Boundaryless Twitter Use: On the Affordances of Social Media

Abstract

For this study, we followed the director general (DG) of a large Swedish public authority on Twitter. We analyzed the data from Twitter and from interviews in terms of four affordances that distinguish social media from more traditional technologies: visibility, persistence, association and editability. We suggest that to understand social media affordances, it is necessary to consider the medium and the situation it creates and how this increases the range of possible interpretations. Therefore, we propose counterparts to the affordances of visibility, persistence, association and editability, in the form of invisibility, fluidity, dissociation and indeterminacy, to be included in an analysis of social media affordances and, as we argue, the creation of a persona through Twitter communication.

1. Introduction

Social media, such as Twitter, promise symmetric, two-way communication between organizations and their stakeholders (Etter 2014) by enabling more interactive and personalized communication, heightened intensity (Kelleher and Miller 2006), and few gatekeeping mechanisms.

Organizations can thus converse with stakeholders who ask questions and voice concerns (Greenberg 2010) and even address issues relating to trustworthiness (Coyle et al. 2012). However, research indicates that organizational communication through social media is still almost always one-way communication (James 2007; Kavanaugh et al. 2012; Macnamara 2010; Valentini 2015). Moreover, while arguing that social media can open up new communication pathways and engender a constant communication flow, research on social media remains, for lack of empirical examples and/or a focus on social media as platforms for spreading and accessing information, grounded in a conception of communication as transmission (Albu and Etter 2016). Accordingly, most research on social media underplays the way it enables and restricts communication and the role it plays in the constitution of identity.

From a sociomateriality perspective, however, social media is understood to play a performative role in creating identities. The affordances approach posits, for example, that social media afford opportunities for action and moves the focus away from the functional aspects of how technology is shaped. Instead, it emphasizes the role of technology in shaping social interaction, i.e., how social media technologies enable and constrain communication (Boyd 2010; Fayard and Weeks 2007; Gibson 1979; Treem and Leonardi 2012). “Sociomateriality research [then] typically focuses on the entanglement between humans and technology and the way affordances of social media technologies offer possibilities and constraints for communicative activities” (Albu and Etter 2016, p. 9).

Research on the affordances of social media, especially Twitter, has, among other things, dealt with the risks generated for organizations by the visibility, persistence and association of sensitive information that external parties may access and manipulate (Vaast and Kaganer 2013); the advantages and disadvantages of visible and persistent enterprise social media (Leonardi 2014; Leonardi et al. 2013); how the affordances of social media can have contradictory effects on knowledge sharing (Ellison et al. 2015; Majchrzak et al. 2013); and how the affordances of visibility, association, persistence, and editability allow for both open and closed approaches to knowledge sharing (Gibbs et al. 2013). We also see a quite significant stream of research on political Twitter, which investigates, for example, the effects of the increase in the number of characters allowed in a tweet on the climate of the discussion in the replies (Jaidka et al. 2019) or the dissent empowering affordances of Twitter used by rogue and alt government agency Twitter accounts (Oltmann et al. 2020). An increase in the interest is no surprise given the changes in the political landscape (increased polarization and election meddling through social media), and the assumed political nature of Twitter. Our interest, however, focusses on the boundaries between the organizational and the personal and how the particular situation on Twitter is shaped through the affordances of the technology.

The understanding of the affordances of social media remains mostly grounded in a conceptualization of affordances that privilege presence over absence, i.e., research describes affordances that are observable and ignores that even the non-observable aspects could be considered affordances. In effect, the concept of affordances points to an ability or to a state of being able to do something, which connotes the presence of that ability. The affordance of visibility relates, for example, to the ability to make behaviors, knowledge, preferences and connections visible (Treem and Leonardi 2012). The absence of that ability is in turn implicitly conceptualized as the absence of the affordance of visibility or simply as a reaction by the user to the constraining aspects of visibility, i.e., the user responds strategically to the affordance of visibility by “going invisible” (Gibbs et al. 2013). Such a response points, however, to another possibility, namely that invisibility itself can be conceived of as an affordance. Invisibility is therefore not an accident or by-product but an affordance of the technology.

To summarize, even though the affordances of social media technologies identified in the literature are usually conceptualized as both enabling and constraining, they all seem to privilege the presence of an ability, e.g., visibility, persistence, association, editability (Treem and Leonardi 2012); reviewability, recombinability (Faraj et al. 2011); scalability, searchability and replicability (Boyd 2010). From such a vantage point, the possibilities and constraints of social media affordances are therefore the positive and negative effects of a conception of affordances that glosses over absence and the possibilities and constraints that it might imply.

Thus, although research has revealed the enabling and constraining character of social media affordances, it has precluded the possibility of the “presence” of absence, i.e., what the absence of visibility might, for example, afford the user of social media. In other words, an important part of the performative role of social media technology has either been lessened or ignored because of the precedence, prevalence and privilege afforded to presence in our language and the way we tend to refer to “things”—in this case, social media and the opportunities for action they afford.

In what follows, we focus on Twitter as a social media technology that affords visibility, persistence, association and editability (Treem and Leonardi 2012), and identify opposite affordances and the possibilities and constraints they entail. We do this by studying how the director general (DG) of a Swedish public authority acts through and with Twitter to communicate with his/her followers and thus co-constitutes meaning. We conceptualize the observed reality as a synthetic situation, without a physical location but the constant possibility for destabilization and fluidity (Knorr Cetina 2009). In doing so, we aim to contribute to the literature on social media affordances by expanding the affordance perspective on social media, especially Twitter.

Our paper is structured as follows. We start with a discussion of the literature concerning the affordances of social media, followed by our theoretical framework. We then describe our method, present our results and discussion and end the paper with conclusions.

2. Social Media and Their Affordances

The affordance perspective was advanced by Gibson (1979) and is based on the idea that humans relate to inanimate objects differently, depending on their background, experiences and environment. Objects are thus perceived in terms of the opportunities for action that they afford. Affordances can vary significantly, depending on the situation and the person assessing them (Hutchby 2001). As Treem and Leonardi (2012, p. 146) put it: “materiality exists independently of people, but affordances do not”. Affordances are therefore not the features or physical attributes of a given technology; they develop through the interplay between humans and technologies.

