Discussion Protocol for Alleviating Epistemic Injustice: The Case of Community Rehabilitation Interaction and Female Substance Abusers

: Substance-abusing women are vulnerable to speciﬁc kinds of epistemic injustice, including stigmatization and discrimination. This article examines the development of the epistemic agency of female substance abusers by asking: How does the use of a formal discussion protocol in community rehabilitation interaction alleviate epistemic injustice and strengthen the epistemic agency of substance abusers? The data were collected in a Finnish rehabilitation center by videotaping six group discussions between social workers, peer support workers, and rehabilitation clients with substance abuse problems. Of these data, one recorded group discussion between four female participants—two rehabilitation clients, a peer support worker, and a social adviser—was used in this paper. Using conversational analysis, the ﬁndings indicate that, through the collaborative activities of sharing experiential knowledge about substance abuse and discussing the experiences of abuse in the rehabilitation interaction, substance abusers can develop novel ways to strengthen their epistemic agency by enhanced self-awareness. The discussion protocol is an epistemic tool that professionals and clients can learn to use in ethically and epistemologically successful ways in interaction. The use of a discussion protocol is an example of social professionals’ clinical knowledge of intensifying collaboration and sharing experiential knowledge in community rehabilitation and other substance abuse services.


Introduction
Recent discussions of moral social epistemology (Fricker 2007;Townley 2011) have provided fruitful theoretical concepts and insights into the analysis of epistemic injustice toward marginalized individuals and groups in society. Miranda Fricker's notion of epistemic injustice refers to the unfair downgrading of credibility as a type of harm done to individuals or groups regarding their ability to contribute to and benefit from knowledge. Epistemic asymmetries between social-work professionals and clients can easily make the client epistemically vulnerable in interactions if professionals are not sensitive to and aware of the power of their authority, which exists as a result of their scientific and clinical knowledge. While discussions on patients' testimonies being dismissed and epistemically downgraded are well documented in health care (e.g., Carel and Kidd 2014;Kurs and Grinshpoon 2018), there is far less empirical and theoretical research on how clients' experiences and subjective testimony are received and recognized in rehabilitation and other social services (Baillergeau and Duyvendak 2016).
In this article, we examine the social production of the epistemic agency of female substance abusers in community rehabilitation interaction. Our starting point is the observation that substance-abusing women are vulnerable to specific kinds of epistemic injustice, such as testimonial injustice, stigmatization, and discrimination. In addition, substanceabusing women's self-distrust in their experiences and knowledge as epistemic agents is a complex combination of shame, self-accusation, feelings of being oppressed, and a lack of proper concepts to express their own feelings. This, in turn, is an aspect of hermeneutic injustice, in Fricker' (2007) terms. Frequent experiences of testimonial injustice may exacerbate the hermeneutic injustice by crushing the epistemic agent's self-confidence and leading to their taking the pronouncements and conceptual frameworks of social and healthcare professionals and other authorities as the only possible viewpoints. This vicious circle is exacerbated by another aspect of hermeneutical injustice: dominant epistemic resources and conceptual frameworks are dominated by the neoliberal conception of the individual agent as self-sustained, entrepreneurial, and economically rational (e.g., Adams et al. 2019). This conception leaves little room for understanding and articulating the epistemic value of the lived experience of addiction.
Social professionals' epistemic humility, as a form of epistemic justice, may be the kind of disposition required to correct harmful prejudices, but it is hardly enough to restore substance-abusing women's self-confidence as epistemic agents. Instead of stressing here professionals' personal knowledge or empathy skills, our key question is: How can the use of a discussion protocol in community rehabilitation interaction alleviate epistemic injustice and strengthen the epistemic agency of substance abusers? We assume that rehabilitation practices and protocols may prioritize some kinds of evidence and ways of knowing while excluding certain clients' testimonies from epistemic consideration. Thus, the protocol for community rehabilitation is considered here as a peculiar epistemic device that professionals and clients can learn to use ethically and epistemologically.

From Epistemic Injustice to Experiential Knowledge
Built on feminist epistemology, sociologist Ettorre (2015) discusses substance-abusing women's experiences, showing how "normal" everyday life and interactions with other people are foreclosed to women substance abusers. Substance-abusing individuals appear to be those who have failed in regulating their external behaviors and internal desires. Women's substance abuse has been considered an emblematic failure of gendered performativity, stigmatizing substance-abusing women as "deviant bodies" (Ettorre 2007). As Campbell (2000) and Ettorre (2015) argued, these women are frequently deemed epistemologically incredible; they continue to be constructed as willfully wayward women who are morally corrupt and deviant in socially unacceptable ways. Even though social professionals may assume that substance abusers provide relatively accurate accounts of their own situation, their capabilities and motives in seeking better living conditions are frequently considered unrealistic or unreliable (see also Leung and Lam 2019).
