Vernacular and discursive harm reduction: an ethnography of an online drug community

Abstract

Background

In the context of online drug communities, the ethos of harm reduction comprises a set of ethical and practical principles that enable drug enthusiasts to shift stigma by normalizing certain behaviors to the detriment of others. The HR ethos is the dominant discourse in drug forums, which may give the impression that it is a natural expression of the online sociability of drug enthusiasts. However, according to the interactionist theory of deviance, there is a process behind the demarcation between deviance and normality; this would suggest that before becoming the dominant discourse, the HR ethos has had to assert itself in drug forums in the face of other attitudes. The study aimed to follow the progression of the HR ethos in an online drug community, in order to identify the process that led to both its generalization by the community’s members and their labelling of deviant behavior within the community.

Methods

Discord is a recently created social media platform that provides spaces for various online communities called ‘servers’. The connective ethnography method was used to select three Discord servers (Blue, Green, and Yellow) for drug enthusiasts. Participant observation was utilized to collect data because it allows us to understand the meaning of interactions between players. Data were collected over a 14-month period using ethnographic field notes. The present article analyzed a 3-month period of identity crisis and resolution on the Blue server using concepts from the sociology of deviance.

Results

The process began with the problematic situation: statutory members of the Blue server were exhausted by their own concern that members of the community were being hurt by their drug use. As a solution, members who had already participated in other drug communities, acting as moral entrepreneurs, proposed implementing the HR ethos on the Blue server. The statutory members derived rules from this ethos that created a vernacular HR specifically adapted to the Blue community. In particular, they established rules for discursive HR which enabled them to set boundaries for conversations without judging members’ practices. In conclusion, drug enthusiasts in the Blue community took care of each other, and developed a vernacular HR with the aim of ensuring that being part of their community was beneficial for its members.

Background

Drug enthusiasts communities

According to interactionist theory, behaviors are designated as deviant through a labelling process that separates them from what is considered normal. In Western societies, where dominant values mandate that people be responsible and take care of themselves [1], people who use drugs are labelled as irresponsible and dangerous to themselves. This is reflected in the criminalization and pathologization of drug use [2]. To maintain their practices, people who use drugs must acquire ‘rationalizations’ that normalize their use, making it compatible with values of responsibility and rationality. These rationalizations are developed and maintained by interacting with peer groups [3].

Computer mediated communication (CMC) has become an important vector for discourses about drugs. In 2015, 63% of Europeans aged 15 to 24 used the internet as their primary source of information about drugs [4]. Starting in 2000, some of the discourses on websites providing information about drugs were characterized by a tolerant, even positive, attitude towards drugs [5]. This raised concern in medical and governmental institutions; for example, the European Union financed the mapping of drug information sites [6] including several drug forums.

Davey et al. were the first authors to explore the key features of drug-related internet forums and drug forum communities. They described them as hierarchical digital spaces managed by drug enthusiasts who exchange advice, experiences and practices [7]. Akrich called them communities of practice that ‘build[ing] common objects of interests and a shared understanding of the purposes of the group and of appropriate forms of participation’ [8]. In the present article, we use the terms ‘drug community’ to refer to this type of online community of practice, and ‘drug forum’ to refer to communities studied in the literature on drug forums. Moreover, we will refer to members of drug communities as ‘drug enthusiasts’ rather than ‘people who use drugs’, to emphasize that what brings them together is more an interest in drugs than drug use itself. Finally, the term ‘user’ refers to internet users.

In a context where drug use is stigmatized globally, drug communities enable drug enthusiasts to come together and to communicate without stigma [9]. Like other online health practice communities, drug communities enable “each user to reopen the space of possibilities, not to find him/herself locked on the paths traced by the medical and institutional world” [8]. Rosino transposed the interactionist theory of deviance to a drug forum devoted to the use of dimethyltryptamine (DMT) [10]. He showed how drug enthusiasts learned how to use DMT, and how to formulate rationalizations which aligned with the dominant values (see above). Rosino’s work highlights how drug communities constitute peer groups which help to normalize drug use by making it compatible with a normal life. In this context, one of the founders of the French-language forum Psychoactif explains how drug forums allow their members to think of their use as something other than deviant: “these communities widen the social space between the categories of ‘patients’ and ‘delinquents’ carved in the 1970 law penalizing drug use. They allow people to think differently and to experiment with other ways of living with drugs” [11].

Harm reduction as a moral code

In public health, the concept of harm reduction (HR) refers to an approach to drug policy which aims to reduce the harm caused by drugs, rather than reducing drug use itself [12]. HR considers that people who use drugs are willing and able to take care of themselves (which goes against their label as irresponsible, see above) [13]. However, the current institutionalized HR model, which is presented as pragmatic and apolitical by its supporters, is nonetheless linked to repressive policies [14]. In France, institutionalized HR is part of the Public Health Code [15]; it aims to combat the harm caused by drugs, without questioning the ongoing legal context in France where drug use is repressed and medicalized [16]. The framework of legal HR in France is strictly defined by the State, and in particular, covers the provision of clean equipment (e.g., syringes) and information on risks [17].

HR has a different meaning in the context of drug communities. In a previous article based on the same field research outlined in this article, I argued that HR in the drug community context refers to a set of rules and principles regulating use which are developed by drug enthusiasts themselves through community discussions [18]. In this context, HR is not framed by an institution. Community members compare their practices and come to an agreement on the best ways to use drugs. This “bottom-up” understanding of HR has been called “vernacular HR” [18, 19], a concept which distinguishes the HR on drug forums from that of “professional drugs workers, criminal justice officers or medical staff” [20]. HR also has a moral meaning; the ethos of HR refers to a “culture of caution, care, and responsibility” [21]. By exchanging techniques and advice, internet users strive to build an HR framework that reconciles the use of drugs with a moral and fulfilling life [21]. In this way, drug use aligns with the dominant values of responsibility and rationality (see above) [1]. Bancroft describes a “counter-public health”: a shift from the institutional approach of HR to an approach of “responsible harm”, where pleasure and risk coexist, and where the potential for damage becomes a challenge that the informed person who use drugs can manage [22]. In this way, drug communities appropriate the concept of HR, “allowing each user to reopen the space of possibilities, not to find him/herself locked on the paths traced by the medical and institutional world” [8].

