Don’t Work for Free: Online Discursive Resistance to Precarity in Commercial Photography

While increasing academic attention has been paid to the precariousness of contemporary work, less research has examined how workers organise in response. This article examines how a group of precarious workers – commercial photographers – use an online forum to resist changes to their working conditions. Our findings illustrate how the forum enables photographers to share knowledge, debate rules and organise collectively. We discuss two implications: firstly that the forum performs many of the functions of a professional association, and so gives us a new insight into how traditional forms of worker organisation may be translated in the digital realm; and secondly, that the form of collective resistance produced by the group may constitute a move beyond existing understandings of online resistance as relatively ineffectual. Our work contributes a new perspective on how precarity is reshaping workers’ collective organisation and resistance mechanisms.

to knowledge-intensive forms of 'gig working' (e.g. Sutherland et al., 2020) and this article extends existing work to consider how photographers support one another in an increasingly precarious occupation. As the technology required to take and share photographs has become cheaper and more available, the boundaries of 'professional' photography have eroded, and the rules of how photographical work is remunerated have become ambiguous and unfixed. Many aspiring photographers offer their services below cost price (or even free) in order to build a portfolio and client list. Full-time positions for photographers are few and diminishing; most work is done on a project basis. As a result, over 40% of photographers recently surveyed stated that their financial circumstances were 'difficult' or 'very difficult' (Hadland and Barnett, 2018). Photography is in a period of flux, where industrial conventions have broken down and the consumers of photographic work are able to exert significant pressure over the price of creative labour and goods. Our article considers how professional photographers alienated by these changes are using social networks to empower themselves.
Methodologically, this study is located in the transdisciplinary field of discourse studies (Wodak and Meyer, 2016). We conduct an empirical examination to discover how photographers produce and utilise discourse in online forums. Our sample is drawn from a large global Facebook group 'dedicated to removing unpaid labour from the Creative Industries' (Group Manifesto). We examine threads relating to photographic work posted between August 2017 and July 2018, asking how photographers use the forum to discursively determine and promote rules regarding the occupation. These findings allow us to contribute insights into how workers use online communities to evolve occupational rules and resist the undermining of work conditions. Artistic labour, occupational resistance and the role of professionalism Standing (2011) defines precarious work as lacking seven forms of security (income, representation, labour market, employment, job, skill and work security). With decreasing numbers of full-time, permanent jobs (Dex et al., 2000), falling pay rates (Kim, 2014), labour oversupply (Menger, 1999) and the difficulties traditional unions face in trying to represent freelance workers (De Peuter, 2011), work in the contemporary creative industries is often identified as precarious (Gill and Pratt, 2008). However, it has been argued that creative work has always been distinguished by an 'art for art's sake' orientation (Bourdieu, 1993) whereby the work is seen in some way to be its own reward (McRobbie, 2003). This idea of creative work being a vocation (Svejenova, 2005) can be a complicating factor in organising, and conceptualising, resistance to poor working conditions. It also means that workers' perceptions of their own precarity are often better understood through an identity lens than through a straightforward examination of wage rates (Patrick and Elks, 2015).