According to Fricker (2007), epistemic injustice means that an agent is wronged specifically in their capacity as a knower, as someone who can learn things about the world and can be relied on to describe those things to others accurately-in social-epistemological terms, they give testimony. There are two basic forms of epistemic injustice: testimonial and hermeneutical injustice. Testimonial injustice occurs when the wronged party suffers a credibility deficit due to harmful social stereotypes. Their testimony is not believed, taken seriously, or considered interesting or relevant. In a patriarchal society, for example, female substance abusers can be deemed too emotionally unstable to be able to present a neutral and considered point of view. Hermeneutical injustice takes place when an agent (and, by extension, the rest of society) is unable to understand their own situation due to gaps in the shared hermeneutic resources-that is, the ideas taken to be common knowledge and socially constructed conceptual frameworks in a given society (Fricker 2007). The dominant narratives of substance-abusing women, as explained by Ettorre (2015), are implicated in epistemic injustice in both ways. They posit addicts as epistemically inferior and unreliable (testimonial injustice), and their explanation for addiction forecloses the possibility of substance abusers understanding their experiences and articulating them to others. They, thus, both maintain and create gaps in the social hermeneutic resources we have for understanding addiction. Fricker (2007, p. 3) reminds us that being sensitive to a "socially situated account" means that all humans are recognized as epistemic agents, as beings who can know things and be responsible for their own knowledge to a reasonable extent, and at the same time as "operating as social types who stand in relations of power to one another". Therefore, all individuals are conceived of as epistemic agents possessing information and skills relevant to their everyday lives (Carel and Kidd 2014;Kurs and Grinshpoon 2018). As epistemic agents, substance abusers' knowing falls into categories of tacit/personal knowledge (Polanyi 1967), experiential knowledge (Borkman 1976), and embodied knowledge (Oliver and Lalik 2000). Borkman (1976) considers experiential knowledge a resource that helps people in precarious situations to respond to uncertain futures. Experiential knowledge relies on lived experiences that are difficult for outsiders to comprehend, but its significance becomes more accurate when it is discussed in the context of embodied knowledge. Addressing sensations, feelings, and emotions, the notion of embodied knowledge captures knowing in and through the lived body (Parviainen and Aromaa 2017), in particular, in relational situations, when the body is physically controlled, abused, or judged by others (Oliver and Lalik 2000;Ettorre 2015). Baillergeau and Duyvendak (2016) stated that the rise of experiential knowledge as a resource for intervention was born of the critique of the hegemonic power of "expert knowledge". Social professionals can utilize their knowledge and skills to provide advice and guidance in their clients' best interests within the standards of normative citizenship, but their expertise does not always result in effective interventions. In many cases, social workers are seen as using their expert knowledge to "treat" and "handle" clients in precarious situations, but they cannot provide embracing experiential knowledge as a resource for action, as do peer support workers. The latter's embodied knowledge enables clients to recognize similarities and differences to their own experiences of hardship or maltreatment. Peer support workers' knowledge provides resources that are helpful in building trust with marginalized people and that aid them in tackling social problems more effectively than do professionals (Baillergeau and Duyvendak 2016).
While embodied knowledge is developed along with an understanding of one's own substance abuse and closely grounded in rehabilitation processes, professionals need tools to combine peer support workers' embodied knowledge in a way that resonates with their clients' needs and motivations. We hypothesize that, in community rehabilitation, the use of a discussion protocol is one way that social professionals' clinical knowledge intensifies collaboration and the sharing of experiential knowledge in rehabilitation interaction.

Materials and Methods
The research data were collected in a Finnish rehabilitation center for substance abusers in summer 2015. The total data include video recordings of six group discussions between social work professionals, peer support workers, and rehabilitation clients with substance abuse problems. Of these, one video recording lasting an hour and sixteen minutes was used in the analysis. There were four female participants in the recorded group discussion under scrutiny: two rehabilitation clients, a peer support worker, and a social adviser. The video recording chosen for analysis is the only one that includes female substance abusers. The rehabilitation clients in the other five recordings were males, and the discussions used different protocols than those used with the females. During the data collection, no other female clients agreed to participate in the research-presumably because women found it difficult to deal with their addiction and the shame associated with it in the presence of researchers.
The rehabilitation center in question follows the principles and practices of community rehabilitation to assist clients in recovering from substance addiction. The clients are not treated as passive agents but as people who are able to be responsible for the community and for each other's rehabilitation together with the staff. Decisions about the community's affairs are made collaboratively in regular community meetings, and open communication between professionals and clients is emphasized. Social interaction, sharing experiences, and peer support are utilized as essential resources in treatment and recovery (De Leon 2000).
