Does context really collapse in social media interaction?

Jan Blommaert & Gosia Szabla

(Plenary paper, conference on Moving Texts: Mediations and Transculturations. Aveiro, 12 July 2017)


‘Context collapse’ (CC) refers to the phenomenon widely debated in social media research, where various audiences convene around single communicative acts in new networked publics, causing confusion and anxiety among social media users. The notion of CC is a key one in the reimagination of social life as a consequence of the mediation technologies we associate with the Web 2.0. CC is undertheorized, and in this paper we intend not to rebuke it but to explore its limits. We do so by shifting the analytical focus from “online communication” in general to specific forms of social action performed, not by predefined “group” members, but by actors engaging in emerging kinds of sharedness based on existing norms of interaction. This approach is a radical choice for action rather than actor, reaching back to symbolic interactionism and beyond to Mead, Strauss and other interactionist sociologists, and inspired by contemporary linguistic ethnography and interactional sociolinguistics, notably the work of Rampton and the Goodwins. We apply this approach to an extraordinarily complex Facebook discussion among Polish people residing in The Netherlands – a set of data that could instantly be selected as a likely site for context collapse. We shall analyze fragments in detail, showing how, in spite of the complications intrinsic to such online, profoundly mediated and oddly ‘placed’ interaction events, participants appear capable of extraordinarily ‘normal’ modes of interaction and participant selection. In fact, the ‘networked publics’ rarely seem to occur in practice, and contexts do not collapse but expand continuously without causing major issues for contextualization. The analysis will offer a vocabulary and methodology for addressing the complexities of the largest new social space on earth: the virtual space of online culture.


  1. Introduction

In social media studies, the notion of “context collapse” has acquired considerable currency.[1] It is part of an – often tacitly adopted – theory of communication grounded, in turn, in a particular imagery of the social world, and stands for

“the flattening out of multiple distinct audiences in one’s social network, such that people from different contexts become part of a singular group of message recipients”. (Vitak 2012: 541)

This is generally seen as a problem, something that distorts “normal” assumptions about communication and requires caution and repair strategies. This problem is an effect of the specific features (affordances as well as constraints) of social network communication, the technology of which “complicates our metaphors of space and place, including the belief that audiences are separate from each other” (Marwick & boyd 2010: 115), and has taken us from a world of relatively transparent audiences to that of far less transparent “networked publics” (boyd 2011). Users on social network sites (SNS) have assumptions about whom they are addressing and interacting with, but the features of SNS do not correspond to these assumptions and create indeterminacy in audience selection, with confusion and uncertainty of users as one effect.

While the notion of context collapse certainly has its merits and should not be dismissed entirely – the indeterminacy of addressees is irrefutable – it invites critical scrutiny. In what follows, we shall engage in such an examination, aimed, specifically, at the assumptions about the social world and communication carried along with the notion. And we shall do this by means of a relatively straightforward approach: confront such assumptions with a detailed analysis of a sample of SNS interaction. The latter, we undertake by means of well-established methodological tools drawn from the interactionalist discourse-analytical tradition, notably linguistic ethnography and interactional sociolinguistics (e.g. Cicourel 1973; Gumperz 1982, 1992, 2003; Rampton 2017; Blommaert 2018).

Let us first look somewhat closer at how the problem of context collapse is sketched by some prominent authors and highlight some of the more questionable assumptions underlying such sketches.

The problem called context collapse rests on a general imagination of communication – in earlier times – as not (as) sensitive to context collapse. Before we had SNS, communication was relatively simple. Davis & Jurgenson (2014: 477) speak of “the relative segmentation [of communication] of earlier times”, and this has to do with a presumed clarity of audience and situation. People (it is presumed) used to know quite clearly with whom they interacted and, thus, how they should interact. The big problem caused by SNS lies in the latter’s unique affordances: communication through SNS is persistent, replicable, scalable, searchable and sharable – features, all of them, that characterize communication beyond the immediate interactional situation (or beyond the single speech event, to quote the title of an excellent recent study of such phenomena: Wortham & Reyes 2015).

This, of course, has effects on who can be addressed by SNS messages, and how such unintended audiences might respond and react to them. People tend to get confused on SNS in a specific way:

“While Facebook and Twitter users don’t know exactly who comprises their audience addressed, they have a mental picture of who they’re writing or speaking to – the audience invoked. Much like writers, social media participants imagine an audience and tailor their online writing to match”. (Marwick & boyd 2010: 128)

This analogy with professional writers turns SNS interaction into something special, exceptional. While SNS “combines elements of broadcast media and face-to-face communication” (id: 123), spoken face-to-face interaction is the normal default mode of communication, the source of people’s expectations and norms in interaction, also in scholarship:

“Most of these studies [on ‘normal’ interaction] draw from data and observations that involve people interacting face-to-face, where it is fairly easy to gauge the gender, race, status, etc. of the audience. Removing this ability creates tensions”. (ibid)

To be more precise:

“The requirement to present a verifiable, singular identity makes it impossible to differ self-presentation strategies, creating tension as diverse groups of people flock to social network sites”. (id: 122)

These tensions often have to do with issues of “privacy” and have effects in the ways in which people handle issues of unintended addressees, by means of privacy settings, self-censorship or “unfriending” and “blocking” (e.g. Marwick & boyd 2014; Sibona 2014; Dugay 2016).

We can pause now and take stock. Underlying discussions of context collapse, there is a social imagination of communicative simplicity and determinacy. SNS communication complicates a world in which “normal” interaction was:

  • Dyadic and spoken, with clear, transparent, “authentic” and verifiable (singular) identity positions deployed
  • in a linear, simple and bounded activity, not replicable beyond the speech event, not shareable, not searchable etc.
  • and with a maximum of social sharedness, relating to the nature and identities involved and the audiences addressed.

People, so it seems, had just one set of common assumptions about communication: those directing simple dyadic face-to-face conversation in a world known to both participants. Complex and non-homogeneous audiences used to be exceptional and only familiar to specialized practitioners: “professional writers” (Marwick & boyd 2010: 115). Within acts of communication, ordinary people performed simple bounded activities resting on shared assumptions and conduct-and-meaning frames circulating in a “real” community; all of this, together, constructed the “context” with which people were familiar. Such simple contexts are no longer afforded in the blended, complex networked publics of SNS, and tensions arise. As we can see, people are, in a way, “stuck” in specific contexts: “people from different contexts become part of a singular group of message recipients” (Vitak 2012: 451). And even in more sophisticated discussions, where the assumption of a “verifiable, singular identity” is replaced by a more Meadian-Goffmanian emphasis on specific and diverse forms of social roles and role expectations, such roles and expectations appear to “belong” to specific networks:

“These expectations inform appropriate – and inappropriate – lines of action and identity performance. In these terms, collapse refers to the overlapping of role identities through the intermingling of distinct networks”. (Davis & Jurgenson 2014: 477)

Groups – “audiences”, “networks” or “publics” – appear to have amazing degrees of stability and persistence, and “contexts”, in that sense, are features derived from group membership. It is the presence of such unintended audiences that generates context collapse.