Various sociologically informed literatures explore how occupations may exert pressure on employers when work becomes precarious. Critical Management Studies offers a useful lens to theorise the role of workplace resistance, defined by Collinson (1994: 49) as activity designed to 'challenge, disrupt or invert prevailing assumptions, discourses and power relations'. Fleming and Spicer (2007) clarify that while resistance may take many forms, each is aimed at disrupting whatever rules and processes establish power in the workplace over a dominated group. In traditional workplaces, the four basic agendas of 'deflecting abuse, regulating the amount and intensity of work, defending autonomy and expanding worker control' may be employed by workers in their resistance (Hodson, 1995: 79), but such tactics are not necessarily appropriate for creative workers, or in the broader disaggregated contemporary workplace. Spicer and Böhm (2007) address this inconsistency by conceptualising the role of broader resistance movements, including extra workplace resistance, defined by Thomas and Davies (2005: 687) as a 'constant process of adaptation, subversion and reinscription of dominant discourses [that] . . . takes place as individuals confront, and reflect on, their own identity performance'. They are inspired by Foucault's (1977) theorem that social reality is produced through micro-practices constructing subjectivities such as the worker, the prisoner, the patient. A similar perspective is adopted by Fournier (1999), who argues that professionalism is a disciplinary discourse deployed across a wide variety of traditionally non-professional occupations as a means of responsibilising autonomy, making work behaviours and work identities governable. Spicer and Böhm (2007) stress that such 'Foucauldian approaches have often side-stepped the collective struggles against the discourses of management in the wider realms of society' Böhm, 2007: 1671). Combining Fournier with Spicer and Böhm raises the question of what role subjectivities and discourses of professionalism may play in new forms of worker resistance, and whether previously oppressive subjectivities of professionalism may be mobilised in favour of worker collectives.
However, Pichault and McKeown (2019) and Bain (2005) highlight how complicated it is to construct and communicate professional norms in precarious occupations. Bain's artists lack the 'shared workplace cultures and everyday social interactions' necessary to establish shared norms (2005: 26) and both studies note the difficulty 'professional' workers may face in distinguishing themselves from amateurs or hobbyists (Pichault and McKeown, 2019). Bain argues that artists lack 'the degrees or licences, prerequisites or credentials to authenticate occupational status' and claims that '[w]ithout clear definitional parameters . . . there is little incentive for society to compensate artists adequately for their labour ' (2005: 26). Despite the solitary artist trope, creative workers regularly support one another, but the literature is yet to consider how collective support and professional discourses might enable creative workers' resistance.
Online spaces present opportunities for worker organisation and the methods and sites of such activity are highly diverse, from online forums revitalising traditional unions (Edwards, 2009), to whistleblowing weblogs exposing working conditions (Peticca-Harris et al., 2015), to digital campaigns used to resist precarious contracts (Salamon, 2016). While online resistance is highly visible, it is argued to be 'subtle' (Peticca-Harris et al., 2015: 576) or 'decaf' (Contu, 2008: 370) because the workers are often anonymous. However, online platforms may facilitate resistance, not just by revealing unfair labour practices, but also through supporting workers. Sutherland et al. (2020) suggest that forums allow online freelancers to develop 'platform literacy', defined as 'the knowhow required to leverage platform resources in order to minimise the precarity of independent work, while retaining as much autonomy as possible from those structures themselves' (p. 470).
In summary, creative work is precarious. The resistance literature offers a theoretical framing for understanding how workers combat deteriorating work conditions, and the role professional discourses play in creating work subjectivities of oppression and potentially for resistance. As worker collectives evolve in the disaggregated workplace, contemporary studies have begun to explore the sites and mechanisms of online resistance. However, there remains little empirical evidence of how creative workers such as photographers use social networking websites (SNSs) to collectively organise. This is the gap to which our study relates. Following Fairhurst and Putnam (2014), we understand discourse analysis first and foremost as a methodology (rather than method) that operationalises knowledge and communication-related phenomena. This is rooted in the linguistic turn in the social sciences, which led to an understanding of language as a central resource to shape social reality (Berger and Luckmann, 1966), based on the insight from linguistics that language use is a form of social action (Austin, 1962;Searle, 1969). Because of these 'constitutive effects of language' (Fairhurst and Putnam, 2014: 271), methods under the umbrella term 'discourse studies' concentrate on the linguistic and other semiotic elements of the social (Fairclough, 2005: 916), as well as the contestation and production of meaning.