A number of social media affordances have been advanced, such as visualizing entire work processes, real-time/flexible product and service innovation, virtual collaboration, mass collaboration, and simulation/synthetic reality (Zammuto et al. 2007); tagging, links and authoring (McAfee 2009); scalability, searchability and replicability (Boyd 2010); reviewability and recombinability (Faraj et al. 2011); and visibility, association, persistence and editability (Treem and Leonardi 2012). Bucher and Helmond (2018), when discussing Twitter, even mention hidden affordances, i.e., affordances of the technology that are not obvious to the end-user. In the case of Twitter, they mention the user and non-user tracking through, e.g., cookies and buttons across the web. In what follows, however, we concentrate on Treem and Leonardi’s (2012) four social media affordances, because their typology is based on a meta-review of studies of internal social media used by private companies. Moreover, most of the affordances presented by other researchers fall within, or at least overlap with, the four categories presented by Treem and Leonardi (2012), which makes them a particularly appropriate starting point for our analysis.

Treem and Leonardi (2012) identify four affordances on which social media rank consistently highly: visibility (ease of access to information), persistence (continued presence of information), association (link to people, organization, and information) and editability (information can be edited continuously). Other technologies used by organizations may also provide these affordances, but not all four of them consistently.

The affordance of visibility relates to how easily information and knowledge can be found in a network and how visible a person’s own network is. In effect, “social media afford users the ability to make their behaviors, knowledge, preferences, and communication network connections that were once invisible (or at least very hard to see) visible to others” (Treem and Leonardi 2012, p. 150). Gibbs et al. (2013) do suggest that there is an affordance of invisibility when the people in their study use the function of Skype to be shown as offline even though they are still working. They suggest it is a strategic use of the affordance of invisibility.

In general, this is applicable to public social media such as Facebook, Ello, Plague, Mind, Twitter, Instagram and LinkedIn, though in varying degrees. Twitter, the focus of this paper, is particularly conducive to making an individual’s social capital visible both quantitatively and qualitatively. The number of followers is easily observable and operates as a form of Twitter currency. The more followers an account has, the more valuable their mentions or retweets become. Subtle differences in the use of retweets, quotes and direct mentions enable people to draw conclusions about relationships between people and even their position in a debate or their political opinion. The knowledge gained about a person through social media is thus not limited to any one area of life, such as work, leisure interests or personal information; it includes all of the above (Treem and Leonardi 2012).

Twitter offers users the opportunity to share work-related behavior, meta-knowledge about themselves, and practices, such as to how to do things. All this is similar to an internal social media tool. Particularly for leaders, Twitter offers a tempting opportunity to combine branding the organization, working on their own career, and influencing their subordinates.

Persistence, the second affordance identified by Treem and Leonardi (2012, p. 155), is defined as follows: “communication is persistent if it remains accessible in the same form as the original display after the actor has finished his or her presentation.” Persistence has, moreover, two dimensions in public social media. First, once information is online it becomes everlasting. There is no taking back a badly formulated tweet. Of course, people sometimes delete tweets and try to erase the traces of possibly poor judgement. However, if the person is in any way of public interest, there is little hope that the deletion from their timeline will erase the tweet from the wider Twitter universe and/or the web. Once it is on the Internet, it stays there forever in one form or another, even without the tweeter’s consent, either as cached data, scrapped data or in the easily shared form of a screenshot. In the case of certain politicians, it has even become archival data, e.g., Barack Obama or Donald J. Trump. Secondly, the persistence of online communication also helps to transcend time boundaries, as tweets posted during working hours can be reacted to, shared and discussed when work stops. Karahanna et al. (2018) suggest that this is not an action affordance and therefore they exclude it from their study. We would agree but conceptualize the affordances differently within our theoretical framework outlined in Section 3.

Association, the third affordance, is defined as “establish[ing] connections between individuals, between individuals and content, or between an actor and a presentation” (Treem and Leonardi 2012, p. 162) and, as Vaast and Kaganer (2013) remark, between individual and organization. The public display of any kind of association becomes particularly significant on widely used social media like Twitter, for two reasons. Firstly, the world can see whom the person follows, what their interests are and who is following them. This creates a network of direct or indirect connections and associations. For example, by retweeting and following Twitter accounts that have a clear alt-right agenda and/or connection, the current American president places himself in an assumed ideological proximity. Secondly, all social media platforms suggest new connections using algorithms based on people’s existing connections, interests and likes. This becomes particularly consequential, for example, when one has a close association with one’s employer. Many suggestions will then be work-related, turning the personal Twitter account into an extension of work and thus the organization.

The final affordance is ‘editability,’ which refers to the ability to spend a good deal of effort and time crafting a message before it is seen by others (Treem and Leonardi 2012, p. 159). This is enabled by the medium’s asynchronicity. It is also possible to edit after posting, even though this will usually leave a trace of the original.

Most research on social media technologies focuses on their positive effects or enabling role (Ellison et al. 2011; Kaplan and Haenlein 2010; Shirky 2008). As Majchrzak et al. (2013) note, this is perhaps not surprising, given that the ubiquity of social media is comparatively new. An affordance perspective becomes all the more important in this context, insofar as it not only focuses on the positive impact and enabling character of social media, but also their constraining aspects and impact on organizations.

Analyzing the social media policies of a broad sample of organizations, Vaast and Kaganer (2013) conclude, for example, that organizations perceive ‘visibility’, ‘persistence’, and ‘association’ as generating risk. These affordances suggest that the organization worries about losing control over how, when, and where information is shared. Majchrzak et al. (2013) identify contradictory influences of social media affordances on knowledge sharing. They do so by pointing out tensions “that point to the paradox of media in-use: coexistence of opposite tendencies, unexpected consequences, and contradictory findings” (Majchrzak et al. 2013, p. 48f.), i.e., the tension generated by positive and negative aspects of social media affordances. For example, they write of one affordance they identify in their study:

The affordance of metavoicing provides one way in which knowledge workers can engage in the communal online knowledge conversation of the workplace. Metavoicing can foster productive knowledge conversations when the mechanism of critical mass is invoked. Simultaneously, though, metavoicing can inhibit the productivity of these knowledge conversations when they promote biased and inaccurate information.

