Friendship and trust in the social surveillance network

Abstract

This article draws on findings from semi-structured interviews to investigate an important and previously unexplored aspect of the social networking site Twitter, which is the ways that users can employ lateral surveillance to initiate friendships. Social penetration theory (Altman and Taylor 1973) is used to investigate the ways that Twitter provides a platform to technologically augment friendship evolution. Three consecutive phases of online surveillance are proposed, which are: unidirectional 'studied' following; 'reciprocal' following, when interaction between the user and the Tweeter only occurs online; and fully fledged friendship, when online relations are supplemented by offline meetings. The visibility of users via their Tweets creates information that others can use to evaluate character in what might be considered a phase of pre-friendship screening; this information also generates trust and solidarity prior to an offline meeting. Perceived similarity between the user and the Tweeter was contributory to an assessment of their worth as a potential friend. Geographic proximity and 'transferable trust' were also factors that helped offline friendship to develop. Ultimately, Twitter was an important tool for sociality, and provided a wider source of potential friends than was available offline. [PUBLICATION ABSTRACT]

Headnote

Abstract

This article draws on findings from semi-structured interviews to investigate an important and previously unexplored aspect of the social networking site Twitter, which is the ways that users can employ lateral surveillance to initiate friendships. Social penetration theory (Altman and Taylor 1973) is used to investigate the ways that Twitter provides a platform to technologically augment friendship evolution. Three consecutive phases of online surveillance are proposed, which are: unidirectional 'studied' following; 'reciprocal' following, when interaction between the user and the Tweeter only occurs online; and fully fledged friendship, when online relations are supplemented by offline meetings. The visibility of users via their Tweets creates information that others can use to evaluate character in what might be considered a phase of pre-friendship screening; this information also generates trust and solidarity prior to an offline meeting. Perceived similarity between the user and the Tweeter was contributory to an assessment of their worth as a potential friend. Geographic proximity and 'transferable trust' were also factors that helped offline friendship to develop. Ultimately, Twitter was an important tool for sociality, and provided a wider source of potential friends than was available offline.

Introduction

Participants on social networking sites such as Twitter are embedded in a form of lateral, mutual or participatory surveillance (Albrechtslund 2008; Andrejevic 2005: 481; boyd and Ellison 2007). As the use of Twitter on computers and mobile devices increases, there is a need to understand the ways that lateral surveillance impacts on social relationships, particularly friendship (Nielsen 2012). Launched in 2006, Twitter is a micro-blogging social networking site that is used in a range of ways, including as a broadcast platform for news media organisations and celebrities, and as a way for individuals to socialise online (Gruzd et al. 2010; Marwick and boyd 2010; Nielsen 2012). Existing research on surveillance and social networking sites such as Twitter has focused on aspects such as: the differences, if any, between on- and off-line friends (Boase and Wellman 2006; Zoppos 2011); the 'imagined audience/ community' (Gruzd et al. 2010; Marwick and boyd 2010); use by celebrities (Van Krieken 2012); surveillance in romantic or intimate relationships (Gershon 2011; Miller 2011); privacy and risk (Andrews 2012; boyd 2011; Coll et al. 2011); and policing and security (Trottier 2012). However, the ways that Twitter can facilitate friendship and generate relationship trust has been neglected as a topic of inquiry.

This article draws on empirical findings from semi-structured interviews to engage with a relatively novel but substantive theme, namely the ways that some individuals have adopted lateral-or peer-to-peer- surveillance practices on Twitter as a way of identifying and screening new friendships. It extends Altman and Taylor's (1973) social penetration theory to show how online Twitter 'following' can be a starting point to offline friendship rituals. This article will state how and in what ways online surveillance is contributing to friendship formation and hence suggests that contemporary modes of friendship are increasingly technologically augmented. Twitter has become well known as a platform for the online branding of the self (Bauman and Lyon 2013: 40; Gershon 2011; Marwick 2012; Marwick and boyd 2010). However the findings show a different use of Twitter, which is when following develops in stages that are similar to the procession from acquaintance to friend that occurs in offline friendships. Three discrete stages of following are identified as: studied following, which is unidirectional; reciprocal following, where online interaction is initiated and shared, and; fully fledged friendship, where online mutual surveillance relations are a precursor to offline co-present meetings. In the first two stages the ongoing visibility of the individual via Tweets is discussed to investigate the ways that this is important for the development of trust. This article starts with a review of the literature, and then presents analyses of excerpts from the interviews.

Friendship rituals

Co-present friendships develop in stages and the beginning or acquaintance stage is when two individuals find out about and get to know each other, and decide whether or not they want to continue the friendship (Blieszner and Adams 1992; Duck 1977; Newcomb 1961). The process of making friends is subject to many factors and involves investments of time, emotional energy, and very often, material resources. Altman and Taylor's (1973) social penetration theory is useful to understand the complexity of friendship initiation as it charts the movement from co-present stranger, to acquaintance, to friend. The theory proposes three key factors that 'hasten' social relationships (Altman and Taylor 1973: 4). Firstly, personality which can include social preferences, for example choosing to socialise in groups or one-onone. To enact social preferences, an individual also needs the requisite social skills as well as access to others who are available for friendship. Secondly, liking or feeling 'attraction' to another person has been identified as a reason to pursue an acquaintance (Duck 1977: 16; Newcomb 1961: 6). Usually people feel affinity to others with whom they share hobbies and values, or with those identified as being similar to them, for instance, the same life-stage (e.g. starting a family). Altman and Taylor (1973: 53) note that similarity is a proxy for liking because there is an association with greater relationship rewards in future. Of course, friendship is a two-way process, and liking another person does not necessarily mean that an individual will be liked equally (or at all) in return. For this reason, friendship initiation is risky, as rejection may lead to loss of face, meaning feelings of shame or embarrassment (Cupach and Metts 1994). Thirdly, the environment or context can contribute to and impede the development of friendship. Until relatively recently, the context of friendship was marked by co-presence. Prior research has identified the development of offline friendships in a range of contexts, such as local neighbourhoods, at work, or via volunteering (Bulmer 1986; Wilkinson 2010). Geographic proximity, for instance, being neighbours, can increase interaction frequency and hence contribute to friendship development, although it is an inducement only when liking is present. The development and proliferation of communication technologies has meant that some friendships have become more mobile and flexible; however, whilst these technologies do make it easier to stay in touch, geographic distances still remain an impediment to their enactment (Allan 2008; Clarke 2005).

