Content uploaded by Ori Schwarz
Author content
All content in this area was uploaded by Ori Schwarz on Nov 25, 2017
Content may be subject to copyright.
1 Schwarz & Shani / Culture in mediated interaction
The definitive publisher-authenticated version was published in American Journal of Cultural Sociology 4(3): 385–421.
Culture in mediated interaction:
Political defriending on Facebook and the limits of networked
individualism
Ori Schwarz
Bar-Ilan University
Guy Shani
Tel Aviv University
Abstract
During the 2014 Gaza war Facebook became a central arena for moral/political boundary work for Israeli
users, resulting in unusually high rates of politically motivated tie dissolution. Cultural criteria were thus
applied to restructure and symbolically cleanse social networks. We analyze Facebook’s visibility-
structures, interview data and public posts to explore this phenomenon. Studying Facebook interaction
reveals cultural mechanisms used offline to sustain heterogeneous social networks and facilitate interaction
despite differences—group style differentiation between circles, differential self-presentation, and
constructing imagined homogeneity—whose employment is impeded by Facebook’s material design. This
case of materiality-informed value homophily introduces materiality to the sociological understanding of
the interrelations between culture and network structure. Interviewees reported dissolving ties following
their shock and surprise at the political views and sacrilegious expression styles of their Facebook friends.
We demonstrate that their shock and surprise derived from Facebook’s design, which converges life spheres
and social circles and thwarts segregation of interactions, group styles and information. Rather than
disembedding individuals from groups within the ‘networked-individualism,’ it makes individuals
accountable for their statements towards all their social circles. In dramatic times, this collapse of
segregation between life-spheres, affiliation circles and group styles conjures Durkheimian sociability and
symbolic cleansing despite commitment to pluralism.
Keywords
boundary work; facebook; group style; imagined homogeneity; materiality; political homophily
In the last decade, the social network site Facebook has turned into an important medium of the social, a central
arena for social action and interaction. Over one billion active users use it to keep informed about the lives of
friends, family, acquaintances and colleagues they include in their list of ‘Facebook friends’ (henceforth F-
friends), and keep communication paths open. Facebook is associated with heightened levels of all kinds of
social capital—bridging, bonding and ‘maintained’ (Ellison et al 2007, 2011). However, Facebook is not
merely a space of interpersonal ties, but also a political sphere used (especially in dramatic times) for political
expression, debate, propaganda, mobilization, activism and organization with impact on dramatic political
events throughout the globe (e.g. Castells 2011, Earl and Kimport 2011).
Two main narratives or theses organize scholarly and popular discourse on this process. The first is
the 'networked individualism' narrative, according to which Facebook represents the demise of well-bounded
national, ideological and cultural groups in favor of dynamic, personalized and loosely-bounded networks
(e.g., Rainie and Wellman 2012). According to the second thesis, the more networked individuals receive
information through networks of like-minded ties and personalized rather than national media, the less they
2 Schwarz & Shani / Culture in mediated interaction
are exposed to different opinions and worldviews, which can no longer penetrate their “echo-chamber”
(Sunstein 2007).
In this article we explore a case that does not sit well with these accounts. During the 2014 war between
Israel and the Palestinian Hamas regime in the Gaza strip (“operation Protective Edge”), an extraordinary
number of Israelis have engaged in dissolution of F-friendships following political online comments and
activity (John and Dvir-Gvirsman 2015). In what follows we present an account of this phenomenon, which
we term political defriending. Elaborating on recent dramaturgical accounts of Facebook interaction (e.g. Boyd
2014, Hogan 2010), we seek to explain political defriending through an analysis of Facebook’s design and the
way it structures visibility and accessibility to interactions and information. By directing our sociological gaze
to failures to sustain heterogeneity on Facebook, we may gain insights on cultural mechanisms that sustain
heterogeneity offline.
We maintain that the political cleansing of egocentric networks derived from the material-algorithmic
design of the website. Facebook’s design offers users a single platform to communicate with members from
different social spheres (professional, friendly and familial ties, strong and weak, past and present), and share
information with all these spheres simultaneously. Facebook thus represents a structural convergence between
the private and public/political spheres and more generally, between different specialized social spheres in
which individuals are embedded and their corresponding social networks. This structure represents a sharp
break from the notion of the modern self and society as conceptualized by Georg Simmel—a society consisting
of multiple isolated social spheres and a personality defined by its embeddedness in a unique composition of
spheres, exposed only partially to each (Simmel 1950, 1955).
Facebook’s design impedes the efficacy of strategies and mechanisms common in offline social life,
which sustain social interaction, solidarity, and perceived homogeneity despite conflicting political views and
orientations. By impeding mechanisms such as switching between interaction ‘group styles’ (Eliasoph and
Lichterman 2003), audience segregation and strategic presentation-of-self (Goffman 1959), Facebook
facilitates (at least in the case we studied) not the rise of a hypermodern networked sociality, but rather the
return of bounded communities that in times of crisis defend their moral boundaries and demand their members
to publicly demonstrate loyalty. While Facebook’s structure renders political talk a high risk form of
interaction (Thorson 2013), actors in different cultural contexts employ different strategies: while in the U.S
this high risk mainly leads to avoidance from engaging in political talk on Facebook (ibid.), Israeli users often
turned to political de-friending.
After a short methodological section, we offer brief reviews of the scope and patterns of political
defriending during the 2014 war; and of Facebook's algorithmic structure that regulates interaction and
visibility. Then we present our data on boundary work during the war, map out the motives behind political
defriending and discuss the role of style in moral boundary work. In the following section, we show how
Facebook’s structure contributed to defriending dynamics by posing a challenge to the construction of
imagined homogeneity. We pay special attention to the influences of these dynamics on Jewish-Arab ties and
on workplace persecution of Israeli-Palestinians. Finally, we discuss two wider issues—placing the network
society within meta-narratives of modernization; and introducing the role of technology into the debate on the
mutual constitution of culture and social networks.
DATA AND METHODOLOGY
The analysis relies on three datasets: exploratory interviews with 13 Israeli-Jews who engaged in political
defriending during the war; 261 public Facebook threads consisting of 4,897 posts and comments, in which
Israeli Facebook users discuss concrete defriending instances and the ethic and logic of defriending in general;
and 20 case summaries of workplace persecution of Israeli-Palestinian employees due to their Facebook
activity. These three empirical qualitative datasets join a structural analysis of Facebook’s infrastructure; and
a collection of news reports on shaming and persecution of Facebook users for their political Facebook activity
during the war.
Interviewees were diverse in terms of age, political orientation, geographic region, ethnicity,
occupation, education, and seniority and intensity of Facebook use. The interviews lasted between 30 and 90
min and gathered data on patterns and experiences of Facebook usage before and during the war; the political
3 Schwarz & Shani / Culture in mediated interaction
orientation of the interviewees and their political engagement online; and on concrete cases in which they
defriended others or were defriended, which is our main focus. The focus on detailed accounts of concrete
cases was used in order to avoid merely reproducing native theory and generalizations in our own account. In
some cases we also read the original conversations that led to defriending (these were not always still accessible
at the time of interview). We coded interviews to identify patterns, themes and mechanisms. These data offered
an initial support and demonstrations for our formal-structural analysis; and allowed us to refine it while giving
us a glimpse into the phenomenologies of political defrienders.
Building on the rich phenomenological data gathered in the interviews, we utilized Facebook’s search
engine to search for public posts in which users discussed defriending. Defriending was often a public act:
users often declared whom and why they defriended; and defended both the moral legitimacy of defriending
and their right to make provocative political statements. These public discussions were thus important sites by
themselves for the study of political and moral boundary work. Search queries were submitted from the
Facebook accounts of four different users with different political orientations, to compensate for the political
bias inscribed in Facebook’s social search, which gives precedent to posts closer to the submitter in the social
graph. We collected 4,897 statements in 261 threads and coded these data for the patterns identified in the
interviews, while adding codes for new patterns that emerged.
Two of the interviewees referred to defriending of Israeli-Palestinian coworkers and workplace
political tension. Persecution of Israeli-Palestinians employees for their Facebook activity was also repeatedly
discussed in our collection of news reports. We thus contacted a workers’ rights and advocacy nonprofit
organization, who supplied us with 20 anonymized case summaries of Facebook-related workplace
persecution, as reported by the prosecuted employees.
These three independent datasets shed light from different perspectives on the issue at hand, allowing
us to elaborate our structural analysis of the cultural logic of political defriending.
BACKGROUND
Dramatic Times
This week I bade farewell to yet another friend: my Facebook posts made her pour fire and
brimstone on me, which I just couldn't stand. What is a loss of friendship compared to the loss
of life, which turned into such a natural part of our lives? But life is not a scale. Technically it
was very simple, “unfriend,” a single mouse-click and that's it. Actually, our lives have been
intertwined for half our lives. A friend with whom I've shared so much happiness and pain and
love. A friend who turned into a part of me, a part which is now gone with all the suffering,
happiness and pain that were part of my own life. As always, the wrath poured over me was not
because I called for ending the war. As always, what made her blood boil were those pictures
of horror, this pain for the suffering of the wo/men of Gaza and their children (Lavie 2014).
Left-wing activist Ruti Lavie is not a typical Israeli: only few Israelis are as engaged in politics and as
radical in their critique as she is. Yet the moment she describes—the dissolution of a social tie (apparently, a
rather strong one) as a result of political activity on Facebook—is anything but unusual. War times are
generally prone to activation of boundaries and collective identities (Brubaker et al 2006), and our data
demonstrate that polarization, activation and dramatization of symbolic boundaries also characterized the 2014
war. Groups—both the nation, in the name of which fighting was carried out, and other moral communities
that traverse national boundaries—summoned their members, demanding that they publicly profess their
membership and defend the group's moral boundaries. The private-public and widely accessible arena of
Facebook was a main arena in which citizens followed this imperative.
Most of our interviewees (left-wing and right-wing alike) reported they have intensified their use of
Facebook during the war and perceived their wartime Facebook activity as an act of citizenship. They mainly
used the website to get news (including from alternative news sources), exchange war-related experiences,
opinions and comments, and even support the state's official propaganda endeavor. Communication on
Facebook has been nationalized, that is, the mundane was suspended (Smith 2005). Simple mundane posts
were sometimes perceived as disrespectfully prioritizing the private over the collective: thus, one interviewee
4 Schwarz & Shani / Culture in mediated interaction
deliberately avoided posting photos from a night out, whereas another defriended a boutique owner for posting
a commercial post unrelated to the war. While the usage of Facebook to share political statements is not
uncommon in Israel beyond times of crisis, during war the share of these contents in users’ feed surged.
1
It is
thus unsurprising that during the war Facebook turned into a space for intensive moral-political boundary work
among Israeli users.
This boundary work had structural consequences: Israelis of different political orientations politically
cleansed their egocentric social networks. This phenomenon was anything but marginal: a survey among a
representative sample of Israeli Facebook users conducted a week after the war ended found that 16% have
either ‘unfriended’ or ‘unfollowed’ (henceforth, defriended)
2
F-friends during the short 50-day war period
(John and Dvir-Gvirsman 2015). These levels are extraordinarily high in both local and international standards.
Although unfriending requires only “a single mouse-click,” it is not a common practice. In the US, only 18%
reported to have ever engaged in political defriending (Rainie and Smith 2012). In Israel, a similar rate was
reached in only 50 days. Furthermore, a third of the Israelis who engaged in political defriending during the
2014 war have never before defriended anyone for any reason (John and Dvir-Gvirsman, 2015), including non-
political reasons such as the common practice of defriending past romantic partners (Smith and Duggan 2013,
Sibona and Walczak 2011).