(Majchrzak et al. 2013, p. 42)

Similarly, Gibbs et al. (2013) show how the affordances of visibility, association, persistence, and editability allow for both open and closed attitudes, while Leonardi et al. (2013, p. 16) present the advantages and disadvantages of enterprise social media by way of three metaphors that have guided perceptions of social media technologies, i.e., social media as a “leaky pipe”, an “echo chamber”, and a “social lubricant.” On the point at hand, they write that “the affordance view reveals that both positive and negative outcomes can result from the use of social media in the enterprise”.

While we agree with previous research that it is important to also study the constraining aspects of social media technologies, and that the affordance perspective contributes to this cause, we also find that the social media affordances identified in the literature tend to privilege presence over absence. For example, visibility, as we have already mentioned above, is presented as a social media affordance, while invisibility is conceptualized as a strategic response to the constraining aspects of the affordance of visibility (Gibbs et al. 2013). Instead, as Ellison and Vitak (2015, p. 110) suggest with regard to enterprise social network sites (ESNS), we believe social media technologies may allow for both visibility and invisibility “by [among other things] enabling users to traverse others’ information streams unobtrusively, such that they may become passive lurkers rather than active participants”. Although they do not expand on the notion, Ellison and Vitak (2015) offer the possibility of two affordances, i.e., visibility and invisibility, or at least for two sides of the same affordance. In what follows, we focus on this view on affordances. In doing so, we attempt not to privilege presence over absence but to include also the “presence” of absence by focusing on how affordances develop in and through the interplay between humans and technologies.

3. Theoretical Framework

Traditional research on social media technologies, such as Twitter, views them typically as diffusion technologies that enable easier access to information and perhaps even transparency (Etter 2014). The affordance perspective shifts the focus toward the enabling and constraining aspects of social media. From this vantage point, social media, in this case Twitter, create “new pathways of communication between individuals” (Albu and Etter 2016, p. 11). Thus, Twitter enables continuous communication, and can be used to enhance individual and organizational reputations, but the outcomes of its use are not necessarily foreseeable and may also be constraining. Notwithstanding its importance, such a view appears to take the character of affordances for granted, and to focus on their enabling and constraining outcomes rather than the affordances themselves and the possibility of alternative affordances that may incorporate the “presence” of absence, a view that is, in itself, both enabling and constraining.

As already mentioned, research has mostly focused on the presence of inherently positive affordances such as virtual collaboration, mass collaboration, and simulation/synthetic reality (Zammuto et al. 2007); tagging, links and authoring (McAfee 2009); scalability, searchability and replicability (Boyd 2010); and reviewability and recombinability (Faraj et al. 2011). In this article, we focus on alternative or opposing affordances by seeking to reinterpret the affordances of visibility, association, persistence and editability identified by Treem and Leonardi (2012). In doing so, we imagine such alternative affordances of Twitter as being enabling and constraining in themselves.

As Albu and Etter (2016, p. 11) remark, “instead of simply facilitating the dissemination of text created by organizational members, Twitter can store information in a fluid and constantly reconfiguring state (outside time, space, or relationships; Jackson 2007)”. This means that any message is potentially open for re-interpretation, contestation, destabilization and subversion. To that effect, the hashtag can serve as “a type of organizational text with dual capacity and open authorship (i.e., hypertexts) due to Twitter’s mix of affordances” (Albu and Etter 2016, p. 25). This, we argue, is potentially the case for every tweet, not only through hashtags. As Murthy (2012, p. 1067) expounds, retweeting re-embeds a tweet into the situational space of other Twitter users who might “feel the utterance to be originating from the retweeter.”

Embedding is a term borrowed from Goffman (1981), who argues that utterances have utterers, but that an utterance does not necessarily preserve the original utterer. The original tweeter may thus not necessarily be preserved in the utterance as it is shared, although it is customary to include the originator in retweets. More importantly, as they are sent, tweets become removed from the situated context in which they are crafted, to be embedded in a time dictated by the Twitter feed. The responses of others in the form of tweets and retweets temporalize, in other words, communication by embedding and re-embedding utterances into the Twitter feed, i.e., the situated present of recipients. Twitter discussions take place in a virtual place, driven by the succession of tweets and retweets and what Knorr Cetina (2009), inspired by Goffman (1981), refers to as “response presence”.

Put differently, Twitter presents us with what Knorr Cetina (2009, p. 69) refers to as a synthetic situation, i.e., “an environment augmented (and temporalized) by fully or partially scoped components—in which we find ourselves in one another’s and the scopic components’ response presence, without needing to be in one another’s physical presence”. Knorr Cetina (2009) based the concept on observations of stock market traders and how the situation was constituted by onscreen information and the scopic role played by computer screens as instruments for seeing and observing. With Twitter, the situation in which the succession of tweets takes place is virtual, it is not situated in the same time and place, as face-to-face interaction is. Instead, tweets are sent to followers so that the virtual place is concomitantly one and a multitude of places. Followers may in turn respond and/or retweet the tweet in what Murthy (2012), building on Knorr Cetina’s (2009) synthetic situation, calls synthetic embedding. “Synthetic embedding, like physical embedding, reformulates the place it is embedded in, which [in the case of a succession of tweets] happens to be a virtual place” Murthy (2012, p. 1068). That is to say that the direction in which meaning is constructed on Twitter depends not on the situated place of interaction, but on the “response presence” that constitutes the succession of tweets and retweets.

In what follows, we build on Knorr Cetina (2009) and Murthy (2012) to produce a theoretical base for our understanding of communication mediated by and through Twitter as a social media technology that affords visibility, association, persistence and editability, but does so in and through a synthetic situation.

Synthetic situations have three prominent features (Knorr Cetina 2009) that are of particular importance for our analysis. First, a synthetic situation is a composite of information. In the case of Twitter, the situation is in effect a Twitter feed, an assemblage of tweets and retweets unique to each account holder. What is on screen is thus the situation that account holders find themselves in. As Knorr Cetina (2009, p. 69) points out, such a situation lacks the cohesive “feel of a taken-for-granted material world that has emerged over time.” In other words, the synthetic situation that Twitter creates is different; it is informational in character and is not something we should take for granted in our analysis of Twitter interactions.