  1. An interaction-centered alternative

There is no need, we think, for a lengthy refutation of the assumptions directing the concept of context collapse. All of them are sociologically and sociolinguistically questionable in a variety of ways.[2] Rather, we would state an alternative general principle and take it through into an analysis of a concrete example.

The principle is that of action, and we adopt it from the interactionist tradition (Goffman 1961; Blumer 1969; Strauss 1993; also Mead 1934). We have seen that some authors refer to this tradition in their attempt to escape the sociological overgeneralizations in the concept of context collapse; we intend to take this line of argument much further.

The literature on context collapse, we have seen, starts from assumptions about groups (‘audiences’), their features and stability in explaining interaction; and the latter is done generally: authors speak of ‘SNS communication’ as one single object, features of which include context collapse. Instead of these, we focus not on groups but on actual practices performed by people, and we focus on specific practices. People do not just communicate, they perform highly specific actions such as ‘asking’, ‘arguing’, ‘shouting’, quarreling’ or ‘storytelling’, and they do so within the space of higher-level social actions such as, for instance, ‘conversation’. It is within the layered structure of such complex actions that we consider ‘context’ and how people deal with it. Such contexts include chronotopic patterns of identity work (a term we prefer over for instance ‘role taking’) based on the genre characteristics of specific activities (Wang & Kroon 2016; Blommaert & De Fina 2017; Karimzad & Catedral 2017). All of this is interactional, i.e. it is driven not by just individual motives and choices but by social (normative) ones that need to be dialogically established and ratified in order to be meaningful in interaction.

We can turn this old interactionist principle into a simple, four-line methodological program for the sociolinguistic analysis of interaction (cf. Blommaert 2017, 2018).

  1. Patterns of communication necessarily involve meaningful social relationships as prerequisite, conduit and outcome;
  2. Such relationships will always, similarly, involve identities and categorizations, interactionally established;
  3. Thus, when observing patterns of communication, we are observing the very essence of ‘sociation’ (Georg Simmel’s term for the continuous evolving of society through social action), and of ‘groupness’– regardless of how we call the groups.
  4. And specific patterns of interaction shape specific forms of groups.

The points of departure underlying context collapse are turned upside down here: we do not start from images of groups, with actions and their features derived from them, but we start from actions and see which kinds of groups might emerge from them. In this sociolinguistic frame we approach groups pragmatically and axiologically, from the angle of the actual observable communication practices and through the values attributed to such practices. Groups, then, are not a priori given collections of human beings but must be taken from patterned sets of communicative behaviors and the relationships with which they are dialectically related. Whenever we see such ordered forms of communicative behavior, there is an assumption of active and evolving groupness – sociation – but the analytical issue is not the nature of the group (or the label we need to choose for it) but the specific social relationships observable through and in communication. All other aspects of sociation can be related to this. So if one needs the definition of a group: a group is a communicatively organized and ratified set of social relationships.

To shift back to context collapse notions: ‘networked publics’ do not exist in any real sense independently of specific patterns and modes of interaction, they are generated by them and they change from action to action, for each action can (and usually does) involve different forms of relationships between actors. When someone tells a story in a conversation, s/he ‘leads’ the event, so to speak; when a few minutes later that same person asks an informative question to the interlocutor, s/he shifts into a subordinate role in the event; and when the interlocutor’s phone rings, s/he changes from participant to non-participant in a moment’s notice. The ‘group’ made up of the interlocutors is, thus, unstable, continuously emerging and subject to dialogical (re-)ratification at any moment in the conversation, depending on what exactly goes on (see Rampton 2006; Goodwin 2007; Goodwin & Goodwin 1992 for excellent examples).

When we apply this frame now, we begin to notice certain things. For instance, we notice that people don’t usually interact with ‘audiences’ or ‘networks’ but with specific addressees placed in specific relationships with them during highly specific forms of interaction. In the examples given by Marwick & boyd (2014) to show the dynamics of privacy control on SNS, thus, we see that much of what people actually do is addressee selection (expressed quite transparently in lines such as “I wasn’t talking to you”, Marwick & boyd 2014: 1057), or more generally the construction of highly specific participation frameworks for specific actions (Goodwin & Goodwin 1992, 2004; Goodwin 2007). Dugay (2016) describes strategies of deliberate simultaneity and ambivalence performed by SNS users, so as to separate specific addressees from the broader audiences; and Sibona’s analysis of ‘unfriending’ on Facebook (2014) is evidently a practice of addressee selection-by-exclusion. Thus, the diffuse (and confusing) ‘audiences’ and ‘network publics’ causing context collapse appear, in actual practice, to be chopped into much smaller and highly specific sets of addressees. The reasons for that may be privacy concerns or anxieties over undesirable disclosure of information on SNS – we do not exclude that possibility. But they may also be an effect of much simpler features of social action on SNS. We shall now attempt to demonstrate that by turning to our case.

  1. Complex compound social action on SNS: A case

The case we shall examine in some detail is a long discussion on a Facebook forum for Polish migrants in the Netherlands.[3] The data, as will become clear, represent a lengthy and complex case of Facebook interaction, starting from an update which then triggers likes, comments and reactions to comments. The interaction ran for five days, from March 14 until March 19, 2016.[4] No less than 65 individuals were involved in the conversation, and the update triggered a total of 192 responses – ‘comments’ as well as ‘replies’ to comments. In our analysis, we shall call the entire interaction the event; the update defines the main action; comments and replies to comments are all actions. We shall need to provide more precise descriptions of those actions later. Thus, the main action, performed by a female journalist whom we shall nickname ‘Ala’, invited 79 comments and 113 replies: a total of 192 actions. In our transcript (available online) the main action is numbered 0, the comments are numbered as 1, 2, 3… etc, and the replies to comments as 6.1, 6.2, 6.2 … etc.

The main action occurred on March, 14, 2016 at 12.37 p.m. when Ala posted this update:

Ala (F): witam, jestem dzienkarka telewizijna I szukam polakow, co pracuja w szklarniach co chca cos opodwiadac o warunkach pracy lubmieszkac I pracowac zagranica bez rodziny. Chetnie infornacie na priw. Krecenjie moze sie stac tez anonymowo.

Translation: Hello, I am a television journalist and I am looking for Polish people, who work in greenhouses who want to tell me about the working conditions or living and working abroad without family. Gladly information on priv. Filming can also happen anonymously.

Due to the initial negative responses, the comment was edited at 01.40 p.m., and from then on appeared online in the following form:

Ala (F): witam, jestem dzienkarka telewizijna I szukam polakow, co pracuja w szklarniach co chca cos opodwiadac o warunkach pracy lub mieszkac I pracowac zagranica bez rodziny. Chetnie infornacie na priw. Krecenjie moze sie stac tez anonymowo. (bo duzo ludzy pyta dlaczego tak zle pisze: jestem urodzona w polsce, ale pracuje dla telewisji niemieckiej i holenderskiej. Przeprazaszam za bledy, ale wyjechalam z polski jakmialam 4 latka. wydaje mi sie jednak, ze kommunikacja w tej grupie powinna byc po polsku, dlatego starams ie..)

