- April
Posted By : Adrian Chan
Theoretical underpinnings of social interaction design

I have been meaning to write this post for a while now in order to provide readers with a big picture introduction to the theory behind social interaction design. I have much of this in the white papers and reading notes on my site, and in a small number of blog posts here. In that many of my posts are anchored in social theory but don’t make explicit reference to it, I thought an overview might be in order.

As my readers will know by now, I view social interaction design as a field that seeks user-centric descriptions of experiences and behaviors on social media, with an eye on emergent social practices. I see user interactions as occurring between users, mediated by social tools — not as interactions with the tools only. Social practices, in the view of social theory, involve users who know what they’re doing, what’s going on, and how to participate. Not on the basis of what’s on the screen, at the level of the interface and its design, but in terms of the social activity in which they are involved.

Social theorists usually refer to cultural norms, traditions, routines, and so on to explain the reproduction of structures and systems of daily life. Macro structures are reproduced by means of these daily repetitions. (See: Anthony Giddens, Constitution of Society). I borrow this idea and apply it to social media: individual user actions add up to social activities and practices. Users produce and reproduce the social reality of social media. Social tools in which users are engaged can be construed as social systems.

I view social interaction design as a paradigm shift in design thinking for two reasons. First, because descriptions and explanations of user activities focus on user experiences, not on design. Second, because the model involves two or more users interacting with each other, not one user interacting with a software application. (It will be said that UX focuses on user experience, and it does, but often as a reflection of design, or as behaviors “influenced” by design. When I anchor social action in the user, I really mean that; and both psychology and social theory are as helpful to the designer in this respect as design skills.)

When social interactions are mediated by software, design is structuring and organizing. Experiences are organized in terms of content and information, navigation, interaction and communication, symbolic actions (think gestures, use of tokens, games, etc). Time and temporality are organized, too, as in the feed revolution and realtime media. (Realtime streams: now and then) I liken social interaction design to urban planning: the use of architecture for social purposes, aware of the tendencies architectures have to produce social outcomes. In short, architecture as social system constraints. At the micro level the user experience may still involve individual user interests, needs, goals and so on. But taken together, social practices correspond to macro effects of individual “uses,” or activities. (I just killed a social game mechanic)

Social theory has been addressing the “instrumentalist” perspective on use over the past decades, expanding concepts of action from rational and goal-oriented action to social action. Social action is action in which the “other” is addressed in the user’s acts. This seems to me the right way to approach social media. I don’t think user actions are explained as interactions with an application, its design or interface, but are rather social acts that explicitly or implicitly, directly or indirectly, intend social consequences.

Communication theory then comes into play when we realize that most social interaction on social media is communication. But there are mediated acts, using tokens, design forms, social objects, and rich media as elements of communication. (SxD: Social Objects ) Things can be said and done online that do not require use of words, or “utterances.” I view this as a rich field for interaction designers for two reasons: first, these elements need to be designed, along with the activity contexts that help to stabilize their meanings. And secondly, because these elements create new ambiguities, new possibilities for both expression and for interpretation.

Interpretation has a special place in social media. Social theory has a long history in the practice of interpretation, dating back to textual interpretation and then through to the interpretation of not only texts, but acts. Interpretation involves meaning, not information. I like this also because I think the best approach to social interaction focuses on the production of meaning in meaningful acts and “exchanges,” not information, data, or “content.” (Social Interaction Design: Leaderboard)

All social action, social theory tells us, is doubly-interpreted. “You know that I know that you know…” and so on. Two subjects in interaction are involved in what’s called “intersubjective” communication. In social media, this communication is mediated not only by language, but by media themselves. As users, we take the medium, or rather the “application,” into account when interacting with others. The medium fundamentally dislocates action and communication from face-to-face co-presence, so it loses both its “situatedness” and its context in place and time.