Second, synthetic situations are temporal rather than spatial. Twitter reality is continuously updated. The user may scroll down to see older tweets, but Twitter’s present, i.e., the latest tweet, is at the forefront, marking constant novelty. The intensity of this fluidity may of course vary, depending on the tweeters in a specific circle, but the fluidity is marked, nonetheless. Instead of the spatial feel of face-to-face interactions, the “synthetic situation […] is inherently in flux; it has none of the durability of a physical situation” (Knorr Cetina 2009, p. 72). The Twitter feed may not be as fluid or volatile as the reflection of the market on screens, in the example provided by Knorr Cetina (2009), but it is still a feed that one has to scroll down to look for what has been and, more importantly, the number of tweets continually increases, and thus is also in constant change. As Jackson (2007, p. 409) puts it: “information does not so much travel from one place to another, as much as it exists in a fluid state, able to flow freely and to be appropriated simultaneously in multiple and undetermined ways.” With every tweet, new elements are made available to consider and to appropriate. Similarly, as Knorr Cetina (2009, p. 72) writes:

The screen reality is a process, but it is not simply like a river flowing from one location to another as an identical mass of water. Rather, it is processual in the sense of an infinite succession of nonidentical matter projecting itself forward as a changing situation.

Third, “synthetic situation features may become symbolic interaction partners for participants” (Knorr Cetina 2009, p. 72). In effect, it is the screen that mediates the interaction between the tweeter and the follower. Specifically, it is first and foremost the tweet on the screen, 140 characters (280 as of September 2017), the informational character of the words and sentences comprised of a limited number of characters with which the follower interacts in the first place. As Knorr Cetina (2009, p. 73) observes: “traders do not engage particular persons but simply hit on a price by typing the instruction on their machine.” Similarly, it is the tweet not the tweeter that followers react to and interact with. The screen indeed projects an Other, but it is an Other formed in a synthetic situation, a synthetic Other. An obvious characteristic of the synthetic situation presented by Twitter is, as Murthy (2011) relates, terseness. The brevity of tweets, which are limited to 140/280 characters, shapes how Twitter is used, and therefore how audiences are constituted by and through the enabling and constraining aspects of the medium. In turn, the informational character of Twitter communication implies promiscuity, in the sense that, as Jackson (2007, p. 411) puts it, “it doesn’t matter whether or not the information was ‘meant’ to be used in a particular way, whether the information is accurate, or even whether it is comprehensible. Databases make content promiscuous by creating the potential for indiscriminate association.” The Twitter feed functions in this manner by juxtaposing tweets from people you follow along a timeline that indiscriminately presents them in what may appear to be a coherent whole (Jackson 2007).

The synthetic situation thus entails a context of interaction where tweets are embedded and re-embedded, and where the meaning of tweets is always open to re-interpretation, contestation, destabilization and subversion by virtue of its virtual time and place. In light of this, we argue that the affordances of Twitter need to be conceptualized to make room for potential re-interpretation, contestation, destabilization and subversion. To do so, we turn to our empirical example, the tweets of the director general (DG) of a Swedish public authority. We join Costa (2018) in considering affordances not as stable but as affordances-in-practice that will change according to context.

4. Method

While research on the use of social media by organizations has focused on the organization as the sender of carefully tailored messages on social media (Albu and Etter 2016; James 2007; Macnamara 2010), we focus on an individual, the Director General of a Swedish public authority, for two reasons.

Firstly, the leader of a public authority is both a private and a public person who represents an organization and her/himself on Twitter. Social media, Twitter in this case, are in turn particularly conducive to the renegotiation of boundaries, insofar as they are a tool to create, shape, maintain and express one or multiple identities (Ollier-Malaterre et al. 2013) that transcend the boundaries between the individual and the organization (Siegert 2015). With their specific affordances of visibility, persistence, association and editability (Treem and Leonardi 2012), they do not just allow for a sophisticated expression of an individual’s identity(ies), they also allow organizations to profit from that expression through association. A top manager’s tweets or posts are thus problematic from the very outset, insofar as they are simultaneously expressed by her/him personally as a non-anonymous private citizen and by him/her as representative of the organization. The difficulty of separating the two provides us with an interesting opportunity to reinterpret the affordances of visibility, persistence, association and editability.

Secondly, choosing a single, identifiable person, the top manager of an organization, also makes it easier to interpret identities, because s/he is a public figure. This is interesting, because research on the external use of social media technologies traditionally focuses on organizational use of social media, that is on the organization as the sender of carefully tailored messages (James 2007; Kavanaugh et al. 2012; Macnamara 2010), and not on tweets and posts by individuals who, as it were, concomitantly represent themselves and the organization they work for in their social media accounts.

Data Collection and Analysis

We collected the primary data between 2015 and 2017. The tweets, 3199 in all, date from 5 November 2015 to 8 May 2017. We also interviewed the DG three times during that period, although not exclusively about social media and Twitter, along with other managers and employees, and we used background material such as guidelines and other documents concerning the social media presence of public sector employees.

We analyzed the Twitter data by reading each tweet multiple times, following the DG’s Twitter discussions, and subsequently coding each tweet as related to work, personal life, or current affairs. We iteratively went back and forth between the data and the theory to establish a link between them. During this phase of the analysis, we coded each tweet as mostly indicative of one or, in rare cases, a combination of the affordances. Once we identified a pattern, we chose specific tweets, which we translated into English, to illustrate the affordances and their reinterpretations.

In our analysis, we found that most of the DG’s Twitter behavior was related to three of the four affordances presented by Treem and Leonardi (2012). The next step involved a [re]interpretation of the affordances of Twitter in terms of the synthetic situation it presents, i.e., its informational character, fluidity, and propensity to become an interaction partner.

In what follows, we attempt to show how tweets are not only both private and public, as is clear from our choice of empirical example, but also that, given the synthetic situation they involve and are involved in, they afford visibility and invisibility, persistence and flux, association and dissociation as well as editability (understood as control over or ability to determine the message) and indeterminacy.