Methodology
As our interest lies in the contestation of business practices and professionalism in online forums for professional photographers, and in the link between the social practices of those professionals and their discursive practices in the forums, we turn to Fairclough's (2003Fairclough's ( , 2009) dialectical-relational approach. Based on critical realism, this approach sees: objects, entities, persons, discourses, organizations and so on as socially produced 'permanences' which arise out of processes and relations [. . .] and which constitute a pre-structured reality with which we are confronted, and sets of affordances and limitations on processes. (Fairclough, 2005: 923) People act in a network of social practices, one of which can be the practice under analysis here: being a self-employed professional. As social life is reflexive, people 'interpret and represent to themselves and each other what they do, and these interpretations and representations shape and reshape what they do' (Fairclough, 2001: 236). These discursive interpretations and representations are present in different types of interactions, such as reports by management consultants, newspaper reports or self-help publications -genres that, by guiding participant expectations, frame discourses in certain ways. Together, these representations of social practices 'may contribute the production of new imaginaries, which may in turn be enacted and inculcated' (Fairclough, 2001: 236) as well as develop moves of resistance against new imaginaries, such as the idea that aspiring photographers should work for free, or the idea to troll exploitative clients.
Our analysis is interested in online forums that contain discourses of advice and selfhelp (Locher, 2013). This genre emerged with the arrival of social networks, which enabled a new 'participatory culture' with 'low barriers to artistic expression and civic engagement' and 'some type of informal mentorship whereby what is known by the most experienced is passed along to novices' (Jenkins et al., 2007: 4). Where self-employed workers in work-related forums form virtual communities of practice (Garzone, 2016;Wenger, 1998), they produce a type of discourse normally observed offline: workplace discourse based on joint enterprise, mutual engagement and shared repertoire of a workplace community (Holmes and Meyerhoff, 1999). Previous studies have shown the value of online forums in providing support for users in healthcare (e.g. Limberg and Locher, 2012;Locher, 2013;Morrow, 2006), but there has been little focus on forums in studies of work, despite indications that online interaction can be an important tool for organising workers (Edwards, 2009;Garzone, 2016). Analysing a self-help forum for a freelance profession -here photographers -we will explore how these forums are also used by workers to share knowledge and organise in response to occupational precarity.
Despite the fact that the research object is, at least in part, constituted by our theoretical approach, we argue that the analytical advantage of a discourse-analytical approach is the opportunity to observe naturalistic data (Potter, 2002) rather than analyse data elicited by the researcher in, for example, interviews or focus groups. We can observe what issues the participants orient to in a genre they co-produce with each other rather than with the researcher. This approach is sometimes referred to as netnographic (Kozinets, 2010) as it treats social media content as a text in the same way as an ethnographer would treat the text produced by their observations (Elvey et al., 2018). Although some netnographies use monologic data to analyse prevailing discourses and attitudes (e.g. Elvey et al.'s online comments), others explore how dialogic online spaces such as forums provide support to users (e.g. Kendal et al., 2017;Kirk and Milnes, 2015). We are also able to observe how the categories created by our participants feed back into reported social practice, as participants reflect on this in the discussions observed. Although we are mindful that we have access only to the identities that the forum users construct and their reporting of action, we can therefore also analyse the influence of the forum discussions on social practice through the discursive construction of these influences in the forum.
Our corpus is composed of threads related to photography posted from August 2017 to July 2018. The public Facebook group we draw this corpus from comprises over 23,000 members worldwide (predominantly US, UK and Australia), working across many content production occupations (predominantly journalists, photographers and designers). The group is the largest and most active of its type with around 100 posts, 1500 comments and 10,000 reactions per month. Of the 50 relevant threads posted during this period, we gained informed consent to examine 35, totalling almost 59,000 words. We gained informed consent by asking the group's 'admins', who manage content on the page, for their permission, and then by approaching the poster of each thread. We only used threads for which we had explicit permission from the original poster (OP), some of whom asked for anonymity and some who asked to be named. We anonymised all comments and checked for traceability wherever they are quoted. In each thread, a photographer -the OP -posts a query or a complaint (generally about payment or contracting), and other forum members -'commenters' -reply with suggestions for resolving payment, comments on the case, or similar experiences. Our corpus therefore gives us live access to photographers making sense of the economic status of their work and the value of their creative content.