Social theory uses the concept of “double contingency,” and I find this very helpful in breaking down the two-sidedness of user-to-user interactions. Double contingency is used in meaning-based systems, where the actions of one take into account the likely interpretations of the other. In fact, I believe interaction design for social needs a two-sided model, based loosely on double-sided accounting (debts and credits), or to use linguistics, speaker-hearer. (Re-framing the problem: SxD) For every user action there is an equally valid user reaction (interpretation of action). (I develop this as an argument for coupled activity streams in my proposal for action streams: Action streams: a blue sky proposal)

Context is an issue for all designers, and in particular context of use. The concept of context implies a constructed-ness of experience and of activity. In face-to-face interactions, this is provided by the “situation.” But in mediated interactions, there is no situation; so we have to deal with loss of context. True, contexts are constructed and reconstructed to a certain degree by the application, how it works, what it’s used for, and so on. And these social contexts inform and shape what we do and how we behave, or act. As well as how we interpret the acts of others. The dislocation of action from place, or situation, is accompanied by a discontinuity in time. Because we experience time in terms of past and future, repetition, habit, and routine emerge as organizing forces in our individual experience of how something works and what to expect of our mediated interactions of others. So where context is lost, the meaningfulness of online activity is explained by the user’s own intentions and interpretations. This meaningfulness translates, in psychological terms, as “expectations.” Expectations of the meaning and consequences of online actions.

These expectations are helped out when we have relationships with others, be these friendships or peer connections. Relationships offer a further constraint on social action, and are again a factor in the social interaction model not directly manifest in design. Some relationships are long and lasting, but many, if they can be called “relationships,” are passing and transient. We might just refer to this as communication &mash; as in the “relationships” we can maintain on twitter. Psychologically, the potential for relationships subsists in all communication, and operates sometimes to motivate, other times to discourage, communication. The point I’m making here is that relationships are a structuring and organizing constraint on individual and social interaction conceptually equivalent to design constraints.

Communication theory plays a profound role in my thinking on social interaction design because so much social action becomes or wants to become communication. Acts of communication that are not picked up are not yet action, even if they have that intent. In fact, “social observation” is under-appreciated. A lot of social interaction occurs only after periods of observation. However, observation alone is not captured by media systems and leaves no result (who’s reading those tweets? No pageviews!). This is one of the reasons that so much of the activity in social media involves getting attention. (Social media: the attention economy explained) Attention that is paid but not shown makes no direct contribution to content on social media. The system is biased to an over-production and redundancy of expression; users are required by lack of social and system feedback to put in more, as compensation for the absence of a return look (of recognition and acknowledgment).

Communication that is taken up by another is action. Action has its own structure and organization, and social situations have been observed and described in great detail by many social theorists and psychologists. Social action clearly sits outside the domain of conventional user experience design. But it operates through the applications which facilitate it, and so design matters. But the design approach should consider social practices and outcomes, not only individual user experiences.

Every action captured in social media involves a selection of some kind. This selection is an act involving an item or element. The act may have user intent, but this intent is lost when captured by the social media system, and so the meaning of these selections, or acts, can only be interpreted by others. Social acts online depend in part on the system’s capture, storage, organization, and re-presentation for their likely meanings. Social activities can supply thematic structure, convention, and much more, but ambiguity and noise will always be amplified by the medium. (Is Clay Shirky on complexity too simplistic? )

Much of the richness in designing for social media involves the wide range of meaning possibilities of the elements and selections offered up for social interaction. (Eleven tips on how to apply social interaction design thinking ) We can borrow from social theory in identifying some different core types: signs, symbols, icons and graphical representations, statistics, objects, words, images, and more (i.e. rich media). Skipping over the distinctions between these design elements and their systems of meaning, we should note that these all have two dimensions of meaning: their “objective” meaning and their meaning in context of use (inter-subjective). Or, what they mean, and why they are used. A four star rating may “mean” something, but it’s meaning differs from why a user chooses to rate something. Again, the use in terms of social action is distinct from the meaning as designed into the element. Ambiguity is again amplified by the two possible ways of interpreting the element. Object worlds and action systems offer up divergent explanations of meaningful action; designers work in both. (SXD: The construction of objective relations and operations)

Time is an interesting dimension of social interaction design because it is hard to represent. Action has a serial order in time: not only chronologically, for one act follows the next, but in meaning also, for each act suggests responses and is a response to previous acts. Social media are discontinuous and the user experience of time is individual and separate. This disables some of the seriality of social action but opens up possibilities for re-sequencing and rearranging the content of communication. This reordering is what we do with search results, lists, filters, and many of the other data sorting techniques and can be applied to conversational content, too.