Logistic regression conducted by John and Dvir-Gvirsman shows that the probability to engage in
political defriending was not influenced by gender, age, income, religiosity level, or political orientation (left-
vs. right-wing). While ideological extremity does show positive correlation with political defriending,
3
even
among centrists nearly 10% engaged in political defriending.
Among our interviewees and posters, defriending patterns were diverse: some ended a single F-
friendship, others defriended dozens or even hundreds; many were moved by spontaneous emotional reactions
to political posts, while others defriended their network calculatedly and systematically. Whereas the weakest
ties were unsurprisingly most likely to be defriended, a substantial share of the defrienders ended strong
(“close”) social ties (12% in John and Dvir-Gvirsman’s sample) or medium-strength ties (12% defriended
coworkers, 13% defriended students peers). Even family ties were defriended, although on a smaller scale. In
our data one poster even admitted abashedly having been unfriended by her mother, and another one recounted
being unfriended by her daughter. This mass political defriending remolded social networks and interaction
structures online and in some cases offline (e.g. affecting workplace relations).
Our data document multiple symbolic boundaries Israelis drew while defriending during the 2014
war—not only between war supporters and a small minority of opponents, but also, for example, along the
lines of the ethic-aesthetic boundary between respectability and unrespectable inciting discourse. Defriending
was an instrument in the contestation over the legitimacy of practices such as criticizing the Israeli government
and military during war; showing empathy to Palestinian civilians; or alternatively, wishing them death. These
political disputes were obviously not endemic to Facebook. However, we suggest that Facebook was not a
passive intermediary of ‘social’ boundary processes but a mediator that influenced these processes (Knorr
Cetina and Bruegger 2002; Latour 2005). Below we show how Facebook’s algorithmic design encourages
1
This subjective impression of our interviewees is supported by data from Vigo, a commercial company
monitoring discourse in social media in Israel. In their sample, which consists mainly on Facebook posts,
the absolute number of political discussions in July 2014 was approximately four times as high as in July
2015 (and more than eight times as high as in July 2013).
2
Both procedures block future exposure to posts by the defriended party without informing her. However,
there are some major differences: unfriending removes both parties from each other’s friends lists, while
unfollowing merely prevents exposure asymmetrically (the unfollowed F-friend remains informed of the
unfollowing party’s posts). While unfollowing can be easily undone, unfriending is irreversible (renewing
the tie necessitates sending a new formal F-friendship request).
3
Apparently people with strong political opinions are more likely to sort their social network for political
homophily. PEW data show that in the US moderates have more heterogeneous F-friends networks: they
are much less likely to agree with most of the posts they read (18%) than ardent liberals (52%) or
conservatives (45%) (Rainie and Smith 2012).
5 Schwarz & Shani / Culture in mediated interaction
extensive political defriending and increases the influence of culture
4
(as demonstrated in political statements,
styles of expression and their judgments) on social structure, that is, sociation networks.
The Structure of Facebook Interaction
For Goffman, “[a] social situation may be defined (...) as any environment of mutual monitoring possibilities
that lasts during the time two or more individuals find themselves in one another's immediate physical
presence, and extends over the entire territory within which this mutual monitoring is possible.” (Goffman
1982:167). Latour (1996) has aptly pointed at the contribution of the material environment to these
interactions: human interaction is framed by material objects such as walls. Without material aids it is hard to
frame and confine interactions, segregate audiences for different performances, distinguish between front-
stage and backstage etc. Facebook is indeed an interaction space where users monitor one another, but in the
digital world visibility and access to interactions are structured by code. The architecture of algorithms such
as Edgerank (which determines which items are presented to each user), ‘privacy settings,’ and interface design
play the role traditionally played by walls, locks and doors.
5
While the notion of materiality derives from matter, the growing recognition of the social power of
algorithms to operate in ways comparable to physical objects encourages scholars to treat algorithms in terms
of materiality (e.g. Beer 2009, Bucher 2012, Scott and Orlikowski 2015) and demand expanding this notion.
In this article we use “materiality” to refer to the power of objects (in this case algorithms and software design)
to enable and restrain action in a way which is not fully reducible to the representations of these objects (and
their associated meanings) in the consciousness of social actors.
Facebook interaction further deviates from the Goffmanian definition in two ways: first, it is scattered
in time and space; posted information is made accessible to an unknown number of people in different times
and places as determined by an algorithm. Facebook users usually cannot tell who their audience exactly is,
which makes strategic self-presentation highly difficult (e.g. Boyd 2014). Secondly, on Facebook interaction
amounts to the production of durable objects, since all the information (verbal and other) exchanged is
documented and can be made accessible to others (Schwarz 2011, 2012; Boyd 2014). For this reason, as our
data clearly demonstrate, users often hold each other accountable for every tiny action or remark, an
accountability level unusual for ephemeral talk. Thus, the presentation of self is not achieved through
performances in front of a bounded cotemporaneous audience but rather through an ”exhibition of the self”
(Hogan 2010), consisting of data-objects (such as conversation protocols, comments and photos) which outlive
their original context.
Facebook interaction consists of four main categories of action: posting, sharing, commenting and
liking, all of which produce durable objects. Posting means sharing content that is usually the starting point of
a conversation. The post (also called “status line”) may be a verbal statement, a photo, a video, a hyperlink or
a combination of the above. Facebook’s Edgerank algorithm presents each post to some of those allowed to
access it (which are by default the author’s F-friends),
6
who may then engage with it in three ways—likes,
comments or sharing. Each such engagement may itself be automatically reported by the algorithms to some
of the engaged user’s F-friends.
Commenting means contributing to a virtual conversation-thread that a post started. Commentators
cannot determine their audiences: their comment may be viewed by all those who have access to the original
post. Like is a durable, public token of support. Users may like posts, comments, or other digital objects (e.g.
pages of politicians, political organizations, cause groups or businesses). While often intended at the original
poster, likes are visible to wide, unknown audiences. Posts defined as “public” may also be shared, that is, re-
4
While Culture has sundry conflicting definitions, most contemporary schools share an understanding of
culture as non-universal shared patterns that are structured by knowledge (including both practical knowhow
and mental representations). While we do not give up the analytical purchase of using culture to
conceptualize individuals' interactions with themselves, our discussion of culture is inspired by and indebted
to Eliasoph and Lichterman’s (2003) treatment of culture as residing in intersubjective interactions that filter
collective representations (rather than in internal values).
5
On the politics of Edgerank see: Bucher 2012.
6
Users may change these default settings and share particular posts with particular segments of their F-
friends, but they rarely do so, possibly because of the cumbersome interface that renders it rather difficult
and time consuming. Users may also extend access to friends-of-friends or the whole public.
6 Schwarz & Shani / Culture in mediated interaction
posted by other users (with or without additional commentaries), thus gradually arriving at wide unexpected
audiences, many degrees of separation away. By traversing contexts, meanings may transform.
When a user logs into Facebook, she is presented with a ‘feed’ composed of a mosaic of conversations
(posts and their comments), each of them taken from a different social context and accessible to a different
audience (usually the original poster’s F-friends), in a way that blurs distinctions among them. Since audiences
mainly consist of passive participants, who monitor the interaction without making their presence noticed,
users cannot practically tell who their audiences are, and often fail to notice this fact (Boyd 2014, Thorson
2013). When a user likes or comments on a post by a like-minded friend, her engagement may be exposed to
her coworkers or to the friend’s relatives. Boyd (2014) conceptualized this phenomenon as “context collapse.”
Having to perform without knowing who your audience is renders Facebook interaction very different from
Goffmanian face-to-face interaction. Only the latter is well-bounded and allows participants to effectively
monitor the number and identity of their interaction parties, and hence to engage in strategic presentation of
self to segregated audiences. For Goffman (1961), this relative freedom of impression management, that is,
the mastery of one’s own persona, the power to decide what to present to whom and when—is at the core of
human agency. Total institutions threaten human freedom exactly because they eliminate this capacity.
Having reviewed survey data on the scale of political defriending during the 2014 war and having
presented the main structural features of Facebook interaction, we now move to the empirical data. We first
portray patterns of wartime interaction and boundary work on Facebook, and the main dynamics that motivated
defriending. We then focus on how these patterns were shaped by Facebook’s structural characteristics.
Finally, we show how these dynamics bear especially devastating consequences for Israel’s Palestinian
minority and for Jewish-Arab relationships.
FINDINGS 1: THE LOGIC OF DEFRIENDING
What makes people cleanse their social network? Interviewees mentioned particular statements that crossed
‘red lines,’ morally staining the speaker and tipping the scales in favor of defriending. This set of statements
allows us to identify the boundary markers around which political boundary work revolved during the war,
these symbols that simultaneously produce Durkheimian solidarity within groups and Weberian closure
towards those excluded (Lamont et al. 2015). Studying these red lines reveals the centrality of style, the surface
of interaction, in boundary work. Importantly, interviewees (right-wingers and left-wingers alike) valued
pluralism and openness to different opinions. Ideological differences were rarely considered a sufficient reason
for defriending. However, each camp considered some forms of expression as offensive and unacceptable
sacrilege and some words as taboos, which evoked strong emotional reactions and justified defriending.
Boundary work took place not only through the almost-invisible act of defriending, but also through
public declarations. Our data include dozens of public posts, in which users publicly mark their red lines,
stating that they would defriend anyone who crossed them, or ask these populations (“whoever believes the
state of Israel behaves unfairly”; “whoever uses the phrase 'death to Arabs',” “whoever cries over Gaza” etc.)
to defriend themselves. Other posters publicly presented their red lines by reporting concrete cases in which
they defriended and publicly defending their decisions. Through these declarations and the dialogues that
followed them users engaged in public display of and negotiation over the boundaries of legitimacy and of
their moral communities.
The schema of boundaries that emerge from our data is organized around two main axes:
civility/rudeness (respectability versus vulgarity) and loyalty to the nation versus cosmopolitan universalism.
However, defriending is not a mere application of abstract typologies but rather an event, a contextually
embedded emerging interaction. Thus, below we go beyond mapping out defriending criteria and also dwell
on the phenomenological-emotional dimension of defriending events.
Almost a fifth of the public threads in our data discussed the value of pluralism: users took pride in
their pluralism and their “self-restraint,” praised the self-restraint of others who avoided defriending and
criticized those who did not (“progressive” liberals in particular were often accused of “hypocrisy” for
defriending). Some users were apologetic about defriending (“I’m not proud of it”), others reflected on their
moral dilemma: while defriending was construed as a natural urge that saves one “pain,” it threatens the
collective self-image of leftists as progressive (e.g. “It’s so depressing, why should we inflict this pain on
7 Schwarz & Shani / Culture in mediated interaction
ourselves? But then I think, well, we should listen to the other side too, we are progressive, aren’t we? But
then it becomes clear that I don’t want to belong to the same people as them. What can I do?”).