5. Revisiting the Affordances of Twitter

The Director General (DG) leads a large Swedish public authority with offices all over the country. In what follows, we focus on the Twitter behavior of the Director General (DG) in terms of the affordances of visibility, association, persistence and editability.

5.1. The Informational Character of Visibility

A significant reason for being on social media is the opportunity to connect with not only friends and acquaintances, but also strangers through and because of shared interests. In the DG’s case, the ratio is about 1:2, with 2066 people followed and 4907 followers (May 2017). If the DG is striving to influence, s/he would be hoping for many followers and certainly more followers than people followed.

Such visibility is enhanced by some of Twitter’s features. The possibility of retweeting, which is a measure of the influence of an account, is one such feature. Most of the DG’s tweets are not retweeted very much, but for the DG of a Swedish public agency, the retweets make a measurable impact that indicates reasonable visibility. Of the tweets that have been retweeted more than 30 times, i.e., high retweets tweets, all but one are, in fact, retweets, and all but four contain links. The DG retweets something that is then retweeted even more, which again emphasizes visibility, not only in the sense that a retweeted tweet becomes more visible, but also in the sense that the DG becomes more visible by embedding him/herself as part of the content that is relayed in the tweet at hand.

The possibility of liking tweets is another function that enhances visibility. Tweets that are ‘liked’ can be filtered on a Twitter account’s profile page and, hence, be made more visible. Twitter users can do this on their own profile but also on any other user’s profile. In the case of the DG, 10% of the tweets with a high ‘like’-count (liked more than 30 times) are native (i.e., written by the DG). In comparison, only 0.5% of tweets with a high retweet count are native.

From the Twitter feed, we gather that the agency the DG directs is undergoing a reform that focuses on leadership issues and digitalization. We also gather that the DG’s favorite football club is Manchester United, that s/he has a partner and a child, watches the Eurovision song contest, and plays Pokémon. We know that s/he owns an iPhone, a private Mac and a work iPad. Twitter affords visibility by relaying information about and from the tweeter and making them available first to his/her followers, and potentially to the followers’ followers. That is the image available to the public through Twitter, and presumably the image that the DG deliberately wishes to convey.

… I cannot tweet as a private person anymore, I do it very little. I have chosen some subjects to tweet as a private person: football and the Swedish Melody Festival and the Eurovision Song Contest.

(IT55, December 2016)

Although it may be obvious, the quote above indirectly makes it clear that the DG and his/her Twitter image are not the same. The difference is marked by a concomitant invisibility that Twitter affords the accountholder. This is a result of the conscious selection of topics to tweet about, which is linked to the affordance of editability (see below), but, more importantly at this juncture, the functioning of tweets as metaphors for the tweeter. Like any metaphor, they are partial and selective (Morgan 1986), and any reflection they convey is therefore necessarily a simplification that emphasizes particular aspects.

The visibility afforded is moreover constrained by a specific format. Tweets are mostly textual and are 140/280 characters long. As such, they are reduced to fit the space allowed by the medium and, more importantly, deprived of the specific context of their conception. The visibility that a tweet affords is that of literal meaning; its character is, as Knorr Cetina (2009) puts it, informational. Consider the following tweet:

Today’s most annoying thing was that I ran out of Pokémonballs. Good night! (31 August 2016 21:02:28)

The literal interpretation of the visibility afforded by Twitter is that the tweeter ran out of Pokémon balls on that specific day and that it was the most annoying thing that happened that day. We know that the tweeter, or the Pokémon player, which is the image suggested, played Pokémon that day, but not whether s/he likes it or plays regularly, or why s/he was playing Pokémon. Without a context, anything beyond the literal meaning of the tweet is difficult to ascertain. For example, is it a good or a bad thing that it was the most annoying thing, given that it is a trivial event? Is it because nothing else happened that it is a good thing and therefore the most annoying thing? And is “Good night!” an expression of well-wishing or a sarcastic remark?

The informational character of tweets leaves a lot of room for interpretation, because tweets withhold the context of their conception. Anything beyond the literal meaning of a tweet needs to be fathomed by the tweet’s reader, and to do so s/he needs a context. As Murthy (2012) argues about re-tweets, and Albu and Etter (2016) about hashtags, we argue that given their informational character, tweets need to be re-embedded into a context if they are to be interpreted beyond their literal meaning. Therefore, the affordance of visibility cannot be understood without its opposite, invisibility, understood as its reverse or as a whole other affordance, and the enabling and constraining aspects it entails.

The Twitter feed, what is on screen, is a composite of tweets that, as Knorr Cetina (2009) argues, is informed by a synthetic situation. It affords visibility, but also invisibility, because it provides bits of information, short tweets limited to 140/280 characters, with little or no context for interpretation to the follower who does not already know the tweeter first-hand. Moreover, tweets need not only be read from the page of the originator, but also from a follower’s unique Twitter feed, as a retweet (Jackson 2007) or via a hashtag (Albu and Etter 2016), etc. The reality that Twitter readers confront on their screens, an assemblage that is unique to the reader in question, is consequently an assemblage that is open to interpretation, given the synthetic situation produced both by Twitter and the reader concerned, and this synthetic situation is very much characterized by invisibility as well as uniqueness. In this sense, Twitter affords visibility to the tweeter to spread a message, but it does so by affording invisibility to the context of conception, the original tweeter (Albu and Etter 2016; Knorr Cetina 2009; Murthy 2012) and to a specific meaning beyond the literal. Next, we turn to the fluidity of Twitter’s persistence.

5.2. The Fluidity of Persistence

Persistence refers to the insistent memory of the online space. Once something is online, it is potentially there for all eternity. The DG refers, for example, to the persistence of tweets when explaining that past guesses about winners in public debates are lost forever, because outlets such as Twitter did not exist.

Maybe, maybe I selected a winner when I worked for politicians... But there was no Twitter so it’s impossible to dig it up... (18 November 2015 21:39:22)

The online eternity that is on offer for all our momentary lapses in judgement is provided partly by the technology itself and partly by the relentless screening of online behavior and the hordes of self-proclaimed online police that take screenshots of everything that can be used against a person. These screenshots turn tweeted mishaps into permanent objects in the questionable art exhibitions of online behavior.