Translation:Hello, I am a television journalist and I am looking for Polish people, who work in greenhouses who want to tell me about the working conditions or living and working abroad without family. Gladly information on priv. Filming can also happen anonymously. (because many people ask why I am writing so badly: I am born in Poland, but I am working for German and Dutch television. I am sorry for mistakes, but I left Poland when I was 4 years old. It seems to me however, that communication in this group should be in Polish, that’s why I am trying).

The update of Ala is a straightforward statement with a request for assistance. Her Polish however is questioned, because it is orthographically, grammatically and pragmatically awkward.[5] The text visible above is understandable, but there are spelling mistakes and grammatical errors, and upper case or punctuation are (not unusually in online writing) missing too. Generally, the text is ‘awkward’, and Ala uses words which sound odd in particular sentences. For example, she says “Jestem urodzona w Polsce”, whereas it would be more expected to say ‘Urodzilamsie w Polsce’ or ‘Pochodze z Polski’ in this context. Her sentence literally translates to English “I am born in Poland”, whereas Urodzilamsie w Polsce would translate: “I was born in Poland”.

We sequentially numbered every participant as they entered into action and marked them as ‘F’ (female) or ‘M’ (male). 34 participants only commented once. Some of them commented as a reply to the event, others only replied to one of the sub-actions. 4 people stood out with their number of comments: Ala (F) posted 24 times on different entries; Participant3 (F) commented 11 times, but all of this as part of the complex discussion following of action 2. Participant4 (M) engaged in the conversation 11 times throughout the event; and prticipant13 (F) engaged 15 times, all of it in actions 2 and 6.  8 other people commented at least 5 times (Participant6 (M), Participant14 (F), Participant30 (F), Participant31 (M), Participant33 (M), Participant53 (M), Participant57 (M), Participant60 (M)). 19 people commented more than once, but less than 5 times. In general, different actions and sub-actions trigger different participation frameworks. The change is clearly visible, but overlap is present as well.

3.1 A complex, nonlinear social event made up of diverse actions

The event is nonlinear. There are thematic shifts, main lines of interactional activity interrupted by stand-alone dyadic interactions, and gender balance shifts. The first part of the event, actions 1-8, is dominated by female participants; actions 9- 34 shows a more gender-balanced profile, while from action 34 up until the final action 79, the interaction is dominated by male respondents.

It is also nonlinear in the sequential sense: people sometimes reply to comments, and thus perform responsive actions to ulterior actions, long after posterior actions had been performed. This is one of the particular affordances of SNS, and the clearest example of it here is the main action itself. Ala posted her original update at 12.37pm on March 14, 2016. This instantly triggered a heated discussion about the spelling and other errors in her text, starting with action 2 at 12.43. Ala edits her update about one hour later, after 16 turns in the discussion, which partly takes the sting out of the discussion on her language proficiency.

This brings us to the issue of actions in need of more precise descriptions. Ala’s update is, as we said, the main action. It introduces a thematic domain and an action format: she launches a request or invitation to Polish people working in Dutch greenhouse industries, to participate in a TV program she intends to make. The thematic domain, from then on, defines what is ‘on topic’ or ‘off topic’, and in that sense establishes the benchmarks for what we could call legitimate participation; the action format – a request – further establishes such benchmarks. The most ‘normal’ response to such an action is to accept or decline the request.

The main action, we can see, draws the main lines of the normative framework that will be deployed in judging the conduct of participants. The main action, in that sense, is always a normatively ratified action frame in which a preferred participation framework and preferred modes of activity are inscribed: people who collaborate with it (respond supportively and stay ‘on topic’) are welcome and legitimate participants; people who deviate from it are unwelcome and illegitimate participants. And at the end of the event, Ala can be satisfied. Of the 79 actions following her request, 31 are cooperative.[6] Participants either straightforwardly volunteer, add encouraging comments or offer further suggestions for developing Ala’s TV program.

But this is not all. Consider Figure 1, a graphic representation of the different actions that occur in the event:

ScreenHunter_575 Jun. 01 11.33

Figure 1: actions in the event.

We can see how the main action sets in motion not one line of action, but several: the event is a complex, nonlinear and composite event, made up of highly divergent actions, legitimate as well as illegitimate ones. And two lines of illegitimate divergence should be highlighted, for both can be said to originate, nonlinearly, from Ala’s main action.

One: Ala’s main action, we have seen, establishes the normative action and participant framework for the event. It is, however, also an action in its own right, the features of which are socio-semiotically salient as reflexive indexicals of identity. Thus, one very salient line of illegitimate participation revolves around the metapragmatics of Ala’s update. And this starts very quickly. After one first collaborative comment (action 1), a female participant (Participant 3) launches a direct attack on Ala’s update in action 2, just a handful of minutes after Ala’s update appeared online:

  1. Participant3 (F): Zajebista dziennikarka co bledy w pisowni robi..

Translation: Fucking great journalist who makes spelling mistakes…

Date:  March 14 at 12:43pm  Likes: 26

The comment, observe, receives 26 likes. In addition, it triggers several things. It triggers the longest series of replies to comments, 57, turning it into a ‘discussion-within-a-discussion’. We get a flurry of 48 replies in the hour following Participant 3’s comment; lower intensity interaction resumes later that night and continues until March 19.[7] Next to that, it also establishes language and identity as a separate line of activity throughout the event. Issues of orthographic stability and language competence are raised throughout this long discussion, and 13 actions directly or indirectly raise issues of language proficiency.[8] And finally, it triggers action censorship as part of the discussion: respondents are identifying linguistic errors of one another, but they are also engaging in self-correction by editing their original messages. Thus we can see that the formal, indexical features of the main action, apart from its thematic contours and action-and-participation frame, become a theme that informs all sorts of other actions, .[9] including general disparaging meta-commentaries such as in action 33:

  1. Participant45 (F): Jakby tak dokladnieprzepytac zasad gramatyki jezyka polskiego tych, co maja zawsze na ten temat duzo do powiedzenia….

Translation: If we only could test the knowledge of the Polish grammatical rules of those who always have so much to tell on this topic…

Date: March 14 at 4:02pm   Likes: 2

Or consider the actions 38 and 39 (and observe the expletives in 38, quite a frequent feature in the more heated parts of the event):

  1. Participant48 (F): Przeczytałam wszytkie te komentarze i dawno się tak… nie zdenerwowałam. Wstyd Wam powinno być wredne i zawistne baby!!! Ala Powodzenie i duzo sukcesów w pracy

Translation: I read all the comments and it has been a long time since I got so frustrated. Shame on you disgusting and envious chick !!! Ala good luck and lots of successes at work J

Date:  March 14 at 6:36pm  Likes:  5

  1. Participant49 (M): To wlasnie robi holandia z polakow

Translations: This is exactly what Holland makes of Poles

Date:  March 14 at 7:06pm  Likes:  2

Two: The topic proposed by Ala for her TV program – Polish workers in Dutch greenhouse industries – likewise becomes a self-standing motif provoking a range of comments and discussions. In several collaborative responses to Ala’s request, participants volunteer information about the labor conditions in such segments of the market, as in action 24:

  1. Participant36 (M): dalbym ci jeden temat jak lokuja ludzi jak swinie w oborach gdzie strumyk gownaplyniesrodkiem pokoju

Translation: I can give you one topic about how they locate people like pigs in barns where a stream of shit flows through the middle of the room

Date:  March 14 at 1:53pm  Likes: 0

Such collaborative responses are complemented by general remarks on The Netherlands, the Dutch people, and the Polish workers as well, and in the second part of the interaction a full-blown discussion develops on what we could call the ‘ethos’ of being a Polish immigrant worker in The Netherlands.[10] Consider the exchange in actions 45-47:

  1. Participant54 (M): skoro wam tak zle to dlacze zgadzacie siewciaz na takie traktowanie? zmiana pracy, poprostu. da sie inaczej troche wiary i samozaparcia a nie tylko narzekac

Translation: If it is so bad why do you still agree to be treated this way? Change job, as simple as that. You can do it differently, a little bit confidence and determination and not only complaining

Date:  March 14 at 11:41pm  Likes:  3

  1. Participant38 (F): Dokładnie zgadzam się całkowicie, trzeba pamiętać gdzie chciałoby się być i dążyć do tego małymi krokami. ….

Translation: Exactly I agree completely, you need to remember where you would like to be and to pursue one’s aim step by step.

Date:  March 15 at 6:04am  Likes:  1

  1. Participant55 (F): wystarczy się nauczyć holenderskiego i trochę postarać,ale wielu polakom się po prostu nie chce i wolą narzekać zamiast się ogarnąć

Translation: One only needs to learn Dutch and needs to strive a bit, but many polish people simply do not feel like it and they prefer to complain than to get a grip.

Date:  March 15 at 9:12am  Likes:  3

Ala is rarely addressed in those exchanges; in that sense they are illegitimate forms of participation in which participants ‘hijack’, so to speak, the broader thematic range of Ala’s update to engage in a discussion among themselves.[11] Such ‘nested’ discussions-within-discussions involve specific participation frameworks. Usually, a handful of participants dominate such divergent lines of action, excluding Ala and others. Yet, it is important to observe that this diverging line of discussion still has its roots in Ala’s main action; it is in that sense a nonlinear extension of it.

In sum, what we see is that over a period of five days, in 193 separate action, a complex social event unfolds in which varying groups of participants create a nonlinear web of actions, most of them rooted directly or indirectly in the main action but several of them involving important thematic and participation framework shifts. If we convert this now to the discourse of context collapse, we see different ‘audiences’ drawn from ‘networked publics’ engage in the interaction, jointly constructing something that looks quite chaotic and may yield confusion and tension. Let us now turn to this issue.

3.2 The rules of a complex game

The question is: given the chaotic mess of diverse actions and shifting participation frameworks, how do participants find their way around all of this? We shall address this question using the simple four-step interactionalist-sociolinguistic methodology mentioned above, and begin by a brief precision to the well-known notion of ‘contextualization’ (Gumperz 1982, 1992; Auer & DiLuzio 1992). Participants in interaction establish the meaning of what goes on in a particular situation by giving off and picking up ‘contextualization cues’. Such cues can be lodged in any and every aspect of communicative behavior: from language or language variety choice, register, style, genre and sequential organization to body posture, pitch, gestures, facial expression and gaze in spoken interaction. In written communication such as the ones we face on SNS, language and language variety (as we have seen) play a role, alongside specific orthographic (or heterographic: Blommaert 2008; Lillis 2013) forms of sign deployment including abbreviations, slang, emoticons and so forth.

Much of what these contextualization cues effectively do is to establish clarity about the action in which one is involved, and more specifically the chronotopic characteristics of the action: the ways in which different actions revolve around different thematic domains, include different kinds of participants and impose different normative patterns of actual conduct (cf. Goodwin & Goodwin 2004; also Blommaert 2015). This is not always a straightforward thing; in a celebrated article, the Goodwins quite long ago (Goodwin & Goodwin 1992) pointed to the fact that quite often, multiple interpretive frameworks (aka ‘contexts’) offer themselves in events, for “within actual interaction it is rare for only a single activity at a time to be on the table. Moreover those present may have competing agendas even within a single activity.” Therefore, according to the Goodwins, “[t]here are great analytical gains to be made by looking very closely at how particular activities are organized” (1992: 96; see also Rampton 2006 for elaborate illustrations). Needless to say, SNS interaction offers its own challenges in this respect, and the event we examine here is a case in point.

Yet, participants appear to be able to draw on a large and quite effective repertoire of forms of interactional conduct for sorting out what really goes on, and for ‘organizing’ their specific parts of the activity, to adopt the terminology of the Goodwins. So, too, in our example. Let us list some of the resources deployed by the participants in our event, starting with the simple ones.

Platform affordances

Facebook, like other SNS, offers a range of technologically configured tools for establishing ‘order’ in interactions. Two such tools demand particular attention here:

  • The system of comments and replies to comments, structuring both a sequentiality to FB discussion and a scaled hierarchical order of superordinate and subordinate comments.
  • The system of name tagging, enabling participants to select and identify direct addressees of an utterance and/or mention indirect addressees.

Both tools have disambiguating functions. The former enables participants to signal thematic coherence and scaled interactional roles. Posting a reply to a comment, for instance, signals a specific (subordinate, low-scale) reaction to the one who posted the (superordinate, higher-scale) comment, while it still, in a more flexible sense, remains inserted in the entire (highest-scale) discussion launched by the update. The assumption in comments and replies is that the superordinate participant is the addressee. Thus, if we go back to the examples above, action 24, the “you” is clearly Ala; and Ala is also the “fucking great journalist” in action 2.

The latter, evidently and explicitly, serves the direct function of addressee selection: from the potentially infinite ‘networked publics’, specific individuals are identified as the direct addressee in interaction. This does not prevent others from interfering, so to speak; but the function of name tagging is obvious, straightforward and effective, as we can observe here:

5.2 Participant22 (F): Participant3 powala mnie Twoja POPRAWNOŚĆ JĘZYKOWA.. A tak szczerze to współczuję takim ludziom jak Ty I Participant13. Miłego wieczoru

Translation: Participant3 I’ am absolutely blown away by your LANGUAGE CORRECTNESS… But honestly I feel sorry for people like you andParticipant13. Have a nice evening

Both tools are abundantly used in our example. We shall discuss an example in which we see both tools in practice in a moment. Let us note, at this point, that while both tools are clear in design and prescribed functions, deviations can be observed. In the event we examine here, people do not always move to the reply-to-comments tool for direct dyadic interaction – see the example of actions 45-47 above in which participants use comments for direct responses and additions to previous turns. And the example of action 38 above shows us that just naming or nicknaming people, rather than tagging them, serves the same function of addressee selection (“Ala” in action 38). Observe also, that the sequentiality offered by these tools may be undone by the non-sequentiality of real actions: a response to an utterance may come several turns after the utterance – other participants having responded more rapidly – which can give rise to misunderstandings as to addressee. We see very few instances of this in our event; one will be documented in the sample analysis below. In general, thus, we do not witness much ‘context collapse’ in our data, and these tools are a major factor in this.