Most interviewees described defriending as a spontaneous reaction to unpleasant encounters with texts
or visual images that have angered, appalled them, or aroused experiences of emotional overload. The contents
that triggered these emotional reactions differed across the political spectrum. Interviewees and posters from
the left mentioned racism, support of martial and political violence, support of collective punishment, support
of boycotting Arab businesses, McCarthyism and treason accusations, and schadenfreude for (or “taking
pleasure” of) Palestinian suffering; whereas right-wing interviewees and posters mentioned comparisons
between IDF and Hamas or Nazism, posting on mundane matters during war, paternalism, support of and
identification with the Palestinian side and its suffering, apathy to the Israeli pain, and critique against the state,
Zionism and the IDF. As one user passionately remarked, “the army is above all, and our soldier’s blood is
sacred sacred sacred,”
7
asking those who disagree to defriend him as he feels ashamed of “sharing the same
blood with them.”
All interviewees, however, also expressed some tolerance towards contested contents. As citizens of
a country with a democratic ethos, most stressed that they are “open to different opinions” and stated their
belief in the right to hold different views of their own (most gave examples). Hence, a closer look reveals that
what characterizes the posts that motivated defriending is foremost their style, their symbolic and emotional
significance. In justifying acts of defriending, interviewees from both sides referred not only to the content of
posts but also to the manner of expression and the timing, describing defriended individuals as either vulgar
and barbarous or provocative and inciting.
David
8
(male, right-wing), who characterized himself as someone who “makes very extreme
comments,” passionately asserted that he can tolerate all viewpoints as long as an appropriate manner of
expression is preserved. Alongside cases in which he got into heated debates with other users ending in the
exchange of harsh language and sometimes defriending, he also shared with us stories in which sticking to
civil interaction has enabled mutual understanding despite disagreements: “it is about form, you know. If you
present things by starting with death wishes and that sort, so you’ll get the same response in return. But if you
present things in a way that is like explaining (...) without wishing all sorts and without swearing (...) then it’s
like a political argument. We’re arguing now. Some action. That’s nice.”
Rules regarding expression styles did not signify a universal standard (like clean language), but were
group-specific. For different groups, specific words and images acted as shortcuts to emotional meaning
(Alexander 2008), the surface level which activates collective schemes and organizes boundary work.
Sharon (female, left-wing) elaborated on her disgust of the expression style of one of her F-friends
(the first one she defriended), while also commenting on the relationship between content and style. She paid
special attention to the expression “Yalla udhrub.” This Arabic phrase, close in meaning to “give it to them,”
carries Orientalist connotation for its Arabic origin.
This guy is an ultra-orthodox [Jew]. ‘Yalla udhrub’ [laughing], it just drove me mad. It’s like
(…)’stick it to them’ plus that word, it was awful. So this ‘udhrub’ was unnecessary. You
know, I can be…. I very much respect… people are entitled to have opinions different from
mine, unfortunate as it might be [laughing] (…). You have people who really think that the
solution is war, and it’s… it’s hard to hear it… [but] there’s something brutish in that “yaaaa”
[growling], that you know [the interviewer repeats her growling], exactly! ‘Udhrub’ [talking
in a crude manly tone], for me it was…so that was my first unfriending…
Interviewer: So it’s not the content but the tone of it?
Sharon: No, no no, it’s also the content, I mean, if somebody would write “let’s donate 100
000 NIS to an animal shelter, udhrub,” I won’t unfriend him.
Interviewer: But had the same content been said differently, then maybe… Sharon: No, no
no, you can’t separate content and style, I mean the tone add another dimension to the content,
I mean, it’s one thing to say “I support a military solution” and another thing to say “Yalla
udhrub, lets beat the crap out of them,” it’s not the same content, it’s not [just] style.
7
A phrase borrowed from a Jewish prayer, originally referring to the sacredness of God.
8
All interviewees’ names are pseudonyms.
8 Schwarz & Shani / Culture in mediated interaction
Left-wing interviewees and posters widely objected calls like the one above and similar ones such as
“death to Arabs” or “wipe out Gaza.” They regarded them as racist and as unjust calls for collective
punishment, which offended their liberal individualist notion of justice. Yet, these statements are often not
actual plans or calls for action. These expressive declarations during wartime are part of a group style. This
ritual talk does not simply reflect opinions located within individual subjects; it rather expresses and revitalizes
solidarity within some sections of Israeli society,
9
and while atrocious to others, for whom it takes a completely
different meaning, it should not be taken too literally. Racist statements are not always good predictors of racist
actions (Eliasoph 1999; Jerolmack and Khan 2014:182) and actors are often not as coherent as sociologists
expect them to be (Swidler 2003). In our case, unlike most affluent liberal Israeli leftists, some lower-class
right-wing interviewees had actual friendly relationships with Arabs, which did not prevent one of them to call
“death to all Arabs” and wish them “all sorts of cruel deaths.” Our two interviewees who made such statements
also expressed some empathy to the Palestinian side. Not everyone who engage in this ritual talk and say they
wish ”death to Arabs” or the destruction of Gaza would necessarily act on these expressions had they had the
power to do so.
10
Many of the elements shocking to left-wingers were perceived not just as extreme but also as vulgar
and lowbrow. We have seen how Sharon described the call “Yalla Udhrub” as “brutish.” In another example
Orna (female, left-wing) defriended a high-school friend who wrote, “like in a totally shallow way, ‘a good
Arab is a dead Arab’.” Others explicitly referred to the vulgar inarticulate style of a right-wing Israeli rapper
as a justification for defriending those who shared or liked his posts. One poster reported having defriended
“two people I know for 25 years because their status lines sounded like the bleachers in Teddy,” referring to
Jerusalem’s infamous right-wing football fans, thus associating right-wingers with lowbrow culture. These
classed undertones repeated in accounts of middle-class left-wingers, as one poster suggested she defriended
in order to have a “boutique account” once again. Form and content, ethics and aesthetics, classed respectability
and politics are closely intertwined.
Right-wing interviewees described the style of left-wingers as inflammatory and as provocative for
the sake of being provocative. They identified as provocative insults against national symbols such as the flag
or “our troops” (this included a derogatory style or describing IDF soldiers as “Judeo-Nazis”). Like in the case
of left-wing defrienders, it was hard to separate content from style. However, right-wing users also stressed
the timing and publicity of the statements. Making these critical statements during the sacred time of war made
them especially provocative and intolerable. One interviewee said, “I mean, be a leftist, what do I care? But at
this time, it did, it struck a nerve.” For right-wing interviewees critique against the Israeli state and army or
expressions of empathy for the other side during wartime were perceived as provocative by definition: “you
have those who know that now it’s really not the time, that it's annoying and oppositional and will hurt others,
and they will post it spitefully, in the name of freedom and all that…” (Rivka, female, right-wing). This was
all the more true for statements made on Facebook, which is perceived as a public arena, e.g., “(...) she told
me that it was her right to express her opinions (...) it's a democracy and she can say whatever she wants. But
I tried to explain it to her, O.K., she may express her opinion, but she didn’t have to do it in public…” (David).
Being summoned
This objection to the timing of statements is significant. It draws our attention to the role of style which goes
beyond classed distinction (vulgarity) to mark symbolic boundaries around sociopolitical collectives,
simultaneously producing social closure and solidarity (Lamont et al. 2015). Timing is important, since the
gap in interaction styles emerges more strongly during times of crisis. During emotionally laden wartime the
mundane is suspended (Smith 2005) and moral collectives summon individuals (Tavory 2016), demanding
them to choose a side. Facebook turned into a central arena for summoning of this sort. Right-wingers felt
compelled to defend Israel from online criticism, thus voluntarily contributing to the official state project of
propaganda (referred to in Israel as Hasbara, literally ‘explaining’). One of the interviewees told us that during
the war he was in Thailand for a vacation, but was still “very much connected.” Through Facebook he
volunteered to Hasbara to foreign F-friends, battled their alleged “Anti-Semitism” and cursed Islam. This
9
In this sense the 'extreme' statements that users often encountered in Facebook during the war cannot be
reduced to 'positions' at the level of individuals: they are produced by groups and their interaction styles.
10
Similarly, Israeli liberal leftists often ritually talk about immigration plans to perform their identity as
cosmopolitans whom nothing attach to Israel, without necessarily acting upon their statements.
9 Schwarz & Shani / Culture in mediated interaction
enthusiastic activity is comparable with being called for reserve duty to defend one’s homeland. Left-wingers
too spent more time on Facebook during the war, summoned to fight “racism.”
Timing was also important for left-wing interviewees. The subject-positions to which they were
recruited was defined by the refusal to be summoned by the national collective call to unity, ‘pulling together’
and Hasbara; refusing "vulgar" collectivism that threatens individual "freedom" and "criticism." Hence,
expressions of ‘pulling together’ (referred to contemptuously as "national erection") were "unbearable" for
left-wing interviewees—e.g., "I couldn’t stand it, as if anyone not having a crazy erection, again, is a traitor"
(Liora, female, left-wing).
Ironically, for structural reasons discussed below, it is particularly during these times that Facebook
users are more likely to be exposed to political contents and styles of expression that deviates from the very
symbolic boundaries they try to defend. It is thus unsurprising that a high rate of Israelis throughout the political
spectrum have engaged in defriending. Encountering contents that function as boundary-markers may
contribute to emotional overload during a time already characterized as emotional and stressful. On the
phenomenological level it was not merely the exposure to contaminating contents but the uncontrolled
overload of exposure that was a key dimension of the experience leading to defriending. Posters frequently
used phrases such as “I’m exhausted,” “Enough!!!,” “I can’t take this shit anymore.” In the interviews the
phrase “I didn’t know how to deal with it” recurred in minor variations. See how Orna described her decision
to hide all F-friends making racists comments such as "death to all Arabs:”
...so it’s not OK to say these things and I don’t wanna see it either. It’s simply incitings, and I was also
torn at this time, because [her partner] was in Gaza [as a soldier] for a month, I wanted to die, I haven’t
slept for a month, not just because I was worried, I was awfully scared, so why do I need it? You know,
I didn’t know how to handle it.”
Symbolic contamination
Boundary-marking contents users encountered on Facebook were not only agitating; they were also
experienced as a symbolic contamination. Posters and interviewees recurrently used a cornucopious amount
of metaphors, borrowed from the fields of bodily fluids ,worldly cleanliness ,religious purity ,and warfare.
They described offensive posts and posters as “dirt,” “shit,” “piss,” “filth,” “vomit,” “ticks,” “air poisoning,”
and “Gestapo”; while defriending was metaphorically characterized as “screening,” “cleaning,” “cleansing,”
“burning chametz” (a Jewish religious cleansing metaphor), “flushing down the toilet,” “performing an
enema,” “friendly fire,” “my private Protective Edge operation,” “operation Iron Fist,” and even “ethnic
cleansing.”
A unique genre of status reports emerged, similar to the body counts published in the news, in which
users proudly reported ‘casualties’ (e.g. “223 unfriends, 396 unfollows (...) we’re back to the normal routine
of dogs, cats and food photos.”)
These acts of cleansing were performed to avoid interaction with both contaminating contents and
contaminating people (“I don’t have racist friends”). 'Sacrilegious' statements are moralized in Tavory’s
(2011) sense: while sociologists may associate them with intersubjective group styles or codes, agents often
perceive them as indicative of the stable essential moral character of those who make them. Many users
interpret these markers as indicative of the essence of their F-friends, and consequently, of the groups and
communities to which they belong. It is because of this moralization that exposure to interactions in different
styles may threaten social ties.
FINDINGS 2: FACEBOOK’S STRUCTURING OF WARTIME INTERACTION
The political identifications and boundary work portrayed above are surely not endemic to Facebook.
However, defriending was not merely a reflection of pre-existing identities. The ubiquity and intensity of
Facebook political conflicts and defriending was influenced by Facebook’s structure and ethics, which
structured interaction and precluded the employment of cultural mechanisms common offline.