Trust and friendship

In addition to perceptions of similarity and liking, a key aspect in developing friendships is perceptions of trust (Altman and Taylor 1973: 12). Trust is coterminous with qualities of benevolence and caring, loyalty, reciprocity and equality (Friedman 1993: 189; Rawlins 2009: 209). Simplistically put, mutual trust is produced when friends show respect for one another and act out of kindness. These behaviours and felt experiences indicate that each person cares for and likes the other. Unlike family or spouse relationships, the obligations towards friends are less clearly articulated in a formal or legal sense, or in societal norms, which is one of the reasons that agreement on what constitutes a friend can sometimes be ambiguous. This ambiguity partially arises from the paradox that friendship is understood as a voluntary relation and hence one without obligations (Spencer and Pahl 2006: 85). Yet, whilst friends are not supposed to keep a record of favours to be returned, the expression of benevolence-being nice for the sake of it and without any expectation of reward-does render expectations of reciprocity (Mauss 1954). The elementary basis of friendship includes both giving, and receiving, in order to generate an exchange relationship that is characterised by equality rather than asymmetry. This reciprocity shows loyalty and commitment to the friend and signals that the friendship is continuing (Spencer and Pahl 2006: 85). Mutual reciprocity is a key requirement of co-present friendship functioning to both maintain and develop the relationship by stimulating feelings of solidarity and trust between the dyad (Rawlins 1992: 12). Although mutual reciprocity can and does fluctuate over time, it is generally a quality that is equal between the two co-present friends. It helps generate ongoing equality and assists in balancing the relationship (Rawlins 1992: 12; Rawlins 2009: 9). Social penetration theory suggests that as individuals become acquainted over time offline, they uncover information about each other's personalities that make it possible to infer trust (Altman and Taylor 1973: 55). Less, however, is known about how individuals perceive and negotiate trust online.

Lateral surveillance and social networking

Within the Twitter site itself, surveillance and the establishment of trust operate on both vertical and horizontal planes. In regards to the vertical operation of surveillance, the 'privacy paradox' goes some way to explaining the somewhat contradictory attitudes displayed by many users. Cohen frames the paradox as "a relatively high generalized concern about privacy but a relatively low level of concern about data generated by specific transactions, movements, and communications" (2012: 108).

While users are generally concerned about their privacy, this concern is surprisingly absent when signing up for, and using, the sites. Users cede a certain amount of privacy in return for access to the site. There is a form of trust in operation because users largely feel that giving information will not be detrimental, and hence that those commodifying their data acquire and process it in a fair exchange. The horizontal operation of surveillance is of most interest here, and can be understood as 'lateral surveillance' (Andrejevic 2005), 'participatory surveillance' (Albrechtslund 2008) or 'social surveillance' (Marwick 2012). To expand, Andrejevic (2005) frames 'lateral surveillance' in regards to the democratisation of access to surveillance technologies as a facet of neoliberal governance whereby a need to monitor others is necessitated by an inherent mistrust and uncertainty about their potential 'risk'. Marwick (2012: 379) distinguishes between the traditional or 'Panoptic' surveillance (top down), or 'sousveillance' (bottom up), by stating that the power is decentralised in this 'domesticated' form of online surveillance. Albrechtslund (2008: 3) elaborates on his concept of participatory surveillance by suggesting that online sociality is characterised by the following: 'eternal friendship' whereby digital trails-what Bauman refers to as a 'digital fingerprint' (Bauman and Lyon 2013: 22)-of online social connections are stored indefinitely; 'searchability', meaning that information may be conveniently located using key words; 'perfectly reproduced', as information may be taken and used in other contexts or even altered, and, finally; 'invisible audiences', whereby even though people 'obviously' communicate with a specific audience of their friends in mind, others who are using the internet may also see this information.

As there are many definitions of online surveillance, for clarity we situate how we use this term. Importantly, social surveillance 'takes place between individuals' (Marwick 2012: 379, emphasis added) and sharing and reciprocity are a positive and empowering function of 'mutual' surveillance (Albrechtslund 2008: 7). Reciprocity occurs as both users send and receive information, a facet of Twitter that facilitates sociality (Marwick 2012: 382, 384). The ways that individuals use Twitter is constantly changing, however, Sysomos's (2009) analysis of 11.5 million accounts provides some insightful data. Firstly it shows that, despite the large number of accounts on Twitter, only 5 per cent of these generate 75 per cent of site content. More than half of all account holders-62 per cent-followed 18 people or less, and 92 per cent followed less than 100 people. Only 0.94 per cent followed 1000 people or more. With regards to the numbers of followers, 94 per cent had less than 100 followers, and only 0.68per cent had 1000 followers or more. Contrary to the popular notion that Twitter users seek to gain an audience and amass as many followers as possible, these figures suggest a more nuanced understanding of user preferences and behaviours.

Surveillance over social networking sites by corporations and government in order to sort and profile population, customers, and opinions, has been well documented (e.g. Gandy 1993, 2009). Of more interest here is the ways that Twitter users screen each other as a precursor to offline friendships. Online surveillance may seem far removed from conventional friendship rituals. However, online screening can assist in the rational evaluation of a potential relationship (Andrejevic 2005). For example, inspecting a user's Facebook profile-known as 'social browsing'-can reveal fragmented information that may be used to make character inferences that influence whether or not to pursue friendship (Gershon 2011; Lampe et al. 2006). Similarly the practice of 'Googling' is a way of finding out about current or prospective partners (Andrejevic 2005; see also Gregg, this issue). Hence, such surveillance techniques, and the knowledge they collectively produce, are used in ways that fashion lines of sight, meaning metrics by which individuals can be assessed and adjudicated, and their prospective behaviours better anticipated.

Method

This research draws on interview transcripts from a separate project that investigated whether Facebook can be considered a 'technology of the self'. This is when individuals use technologies such as writing a diary or self-improvement to performatively construct themself (Foucault 1998). Participants were recruited via a snowballing strategy connected to an online survey that was disseminated on the researchers' social profile on Facebook and Twitter. Individuals that were friends of the researcher were not eligible for the research, while friends of friends were eligible to take part. The survey attracted just over 500 responses internationally, and all participants were invited to take part in follow-up research interviews either face-to-face or via Skype. Twenty Australia-based participants took part in follow-up interviews, eight of whom were demonstrably keen to talk about their use of Twitter. Material from these eight interviews has been analysed and used in this paper. Further, six of the eight respondents participated in a second interview where they were asked specific questions about surveillance, trust and making friends on Twitter in order to explore their perceptions in greater depth. Interviews were analysed using open and thematic coding to capture and contrast the varying responses (Bryman 2008: 542-4, 551; Gromm 2004: 189; Neuman 2006: 461; Liamputtong 2011: 216-7, 282-5). Interview excerpts have been chosen to succinctly exemplify our argument (Gromm 2004: 19, 185). Finally, while the interviews yielded detailed insights, we suggest that, given the sample size, findings are exploratory and do not exemplify the entirety of Twitter users. For ethical reasons, names and identifying details have been changed to protect the participants' anonymity.