The threads were inductively coded using NVivo 11 (QSR International). We collaboratively coded the first three threads within the broad themes of: (1) genre of post; (2) mentions of precarity; and (3) types of advice or support given by commenter. This resulted in an initial set of 63 first-order codes, which were folded up into eight secondorder themes, such as professionalism, rules and response types. These code families were then used by the authors to independently code the remaining 32 threads before comparing to establish multi-coder reliability. At this stage we identified a series of distinctions being applied to the rules by commenters, so recoded inductively until no new codes were produced (12 first-order codes and two second-order themes were identified for distinctions). An illustrative example of a coded thread is shown in Figure 1.

Findings
Before presenting our findings, it is important to reflect on the affordances of the medium of the 'Facebook group' (Bhatia, 1993;Herring, 2007) as an interactive online genre that allows multimodal messaging. In our data, OPs frequently screenshotted requests for work or included images of plagiarised work, and other members responded via multimodal comments. Threads were available to scroll back through or to search, making the forum a de facto compendium of knowledge. The forum was governed by six 'admins' who either founded the group or were selected from the group's most active members. The admins approved group membership requests and occasionally 1 moderated the discussion (typically by removing repetitive, abusive, or off-topic posts). Bearing these factors in mind, this section discusses the three main ways that photographers used the forum to: share knowledge; debate rules; and organise collectively. Source: Thanks to Brian Duffy for providing the post, who asked us to leave it de-anonymised.

Sharing knowledge
The forum enabled photographers to exchange advice, a function well documented for health forums, but less for work-related forums. Knowledge sharing on the forum took two forms: posts by members sharing insights, and posts from members looking for advice. Where insights were being shared, the OP tended to self-identify as an experienced member of the forum, and would describe the situation, providing commentary explaining their actions. Where commenters agreed with the interpretation/actions of the OP, they tended to share similar experiences and state support for the OP's actions. For example, this OP shared his response to being asked to work for free: Hi Brian, I'm looking a Photographer to do a Photo Shoot with . . . Actress ************* (based in Hertfordshire) for a Front Cover for our Online Magazine . . . Would you be prepared to do this for credit-only on the photos + a feature of your past work? Just looking for an expression of interest at this point -thank you! : : : I appreciate the offer of working for free but will politely decline. Best wishes. Perhaps I'm too polite (crying laughing emoji) (post; subcode: sharing experience) Seventeen members 'liked' this post and 14 commented to offer their interpretations of the equity of the offer and the politeness of the OP's response: I note they didn't even offer you the gig if you were generous enough to agree to it -they were 'just looking for an expression of interest at this point'.
As these things go, that was a refreshingly bullshit-free request, and a classy response. It's best to save the swearing for more deserving cases.
(comments; subcode: sharing experience) Where interpretations differed from the OP's, he re-appeared in the comments to justify his course of action over alternatives (e.g. being rude to the client). In many threads this type of engagement with the OP led to discussions or arguments (sometimes comprising up to 50 comments) about interpretations and tactics.
However, the majority of knowledge-sharing posts were instigated by OPs who had confronted a new type of work, client, or client behaviour and were asking how to deal with it (e.g. how much to charge for a shoot). For example, one member started her post with 'Hi all. I need help', and asked for advice on how to price a photograph for sale to a US TV company. Advice was offered from 63 commenters, often prefacing their advice by highlighting relevant experience/expertise (e.g. 'I am in New York, I work in US telly . . .'). In the comment thread, the OP discussed ideas with the commenters, and informed them of the outcome of actions taken offline by screenshotting the conversation with the client, anonymising it, then including it as a comment on her post. Although the advice didn't result in her selling the artwork, she repeatedly commented on how useful the group was (e.g. 'I wouldn't even had known what to put had it not been for this thread!'; 'Thanks everyone for your help!'). These posts allowed newer members to easily source advice, and each discussion enlarged the forum as a compendium of knowledge. Established members on the other hand used this knowledge base to attempt to educate newer members, to prevent them undermining the credibility and/or financial value of the work (e.g. by under-charging).

