as Hugh says. It is an exercise of power on the part of the voice-hearer. However, the voice-hearer’s lack of response is hedged using the modal verb can’t (Extract 7) and the semi-modal try (Extract 8). These, once again, suggest a potential thwarting of the voice-hearer’s ability to behave in a way that they wish to. Can’t suggests a lack of ability where the ability might be welcome, while try implies unsuccessful attempts at doing something, in this case not responding to the voices. This becomes clearer in the following examples.
Extract 9
Alex: On a night time I try not to talk back. . . . Ehm, so sometimes if I try to ignore them, I mean I don’t actually get to ignore them.
Extract 10
Zara: I don’t want to answer them. . . . Because at first I was saying, I would say, hello, and then people would say, there’s nobody there.
These examples appear to be the converse of those where the voice-hearer is silenced by their voices. Here the voices force the hearer into non-silence, when the person would otherwise wish to remain quiet. Being prevented from staying silent has similar implications for the ability to live the life one wishes to, as in the case of being silenced by the voices, but a different aspect of the person’s potential coping mechanism is affected. Above, the voices limit social engagement, and therefore social support, while here they interfere with the person’s ability to exert control through refusing to engage with the voices. This is especially important as voice-hearers sometimes report that speaking back to the voice does not help the situation, but ‘just sort of happens because it can be so overwhelming’ (Dan).
Our discussion of the last three types of silences—between voice-hearers and their support network of family and friends (Voice-hearer silent about the voices), and between voice-hearers and their voices (Voice-hearer silenced by the voices; Voice-hearer silent with the voices)—emphasizes the importance of dialogue in the voice-hearing experience, which is consistent with a range of therapies for psychosis (Avdi et al., 2015). As Avdi et al. (2015, pp. 330–1) explain:
through dialogue . . . the family or network’s psychological resources are mobilized, and participants regain their voice and assume positions of increased agency with regard to the symptoms. Dialogue is considered to allow strong emotions to be expressed, new words for difficult experiences to be jointly created, and new understandings to emerge.
Our analysis has drawn attention to the ways in which voice-hearers are inhibited from participating in dialogue (e.g. because the ‘voice’ does not speak, or because the voice warns them against doing so) or set out to reduce their distress by not engaging in dialogue with the voice, and how this reflects their level of agency. This can have consequences for voice-hearers’ engagement with (dialogically informed) therapies and their ongoing relationships with their voices.
Conclusion
In this chapter, we have described how a (computer-aided) linguistic analysis of the forty VIP interviews first drew our attention to negation as a distinctive feature of interviewees’ contributions, and then enabled us to identify the more specific phenomenon of references to silence via the negation of speech verbs. We have distinguished between four different types of silences and shown that, as has been observed in other contexts, they amount to much more than the absence of sound and speech: depending on who is silent, with whom, and why, silences can reveal phenomenological nuances, as well as pressures and power asymmetries in the relationships that interviewees have with their voices and with important people in their lives. These can, in turn, reflect or lead to distress. In this sense, references to silence deserve as much attention as references to the perception and/or occurrence of speech in accounts of the lived experience of voice-hearing.
- 1
- 2
We combined a measure of effect size, LogRatio, with a measure of statistical significance often used in corpus linguistics, Log Likelihood. The LogRatio values were 0.71 for n’t and 1.32 for not. Log Likelihood was set at a level that corresponds to p <0.0001.
- 3
Forms of the words ‘silent/silence’ occurred only four times in the interviews.
- 4
Relevant sections have been italicized for emphasis.
References
Aston
G., and Burnard
L. (1998). The BNC Handbook: Exploring the British National Corpus with SARA. Edinburgh: Edinburgh University Press.
Avdi
E., Lerou
V., and Seikkula
J. (2015). Dialogical features, therapist responsiveness, and agency in a therapy for psychosis. Journal of Constructivist Psychology, 28(4), 329–41.
Horn
L. R. (2010). The Expression of Negation. New York, NY: Walter de Gruyter.
Jack
D. C. (1991). Silencing the Self: Women and Depression. Cambridge, MA: Harvard University Press.
Jaworski
A. (1993). The Power of Silence: Social and Pragmatic Perspectives. London: Sage Publications.
McEnery
T., and Hardie
A. (2011). Corpus Linguistics: Theory, Methods and Practice. Cambridge: Cambridge University Press.
Roitman
M. (2017). The Pragmatics of Negation. Negative Meanings, Uses and Discursive Functions. Amsterdam: Benjamins.
Varese
F., Mansell
W., and Tai
S. J. (2017). What is distressing about auditory verbal hallucinations? The contribution of goal interference and goal facilitation.
Psychology and Psychotherapy: Theory, Research and Practice, 90(4), 720–34. [
PubMed: 28726350]
Elena Semino is Professor in the Department of Linguistics and English Language at Lancaster University and Director of the ESRC Centre for Corpus Approaches to Social Science. She applies qualitative and quantitative linguistic methods to texts concerned with the lived experience of illness. She has a specific interest in the language used in first-person accounts of psychosis and voice-hearing.
Luke Collins is a Senior Research Associate in the Centre for Corpus Approaches to Social Science at Lancaster University. His research interests lie in the linguistic representations of mental health and illness in clinical contexts. Using methods in corpus linguistics, he examines patterns in the reported user experiences of clinical interventions to assess their impact.
Zsófia Demjén is Associate Professor of Applied Linguistics at University College London. She explores the implications of how language is used to describe experiences of illness, including in contexts associated with cancer, depression, and ‘psychosis’. She recently led a project investigating the extent to which implicit power relationships in the language people use to talk about their voices can predict their likely level of distress.