'Studied following' of an unknown Tweeter

All interviewees using Twitter (hereafter 'users') followed Tweets from individuals ('Tweeters') that were not previously known to them either in co-present or in online encounters. Individuals on Twitter make themselves visible by ongoing Tweets, (messages of 140 characters and less), and following is when a user chooses to view another Tweeter's Tweets. A user may only follow public posts from a Tweeter when the user is on the Tweeter's list of followers, and the default setting on Twitter is that all posts are publically visible. The user may also choose to interact specifically with another Tweeter by direct messaging to them, and these posts are not publically available for anyone to read. The Tweeter and the user may only direct message one another when they are both engaged in mutual following (Gruzd et al. 2010; Marwick and boyd 2010; Murthy 2013). Twitter is often compared to Facebook, yet the design and use of these two sites is markedly different. The architecture of Facebook fosters the reiteration of existing offline social connections (boyd and Ellison 2007; Gershon 2011; Zhao et al. 2008; Trottier 2012: 45). In contrast, Twitter is a 'directed friendship model' because it facilitates the opportunity for a user to follow many Tweeters, a feature that participants in this research welcomed (Marwick and boyd 2010: 3, 14). For example, Kirsten (44 years), a Paralegal from North Sydney, New South Wales, said:

Twitter can be just anybody, anybody at all [laughs] which is probably what's so awesome about it.

Similarly, Tracey (28 years), an academic from Dandenong, North Victoria stated: 'I'm a Twitter whore, I'll follow anybody and anyone, at least for a little bit'. Following involved gradations of participatory surveillance. The first mode of following was epitomised by Kirsten and Tracey, who frequently watched over another Tweeter. Anonymously watching online has been previously identified as 'passive', although we argue that this implies a lack of activity by the follower (Ramirez et al. 2002: 220-221). Instead we propose the term 'studied following' which was marked by deliberate and considered surveillance of another Tweeter, even when this appeared to be spontaneous, because the user had consciously sought to find and view their Tweets. The followed Tweeter (and the user) were complicit in this lateral surveillance by being visible on, actively participating in, and sharing information through Twitter (Albrechtslund 2008; Andrejevic 2005: 481; boyd and Ellison 2007; Lyon 2007: 14; Marwick 2012).

Studied following was also marked by a lack of reciprocal social relations, because the architecture of Twitter facilitates following 'anybody', without necessarily knowing or interacting with them. Dolores (40 years) a freelance writer from Newcastle, New South Wales, explained:

[On] Twitter ... you can just follow someone, and they don't have to acknowledge you in any way.

Following does not necessitate the Tweeter's acknowledgement because as Delores added, they do not have to 'want that connection'. Studied following could be without reciprocity, so it was an uneven or asymmetrical relationship (Gruzd et al. 2011; Marwick and boyd 2010). Studied following alone did not constitute friendship as it was not reciprocal, although it is possible that the user may feel liking for the Tweeter which is particularly important for generating future sociality, as feelings of affinity are the beginnings of a social bond.

Studied following was a one-sided practice that occurred without mutual reciprocity so the balance of friendship was uneven. Yet, studied following also enabled the user to maintain their privacy because they had not divulged anything to the Tweeter which removed the risk of initiating a new relationship. Within Twitter this non-reciprocal, technologically-mediated observance of strangers is legitimated as an acceptable social norm, mimicking the para-social relationship of a celebrity with her or his fan base (Horton and Wohl 1956; Marwick and boyd 2010). The use of Twitter by celebrities such as comedians, politicians and singers has been noted as a way to construct an audience (Van Krieken 2012: 134). Indeed, repeatedly viewing a Tweeter's posts implies the user's loyalty, which is an attribute of a fan, and a friend (Rawlins 1992: 9-10). Studied following did articulate a relationship of sorts, yet the one-sided nature of the gaze was the antithesis of offline friendship initiation that involves the generation of mutual trust and reciprocity. The mediated character of Twitter provided opportunities for users to disregard the normative standards governing face-to-face interaction. Individuals in public operate knowing that they can be seen yet physically following random strangers is socially unacceptable and such behaviour has been categorised as deviant and abnormal. Indeed, stalking, voyeurism, and scopophila are examples of conduct that have negative values ascribed to them. Instead, co-present individuals afford unknown others the expected 'civil inattention', which mitigates against the observing gaze (Goffman 2008: 15). Yet on Twitter, following is manifest in, and mediated by, the site. Visibility is predicated on volunteering information because the information provided in Tweets leads to exposure to others. Whilst in less than a decade this online behavior has become unremarkable to many, it is noteworthy to this research because it identifies new and evolving social norms in relation to the practice of friendship initiation.

Choosing who to follow

Another area that this research set out to explore was the reasons that users chose to follow some Tweeters. For Dolores, following was not a conscious process, she said: 'I really don't know how it happens, you just seem to drifttowards people somehow'. Similarly, Kate (37 years), a PhD candidate, and consultant from Adelaide, South Australia, stated: 'It just starts to snowball in an organic way'. Kate constructed following as a living thing, something 'organic', rendering it in common with nature. This understanding seems to belie the pre-designed, rational and calculated architecture of the site, which had facilitated this rhizomatic surveillant assemblage (Haggerty and Ericson 2000). The surveillant assemblage is when a broad and usually unknown audience can access an individual's online persona or 'data double', which constitutes fragmented information about them. This is because the design of Twitter has been set up to enable a potentially infinite audience to easily follow others online.

Other users seemed to follow particular Tweeters based on perceived commonality. This was consistent with social penetration theory because in co-present friendship individuals tend to make friends with people who have similar needs, attitudes, and circumstances (Altman and Taylor 1973; Friedman 1993: 202). As exemplified by Alison (28 years), a medical physicist from Sydney:

What drew me to her was her comedy of life. She often rants about stupid things that happen to her car, or you know, it's just very funny. She's very funny.

Alison's comment demonstrated that humour was a personality characteristic that drew her to the Tweeter (Friedman 1993: 135-7, 190-1; Rawlins 2009: 183). Kirsten also mentioned humour and politics as being reasons to follow a Tweeter. Twitter users were heterogeneous and interviewees had different motivations for following, however, deciding to follow a Tweeter with whom they perceived a commonality was customary practice.