Those platform affordances are technological resources specific to SNS; participants, however, also draw on cultural resources in the organization of their activities.


Goffman’s work is replete with descriptions of how people who are not necessarily profoundly acquainted with each other construct, observe and police rules for engaging in interaction (e.g. Goffman 1961, 1971, 1981). As soon as people have established the nature of a particular social action and the situation in which it will develop, such rules are used continually to maintain ‘order’ in the event. The most common way of doing that is by simply observing the rules of the game and adjusting one’s conduct to the chronotopic normative framework which has been ratified in the action. A more exceptional way is by ‘policing’ the event: explicitly stating or emphasizing the rules, especially when they have been violated, emphatically pointing to more appropriate modes of conduct for transgressing participants, outright excluding them, or qualifying them with labels flagging illegitimate participation.

In our data, a good deal of such policing occurs. Above, we already pointed to the fact that the event consists of a variety of activities, some directly responding to Ala’s main action (and, thus, ‘legitimate’) while others took a more divergent path only indirectly related to the main action. The latter activities, of course, are possible targets for policing, and Ala does quite a bit of that, particularly when she judges participants to be off-topic or negatively biased towards her:

6.9 Ala (F): jak Pani sie nic nie ma do powiedzenia, to proszesie nie mieszac

Translation: If you do not have anything to say, then please do not interfere

Date:  March 14 at 2:02pm  Likes: 0

Other participants do the same; here, Participant 4 directly addresses Participant 3 in response to action 2 (see above):[12]

2.13 Participant4 (M): Co sieczepiasz?nudzi Ci sie to pozmywaj gary.

Translation: Why are you picking on her? If you are bored, then clean the dishes.

Date:  March 14 at 12:54pm  Likes:  31

A little bit further in the same part of the event a female participant ‘rectifies’ a male one about gender bias in interactions such as those (the start of a self-contained ‘nested’ interaction on gender issues, ultimately involving four participants, 2.38-2.47):

2.38 Participant6 (M): Jakoś mnie wcale nie dziwi, że same kobiety komentują ten wątek;-)

Translation: Somehow I am not surprised that only women comment on this thread.

Date:  March 14 at 3:31pm  Likes: 0  Edited: 2

2.39 Participant13 (F): Participant6 wojnę chcesz rozpętać? O co cho?

Translation: Participant6 would you like to wage war? What’s your problem?

Date:  March 14 at 3:32pm  Likes:  0

We note frequent meta-commentaries dismissive of deviant conduct by participants, such as those:

2.50 Participant4 (M): Adek jak sie wyrwał.hehe

Translation: Adek how you blurt out. Hehe

[Adek is Participant18 who posted a reply earlier and who supposedly changed his name through the course of the conversation, eventually deleting his profile]

Date:  March 15 at 6:09am  Likes: 1

2.51 Participant20 (M): nie umiesz czytac idiotko? chyba dziewczyna wyjasnila czemu popelnia bledy. Niektorzy polacy calkiem zapominaja swojej ojczystej mowy!

Translation: Can’t you read idiot [idiot is in its female form]? A girl explained why she makes mistakes. Some polish people completely forget their native speech!

[This comment does not have a direct addressee, but most likely it is directed to Participant3, as the comment appears as a reply to the original post of Participant3]

Date:  March 15 at 9:49pm  Likes: 0

Adding to that, participants appearing overly aggressive or persistently uncooperative are labeled as ‘trolls’ – a well-known category of illegitimate participants on SNS:

  1. Participant57 (M): Tak tak pochwalcie się jak żałośni jesteścieże was biura walą w rogi. Ale oni walą tylko tych co sobie pozwalają na takie traktowanie. Jak ktoś jest sierota w życiu to trzeba to w tv pokazać.

Translation: Yes, yes, boast about how pathetic you are because the offices put something over you. But they only deceive those who allow them to be treated in this way. If someone is a wimp in life, it is necessary to show on TV.

Date:  March 15 at 12:48pm  Likes: 1  Edited: 2

48.1 Participant53 (M):Taki w gębie mocny jesteś? Widać zes robiony na próbę a starzy zapomnieli cie udusić. Internetowy trollu.

Translation: This is how big mouth you have got? It is clear that you have been made in a trial and your folks forgot to judge you. Internet troll

Date:  March 15 at 10:38am  Likes:  3

48.2 Participant52 (M):Pewnie koordynator pierdolony, ktory sam rodakow w dupe ruche na hajs. Participant57 korwa pozal sie boze

Translation: Probably fucking coordinator, who fucks his compatriots in the ass for money himself. Participant57 [addressed with first name], fuck, pathetic.

Date:  March 16 at 7:14pm  Likes:  0

We also see participants informing others of mistakes in perception, i.e. reshaping a ‘correct’ universe of interpretation for the interaction:

2.53 Participant20 (M): Adek sam walisz literowki cycu a innych uczysz

Translation: Adek you make spelling mistakes yourself, and you try to teach others

Date:  March 15 at 10:01pm  Likes: 0

2.54 Participant13 (F): Participant20 post Ali był edytowany

Translation: Participant20 the post by Ala was edited

Date:  March 16 at 8:43am  Likes:  0

A final form of policing is redirecting interaction. As soon as certain boundaries of information are judged to be reached, instructions are given to move to another form of interaction. When participants respond positively to Ala’s invitation to participate in the TV program, she redirects them towards the personal messaging function of Facebook; in a number of instances, this redirection is proposed to Ala by participants themselves, and of course there may have been people who did not participate in the discussion but contacted Ala directly through personal messaging. This function – another technical affordance – is well known and Ala, in the example below, can use slang to identify it:

  1. Participant25 (F): Ja chętnie opowiem 🙂 mam ciekawe doświadczenia:)

Translation: I will gladly tell my story J I have interesting experiences J

Date:  March 14 at 12:49pm  Likes:  1

7.1 Ala (F): chetnie na priw

Translation: Willingly on priv

Date: March 14 at 12:49pm  Likes:  0

Participants insisting on proof of Ala’s authenticity as a Polish journalist equally get redirected to the personal messaging tool; clear boundaries are being marked between what is allowable and what is not in specific formats of interaction:

76.3 Participant63 (F): jezelijestes uczciwa to podaj swoje nazwisko nie tylko Ala

Translation: If you are honest then give your surname not only Ala.

Date: March 19 at 10:25am  Likes: 0

76.4 Ala (F): Kto do mnie pisze dostaje

Translation: The ones who write to me get it.

Date: March 19 at 12:26pm  Likes: 0

Thus, a very broad and powerful range of norms appears to be at play in this complex event, guiding and directing actions, both specifically in themselves and in relation to more general line of action – Ala’s main action, conventions established within the Facebook group, or rules projected onto appropriate interactional behavior on SNS in general. The event is extremely complex, but not unregulated – on the contrary, there is a continuous articulation, implicitly as well as explicitly, of norms of legitimate participation. And there is an across-the-board exploitation of the platform affordances available to participants, supporting the organization of actions. All of these elements serve the purpose of contextualization, of helping participants understanding what goes on in such complex interactions.