Having said that, drug communities do not offer total freedom of expression. Internet users express strong opinions about what they perceive to be good or bad ways of presenting drug use [23]. HR therefore serves as a normative framework for social practices, and determines what members consider acceptable or normal; as Duxbury says, “[Internet users] use HR as a key criterion for what type of information should be contributed and how the information should be communicated” [24]. The members of an online community are generally organized according to a hierarchy of social and technical powers ranging from the ‘lurker’ (an internet user who is not a member of the community but who reads discussions) to the ‘administrator’ (who has full power to run the community) [25]. In the case of drug communities, power comes with adherence to the HR ethos. Community moderators do not just check the accuracy of the information, but direct the community’s identity towards the HR ethos of moderate, responsible and mature use [9]. Demonstrating responsibility by showing one’s awareness of risks helps members gain credibility within the community [19]. Members high up in the community hierarchy may feel it is their duty to embody the HR ethos in order to encourage other members to adopt it, even if this means presenting a watered-down version of their own use [18]. Conversely, members who prioritize pleasure over safety are labelled as deviants within drug communities themselves, for example with the term ‘trashbag’ [23]. The HR ethos is therefore a strategy of assimilative normalization [26]: the stigma of being a person who use drugs does not disappear within the drug community; it is merely shifted to normalize certain practices, while others are labelled by members as deviant. This strategy of assimilative normalization can also be seen in other concepts. For example, the ill-defined concept of the ‘psychonaut’ can be used to remove the stigma attached to dominant representations [27, 28].

To summarize, in the context of a drug community, ‘HR’ refers to a moral code, combining ethical and practical principles, which enables drug enthusiasts to think of themselves as normal members of society. While some of the elements of this HR ethos are also found in institutional HR, in particular the recognition of the agentivity of people who use drugs, it differs from the latter in that it has a profound moral significance and is defined in vernacular terms rooted in the experience of community members. Accordingly, this article does not study HR in terms of public policy, but in terms of self-regulation by members of a drug community. As such, the term ‘HR ethos’ here refers to this moral code of responsible use; moreover, the term ‘HR discourses’ refers to discourses produced by internet users according to the HR ethos.

Identifying the process behind the generalisation of the harm reduction ethos

Studies of drug forums present remarkably similar results regarding the nature of the information exchanged and the moral values of the collectives involved [11, 19, 21, 23, 24]. HR discourses on drug forums are hegemonic; other discourses, for example promoting prohibition, or putting pleasure before safety, are rarely present [23]. This hegemony may mean that drug forums have reached a form of ‘closure’. Hine states that closure is when “there is a general agreement on what the technology [i.e., drug forums] is and what is it for” [29]. The closure of a technology follows “a period of interpretative flexibility during their development, with the meaning, uses and problems associated to [sic] the technology differ between social groups” [29]. This suggests there was greater diversity in the rules on discourses in drug communities before the HR ethos became the dominant normalized discourse. In interactionist sociology, the relationship between rules and deviance is not static. Rules are produced sequentially, meaning that the process of separation between what is considered normal and deviant proceeds in phases. This process involves values, a problematic situation, and individuals who create and implement rules (i.e., ‘moral entrepreneurs’) [30]. In studies showing the normative dimension of HR, HR discourses can give the impression that they are only motivated by adherence to dominant societal norms [23]. However, to my knowledge, no study to date has investigated the process by which forum moderators choose to adopt the HR ethos rather than alternative discourses on drug use.

One possible way of identifying this process is to use an approach which examines the temporal evolution of a drug community up to the moment when it adopts the HR ethos. Brossard used such an approach when creating a history of French-language self-mutilation forums; he showed that there is a temporal dynamic between the discussion rules on each forum and the ‘market’ of forums which exist simultaneously, at different times. In particular, the personal trajectories of forum administrators as forum members in this market help determine the rules they implement on their own forums [31]. However, temporality is rarely taken into account in research on online communities in general [32]. Most studies of drug forums present snapshots [33,34,35]. A few incorporate a temporal dimension. Although Barratt conducted participant observation of drug forums over an 18-month period, the study was not designed to identify normative changes [9]. Bilgrei showed how the discourse on the use of synthetic cannabinoids on a drug forum evolved as ever more experiences were shared; however, his work focused more on the characteristics of the products (synthetic cannabinoids) than on community standards [36]. Móró and Rácz created a monograph based on 10 years of observation of the Hungarian drug forum Daath; they highlighted distinct moments in time where the organization and rules of the forum changed and underlined that member discussions led to these changes. However, they did not go into detail about these discussions [37]. Finally, the I-TREND report [38] lists and describes the drugs forums frequented by drug enthusiasts in five European countries. Nevertheless, the temporal dimension of the report focused on trends in forum use, in order to monitor users’ interest in various substances [38]. To my knowledge, no study to date has examined the temporal evolution leading to the adoption of the HR ethos as a dominant norm in a drug community.

Study aim

As mentioned above, in the context of online drug communities, the HR ethos comprises a set of ethical and practical principles that enable drug enthusiasts to shift stigma by normalizing certain behaviors to the detriment of others. It is the dominant discourse on drug forums; this may give the impression that it is a natural expression of the online sociability of drug enthusiasts. However, the interactionist theory implies that there is a process behind the demarcation between deviance and normality, and therefore that the HR ethos has had to impose itself over other discourses. Studying this process could shed light on the concrete problems encountered by members of drug communities, and the responses they come up with. In this context, I aimed to follow the progression of the HR ethos in a drug community, in order to identify the process that led to both the generalization of this ethos by the community’s members and their labelling of deviant behaviors.

Methodology

To understand how the HR ethos became the hegemonic discourse in a drug community, I adopted an interactionist approach. This choice was based on the belief that focusing on the interactions between users would provide a greater understanding of how they came to agree on their value system and its meanings.

Earlier, I mentioned that the hegemony of the HR ethos on drug forums may reflect a form of stability (closure) of the latter in terms of their meaning, uses and associated problems, and that this stability would most probably follow a period of interpretative flexibility where various discourses were considered before the HR ethos was finally adopted. This suggests that alternative discourses are more likely to be observed in the early years of a drug community’s development, when it is not yet stabilized.

Accordingly, the choice of methodology was either to analyze the archives of an already stabilized community, or to study a young community through participant observation. The first approach would have provided easy access to a large quantity of material. However, it would have decontextualized the community’s discourses, leading to a poorer understanding of the meaning given to them by participants [39]. In contrast, participant observation comprehensively captures the meaning of interactions between participants [40]. Moreover, I was already familiar with drug forums; this familiarity was both an advantage for finding research sites and understanding how they work, and a disadvantage for gaining a fresh perspective on my research field [41]. This was another reason as to why participant observation was essential: by communicating with the persons involved, study hypotheses could be objectively tested and refined.

To use participant observation, I had to find a community that had recently been created, so that it could be observed in real time. However, in 2020, the year this research began, no recently created drug forum existed. In contrast, discussion groups on drugs were increasingly appearing on proprietary social media, one being the platform Discord. I therefore concentrated my research on drug enthusiasts discussion groups on Discord.

Discord was created in 2015. The platform’s users can chat via direct messaging or by participating in discussion groups called ‘servers’ (fig. 1). Any user can set up a server, which they then become the administrator of. Discord’s servers are organized in a similar way to forums: each one is designed as an independent community, with a theme, rules, a division of space into theme-based chatrooms, and a user hierarchy. The latter is reflected in socio-technical permissions (access to reserved rooms, power to exclude users lower down in the hierarchy, etc.) [59].

