Online distant witnessing and live-streaming activism: Emerging differences in the activation of networked publics

Introduction: NO DAPL and Digital Smoke Signals

In June 2014, the project for the construction of the Dakota Access Pipeline (DAPL) was publicly announced. This 1886-km-long oil-pipeline has been designed to enable the transportation of 470,000 barrels of crude oil a day from North Dakota to Southern Illinois (USA). However, since its formal approval, the DAPL project raised public concern about the environmental sustainability and security of the authorized infrastructure. Among the various groups that contested the construction of the DAPL, the intense resistance organized by the Native American residents of Standing Rock Reservation (Sioux) rapidly attracted public attention. From April 2016, a few months before the official beginning of the construction work, the Sioux of Standing Rock Reservation inaugurated a first permanent camp (Sacred Stone Camp) located near the planned path of the pipeline. The gathered protesters, who named themselves “Water Protectors,” demanded the rerouting of the DAPL. Their demand was motivated by the fact that, according to their claims, the construction of the oil-pipeline would both pose a grave danger to the water supply in the area (Missouri River) and defile terrains held sacred by Native Americans. The number of protesters grew exponentially during summer 2016, leading to the creation of two additional camps. As a whole, the NO DAPL movement represented the largest gathering of Native American tribes in more than 100 years (Gray, 2016). The high level of participation to the protests and their extraordinary media visibility eventually pushed the Obama administration to ask for the temporary suspension of the construction work and a reexamination of the DAPL project (US Department of Justice, 2016). The nonviolent action of the Water Protectors continued throughout the winter of 2016/2017 and was actively supported by a vast array of different groups, perhaps most notably that of 2000 US veterans who joined the Water Protectors in November (Mele, 2016). Nevertheless, the outcome of the 2016 US presidential election marked a critical passage for the NO DAPL movement. On 24 January 2017, the newly elected President Donald Trump signed an executive order that abruptly interrupted the dialogue initiated by the Obama administration with the Water Protectors and paved the way to the forced and complete evacuation of the native camps.

Despite the final completion of the oil-pipeline, the NO DAPL movement succeeded in outreaching and mobilizing various segments of the United States and international public. This important result might be attributed in large part to the adoption of a media-by-default tactic: the participated, strategic, and constant use of Internet and communication technologies (ICTs) as a fundamental component of the resistance activity. From the beginning, each activist-camp launched dedicated accounts on Twitter, Instagram, and Facebook which provided users, both activist and non-activist, with constant updates. Supplementing mainstream social networks, Native American media activists employed community-based news websites (e.g. Indian Country Today Media Network and Indigenous Rising Media) which worked as news providers for social networking sites (SNSs) sharing and discussion. In view of this intense media activity, this article focuses on the strategic use of SNSs by the NO DAPL movement throughout the most active period of the protests, that is, from the beginning of the DAPL construction work to the forced evacuation of the native camps by the Morton County Sheriff Department.

In line with recent studies on users’ online engagement in political movements, this research will investigate how different forms of connective activity are triggered by different forms of online content. More specifically, given the NO DAPL strategic use of cameras and online video-sharing platforms, this study aims at highlighting variations in users’ connective activity in relation to two different forms of distant witnessing: online videos and live streaming. This objective is articulated as follows:

To answer these questions, this study will analyze the users’ activity which took place on the Digital Smoke Signals Facebook page, one of the most widely followed information outlets of the NO DAPL movement. Founded by Myron Dewey, indigenous journalist and Native American rights activist, this Facebook page based its outreach strategy almost exclusively on audio-visual communication. From the end of October 2016 to the forced evacuation of the camps at the end of February 2017, Myron Dewey posted 299 audio-visual items consisting of online videos and live streaming. These contents progressively increased their circulation and impact to the extent that, for example, the video “Protectors shot by rubber bullets, mace and water canon” (uploaded on 28 November 2016) received more than 2.5 million views. The quantitative analysis of users’ activity, as well as its transformation throughout the course of the protest, will shed light on the distinctive forms of connective action activated by different forms of online distant witnessing.