3.3 Navigating multiple contexts

Let us now close this empirical examination with a sample analysis in which the comments and observations made above can be synthetically combined.

Recall the warning provided by the Goodwins: we rarely see just one action in real bits of interaction; more often we observe people making sense of complex overlapping and interlocking activities, through elaborate work of contextualization. We have already seen the particular complications generated by SNS interactions: it is scripted discourse, the sequential occurrence of it does not necessarily mirror the interactional sequentiality. Add to this the diversity of participants and the lack or fragmentation of mutual knowledge among participants, and we get an idea of the tasks of contextualization confronting participants.

In our data, the actions 2.26 to 2.36 generously illustrate the complexity of interaction on SNS such as Facebook. Remember that this fragment occurs in the long interaction following action 2 (performed by Participant 3), quoted earlier. Let us look at the full transcript of this part of the event.

Fragment 2.26-2.36

2.26 Participant15 (F):Participant 3 [Adressed with Miss and only first name], pisze się “nie rozumiem”, a nie “nie rozumie”  to tak w gwoli ścisłości co do Pani znajomości języka polskiego. Pozdrawiam serdecznie

Translation: Miss Participant3, you write [“nierozumiem”] and not [“nierozumie”] I do not understand J This is to the preciseness of your Polish Language competences. The warmest greetings

Date:  March 14 at 2:36pm  Likes:  10  Edited: 2

2.27 Participant16 (F): A ty Participant3? [Adressed with the first name only] może pochwalilabys się znajomością holenderskiego??? Wstyd robisz jadąc po kimś kto wyemigrował dawno temu i być może nie miał styczności w dużej mierze z językiem polskim. Znam wielu takich ludzi… Ala życzę powodzenia!!! I wybacz tym zawistnym ludziom.

Translation: And you Participant3? Maybe you would like to boast about with your knowledge of Dutch??? It’s a disgrace to besmirch someone who emigrated long time ago and maybe was not heavily exposed to Polish language. I know many people like that… Ala I wish you good luck!! And forgive these envious people.

Date:  March 14 at 2:35pm  Likes:  10

2.28 Participant10 (F): Participant15 przeczytaj swój ostatni komentarz i zastanów się nad sobą i nad tym co piszesz. Nie widzisz czubka swojego nosa a innym błędy wytykasz. Straszne chamstwo tutaj. Z pustaka cegły się nie zrobi.

Translation: Participant15read your last comment and rethink your own actions and what you have been writing. You cannot see an inch beyond your nose, but you point out others’ mistakes. Terrible boorishness here. You won’t make a brink out of an air-brick

Date:  March 14 at 2:38pm  Likes:  2

2.29 Participant14 (F): Czubka swojego nosa to nie widzi Pani Participant3. Polecam przeczytać sobie posty od początku, bo chyba jakieś nieporozumienie zaszło. Pani Participant15 wypowiedziała się treściwie I kulturalnie.

Translation: Mrs Participant3 is the one who does not see an inch beyond her nose. I recommend to read the posts from the beginning, because I think that that some misunderstanding arose here. Mrs Participant15 expressed herself concisely and politely.

Date:  March 14 at 2:41pm  Likes: 3

2.30 Participant13 (F): Participant14 ale napisała “w gwoli” i czar prysł haha😛

Translation: Participant14, but she wrote preciseness “w gwoli”[it should be written gwoli scislosci] and the spell broke hahah

Date:  March 14 at 2:42pm  Likes:  0

2.31 Participant15 (F): Pani Participant10, dziewczyna napisała posta, radzi sobie jak radzi w języku polskim, ważne jest jednak, że sobie radzi. Wytłumaczyło się bieżącego nawet na wstępie dlaczego pisze tak, a nie inaczej. Została zaatakowana i wyśmiana przez innego członka grupy, który niestety ani poziomem elokwencji, ani poprawnej poprawnej pisowni poszczycić się nie może. Pani więc wybaczy ale zastosowałam stare powiedzenie: kto jest bez winy niech pierwszy rzuci kamień. Pozdrawiam

Translation: Mrs Participant10 [addressed with first name only], the girl wrote this post, she manages the way she can in Polish language, the important thing is however that she manages. She explained right from the start why she writes in this fashion and not differently. She was attacked and derided by another member of this group, who unfortunately cannot pride oneself with the level of eloquence nor correct spelling. You forgive me miss, but I will use here an old saying: the one who is with no guilt should throw the stone first. Greetings.

Date:  March 14 at 2:43pm  Likes: 5

2.32 Participant10 (F): Przepraszam, źle zrozumiałam, myślałam że Pani atakuje Alę. Przepraszam raz jeszcze

Translation: I am sorry, I misunderstood, I thought that you were attacking Ala. Once again, I am sorry

Date:  March 14 at 2:45pm  Likes:  0

2.33 Participant15 (F): Mea culpa Mrs Participant13  ☺ gwoli ścisłości ☺

Translation: Mea Culpa Mrs Participant13 Jgwoli scislosciJ[corrects her spelling error]

Date:  March 14 at 2:46pm  Likes: 1

2.34 Participant13 (F): Participant15 Amen! Pozdrawiam 😉

Translation: Participant15 Amen! Greetings 😉

Date:  March 14 at 2:47pm  Likes:  1

2.35 Participant14 (F): Participant13, również należę do osób, które lubią ogładę wypowiedzi i ortograficzno-gramatyczną poprawność. Ale nie napastujmy tych, którzy tak pisać nie potrafią.

Translation: Participant13, I also belong to people, who like neat utterances/ statements and orthographic-grammatical correctness. But let’s not harass those, who cannot write like that.

Date:  March 14 at 2:47pm  Likes:  0

2.36 Participant13 (F): Participant14 zgadzam się. Nie napastujmy. W ogóle nikogo nie napastujmy. Ale dla mnie było niejasne jak osoba, która nie potrafi pisać jest dziennikarką. I wyjaśniło się. Dlatego cofnęłam mój wstępny hejt 😂

Translation: Participant14 I agree. Let’s not harass them. In general, let’s not harass anybody. For me it was unclear how a person, who cannot write can be a journalist. And it explained itself. That’s why I took back my initial troll/ hater message.

Date:  March 14 at 2:54pm  Likes:  1

There are four main participants (all of them female) in this bit of interaction: Participants 10, 13, 14 and 15. The interaction starts with two consecutive direct reactions to action 2, in which Participant 3 is directly addressed; in 2.26 Participant 15 reacts, and in 2.27 Participant 16 joins in; both get a large number of likes. Action 2.26 the becomes the object of a response (20.28) by Participant 10, directly addressing Participant 15. Both participants will continue their exchange in 2.31 and 2.32. This, we could say, is one conversation.

But in between the turns of the interaction by Participants 10 and 15, Participant 14 has placed a reaction to 2.28, attacking Participant 3 and defending Participant 15. Again, we can see this as an attempt at ‘correcting’ the context, as a form of policing in other words. This intervention, however, is immediately followed by a riposte in 2.30 from Participant 13, pointing out to Participant 14 a writing error in Participant 15’s earlier utterance. This, we could say, is a second conversation.