Fig. 1

Pseudonymized screenshot of the Discord server called ‘Blue’

Despite this similarity, there are differences between forums and Discord servers in terms of technical infrastructure, and these differences impact user practices. More specifically, Discord is a synchronous (i.e., real-time) medium; that is to say it is based on instant messaging and video-conferencing. This encourages users to be present online at the same time and to send a large number of short messages. Conversely, forums are primarily text-based and asynchronous (i.e., not real-time) media: the pace of participation is slower, messages are longer, and a conversation can span several years [42]. Forums can have their own technical infrastructure and can be self-governed, while Discord servers use the Discord company’s software, and are the property of Discord. The company has full power over the content of its servers. Consequently, although there are a multitude of servers, they have a certain uniformity of organization and must comply with the Discord company’s general terms of use [59].

On forums, the level of privacy varies depending on whether registration is required (or not) to access all or part of the content [42]. In contrast, Discord servers are fundamentally private spaces, because you can only ‘join’ them through ‘invitation’. Having said that, depending on how easy it is to acquire an invitation, servers can effectively be semi-public spaces [43]. Finally, forums are independent of each other: for example, a notification received on one forum does not appear when the user consults another forum. Conversely, Discord users have access to all the servers they are members of through a single interface: this organization makes it easier to visit several different servers simultaneously.

The notion of “virtual community” is often used to designate online discussion groups. This notion was used in my literature review. However, it is not a given that each discussion group is a community in the sociological sense; that is to say, it is not a given that each discussion group generates a bond of belonging to a community that would reflect this group. Internet users rarely confine themselves to one discussion group, and instead circulate between several media [44]. This is important when delimiting the research field, because it is common to delimit this field to a community. Instead, for the present study, I adopted the connective ethnography approach recommended by Hine to delimit the field [45]. More specifically, to delimit the research field, I followed the movement of users from one site to another, rather than their belonging to one site or another. Connective ethnography has the advantage that it delimits a research field which resembles the way users really use the internet [45].

One a Facebook group page, I spotted messages by drug enthusiasts and followed a hyperlink which led me to a Discord server, which I shall call ‘Green’, and an account creation page. After creating an account, the hyperlinks posted on the server led me to various drug information resources including other Discord servers, websites such as Erowid.org and Psychonautwiki.org, French-language based drug forums (in particular Psychonaut.fr and Psychoactif.org), scientific studies, videos on YouTube, etc.

I then separated the research field into two elements, what I call the ‘core field’ and the ‘peripheral field’. More specifically, the core field of discussion groups was defined using the following selection criteria: (1) discussed all drugs (rather than just one type of drug); (2) hosted on Discord; (3) managed collectively (rather than by a medico-social player such as an association); (4) had members in common. Three servers remained after this selection process, which we call here Green (which I joined in October 2020), Blue (which I joined in November 2020) and Yellow (which I joined in April 2021). I conducted participant observation on these three servers (detailed below). As each server had a different population, different governance and different rules, I considered that each one was a community of practice in its own right, and that the type of research being conducted was a multi-sited ethnography [46].

The peripheral field covered the drug-related resources mentioned by the users or which they provided links to. I did not study it per se, but took it into account to contextualize the study’s results.

In the present article, I focus on the Blue server from November 2020 to March 2021. I mention the Green server only to contextualize the results. The Yellow server is not mentioned in the results, as I had not yet joined it during the study period. Other resources, such as the Psychonaut.fr and Psychoactif.org forums [38] are mentioned as elements of the peripheral field.

Observation, participation et position on the research field

This multi-sited ethnography was conducted openly with the permission of the founder and administrator of the Blue server. I spent approximately 10 h a week for 14 months on the study, from November 2020 to January 2022. I followed the daily conversations as they happened and copied interesting conversation sections which (i) illustrated how the Blue community was organized, (ii) discussed the community’s norms, or (iii) because they constituted an event which might have influenced the community’s evolution.

Participation consisted of taking part in the discussions on the Blue server. For ethical reasons, I tried not to have any impact on the type of information circulating in the community. For example, I avoided approving or disapproving of drug use. When a user asked me for information, I quoted HR resources already validated in the Blue community (e.g., the interaction table developed by TripSit.me) (Fig. 2). I also gave general advice such as spacing out drug intake and using drugs in a safe environment. Finally, when a user expressed distress, I expressed empathy and support. This prudent behavior was valued by the community, so much so that 3 months after my arrival on the server, I was awarded with Expert status, which is a specific status on the Blue server. This status is awarded to users who provide what is considered the ‘right’ information about drugs, and who share this information prudently. It was logical that as I adapted to the community’s standards, I would receive the status that reflected that adaptation. I also joined the moderation team 1 month after my arrival on the Blue server, to get a ‘behind the scenes’ insight into the community. Moderating the server involves reminding people of the rules and applying penalties (muting or banning a user). Reading users’ conversations as a moderator while reflecting on the rules of the Blue server, and negotiating with ‘offending’ users, was a very effective way of capturing the tensions around norms. I hardly ever applied any sanctions, because I did not want to have a punitive attitude towards the people I studied. As an expert and moderator, I gained access to chatrooms reserved for users higher in the server’s hierarchy. This proved very useful for the study. Eight months after starting fieldwork (i.e., October 2021), I resigned from my status of moderator, as data saturation had been reached and the moderating activity required a great deal of time and attention.

Fig. 2

The drug combination chart edited by TripSit

My 14-month immersion into the heart of the Blue server’s activities greatly affected me [47]: I felt enthusiasm, anger, fatigue, among other emotions. That is why I kept daily field notes describing my impressions, feelings and thoughts. These notes were used for field work, recounting how it felt to participate in a community of drug enthusiasts, and as tools for objectifying my subjectivity, with a view to ensuring that my interpretation of the results would be as objective as possible.

Interviews and conversations

I conducted semi-structured interviews with seven members of the Blue, Green and Yellow servers to reconstruct their trajectory as drug enthusiasts, their view of the different drug servers they were members of, and their perception of the evolution of these servers over time. I also conducted instant online discussions by private message on ad hoc topics with members of the three servers. The latter method was particularly well-suited to Discord servers, which are based almost exclusively on a sequence of real-time text messaging, whose threads are initiated, followed and ended abruptly, depending on the level of attention the users give to them. I labeled these discussions as ‘discussions with a member’.

Data analysis

All the ethnographic materials and data collected were analyzed as the fieldwork progressed. Daily notes included reflections about the meaning of the observations. As the fieldwork progressed, I noticed recurring events, contradictions and changes.

Once the fieldwork was finished in October 2021, I organized the field notes to draw up a chronology of events. I analyzed these events using concepts from the interactionist theory of deviance and following the requirements defined by Akrich for the study of forum dynamics as follows: (i) articulating the organization of exchanges, their content and the links between internet users; (ii) taking into account unequal distribution of participation; and (iii) considering temporality in the organization of the community [32]. For the purposes of the aim of the present article (i.e., understanding the adoption of the HR ethos by a drug community and the process behind it), the results described here focus on a period during my participant observation when a great deal of conflict emerged on the Blue server as to how drugs should be discussed on it.