The participants in the group discussion used a specific discussion protocol as a tool for sharing and conversing about their views and experiences related to the four main themes of the protocol: advantages of substance use/non-use and disadvantages of substance use/non-use. The protocol also included ten sub-themes to support participants in structuring their experiences: human relations; emotional life; use of time; studying/working life; health; sexuality; finances; attitudes and values; beliefs about oneself, about other people, and about life; other issues. The rehabilitation clients' task was to ponder and write down their personal thoughts and experiences about the themes before discussing them in the group. In considering the advantages of substance use, for example, the clients could think about the positive effects of using substances on their human relations, emotional life, and so on.
In the group discussion, the participants were seated informally around a table, and all had two sheets of paper. One paper included a four-part "grid" classified according to the four main themes with fixed spaces for clients' advance notes. The other paper contained concise and practical instructions on the use of the protocol and a list of the ten sub-themes. At the beginning of the group discussion, the social adviser, with the help of the peer support worker, verbally instructed the rehabilitation clients on the use of the protocol. The instructions were presented as loose guidelines rather than strictly defined rules. The idea was to go through and discuss one main theme at a time by means of various sub-themes. Instead of time limits, the protocol use was meant to last as long as was appropriate and pleasant for the clients.
The discussion protocol was also used as a tool to conduct motivational interviewing in the group discussion. Miller and Rollnick (2013, p. 12) define motivational interviewing as "a collaborative conversation style for strengthening a person's own motivation and commitment to change". One vital aspect of motivational interviewing relates to partnership. Motivational interviewing is not something done by an expert with a passive recipient, but is seen as active collaboration between professionals and clients, as experts on themselves. The interviewer sometimes asks questions, but mostly they listen to what the clients have to say. In this way, the interviewer aims to see the world through the client's eyes rather than superimposing their own vision on the client (Miller and Rollnick 2013).
At the time of data collection, the use of the discussion protocol was an established practice in the rehabilitation center, and it was still in use in summer 2019. We became interested in analyzing the protocol because it was oriented as a prominent resource in the organization of rehabilitation interaction. We wanted to know, specifically, how using a discussion protocol helps participants to talk about the experiences of substance abuse and therefore to alleviate epistemic injustice.
The method used to analyze the recorded data was conversation analysis (CA). CA is an approach that investigates how people accomplish social activities through talk and embodied actions (Sidnell and Stivers 2013). The recording was transcribed following Jefferson's (2004) transcription system (see Appendix A). The transcripts include translations from Finnish to English. Our analysis focuses on the phases where the participants in the group discussion first used the protocol, then disengaged from using the protocol, and finally reflected on its use in situated rehabilitation interaction. The analysis considers the various multimodal resources employed by the participants in the production of mutually intelligible social action, including talk, gesture, gaze, and embodied manipulations of objects, such as protocol papers (Goodwin 2000;Arminen et al. 2010).

Results
The empirical analysis starts from the premise that substance-abusing women face epistemic injustice and that, through collaborative activities of sharing experiential knowl-edge about substance abuse and discussing their experiences of abuse in community rehabilitation interaction, they find ways to strengthen their epistemic agency. In line with this, the analysis focused on examining participants' use of the discussion protocol in relation to sharing and talking about their lived experiences in rehabilitation interaction. The analysis aims to show, first, how rehabilitation clients (RC1 and RC2) together with the social adviser (SOA) and peer support worker (PSW) use and orient to the protocol in describing their thoughts and experiences on matters related to substance abuse. Second, the analysis describes how participants, instead of sticking to the use of the protocol, occasionally and jointly detach from using it to create conversational space for a more extensive collaboration and, therefore, to alleviate hermeneutic injustice through articulating their tacit experiences. To achieve these aims, the study is divided into three subsections that include five data examples presented in chronological order. The first subsection introduces a case in which the sharing of personal and intimate experiences is jeopardized due to problems with the use of the discussion protocol in rehabilitation interaction. The second subsection presents two cases in which hermeneutic injustice is redressed successfully. These cases introduce the methods the participants employ to disengage from using the protocol to create opportunities for more collaborative experience sharing in rehabilitation interaction. The final subsection outlines two cases in which the participants, together, reflect on the use of the protocol in relation to talking and sharing in rehabilitation interaction.