Online political engagement and the emergence of real-time distant witnessing

This study lies at the crossroads between two recent yet fast-growing lines of academic inquiry: the investigation of online political engagement and that of emergent forms of media witnessing on digital platforms. Since the beginning of the 1990s, the tension between more or less explicit forms of digital utopia (Turner, 2010) and digital catastrophism (Jones, 2013) has characterized not only early academic debate but also popular cultural production. Nevertheless, besides this polarization, the pervasive and increasingly unavoidable use of ICTs for the accomplishment of everyday tasks has raised questions about users’ privacy, freedom of information, and the hegemonic position of Internet giants (Chadwick, 2017; Couldry and Hepp, 2016; Pariser, 2011; Vaidhyanathan, 2012). In particular, the role played by social networks in recent key political events, such as the 2016 US election and the Brexit referendum, have drawn academic and governmental attention to the problems arising from the online interaction between human and non-human actors (e.g. bots and botnets) for the functioning of democratic systems (Bessi and Ferrara, 2016; Howard and Kollanyi, 2016; US Office of the Director of National Intelligence, 2017).

The perceived negative effect of SNSs on political participation reignited the time-honored debate concerning the potential of digital platforms to support direct democracy and boost citizen political engagement. While some scholars went as far as throwing into question the survival of democracy in the digital age (Helbing et al., 2017; Persily, 2017), others took this chance to critically reassess and problematize the potential of new media for social advancement (Jenkins and Zimmerman, 2016; Marichal, 2016). Until recent times, studies concerning the political impact of SNSs have been limited to their strategic use by social movements as a tool for social mobilization. Emphasizing the relationship between online and offline political activity, scholars approached digital environments as new venues for sharing information, coordinating activities, and taking decisions (Warren et al., 2014). This line of research highlights the existence of an intense political activity online which is characterized, however, by different levels of efficacy (Conroy et al., 2012; Miller, 2017) and highly blurred boundaries (Vissers and Stolle, 2014a). To outline the limits of the politics on the Internet is a complex task which has not produced general rules. This is due not only to the existence of a spill-over effect between online and offline engagement but also to the constant emergence of new modes of participation resulting from this double movement (Vissers and Stolle, 2014b). Accordingly, the literature concerning the use of SNSs for political participation has mostly focused on the analysis of specific cases, such as that of the Jasmine Revolution (Müller and Hübner, 2014) or Chinese digital activism (Tai, 2015), and described the participated construction of shared values, collective agency, and political efficacy on digital platforms (Velasquez and LaRose, 2015).

The study of online engagement requires an understanding of how affordances offered by different digital architectures are utilized to create relationships, spaces, and identities which are perceived and recognized by users as political. Reviewing the extensive body of literature on this topic, it should be noted that, in the last decade, more and more scholarly attention has been devoted to the rapid emergence of online visual communication. While, on the one hand, this change is the result of a technological and industrial development (e.g. popularization of smartphones and broadband Internet connections), on the other, its implications go far beyond these technical aspects. From the commercialization of the first portable cameras, the grassroots production and circulation of visual material have been often perceived as a politically charged activity (Apel, 1999; Martini, 2017). With the advent of the Internet, however, the massive proliferation of video-related media practices has progressively transformed the nature of visual communication. Today, media production, circulation, and consumption play a central role in constructing the social value of a certain occurrence (Andén-Papadopoulos, 2008; Feldman, 2005). This ubiquitous, continuous, and networked media activity has eventually affected users’ perception of everyday experience: reality constantly appears to them as a potential object of mediation (Miller and Sinanan, 2017; Nair and Aram, 2014).