Results

Server’s characteristics, HR rules of conversation, and structure of the Blue community

The first results presented here are the data used to describe the different communities studied, and the analysis of these data to establish the organizational structure of the Blue server. The results presented are valid for the entire study period (from November 2020 to January 2022).

Characteristics of the communities and the relationships between them

The Blue server is the main focus of this article. It was created on 28 April 2020. The number of members grew exponentially until March 2021. This was primarily due to the fact that the server was linked to the account of a content creator on another social media platform. This account publishes original educational and entertainment content on drugs. In that period, with each publication, its popularity increased and consequently, the Blue server saw a large increase of new members (Fig. 3).

Fig. 3

Evolution of Blue server member count over time

Most members of the Blue server are under 25 years of age. Although the community is closed to under-18s, moderators are certain than a proportion of users are minors who conceal their true age to take part in discussions. Most Blue users identify themselves as masculine. Most work, are apprentices or are university students. Some live in socially or mentally precarious situations (living on the street, in a psychiatric hospital, etc.). All psychoactive substances can be discussed on Blue. People with different consumption habits meet and chat. Some members are not people who use drugs and participate out of intellectual curiosity. Most exchanges concern first-time consumption: for example, explanations of the psychoactive effects of a product, or advice on taking a drug for the first time.

The communities presented in the following paragraphs are not the subject of this article, but they must be taken into account to contextualize the results for the Blue server.

The Green server has not been active since 2022. It was a small community of approximately 300 members, with few new arrivals during the study period. This dimension of ‘intimacy’ fostered strong bonds between users. Active members were between 20 and 25 years old. Most used drugs before becoming part of the Green community.

I also had access to a few small ‘private’ servers, created around friendship rather than on drug use themes. One example is the X server which comprised approximately 25 members of the Green and Blue servers.

The French-speaking forums Psychonaut.fr and Psychoactif.org are independent web forums, which have existed since 2001 and 2006, respectively [38]. Their membership and the issues they tackle have evolved over time as a result of different internal events and to developments in the drugs market, internet geopolitics and drug policies. Therefore, these collectives have long and complex histories; the forums’ archives are accessible on the web.

Communities as part of a drug enthusiasts network

Knowledge about drugs circulates in the discussions on drug communities. Users post links to or paraphrase information from other websites, some of which have already been described in the academic literature, such as Erowid.org [5] and Bluelight.org [48]. In the example below, a Blue server’s member posts a link to a ‘trip report’ from the psychonaut.fr [49] forum in order to convince another user not to consume datura. This is an example of how the same ‘tale of caution’ [33] can circulate between different communities.

‘Hobbit: Hi, I’d like to know how to take datura please; it grows close to to where I live; do you have to take the leaves or the flowers and how much do I need for a trip?

Etan; Hi, I wouldn’t advise you to take datura unless you want to die or end up in a psychiatric hospital. It’s a very powerful and deadly delirogenic plant. Reading this TR [trip report] should calm you down. https://www.psychonaut.fr/threads/rdr-datura-la-folie-dans-les-bois.30558/’

(Conversation on the Blue server, 15 May 2021)

In this way, drug servers are connected to a drug enthusiasts network of documentary resources and communities [18]. The opinions of server’s members are influenced by the information available on this network: there is a level of interplay between different communities.

Some members of the Discord servers studied in the present work participated in several communities at the same time (hereafter, ‘ubiquitous members’), and transmitted information, norms and values. For example, BlueBlad, a member of the Green server, also participated in Blue and in the Psychoactif.org forum, and was aware of the Psychonaut.fr forum. He thought that meeting internet users in real life to consume together was dangerous, and used his personal interpretation of forum archives as an argument against such meetings.

Yeah, but you should know that it’s a downward spiral consuming together, especially with stimulants; we’ve already seen it many, many times on [forum name].

(Conversation on the Green server between three ubiquitous members—12 January 2021)

In this article, I use the term ‘to meet’ to indicate when one server was discussed in the public space of another. The Green and Blue servers ‘met’ on 2 November 2020. Specifically, Green ubiquitous members talked about Blue in Green’s public space, which led curious Green members to join Blue (and therefore become ubiquitous). This was also how I became aware of Blue. Moreover, in January 2021, active Blue’s members were invited to join Green through personal invitations.

Practical HR and discursive HR

The Blue server is regularly defined by its members as an ‘HR server’; this definition also appears in the rules and discussions on the server. This definition means that the notion of HR acts as a framework for defining what is normal in the Blue server’s community.

The notion of HR on Blue first and foremost concerns regulating consumption practices; in this article, I call this ‘practical HR’. For example, consuming less than 125 mg of MDMA and waiting 6 weeks between each dose is considered to ‘be HR’, while taking unknown doses of MDMA every weekend is considered ‘not HR’. To find out what HR is or is not, users refer to resources on the wider drug enthusiasts network (see above). For example, the websites TripSit.me and Mixtures.info are used to learn about the risks of interactions between drugs. By establishing risk hierarchies, these resources allow users to state ‘HR rules’ such as ‘do not mix two depressants’. It is important to point out that users perceive that one cannot be completely HR; one can reduce the risks ad infinitum but one cannot eliminate them.

Moreover, the notion of HR on Blue concerns regulating discussions in terms of distinguishing what is permitted to say on the server from what is not: in this article, I call this ‘discursive HR’. In this discursive case, ‘being HR’ means following rules of good conduct on the server; for example, it is forbidden to brag about one’s consumption. Some rules are seen as common sense and are not explicitly spelled out in the community’s rules. In general, the aim of discursive HR is to demonstrate a commitment to practical HR, even if users do not apply it in reality. It is therefore possible not to be HR in terms of actual consumption, but to be HR in terms of discourse. In the verbatim below, a moderator on the Blue server explains why he prefers to talk about his consumption on the X server (see above):

The cool thing about the X server is that there’s nothing to hide; it’s the only place where I can say ‘I was a pig and I took 500mg of 3M’.

Why not elsewhere [on the Blue server]?

Well, firstly because elsewhere I’m a moderator, and secondly because we do HR, that’s all.

(Conversation with a Blue server moderator, 18/05/2021)

The aim of discursive HR is to make people understand that following the rules of practical HR is something normal and is seen in a positive light. Accordingly, ‘doing HR’ consists of both transmitting the rules of ‘good’ consumption practices and normalizing these rules through server discussions. The Blue server is then as much a space for education as it is for information: experienced users transmit to less experienced users notions of what ‘doing drugs well’ could be.

Finally, French law plays a part in the HR rules of the Blue server. Given that the server’s theme regards a practice that is illegal in France (i.e., drug consumption), participation could expose its members to official repression. Consequently, the server’s regulations demonstrate compliance with French law. For example, in line with French law, the server’s rules prohibit users from encouraging drug use and dealing [50]. Moreover, HR is integrated into public health policy in France [15]. By framing the community’s activity within the notion of HR, the Blue server strives to ensure its own legality.