The ethic of transparency and its structural anchors
10 Schwarz & Shani / Culture in mediated interaction
Above we describe the algorithmically inscribed difficulties in identifying and segregating audiences and
engaging in differential impression management on Facebook. These difficulties are not a coincidence:
Facebook’s founder and CEO, Mark Zuckerberg, declared the era of “transparency” and “sincerity” that will
obliterate the human capacity to designate different self-presentations in front of different audiences (quoted
in Rainie and Wellman 2012:141). This reflects not only a philosophical stance but also a clear commercial
interest not to reduce the volume of audiences exposed to a typical Facebook post.
This ethic of authenticity is well reflected in our empirical data. Many interviewees and posters
adhered to an ideal matching Zuckerberg’s ideology, striving to act authentically and fully express their whole
personality on Facebook while avoiding self-imposed censorship. One poster defriended an F-friend because
he requested that she avoid replying publicly to his political posts. “There can’t be one private truth and a
different public truth, because then it isn’t the truth,” she explained in her post. In other cases, living up to this
ethic bore a heavier price.
Yuval (male, left-wing) regularly posts on current events, sometimes several times a day. “During the
war I posted all day long.” His attitude was non-apologetic: “I believe my feed is like my private home” (this
metaphor repeated in data), and therefore “In my home I publish whatever I want, however much I feel like
(...) I always say: hide me, block me, unfriend me, do whatever you want—I publish as much as I feel like,
whenever I feel like it, and about whatever I feel like...” Many posters made similar declarations publicly. Yet
this ethics has its difficulties since offline, even the most “authentic” people usually adapt their self-
presentation between contexts. For Yuval, this ideal of free speech was especially challenging. Raised in a
right-wing family in Israel's socio-geographic periphery and currently living among a leftist milieu in the
country’s urban center, his Facebook posts were written to a mostly sympathetic audience. Yet a small portion
of his audience, including his relatives, viewed his posts as crossing the line. This stirred conflicts with his
family and eventually his aunts and other relatives defriended him. “I had a problem with my family, because
my mother has three sisters and they all unfriended me, it got so far that they didn’t wanna talk to me because
my cousin was [a soldier] in the war, so how come I talk against him?, and I tried to explain to them that it’s
not against him, on the contrary, I want him to come back home, I … I want him to live. But no. ‘You are
against him’.” His relatives claimed that one should not criticize the war while it is ongoing and there was “a
real family bedlam (...) it reached a point that my brother called me, saying, ‘listen, if tomorrow he dies at war,
how would you look your aunt in the face?’ like, it got so cynical that I couldn’t win whatsoever, like, what
am I supposed to say? There’s really nothing I can say.” Following his Facebook posts, his three aunts
unfriended him. One of them, more left-leaning, called to apologize, telling him that “even if you’re right and
we’re not the just side, 'it’s still our people', and all this rubbish, so we must support our people.” His brother
did not unfriend him: instead, he phoned him regularly trying to convince him to stop posting critical posts:
“two phone-calls a day, about what I write [on Facebook], and why I do it, and that I should stop, write less,
stop writing, and ‘your mom doesn’t feel well’, ‘she has heart pains’.” His brother and aunt pressured him to
“shut up a little” and “not say everything.”
Since he lives in Tel-Aviv, far away from his family, they do not meet too often, and when they do
meet, it happens “very rarely” that they engage in political debates. However, on Facebook he was rendered
accountable to his family for his political conversations with his friends in a different part of the country.
Similarly, many posters in our data publicly invited all F-friends who cannot accept them the way they
are to unfriend them, stressing that “I’m not going to change.” Their F-friends often praised them for this
demonstration of integrity. Others took pride in avoiding self-censorship, which one user described as “a trial
to keep my independence.” Not speaking your mind is perceived as a threat to one’s moral integrity. Others
celebrated their freedom of expression on their “private” wall, which was repeatedly compared to “home,” and
used the fact of being defriended as a badge of honor (e.g., “Congratulate me, yesterday [I was defriended for
the first time]. I’m happy to pay this price for my freedom to express my opinions.”)
Zuckerberg’s ideal of authentic unrestrained self-expression that Yuval and others followed on
Facebook reverberates indeed a highly common sensitivity in late-modern societies. However, for Yuval it
was medium-specific. He has never posted political posts in his family WhatsApp group (a cellular application,
which—unlike Facebook—is used for sharing messages with segregated, well-bounded groups). On
WhatsApp he never went further than asking them to avoid racist comments, and even on this point he gave-
up easily. The WhatsApp group was defined in advance as familial—belonging to a specific social sphere,
which by definition concerns only particular aspects of each member’s life and personality, and where
11 Schwarz & Shani / Culture in mediated interaction
interaction is regulated by clear local rules. Facebook, on the contrary, asks its user to create an exhibition-of-
self that fully represents her, a single symbolic representation to be presented to all one’s social circle. For this
reason, avoiding political debates may be experienced as obvious and unproblematic in WhatsApp groups, but
as an unforgivable concession on Facebook, an arena which is simultaneously highly public and deeply
personal—“my home.”
Ideology embedded in materiality influences not only those who consciously subscribe to it. The
undifferentiated sincere self ideology is subtly embedded in Facebook’s design, which encourages users to
share information—be it holiday family photos or political statements—with wider and less specialized
audiences than in face-to-face communication. Sharing the same conversations with highly different audiences
(family and friends, coworkers and remote acquaintances), they often present information they would not
present to some members of their audience face-to-face, often without fully acknowledging the consequences.
As a result, many interviewees reported learning from Facebook about previously hidden aspects of
their F-friends’ lives, including their politics, values and sexual orientation. This challenge is often framed in
terms of privacy, i.e. sharing some information with some people while withholding it from others. We suggest
that the challenge is actually wider and may be more accurately framed as a challenge to the segregation of
group styles (Eliasoph and Lichterman 2003)—controlling not merely access to sensitive facts, but access to
modes of expression, to the fabric of interactions which follows local rules and may change its meaning while
taken out of context.
The conflicts that led to defriending in our data often derived from the gap between perceived and
actual audiences, and the ensuing mismatch between performance and audience. Users who engaged in
Facebook-conversation with a clique of like-minded F-friends often failed to notice that this conversation was
exposed to F-friends from other life spheres. Adam (male, left-wing) had a coworker who published posts
blatantly attacking the leftists who “disseminate lies” and “defame the army,” yet he has never heard her (or
other right-wing coworkers) making similar statements at work, and was convinced “they wouldn’t do it that
way” there; in face-to-face communication “the most extremists would prefer to avoid it rather than get into
political arguments.” When he replied to a blatant post of his coworker by writing “do you remember that I
also see these feeds? You are talking about me,” she responded by sending him and other leftist coworkers a
private message, to “remind you that I love you very much.” At work she was more conscious of political
heterogeneity, and took it into consideration, yet while making online comments against the “leftists” who
demonstrate against the Israeli army, she failed to notice that some of these leftists were among her audience.
Untypically, Adam did not interpret her Facebook comments as revealing her true self, but instead as
“being carried away by a group I cannot see.” However, this did not make exposure any more pleasant. When
the same coworker later posted a racist caricature that hurt his feelings, he defriended her. Similarly, Inbal
(female, right-wing) avoided talking about politics at work, “to keep things pleasant,” yet on Facebook she
and her coworkers (Jews and Arabs alike) were exposed to each other’s political statements. During the war,
she defriended all her Arab coworkers, a case we discuss in detail later.
Ostensibly, users could have restricted themselves to statements they find suitable to a general
audience. Algorithm design may help us understand why this is often not the case. Users are presented with a
feed of different conversations which can be easily experienced as private conversations, as they are rarely
accessible to just anyone. Yet confusingly, each conversation is shared with a different, unknown set of users
that those who join cannot restrict. This dynamic is exacerbated by the asymmetrical visibility shaped by
Facebook’s algorithms: While in face-to-face interaction visibility and audibility are usually symmetrical,
algorithmic curation (Hogan 2010) produces asymmetry. User’s posts are seen by different users than those
whose posts are shown to them. For example, posts by highly popular users are usually presented to their weak
ties, but not the other way around. Thus, popular users may get false impression of their network homogeneity.
In addition, posts on contested issues often attract many comments and are consequently presented to wider
audiences than usual. Anat (female, left-wing) who has thousands of F-friends reported she had not been
exposed to political views that annoyed her at first :“I really don’t see anyone in my feed (...) those I do see
are usually those people with whom I have more interaction, that’s how Facebook makes it to be, it puts you
in an ever smaller bubble.” This changed however when she posted a controversial political post that received
multiple and largely violent comments.
12 Schwarz & Shani / Culture in mediated interaction
However, national cultural norms (reflected in users’ experience of their Facebook feed) also played
a crucial role in shaping users’ strategies to cope with the challenge posed by algorithmic design. Hogan (2010)
suggests that when contexts collapse users may restrict themselves to the lowest common denominator.
American and British users with multiple “distinct friend communities” are statistically more likely to engage
in self-censorship on Facebook (Das and Kramer 2013). Thorson (2013) found that among American “young
citizens” the ambiguity and fluidity of audiences and the consequent lack of clear group-specific interactional
norms often led to self-censorship and avoidance of political talk on Facebook.
11
While in the US Facebook
users felt out of line to make political statements (Thorson 2013), in Israel, where political talk is generally
more common and legitimate offline and online (and even more so during wartime),
12
politics could hardly be
avoided while entering Facebook. Thus, many Israeli users adopted a different strategy: rather than avoiding
altogether political talk (which they did in some offline contexts), they took some control of their designated
audiences by defriending, while taking the risk of being defriended themselves.
Some interviewees even described defriending as a strategy to repair the damage done to relationships
by the convergence of multiple audiences and life-spheres. They preferred to block or hide F-friends out of
fear that unsegregated Facebook communication would create conflicts that could affect the relationship. Thus,
a right-wing interviewee temporarily blocked her (left-wing) brother and sister to avoid insulting them. In two
other cases interviewees explained blocking F-friends as pre-emptive strategy aimed at saving the relationship
by avoiding exposure to statements that would threaten their positive opinion of the other party. One of them
blocked a coworker, because “when it’s clear for you that the person is a racist (...) it’s hard to see him at work
later”; the other blocked his “bigot” uncle, in order “not to let him have my opinion of him deteriorate any
further.” In these cases, defriending does not merely reflect the damage already done to a relationship, but also
used to minimize it by re-introducing segregated interaction.
Like and punishment
Writing original posts is but a tiny minority of Facebook activity, which consists mainly of engagement with
other users’ posts, most often by 'liking' them. This action is often carried out casually and unreflexively, with
meaning and function comparable to that of nodding in offline communication. Yet users are held accountable
for it: liking contested contents may be enough to justify defriending. A large share of the political defriending
documented in our data was reactions to likes, comments and shares. This happens because in some cases
liking could also mean an enthusiastic public endorsement of the liked contents, and audiences cannot tell the
subjective intentions beyond each like.
Offline, nodding plays a central role in symbolic interaction as a way to show interest and sympathy.
Nods may have implicit or ambiguous meaning, and they are aimed mainly at the speaker and occasionally at
other interaction parties. Nods are rarely monitored by those not present in the interaction. On Facebook likes
are the conventional and algorithmically inscribed way to show support or recognition. However, unlike
ephemeral nods, likes are documented, counted, and reported to F-friends of the liker, the liked and anyone
exposed to the post. This is another dimension of the incapacity to segregate audiences and the convergence
of social spheres.