Doubts about the Protocol
The first subsection introduces an example in which participants doubt the protocol because the practice of sharing one's own experiences is at risk while using it. The case is presented in three parts, each of which highlights a different aspect of the phenomenon under scrutiny. In Excerpt 1a, the participants orient to using the protocol in a way that-rather than maximizing their collaboration with each other-minimizes collaborative activities between them. In Excerpts 1b and 1c, where minimal collaboration continues, the rehabilitation client's problems in her interaction with the protocol impede the activity of experience sharing, simultaneously posing a threat to the usefulness of the protocol. Following CA procedures, the excerpts are analyzed by asking: How does the rehabilitation client describe her lived experiences while using the protocol? How do other participants receive the client's lived-experience descriptions? What implications does the use of the protocol (i.e., the production and reception of lived-experience descriptions) have for the sequential organization of rehabilitation interaction?
In Excerpt 1a, the participants begin using the discussion protocol in rehabilitation interaction. The excerpt shows how the ways in which participants orient to the use of the protocol can minimize collaboration among them, thereby jeopardizing their chances of strengthening their epistemic capabilities. At the beginning of the excerpt, the participants negotiate in which order they will use the protocol, and RC1 agrees to start (lines 1-10 Before describing her experiences concerning the advantages of substance use, RC1 humorously criticizes the idea that using substances could have any advantages at all. RC1, RC2, and SOA subsequently laugh aloud together, simultaneously displaying their mutual affiliation and alignment (lines 10-16). In lines 17-18, RC1 starts to describe the advantages of substance use in terms of human relations. She views substance use as a social activity-an activity we may easily engage in in many social situations and events (lines 19-27). RC1 continues her description by referring to the positive effects of substance use, including feelings of relaxation and pleasure; she also associates substance use with having fun (lines 27-34). RC2, PSW, and SOA indicate their involvement and attention in the ongoing conversation with various non-verbal response tokens, such as regularly gazing toward the speaker, RC1, and nodding their heads (data not shown); SOA also uses the verbal response token "mmm" (lines 25 and 32) to maintain her conversational position as a listener. In using the protocol, therefore, the verbal-level communication is organized as a joint effort between two participants in interaction: SOA and the client using the protocol. In addition, and interestingly, the inner experiences described by the client through the protocol are not discussed or elaborated further but are left hanging. It is important to note that minimal collaboration of this kind goes on throughout Example 1. In Excerpt 1b, RC1 continues discussing the theme of the advantages of substance use by means of various sub-themes, including use of time, studying/working life, health, sexuality, finances, and attitudes and values. As shown, her inner experiences with the advantages of substance use do not fully resonate with the structure provided by the protocol.
Example 1 (.) 112 RC1: certain things so that you take it more relaxed so 113 (1.5) ((7 lines omitted)) In Excerpt 1b, RC1 highlights the key problem she experiences in using the protocol: certain sub-themes included in the protocol, such as use of time (lines 62-66), studying, working life, health (lines 73-76), and finances (lines 87-92), do not resonate or align with her lived experience of the advantages of substance use. RC1 presents these problematic sub-themes by systematically using the first-person pronoun ("I don't think", lines 62-64; "I don't feel", line 75; "I don't include", line 90) and humor (see laughter particles in lines 66, 74, and 92), thus mitigating the criticism she simultaneously voices of the structure of the protocol. In lines 76-79, however, RC1 revises her description concerning the advantages of substance use for health by providing the general idea that a small amount of alcohol consumed daily can be "quite healthy too". Other sub-themes RC1 relates to her experiences include sexuality as well as attitudes and values: using substances helps her to become sexually "freer" (lines 82-83) and to adopt a "more relaxed" attitude toward "certain things" (lines 93-112). SOA, PSW, and RC2 participate in the conversation through embodied means by, for example, regularly shifting their gaze between RC1 and their own papers (data not shown). Once again, it is SOA who uses minimal verbal responses to keep the conversation going: "yeah" (lines 67, 91), "yes" (line 70), "mmm" (lines 77, 85, 110), "oh yeah" (line 80). In Excerpt 1c, RC1 concludes discussing the theme of the advantages of substance use with the sub-theme of beliefs about oneself, about other people, and about life. The excerpt shows how there is a lack of shared understanding between the protocol and the client using it.