The Blue community’s core members and HR hierarchy

Despite the large number of members on the Blue server, most of the discussions are led by a community core of approximately 50 members. In the following paragraphs, we describe this core in greater detail and show that it is mainly structured by two criteria: participation in the community and adherence to the HR rules (both practical and discursive).

As with all online community, the Blue server is a hierarchical space: certain members have a status which gives them more social and technical rights. In this article, we mainly talk about users with the status of ‘manager’ (i.e., administrators and moderators) or ‘expert’.

Experts are members recognized as having mastered the rules of HR and being very knowledgeable about drugs. They have a say in the organization of the server because it is their job to moderate discussions about drugs according to HR criteria. New experts are appointed from among the active members, but an expert may withdraw from the community without his or her status being called into question. During the study period, there were approximately fifteen experts on Blue. Managers define the direction of the community and its rules, and enforce them by applying sanctions. They are particularly active in the community, reading all the messages and participating several times a day. Over the period from November 2020 to March 2021, Blue had between 5 and 9 managers. Managers are often experts but not necessarily, and they follow and apply experts’ recommendations. Hereafter, the umbrella term ‘statutory member’ covers both managers and experts.

Moreover, regular participants in the Blue server are rewarded with ‘active member’ status. This status highlights that members are differently treated depending on their involvement in the community; but for the sake of clarity, in this article, active members will not be included in the term ‘statutory member’. Active members’ frequent use of the server often leads to a bond between them and statutory members. All of the statutory members were active members before being promoted.

As mentioned above, having knowledge about drugs is not enough to achieve expert status. One also needs to know how to present its knowledge from a HR perspective. For example, a person who is knowledgeable about the active substances of many plants but who enthusiastically talks about experiments deemed dangerous (e.g., using datura) will not receive expert status. As we saw in the introduction, the HR ethos is hegemonic over other discourses in drug communities, such as pro-prohibitionist discourses and discourses privileging pleasure over safety. As a result, ubiquitous members, who have more experience of drug communities, have an advantage over other members, because they already know the rules of both practical and discursive HR. Throughout the period of my observations, and despite a large turnover of managers, a large proportion (between one and two thirds) of the latter were already ubiquitous members before joining the Blue’s server’s team. This was also true for myself: my previous experience of drug forums helped me to fit into the Green, Blue, and Yellow servers, as I already knew how to present my knowledge so that it would be appreciated.

The HR hierarchy is also expressed through the term ‘level’, a notion used by the members to designate to what extent a discussion globally conforms to HR rules. The level is especially evident when communities ‘meet’ for the first time (see above). In November 2020, when the Green and Blue servers met, members of the Green found the Blue discussions to be riddled with errors, poorly structured, and therefore dangerous for the server’s readers: they saw Blue as having a ‘very low level’. This voyeurism quickly turned into interventionism: members of Green joined and actively participated in Blue in order to correct statements which they believed to be erroneous. They established themselves in the Blue community, providing an educational approach for people using drugs drugs whom they considered to be little or poorly informed. The following verbatim show how ubiquitous Green members considered their ‘mission’ on Blue.

Be nice to the members of Blue

And be aware that most of our HR work is done right there

Because [Blue has] a younger audience, more impressionable, less informed

(Conversation in the Green server—15 January 2021)

Blue’s members also perceived the difference in the ‘level’ of both communities. Greens were seen as more mature and informed, and therefore as having a higher level. The most active Greens participating in Blue became experts and/or moderators. The opposite of this educational movement was observed when members of Blue were invited to join Green: it meant that they were ‘elected’, in the sense of being deemed worthy of joining a community that considered itself to be more mature when it came to drug use. This election was based on Green members’ esteem of certain Blue members, mostly statutory members, and on friendship, which prioritized the invitation of active members. This is how the community core was formed, and why most of Blue’s statutory members were ubiquitous.

To summarize, the community core comprises ubiquitous members, statutory members, and active members who developed close personal relationships with the statutory members. These categories tend to overlap, as it is common for a member to fall into more than one category.

From ‘buddy’ server to ‘harm reduction’ server

Above, I presented the results of the data analysis regarding the general characteristics of the Blue server. In the following sections, I present results for the period between November 2020 and March 2022, as it covers a time when Blue experienced a crisis of identity and its subsequent affirmation as an HR server. The events surrounding these changes are described chronologically.

Chronology of the tightening of rules on the Blue server

At the end of October 2020, the Blue server saw its membership grow exponentially. Field notes taken on my arrival on November 2, 2020, highlight a very active server. Several thousand messages were being exchanged every day in approximately twenty chatrooms on the server. Users wrote short messages, where they expressed their enthusiasm for drug use, asked for advice and information, replied to their peers and shared their difficulties. All this flurry of activity was painfully tempered by a small number of experts (see above) who wrote longer messages, corrected comments, and urged caution about drug use. The server’s rules stipulated that it was obligatory to be of legal age, to anonymize your profile, and to respect the themes of the chatrooms. Furthermore, it was forbidden to be rude or discriminatory, to brag about one’s consumption and to encourage others members to consume drugs.

Very quickly, ubiquitous Green members became experts on Blue. Shortly afterwards, experts on Blue began to worry about what they called “disinformation”: the dissemination of information which seemed to contradict the knowledge established in the wider ‘drug enthusiasts network’. For example, on 18 November 2020, a user named Amibo promoted mixtures already considered dangerous (e.g., DXM + MDMA; see Fig. 2). He then invited a young user to take drugs in real-life with him. Experts perceived these discussions as dangerous, as an uninformed user might have trusted Amibo and consequently be harmed by a dangerous drug mixture. Experts believed that simply contradicting what Amibo said was not enough to deal with the situation; a way had to be found to prohibit such interventions. On December 11 2020, in order to cope with the influx of new members and the consequent increase in the number of discussions on the server, the managers posted a recruitment advertisement for new moderators. My application to the latter was accepted just when a discussion about how to change the server’s organization was in full flow. For example, on 14 December 2020, one moderator suggested that people who spread ‘disinformation’ should be penalized. This shows how managers were listening to experts’ concerns.

Field notes reflect the experience of moderating the Blue server as exhausting and disorientating. Thousands of messages were posted every day, many of which I found incoherent. There were arguments, calls for help, trolling (i.e., provocation). At that time, the moderation processes was to act on a case by case basis, dialogue rather than sanctions, and consensus between team members. However, with the huge demographic growth of the server, this way of management became unsustainable: decision-making was too slow, and the regulations were not precise enough.