Let's imagine a world in which a sociology professor nodding in agreement during a departmental
seminar would automatically send the presented paper to the parents who send their children to the same
kindergarten as the professor. Curiously enough, this is how Facebook works. Our professor may now be held
accountable for her critical academic worldview in front of the parents of her son's peers, who might feel
uncomfortable with contents deriving from a different social sphere. In Wellman’s (2002) terms, this structural
change is a movement towards bigger little-boxes, not towards networked individualism.
Thus, Inbal recalled that “some Arab guy I've worked with in a café liked something, and I saw it and
I said, 'I don't want to have him on Facebook.” She defriended him for liking these war-related materials,
11
However, this does not render Americans immune to context collapse in political and other morally
contested issues, as demonstrated by the public rage that hunting photos or private jokes published by
American Facebook users occasionally evoke while reaching beyond their intended audience.
12
While US political defrienders often defriended those who “Posted too frequently about politics” (Rainie
and Smith 2012), this motivation accounted for a modest 13% of defriending cases in Israel during the war
(John and Dvir-Gvirsman 2015).
13 Schwarz & Shani / Culture in mediated interaction
because “I just didn't want to see the likes he gives to all sorts of things.” While nods fade and cannot exceed
the original interaction, Facebook likes do exceed it and produce accountability for much wider and more
diverse audiences. Objectified nods—likes and even more so shares—are added to the exhibition of the self
(Hogan 2010) and are interpreted as taking responsibility for the endorsed materials;
13
or even for other
statements made by the original authors. Interviewees and public posters reported defriending those who shared
or liked posts by controversial authors or organizations, as endorsing the post was often interpreted as
endorsing the author.
Importantly, interviewees reported sharing and liking texts they would not have formulated
themselves, and being defriended for it. For example, on the war's second day Anat had shared a list of the
names and ages of Palestinians killed by the Israeli army on the first day of bombing, including young children.
The post she shared stated that the children were surely not terrorists, and expressed hope that “no further
names from any side will be listed.” The reactions were tough: more than 15 F-friends unfriended her within
a single day. It aroused a fierce debate with hundreds of comments: some accused her of “treason,” others
accused her accusers of “Nazism.” F-friends expressed their “disappointment” of her, and denied her right to
make such statements. Having thousands of F-friends, the post was exposed to a wide audience, and the
hundreds of comments, shares and likes encouraged Facebook's algorithm to keep presenting it to ever wider
audiences.
In the interview Anat claimed that she would not have made such a statement herself hadn't she
stumbled upon the post she shared. Her original posts during the war were indeed vague and ambiguous
expressions of frustration and despair without having herself classified or arousing antagonism: “I'm not sure
I would have consciously stepped into this filth out of choice (...) but even had I said something, I would have
said it differently, these were not my words (...) I would have done it more elegantly, more cautiously, with
the same message,” yet “once I pushed the button, I completely stood behind what was written there.” She
“didn't think of the consequences," having reacted spontaneously to a post that “touched” her. She was not the
only interviewee to claim she had shared posts of more blatant style than those she wrote herself. Sharing is
like quoting word-by-word someone else’s argument in front of a different audience, without making
accommodations. However, those sharing posts are held by their audiences as authors accountable for their
content and style.
Imagined Homogeneity
Interviewees’ experiential narratives were organized around a few key emotions: surprise, shock and
disappointment. A central source of this disappointment was the collapse of the ability to maintain an image
of one’s egocentric network and of the collectives to which one belongs as homogeneous. We know people
tend to overestimate similarities between themselves and their friends (McPherson et al 2001) and particularly
report higher levels of political homophily than actual levels (Knoke 1990). People also imagine themselves
as members of homogeneous sociocultural groups, the homogeneity of which is a discursive and cognitive
achievement. Richard Jenkins (2008, ch.11) suggested that group membership always relies on imagined
similarity, a “mask” or “umbrella” of homogeneity which disguises heterogeneity without necessarily
eliminating it. These groups are used by individuals as anchors in personal identification. Similarly, for Simmel
people look upon one another through the veil of typifications, as we can only know one another through
typifications based on group affiliations (Simmel 1910:380). Thus, interviewees were shocked to find out that
the imagined collectives and concrete egocentric networks with which they associated themselves were not as
homogeneous as they believed them to be. As one poster who defriended all her “racist” friends and remained
with a leaner network pondered: “what does it say about me that these were most of my Facebook friends?”
Posters and interviewees who could still perceive their networks as pure expressed pride of their
homogeneous imagined community, as F-friends were viewed as reflecting on one’s personality. One poster
boasted on his purely patriotic feed, another proudly declared he “couldn’t find any leftist friends to kick them
out,” while a third poster encouraged wartime defriending by saying: “Tell me who your friends are and I'll
tell you who you are: that’s the time to examine and weed out.”
13
Similarly, the objectification of social networks as F-friends-lists renders users accountable for their social
ties in one social spheres to members of others: some users were defriended for failure to unfriend polluting
F-friends.
14 Schwarz & Shani / Culture in mediated interaction
Interviewees and posters thought about their imagined communities (lifestyle communities such as
‘vegans’ or the ‘reggae community’ or ascribed categories such as Kibbutz-members) as moral communities
with shared values. These images relied on common stereotypes and on interviewees’ own interpretations of
the group’s moral ideals. Orna, who was raised in a Kibbutz, was shocked by the “racism” of Kibbutz residents,
which conflicted with the common image of the Kibbutz movement, historically rooted in the Israeli left: “I
was sure that all Kibbutz-members are left-wing Ashkenazi. Well, no! They are all racist, all of them terribly
hate Arabs, and that kind of stuff. So I was terribly surprised.” Similarly, a vegan interviewee was disappointed
since “when someone is vegan, you say, ‘well, animal rights, he must care’… you assume that he also cares
for humans, humans are animals too.” Right-wingers also reported disappointment, which may be interpreted
as deriving from similar sources—the collapse of their ability to imagine the national collective as at least
somewhat homogenous, unified around the people’s army during the war.
A common response was challenging the status of the defriended as authentic group members, blaming
them of merely superficial affinity with the group without commitment to its moral ideals. Thus, the revealed
heterogeneity redefined and sharpened group boundaries. Many right-wing posters characterized left-wingers
as traitors, or as not real Jews. One left-wing interviewee, belonging to the reggae community described
reggae-lovers who made right-wing comments during the war as “people who only like the music, and you
know, the smoking, but who aren’t really related to the very essence [of reggae].” He keeps on believing that
real reggae lovers are committed to pacifism, even after having found out that many community members were
not. Anat paraphrased the vegan slogans “justice, compassion, veganism” and “in suffering we are all equal”
while discussing the deviation of right-wing vegans from the imagined ideal of the group she used to hold:
“you reveal that we are not all equal in suffering,” “it’s impossible to say ‘justice, compassion, veganism’
when these justice and compassion do not apply to all.”
Imagined homogeneity and group styles
In order to understand this collapse of imagined homogeneity, we must pay attention to the cultural
mechanisms that maintain it in offline sociality. The capacity to imagine one’s egocentric social network and
imagined communities as homogeneous relies on a set of social mechanisms that regulate social interaction,
which Facebook’s material infrastructure renders difficult to employ. In daily life people maintain many weak
social ties, in which interactions hardly include any reference to politics: coworkers talk about work or
exchange office gossip, parents talk about the schools their children share, acquaintances talk about common
friends, celebrities, T.V shows, books, etc. The subjects of discussion vary depending on the interlocutor.
Political discussions were found to take place more often in strong ties, which are more resistant to
disagreements and where disagreements are often smaller, usually confined to the same political camp (Beck
1991; Morey et al 2012). Local ‘group styles’ preclude discussing politics in sundry contexts, including groups
whose activity has an obvious political dimension. Even in contexts where talking politics is legitimate, group
styles define the register that should be used for political talk: intellectual and detached or emotionally engaged,
light-hearted and teasing or utterly serious, inciting and dividing or stressing shared perspectives (Eliasoph
1998; Eliasoph and Lichterman 2003). Offline individuals may switch styles, discussing politics to different
degrees and in different ways in different groups to sustain solidarity. On Facebook, context collapse
challenges this strategy.
Again we can see the importance of ‘style,’ which goes well beyond classed distinction. Sticking to
group styles allow surface coordination and even deep solidarity despite ideological differences. Group styles
inhibit opening-up contentious issues, thus preventing disagreements from surfacing in many interactions.
They also regulate the ways in which interlocutors tackle contentious issues, allowing them to surface in a
pacified way that takes their sting away. Thus, as long as interlocutors are aware of the interaction’s context
and comply with the appropriate group style—which is usually the case in face-to-face interaction—local
styles often channel speech to support the maintenance of group cohesion and boundaries.
On Facebook, as a result of the incapacity to distinguish between audiences and their group styles,
group styles could no longer do this important job. Users lose their ability to imagine groups as homogeneous,
and could no longer assume that homophilous sort on one dimension produces homogeneity on others.
Veganism, love for reggae, being a Kibbutz-member or a past in the same youth movement are parameters in
a multidimensional space, which do not converge with other parameters: while the percentage of right-wingers
and left-wingers among these groups may deviate from the national average, they are far from being politically
homogenous. Facebook’s visibility structure reveals to users the unidimensionality of these communalities,
15 Schwarz & Shani / Culture in mediated interaction
whose members do not necessarily share a wider set of values. The way these users perceive and experience
their social worlds is thus brought in line with the structuralist sociological insight that in a world of multiple
network affiliations, homophily in one attribute often creates heterogeneity in others (Blau and Schwartz
1984:196). This is crucial, since the mechanisms of homophilous sort depend on the surface of interaction,
14
that is, on perceived similarity, which is necessarily a cognitive achievement (Brubaker 2004). Under these
conditions the umbrella of imagined similarity breaks.
This heterogeneity (rendered visible by Facebook’s visibility structure) often “disappointed” users,
and eventually led to defriending. In particular, Facebook’s structure of visibility exposes users to political
comments of F-friends, which are no longer regulated by local styles of groups they share. Even in the U.S
(where the legitimacy of political discussions on Facebook is lower than in Israel) 38% of the users report
having discovered new information about their F-friends’ political views through their posts (Smith and
Duggan 2013). Similar revelations repeated in our data and motivated defriending.
The salience of classification systems varies across contexts, and war obviously increases the salience
of national and political categories in evaluating similarity and difference at the expense of others (see
Brubaker 2004, Brubaker et al. 2006). Unable to maintain imagined homogeneity, users engaged in political
defriending to achieve actual political homogeneity. Political defriending applied political sorting even to ties
that were conceived based on other similarities, thus increasing political segregation within social networks.
After this war I don’t have any Arab Friends
While the ‘context collapse’ of political talk among Jews had in many cases somewhat limited offline impacts,
the access Jews gain to intra-national political talk of their Arab acquaintances resulted in many cases in moral
panic and persecution.
Israeli society is characterized by an almost complete national socio-spatial segregation, and ties between Jews
and Arabs (Israeli-Palestinians constitute 20% of Israel’s population) are few, weak and specialized (e.g. Falah
1996). Within this context, Facebook plays a dialectic role: on the one hand, it helps transforming narrow
specialized ties between coworkers in binational workplaces (e.g. call centers) or between student peers into
wider multifaceted ties. On the other hand, it exposed Jews and Arabs to the intra-national discourse of the
other party, with explosive potential, especially in times of war.