Debating rules
Four types of forum rules were identified in our analysis: (1) not working for free; (2) pricing work; (3) defining bad client behaviour; and (4) dealing with bad client behaviour. Rule 1 was the most cited and least contested, while rules 2, 3 and 4 were subject to debate, with 4 being the most regularly debated and most contentious rule. While the application of categories and rules is typical for institutional language use (Sarangi and Candlin, 2011), we could, in the forum, observe the usually hidden process of their making. This section explains how these rules were constructed and debated, through four mechanisms: reference to personal experience; comparisons with other forms of commerce/professional work; deconstructing 'client speak'; and the use of memes. Each is illustrated using the second rule type, which concerned whether photographers should set prices based on business factors (such as supply, demand and the potential future value of the client) or using a professional logic of set prices.
Members who advocated using a business logic often drew on personal experience which emphasised the need for pragmatism when pricing work, as shown in the example below where a commenter is offering advice to a photographer who was asked to shoot a corporate awards dinner for free: In my experience almost every job involves negotiation. Today -literally today -I shot the first part of a corporate event that I initially bid more than $3500 to do. The client came back with, 'I have $2500 to spend on photography'. She trimmed the hours of coverage needed to make it work. When a client is willing to do that, I am more than willing to slightly trim my rate to meet a budget.
(comment; subcode: negotiation) However, other members in the forum argued for a professional approach with fixed prices and justified this rule using the second mechanism -making comparisons to other forms of commerce/profession. For example, when a restaurant owner asked for free photography in exchange for reciprocal advertising, the forum members trolled 2 the restaurant's page by posting joke reviews which presumed the restaurant owner would offer them free food in return for exposure: Had to . . . listen here . . . PAY for my food. Absolutely unheard of. I even offered to share their page on my Instagram story every day for a week.
Drawing this comparison allowed forum members to demonstrate to outsiders (e.g. the restaurant owner) that creative labour and products (e.g. photographs) should be valued in the same way as their own labour and product (e.g. food). They also drew comparisons to professions, such as this recommendation to an OP whose client was a bridegroom offering a very low fee for a wedding shoot: '[Tell him] "Okay . . . call your local plumber and ask him for a quote on a 5 hour job" (or however long they are asking you to work for). [Then say] "I'll do my job for the same price he is quoting"' (comment; subcode: profession comparison).
Often, the application of pricing rules depended on 'decoding' the offer made by the client. This third mechanism was commonly employed to analyse and debunk offers of monetary 'equivalents' (most commonly 'exposure', 'credit', 'quid pro quo', 'portfolio building' and 'time for pictures'). In many cases, forum members highlighted that apparently legitimate job adverts were not offering paid work, but instead substituting a monetary equivalent. As well as identifying such adverts, commenters explained the true value of monetary equivalents, such as this commenter who explained the reality of working celebrity events for exposure: Aha, ignoring the wanting you to work for free bit, shall we disect [sic] the claims of benefit? Like being able to mingle and network while at all times being made to take photos of these people? Of that, at these events you really can't mingle with celebs, it's too big, busy and [they are] directed by their managers to be at a place at a specific time. Or that if you don't know anyone to introduce you, parachuting yourself in can, well, be received badly. If you could even get near for bodyguards. So, the reason you are there outside of taking photos? For their ego and self-promotion. There really isn't any benefit.
(comment; subcode: decoding client speak) In some cases, as below, where a TV company were trying to buy a photograph, forum members judged the intent behind the client's offer based not only on deconstructing the offer itself, but also on the characteristics of the client: The phrasing is insinuating they're being reasonable by offering £15 and OP is being unreasonable, and they're mystified by it.
They're a show, it's scripted. They have experienced, professional writers there. This is specifically written to try to hand-wave away how badly they're lowballing.