The impact of distant media witnessing on networked publics and, by extension, on democratic governance is the object of recent scholarly investigation. However, in this line of study, the research on the potential dangers of networked media witnessing, such as its strategic use for propaganda (Seo, 2014; Seo and Ebrahim, 2016), the lack of shared ethical guidelines in citizen journalism (Allan and Peters, 2015; Mortensen, 2014), or the non-consensual circulation of private materials (Albury, 2017; Stroud, 2014), is still at a seminal stage. When discussed in relation to its political dimension, new media witnessing is still generally regarded as a positive form of emancipation from the mass media construction of reality. In other words, the political potential of networked video-mediated communication is presented in relation to its strategic use by activist groups which embrace grassroots video production for documentation and counter-surveillance purposes (Farrell and Allan, 2015; Gregory, 2015). In these cases, the technology-enhanced mediability of reality is strategically employed by media activists to support collective action and expose repression (Andén-Papadopoulos, 2014; Lo, 2016).

Besides its political impact, however, grassroots production and online circulation of audio-visual materials raises several problems concerning authenticity and the ethics of witnessing. Reflecting an established scholarly tradition, the tension between the mobilizing potential and the moral implications of witnessing other people’s suffering is often brought into question (Andén-Papadopoulos, 2017; Chouliaraki, 2010; Kyriakidou, 2015; Martini, 2018). This tension, particularly clear in the case of armed conflict, brings back the issue of who has the right to videotape and circulate somebody else’s suffering, in which ways, and for which purposes. The breaking of the professional monopoly of journalism has opened, de facto, a struggle for authority between well-established news companies and smartphone-armed citizens on the ground (Bock, 2012; Chouliaraki, 2015; Mortensen, 2014).

This rising tension between top-down and bottom-up forms of truth-making is indicative of a radical change in the relationship between video-makers and networked publics. This relationship, reshaped each time by the videotaped occurrence, is often discussed in relation to users’ perceived proximity to the event. While some scholars state that the symbolic management of distance guaranteed by professional journalism is fundamental to trigger feelings of moral responsibility in the public (Andén-Papadopoulos, 2013), others propose the development of strategic media witnessing as a structured form of social activism (Ristovska, 2016). Within this debate, the advent of the live streaming and its incorporation in major SNSs as part of the user’s communication toolkit represents a turning point. Studied much less than circumstances warrant, live streaming is a form of audio-visual communication which radically differs from online video (Golan and Martini, 2017; Lenzner, 2014; Thorburn, 2014). As some recent studies suggest, live streaming generates not only a higher level of engagement but also a qualitatively different kind of engagement (Fichet et al., 2016; Haimson and Tang, 2017). Networked liveness implies the interpenetration of physical and virtual spatio-temporal coordinates through the communicative performances of users (Mostafa et al., 2016; Webb et al., 2016). In this line of studies, this research intends to understand how users’ experience of networked media witnessing changes in relation to synchronous (i.e. live streaming) and asynchronous (i.e. online video) forms of audio-visual communication.

Connective witnessing: bridging connective and collective modes of social action

The need to intersect the two above-mentioned lines of research arises from the widely accepted yet rarely empirically explored assumption that media witnessing on social networks triggers a different kind of user engagement. To this aim, Mortensen’s (2015) concept of “connective witnessing” will be assumed as a main theoretical framework and further developed in relation to the findings. Elaborating on Bennett and Segerberg’s (2013) idea of connective activity, Mortensen states that connective witnessing is a new way of relating to audio-visual material through SNSs’ digital architecture. Conceived as a result of the massive participatory circulation of user-generated footage, this form of networked experience raises several issues concerning social responsibility, personalized political participation, and user agency (Dolata, 2017; Mercea and Funk, 2016). Elaborating on this theoretical background, this study aims to sharpen the idea of connective activity. At the core of this concept, developed in contrast to more traditional forms of collective activity, “is the recognition of digital media as organizing agents” (Bennett and Segerberg, 2013: 752). This general acknowledgment, as well as its adaptation to media witnessing, invites a more empirical analysis of how different online environments reorganize user activity and how, in turn, such activities are performed and experienced by users.