Being followed

The ways that social networking site users manage their audiences have been previously examined (e.g. Marwick and boyd 2010). The main purpose of this paper is to explore the ways that Twitter users developed friendships that were initiated from a platform of online surveillance. However, as we acknowledge that managing followers is an aspect of Twitter use, we briefly discuss this here. Users seemed to have a high degree of awareness that they were exposed in a 'synoptic' surveillance relationship being surveilled by a potentially large audience of unknown others (Ball 2009; Mathiesen 1997). Users understood that their audience could include friends, family and potential employers, which seemed to influence their need to be cautious. Congruent with findings from Marwick and boyd (2010) users managed their followers in a range of ways. Anonymity was one tactic, for instance, Kirsten and Tracey used an anonymous account and did not provide identifying details. Kirsten and Winston (57 years), a retired business consultant from Newcastle, New South Wales, were cautious about what content they posted and omitted potentially inflammatory remarks. Alison did not reply or limited her interaction with unknown others to maintain social distance from them. Tracey deleted messages after they had been posted:

Some of the things I've said on Twitter, or at least on DM [Direct Message] ... [I've thought] the next day: 'Ooh I might delete that'.

Whilst deleting messages may have made Tracey feel better, it was likely an ineffective strategy. This is because messages can be 're-Tweeted' or re-sent either fully or partially, which makes it difficult for the user to keep track of content and their audience (Marwick and boyd 2010). It is unclear whether users had learnt how to manage followers from peers, or were simply using common sense. However, taken in general the tactics employed online were those of avoidance and were similar to strategies used offline- such as hedging a response to a verbal invitation-to discourage the development of friendship (Cupach and Metts 1994: 7).

From 'studied following' to 'reciprocal following'

Friendship initiation is two-way because individuals evaluate each other to assess their future compatibility (Altman and Taylor 1973: 36). Studied following seemed to facilitate the evaluation process online, rather than it occurring face-to-face, as with traditional acquainting. Alison stated:

You get to know much more about a person's interests and what they like, and that's when you get to start talking [on Twitter] and you're both like: 'Ahhhh we both love this stupid little thing that no one else loves, so maybe we should meet up and talk about it over beer'. And then [...] you do, and generally it works.

Studied following was a preliminary to friendship initiation because it operated as a tool for Alison to glean information 'about a person's interests and what they are like' as a prelude to commence interaction online. Just as individuals evaluate another person when they meet them face-to-face, the architecture of Twitter had enabled Alison to undertake acquainting online (Altman and Taylor 1973). We suggest that when online interaction becomes mutual, the mode of following moved from studied to reciprocal. Reciprocal following related to the 'sharing' aspect of participatory surveillance, because both users revealed their interests and identities with each other and this was characterised by ongoing interaction (Albrechtslund 2008: 7; Marwick 2012: 379). Sharing information was contributory to a positive evaluation of trust that demonstrated a move towards relationship equality, in contrast to the asymmetry of studied following. In Alison's case, following on Twitter was a form of social surveillance that facilitated the rational honing of friendship choice, because it gave her some minimal details about another Tweeter prior to commencing contact online or instigating a face-to-face meeting.

Following to friendship

Commonality can dispose a user to get to know a Tweeter more, however, friendship initiation and maintenance is also dependent on structural factors of geographic proximity and on having free time to enact social relations (Aristotle 2002; Little 1993; Montaigne 1991; Pahl 2000; Spencer and Pahl 2006: 99-100; Willmott 1987: 21-2). Geographic proximity was a factor that enabled Tracey's friendships to transition from on- to offline:

Well, [...] when we discovered that we work close to each other, or lived close to each other, we were like 'Oh! We should meet up for coffee!'[...] Yeah, so it's been geographical proximation of people that I was already relatively chatty with. So kind of like, it hasn't been I've met them on Twitter one day and then the next day let's meet up for coffee. I had already established friendships with them already.

Tracey had been 'relatively chatty' with some people for a substantial period of time during reciprocal following. This pre-friendship phase was sufficient to generate some online rapport and mutual trust that made Tracey feel comfortable enough to meet in person. The period of online following also demonstrated her loyalty and commitment to them as a friend (Friedman 1993: 194; Rawlins 1992; 9-10). Indeed, the commitment and dedication that was required of studied and reciprocal following is a reminder that whilst friendship is voluntary, it does involve an investment of time and energy. When Twitter mediated friendship, this work also involved constant monitoring and emotional performance. Geographic identifies new and evolving social norms in relation to the practice of friendship initiation. Choosing who to follow Another area that this research set out to explore was the reasons that users chose to follow some Tweeters. For Dolores, following was not a conscious process, she said: 'I really don't know how it happens, you just seem to drifttowards people somehow'. Similarly, Kate (37 years), a PhD candidate, and consultant from Adelaide, South Australia, stated: 'It just starts to snowball in an organic way'. Kate constructed following as a living thing, something 'organic', rendering it in common with nature. This understanding seems to belie the pre-designed, rational and calculated architecture of the site, which had facilitated this rhizomatic surveillant assemblage (Haggerty and Ericson 2000). The surveillant assemblage is when a broad and usually unknown audience can access an individual's online persona or 'data double', which constitutes fragmented information about them. This is because the design of Twitter has been set up to enable a potentially infinite audience to easily follow others online.

Other users seemed to follow particular Tweeters based on perceived commonality. This was consistent with social penetration theory because in co-present friendship individuals tend to make friends with people who have similar needs, attitudes, and circumstances (Altman and Taylor 1973; Friedman 1993: 202). As exemplified by Alison (28 years), a medical physicist from Sydney:

What drew me to her was her comedy of life. She often rants about stupid things that happen to her car, or you know, it's just very funny. She's very funny.

Alison's comment demonstrated that humour was a personality characteristic that drew her to the Tweeter (Friedman 1993: 135-7, 190-1; Rawlins 2009: 183). Kirsten also mentioned humour and politics as being reasons to follow a Tweeter. Twitter users were heterogeneous and interviewees had different motivations for following, however, deciding to follow a Tweeter with whom they perceived a commonality was customary practice.

