The title of this thesis is inspired by Benedict Anderson's Imagined communities: Reflections on the origin and spread of nationalism (1991/1996). In this book he argues that a certain form of community, the nation, is merely an imagined community, since the vast majority of people who belong to the same nation are unknown to each other and will never meet. In this thesis I have argued that the conception of e-communities as virtual (in the meaning 'non-real') is a result of the particular form that the discourse around them has evolved. This 'virtuality' is therefore in a sense imagined, and as an analytical term it represents more of an obstacle than a tool for the understanding of these types of communities.

Furthermore, arguments have been produced in favour of regarding e-communities as real (not imagined), since real social relations are built up between real people. The fact that these relations are mediated by technology does not alter this fact, a view corroborated by the demonstration that all kinds of communication, even face-to-face, are mediated. This mediation ultimately occurs through the socio-culturally contingent process of converting sense impressions to meaning.

This insistence that e-communities are not imagined goes counter to Anderson's view that what distinguishes different forms of communities is the way in which they are imagined. It is also in opposition to Fornäs' suggestion that 'virtual' communities are imagined (or imaginary) communities (1998: Ch. 5). This opposition is based on keeping imaginary entities and mental representations apart. Communities involve a great deal of mental representation from their participants, and imagination activities are central in some forms of community (e.g. off-line role-playing groups or on-line MUDs), but the salient feature of a community is the existence of real relations between the participants, relations built up during real interaction.

Analyses of community have traditionally been place-centred and frequently focussing on pre-modern communities (or idealised versions thereof). I have given reasons why a shift in emphasis away from place and historical forms and towards a focus on essential and generic properties might be advantageous. Likewise, the notion that face-to-face communication is inherently better than other forms has been contested, leading to a more symmetric view of 'socially mediated' and 'technically mediated' communication.

E-communities have some unique characteristics due to the specific way that communication is mediated by ICT. The technology can be regarded as a filter of social cues; some cues are readily conveyed, some are transformed into other forms, and some are altogether lost. This transformation process has both negative and positive aspects, and the technical limitations of the medium are to some extent bypassed through creative social practices.

The characteristics of e-communities are ultimately rooted in the wider 'off-line' society, but they often acquire distinguishing features that are specific to the medium through which they are formed. It has proven difficult to find fundamentally new features in the on-line realm, features that have no analogous form off-line. This indicates that the factors that influence peoples lives and thoughts most profoundly are to be found in the wider society, even in the case for computer enthusiasts who spend 80-120 hours a week in e-communities (Turkle 1995: 201). On-line activities influence people to some extent, of course, and on-line phenomena do to some degree manifest themselves off-line (this thesis is itself an example of that), but the alleged revolutionary powers of Internet-based phenomena has failed to see the light of day. Instead, any revolution is likely to go the other way, for example through the on-going 're-interpretation' and re-shaping of the Internet as an economic tool.

Such re-interpretation and re-shaping has been a reoccurring feature of the Internet since its conception, and I have pointed out that not only technical developments can be analysed in this way, but also social processes, like the formation of e-communities, follow similar patterns. Infrastructure technology exhibits a particular dynamic pattern in this respect, with standardisation or closure of a flexible artefact on one level giving rise to prolific growth and variation on another. These 'multi-artefactual' dependencies possibly indicate directions for further development of socio-technical theoretical frameworks like SCOT.

Further work may seek to develop the tentative generic community criteria that were arrived at in chapter 3, and give a more thorough empirical foundation for the claim that many on-line social aggregations are indeed communities. How the off-line society becomes mirrored on-line and the nature of this transformation may also be further elaborated. Another issue is the democratic content of e-communities; how do the formal equalities (equality in opportunity - given that one has the socio-technical preconditions to take part) compare with the actual state of affairs (equality in outcome). I have used SCOT fairly sketchily, and a more thorough application of this framework, both for analysing e-communities themselves and for further elaboration of the standardisation aspects is desirable.

To conclude, it can be handy to merge some of the partial views of socio-technical dependencies presented at various points in this thesis. Fig. 6.1 is one way of doing this. Community A, B and C are the broad, weakly connected, overlapping communities based on assorted fundamental socio-cultural and geographic commonalities; they are 'the society', more or less. E-communities arise from these and inherit many of their traits. E-communities can be formed due to the availability of a technological infrastructure, the Internet, which is a technical 'basic service'. This technology is developed by other communities, and the design of the technology is influenced by the socio-cultural paradigm of the society from which they arise. E-communities also use social 'basic services', e.g. language, that has been 'developed' in the fundamental community.

By highlighting all these dependencies and interactions, it is my hope that this thesis will contribute to an attenuation of the swings of the discursive pendulum. E-communities are not revolutionary entities that exist in a metaphysical space separated from the rest of the world. They are not inherently good or bad. They are real social aggregations that share a lot of features with other community forms, and they also have their own idiosyncrasies. In the words of Sherry Turkle (1996), "in five or ten years we won't even be talking about the real and virtual, we'll just have relationships, some of which are on-line and some of which are off-line".