Example 1. Excerpt 1c. (.) 249 RC2: yeah (0.6) well ((oriented to her papers)) 250 (1.0) 251 RC2: relieves anxiety tension ((continues)) In Excerpt 1c, RC1 raises the issue of not knowing. In other words, she does not know how to describe her personal experiences with the ways that substance use could be advantageous for her beliefs about herself, about other people, and about life (lines 121-131). In an effort to resolve the problem, SOA intervenes by explaining that, if drinking makes the client "more relaxed and funnier", then she can also see herself in the same way "if that is like the thing" (lines 160-171). In her description, SOA uses wording-"more relaxed" (lines 164, 166)-similar to the phrase used by RC1 earlier (see Excerpt 1b, lines 106, 112) to provide some evidential foundation for her interpretation. SOA also downgrades her epistemic access to RC1's inner experiences with uncertainty markers, including the use of the adverb "probably" in line 162 and the "If" clauses starting in lines 162 and 168 (Weiste et al. 2015). RC1 completes her lived-experience description of the advantages of substance use with "so that kinds", to which RC2 responds in a low volume, " • yeah • " (lines 239-241). With their gazes directed toward RC2, RC1, SOA, and PSW prompt her to continue her turn at talking (line 242). RC2's second "yeah#", produced in line 243, exhibits her readiness to shift from recipiency to speakership (Jefferson 1984). Accordingly, through the coordination of her talk and non-verbal activity, RC2 seeks confirmation that it is her turn to start describing her personal experiences in terms of the protocol (line 243). Both PSW and SOA confirm RC2's understanding of the upcoming course of action (lines 244-247). RC2 acknowledges their talk with minimal responses, including "yeah" (lines 246, 249). With the particle, "well", in line 249, RC2 indicates that the transition to the activity of describing her lived experiences is about to take place (Sorjonen and Vepsäläinen 2016; see Example 2).
In Example 1, the activity of using the discussion protocol overrode collaboration and experience-sharing between participants, simultaneously jeopardizing the reduction of epistemic injustice in rehabilitation interaction. The use of a discussion protocol was accomplished through minimal collaboration among participants. Although all participants contributed to the client's protocol use through their talk and non-verbal activities, it was the social adviser who had the most essential role in verbally receiving the client's lived-experience descriptions. The verbal-level communication proceeded mainly between the social adviser and the client using the protocol, the former contributing with minimal responses to the lived-experience descriptions by the latter. While being heard and understood, the inner thoughts and experiences described through the protocol were not further discussed or elaborated on in the rehabilitation interaction. The minimal collaboration described can be partly explained by the participants' orientation toward the protocol use on a turn-by-turn basis (Sacks et al. 1974). As shown, when one client was describing her experiences using the protocol, the other client and the peer support worker were mainly silent, showing their attention and participation through gaze and other embodied behaviors. Only when one client had explicitly finished describing her inner experiences through the protocol did the other client take the floor and describe hers.
From an institutional point of view, the discussion protocol is a professional artifact used under the guidance and direction of institutional representatives, including the social adviser and peer support worker. Thus, since the participants' main orientation was toward the use of the protocol as a turn-by-turn activity, the verbal interaction in the group discussion was principally accomplished as a collaboration between two participants: the social adviser (representing the institution) and the client using the protocol.
Instead of stories of experiences, the situated use of the protocol seemed to evoke relatively short descriptions of the client's personal thoughts and experiences in a list-like manner. There were also some observable problems in the situated production of the livedexperience descriptions. For example, there was a lack of resonance between the structured discussion protocol and the world of the lived experience of its user, as well as a lack of shared understanding between the protocol and the client using it. Due to the problems in her interactions with the protocol, the client occasionally became incapable of delivering descriptions of her personal experiences of substance abuse. This, in turn, blocked some opportunities for alleviating epistemic injustice in rehabilitation interaction while at the same time made it possible to take critical distance from the protocol, to approach it as a tool to be controlled by the participants rather than the participants being controlled by the protocol.

Methods for Alleviating Hermeneutic Injustice
In the following data Examples (2-3), the group discussion proceeds beyond the protocol in a way that creates space for more extensive collaboration and experience-sharing between participants in rehabilitation interaction. The examples introduce the methods or strategies that the participants use to steer the conversation toward collaborative discussion and the elaboration of participants' personal experiences and, therefore, to alleviate hermeneutic injustice. As in the prior case, the analysis of the Examples (2-3) focuses on the situated use of the protocol and asks how rehabilitation clients' lived-experience descriptions are produced and received in group discussion and what implications the production and reception of lived-experience descriptions have for the sequential organization of rehabilitation interaction. In Example 2, it is RC2's turn to describe her experiences on the theme of the advantages of substance use. To enhance collaboration, RC1 raises one element from RC2's description-namely, the temporal expression, "momentarily"-for further discussion in rehabilitation interaction.
