Of Lightweight Crowds and Heavyweight Communities

Krems.
The second round of keynotes at CeDEM 2011 starts with Caroline Haythornthwaite, whose focus is on making sense of online community structures. She begins from a social network analysis perspective, which understands social networks as constituted of relations between actors. Such social networks transcend online social networks, of course; rather, we now need to take a whole-of-system perspective in which social networking takes place across a range of networks, including online networks.

What’s especially important here, too, is a focus on new forms of collaborating and organising; with the shift towards Web 2.0, but also with many other concurrent shifts, there’s been a transition in attitudes and practices towards collaboration. Indeed, Caroline suggests that we’ve entered a Web 2+ period now. Alongside this are shifts towards user-driven practices, the perpetual beta where things are constantly in flux, and where data and information are mashed up and remixed all the time.

How do collective wholes emerge from individual actions, then? What motivates people to participate in crowds, how do individual actions relate to the crowd, and how (and when) do crowds turn into communities? How can movements be organised to engage individuals?

Caroline has researched much of this especially from an e-learning perspective, where such questions also apply – with a particular focus on crowdsourcing. Crowd-based activities are a centralised effort by anonymous strangers, with people dropping in and out. Community activities, on the other hand, are based on mutual knowledge and understanding. The first of these constitutes a lightweight, the other a heavyweight end of the overall spectrum of peer collaboration and production. Weight here refers to the commitment and engagement with the production process, not to the quality of the outcomes themselves.

Lightweight processes depend on people dropping in at random, making small, discrete, and unconnected contributions. What’s necessary here is for the process to be set up and controlled by an institution outside of this crowd of contributors; through this, people are cooriented towards a particular common purpose. On the heavyweight side, there are fewer, diverse, and connected contributions, and named and visible attribution; contributors are fewer and more heavily interlinked and networked individuals, with control internal to the community itself.

Key dimensions of this are authority and control, motivations and coorientation, contributions, contributors and networks, learning and commitment required, and recognition, reputation, and reward. On the lightweight end, examples for this include SETI@Home, which merely requires participants to provide their idle computer time; human sensemaking projects like the NASA Clickworkers project, the Hominid Fossil Venture, or the use of CAPTCHA systems for OCR corrections; knowledge aggregators like del.icio.us or simple contributions to Wikipedia.

There are also hybrid light and heavy forms, of course – including Wikipedia or OpenStreetMap. Basic behaviours in such projects are largely impersonal and allow for random participation. Citizen science projects are also good examples here – with such projects cooriented towards shared goals and ambitions, like contributing to the study of genetics.

Heavyweight examples include communities of practice of all types, open source projects, committed online communities, learning communities, gaming communities, as well as academic disciplines themselves. What are we ‘sourcing’ in such cases, then? Information, in the first place, but also experience and opinion, debate, evaluation, behaviour, and practices.

This model can also be applied to e-democracy, of course. Lightweight, individual actions here might include flash mobs, elections, institutional repositories, citizen journalism, and citizen diplomacy; heavyweight actions are the actions by social movements, such as environmental groups, political parties, open source communities, or open access alliances.

How may we design for such a model, then? At the crowdsourced end, latent tie structures need to be supported: individuals need to be gathered around specific activities, and on those latent ties stronger social networks and activities can then be built; the authority or designer, and the design itself, have great influence over the form of organising that results from such latent tie structures, then. Authorities need to learn to exercise a ‘soft power’, allowing communities (and their leaders) to emerge out of this. Social connections also arise spontaneously, of course, and this also needs to be harnessed.

Lightweight coordination builds on weak ties; heavyweight collaboration builds on a strongly tied core. Lightweight projects are controlled by a central authority, and therefore rely on users’ trust in that authority; heavyweight projects are controlled by the community (and especially its core members). Lightweight projects are assessed quantitatively; heavyweight projects qualitatively; consensus is similarly formed quantitatively (statistically) or qualitatively, by comparison to the community norms.

Motivations are variously personal, shared (internal), or shared (external): personal motivations include stimulation and entertainment, an interest to work, learn, or play, or self-promotion and career outcomes; internal shared motivations are sharing with the community, social presence, and attention to others’ opinions to one’s own work. Such motivations can be observed in a number of projects – including OpenStreetMap or participation in online learning communities.

Issues around trust and the social contract in all of this have been exposed especially in academia recently: academic publishers have been increasingly perceived as no longer being committed to sharing knowledge, but simply to making money; as a result, more academics have moved to publishing their work directly online. Lightweight projects are better suited to questions which are clearly described, but lack an answer; questions which themselves still need to be defined are better addressed by heavyweight approaches.

Communities generally depend on the presence of a strong-tie core, but also include a weak tie periphery, potentially even including lurkers. Such status within the community are sometimes made explicit by community markers, which also generates bragging rights for members – instituting a certain sense of competition within the community that can be harnessed to encourage more active and constructive participation. There is a gameplay element in all of this as well, then. (Social and serious games – like those produced by Zynga – directly utilise such phenomena.) Lightweight groups are easy to join: Distributed Proofreading and Wikipedia are good examples of this. Membership in heavyweight groups is much harder to attain.

Lightweight processes are much more useful for (pre-) structured contributions; heavyweight processes can be harnessed for unstructured contributions, smoothing knowledge and creating a common voice. Both, ultimately, need to be connected, too, and used as appropriate in each case.