Moreover, we’ll soon reach 1000 members, and there’s a risk it’s going to be a mess if we don’t take clear measures

(Conversation in the managers’ chatroom—29 December 2020)

Subsequently, throughout the month of December 2020, the server’s rules were amended incrementally. Specifically, on 3 December 2020, the rules were modified to introduce a ban on dealing (which until then was a ‘self-evident’ rule). Later in December 2020, managers modified the technical structure of the server, with the aim of making discussions about drugs more serious. For example, they implemented a separation between ‘community channels’ dedicated to discussions not related to drug use, and ‘drug channels’ where the rules would be strictly applied. On the drug channels, they introduced a ‘cooldown’ period where users had to wait before being able to write a follow-up message, in a bid to slow down the pace of discussions. December 2020 also saw the implementation of the monthly deletion of all public messages in order to protect the privacy of the community and its members. On 2 January 2021, giving precise dosages for consumption was prohibited, as statutory members feared a situation where a user would advise another to use too high a dosage, the possible consequence being a serious accident for which the community could be held responsible.

On 17 January 2021, the experts requested and obtained the exclusion of Amibo, on the grounds that he was going to “end up killing someone”. This was the first time a Blue member was banned from the server because of his/her negative influence on the community. Shortly afterwards, the experts opened a discussion about their own influence on users. One expert shared a topic from the psychonaut.fr forum, in which forum users questioned the influence of drug communities on their consumption [51]. From the quote below, one can see that that Blue’s experts were not just thinking about active members, but also about lurkers.

You can’t tell someone not to take 300 mg of MDMA and then say another night, in public, that you once did 500 mg and that it was so cool. Let’s not forget our etiquette; it is a source of influence, particularly for the silent majority which is, I think, made up of at least 30% minors.

(Conversation in the experts’ chatroom—25 January 2021)

From January 2021 onwards, the communities began to mix to the point where it became difficult to remember who came from which community. It was really at this point that a community core of statutory, ubiquitous and active members came together. Some of them started to meet up, leading to BlueBad’s concerns about meetings ‘in real life’. On 31 January 2021, an argument broke out on the Blue server between two ubiquitous members about the possession of drugs. Two experts—ubiquitous members who came from Green—immediately advised Blue’s administrators to take restrictive measures. The next day, a new rule appeared:

As the rules now state, server members are prohibited from meeting each other. The staff takes no responsibility for problems generated following a breach of the regulations.

For members who know each other in real life and are on the server, it is FORBIDDEN to talk about your real-life meetings here. [...]

In the coming days the regulations will be updated and moderation will become stricter. We can no longer tolerate this type of behavior which could take on terrible proportions. Be aware of the weight of what you write, over 1200 people see it.

BE RESPONSIBLE, YOUR WORDS HAVE CONSEQUENCES, AND WE WILL SANCTION MISBEHAVIOR

(Extract from the regulations—1 February, 2021)

In the second half of February 2021, two ubiquitous members encountered problems linked to their consumption and expressed suicidal desires and depression. During the following week, the experts’ chatroom was the scene of heated debates. Several experts, affected by the very tense climate, indicated that they wanted take a break from the server. Other experts called for a structural response. All the experts were worried or angry, and all agreed on the need to urgently change the server’s organization. To put an end to this, on 28 February 2021, all statutory members had a vocal conversation with the intention of establishing new rules. On 1 March 2021, a new set of rules was published.

RULE #5

BE RESPONSIBLE

Any form of inappropriate or irresponsible behaviour is prohibited.

5.1: Any form of flooding, trolling, or repeated incoherent comments will be punished.

The same goes for disinformation and for pornographic, gore, and negationist/revisionist images/comments, as well as for any type of apology for violence and/or hatred.

5.2: Do not give information which you are unsure about. Any approximate or false information about a product, its use or its properties could have disastrous consequences.

5.3: Differentiate between what is information, verifiable fact, and what relates to belief or spirituality. Blue is an HR server, nothing else.

5.4: All forms of self-destructive comments and behavior are prohibited. The server does not provide psychological support; we are neither doctors nor psychologists. Claiming to be so would be the illegal practice of medicine which is formally prohibited by law. Dubious discussions around death, suicide or self-harm are prohibited, because it is an extremely sensitive subject and everyone reacts to it differently.

(Extract from the regulations—1 March 2021)

On 17 March 2021, statutory members noticed that the atmosphere on the server had improved and described it as “pleasant”. The rules of the Blue server changed little after these events, at least until the end of the study period (i.e., until January 2022). Despite a significant turnover of statutory members during this time, I never saw them question these decisions. In contrast, some non-statutory members criticized the server’s rules for their length and their rigorous application.

In the previous paragraphs, we outlined the timeline of changes that occurred on the Blue server during the period between Novembre 2020 and Mars 2021. In the following sections, we shall describe results regarding cross-sectional themes which emerged during this period, because they help to shed light on the meaning of the changes which occurred.

HR-related negative feelings

Members of drug communities enjoyed discussing drugs, a topic they were passionate about. They liked to be recognized for their knowledge, and to feel that they were useful. However, their commitment to HR was also motivated by concern. They felt compelled to protect and take care of their peers if the latter seemed to be in danger. It was this mix of pleasure and duty that motivated members of the Green server to get involved on Blue (see above).

For Blue’s members, ‘doing HR’ meant reading discussions with a critical eye, answering questions from other users and correcting statements considered to be false. In the context of instant messaging, where up to 5000 messages were exchanged every day, doing HR could take up a lot of time and energy. Furthermore, users were frustrated when their conversation partner did not listen to them, and when their efforts seemed to be wasted. Users also had negative emotions when they felt like they had not accomplished their ‘task’. For example, they might blame themselves for not knowing how to help, or be frustrated at not being able to convince someone to give up a dangerous practice. The demographic growth of Blue between November 2020 and March 2022 multiplied these frustrations.

These negative feelings were exacerbated by a sense of responsibility. Then, statutory members in particular experienced negative feelings, because they felt responsible for the other members of the collective. Furthermore, although the server was prohibited to minors, the language used and ideas conveyed by some users strongly suggested they were in fact minors; this added to statutory members’ sense of responsibility and anxiety. This created ambivalent feelings and frequently led to the impression of having failed.

Not long ago I talked about 3-MMC with someone; I learned afterwards that he is 16 years old. I wanted to kill myself

(Conversation in the Green server talking about the Blue server—29 December 2020)

Finally, the bonds formed by users made this sense of concern and responsibility even stronger, making them more vulnerable to negative emotions when the HR they proposed did not work. In particular, active members acquired an overview of each other’s consumption and were able to identify negative dynamics in the trajectories of other members. When they failed to help each other, they experienced a great deal of fear, anger and frustration. The growing porosity between the Blue and Green servers strengthened the bonds between members of the community core. This explains the distress felt by statutory members when the two ubiquitous members mentioned above encountered problems with their consumption and expressed suicidal desires and depression.

How is the role of moderator of the HR theme different from being a moderator of other themes?

Well, we’re already tackling sensitive subjects that border on legality. Then you have to monitor people in different ways; it’s moderation over the long term with several factors to take into account. Long-term means understanding, mediation, and knowing how to put things into perspective... If you are empathetic, it’s difficult to keep your distance from the suffering and distress of certain cases. We must never ignore the fact that the HR servers are refuges for a lot of people, a place to fall back on...

