Polyvocality, according to Bakhtin (1986), is integral to all human communication: our understanding of any text is informed by our knowledge of its history of use. Kristeva's work on intertextuality elaborates this picture by emphasising the relationships between texts at a single point in time (Moi, 1986). Between them, these two influential critics have highlighted the importance both of diachronic and of synchronic textual relationships in understanding language in use.
In the emergent present of any act of communication, the generally unconscious but still shaping memories of history of use are entwined with the configuring effects of the nature of the communication channel. Just as the model of language as a conduit for ideas has been quizzed and found seriously wanting (Reddy,1993), so 'context' can no longer (if it ever did) work as a container metaphor to identify what lies around the linguistic code. The study of language as social interaction demands examination of attempts at intersubjectivity via texts that can only be approached through active engagement with their spatial and temporal characteristics and, therefore, through active engagement with culturally embedded meanings.
In this paper we take data from two kinds of sources in order to explore the issues raised above. Both sets of data are from novice users of a technology communicating with their peers: (a) three- and four-year old children talking with others on a simplified telephone system; and (b) undergraduate students engaged in Internet Relay Chat. Our investigation points up interlocutors' creative deployment of language resources in their encounters with the constraints (understood here as the affordances and limitations) of the mode of communication, and demonstrates ways in which those language resources are permeated by participants' sociocultural understandings. This is evidenced through three lines of inquiry:
• the acquisition and manipulation of socially constructed routines in openings and closings;
• phonological, graphical and other means of language play;
• pursuit of interpersonal goals through negotiation.
Finally it is proposed, partly through this paper as an enacting exemplar, that the study of young children's discourse need not proceed from an assumption of deficit and need not be shunted into (cognitivist) psycholinguistics. From the viewpoint of linguistic or discourse theory as centre ground, data from children's discourse need not be separated off as at best a specialist area, at worst an irrelevance to mainstream theorizing appertaining to issues of linguistics, language and the real world. Engaging with new communication channels entails learning, language development, and hence modification to social identity for all. Consideration of such changes must be at the heart of theorizing discourse.