In this chapter I consider some important structural pressures on computer-mediated social interaction, associated with the computerised and textual medium of CMC. I do not argue that any of these pressures precludes friendships or other relationships, nor that they prevent the use of CMC for fostering political activities such as democratic decision-making. However, I argue that they do collectively make the CMC experience significantly different from relationships in which people spend time physically together. CMC social interactions are unique in the degree and type of mediation involved, and in the types of attention required to understand the circumstances of others, so altering the sorts of activities required to build reciprocal relationships. A parallel with relationships in letters would capture many of the social qualities of CMC relationships.
For example, CMC limits the possibilities for many practices that are characteristically associated with friendship. Spending time together, living through situations together, and sharing common projects are all either impossible or substantially limited by CMC. In turn, CMC directs interpersonal interaction towards other types of activity, namely towards discussions abstracted from any physically shared experience and towards action in the forms of discussions, formulation of public statements and sending of petitions. In particular, social interactions via CMC are more oriented around textual activities than in any other relationships save those conducted in letters. What people write to one another constitutes not only their interaction, but also partially constitutes their social environment, in the absence of spatially located places which participants inhabit in common.
It is important to keep in mind throughout this section my earlier argument (in Chapter One) that people's selves are constituted by and through their ongoing activities, rather than being a fixed substrate of activity. Character as style is argued to be (socially situated) activity or practice, condensed for comprehension into simpler categories of personal types. This claim in turn suggests that verbal activities, such as deliberation, correspondence and discussion are best treated as varieties of activity, rather than as non-activity or as reflective stances that are not themselves active. Reciprocally, any person is more than a linguistic subject, and indeed the subject treated as linguistic (or linguistically interpellated) cannot be but an