Being followed

The ways that social networking site users manage their audiences have been previously examined (e.g. Marwick and boyd 2010). The main purpose of this paper is to explore the ways that Twitter users developed friendships that were initiated from a platform of online surveillance. However, as we acknowledge that managing followers is an aspect of Twitter use, we briefly discuss this here. Users seemed to have a high degree of awareness that they were exposed in a 'synoptic' surveillance relationship being surveilled by a potentially large audience of unknown others (Ball 2009; Mathiesen 1997). Users understood that their audience could include friends, family and potential employers, which seemed to influence their need to be cautious. Congruent with findings from Marwick and boyd (2010) users managed their followers in a range of ways. Anonymity was one tactic, for instance, Kirsten and Tracey used an anonymous account and did not provide identifying details. Kirsten and Winston (57 years), a retired business consultant from Newcastle, New South Wales, were cautious about what content they posted and omitted potentially inflammatory remarks. Alison did not reply or limited her interaction with unknown others to maintain social distance from them. Tracey deleted messages after they had been posted:

Some of the things I've said on Twitter, or at least on DM [Direct Message] ... [I've thought] the next day: 'Ooh I might delete that'.

Whilst deleting messages may have made Tracey feel better, it was likely an ineffective strategy. This is because messages can be 're-Tweeted' or re-sent either fully or partially, which makes it difficult for the user to keep track of content and their audience (Marwick and boyd 2010). It is unclear whether users had learnt how to manage followers from peers, or were simply using common sense. However, taken in general the tactics employed online were those of avoidance and were similar to strategies used offline- such as hedging a response to a verbal invitation-to discourage the development of friendship (Cupach and Metts 1994: 7).

From 'studied following' to 'reciprocal following'

Friendship initiation is two-way because individuals evaluate each other to assess their future compatibility (Altman and Taylor 1973: 36). Studied following seemed to facilitate the evaluation process online, rather than it occurring face-to-face, as with traditional acquainting. Alison stated:

You get to know much more about a person's interests and what they like, and that's when you get to start talking [on Twitter] and you're both like: 'Ahhhh we both love this stupid little thing that no one else loves, so maybe we should meet up and talk about it over beer'. And then [...] you do, and generally it works.

Studied following was a preliminary to friendship initiation because it operated as a tool for Alison to glean information 'about a person's interests and what they are like' as a prelude to commence interaction online. Just as individuals evaluate another person when they meet them face-to-face, the architecture of Twitter had enabled Alison to undertake acquainting online (Altman and Taylor 1973). We suggest that when online interaction becomes mutual, the mode of following moved from studied to reciprocal. Reciprocal following related to the 'sharing' aspect of participatory surveillance, because both users revealed their interests and identities with each other and this was characterised by ongoing interaction (Albrechtslund 2008: 7; Marwick 2012: 379). Sharing information was contributory to a positive evaluation of trust that demonstrated a move towards relationship equality, in contrast to the asymmetry of studied following. In Alison's case, following on Twitter was a form of social surveillance that facilitated the rational honing of friendship choice, because it gave her some minimal details about another Tweeter prior to commencing contact online or instigating a face-to-face meeting.

Following to friendship

Commonality can dispose a user to get to know a Tweeter more, however, friendship initiation and maintenance is also dependent on structural factors of geographic proximity and on having free time to enact social relations (Aristotle 2002; Little 1993; Montaigne 1991; Pahl 2000; Spencer and Pahl 2006: 99-100; Willmott 1987: 21-2). Geographic proximity was a factor that enabled Tracey's friendships to transition from on- to offline:

Well, [...] when we discovered that we work close to each other, or lived close to each other, we were like 'Oh! We should meet up for coffee!'[...] Yeah, so it's been geographical proximation of people that I was already relatively chatty with. So kind of like, it hasn't been I've met them on Twitter one day and then the next day let's meet up for coffee. I had already established friendships with them already.

Tracey had been 'relatively chatty' with some people for a substantial period of time during reciprocal following. This pre-friendship phase was sufficient to generate some online rapport and mutual trust that made Tracey feel comfortable enough to meet in person. The period of online following also demonstrated her loyalty and commitment to them as a friend (Friedman 1993: 194; Rawlins 1992; 9-10). Indeed, the commitment and dedication that was required of studied and reciprocal following is a reminder that whilst friendship is voluntary, it does involve an investment of time and energy. When Twitter mediated friendship, this work also involved constant monitoring and emotional performance. Geographic proximity facilitated co-present friendship because being in the same physical place and sharing times together during the beginning stage of friendship is essential to building solidarity and generating trust (Little 1993; Pahl 2000: 121).

Whilst similarity created affinity and added to relationship trust by degrees, some interviewees mentioned that they had chosen not to follow others based on a perception of difference. For example Kirsten said:

Well, you end up sort of not following anyone who is not like-minded I guess. [...] I'm sure there are a lot of racist redneck yobbos on Twitter, but I just don't follow them so I don't have anything to do with them. You don't sort of have to see them.

Kirsten followed Tweeters whose views reinforced and validated her offline identity and interests, and that she considered 'like-minded', a quality that indicates a deeper or more intimate level of friendship connection (Aristotle 2002; Little 1993). Kirsten filtered Tweeters that conflicted with her social and political values and beliefs, creating distance from those that she considered to be 'yobbos'. In doing so, Kirsten made a specific category of people unvisible, (see also Marks, this issue), as in her words: 'You don't sort of have to see them' [emphasis added].

After the first face-to-face meeting, the user needed to negotiate other aspects of friendship, which were deciding whether to continue the friendship, and how to maintain it. When the user first met the Tweeter face-to-face they already had expectations about them based on their interaction on Twitter and the next phase of friendship involved assessing if the ideal friend that they had imagined matched reality (Rawlins 1992: 11). Alison stated:

I don't think I've met anyone on Twitter where I've just really gone 'You are not what I expected'.

Alison's use of Twitter to screen potential friends had enabled her to securitise friendship to an extent, because she was able to minimise some of the risks associated with offline acquainting. The ongoing visibility of other Tweeters made it possible to develop trusting relationships over time. On Twitter, because online contact came first, exchanging additional details like a phone number and deciding whether or not to continue to meet were new stages that needed to be negotiated. Alison said:

It kind of goes from, you know, 'Oh let's meet up and have a beer' or whatever, and then, you know, there's that delicate balance, that dance around, you know, phone numbers. We're actually friends now, we're real friends, you know, how does this work? [...] To keep that up to start with I think there's a bit more of an interaction on Twitter. [...] I'd probably focus more on somebody if I'd met them than other people. [...] Generally if I meet someone I will interact with them a bit stronger.