Action 2.30, next, becomes the point of departure for two more conversations. Participant 15 responds in 2.33 to Participant 13 with “mea culpa”, to which Participant 13 adds “Amen” in 2.34. Remember that participant 14 was mentioned in 2.30, but was only an indirectly addressed participant in that sense. The direct addressee of 2.30, Participant 14, responds in 2.35, and this conversation ends with conciliatory words from Participant 13 in 2.36. Observe how in 2.36 mention is made of a feature of SNS interaction we already encountered: Participant 13 refers to an earlier comment she had removed from the interaction.

Each of the four participants is involved in two separate conversations in this fragment, and the response in one conversation (viz. 2.28 and 2.30) can serve as the point of departure for another one – thus action 2.30 is the point of departure of two separate conversations. Shifts from one conversation into another are swiftly made, mostly by means of name tagging, and no misunderstandings occur, in spite, even, of the odd sequentiality of written texts in the reply tool.

Graphically, the different actions – four interlocking conversations, in which each participant is involved in two of them – can be represented as follows:

ScreenHunter_610 Jun. 25 12.35

Figure 2: Four interlocking conversations.

Each conversation, needless to say, demands its own small chunk of specific context (and, thence, its specific forms of contextualization); each one needs to be marked indexically by participants as separate from others, while still in some way connected to higher-scale ones; and all need to be sustained and concluded in collaboration with people who might be, and often are, strangers in offline life. This complex work is done by the participants without much apparent difficulty. The participants in this bit of SNS discourse (of whom we cannot assume much mutual knowledge) successfully navigated multiple contexts activated in overlapping, interlocking actions, awkwardly occurring as written signs on a screen.

Does context really collapse?

Let us summarize what we have seen in our case analysis.

  1. We have observed a complex and compound social action, the ‘event’ as we called it. This event is non-homogeneous thematically, in terms of modes of interaction and styles of expression, and in terms of participation frameworks.
  2. This means that this event was made up of an intricate web of nonlinearly organized sub-actions: comments, replies to comments and so forth;
  3. This web of actions displayed specific interaction modes and participation frameworks, all demanding normative enactment. Participants appeared to have a high awareness of the rules of the game, most clearly when they explicitly policed parts of it.
  4. Each of these actions showed a relatively unproblematic ‘context’: participants used various mechanisms to solve possible complications in addressee selection, provided useful correcting information to each other, and completed complex interactional tasks.
  5. All in all, participants displayed an acutely accurate sense of the specific actions they were involved in, adjusted their conduct accordingly and sanctioned that of others.

The event, recall, did not take place on a personal Facebook wall; it happened on a forum serving a large community, and it is safe to assume that the administrators of the forum do not personally know every member of the forum. In that sense, the case we have analyzed could have been sensitive – even typically so – to context collapse as a feature of SNS communication. We did not see any evidence of that; we saw a good deal of evidence to the contrary: that participants have a pretty well developed sense of what they are involved in, with whom, and how – their contextualization skills were rather advanced and did not seem to slacken in the face of a lengthy, meandering and often high-tempered SNS discussion. Contexts did not collapse; if anything, they multiplied and expanded into a mountain range.

They are, however, specific contexts characteristic of specific forms of action. Responding to a question involves a different kind of context than launching expletives to a participant whose conduct was judged to be inadmissible; and volunteering to collaborate with Ala in her TV project involved yet another context than challenging her credentials as a Polish journalist. Regarding SNS interaction, to repeat Goodwin & Goodwin’s (1992: 96) words, “there are great analytical gains to be made by looking very closely at how particular activities are organized”. Too general a picture leads to superficial, and sometimes factually unsubstantiated claims and insights. We found such aims and insights in studies on context collapse.

As we said at the outset, it is not our intention here to dismiss or disqualify what scholars have described as context collapse. Our intention was to bring a more precise picture to the table, and what we hope to have shown is that the term perhaps stands for a smaller set of actual SNS communication phenomena than often suggested. Yes, there may be moments where SNS users experience discomfort by the indeterminacy of addressees and that issues of privacy determinate the choice of modes of interactions and of participation frameworks. Let us use the term context collapse for such phenomena. But let us remember that in the data we presented here, addressee selection as well as the segmentation of, and shifts between, participation frameworks did not lead to substantial difficulties. People do usually not address “audiences”, they select specific addressees and, depending on the specific nature of the action they are involved in, are not overly disturbed when others join in.

Let us therefore not use context collapse as a general feature – a defining feature – of SNS communication. Even if the Web 2.0 has shaped tools affording the construction of terribly complex modes of interaction (such as the one we documented here), and even if such degrees of complexity have no equivalent in the offline world of interaction, people actually appear to know their way around. They appear to have built forms of competence for maneuvering such complex interactions, and for determining their possible (and desired) roles in them. The sociation processes shaped by SNS are new and have no precedent. But they can still be described as forms of social action collaboratively performed by people drawing on the available resources and the normative expectations they hold with regard to specific forms of social action. It is this capacity that we call ‘contextualization’, and this capacity appears to be quite flexible, expandable and dynamic when we look at actual instances of SNS communication.


Auer, Peter & Aldo DiLuzio (eds.) (1992) The Contextualization of Language. Amsterdam: John Benjamins.

Blommaert Jan (2008) Grassroots Literacy: Writing, Identity and Voice in Central Africa. London: Routledge.

Blommaert, Jan (2015) Chronotopes, scales and complexity in the study of language in society. Annual Review of Anthropology 44: 105-116.

Blommaert, Jan (2017) Four lines of sociolinguistic methodology. Ctrl+Alt+Dem 8 March 2017. Retrieved 20 June 2017 from

Blommaert, Jan (2018) Durkheim and the Internet: Sociolinguistics and the Sociological Imagination. London: Bloomsbury (in press).

Blommaert, Jan and Anna De Fina (2017), ‘Chronotopic Identities: On the Timespace Organization of Who We Are’, in Anna De Fina, Jeremy Wegner and Didem Ikizoglu (eds) Diversity and Super-Diversity. Sociocultural Linguistic Perspectives, 1-15, Washington DC: Georgetown University Press.

Blumer, Herbert (1969), Symbolic Interactionism: Perspectives and Method, Englewood Cliffs: Prentice Hall.

boyd, dana (2011), ‘White Flight in Networked Publics? How Race and Class Shaped American Teen Engagement with MySpace and Facebook’, in Lisa Nakamura and Peter Chow-White (eds), Race after the Internet, 203–22, New York: Routledge.

Cicourel, Aaron (1973), Cognitive Sociology: Language and Meaning in Social Interaction, Harmondsworth: Penguin Education.

Davis, Jenny & Nathan Jurgenson (2014) Context collapse: Theorizing context collusions and collisions. Information, Communication & Society 17/4: 476-485.

Dugay, Stefanie (2016) ‘He has a way gayer Facebook than I do’: Investigating sexual identity disclosure and context collapse on a social networking site. New Media & Society 18/6: 891-907.