Witnessing is a practice which radically differs from observing or being somehow aware of the occurrence of an event. Witnessing is primarily a practice of mediation composed of two main phases: the act of perceiving a certain occurrence and the act of communicating the perceived occurrence to others. In these terms, the act of witnessing expresses the establishment of a specific relationship between an individual and his community. Through the act of telling, a private experience becomes a public account (e.g. court or journalistic interviews). This communicative performance retroactively reframes the act of seeing as witnessing and the individual as a witness, emphasizing the latter’s role as a medium between the past event and the community (Peters, 2009). With the advent of mass media, this role has often been assumed by professional witnesses, such as journalists, who observe and report an event on behalf of an entire social group. This process of professionalization and massification of witnessing has raised several issues, focusing often on the passive role assumed by viewers in relation to televised events (Ellis, 2000). However, as Mortensen (2015) notes, the act of witnessing has been considerably transformed by the popularization of digital technologies. On the one hand, the practice of witnessing has been re-appropriated by single individuals who can now independently produce and circulate contents among their networked communities. On the other, media witnessing has become a strategic social practice which transforms the witness and his network in a composite agent taking active part in the unfolding event.

Given its dramatic nature, media witnessing accentuates, rather than mitigates, the existing tension between connective and collective modes of social action. Indeed, the “virality” of a certain content is tightly connected to its ability to migrate within the existing network and partially redefine its structure (Berger and Milkman, 2012). The connective activity triggered by certain content represents, metaphorically speaking, the propellant of such movement: a propellant which must be constantly reignited by the activation of new publics. Nevertheless, the encounter with different groups and their systems of values may or may not be successful: an iconic picture circulated within a certain community for its perceived affective impact might deliver little information to others, be considered offensive, or even completely misinterpreted (Bruns and Hanusch, 2017). In this context, the authority of media activists as alternative sources of information represents a central issue. Social media teams, often composed by a limited amount of people, are the link between collective actions conducted on the ground and connective activity online (Gerbaudo, 2017; Toepfl, 2017). In a cross-dynamic, the legitimacy of these groups is constructed through the balance between two converging factors. On the one hand, media teams are legitimized by social movements for their capacity to effectively circulate information and gather support while, on the other, users recognize them as part of the movement and therefore able to provide firsthand coverage and a channel for direct interaction.

To conclude, contrary to what the case under scrutiny may suggest, this research is only tangentially related to the field of Indigenous Studies. Nevertheless, the migration of indigenous political discourse and practice to social networks, as well as the corresponding cultural redefinition of the Internet as a legitimate domain for communal activity, indicates a transformation in the ways these groups construct their identity. For example, despite a very low level of engagement in established democratic praxis, Native American advocates utilize social media to disseminate political information and discuss issues perceived as of importance to the community (Vigil-Hayes et al., 2017). The restructuring of indigenous social life to digital architectures affects both established forms of social engagement and the culturally defined limits of such engagement (Duarte, 2017; Lindgren and Cocq, 2017). As this study indirectly suggests, these transformations may foster the rise of secondary authorities which, unlike traditional ones, are able to effectively reach and connect various publics (e.g. indigenous communities, activist groups, or governmental organizations). In this context, the appearance of indigenous media activists that facilitate communal and extra-communal engagement is symptomatic of an ongoing change. While this quantitative study focuses uniquely on the relationship between online witnessing and connective activity, the success of Native American activists in mobilizing various sectors of the national and international audience invites further investigation. Studies on the incorporation of the Internet and the rise of technology-savvy leadership in native communities would enable better understanding of the changing indigenous identity in the global age.

Methodology and ethics

Online video and live streaming: differences and similarities in user behavior

Video production and user activity

In 4 months, Myron Dewey produced and uploaded 299 videos on the Digital Smoke Signal Facebook page. This media activity is represented in the following chart (Chart 1).

OPEN IN VIEWER

Chart 1.

Digital Smoke Signal’s video production from November 2016 to February 2017.

As this chart shows, live streaming constitutes the majority of Dewey’s media production (84%—250 contents). The uploading of online videos (16%—49 contents) was mostly concentrated in the first 2 months of his activity and was progressively abandoned in the last two. This change in the media strategy can be related to several factors, which includes not only the unfolding situation on the ground but also the different affordances offered by live-streaming technologies. Indeed, it should be noted that as the use of live streaming increased there was a corresponding increase in video production which culminated, during the forced eviction of the camps, in the circulation of 20 videos in a single day. Not requiring any kind of editing or post-production, live-streaming feeds met the activists’ need to cover fast-evolving events and secure their footage without interrupting the ongoing action or leaving the scene. Users’ reactions highlight some interesting dynamics.