Usually at the first face-to-face meeting individuals exchange contact details so they can stay in touch. Alison was unclear about appropriate social norms in this context, as was evidenced by her question: 'How does this work?' Her description of 'that dance' was reminiscent of the uncertainty and excitement sometimes experienced at the beginning of a romantic relationship, when both participants attempt to ascertain interest without losing face (Goffman 1969; Sternberg 1998: 72). Winston suggested a similar analogy: 'It's like a courtship ritual [laughs] but without the love interest part'. Moving from online correspondence to offline encounters was a process that required people to place a degree of trust in Tweeters with whom they wished to develop a friendship. Like a new romantic relationship, friendship initiation was a 'delicate balance' between emotional distance or independence and closeness or interdependence (Cupach and Metts 1994: 97; Rawlins 1992; Tannen 1986).

Once friends have decided to continue the friendship there is a transition from the initiation to the maintenance stage, whereby individuals undertake small rituals of supportive interchange to keep the friendship going (Goffman 1971). Alison indicated that she made an extra effort to: 'interact with them a bit stronger' after meeting face-to-face, by using a form of phatic communication to iterate their connection on Twitter (Miller 2008). This supportive interaction online sent a positive signal that reinforced their relationship.

For Kirsten, the movement from on- to offline friendship resulted in a deeper level of friendship than simply online:

In a way, yeah, because you can read more into them. Because during coffee they'll have told you about their miserable life or something [laugh] and so you'll see that this cheerful facade is not what it is, [...] or you'll understand what they mean when they say something that is quite oblique or a bit sort of cryptic. And then you'll say 'Oh no, I know what she means by that' and you'd never have known if you hadn't met them in person. Because you don't have the same conversation obviously as face to face.

Telling a friend a secret is a way to make them a member of a smaller or more exclusive social group (Goffman 1971: 142, 158). When the Tweeter confided about her 'miserable life' she showed willingness to trust that Alison would not expose or betray her personal details on Twitter (Friedman 1993: 130; Rawlins 1992: 22). Kirsten's comment highlighted that Twitter users have online personas that may contradict the realities of their life circumstances or that their on- and off-line personalities may be incongruent. Kirsten colluded with the Tweeter's online personality, by not challenging it or offering information on Twitter that would be viewed as discrepant, which is an example of relationship trust (Goffman 1971).

Reciprocal following as an antidote to difficulties meeting new co-present friends

Meeting people that are available for friendship can be a challenge as an adult, for a range of reasons, including: lack of access to a pool of potential friends; social confidence and social skills; cultural or language difficulties, or stage of life, for example, being retired and hence no longer interacting with work colleagues who provide social interaction. In co-present life, formal social rituals such as birthday celebrations, or informal gatherings like after-work drinks or 'parties', exist to provide excuses, no matter how tenuous, to meet socially and to invite others along (Altman and Taylor 1973: 10; Goffman 1966: 78). As an adult, approaching a stranger solely on the basis of perceived affinity is an odd, insufficient and likely unsuccessful approach to friendship initiation, because it attempts to skip the fundamental rapport building that necessarily generates trust before face-to-face relations can proceed and deepen. For both Tracey and Winston, Twitter offered a small but significant window into a social world that was otherwise inaccessible. Tracey explained that Twitter provided her with a pool of potential friends from which to draw:

I am Muslim so I don't go to the pub [...] There are avenues that I won't go to which would be the social avenues of some types of people. So I wouldn't have a lot of social opportunities to meet them, so I can meet them through Twitter.

The Australian drinking culture posed a cultural and religious barrier to Tracey participating in social life because she was excluded from events where alcohol was served and consumed (Falahey 2001: 137, 145). This made it hard for her to interact with people offline simply because of the lack of opportunity. Twitter offered Tracey a virtual meeting place for a broader range of people than were otherwise available to her offline, so it was a catalyst for the development of co-present friendships.

Assessing the worth and /or risk of Twitter friends

Andrejevic (2005: 489) has highlighted that online surveillance can be a tool to facilitate 'screening' of individuals in order to ascertain their potential 'risk'. It is acknowledged that as with any social relationship, meeting a Tweeter for the first time could be potentially risky. We have demonstrated that users would follow and reciprocally interact with another Tweeter for some time in the pre-friendship phase to help establish expectations and trust. Taking time to get to know a Tweeter online was a key component in the assessment of their character, and hence whether or not they were trustworthy, and this was a strategy to minimize the potential for negative outcomes should interaction proceed offline. Winston said:

I think if you've been chatting with people in a hundred and forty characters over a 10 month period, I guess you kind of learn what people are like. In the same way as ... [going] down to a coffee shop ... and overhearing conversations. I think you'd pretty quickly ... work out the people that were nice. And by nice, I mean the same as me.

Winston compared using Twitter to eavesdropping in a coffee shop because it enabled him to 'learn what people are like' over time. Whilst 'chatting' online, Winston evaluated whether he could trust that people were authentic and that their personalities were constant (Friedman 1993: 205; Goffman 1969: 180). Winston was drawn to people that he perceived to be similar, because they were 'nice', meaning (for Winston) people who were 'the same as me'. While other users were not so explicit in their narratives, often their preferences for similarity suggested that a trusted friend was one who was perceived to be as near to the user as possible-in terms of character, humour, politics, hobbies and interests. Twitter was a tool that broadened the net of potential friends that users could view, whilst also providing a means by which their opinions could be disregarded, or filtered and which hence securitised potential friendships. In this respect, Twitter enabled friendship initiation to be technologically augmented. The extent to which a Tweeter consistently presented as nice related to how Winston understood trust. He said:

Trust is something that's earned [emphasized by participant]. And through your actions and your behaviour, and which includes therefore the things that you might say on Twitter and re-Tweet, all that, all of those things build up an image of what you stand for.

Winston's narrative articulated the sentiment that over a period of studied and reciprocal following he ascertained character consistency and appropriate behavior according to the visibility of the Tweeter as presented in their Tweets. When these qualities were present, trust and friendship were forthcoming (Rawlins 2009: 209).

Another way to gain trust about a Tweeter was to garner verification of their authenticity from an online associate, a bit like an informal reference. Alison described:

Another guy I started following ... I met one on one. He's a Melbourne guy and he actually just came and stayed with me the first time I met him. But ... he was vouched for by someone else who is a Twitter friend of mine.