During the war Israeli media and workers’ rights organizations reported dozens of cases in which
employers fired, suspended or initiated dismissal proceedings against Arab employees for a wide range of
Facebook activities. These included texts accusing the Israeli army of killing innocent citizens and committing
war crimes (accusations also made by many left-wing Israeli-Jews); posting photos showing Israeli policemen
hitting Arab anti-war female demonstrators; using national Palestinian flags as profile pictures; blatant
caricatures presenting Israeli army and PM as vampires or Nazis; pro-Palestinian statements; and expressions
of joy at the death of Israeli soldiers and hope for further deaths. Some were persecuted not for posting original
materials, but for endorsing (liking or sharing) posts by others; or even for being tagged by others in political
posts.
This tide of persecution was directed at Arab businesses and employees at all levels: starting with
cleaning and maintenance employees, through an X-ray technician and an assistant branch manager in a retail
chain and ending with a hospital doctor. In some cases persecution was at the employer's initiative; in other
cases Jewish customers or anti-Arab activist groups demanded it (popular Facebook groups formed during the
war were fully dedicated to shaming Arabs for their “extremist” statements, pressuring their employers and
demanding their dismissal). Some persecution attempts were eventually blocked by courts, lawyers and public
figures. The Israeli higher education system was not immune to these trends: Israeli universities have warned
professors and students of making “extremist” statements, and one college punished Arab students for their
Facebook activity.
14
Merton, while coining the term “value homophily” in his seminal 1954 paper , (Lazarsfeld and Merton
1954), already suggested that this pattern resulted from the gratification from initial encounters with similars
and the ensuing incentive to maintain ties with them, thus anchoring homophilous sort in the surface of
social interaction.
16 Schwarz & Shani / Culture in mediated interaction
While most media reports focused on the formal sanctions, our data reveals the process leading to
them and the centrality of Facebook’s structure in those processes. Formal workplace persecution and informal
sanctions (defriending) were produced by the same mechanisms. Facebook hardly allows users with binational
F-friends networks to keep their political conversations within the confinement of their national community.
Facebook's embedded translation feature gave Jews some access to their F-friends conversations even when
conducted in Arabic. Jewish colleagues (and occasionally customers) reacted to their exposure to Israeli-
Palestinian intra-national discourse and communication on Facebook by both political defriending and
complaints to higher-level managers in demand for institutional sanctions, in order to cleanse their virtual and
non-virtual environments from their Palestinian former F-friends.
Two right-wing interviewees reported having unfriended all their Arab F-friends during the war, and
no one else. One of them told us that in general, Jewish-Arab friendships in her workplace did not survive the
war. Offline, Jewish and Arab coworkers may be careful about the political statements they make, although
political debates did take place in some workplaces (Arabs in particular tend to avoid making political
statements at work: Shehade-Switat 2015). By exposing users to their F-friends’ conversations in non-work
contexts, Facebook’s visibility structure resulted in exposing Arabs to inner-Jewish ‘backstage’
communication and vice versa. For example, when David (male, right-wing) stated on Facebook that he wishes
“cruel deaths” to Arabs, his Arab F-friends could read it. Consequently, users could no longer keep believing
a-priori that their F-friends are untypical Arabs or Jews in their political and national affiliations. Users’
participation in their own national debate marked them as ‘the enemy’ for coworkers of the other nationality.
Before the war, Inbal has never unfriended anyone (“it seemed infantile”). However, during the war
she unfriended a few F-friends, who happen to be all her Arab F-friends (“in Protective Edge, specifically, I
didn’t want to see what they post. I didn’t want to see it”), and only them, even though political statements by
Jewish F-friends also annoyed her. She told us: “after this war I don’t have any Arabs on Facebook (...) It’s
not that I had many, let’s say I had 4, 5? I have no-one.”
She came to know all of her Arab ex-F-friends as coworkers in two different workplaces. In one case,
unfriending followed an offline political dispute. However, she stressed that “at work I tried (...) not to talk
with them politics, to keep things pleasant.” Facebook’s visibility structure rendered this attempt futile. Thus,
a tie with an Arab coworker was dissolved after Inbal wrote on Facebook: “like the sea, Arabs will always
remain the same.” One of her F-friends commented that “not everyone is the same,” expressed his surprise and
disappointment about her generalization and suggested she should “open up.” She interpreted it as an
illegitimate disciplining attempt (“I wanna write whatever I feel like”), and defriended him immediately. While
she admitted that “when I tell it now, it sounds infantile,” this social tie has not been renewed since (“he doesn’t
talk to me since”).
She also unfriended an Arab ex-coworker for a hate post: “he used to talk to me, we were friends, like,
friends from work but still cool; and suddenly finding out that he’s like ‘I hate Jews, or Israelis, like, get out
of this land, I wish I could kill you one by one with a gun’—so (...) why are you talking to me at work if you
hate me?” This was her first defriending ever. She also contacted his employer, unsuccessfully trying to have
him dismissed. Despite being “shocked” to find out that some of her old Jewish classmates posted “pro-Arab"
posts and that one of her childhood friends was a leftist, Inbal defriended only Arabs (and all her Arab F-
friends who did not defriended her first). War is a sensitive time: Inbal’s brother fought in Gaza as a combat
commander, and one of her high-school classmates was killed during the war. The threat that “Arabs” posed
to her dear-ones’ lives was experienced as highly concrete. Consequently, her defriending policy followed
national lines. This was no exception in her workplace: since the war, she comments, “we are not their friends.”
Sigalit (female, right-wing) dissolved ties with a coworker who had become her “friend” (“I went out
with her (...) I came to her wedding”), because of a comment this Arab friend left on another person’s post. In
this comment, an engagement in a conversation in a different context, her friend "pointed a finger to the state
of Israel," and mocked IDF soldiers.
The NPO dataset demonstrates a similar pattern, although with more severe consequences. Like in
Inbal’s case, Jewish coworkers played a key role in workplace persecution: as F-friends, same-rank Jewish
coworkers were exposed to their coworkers’ Facebook activity and often were the ones who reported it to the
management and demanded action. Employers cited the hurt feelings of coworkers as explicit justification for
persecution. Thus, employees were rendered accountable for their political statements to their bosses, who
17 Schwarz & Shani / Culture in mediated interaction
were not their F-friends. These statements, often made in the context of a heated debate, where stripped out of
their dialogical context, e.g. in the case of an educational psychologist dismissed for replying the comment
“Only 200-and-so [Palestinians] were killed? Let many follow LOL” by paraphrasing it, writing “Only 13
[Israeli soldiers killed]? Let many follow LOL.”
In the cases above, Facebook’s visibility structure facilitated ‘ethnic cleansing’ of egocentric friend-
lists and workplaces alike. During wartime, when each person is called to duty and demanded to publicly
declare her opinions without being able to adapt performances to audiences, conflicts become much more
visible. Under these conditions, Facebook may threaten the livelihood of the weakest members of society,
15
as
well as the most fragile social ties within Israeli society.
DISCUSSION
Back to Durkheim? Rethinking the networked individualism
We opened the article by pointing at the convergence that takes place between the private and the public
spheres in Facebook and between different segments of users’ social networks. We then presented empirical
data that demonstrated some of the impacts of this convergence, the challenges it poses to actors and the
collapse of cultural mechanisms that facilitate cooperation despite cultural heterogeneity (formulated in terms
of interaction styles rather than in terms of intra-subjective values). While these findings are drawn from a
unique cultural context, they may still encourage us to rethink the sociological diagnosis of our time vis-à-vis
classical narratives that have conceptualized modernity as a unique transformation of sociation patterns.
Different founding fathers of sociology described modernity as a process of growing social differentiation—
the shift from Gemeinschaft to Gesellschaft in Toennies’ work, or from mechanical to organic solidarity in
Durkheim’s work—but it was Simmel who offered a purely formal account of this process, conceptualizing
modernity as a transformation in the structure of group affinities and social ties of individuals (Simmel
1955[1922]).
Simmel conceptualized society as a network of sociations and located modern individuals at the
intersection point of multiple circles of affiliation. Simmel’s modern society is thus based on informational
segregation—unlike pre-modern communities in which secrets were rare and social spheres nested within each
other, modern societies consist of multiple (and increasingly voluntary) ties and spheres, in each of which
individuals display different aspects of their identity and share different secrets. These multiple affinities pose
individuals conflicting demands for action and solidarity. For Simmel, modern individualism is thus an effect
of a shift in the structure of social networks (Simmel, 1950, 1952). One can draw striking parallels between
Simmel’s discussion and Goffman’s presentation of self: Goffman’s social actor may present different
personas in different spheres. Her capacity to differentiate between front-stage and back-stage, adopt her self-
presentation to her audience and generally control her personal information are the sources of individual
agency (Goffman 1958, 1961; see also Cahill 1998; on Goffman’s debt to Simmel: Smith 1989).
How shall we characterize our time in light of Simmel’s conceptualization of modernity? The most
ambitious attempt to do so was Barry Wellman’s. Wellman suggested that in late modernity, social structure
consists no longer of well-bounded intersecting groups, but rather of egocentric networks based on person-to-
person communication rather than on inter-institutional ties. Wellman’s late modernity is thus an extreme
version of Simmel’s first modernity: individuals are no longer the intersection points of bounded spheres but
rather nodes in a network, a metaphor that symbolizes the demise of the social group. Communities have
allegedly experienced glocalization and eventually dissolved into networks in the era of "networked
individualism," in which individuals are engaged with different vague and sparse networks, engagement which
is always partial and limited. The social control that each of these different milieus can employ on individuals
15
This threat is especially relevant to Arab employees, but low-rank right-wing Jewish employees may also
be affected—some have expressed fear of being labeled as “racist” (due to their expression style) and having
their livelihood jeopardized. Some even defriended leftist F-friends and coworkers to avoid this type of
scrutiny. Fear of persecution that may lead to defriending and self-censorship is thus closely linked to
symbolic and institutional power inequalities.
18 Schwarz & Shani / Culture in mediated interaction
and the social solidarity they offer are allegedly equally limited (Wellman 1999, 2002; Rainie and Wellman
2012).
This focus on the demise of well-bounded social groups is shared by other prominent popular and
sociological accounts of our time, such as those of Robert Putnam (2001), Anthony Giddens and Ulrich Beck
(e.g. Giddens 1994). Similarly, in political sociology, new social movements are often interpreted as
representing a structural shift from concentric circles towards a looser network structure of political affiliations
(Diani 2000).
Wellman stresses that this “social network revolution” has its roots well before the internet: in
globalization processes and structural changes in varied social institutions, some apparent already in the 1950s.
Yet, for Wellman the internet in general and social network sites in particular represent the ultimate epitome
of networked individualism. Wellman passingly admits that Facebook’s design renders maneuverability of
individuals between varied social worlds and relationships harder, yet he insists on interpreting Facebook as
the embodiment of networked individualism rather than as a threat to it.
Our findings are not in line with Wellman’s networked individualism thesis. Indeed, Facebook enables
person-to-person interaction across social and institutional boundaries, but it objectifies interaction and its
contents, exposes them to all of one’s social circles, and renders them subject to scrutiny by one’s F-friends.
The convergence and spillover of information across spheres render individuals accountable to collectives to
which they belong for their activity in other life-spheres, thus enabling (even encouraging) collectivities to
demand increased loyalty and public recognition of their collective sacred embodied in its boundary markers.