(Interview no. 3–21 November 2021)

Statutory members tried hard to repress their emotions in order to maintain a caring atmosphere. However, they did not always succeed and were often unpleasant, sometimes even exploding with rage. This failure to maintain a positive atmosphere added to their frustration.

Consequently, during the 2020–2021 winter, statutory members found themselves deeply dissatisfied with the Blue community and started thinking about how its organization could be changed for the better. During the meeting of statutory members on 28 February 2021, it was agreed that in order to preserve their mental health, they had to learn how to limit their concern for other members, the reasoning being that it would also limit their negative emotions of ‘failure’. The following excerpt from my field notes shows the type of decisions statutory members made to limit their attempts to help server members.

If it turns out that the [user’s] approach [to drug use] is self-destructive (“I use because life is shit,…”), [we the statutory members] can advise against it. If the person persists, we understand that we cannot stop them; we warn them that we do not accept these comments on the server, and we stop there.

(Note taken during a meeting of statutory members—28 February 2021)

Risk publicization

Some users regularly described ongoing or planned risk-taking (‘risk publicization’ hereafter). What distinguished this from other participations was the deliberate nature of the risk-taking and its expression. For example, a user might say every day that he/she planned to consume drugs considered dangerous, such as smoking datura or injecting heroin, despite knowing that other members would worry and would try to dissuade him/her. The two ubiquitous members who expressed their depression in February 20, 2021 (see above) also engaged in risk publicization by manifesting their desire to use drugs to “destroy themselves”.

Risk publicization stressed the community. When risk-taking was deliberate, HR advice was not listened to and other members began to worry. They were also concerned about the context of risk-taking, particularly when the member publicizing risk appeared to be in poor mental health and/or seemed to be young. These discussions could last several hours or days and could provoke very intense emotions. Statutory members were skeptical about risk publicization, especially when it was repeated. They saw it more as a desire for attention than a request for information and guidance. They attributed this attention seeking to adolescent narcissism. They believed that publicizing risk in an educational space could influence certain readers by altering their perception of what ‘good’ drug use entails. They also wondered whether risk publicization could escalate risk-taking through imitation by some users in order to get the attention of others. Accordingly, managers finally responded to risk publicization on the Blue server with rule 5.4: “All forms of self-destructive comments and behavior are prohibited.” (see above).

Everyone does what they want at home, but in a public space it’s not the same! People watch a guy boast about his life, where he fucks himself up, [a guy] who doesn’t give a fuck about HR, who isn’t sanctioned. That’s going to have bad repercussions on some people’s view of things.

(Conversation in the Blue server expert chatroom—27 February 2021)

One expert, FluxY, thought that risk publicization contributed to a vicious circle where people who use drugs only have negative representations at hand to explain their use, but also sustain these negative representations by aligning with them. Thus, according to FluxY, risk publicization contributed to the stigmatization of people who use drugs throughout society, and helped to justify their repression.

If we want things to evolve, we have to make efforts. And that also starts here. I want the decriminalization of certain products, that we stop stigmatizing. But for that, you have to invest and do what is necessary so that people have the best possible image of the issue.

(Conversation in the experts’ lounge—27 February 2021)

Regulation of discussions

Between 20 and 28 February 2021, the statutory members debated the organization of the community. Contrasting notions regarding the role of a drug enthusiasts community in the consumption patterns of its members emerged. For some, the community must impose standards on consumption, in order to prevent users from adopting at-risk behaviors. For others, imposing standards on consumption was neither legitimate nor realistic; one had to accept that the community was fundamentally impotent in such matters.

During the meeting on 28 February 2021, the statutory members decided that the server could not control risk-taking, but it could control risk publicization. For example, the managers could not sanction a member because he/she regularly consumed dangerous mixtures, but they could prohibit him/her from writing detailed descriptions of them. It was at this point in the evolution of the Blue server that I began to observe and analyze the distinction between ‘practical HR’ and ‘discursive HR’ (see above). Practical HR rules could be communicated but not imposed: the responsibility for applying them lay with the individual user; this decision meant that the user’s free will would be respected and at the same time, the statutory members would not have to worry about this person’s consumption. In contrast, the rules concerning discursive HR were written into the server’s overall set of rules. The entire community was therefore responsible for their application: experts had to share them, users had to implement them, and managers had to punish any deviation. In this way, all Blue’s members were responsible for protecting the community from the effects of deliberate risk-taking.

The concept of risk publicization goes hand in hand with the identity of the Blue server as a public space. When the server had fewer than 500 members, it was thought of as a private space, a friendly, “buddy” server, where users expressed themselves with little hierarchy in a relaxed atmosphere. However, the exponential demographic growth of the server led to the belief that active members would address a largely silent audience (i.e., lurkers), in a mostly vertical dynamic. This led to a new identity for the Blue server: specifically, that of a server following the HR ethos, with a mission to provide drug information and education on safe drug use. The following extract from a conversation highlights the hesitation between these two identities.

Ean: Do you want a kind of a “buddy” thing, or to be really into the reflective side of HR?

Blue: The thing is, I wish I could have both in the same time, but I understand that it’s complicated. It was going well until now, but there are too many of us

(Conversation in the Blue managers’ chatroom. Blue is the founder of the server—04/01/2021)

This hesitation between two identities may explain the intensity of the existential crisis which shook the Blue server at winter 2020–2021. While most of the statutory members considered the server to be a public space requiring a certain level of restraint, the bulk of the community still considered it to be a private space where one could express oneself with no filters.

Discussion

To analyze the HR practiced on Blue, we shall first return to the notion of ‘vernacular HR’ [19]. This brings us back to Illich, who suggested that ‘vernacular’ designates ‘unpaid activities which provide an improved livelihood, but which are totally refractory to any analysis utilizing concepts developed in formal economics’ [52]. Despite the resources of the drug enthusiasts network, vernacular HR is not standardized: it is ‘shaped anew by each small community’ [52]. It takes the form of ‘practical wisdoms’ that are sensitive to situations and rooted in particular socio-technical spaces [20]: in this case, a Discord server on the theme of drugs, frequented by a large number of young users. This is why the vernacular HR of the Blue server is not exactly transposable to other drug communities, and cannot be confused with the medico-social HR provided for by law [15]. Accordingly, one of the limitations of this study is that it do not describe a generic online type of HR. Instead, it analyze the process by which a drug community adopts the HR ethos; in other words, how a vernacular HR is constructed.

In the following paragraphs, this process shall be analyzed using the concepts proposed by Becker to study the imposition of norms [53]. First of all, norms originate from values. In the case of Blue, we can identify two such values: (1) ideally, the community should positively influence its members, or at the very least not negatively influence it. (2) ideally, the community should be inclusive, welcoming all kinds of people without judgement. The statutory members created a discursive HR norm from these two values: specifically, this norm stipulated that discussions had to reflect the rules of practical HR (in order to improve users’ health) but it did not impose practical HR (in order to ensure that users with at-risk practices would still feel welcomed in the community).

