Academic studies of computer-mediated communication (henceforth CMC) are currently flourishing in a variety of social science disciplines. Within the broad linguistic areas of discourse studies and pragmatics alone, the range of topics and methodologies, as well as the pace of advances can only be rivaled by the rapid changes in the ever-growing medium itself. Research fascination with human-to-human interaction via computer networks such as the Internet cannot but be related to the fact that, particularly in the last decade, this type of interaction has taken the world of communication by storm, thus by now having secured a legitimate place for itself next to the older and well-established forms of mediated communication (e.g. telephone, television). As a result, it is hardly an exaggeration to claim that CMC has truly revolutionized social interaction, at least in technologically advanced societies. In the light of the above, this overview will single out the main themes and preoccupations of linguistic studies of CMC with a discourse-pragmatic focus. As will be shown, research on CMC in this area has come a long way in the span of a decade: from treating everything that takes place on the medium as an undifferentiated whole to acknowledging and exploring distinctions amongst computer-mediated discourses that are related to register, style, and genre or, equally, to system specifications that determine the manner, speed, length, ability for storage, etc. of different types of CMC. In similar vein, increasing emphasis is being placed on contextual and particularistic analyses that shed light on how different contextual parameters shape and are invoked in the discourse of various types of CMC. These have brought to the fore the importance of expressive and affectionate aspects of CMC at the expense of an earlier emphasis on transactional and organisational aspects. At the same time, the area has been successful in claiming a legitimate position for itself away from the periphery of linguistic enquiries, as can be attested to by the increase of research presentations and publications forums for CMC. In the spirit of the area’s concerns, the two main journals (Journal of Computer-Mediated Communication and Language @ Internet) are published online. Despite this healthy picture of a rapidly growing area, several questions remain unanswered and avenues for further exploration are still far from exhausted; similarly, any cross-disciplinarities with other domains and topics of the related areas of sociolinguistics, discourse studies, and pragmatics are still at initial stages. Thus, if the calls for systematic research on the various discourse types of CMC (e.g. see Foertsch 1995) currently lack the urgency they had in the ‘90s, they are still far from redundant or misplaced. The discussion attempts to demonstrate the main strands of research in CMC. In terms of textual features of CMC, it was argued that the focus has been on the relations of CMC with the prototypical style of speaking and writing. A consistent finding in this respect is that CMC discourses are hybrids of spoken and written stylistic features, adapted strategically and modified accordingly to suit various medium-specificities. In terms of the formation of interpersonal relations on CMC, there seem to be two interrelated foci of inquiry: the first has to do with the study of expressive features and has reported the abundance of play and performance which, in addition to its aesthetic and creative aspects, seems to reaffirm intimacy relations. The second is concerned with the ways in which participants interact as members of online communities. Intimacy is explored here as part of the creation over time of shared codes and norms of behaviour. At the opposite end of the spectrum, power relations and asymmetry on online communities seem to be intimately linked with degrees of control over and knowledge of the systemic environment in which interactions take place. The volume of the work discussed above attests to a recent decisive break of CMC from an earlier technological determinism, which, informed by social psychology, tended to attribute communication strategies and stylistic features solely and in an absolute way to medium characteristics. In its place, a shift towards a contextualized view of the medium as part of as well as in interaction with situational and sociocultural parameters is expected to be consolidated and expanded. There are many ways in which this can be achieved, not least through in-depth ethnographic studies of CMC environments. So far, one of the main features of the medium, which has more often than not proven to be of pivotal importance is the lack of an immediate, shared physical context and the cognitive as well as interactional constraints and exigencies that this poses. This seems to be linked with the finding that certain code-centered choices (e.g. code-switching, register shifts, repetition) gain enhanced force in CMC (Baym 1995b; Davis & Brewer 1997: 154–157; Georgakopoulou 1997, forthcoming). That said, the range and kinds of medium-preferential cues that participants mobilize in the absence of non-verbal ones remain to be explored. At the moment, what is beyond doubt is that participants in CMC have developed — and will continue to do so — various cues, suited to a CMC environment, that help them create situated interpretations, draw inferences, and make choices keyed to the understanding of their context. A systematic investigation of such cues would enable us to shed light on how the absence of paralinguistic and prosodic ‘contextualization cues’ (Gumperz 1982), so important in FtFs, is made up for in CMC. In addition, it would help us uncover the forms and types of interaction and synergy that different contextualization cues form at different levels. Another avenue for future research in CMC involves comparative studies of different CMC types as well as studies that will seek to explore links of on-line communities and relationships to off-line settings. This is imperative, since, as already suggested, CMC analogies with FtF seem to have monopolized research on the conceptualization of CMC discourse phenomena. To take steps in that direction, CMC instances should not only be seen as mediated and cultural context-shaped and -shaping activities but also as temporalized and relational practices, interrelated with the participants’ trajectory of different types of, past and future, mediated and, where applicable, non-mediated, interactions. In a similar vein, it is to be hoped that CMC research will expand its links and alliances not only with discourse studies of other interactional contexts (professional, institutional) but also with a range of analytical paradigms within discourse studies. It is, for instance, remarkable how little explored CMC instances are from the point of view of genre theory that would shed light on the relative importance and on the conventionalized features of genres that have proven to figure prominently in other contexts of ordinary communication (e.g. storytelling, see Georgakopoulou 2005). That said, the breadth and scope of new directions of research are not easy to predict, as they will be shaped by changes in an ever-growing medium. Herring’s (1999) list of desiderata of system changes includes: two-way message that would enable simultaneous feedback during message production; enhanced logging and archiving capabilities to reduce the incoherence caused by insufficient feedback; facilitated tracking and linking of logically connected and related turns so that there are no intervening irrelevant messages. Finally, the time is ripe for research that will seek to provide a deeper understanding of the relation between CMC and sociocultural processes. An added benefit from that type of research will be the uncovering and celebration of linguistic diversity in on-line communities, where English is refreshingly not the only or main language of interaction. In turn, this will advance our understanding of what is globalized and, equally, what is parameterized and unique to local cultures on CMC.