The ability to take screenshots, retweet to new and increasing audiences, ‘like’, and reproduce otherwise endlessly has an upside. This reproducibility enables persistent visibility and impact, which is helpful for all kinds of causes that depend on the spread of information. Part of this impact can be seen in the relatively new practice of traditional media referring to tweets and Twitter conversations. This reference from one medium to another increases the visibility and persistence of tweets. The early morning and late-night tweets by President Trump that keep the media, US citizens, and certainly some world leaders on their toes illustrate how tweets cross boundaries to become even more visible and persistent.

Twitter is also persistent in the sense that tweets can follow one throughout the day. The DG might post a work-related tweet during the day but other people on Twitter do not stop engaging with it just because the DG has left the office at 6 p.m., or because the DG is on a plane with no internet access. Twitter crosses space and time boundaries. The DG tweets, for example, directly to an employee that s/he will answer their earlier question with a short message the next day. S/he crosses and protects boundaries at the same time by answering in the present and bringing both people back to the work issue by indicating that it can be dealt with the next day.

@[ManagerAtAgency] I realize that I did not reply to your text from yesterday. I’ll do it tomorrow. (4 February 2017 19:48:52)

We know that once something is online, it stays online, and that tweets are asynchronous and therefore need not respect time. The temporal situation presented by Twitter in effect consists of and is defined by the latest tweets. It is run by what Knorr Cetina (2009) refers to as ‘response presence’. In other words, the present on Twitter is constituted by the latest response. However, this emphasizes fluidity rather than persistence, because the present of the synthetic Twitter situation is based on the succession of tweets, of which the latest constantly pushes back older ones (and since the latest change, the most popular tweets push down more recent tweets). During an interview, the DG referred to as the fleeting or right moment to write and send a tweet:

… many times, I have something really good but I hesitate and Twitter is so fast. The day after it may have become too boring to send. But that’s because I check it out in my mind, I think about what image it may relay.

(IT55)

The quote above refers to the affordance of editability, which we will turn to below, but also to the inherent fluidity of Twitter and the difficulty of responding to a tweet at the right time. The right moment to send a tweet is indeed fleeting, because the long-term visibility of tweets is not the same as the visibility they enjoyed when they were first tweeted. The right moment to tweet may therefore pass as tweets fall away in the Twitter feed. They are still there but are not as visible and therefore not as persistent. Old tweets are not readily available but may require considerable scrolling, with loading times increasing as they get older, or specific searches if the tweet is very old (anything older than six months may as well be a lifetime on the Internet). Old tweets do not disappear completely, hence the persistence that Twitter affords, but they are increasingly difficult to unearth in the constantly growing composite of information that comprises the Twitter feed.

Moreover, the ability to make old tweets relevant and up-to-date once again is not only proof of the persistence that Twitter affords, but also of its asynchronous character and the fluidity that this composite of information implies through the ability to re-embed an old tweet in the present, placing it on top of the Twitter feed by tweeting an answer or retweeting it. In effect, as Knorr Cetina (2009) writes, the synthetic situation is temporal and not spatial. It is continuously real as a virtual place, but it is also continuously updated and thus always in flux. The now is the tweet or retweet at the top of the feed. More importantly, fluidity comes from the constant possibility of re-embedding (Jackson 2007), i.e., the freedom to use the fluidity of persistence to appropriate tweets in multiple, undetermined ways.

5.3. The Interaction of Associations

Through online behavior, a person can establish connections with individuals, content (Treem and Leonardi 2012), and with organizations (Vaast and Kaganer 2013). In other words, tweeters can associate themselves with one or multiple organizations, people, ideas, and/or knowledge.

The DG joined the organization in 2014, with one particular task, to develop the organization and lead it into the 21st century, after its development had been neglected for years. In this context, his/her tweets focus on the development of new, particularly digital, services. For example, the DG makes his/her association with his/her work organization clear in the ‘about’ section, which states that the accountholder is the DG of the agency. In the example below, the DG associates work at the agency with Google teams.

At the [Agency] meeting the management team right now: Inspired by Google to develop successful teams. #reform (3 May 2016 07:13:32)

The association with Google highlights the fact that they are also going to prioritize teamwork. It reflects positively on the DG and the Agency, by associating him/her with the cutting-edge leadership of Silicon Valley and with renewal. Moreover, the visibility of this association is made broader by the hashtag ‘#reform’, which the DG uses so that people not following him/her on Twitter can access this important information. Similarly, praising another public agency for donating funds to a good cause instead of giving its staff Christmas presents is a way of spreading positive vibes and emphasizing his/her own organization’s culture change.

I suppose the @[OtherStateAgency] is so well-managed that it gives room for some things. We at [Agency] are renewing ourselves to be able to do the same… @[AgencyWorker1] (11 November 2015 21:44:07)

The DG also associates with the other leaders of national and foreign public authorities, indicating that his/her personal network extends beyond Sweden, and s/he uses Twitter to praise individual employees for their achievements and to announce when the organization wins awards.

Writing together with my colleague [Firstname Lastname, DG of another agency]: [Link to article] (8 May 2017 05:21:55)

On my way to London to meet colleagues and private players. The idea is to dig into how the UK uses private actors. #[Agency] (7 September 2016 12:32:18)

Fun when we and our co-workers are getting awards! #[Agency] [Link to article] (30 March 2017 09:17:34)

S/he also engages regularly with the Agency’s employees. When addressed by employees, the DG answers with direct messages on Twitter, making it an important space for dialogue. By tweeting in this way, the DG also strengthens his/her association with the whole organization. This can be understood as an integral part of his/her attempt to establish a new leadership style. S/he wants to be seen as being in touch with the organization. The signal is that s/he is always available as a DG, even when tweeting about the Eurovision Song Contest, to receive important feedback, especially from the base of the organization.

… and when I have tweeted stuff, private messages pop sometimes up and they say: ‘I saw that you said so and so. Think about this and that. This is what it is really like’. So I get a reality check.