Nevertheless, for a norm to be deduced from values, a problematic situation had to prompt someone to make this deduction [53]. This problematic situation was the crisis situation during the winter of 2020–2021 described above, where statutory members were morally exhausted, as a result of the conflict between the relaxed atmosphere of the community and the seriousness of the theme of drug use. The statutory members feared that, because of the lack of a clear, rule-bound discursive framework, other members might underestimate the risks of drugs and injure themselves. The issue of discursive HR was raised in meetings organized to deal with this conflict: how can we reduce the risk of the community harming its own members? How can we talk properly about drugs?

Two factors made this conflict untenable: the exponential increase in the number of Blue’s members, and the bonds of friendship between members of the community’s core (approximately 50 members). Doing HR activity is emotional labor [54]: you must exercise concern while maintaining your distance, be attentive without getting involved, give information and advice without presuming that you will be heard. Poor working conditions hinder emotional labor and make workers vulnerable to emotions which need to be controlled. The events of the 2020–2021 winter correspond to a deterioration in the working conditions of the statutory members of the Blue server; they had to manage an ever greater number of problems, and some of these problems affected people with whom they had created a bond.

Although the experts drew attention to the risk of accidents, and the managers highlighted the shortcomings of the then current regulations, it was the ubiquitous members in particular who acted as moral entrepreneurs [53]. In November 2020, members of the Green server joined Blue to spread the HR ethos. This moment when some Green members became ubiquitous, coincided with the start of Blue statutory members’ concerns about the organization of the server. Ubiquitous members had a long-term vision concerning discussions on drugs. Some were able to mobilize archives from older forums. Statutory members were receptive to their experience. Some measures, such as banning ‘real life’ meetings, were directly requested by ubiquitous members. In this way, the interaction between the two drug communities played a major role in the creation of a vernacular HR on the Blue server. This shows that the HR ethos was transmitted from one community to another. One of the key takeaways here, is that we do not know how the statutory members of Blue would have identified the problematic situation with their server, or how they would have responded to it, had ubiquitous members from Green not joined them.

It is important to ask what the personal interest of these moral entrepreneurs was [30]. Barratt suggests that such an interest reflects the drug’s desire to establish him/herself as a rational agent and to foster his/her integration into neo-liberal societies [23]. This would involve rejecting the speeches of drug enthusiasts whose self-presentation does not meet fundamental imperatives of “being concerned for one’s own health, one’s own equilibrium, and behaving reasonably” [40]. Different elements in the study confirm this analysis. For example, risk publicization was prohibited not only because it represented a danger to other members, but also because it participated—in the opinion of some experts—in continued social stigmatization of people who use drugs by feeding negative representations. There was also a link between the creation of the Blue server as a public space and the adoption of HR rules. As Blue’s membership grew, it became increasingly public, and therefore increasingly visible; this led to greater accountability in terms of the dominant values of general society. The managers feared that the Discord company would delete the server, that they would be prosecuted, and that other digital players would stigmatize their community. It must also be pointed out that HR as a normalization strategy shifts the stigma towards uses of drugs deem which are ‘irresponsible’ [9]. Accordingly, if the rules of practical HR make it possible to ‘do drugs well’, then not following these rules means ‘doing drugs badly’. This shifting of stigma on others can be seen in the comments of some experts, who wanted to exclude users who did not respect practical HR. Moreover, the hierarchy between communities based on the notion of ‘level’ (see above) certainly helped to normalize the practices of users who frequented ‘high level’ communities—to the detriment of those who frequent ‘low-level’ communities (at least in the eyes of the former). The labeling of Blue as ‘low-level’ helped some Green members to justify their moral undertaking [30]. Becker and Gusfield noted that this approach is typical of the reformism of the dominant classes towards the dominated classes [30]. It is therefore interesting to note that the notion of “responsible use” is characteristic of middle-class people who use drugs, who are integrated into neo-liberal societies [26].

However, this analysis has the drawback of distancing us from the meaning that the actors themselves give to their actions. It can give the impression that the HR ethos is an arbitrary norm supported by fanatical reformers. To really understand the situation studied, one needs to look at the facts as expressed by the actors. The problematic situation Blue faced was that the statutory members were morally exhausted by their concern for other community members. This concern regarded concrete situations involving high-risk activities, where people could be injured. Each new rule introduced on the server responded to a specific problem: for example, the ban on indicating what dosages of a drug to take was a response to a problem where users were indicating high dosages, something which could lead to a risk of overdose. HR rules are “subsistence activities” [55]: drug enthusiasts directly improve each other’s’ conditions by using tools developed in their own communities. Certainly, the vernacular HR of the Blue server owes a great deal to ubiquitous members, who imported a framework of thought from other drug communities. Moreover, by adopting the HR ethos, managers were able to align with French law, thereby ensuring the server’s legality. However, the rules adopted were first and foremost developed to meet the specific needs of the server itself.

Moreover, the emotional ties between members were decisive in the shift towards adopting an HR identity. Members’ care for each other is at the root of their HR activities: they take pleasure in making themselves useful, and suffer when they fail to do so. The more they become attached to each other, the more they care for each other, and the more unbearable failure becomes. Professionals often present HR policy as an amoral, dispassionate activity [56]. Our findings suggest the opposite: vernacular HR appears above all to reflect the collective care between members. Certainly, on the Blue server, the implementation of discursive HR may seem the opposite of a care-based approach, since it was supposed to reduce the concern of statutory members. However, its aim was also to reduce the violence that arose from the frustration of statutory members, and to ensure that users who did not follow the practical HR rules were not excluded. Accordingly, despite their restrictive nature, HR rules improved the conditions of Blue’s members. This was also observed by Van Schipstal et al. in a study that observed “peer HR from below” practices, both offline and online. When drug enthusiasts disseminate norms on “what can be done”, they help to make everyone’s experience less dangerous and more enjoyable; indirectly, they take care of the drug-using community [60].

Conclusion

This study shows the process by which an online drug community, specifically the Blue ‘server’ on the social media platform Discord, adopted rules on discussions inspired by the HR ethos. This process was the result of two simultaneous situations in a time of crisis for the community: first, the moral exhaustion of the server’s statutory members, who were worried about the impact of discussions on the health of its members, and second, the reform initiatives put forward by ubiquitous members (i.e., who were part of other communities in the drug enthusiasts network). Based on the HR ethos, which calls for responsibility and prudence on the part of each individual, the statutory members of the Blue server developed a vernacular HR adapted to the server community’s needs. In particular, they recognized the community’s role in educating inexperienced drug enthusiasts. Due to the dematerialized nature of discussions, this type of HR is essentially a ‘discursive HR’, which consists in watching out for the negative effects of discussions. In this way, members imposed rules for discussions on themselves, so that the collective expression of their interest in drugs improved the experience of everyone.