Users’ connective activity seems to be tightly related to the activists’ video production. Comparing Charts 1 and 2, we can see how both video production and users’ social activity have a peak toward the end of November and February while it is much lower between December and January. However, while the liking and commenting activity follow a similar evolution in the course of time, the sharing activity presents some evident discontinuities. On this Facebook page, sharing is by far the most common activity among users (510,828), while liking (328,568) and commenting (349,571) have approximately the same weight. Nevertheless, the distribution of the sharing activity is also much more uneven: it concentrates around certain contents (peaks), while it is much lower in the rest of the cases. This particular dynamic indicates the existence of a large group of users which gets activated only in relation to particularly relevant events.

OPEN IN VIEWER

Chart 2.

Users’ connective activity on the Digital Smoke Signals Facebook page from November 2016 to February 2017.

In short, Charts 1 and 2 highlight the centrality of the streamer’s initiative as a prerequisite for the emergence of connective activity. Indeed, it is the creation of new content which enables user interaction, consequently affecting its frequency and focus.

Users’ connective activity around online videos and live-streaming feeds

The described differences in the video production and connective activity can be further articulated and analyzed in relation to the distinction between online videos and live-streaming feeds. This distinction is necessary to describe if and how these two forms of online video-mediated communication activate users in different ways. The following chart shows how user connective activity is distributed around online videos and live-streaming feeds.

As Chart 3 shows, there are relevant differences between the social activity generated by users around online videos and live-streaming feeds. The social activity generated in relation to online videos (Chart 3(a)) consists largely in sharing (192,300 shares) and significantly less in liking (54,842 likes). The fact that commenting represents, in this case, a very limited part of user activity (7090 comments) is particularly relevant. Indeed, if compared to users’ social behavior on live-streaming feeds, we can see how in the second case the ratio between the three examined activities is much more balanced and that commenting is the most practiced activity (342,427 comments). The highlighted difference between these forms of connective activity suggests that online videos and live streaming are perceived by users as different digital objects. In particular, it could be argued that, whereas online video is perceived as a piece of information to circulate, live streaming triggers a much more dialogic relationship with users. To further investigate this assumption, the following chart describes the temporal distribution of the commenting activity.

OPEN IN VIEWER

Chart 3.

Social activity around online videos and live-streaming feeds. (a) online videos and (b) live-streaming feed.

In both cases, a substantial part of the commenting activity takes place in the 24 hours following the creation of the content which indicates the existence of an attentive and reactive public. Nevertheless, Chart 4 also highlights the existence of relevant differences between online video and live streaming. Online videos are characterized by a significant number of comments left as replies to other users (31%—2871 replies) and after 24 hours from the time of the original upload (45%—4462 comments). Proportionally, live streaming produces a very limited amount of replies (3%—5106 replies) and after 24 hours comments (6%—10,950 comments). Instead, they trigger an extremely intense commenting activity during the live transmission which produces, often within a few minutes, 36% of the total comments.

OPEN IN VIEWER

Chart 4.

Temporal comment distribution around online videos and live-streaming feeds. (a) online videos comment distribution and (b) live-streaming comment distribution.

In short, Charts 3 and 4 highlight the existence of significant differences in the connective activity generated by users around online videos and live-streaming feeds. Whereas online videos are mostly shared as pieces of news, live-streaming feeds trigger an immediate, high-intensity yet short-lived form of user engagement.

Reflexive and user–streamer interaction

User engagement on SNSs has often been described in terms of produsage (Bruns, 2008). However, live-streaming communication seems to be characterized by a local yet clear separation between the streamer and its public. While the former produces the feed, the latter generates connective activity around it. Such activity can be directed toward unengaged users (i.e. share), the existing public (i.e. reflexive activity), or the streamer (i.e. user–streamer interaction). This cluster of data investigates the latter two facets.