(IT55)

The DG’s references to Star Wars and the Eurovision song contest help build an image of an approachable, down-to-earth leader who is ‘in touch’ with pop culture. However, this openness may attract criticism, which s/he addressed as follows on the night of the 2016 Eurovision Song Contest, after commenting on the songs and performances in numerous tweets:

To those who occasionally show up and comment I should work instead of watching the #eurovision, greetings, I worked 12-2. Ok? (14 May 2016 18:13:01)

This example shows how the informational character and the fluidity inherent in the synthetic Twitter situation enable not only association, in this case with the popularity of the Eurovision Song Contest, but also dissociation. The tweet is an answer to private messages that apparently reduce all previous tweets by the DG about the show into the informational bit “the DG is watching the Eurovision Song Context and consequently not working”. The DG’s tweets were thus dissociated not only from the context of their creation, i.e., the weekend, but also transformed into a minimal expression that re-embedded them into a new, critical context.

Visibility allows associations to shape the public image of an accountholder. Invisibility in turn reduces the accountholder into his/her tweets and may therefore dissociate the tweet even further from its full informational content. In effect, beyond the specific tweet and its literal meaning lie manifold interpretations that can be made visible through ensuing critical, tendentious, ironic and/or humorous tweets. To reiterate and expand on this, consider the following Twitter conversation[s] that started with a tweet about the DG’s new Ipad:

DG:	My new Ipad which can show two apps side by side is like made for Twitter. Very good Apple.. (30 December 2015 10:04)
Tweeter1:	@[DG] Christmas present? (30 December 2015 11:27)
Tweeter2:	@[DG] Wishing the [Agency] would work more in the field and perhaps use Ipad Pro as its main tool (requires continued high pace development around IT) (30 December 2015 13:53)
DG:	@[Tweeter2] more mobility is surely coming. The question is what platform we are going to work with (30 December 2015 13:56)
DG:	@[Tweeter1] not really, work’s (30 December 2015 13:57)
Tweeter2:	@[DG] preferably systems that are as platform independent as possible according to me. (30 December 2015 13:59)
Tweeter1:	@[DG] good employer (30 December 2015 14:06)
DG:	@[Tweeter3] It’s actually from work, but roaming is turned off. Private mobile phone though with a local prepaid SIM card. (31 December 2015 13:25)

The conversation[s] above not only show association and dissociation along the conversation, but also the proliferation of meaning. Three different associations/dissociations of the original tweet develop as the tweet is re-embedded into new contexts, i.e., Christmas, technological development at the agency and the ethics of using communication devices paid for by public authorities. As is clear from the example above, the informational and fluid character of the synthetic situation appears first and foremost, as Knorr Cetina (2009) writes, in the interaction between the reader and the tweet. This opens up the realm of interpretation and opportunities to embed and re-embed utterances (Murthy 2012) into new meaning-making possibilities driven by the synthetic situation of reading, rather than writing, a tweet. Now we turn to the final affordance presented by Treem and Leonardi (2012), editability.

5.4. Editability as a Twitter Persona in the Making

Descriptive statistics of associations indicate the distribution of ideas and knowledge and offer another way to address the affordance of editability. In effect, of the 3199 tweets, 1297 (41%) have links in them and more than half (59%) of these are retweets. Secondly, we found that of the 1083 retweets (34% of all the tweets), 71% have links in them. When we exclude all the replies from the data, we find that 55% of all tweets are retweets.

We draw two conclusions from this. Firstly, the DG does not randomly retweet other people’s musings, but focuses on sharing links that help to inform/shape public discourse. Our analysis reveals that the DG mainly shares links related to the Agency and to news articles from Sweden’s leading newspapers. The following word cloud indicates which words appear most often in all his/her tweets, demonstrating that the associations with ideas and knowledge are far from random. Figure 1 is a word cloud for all tweets. “Agency” is a popular word, along with “PressAgency,” but others are private, such as “Eurovision” [Eurovision song contest] and “mufcse” [Manchester United Football Club Sweden].

Secondly, the DG’s tweets are very ‘link’ focused. Overall, 41% of all tweets contain links, which indicates that s/he is not only sharing his/her own musings but attempting to broaden the discussion. This argument is supported by the statement “will RT [retweet] without sharing opinion” in the ‘about’ section of her/his Twitter page. The ideas and knowledge shared are mainly connected to work. Even the current affairs tweets that the DG shares are often connected to the organization’s aims (e.g., numbers, statistics), and s/he makes a lot of positive retweets. All tweets related to leadership, the ‘reform’ [culture change], and digitalization are native tweets.

This work is about the technology behind the website, and about new digital services, such as an improved [Agency’s goal] feature. (9 March 2016 18:24:15)

In a way, the DG engages in very carefully crafted form of impression management. It is not important to tweet directly about everything that the organization does; the focus needs to be on what the DG wants to be associated with, and that seems to be leadership, cultural change and digitalization. His/her plan is for cultural change within the agency, and s/he uses every opportunity to explain what that means and to spread the word. Regularly, s/he will post a succession of tweets explaining the reform and cultural change.

#reform also includes new leadership, a new value base and a strategic map of what we want to improve year by year. (3 May 2016 06:09:50)

Editability refers to the ability to craft a message before it is seen by others (Treem and Leonardi 2012). Although we did not observe the DG when s/he was tweeting, some things can be inferred from the analysis of his/her Twitter feed and the interviews. Recall, for instance, the DG’s quote about how fast Twitter can be and that responses can “become too boring” if one waits too long to think about what an image might relay. From the section on visibility we gathered that the DG’s image on Twitter is carefully crafted by his/her editorial choices. As indicated above, we know that the DG focuses on digitalization, self-leadership, and the agency’s reform in work related tweets, and that s/he is interested in Manchester United, the Eurovision Song Contest and the Swedish Melody festival. From the section on association, it is also clear that the choice of topics to tweet about as a private person is conscious and that s/he has chosen to restrict her/his tweeting as a private person to only a few subjects (football, the Eurovision Song Contest and the Swedish Melody Festival). Of the latter, s/he said in one of the interviews:

I am not that interested in the Melody Festival, but I tweeted about it, and I don’t remember why, and it drew so much attention that I thought, why not?