(comments; subcode: bad client behaviour) Finally, much of the deliberation was encoded as textual and image memes (the fourth mechanism), 'a popular term for describing the rapid uptake and spread of a particular idea presented as a written text, image, language "move", or some other unit of cultural "stuff"' in an online space (Knobel and Lankshear, 2007: 202). The most prolific textual meme 'No is a whole sentence' was used when an OP or commenter complained about a client offering unfair pay. Image memes were also used extensively, such as in Figure 2 when an OP described being asked to work 'time for pictures' (the model and the photographer work for free and each gets access to the pictures afterwards for their portfolio). Memes allowed commenters to quickly, clearly and often humorously signal their position in debates over rules (e.g. by using the image in Figure 2 to make fun of commenters who advocated working for free). However, they also seemed to play an important role in building and expressing collective outrage over precarious work, as explored in the next section.

Organising as a collective
The forum was also a political space, used to establish consensus on issues and actions for the occupation. The basis for collective action is usually a shared grievance, and the forum allowed a diverse range of freelance photographers to identify common gripes (e.g. being undercut in pitches by novices who were working for free to build a portfolio). Alongside previous research, our analysis showed that memes were important not only in promoting or dissuading offline acts, but also in building a sense of collective identity (Gal et al., 2016). For example, the image in Figure 3 was commonly used when a client offered to pay with 'exposure', and the 'get the popcorn' textual meme was used when a naïve client appointed a photographer purely on price: 'Now I'm ready to get the popcorn and watch as someone has entrusted a brand identity to the hands of someone with no experience and [who] is being paid considerably less' (comment; subcode: get the popcorn). The use of these memes across the wide range of photographers on the forum built a sense of shared identity across regions, photography types and experience levels, allowing them to establish a shared mission to confront common issues. Memes around the value of 'exposure' and working 'time for pictures' all serve to reinforce the message that photographers should be paid fairly for their work.
The forum precipitated offline action by facilitating the exchange of tactics used to tackle bad client behaviour. Members advocated either educating the client (if they were seen as naïve) or punishing them (if they were repeat offenders). As we might expect in an online setting, commenters were often emphatic about their proposed course of action, as seen in this reply to an OP who complained about being asked to re-shoot a location to get a repeat job: This is insane -you already photographed the exact same hotel. Offer a meeting to go over your portfolio, but no shooting. This is not a fucking episode of -Our Hotel's Next Top Photographer. Extremely disrespectful and unprofessional.
(comment; subcode: respect) It was common for OPs to update the forum on offline client negotiations as they unfolded, via the comment thread. For example, one OP asked the forum for advice when a UK tabloid newspaper used her photo without attribution or payment, and after discussing tactics with a number of commenters she commented a copy-paste of the newspaper's response, which commenters then evaluated. Commenters were always interested in the outcome of the offline transaction and sometimes prodded for an update when the OP was not forthcoming (e.g. 'Don't be shy. Please let us know how it turns out'). As such, while the offline action was undertaken by the individual (the OP), it was often collectively guided by the forum members to achieve their own interests of improving labour conditions in the industry.
This sometimes spilled over into collective action when OPs 'named and shamed' bad clients, of their own volition or by request (e.g. 'That's why we have places like this to expose what really are con artists, who indirectly steal your time and money') (comment; subcode: name and shame). Naming was sometimes aimed at making other photographers aware of bad clients, but it also led to collective and overt political actions such as 'trolling' clients' social media pages to 'mess with them' or 'wind the buggers up and waste their time' (as they have wasted the photographer's time) (comments; subcode: trolling). The group trolled the Facebook page of celebrity tattoo artist Jay Hutton after his wedding organiser advertised for a free wedding photographer in exchange for social media advertising. This was picked up by various news websites, including the website of British tabloid the Daily Mirror (Romero, 2018). A further example occurred after a long thread about Surrey Police advertising for a volunteer photographer. Despite considerable debate on the thread regarding whether volunteering was permissible in this case, 3 members of the group trolled the police force's Facebook page, causing a long online debate between followers of the force's page and others (including forum members), which prompted a reaction from the force's digital communications manager -picked up and reported by the Press Gazette, a British media trade magazine (Tobbitt, 2018) (Figure 4).