The measurement of the level of reflexivity expressed by an online group, however ephemeral it may be, enables the description of users’ relationship with the hosting digital environment, that is, the strategic and collective use of the available tools and the shared values ascribed to such practices. Accordingly, reflexivity codes were conceived in order to classify all those comments that expressed users’ perception of their own connective activity and its value in relation to perceived communal goals. This included explicit mention of social networking activities (e.g. “Hello! Minnesota watching!!” or “SUBMIT YOUR OWN COMMENT (easy to do) for the Army Corps of Engineers & say NO to DAPL”) as well as the active contribution to the streamer’s efforts (e.g. “You’ve got to share this if you want to support #NoDAPL” or “Please share videos otherwise we can’t see anything. Videos are getting jammed. Share, share, share for us!!”).

The second set of codes, the user–streamer codes, was designed to highlight users’ efforts to reach back to the streamer on the ground. Codes have been divided into three families: direct, cultural, and symbolic. Direct codes included comments mentioning the streamer’s name (e.g. “So good to hear your voice Myron” or “welcome back Myron missed you”) or addressing him in dialogic terms (e.g. “You are so close to us!!!” or “Can we donate to get you new drones?”). Cultural codes identified comments greeting the streamer utilizing terms which belong to the Native American cultural heritage (e.g. “A’HO from Florida” and “Together we WILL get justice and KILL THE BLACK SNAKE. A’ho Brother”), ² while symbolic codes marked comments stating verbal acts (e.g. “Prayers to water protectors!!” or “Many, many blessings and prayers for you great warriors.”).

The coding, differentiated between online videos and live streaming, produced the following results.

As Chart 5 shows, in terms of reflexive activity, the differences between online videos and live streaming are significant. On average, the reflexive activity around online videos is extremely limited (17.10), while that around live-streaming contents is much more intense (316.42), in particular during the live transmission and the following 24 hours. These data suggest that the synchronous relationship with events allowed by live-streaming feeds triggers users’ reflexive activity considerably more than uploaded videos. Besides, it should be noted that after 24 hours from the transmission, the reflexive activity around live-streaming contents drastically drops and reaches levels similar to that of online videos (live streaming: 5.11; online video: 5.18). These data also reveal an interesting qualitative aspect of this dynamic: users’ reflexive activity concentrates around the act of watching and sharing, while those of liking and commenting are far less common (Table 1).

OPEN IN VIEWER

Chart 5.

User reflexive activity around online videos and live-streaming feeds (average). The two sets of data (orange and green) are presented in three different temporally defined clusters. For each data set, the first line represents the “Live” comments, the second line the “24 hours” comments and the third line the “After 24 hours” comments.

OPEN IN VIEWER

Table 1.

User reflexive activity around online videos and live-streaming feeds (average).

		Watch	Share	Like	Comment
Online video	Live	0	0	0	0
	24 hours	3.387755	3.816327	3.632653	1.081633
	After 24 hours	1.326531	1.306122	2.020408	0.530612
Live streaming	Live	15.19008	11.46281	4.033195	1.479339
	24 hours	23.20248	15.35124	5.322314	2.764463
	After 24 hours	2.53719	1.285124	0.764463	0.524793

As will be discussed in the following section, these differences are probably related to the kind of relationship networked publics establish with activists on the ground through live streaming. In this respect, it is interesting to note that data resulting from the coding of the user–producer interactive activity seem to reinforce this assumption. Indeed, while quantitatively more intense than the reflexive activity, the interactive activity is distributed in a similar manner.

In this case, the peak of user activity takes place during live-streaming transmissions and the following 24 hours (average: 217.23). Online videos present a very limited amount of user–producer interactions (average: 28.89), mostly concentrated in the first 24 hours from the uploading. Symbolic and direct statements constitute the majority of user–streamer interactions in both live streaming and online videos, while cultural expressions represent a minor component.

In short, Charts 5 and 6 show a high level of user reflexivity and direct user–streamer interaction which corresponds to an intense connective activity triggered by live-streaming feeds. Accordingly, live-streaming publics actively recognize their existence as collectives, which, to a varying extent, include the streamer and the videotaped event (Table 2).