Example 2. Compared to the prior lived-experience description offered by RC1, RC2 produces in lines 1-5 a more concise and compact description about how substance use is advantageous for her. As in the previous case, the SOA displays engagement and attention with her minimal response token, " • yeah • " (line 7). With the continuation of her turn ("I didn't really find, like, other advantages", lines 9-10), RC2 not only completes the activity of describing her inner experiences of the advantages of substance use, but also provides an explanation for her relatively short description. Subsequently, in line 12, RC1 displays her orientation to collaboration and the sharing of experiential knowledge by addressing her turn to RC2. With her pre-announcement ("Do y'know what came to me when you said momentarily?"), RC1 lays the groundwork for the upcoming production of an announcement (Schegloff 2007). After RC2's go-ahead response, "mmm" (line 13), RC1 announces that the temporal expression "momentarily", used by RC2 in her earlier turn, is "the thing exactly" (line 15), simultaneously demonstrating her identification with RC2's experiences. RC2 agrees with the prior announcement ("mmm", "precisely", line 17) and further reinforces her agreement by elaborating the temporal aspects of substance use (lines 17-21).
For the following six and a half minutes or so (data not shown), the participants continue discussing and sharing their inner experiences about, for example, the reasons for their substance abuse, their mental problems co-occurring with substance abuse, and the operation and usefulness of the particular peer support group. In this way, they are simultaneously capable of alleviating epistemic injustice through practices of social interaction. Compared to Example 1, this conversation proceeds more spontaneously and interactively, without the participants' observable use of the protocol. RC2 then reorients to her papers and repeats that she has no more advantages of substance use to discuss (see also lines 9-10 above). The participants subsequently shift their orientation back to the activity of protocol use by starting to discuss the theme of the advantages of substance non-use, as shown at the beginning of Example 3. There, SOA's treatment of RC1's livedexperience description as deficient and in need of further discussion is the method for extending collaboration and minimizing the risk of epistemic injustice in rehabilitation interaction.
Example 3. [yes:: yeah:: ((continues)) In lines 1-2, SOA reads aloud from her papers the theme to be discussed ("advantages of substance non-use") and reformulates her prior turn as a question ("Why • is it worth it • ?"). By clearing her throat in overlap with SOA's question, RC1 indicates that she is about to take the floor (line 3). Then, in lines 5-9, RC1 provides a lived-experience description of the ways in which substance non-use is advantageous for her, followed by SOA's verbal response token, "mmmm" (line 10). In lines 12-13, instead of sticking to the theme previously presented by SOA, RC1 uses the negative consequences, or disadvantages of substance use (stomach symptoms, lost appetite), as a resource in producing her description. RC1 then completes her description with the concluding utterance ("so all these kinds of, like, basics", lines 13-14). Next, SOA acknowledges RC1's prior turn with the minimal verbal response, "mmm" (line 16). In line 18, RC1 extends her prior turn by emphasizing the great importance of the basics. To check the level of shared understanding, SOA offers an interpretation, or a candidate understanding, of the basics that RC1 referred to in her prior turn ("so that (.) wellbeing and health", lines 20-23). After RC1's confirmation of the candidate understanding in line 22, RC2, SOA, and RC1 verbally indicate that they have an intersubjective understanding of the advantages of substance non-use for RC1 (lines 24-27).
SOA subsequently treats RC1's lived-experience description as incomplete by highlighting the sub-theme "human relations" from the protocol (lines 29-30). In this way, she is simultaneously displaying her orientation to further collaboration and sharing experiential knowledge in rehabilitation interaction. In overlap with RC1's minimal response, "mmmmm", SOA asks her whether substance non-use helps in human relations (lines 32-33). However, RC1 does not answer the question (pause in line 34). SOA orients to the absence of an answer by pursuing it with a reformulated question, "Ha-(.) have you screwed up • (a)ny relationssh(ips) • ?" (lines 34-35). Implicit in SOA's question is the premise that substance use (not substance non-use) is detrimental to human relations. RC1 delivers an overlapping negative response (" • no • ") in line 36. With the surprise token "whe::w" (line 38), SOA treats RC1's prior talk as unexpected or counter to expectations (Wilkinson and Kitzinger 2006). In line 39, RC1 starts to explain why she has not screwed up any relationships due to substance use, with the reason being that drinking "a bit too much" happens when she is "alone at home" (line 47-48). The talk in line 52 ("yes:: yeah::") indicates that SOA understands, here and now, why heavy drinking has not had a negative effect on RC1's relationships.