(IT55)

The DG’s public image on Twitter is to some extent what s/he wishes to convey by carefully choosing subjects to tweet about to relay a certain image and consciously crafting the tweets. Editability on Twitter can thus be understood as the careful crafting of specific messages. More importantly, however, editability refers to the editorial role a tweeter can exercise through conscious choices. In the case of the DG, it is clear that s/he placed an editorial boundary around subjects to tweet about as a private person. Rather than pointing to the editability of specific tweets, this relates to the whole collection of tweets and the image that is constructed in and through them.

As argued above, the whole person does not become visible, just a negotiated, polished Twitter persona. In that sense, Twitter, or any social media, can be used to create a form of ‘attractive visibility,’ a version or sample of the person tweeting, a Twitter persona that is detached from the accountholder since it is not the whole person, but just a collection of tweets open to interpretation. Insofar as the Twitter persona is a reduced form of the accountholder, invisibility also opens up for a counterpart to association to emerge through the dissociation of the accountholders and their contexts from the Twitter persona. In effect, this means that association beyond the literal meaning of a tweet is open to interpretation and thus potentially dissociative from any associations intended by the accountholder.

Understood in light of the affordances of visibility, persistence and association, editability, i.e., the editorial role on Twitter, can be understood as editorial control over which topics are permitted, and over the crafting of specific tweets. However, as the DG once put it in a tweet:

The fascination with Twitter is that anything can happen… (Ping @moabrglf @osinca @oisincantwell) [screenshot] (2 December 2015 16:24:38)

Thus, aside from the obvious editability of and control to determine any written message, Treem and Leonardi’s (2012) affordance of editability could thus be better understood not only as the editorial role exercised by a tweeter to create an online persona, but, in the light of our discussion about the affordances of visibility, persistence and association, as well as their counterparts invisibility, fluidity and dissociation, as the collective online creation of that persona. In that sense, editability can be understood as the total crafting of an online persona, in and through the visibility/invisibility, persistence/fluidity and association/dissociation afforded by tweets, retweets, hashtags, etc. and the editability/indeterminacy that such a process of collective co-creation involves.

6. Conclusions

While most research on the affordances of Twitter focuses on the enabling and constraining outcomes of inherently positive affordances (Gibbs et al. 2013; Leonardi et al. 2013; Majchrzak et al. 2013), we have focused on the synthetic situation (Knorr Cetina 2009) that arises through Twitter communication, to explore, reinterpret, and derive counterparts to the affordances of visibility, persistence, association and editability presented by Treem and Leonardi (2012): invisibility, fluidity, dissociation and indeterminacy. In doing this, we contribute to research on the affordances of social media technologies by focusing on the synthetic situation that Twitter communication entails and producing a reinterpretation of the affordances presented by Treem and Leonardi (2012).

Thus, we have shown that in a synthetic situation, such as Twitter, interpretation is influenced by the visibility/invisibility and persistence/fluidity of the Twitter feed, and the associations/dissociations that they enable/constrain. This view extends the work of Albu and Etter (2016) on hashtags, and their ability to transpose conversations from one context to another, by describing the whole of Twitter and its synthetic situations as fluid and indeterminate processes of meaning-creation, in which editability as the crafting of a Twitter persona cannot be understood as an individual endeavor, but as collective, involving the embedding and re-embedding of tweets, retweets (Jackson 2007), and ‘likes’ into ever-changing synthetic situations. Moreover, this endeavor is characterized as much by editability as by indeterminacy. Seen this way, Twitter becomes, as Jackson (2007) writes of the web 2.0, a virtual place with its own time, which promotes the promiscuity of content through indiscriminate association/dissociation.

By pointing to the asynchronous character of Twitter and the fluidity of the Twitter feed, which creates opportunities for multifariousness and the indeterminate constitution of a Twitter persona, it also becomes clear that, although synchronic and diachronic perspectives are very much conflated into the present of a tweeter’s feed, they are essential for the study of social media communication. In effect, individual tweets and Twitter conversations are elemental for understanding the synchronicity inherent to the synthetic situation. In turn, a diachronic perspective is important not only as an outcome of Twitter communication, but also as part of the input to the synthetic situation, insofar as a follower not only follows one tweet but a succession of tweets that are incorporated into their Twitter feed.

In conclusion, using social media involves a certain boundarylessness that challenges our understanding of what a person is and becomes online. Through the affordances of Twitter, we can watch almost in real-time how an account becomes a certain version of a person through their own doing and the doings of the wider Twitter universe. Technologies such as Twitter amplify many of our capacities and enhance our possibilities by negating and conflating time and distance and allowing for promiscuity of content (Jackson 2007). This means, as Albu and Etter (2016) write, a disturbance of the notion of authorship and the emergence of an existence on its own. In terms of the affordances of social media, this entails, in turn, a disturbance in the notion of affordances, characterized in the beginning of this paper as the play of “absence” absent in earlier conceptualizations of affordances, and in extension the need to include opposites in the analysis of the affordances of social media. To that purpose, as we have tried to show, we have proposed alternatives or counterparts to Treem and Leonardi’s (2012) affordances of social media.

One limitation of this study is the focus on only one Twitter profile which limits our ability to speak to this as a broader phenomenon. By choosing a person in a semi-public position, valuable insights can be gained about this particular group of people, but it might also be less representative of the wider population. The study is placed in the Swedish context, which is different due to strong culturally anchored expectations of transparency but also flat hierarchies.

Given the results of this study and the limitations, we suggest future research directions with both quantitative and qualitative methods. Firstly, large-scale quantitative studies (i.e., multiple Twitter users with different roles within organizations, over longer periods of time) looking at the form and shape of the synthetic situation on Twitter are encouraged. Ideally, these studies would seek to understand the dissemination of the online synthetic personality through various computational frame works (e.g., Tasgin and Bingol 2019). Secondly, we see room for in-depth studies of few profiles but with comparative elements. Thirdly, studying the synthetic situations arising between Twitter users that belong to the same organization could tell us more about the other side of the affordances i.e., invisibility, fluidity, dissociation and indeterminacy, particularly if combined with observational data from interactions between those users. Fourthly, studying the emergence of the synthetic situation through including the followers of one or multiple profiles would build further on our understanding of this theoretical framework. Given the rapidly changing nature of social media, there are certainly many more studies that could provide useful insights, especially for organizational policy formulation.