There were some risks to Alison in this situation. As a single woman inviting an unknown man into her home, Alison potentially put her property and person at risk of harm. Next, negotiating a co-present crosssex heterosexual friendship can be complex because there is the potential for intentions-that is, for sexual intimacy (or not)-to be confused (Goffman 1969: 205; Rawlins 2009: 119). Alison trusted her judgment that the on- and offline personalities of this man would be congruent, so she extended her benevolence, and a large degree of trust to him. She showed willingness to move their relationships from the online sphere to off-line, which can be challenging because it raises privacy issues. However, Alison had been following the Tweeter for 'a very long time' during which she had been able to ascertain his online personality and credibility from his visibility on Twitter. They had generated rapport, solidarity, loyalty and trust which was a pre-friendship phase and a period of shared history online that cemented the trust of the friendship (Rawlins 1992: 12). In keeping with the rhizomatic spread of the surveillant assemblage, overlaps in Twitter communities have been identified in previous research, and in this example Alison already had a connection with the Tweeter, no matter how tenuous, via her Twitter network because he was 'vouched for by someone else who was a Twitter friend' (Gruzd et al. 2010). When friendship is part of a larger community of others, this strengthens trust between them, because accountability to a group is greater than to one person. The surveillant assemblage operated in ways that permitted individual users to mutually co-construct trust. We define the practice of 'vouching' for someone online as 'transferable trust', which is a trust by association that can originate on- or offline, but which is manifest in and mediated by the social surveillance processes occurring in the Twitter realm. Transferable trust requires an honesty obligation as it necessitates that the knowing Tweeter will reveal if there is a reason not to trust someone.

Conclusion

This article has highlighted an important but previously unnoticed aspect of Twitter, which is the ways that everyday lateral surveillance on the site can operate as a vehicle to friendship initiation. Despite a small sample and exploratory findings, we have shown that following on Twitter can be understood as a nuanced practice that encompassed a range of friendship rituals that both mimic and differ from co-present friendship initiation rituals as per Altman and Taylor's (1973) social penetration theory, whereby individuals find out about and evaluate each other when they first meet. Indeed, there are some novel differences between acquainting on Twitter and acquainting offline. Studied following was a unidirectional relationship without the reciprocity characterised by offline friendship. As the visibility of Tweets enabled users to follow Tweeters without interacting with them, this meant that users could assess their character without the risk of friendship rejection that can occur offline. This unidirectional surveillance involved deliberately overseeing another Tweeter, but lacked any interaction with them beyond viewing their Tweets, hence Twitter enabled friendship screening to occur. Nonetheless, studied following involved a commitment of time and loyalty. Studied following could generate affinity and low levels of one-sided trust that enabled the user to evaluate the character of the Tweeter via their ongoing visibility so it technologically augmented rituals of acquainting. Affinity is characteristic in offline acquainting and during the course of studied following the user could feel affinity for the Tweeter based on their Tweets despite a lack of interaction. We highlighted that Twitter has enabled mediated surveillance and hence normalised the practice of deliberately observing another person and finding out about them in a way that is not acceptable in the offline world.

Sometimes studied following progressed to reciprocal following, where interaction was shared online. For these participants this process of sharing interests via direct messages online generated mutual trust, rapport, and expectations about character and personality, prior to an offline meeting. Both studied and reciprocal following facilitated an important pre-friendship phase that is hard to achieve or replicate when initiating friendship offline. Twitter gave the user a wider pool of people to meet with a view for potential friendship than offline, and these friendships were often initiated both on and offline based on perceptions of similarity. This was particularly important for Tracey, as cultural barriers around alcohol limited meeting new people offline.

Shared interests were a motivation for some users to meet offline, and geographic proximity made this easier for Kirsten, whilst Alison hosted a Tweeter who visited from inter-state. Meeting a Tweeter offline entailed the need to ascertain whether their online and offline personalities matched and if their similarities were sufficient to continue a friendship. Meeting face-to-face added a new dimension to the online friendship for Kirsten. Following a Tweeter for a long time meant that Winston could make an informed assessment of their character. For Alison, the risk of meeting a stranger offline was minimised because she had been following him online for a long period of time; in addition, he had been vouched for by another user, what we call 'transferable trust'. Reciprocal following enabled users to interact prior to meeting face-to-face, hence when a co-present meeting occurred this meant that they needed to negotiate whether or not to continue the friendship. This is slightly different to co-present friendship initiation where exchanging contact details and evaluating whether to continue the friendship usually occurs at the first co-present meeting. Not surprisingly then, this was a bit confusing for Alison, as she was uncertain about the social rules for this situation. As has been found in other research, Tweeters that generated posts that were incongruent with the user's values were not engaged with and were deleted from her or his list of followers. Users had various strategies to manage being followed by others such as maintaining anonymity and with-holding or deleting inappropriate comments.

In conclusion this research has explored how social surveillance via following on Twitter facilitated the process of making friends. This was because Twitter users were highly visible, and their presence and online behaviours generated, iterated and reiterated friendship qualities of reciprocity, loyalty, and trust. Put simply, the use of the social network site Twitter broadened the net of potential friends far wider than that offered in the offline world. Further, this type of Twitter use can lead to new ways of defining the phases of friendship and this article has highlighted a pre-friendship screening phase on Twitter that does not exist offline.

Acknowledgements

Thanks to two anonymous reviewers and Catriona Elder for their insightful comments on earlier versions of this paper. Special thanks to Gavin Smith for supporting this article.

References

References

Albrechtslund, A. 2008. Online social networking as participatory surveillance. First Monday 13(3). http://firstmonday.org/htbin/cgiwrap/bin/ojs/index.php/fm/article/view/2142/1949

Allan, G. 2008. Flexibility, friendship, and family. Personal Relationships 15(1): 1-16.

Altman I. and D. Taylor. 1973. Social Penetration: The Development of Social Relationships. America: Holt, Rinehart and Winston, Inc.

Andrejevic, M. 2005. The work of watching one another: Lateral surveillance, risk, and governance. Surveillance & Society 2(4): 479-497.

Andrews, L. 2012. I Know Who You Are and I Saw What You Did: Social Networks and the Death of Privacy. US: Free Press.

Aristotle. 2002. Nicomachean Ethics. Oxford: Oxford University Press.

Ball, K. 2009. Exposure. Information, Communication and Society 12(5): 639-657.

Bauman, Z. and D. Lyon. 2013. Liquid Surveillance: A conversation. Oxford: Wiley.

Blieszner, R. and R. Adams. 1992. Adult Friendship. California: Sage.

Boase J. and B. Wellman. 2006. Personal relationships: On and offthe internet. In: Cambridge handbook of personal relationships, eds A.L. Vangelisti and D. Perlman, 709-723. Cambridge: Cambridge University Press.

boyd, d. 2011. Social network sites as networked publics: affordances, dynamics, and implications. In: A Networked Self: Identity, Community, and Culture on Social Network Sites, ed. Z. Papacharissi, 39-58. New York and London: Routledge.

boyd, d. and N. Ellison. 2007. Social network sites: definition, history and scholarship. Journal of Computer-Mediated Communication 13(1): 210-230.