Discussion and conclusion
As the model of traditional employment declines and labour unions struggle to meet the demands of newly freelance workforces, it is important to identify and examine the new mechanisms by which workers seek to challenge the sources of their precarity. Existing literature shows how online spaces such as weblogs (Peticca-Harris et al., 2015) and web forums (e.g. Edwards, 2009;Sutherland et al., 2020) facilitate whistleblowing, information sharing and mutual support, and provide new mechanisms for the renewal of collective worker organisation. We extend this work by examining how precarious workers use social network sites to challenge their precarity by trading advice regarding client relationships, debating rules around professional behaviour and organising collectively. Our study shows how the relatively young technology of SNS can support the formation of occupational communities in groups that do not have a shared, offline workspace and that lack the established institutional structures needed to defend deteriorations of working conditions (e.g. professional bodies or unions).
The discourse-analytical lens used in this article allows us to gain an insight into the reflexivity of social life and the dialectic between discourse and social practice (Fairclough, 2001): We see how members of the forum reflect on professional practice, and how they develop rules of professional behaviour. The tensions over rules and their application make SNS forums a heteroglossic genre (i.e. a genre that overtly represents many voices). This makes them an ideal pool of data to demonstrate the complexity of collective worker organisation, as the written nature of the naturally produced data allows the observer to observe without intervention.
Our data show that the Facebook forum allows workers to challenge precarity in two ways. First, the forum performs some of the functions of a professional association, enabling the development of a shared repository of knowledge around work practices and client transactions, and allowing forum members to use professionalism as a means of exerting disciplinary pressure on other workers. Second, the forum facilitates collective action on and beyond the Facebook platform to exert pressure on clients to pay appropriately. Taken together, these functions represent a variety of discursive work, enabling extra-workplace resistance not yet discussed in the literature (Spicer and Böhm, 2007), seeking to rectify the skewed balance of power in client-photographer negotiations by decoding client speak, sharing negotiating tactics, naming and shaming bad clients, and organising politically to punish bad client behaviour. We will now explain the two functions of the forum in more detail.

Comparing the forum to a professional association
Forum members use the forum to challenge their precarity by undertaking some of the activities traditionally undertaken by professional associations (PAs); a form of regulatory agent which has a number of exclusive rights over a domain of activity (Greenwood et al., 2002). Drawing on Abbott's (1988) seminal work, the forum performs three of the core functions of a PA. First, establishing membership criteria by defining what a 'professional' is and does (Abbott, 1988). Forum members set rules which delineate professional conduct and reinforce them by writing posts which show the rules in action, including using memes to draw attention to commonly held rules (such as 'don't work for free'). PAs also educate new entrants (Abbott, 1988), as occurs informally in the forum when new members ask the forum for advice on the 'rules of the game'. However, these rules are not universal or fixed for photographers in this global forum (and workplace); in the data, we see debates over rules (e.g. due to regional customs and legal regimes). Accordingly, the learning function of the forum is live and continually reenacted, rather than being codified in a set of standards and training courses.
Finally, PAs enforce standards through monitoring (Abbott, 1988), and the forum performs this function as commenters approve or sanction the OPs' interpretations and actions, textually or using image memes. However, the lack of codified standards and sanctions means that the policing function of the forum is not legalistic (like a PA), but normative. The use of normative control in the forum, particularly how the members lionise the subject position of 'professional' in opposition to the pejorative 'hobbyist', illustrates Fournier's notion that 'the appeal to the discursive resources of professionalism in new occupational domains potentially acts as a disciplinary mechanism that serves to profess "appropriate" work identities and conducts ' (1999: 280). This disciplinary mechanism is rendered all the more effective because the normative function of the forum can operate at a distance in the globalised occupation of commercial photography. This use of professionalism as a form of boundary work has been identified in previous studies of media workers (e.g. Lewis, 2012), but not in the context of collective resistance.