Our findings demonstrate that collectivities (such as the nation) still hold power to summon individuals and
demand them to express their loyalty. However, this power was expanded by Facebook’s structure. When the
segregation of interactions collapses, this power is no longer bounded to interaction within bounded sub-
networks of one’s egocentric network. Our argument is thus in line with recent dramaturgic accounts of
network sociality, which have identified some of the challenges posed by the internet in general and Facebook
in particular to impression management (Boyd 2014, Hogan 2010, van den Berg and Leenes 2011).
The social drama experienced by interviewees and posters, which often led to defriending, derives
from the structural tension between subjects who demonstrate modern subjectivity (as analyzed by Simmel
and Goffman) that presupposes a multiple self (maintained by privacy) and freedom of expression; and a
technological infrastructure that presupposes, as Zuckerberg openly declared, a unified coherent self across
social spheres, while promoting nearly total accountability to social collectives.
Moreover, when performances and interactions are rendered accessible beyond their original contexts,
interactants may be subjected to severe, disproportional punishment. Solove (2007) compared online shaming
to medieval shaming in the town square, perceiving it as a withdrawal from the urban anonymous and
differentiated world of modernity. In the 2014 war, some users were severely punished when online political
statements they made reached unintended audiences. They suffered multiple formal and informal sanctions,
including dismissal, public shaming, and interpersonal sanctions such as defriending. As demonstrated above,
the incapacity to maintain imagined homogeneity encouraged actors to try to impose actual political
homogeneity.
Do we revert then to a world in which people belong to total collectivities, which hold the power to
demand loyalty and punish those who commit sacrilege and offend collective moral emotions, that is, a
Durkheimian society? If this is the case, it stands in sharp contrast with most historical accounts of modernity
and late/post-modernity, including those discussed above, that portray modernization as shifting away from
this model. While Facebook allows individuals to represent their unique individual mixture of social
commitments, identities and interests, we found that in times of collective effervescence such as war,
Facebook’s structure of sociation and visibility allows collectives to exert power of coercion and demand
loyalty, since it removes the boundaries between different interaction contexts and social spheres.
There are however some important caveats. First, we should not forget that Facebook is only one of
many arenas of sociality. According to educated assessments the average American Facebook user has F-
friendships with only a third of her egocentric social network (Rainie and Wellman 2012:141). Even within
Facebook, users are aware of some of the difficulties that come with the dissolution of boundaries between
contexts and life-spheres. Some avoid discussing sensitive topics on Facebook and discuss them elsewhere
19 Schwarz & Shani / Culture in mediated interaction
(although they cannot make all their F-friends do the same). However, as the phenomena of political
defriending suggest, notwithstanding these coping strategies, the structural challenge remain.
Secondly, while our findings demonstrate a case of renewed coercive power of groups, this power
does not inhere in pre-existing groups; it derives at least partly from a material/technological structure. In this
sense, this new social reality diverts from the Durkheimian account both empirically and theoretically.
16
More
generally: while our empirical findings bring to mind Durkheimian and neo-Durkheimian accounts on the
continuous role of the sacred in modern solidarity, we do not explain these findings by relying on Durkheimian
ontological presuppositions, as our basic units of analysis are interactions and sociations rather than reified
societies.
Lastly, the direction of history cannot be reversed. The rise of modern differentiated societies was
ensued by the emergence of specific cultural ideals and specific subjectivities. Facebook users think of
themselves as multifaceted persons and strive to authentically express themselves and freely choose their
identities. The drama in the narratives in our data derives from this very tension between a subject molded by
modern individualist culture and a new environment. It is from this tension that a new type of sociality might
emerge.
Culture and networks: beyond mutual constitution
The case of political defriending also urges us to rethink the core theoretical question of the interrelations
between social structure and cultural contents (in our case, political orientations, values and styles).
Sociological literature increasingly stresses the mutual constitution of culture and social networks (Pachucki
and Breiger 2010; Vaisey and Lizardo 2010): while culture often flows in social networks, network structure
is shaped by value homophily and moral boundary work. However, as demonstrated above, cultural similarity
is a discursive and cognitive achievement, and is thus informed by the material organization of visibility and
access to information. As shown above, Facebook’s architecture reveals hitherto unnoticed differences and
restructures the network consequently. Hence, the dyadic relationship of culture and structure should be
reframed as a triad by introducing materiality as a third node that informs the interrelations between culture
and social structure.
Originally, social network analysis tried to explain contents by forms, culture by structure. Its "anti-
categorical imperative"—to study merely structural ties, without classifying people into groups—relied on a
reasoned and consistent ontology and epistemology. However, during the last decade, social network scholars
have increasingly reintroduced culture into their models. The cultural turn in SNA represents a growing
recognition that cultural dispositions, styles, and (inter)subjective meanings play a key role in the design,
reproduction and transformation of networks.
People actively shape networks by creating and dissolving social ties, and understanding these
networks dynamics requires taking into consideration boundary work (Barth 1969; Lamont and Molnar 2002;
Wimmer 2008): cultural processes in which a bunch of actors understand themselves as belonging to the same
group while relying on cultural styles and meanings to mark its boundaries. As a result, the density of ties
within the cluster increases, while outgroup ties grow weak, gradually turning this bunch into a cluster or even
a clique in SNA terms and creating structural holes. From this perspective, culture is not epiphenomenal:
discourses, worldviews and styles shape networks just as much as they flow within them. Network structure
and cultural meanings and dispositions are mutually constituting in a recursive manner (Pachucki and Breiger
2010; Vaisey and Lizardo 2010).
On first sight, the dynamics of defriending seems to fit this dyadic model: the same online egocentric
networks that shape users’ culture by exposing them to information and influences are also shaped by culture,
as cultural features informed defriending patterns. However, such a humanocentric account would miss a third
protagonist, Facebook’s algorithmic infrastructure that mediates and regulates social interaction, presentation
and representations of the self, thus intervening in the co-constitution of network and culture. By rendering
some cultural features of co-users visible in certain times, Facebook encourages the reorganization of
egocentric social networks on cultural and political lines. While lay discourse and academic literature usually
stress Facebook’s capacity to strengthen and maintain weak ties (Ellison et al 2007; Lewis and West 2009;
16
Since in our case the coercive power of society relies on non-humans, society is no longer a purely human
thing as in most interpretations of Durkheim (although see Latour 2005:37-8).
20 Schwarz & Shani / Culture in mediated interaction
Rainie and Wellman 2012), the case of political defriending demonstrates that Facebook’s design also subject
weak ties to value homophily, of which they were often exempt in purely-offline relationships. Our data thus
join other recent works that demonstrate how algorithmic structures inform social sorting and homophily in
sites such as online dating, school choice and hiring (e.g. Marchal et al. 2007, Weininger 2014). Like ours,
these cases reveal the mediating role of the material (and particularly, the digital) in the relationship between
culture and social structure.
CONCLUSION
As we have shown, in the dramatic times of war Facebook operated as an arena, to which Israelis were
summoned by different moral collectives. However, many found that following these imperatives and referring
online to the politically charged issue of war may result in tie dissolution. This is a result of Facebook’s design,
which renders audience segregation, audience accommodated presentation of self, and differentiation of group
styles highly difficult to pursue; objectifies interactions and makes them accessible to diverse unexpected
audiences; and creates increased accountability for statements and interaction gestures to those not present
during the original interaction.
By directing our attention to Facebook we could thus learn about cultural mechanisms that regulate
offline sociality and sustain interaction despite cultural heterogeneity. The constant spillover of information
between interactions, groups and audiences on Facebook challenges the capacity to maintain relationships
despite heterogeneity. This highlights the mechanisms, often kept unnoticed, which are used offline to allow
heterogeneous sociality to exist. Offline, when interactions are relatively well bounded, actors can maneuver
between multiple group styles. When the segregation of interaction collapse, individuals are not only exposed
to knowledge about other aspects of their F-friends (thus thwarting the constructing of imagined homogeneity),
but also gain indirect access to interaction styles of other groups to which these F-friends belong. While
interviewees largely wished to be tolerant to other views, it was the styles of expression they found harder to
tolerate.
Indeed, defrienders often believed they were purifying their social network from morally flawed F-
friends who hold different values. However, the hidden heterogeneity Facebook revealed to them cannot be
fully traced back to culture at the level of individual interiority. The heterogeneity uncovered was rather at the
level of culture in interaction, as in mediated interaction users were exposed to the textures of interactions in
which their F-friends engaged within other social circles.
Traditionally critiques of social differentiation concentrate on the mechanisms that separate groups
from each other, which are conceived as a threat to either the conservative sentiment of national unity or the
liberal ideals of diversity and freedom to cross intergroup boundaries. This is why ideas such as the "echo
chamber effect" of digital media (Sunstein 2007) and Bishop’s (2009) findings on political spatial segregation
in America attract much attention. However, the challenge our interviewees experienced were not an echo-
chamber effect delimiting them to familiar views, but rather the exposure of hidden political heterogeneity
within their social networks. Our findings thus point at a different set of mechanisms than those usually
identified as reducing tolerance to diversity. These are mechanisms that enable diversity and social ties across
social groups and boundaries through delimitation of interactions and social spheres.
It was Fredrik Barth who wrote in his seminal work on boundary work that intercultural
communication relies on "a structuring of interaction: a set of prescriptions governing situations of contact,
and allowing for articulation in some sectors or domains of activity, and a set of proscriptions on social
situations preventing inter-ethnic interaction in other sectors." Agreeing on what can be made relevant to
interaction renders further agreement on codes and values unnecessary (Barth 1969:16). Yet, the structuring
of interaction and the employment of such cultural strategies is informed and constrained by material
infrastructures.
The densely populated modern city, along with the physical separation of work, home and third places,
are the physical infrastructure, without which we cannot understand the emergence of Simmel's modern
individuals and their segregated ego-centric social networks. On Facebook, the algorithmic structuring of
interaction, audiences and visibility makes it almost impossible for individuals to segregate their presentation
of self for each of their specialized networks. However, in different cultural (and situational) contexts this
21 Schwarz & Shani / Culture in mediated interaction
infrastructural transformation may result in either avoidance of political talk online (as documented in the
USA: Thorson 2013), or political defriending and increased cultural homophily (as happened in our case).
The convergence of social spheres and the spillover of information between them did not promote
networked individualism and decline in the coercive power of social groups but rather the opposite.
Specifically during war, intragroup political conversations have violated the sacred of other political groups
(e.g. “progressive” universalism vs. national ‘‘loyalty”), encouraging the intensification of boundary work.
Taken out of context, war-related statements received new meanings that rendered them sacrilegious, and
hence users faced both informal and formal sanctions (defriending, shaming, persecution, and dismissal) for
their informal political comments in interpersonal Facebook communication.
The inability to differentiate and distinguish between public debate and conversation with like-minded
people has led to dissolution of (online and occasionally offline) ties between right-wing and left-wing Israeli-
Jews, and between Jews and Arabs. By increasing the visibility of political styles across contexts; and by
undermining the imagined homogeneity of collectives to which individuals subscribe, Facebook promotes
actual cultural homophily.
This trend represents a sharp break from modernization processes as conceptualized in the Simmelian
tradition. The separation of life spheres, which Weber and Simmel perceived as substantial to modernity and
Goffman viewed as substantive of human agency, is being challenged by commercially-driven digital social
arenas promoting the ideology of sincerity and unified selfhood. The material world is thus revealed as a
powerful actant in the interrelations between network structure and culture.