In the first half of Example 3, the verbal conversation once more appears as a two-way street between the client using the protocol (here, RC1) and SOA. In the latter half of it, where the rehabilitation interaction proceeds beyond the protocol and moves toward collaboration and the alleviation of epistemic injustice, RC2 (line 40) and PSW (lines 50-51) also contribute to the ongoing conversation with alternatives to non-verbal means, or minimal verbal responses only. For the next couple of minutes (data not shown), RC1 describes her problematic drinking and how it causes her close relatives (e.g., her children) concern. She simultaneously uses the protocol sub-theme "emotional life" as a resource in describing her feelings of guilt that are aroused by drinking. Soon after, RC1 reorients to her papers and asks whether they should take "some of these", probably referring to the sub-themes of the advantages of substance non-use. SOA dismisses the use of the protocol by directing the conversation to the relationship between substance use and non-use in RC1's life. During the following twenty-two minutes or so, the group discussion continues, focusing on the participants' thoughts and personal experiences related to the integration of working life and substance abuse, their substance abuse habits, their family histories, the points of acknowledging their problem, and so on. After this lengthy conversation on sharing experiential knowledge and articulating experiences, the participants reflect on the use of the protocol in collaboration with each other.

Learning to Use the Protocol as an Epistemic Tool
The final section presents two data Examples (4-5) in which the participants in interaction reflect on the use of the protocol in relation to the activities of talking and sharing. As the participants point out, sharing and talking about experiences of substance abuse and, therefore, recognizing their role as epistemic agents in the rehabilitation process, requires a situated and joint detachment from protocol use in the rehabilitation interaction. At the beginning of Example 4, RC1, looking ahead and away from her papers, suggests how in the current conversation they are digressing from the topics (line 1).
Example 4. In line 2, PSW treats RC1's previous turn as somewhat of a complaint by responding that digressing from the topics is just fine. With their overlapping responses in lines 3-6, SOA and RC2 align with PSW's prior turn. SOA also provides a reason for why digressing is appropriate by pointing out that, if the participants would focus only on using the protocol, the time spent talking would be minimized (lines 6-10). SOA therefore implies that, when it comes to talking, the temporal detachment from using the protocol is preferable to totally sticking to its use. Subsequently, RC1 and SOA laugh aloud together (lines 11-12).
For the next half hour (data not shown), the participants discuss and share their experiences on various issues, including the hereditary nature of alcoholism and the prevalence of alcoholism in the family, the role of experience in understanding alcoholism and other drug addictions, the positive and negative aspects of community rehabilitation, and the preparation for an abstinent future. During that time, the discussion protocol is explicitly used once in the rehabilitation interaction. This happens when RC2 describes how, because of her "spiritual awakening", she has been able to change her former distorted way of thinking as a drug addict. Orienting to her papers, RC2 mentions in passing that she has written this thing down under the theme "advantages of substance non-use". In any case, the thirty-minute conversation mainly proceeds without the participants using the protocol as a resource in talk and, after that, the participants continue reflecting on the use of the protocol in rehabilitation interaction (see Example 5).
Example 5. Prior to the beginning of Example 5, RC2 reflected on whether or not the participants had sufficiently used the protocol. All participants agreed that their protocol use had been adequate. In line 1, SOA continues with the topic of the protocol by specifying its function in rehabilitation interaction. In SOA's words, the "papers" are used to get the conversation started (lines 1-3). Other participants respond with overlapping dialogue particles, "mm", produced in various sound lengths (lines 5-7). At the beginning of her turn in line 9, SOA uses the conjunction "but" to indicate contrast between her prior utterance and the one to follow. She states that "the idea" of group discussion is to talk about and share things of importance with others (lines 9-15) and, as her talk implies, the discussion protocol is used as a tool to serve that purpose (lines 1-15). In this way, she also accounts for the earlier discussion about the proper use of the protocol in interaction. For the last five minutes (data not shown), without any references to the protocol, the participants mainly orient to the rehabilitation clients' future by discussing the importance of fostering positive thoughts about oneself, others, and life in general.

Discussion
The community rehabilitation practice of sharing thoughts, feelings, and experiences of substance abuse through a discussion protocol aims at enhancing clients' self-knowledge and sense of personal agency. According to our epistemological interpretation, one's own self-confidence can be improved, first, as creating a space free from testimonial injustice, a space where all voices are given an equal opportunity to be heard and discussed. This happened in Examples 2 and 3 in which the participants disengaged from using the protocol to be more collaborative in sharing and discussing their experiences. Such spaces, if successful, can resist some of the effects of epistemic oppression-that is, exclusionary practices that make it more difficult for the oppressed to take part in knowledge production (Dotson 2014) and epistemic smothering (Dotson 2011). Epistemic smothering refers to being silent when one does not believe that one's contribution to the discussion receives a fair hearing; it is "the truncating of one's own testimony in order to insure that the testimony contains only content for which one's audience demonstrates testimonial competence" (Dotson 2011, p. 234). Another epistemological aspect of the communal rehabilitation practice is that it can help substance abusers to cultivate epistemic courage to honestly assess their situation and the effects of their substance abuse on themselves and others. This element was also found in the case discussed here.