However, the forum cannot (or has not yet found a way to) replicate all of the functions of a PA, such as representing the profession to outside audiences (Greenwood et al., 2002). PAs are traditionally funded through the organisations or individuals they represent, so the work of regulating the profession is properly funded. The forum, however, is 'staffed' by members who offer their advice (and some who serve as 'admins') all without payment (ironically given the aim of the forum is to eliminate unpaid labour). As such, while we would not argue that a Facebook group operates as a de facto professional association, such forums may represent a form of worker organisation and collective resistance which, over time, come to take the place of more traditional worker institutions, such as PAs.

The online forum as a means of collective resistance
The traditional agendas of workplace resistance (Hodson, 1995) do not appear representative of photographers, or of other freelance occupations. While workers can still deflect abuse (and they do so regularly on the forum through sympathy and humour), the lack of bargaining power that freelancers often have over their clients means that regulating work, defending autonomy and expanding control (the traditional methods of resistance) are unlikely to be effective (see also Pichault and McKeown, 2019). As such, the collective discursive struggles of online groups are an important mechanism for worker resistance which are under-explored in the literature (Spicer and Böhm, 2007).
Our research shows that SNS forums can play an important role in creating communicative spaces for workers which address the imbalance of transactional power between freelance workers and their clients by enabling workers to share experiences, exchange tactics and to guide the transactions of others in live time. These communicative spaces are also, as Edwards (2009: 443) notes, 'essential conditions for the building of collective identities and solidarities'. The sharing of similar experiences and the performing of collective outrage on the forum establish the solidarity necessary to underpin collective action, both on client Facebook pages and across a range of online platforms (such as Twitter, TripAdvisor and client websites). This shows that online worker resistance is not always 'subtle' (Peticca-Harris et al., 2015: 576) or 'decaf' (Contu, 2008: 370), but can be overt, public and may be effective in producing offline change (e.g. by provoking public apologies). Therefore, our findings support a positive view of SNSs' potential to shape the politics of freelance work and to resist precarity.
These findings open up a range of avenues for researching online collective resistance. Future research could adopt the methods used to study consumer and activist-led social media campaigns (e.g. Albinsson and Perera, 2013), to examine the effectiveness of different tactics on the views of clients, client stakeholders and other workers, and on offline client behaviour (e.g. Champoux et al., 2012). Such analysis would be able to unpick the value of online collective resistance in changing the outcomes of offline transactions with employers and clients. While our research focuses on photographers, similar activity takes place on the forum related to other content production professions (in particular, journalism and web design) and the applicability of our findings to these other occupations would be a valuable area for future research. In addition, the rise of workbased SNSs (e.g. Facebook's 'Workplace') raises the question of whether similar online mechanisms of collective resistance may be identified in traditional workplaces.

Limitations
One of the limitations of studying worker deliberations online is the inability to fully understand the characteristics of participants (e.g. do they identify as a photographer?). We do not know whether the performativity of online forums affects how members post or comment. To some extent, we mitigated these concerns by the first author dwelling in the forum for over a year prior to data collection, to build knowledge on active participants and on the conventions by which people express themselves. In line with generic limitations of online data gathering, we acknowledge that forum members' offline behaviour may differ from that reported (Jenkins, 2008) and we do not claim representativeness of the forum, as there are no data available on which photographers do not participate in the forum, and forum users may be more politicised than the general population.

Concluding comments
As people utilise digital technologies to move more of their work activities online, and as traditional forms of worker collective are challenged in increasingly precarious occupations, we see our research as being a step forward in understanding which of the conventions of traditional worker organising persist in an online workplace, and which will be re-shaped or made redundant in the new world of work. While we associate the Facebook forum studied with some of the functions of a traditional professional association, its disaggregated nature makes compliance with its collective judgements entirely voluntary, and also means that 'guerrilla' campaigns aimed at tackling bad client behaviour (such as wealthy celebrities asking photographers to donate their labour in return for exposure) take the place of more traditional collective action. Unlike previous work which has highlighted the subtle and covert nature of online resistance, our findings indicate that occupational fightback in the digitalised workplace can be both deliberative in origin, and immediate, direct and public in effect.