OPEN IN VIEWER

Chart 6.

User–producer interactive activity (average). Similar to Chart 5, also in this case the two sets of data (orange and green) are presented in three different temporally defined clusters. For each data set, the first line represents the “Live” comments, the second line the “24 hours” comments and the third line the “After 24 hours” comments.

OPEN IN VIEWER

Table 2.

User–producer interactive activity around online videos and live-streaming feeds (average).

		Symbolic	Direct	Cultural
Online video	Live	0	0	0
	24 hours	13.67347	7.632653	1.408163
	After 24 hours	4.387755	1.530612	0.265306
Live streaming	Live	37.08264	35.93802	8.516529
	24 hours	75.06198	48.82231	11.80992
	After 24 hours	8.847107	5.169421	1.256198

Connective witnessing: user activity and digital co-presence

The data presented above attempt to describe how synchronous and asynchronous forms of digital witnessing trigger different forms of connective activity. The case of Digital Smoke Signals is particularly suitable to highlight these variations because, while allowing the analysis of both online videos and live-streaming feeds, it maintains relatively clear boundaries: it is temporally demarcated, limited to a single digital platform (Facebook), and focused on the development of a specific protest. These limitations suggest that, within this case study, users who were exposed to live-streaming feeds were probably also exposed to online videos and vice versa. In addition, all the analyzed audio-visual contents have been produced by the same media activist. This ensures a relatively strong consistency in terms of media production and user–producer interaction. Accordingly, it is unlikely that the revealed changes in user behavior are connected to transformations in audience composition or communicative style.

Nowadays, witnessing has become a diffuse, participatory, and reflexive practice enabled and regulated by the digital architecture of social networks. As described in Charts 1 and 2, the uploading of a content is the necessary act which creates, each time, the conditions for the emergence of connective activity. This is due to the fact that the necessary tools (e.g. share button and comment space) are provided to users only in relation to a specific content which, accordingly, becomes the temporary digital venue of their activity. In these terms, the act of witnessing is redefined as a connective performance triggered and catalyzed by the appearance of a digital content which literally “makes space” for user action. Accordingly, variations in the perception and use of the generated digital spaces are particularly relevant for they can inform us about the ways in which these contents impact users’ everyday life.

From this perspective, the data presented in Charts 3 and 4 are noteworthy. As Chart 3 shows, while the liking activity is proportionally the same both in online videos and live streaming, commenting and sharing differ significantly. Online videos are treated by users as temporally isolated packets of information which need to be circulated for advocacy purposes (Share 76%) and only secondarily as a venue for interaction (Comments 3%). Moreover, Chart 4 shows the existence of a large share of replies (31%) which is not present in live-streaming feeds (3%) despite the much more intense commenting activity triggered by this form of content. If related to the data presented in Chart 6, which highlights a low presence of user–producer interactions around online videos, the high level of replies and after 24 hours comments (24%) suggests that users perceive these contents as venues for user–user and long-term interactions.

The data describing user activity around live-streaming feeds highlight an almost opposite dynamic. The substantially higher level of commenting activity, concentrated mostly during the live transmissions (36%) and the following 24 hours (55%), suggests the emergence of a different kind of connective process. Characterized by a low level of replies (3%) and after 24 hours comments (6%), the observed connective practices seem to rely on a perceived and shared sense of co-presence. This co-presence is expressed on two different yet tightly intertwined areas: user–streamer and user–user interaction. The first form of co-presence is expressed by the high incidence of live user activity. As Chart 6 shows, comments left during live transmission and the following 24 hours are characterized by a high frequency of direct and symbolic user–producer interactions. These data suggest that users actively attempt to reach back to the streamer and the activists. The streamer is usually addressed by name and engaged on a personal level. In this respect, it is interesting to note that Myron Dewey seems to regard online users as a strategic component of the movement. In an attempt to enhance their agency and maximize the impact of their online presence, he occasionally asks them to actively help him (e.g. “I need help fact-checking this …” or “We need to organize to counter what mainstream media says. Corporate media posts spread like wild fire”). Nevertheless, the high presence of symbolic interactions (e.g. prayers or blessings) is indicative of a lack of agency perceived by users. Indeed, the possibilities of action offered by the Facebook platform become an object of communal redefinition in the second form of co-presence: the user–user interaction. This form of interaction is characterized, perhaps paradoxically, by a low level of replies (Chart 4) but a high level of reflexivity (Chart 5). These data suggest that users interact intensively with each other in real time in the comment section but do not abandon “the stream” to initiate private sub-discussions. The fact of appearing to each other, of being active at the same time on the same page, is sufficient to express the existence of a community of individual users. By stating their connective performance (e.g. “California is watching” or “Please share, be aware”), this ephemeral community strategically constructed the acts of watching and sharing, that is, bearing witness and circulating the content, as communal values (Chart 5).