In addition to raising self-knowledge and cultivating the epistemic courage of recovering addicts, the successful use of discussion protocols may alleviate hermeneutical injustice through the invention of new conceptual frameworks for understanding addiction. In Example 2, articulating the elements of temporality, of moments and durations as key to discussing the "advantages" and "disadvantages" of addiction, the discussants were able to create a new framework for understanding their situation. Even if concepts such as short-term satisfaction and long-term effects are widely used in discussing addiction, there is definite epistemic value in creating concepts from the ground up rather than mechanically applying received abstract concepts of experience. Such organic concept creation can be both useful for ensuring the uptake of such concepts by peers with similar experiences and for cultivating the epistemic courage of discussants.
The element of temporality played another crucial role in the discussion. By first using the protocol, then detaching and taking critical distance from it, then speaking freely, then creatively using the protocol on their own terms, while also reflecting on the use of the protocol on a metacognitive level, the discussants were able to create a process for taking control of the protocol rather than being controlled by it. In future research, it would be fruitful to (1) analyze other cases of taking control of discussion protocols from the perspective of alleviating epistemic injustice and (2) assess what makes such processes successful. In this case, the key to success was a combination of the skills and epistemic virtues of the social adviser, the epistemic courage of some of the discussants to challenge the protocol, the simplicity of the protocol itself, the favorable background conditions, and probably a good measure of luck. While epistemic virtues cannot be detached from individuals, there may be good encouragement strategies that can be used in various situations. However, in situations where power differences exist between facilitators of discussion and discussants, such strategies are always risky in that they may end up perpetuating the epistemic injustice they were supposed to alleviate.
Individual self-improvement and new conceptual frameworks and processes in local situations are, of course, not enough, as epistemic injustice is a wider social and structural phenomenon. Moving toward an epistemically just treatment of addiction requires that the viewpoints expressed, as well as the conceptual innovations created in community rehabilitation practices, are taken seriously by professionals, academics, politicians, the criminal justice system, and society at large.

Conclusions
In this article, we examined the social production of the epistemic agency of female substance abusers in community rehabilitation interaction. Our starting point was the observation that substance-abusing women are vulnerable to specific kinds of epistemic injustice, such as stigmatization and discrimination. These women are frequently deemed epistemologically incredible and constructed as willfully wayward women who are morally corrupt and deviant in socially unacceptable ways. These persistent prejudices make it difficult for female substance abusers to recover from their illness. Following Fricker's (2007) social epistemology and one of its principal ideas that all human beings can be recognized as epistemic agents, we formulated our leading question: How can community rehabilitation practices alleviate epistemic injustice and strengthen the epistemic agency of substance abusers?
Our research data were collected in a Finnish rehabilitation center by videotaping six group discussions between social work professionals, peer support workers, and rehabilitation clients with substance abuse problems. Female substance abusers participated in only one recorded group discussion. Using CA as the research method, we focused on the role of the discussion protocol as both a hindrance and facilitator of the conversation. The element of temporality appeared to have a crucial role in the discussion. By first using the protocol, then detaching and taking critical distance from it, then speaking freely, then creatively using the protocol on their own terms, while also reflecting on the use of the protocol on a metacognitive level, the discussants were able to create a process for taking control of the protocol rather than being controlled by it. This resonates with how the discussants were able to share their feelings and deal with their own addiction.
Theoretically, our analysis relied on feminist epistemologies and Fricker's (2007) two epistemic concepts of injustice-testimonial and hermeneutical injustice. By testimonial injustice, we mean how substance abusers suffer a credibility deficit due to harmful social stereotypes. Their testimony is not believed, taken seriously, or considered interesting or relevant. By hermeneutical injustice, we mean that substance abusers are unable to understand their own situation due to their inadequate social resources, or the reality that language does not provide concepts that could describe feelings of injustice.
Our results suggest that the use of the discussion protocol in community rehabilitation has the potential to alleviate epistemic injustice for female substance abusers and increase their confidence as epistemic actors in society. The discussion protocol was understood as an epistemic tool that professionals and clients can learn to use in ethically and epistemologically successful ways in interaction. Through the collaborative activities of sharing experiential knowledge about substance abuse and discussing the experiences of abuse in the rehabilitation interaction, the substance abusers were able to develop novel ways to strengthen their epistemic agency by enhanced self-knowledge.

Data Availability Statement:
The research data used in this study is not available to other researchers due to ethics and confidentiality issues.

Conflicts of Interest:
The authors declare no conflict of interest. The funder had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.