Bryman, A. 2008. Social Research Methods. Oxford University Press: Oxford.

Bulmer, M. 1986. Neighbours: The Work of Philip Abrams. Cambridge University Press: Cambridge.

Clarke, N. 2005. Detailing transnational lives of the middle: British working holiday makers in Australia. Journal of Ethnic and Migration Studies 31(2): 307-322.

Coll, S.O. Glassey and C. Balleys. 2011. Building social networks ethics beyond 'privacy': a sociological perspective. International Review of Information Ethics 16: 47-53.

Cohen, J. 2012. Configuring the Networked Self: Law, Code, and the Play of Everyday Practice. New Haven and London: Yale University Press.

Cupach, W. and S. Metts. 1994. Facework. California: Sage.

Duck, S. 1977. The Study Of Acquaintance. Farnborough: Saxon House.

Falahey, A. 2001. Resident Action Collectives as Social Movements? Investigation of oppositions to Alcohol Regulations and Drinking Practices in Sydney, Australia 1788-2001. PhD Thesis, UNSW.

Foucault, M. 1998. Technologies of the self. In: Technologies Of The Self: A Seminar With Michel Foucault, eds L.H. Martin, H. Gutman and P.H. Hutton, 16-49. Amherst: University of Massachusetts Press.

Friedman, M. 1993. What are friends for? Feminist Perspectives on Personal Relationships and Moral Theory. Ithaca, New York, and London: Cornell University Press.

Gandy, Jr. O.H. 1993. The panoptic sort: A political economy of personal information. Boulder: Westview.

Gandy, Jr. O.H. 2009. Coming to terms with chance: Engaging rationale discrimination and cumulative disadvantage. Farnham: Ashgate.

Gershon, I. 2011. Un-Friend My Heart: Facebook, Promiscuity, and Heartbreak in a Neoliberal Age. Anthropological Quarterly 84(4): 865-894.

Goffman, E. 1966. Encounters: two studies in the sociology of interaction. USA: The Bobbs-Merrill Company Inc.

Goffman, E. 1969. Where the Action Is: Three Essays. London: Allen Lane.

Goffman, E. 1971. Relations in Public: Micro-studies of the Public Order. London: Allen Lane.

Goffman, E. 2008. Interaction Ritual: Essays in Face to Face Behaviour. US: Interaction Publishers.

Gromm, R. 2004. Social Research Methodology: A Critical Introduction. Basingstoke: Palgrave Macmillan.

Gruzd, A., B. Wellman and Y. Takhteyev. 2010. Imagining Twitter as an imagined Community. American Behavioural Scientist 55(10): 1294-1318.

Haggerty, K.D. and R.V. Ericson. 2000. The surveillant assemblage. British Journal of Sociology 51(4): 605-622.

Horton, D. and R.R. Wohl. 1956. Parasocial Interaction: Observations on Intimacy at a Distance. Psychiatry 19: 215-29.

Lampe C., N. Ellison and C. Steinfield. 2006. A Face(book) in the Crowd: Social Searching vs. Social Browsing, Proceedings of CSCW-2006, 167-170. New York: ACM Press.

Liamputtong, P. 2011. Qualitative Research Methods. Oxford: Oxford University Press.

Little, G. 1993. Friendship: Being Ourselves with Others. Melbourne: The Text Publishing Company. Lyon, D. 2007. Surveillance Studies: An Overview. Cambridge: Polity.

Marwick, A. E. 2012. The public domain: social surveillance in everyday life. Surveillance & Society 9(4): 378-393.

Marwick, A. and d. boyd. 2010. I Tweet Honestly, I Tweet Passionately: Twitter Users, Context Collapse, and the Imagined Audience. New Media & Society 13(1): 113-144.

Mathiesen, T. 1997. The Viewer Society: Michel Foucault's 'Panopticon' Revisited. Theoretical Criminology 1(2): 215-234.

Mauss, M. 1954. The gift: forms and functions of exchange in archaic societies. Cohen and West: London.

Miller, D. 2011. Tales from Facebook. Cambridge: Polity.

Miller V. 2008. New Media, Networking and Phatic Culture, Convergence. The international Journal of Research into New Media Technologies 14(4): 387-400.

Montaigne, M. 1991. On Friendship. London: Penguin.

Murthy, D. 2013. Social Communication in the Twitter Age. Oxford: Wiley. E-book.

Nielsen. 2012. State of the Media: The Social Media Report, 2012. The Nielsen Company and NM Incite.

Neuman, L.W. 2006. Social Research Methods: Qualitative And Quantitative Approaches. Boston: Pearson International.

Newcomb, T. 1961. The Acquaintance Process. America: Holt, Rhinehart and Winston.

Pahl, R. 2000. On Friendship. Oxford: Polity.

Ramirez, A., J. Walther, J. Burgoon and M. Sunnafrank. 2002. Information-Seeking Strategies, Uncertainty, and Computer- Mediated Communication: Toward a Conceptual Model. Human Communication Research (28(2): 213-228.

Rawlins, W. K. 1992. Friendship Matters: Communication, Dialectics and the Life course. New York: Aldine De Gruyter.

Rawlins, W. K. 2009. The Compass of Friendship: Narratives, Identities and Dialogues. California: Sage.

Spencer, L. and R. Pahl. 2006. Rethinking Friendship Hidden Solidarities Today. New Jersey: Princeton University Press.

Sternberg, R. J. 1998. Cupid's Arrow: The course of love through time. Cambridge: Cambridge University Press.

Sysomos Inc. 2009. Inside Twitter: An In-Depth Look Inside the Twitter World, Sysomos.

Tannen, D. 1986. That's Not What I Meant! How conversational style makes or breaks relationships. New York: Ballantine Books.

Trottier, D. 2012. Social Media as Surveillance: Rethinking Visibility in a Converging World. Ashgate E-book.

Van Krieken, R. 2012. Celebrity Society. Routledge Ebook.

Wilkinson, J. 2010. Personal Communities: Responsible Individualism or Another Fall for Public [Man]? Sociology 44(3): 453- 470.

Willmott, P. 1987. Friendship Networks and Social Support. London: Policy Studies Institute, Report No. 666.

Zhao, S., S. Grasmuck and J. Martin. 2008. Identity construction on Facebook: Digital empowerment in anchored relationships. Computers in Human Behaviour 24(5): 1816-1836.

Zoppos, E. 2011. 'Augmented friendship' and new considerations in the age of social media. Paper presented to The Australian Sociological Association Conference, Newcastle, Australia, 29th November to 1st December.