Conversational media create a huge number of opportunities for social media designers, and we are only scratching the surface as of now. Many twitter-related apps are feed readers and facilitate content consumption but not conversational participation. The open space here is in relationships, use of the social graph, user interests and group affinities, and of course the navigation and representation of conversational activities. The temporal dimension of social interaction design can be explored much further, along two axes: the flow of user experience and experience of time (sequence, order, seriality, chronology, but also navigating time periods, snapshots of time, etc); and the linking up of stream items to connect and merge streams for new results. The former involves user experience solutions, the latter, interactive representations of time-based content and interaction.

Linguistics here is an interesting theoretical source, for linguistics, and pragmatics in particular (speech and performance of communication), have long examined the order of linguistically-mediated meanings. One such use of linguistics available to social interaction design is the distinction among different kinds of linguistic expression. We have, for example, differences among requests, invitations, questions, answers, greetings, statements of fact, opinion, recommendation, and many more. All of these come into play, and some efforts have even been made to codify them with microsyntax (in the tweet), activity stream tagging (tagging the feed item), even twitter’s codification of replies and retweets. Important to note here is that language is structure and speech is structuring (of activity and communication). Quora is an excellent example of a simple modification of feed-based communication taking advantage of the structure offered by language (to wit, the Question and the Answer, or more broadly, statements and responses). Quora is an answering machine. (Answer services: satisfying two user experiences)

Social media not only produce interesting possibilities for novel uses of language (its production and consumption) but also for how language and communication reflect the mediated social space. Social media create a kind of “second public,” or audience, one that the user may have in mind even if communication is directed at friends, or at nobody in particular at all. This aspect of the mediated social and public space can create a sense of being seen and of being visible. It is in part responsible for the attention economy and perceptions of influence, status, and “social capital.” And it is certainly a deep motive at work in how and why social media engage users, for it is based on absence, dissociated from situation, and dislodged out of time.

Psychology enters the picture here, for social media not only capture action and communication but create representations of the self, also. We can think of the social interface as having three modes: mirroring (the self), surface (on which is content), and window (through which we interact with others). And related to these three modes we might identify three core user types: Self-oriented, Other-oriented, and Relationally-oriented. In short, people who talk, people who comment, and people who like a feeling of something going on. Psychological factors come into play in how we relate to the medium, to others, to content, and to mediated activities. Relations, not relationships, come into play here in psychological modes of identity and identification, as well as reflection, projection, introspection, internalization, and triangulation. (“You’re OK. How am I?” Reading through twitter.) Some, no many, of these can be applied to social dynamics.

The social dynamics which emerge around social media may be attributed to different kinds of user personality types and behaviors, as these are facilitated by the medium. Experts and their fans, stars and status seekers who want to be around them, critics and their peers, inviters and socialites… There are many social couplings and social formations that bring a social site to life, and whose participation can be served by smart social design. An agile approach to social architecture may come in use here, as a model for design planning as well as product execution and iteration. Social practices can be anticipated and iteration need not be a matter of tweaking, for all interactions are contingent on design, and design can do better than launch and see what happens. (Connections: A reflection the development of social tools)

A table of contents to all blog posts is here.

Just a few suggested readings

Slideshare presentations

White papers



  • The information that Facebook lays bare is fertile ground for the disciplines you cite. I’d add Sociology and Antropology to the fields that contribute to social interaction design. I hope the study of social interaction invigorates liberal arts across the board.

  • It is my great delight to pay a visit to your website and to enjoy your superb post here. I like them quite a lot. I can really feel which you paid significantly attention to those articles, as all of them make sense and really informative.

Leave a Reply