Goffman, Erving (1961), Encounters: Two Studies in the Sociology of Interaction, New York: Bobbs-Merrill.

Goffman, Erving (1971) Relations in Public: Microstudies of the Public Order. New York: Basic Books.

Goffman, Erving (1981) Forms of Talk. Philadelphia: University of Pennsylvania Press.

Goodwin, Charles (2007), ‘Participation, Stance and Affect in the Organization of Practice’, Discourse and Society, 18 (1): 53–73.

Goodwin, Charles & Marjorie Harness Goodwin (1992) Context, activity and participation. In Peter Auer & Aldo DiLuzio (eds.) The Contextualization of Language: 77-99. Amsterdam: John Benjamins.

Goodwin, Charles and Marjorie Harness Goodwin (2004), ‘Participation’, in Alessandro Duranti (ed.), A Companion to Linguistic Anthropology, 222–44, Malden: Blackwell.

Gumperz, John (1982) Discourse Strategies. Cambridge: Cambridge University Press.

Gumperz, John (1992) Contextualization revisited. In Peter Auer & Aldo DiLuzio (eds.) The Contextualization of Language: 39-53. Amsterdam: John Benjamins.

Gumperz, John (2003) Response essay. In Susan Eerdmans, Carlo Previgniano & Paul Thibault (eds.) Language and Interaction: Discussions with John J. Gumperz: 105-126. Amsterdam: John Benjamins.

Karimzad, Farzad & Lydia Catedral (2017) ‘No we don’t mix languages here’: Ideological power and the chronotopic organization of ethnolinguistic identities. Tilburg Papers in Culture Studies, paper 179. Retrieved 20 June 2017 from

Lillis, Theresa (2013) The Sociolinguistics of Writing. Edinburgh: Edinburgh University Press.

Marwick, Alice & danah boyd (2010) I tweet honestly, I tweet passionately: Twitter users, context collapse, and the imagined audience. New Media and Society 13/1: 114-133.

Marwick , Alice & dana boyd (2014) networked privacy: How teenagers negotiate context in social media. New media & Society 167: 1051-1067.

Mead, George Herbert (1934) Mind, Self and Society. Chicago: University of Chicago Press.

Rampton, Ben (2006) Language in Late Modernity: Interactions in an Urban School. Cambridge: Cambridge University Press.

Rampton, Ben (2017) Interactional Sociolinguistics. Working Papers in Urban Language and Literacies, paper 205. Retrieved 19 June 2017 from

Sibona, Christopher (2014) Unfriending on Facebook: Context collapse and unfriending behaviors. Paper, 47th Hawai’I International Conference on Systems Science. Retrieved 19 June 2017 from

Strauss, Anselm (1993), Continual Permutations of Action, New Brunswick: Transaction Publishers.

Vitak, Jessica (2012) The impact of context collapse and privacy on social network site disclosures. Journal of Broadcasting and Electronic Media 56/4: 451-470.

Wang, Xuan & Sjaak Kroon (2016) The chronotopes of authenticity: Designing the Tujia heritage in China. Tilburg Papers in Culture Studies, paper 169. Retrieved 20 June 2017 from

Wortham, Stanton & Angela Reyes (2015) Communication Beyond the Speech Event. New York: Routledge.


[1] The data for this paper are drawn from Gosia Szabla’s fieldwork on the online and offline networks in the Polish community in Belgium and The Netherlands.

[2] The assumption that dyadic spoken conversation is the most ‘elementary’ kind of interaction is a widespread one in several branches of language and communication studies – Conversation Analysis, of course, being the most prominent one. The assumption is however vulnerable to a broad set of critical objections, and we can distinguish some broad lines of critique: (a) a ‘primordialist’ critique in which one might argue that rather than ‘conversation’, ‘narrative’ might as well be considered the most elementary form of interaction (many narratives are conversationally organized), or ‘argumentation’ (many conversations are argumentative); (b) a culture-historical one revolving around the observation that communication cultures today are, almost everywhere, marked by spoken and written forms of communication, where the suggestion that 21st century adolescents in, say, Copenhagen, would still draw their cultural assumptions about communication from spoken forms only is hard to sustain; and (c) an analytical one observing that ‘conversation’, as an activity type, can be broken up into several sub-types such as narratives, question-answer sequences, silence and so forth – ‘conversation’ is too rough a label to cover such diversity. We adopt and shall use this latter objection in our analysis.

[3] Though we cannot belabor this point to any satisfactory degree in the space of this paper, the event presented as a case here is in itself, of course, an abstraction. The Facebook discussion we examine here appeared on a forum, and the histories of themes, modes of interactions, shifting ‘camps’ and conflicts on this forum evidently provide a backdrop – a higher-scale context – for what happened in the case we focus on. The sensitivities regarding the ‘correctness’ of the Polish language, for instance, were frequently articulated on the forum, as was, more generally, the issue of what it means to be a ‘true’ Polish migrant.

[4] Initially, the conversation received 75 likes; on June 22, 2017 (date of data retrieval) there were 73 likes. 11 to the initial message, and 65 to the edited one. There were no new comments after March 19, 2016. The entire data set, in transcript, can be consulted on

[5] The translations from Polish into English were very challenging, due to (a) the features of online SNS writing (abbreviations, erratic case usage and punctuation, emoticons, slang); (b) the thematic salience of ‘correctness’ in Polish language display in these data, which caused participants to edit their comments or willfully play with it. Since this thematic issue is not the core of our argument here, we decided to render the essence of the utterances in our translations, but without trying to reproduce the grammatical errors in English.

[6] Actions 1, 7, 10-13, 15-19, 23-29, 31-32, 34-35, 42-44, 48-49, 60, 74-75, 77.

[7] This observation points to a different feature of SNS interaction: the fact that phases of high-velocity interaction are alternated with phases of slower and more fragmented interaction. We must reserve a fuller discussion of this feature for later work.

[8] Actions 2, 3, 5, 6, 8, 22, 27, 33, 36, 41, 60, 65, 72.

[9] At one or two points in the discussion, participants suggest that Ala is not a journalist at all, that she is a fraud, an unreliable person and so forth. We did not include those items in our count of actions related to language proficiency and its relationship to Polish identity, although there might be a case for seeing it as a further branching of the same theme. As noted earlier, the broad theme of Polish identity (and its defining forms of conduct) is a recurrent one in this Facebook group, and this is where we observe the broader context seeping into this particular event, creating indexical links across separate events (cf. Wortham & Reyes 2015).

[10] A Total of 28 comments (not counting replies) can be listed in this category: actions 15, 19-21, 23-28, 30-31, 39-40, 42-47, 51-52, 54-58, 74.

[11] Similarly, the actions 2.38-2.47 are a self-standing, quite combative discussion between a male and female participant in gender issues in the discussion.

[12] Observe the number of ‘likes’ attached to 2.13 in spite of the gender bias of the utterance. The term for ‘(nit)picking’ introduced in this utterance was adopted by several other participants in later actions.



Author: jmeblommaert

Taalkundig antropoloog-sociolinguist, hoogleraar Taal, Cultuur en Globalisering aan Tilburg University. Politiek publicist.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s