To summarize, the discussed data suggest that Facebook online videos and live-streaming trigger different forms of connective response (RQ 1). These differences concern both user–streamer and user–user interaction. The connective activity generated by online videos is a long-term and low-intensity form of engagement, characterized mostly by a high level of passive sharing and a minimal volume of reflexive activity and user–producer interaction. Conversely, the connective activity generated by live streaming is typified by a high-intensity user engagement during live transmission and the following 24 hours, with significant levels of reflexive activity and user–producer interaction (RQ 2). The presented data describe the development of a specific series of events and have been collected on a single Facebook page during a well-defined, uninterrupted, and significant period of time. Accordingly, it can be reasonably implied that users who were exposed to the live-streaming feeds were probably also exposed to online videos and vice versa. For this reason, the presented statistics hint that transformations in user behavior should be related primarily to the different relationship online videos and live streaming construct between online users and activists on the ground (RQ 3).

Conclusion

This study assessed the hypothesis that online videos and live-streaming trigger different forms of connective activity. To this aim, a quantitative analysis of the user activity taking place on the Digital Smoke Signals Facebook page was conducted. As discussed, findings substantiated the starting hypothesis. However, rather than a definitive concept of connective activity, the obtained results call for more systematic investigation. The existence of significant differences in the connective activity triggered by synchronous and asynchronous forms of online audio-visual communication highlights three main directions for future research.

First is the need to develop a typology of connective action. As indicated by Bennett and Segerberg, the simple recognition of digital media as organizing agents, while fundamental, is far from satisfactory. If digital architectures and user interfaces organize user activity, then the meanings these systems construct in terms of user empowerment, possibilities of action, social engagement, and even bottom-up redefinition of the platform’s functions represent significant variables. From this perspective, a typology of connective action should enable the mapping of user activity in relation to different digital outlets.

Second, the presented findings indicate that, on the same digital platform, different forms of communication trigger different forms of connective activity. Accordingly, within a defined typology, variations of user behavior in relation to the enabled forms of communication should be investigated. This is particularly important because user activity represents the core business of social networks, and consequently, changes in the connective activity are likely to influence the development of platforms in both technical and economic terms.

Third, the socio-political implications of the interconnection between connective activity and digital platforms (and companies) deserve scholarly attention. While the emancipation from mass media marked a radical change in the global information economy, such opposition may not be sufficient to understand today’s transformations in the digital landscape. The presented data not only indicate that users personally engage in distant witnessing but also that they collectively constructed its political value through intense real-time interaction. Accordingly, this study hints at a structural change in the ways networked communication and political participation is experienced in the digital age.

Looking at the examined case through the lens of these challenges, it can be concluded that live streaming on SNSs redefines the act of witnessing as a networked, collective, and political experience. This transformation relies on three key points: the introduction of live streaming on Facebook (i.e. digital architecture), the users’ high reactivity to live feeds (i.e. form of communication), and the collective negotiation of the value of such experience (i.e. political relevance). Accordingly, these findings invite further exploration on how the role of the user is constructed through the mediatized occurrence and, in turn, how such a position is renegotiated (e.g. accepted or rejected) by the community of users engaging in connective witnessing.