Acknowledgements
We wish to thank Sagit Festman for research assistance; Nicholas A. John, Gadeer Nicola and Kav LaOved–
Worker's Hotline for generously allowing us access to data; Lior Gelernter, Ido Yoav and the AJCS anonymous
reviewers for their helpful comments on earlier drafts; and the interviewees for their helpful cooperation.
About the Authors
Ori Schwarz is a senior lecturer of sociology at Bar-Ilan University. He has published on various topics,
including the role of sonic styles and sonic sensitivities in boundary work; the cultural sociology of
authenticity; and the influence of digital data-objects on notions of value, attention, and memory.
Guy Shani is a PhD candidate in the Department of Sociology and Anthropology at Tel Aviv University. His
research focuses on major life choices and middle-class identities from a cultural sociology perspective. He
studied university students' choice of academic field, and is now studying residential choices of middle-class
households in different urban settings.
22 Schwarz & Shani / Culture in mediated interaction
References
Alexander, J.C. (2008) Iconic Consciousness: The Material Feeling of Meaning. Environment and Planning
D: Society and Space 103(1): 10-25.
Barth, F. (1969) Ethnic Groups and Boundaries. Boston MA: Little, Brown and Co.
Beck, P.A. (1991) Voters' Intermediation Environments in the 1988 Presidential Contest. Public Opinion
Quarterly 55: 371-394.
Beer, D. (2009) Power through the Algorithm? Participatory web cultures and the technological unconscious.
New Media & Society 11(6): 985-1002.
Bishop, B. (2009) The Big Sort: Why the Clustering of Like-Minded America Is Tearing Us Apart, Boston:
Mariner.
Blau, P.M. and Schwartz, J.E. (1984) Crosscutting Social Circles: Testing a Macrostructural Theory of
Intergroup Relations. Orlando, FL: Academic Press.
Boyd, D. 2014 It's Complicated: The social lives of networked teens. New Haven, CT: Yale UP.
Brubaker, R. (2004) Ethnicity without Groups, Cambridge, MA: Harvard UP.
Brubaker, R., Feischmidt, M., Fox, J. & Grancea, L. (2006) Nationalist Politics and Everyday Ethnicity in a
Transylvanian Town. Princeton, NJ: Princeton UP.
Bucher, T. (2012) A Technicity Of Attention: How Software ‘Makes Sense’. Culture Machine 13:1-23,
http://www.culturemachine.net/index.php/cm/article/view/470/489.
Cahill, S.E. (1998) Toward a Sociology of the Person. Sociological Theory 16(2):131-148.
Castells, M. (2011) Communication power. Oxford: Oxford UP.
Das, S., and Kramer, A. (2013) Self-censorship on facebook. In Proceedings of ICWSM 120-127.
Diani, M. (2000) Simmel to Rokkan and Beyond: Towards a Network Theory of (New) Social Movements.
European Journal of Social Theory 3(4):387-406
Earl, J., and Kimport, K. (2011) Digitally Enabled Social Change: Activism in the Internet Age. Cambridge:
MIT Press.
Eliasoph, N. (1998) Avoiding Politics: How Americans Produce Apathy in Everyday Life. Cambridge:
Cambridge UP.
Eliasoph, N. (1999) "Everyday Racism" in a Culture of Political Avoidance: Civil Society, Speech, and Taboo,
Social Problems, 46(4):479-502.
Eliasoph, N. and Lichterman, P. (2003) Culture in Interaction. American Journal of Sociology 108:735-794.
Ellison, N., Steinfield, C. and Lampe, C. (2007) The Benefits of Facebook “‘Friends”: Social Capital and
College Students’ Use of Online Social Network Sites, Journal of Computer-mediated Communication
12(4):1143–68.
Ellison, N. B., Steinfield, C., and Lampe, C. (2011) Connection strategies: Social capital implications of
Facebook-enabled communication practices. New Media & Society, 13:873–892.
Falah, R. (1996) Living Together Apart: Residential Segregation in Mixed Arab-Jewish Cities in Israel, Urban
Studies 33(6):823-857.
Giddens, A. (1994) Risk, Trust, Reflexivity, in U. Beck, A. Giddens and S. Lash, Reflexive Modernization,
Cambridge: Polity, pp.184–97.
Goffman, E. (1959) The Presentation of Self in Everyday Life. New York: Anchor.
Goffman, E. (1961). Asylums: Essays on the social situation of mental patients and other inmates.
Harmondsworth: Penguin.
Goffman E. (1982) Interaction Ritual: Essays on Face-to-Face Behavior. New York: Pantheon.
Hogan, B. (2010) The Presentation of Self in the Age of Social Media: Distinguishing Performances and
Exhibitions Online. Bulletin of Science, Technology & Society 30(6):377–386.
Jenkins, R. (2008) Social Identity. New York: Routledge.
Jerolmack, C. and S. Khan (2014) Talk is Cheap: Ethnography and the Attitudinal Fallacy. Sociological
Methods & Research 43(2):178-209.
23 Schwarz & Shani / Culture in mediated interaction
John N. and Dvir-Gvirsman, S. (2015) ‘I Don’t Like You Any More’: Facebook Unfriending among Israelis
During the War of 2014. Journal of Communication 65(6):953-974.
Knoke, David. 1990. Political Networks: The Structural Perspective. Cambridge: Cambridge UP.
Knorr Cetina, K., & Bruegger, U. (2002). Global microstructures: the virtual societies of financial
markets. American Journal of Sociology 107(4): 905-950.
Lamont, M. and Molnar, V. (2002) The Study of Boundaries in the Social Sciences. Annual Review of
Sociology 28:167-95.
Lamont, M., Pendergrass, S. and Pachucki, M. (2015) Symbolic Boundaries, in Wright, J (ed.) International
Encyclopedia of Social and Behavioral Sciences, Oxford: Elsevier, pp.850-855.
Latour, B. (1996) On interobjectivity. Mind, Culture and Activity 3(4): 228-245.
Latour, B. (2005) Reassembling the social. Oxford UP.
Lavie, R. (2014) In Every War I lose Friends, Haokets, www.haokets.org/2014/08/05/-תדבאמ-ינא-המחלמ-לכ
תורבח/
Lazarsfeld, P. and Merton, R. (1954) Friendship as a Social Process: A Substantive and
Methodological Analysis, in M. Berger, T. Abel, and C.H. Page (eds.) Freedom and Control in Modern
Society, New York: Van Nostrand, pp.18-66.
Lewis, J. & West, A. (2009) ‘Friending’: London-based undergraduates’ experience of Facebook. New Media
& Society 11(7):1209–1229.
Marchal E., Mellet, K. and Rieucau, G. (2007) Job board toolkits. Internet matchmaking and changes in job
advertisements. Human Relations 60(7):1091-1113.
McPherson, M., Smith-Lovin, L., and Cook, J.M. (2001) Birds of a feather: homophily in social networks.
Annual Review of Sociology 27:415–44.
Morey, A.C., Eveland, W.P. and Hutchens, M.J. (2012) The “Who” Matters: Types of Interpersonal
Relationships and Avoidance of Political Disagreement. Political Communication 29:86–103.
Putnam, R. D. (2001). Bowling Alone: The collapse and revival of American community. New York: Simon
and Schuster.
Pachucki, M.A. and R.L. Breiger (2010) Cultural Holes: Beyond Relationality in Social Networks and Culture.
Annual Review of Sociology 36:205-24.
Rainie, L., and Smith, A. (2012) Social networking sites and politics, http://www.pewinternet.org/files/old-
media//Files/Reports/2012/PIP_SNS_and_politics.pdf.
Rainie, L. and Wellman, B. (2012) Networked: The New Social Operating System. Cambridge, MA: MIT Press
Schwarz, O. (2011) Who Moved my Conversation? Instant Messaging, Intertextuality, and New Regimes of
Intimacy and Truth, Media, Culture & Society 33(1): 71-87.
Schwarz, O. (2012) The New Hunter Gatherers: Making Human Interaction Productive in the Network Society,
Theory, Culture & Society 29(6): 78-98.
Scott, S.V. and Orlikowski, W.J. (2015) The Algorithm and the crowd: Considering the materiality of service
innovation. MIS Quarterly 39(1): 201-216.
Shehade-Switat, M. (2015) The influence of national identity of Arabs Israel on the emotion management
strategies in ethnically mixed workplaces in Israel, unpublished MA thesis, Haifa University.
Sibona, C. and Walczak. S. (2011) Unfriending on Facebook: Friend Request and Online/Offline Behavior
Analysis. HICSS 2011:1-10.
Simmel, G. (1910) How is Society Possible? American Journal of Sociology, 16(3):372-391.
Simmel, G. (1950) The Sociology of Georg Simmel, Glencoe: Free Press.
Simmel, G. (1955[1922]) Conflict and the Web of Group Affiliation, New York: Free Press.
Smith G.W.G. (1989) Snapshots 'Sub Specie Aeternitatis': Simmel, Goffman and Formal Sociology. Human
Studies, 12:19-57.
Smith, P. (2005) Why War? The Cultural Logic of Iraq, the Gulf War and Suez. Chicago: Chicago UP.
Smith A. & Duggan, M. (2013) Online Dating & Relationships, http://www.pewinternet.org/files/old-
media//Files/Reports/2013/PIP_Online%20Dating%202013.pdf
Solove, D.J. (2007) The Future of Reputation: Gossip, Rumor, and Privacy on the Internet. New Haven: Yale
UP.
Swidler, A. (2003) Talk of Love: How Culture Matters. Chicago: Chicago UP.
Sunstein. C.R. (2007) Republic.com 2.0, Princeton: Princeton UP.
Tavory, I. (2011). The Question of Moral Action: A Formalist Position. Sociological Theory 29(4):272-293.
Tavory, I. (2015) Summoned: Religious Life in an Orthodox Jewish Neighborhood, Chicago: Chicago UP.
24 Schwarz & Shani / Culture in mediated interaction
Thorson, K. (2013) Facing an Uncertain Reception: Young citizens and political interaction on
Facebook. Information, Communication & Society 17(2):203-216.
Vaisey, S. and Lizardo, O. (2010) Can Cultural Worldviews Influence Network Composition? Social Forces
88(4):1595-1618.
van den Berg, B., & Leenes, R.E. (2011) Keeping up appearances: Audience segregation in social network
sites. In S. Gutwirth, Y. Poullet, P. de Hert and R. Leenes (Eds.) Computers, privacy and data
protection: An element of choice. Heidelberg: Springer, pp.211-232.
Weininger, E. (2014). School choice in an urban setting. In A. Lareau & K. Goyette (Eds.), Choosing Homes,
Choosing Schools, New York: Russell Sage, pp. 268–294.
Wellman, B. (1999) Networks in the Global Village, Boulder, CO: Westview.
Wellman, B. (2002) Little Boxes, Glocalization, and Networked Individualism. In M. Tanabe, P. van den
Besselaar & T. Ishida (Eds.), Digital Cities II: Second Kyoto Workshop on Digital Cities. Heidelberg,
Germany: Springer, pp. 10-25.
Wimmer, A. (2008) The Making and Unmaking of Ethnic Boundaries. A Multilevel Process Theory. American
Journal of Sociology 113(4):970-1022.
- A preview of this full-text is provided by Springer Nature.
- Learn more
Preview content only
Content available from American Journal of Cultural Sociology
This content is subject to copyright. Terms and conditions apply.
