The Internet provides access to a range of virtual communities accessible by individuals across the world. While social interaction is likely to be a feature of all virtual communities to varying degrees, some virtual communities have as their basis software applications with an explicit focus on supporting “chat.” In this article we examine social interaction in one such communication medium, Internet Relay Chat (IRC). Following a brief review of the literature, we present our research findings on the interplay between context, the virtual representation of individuals, and disinhibited social interaction in IRC. IRC is a communication medium that began in 1988, based on an initial software program written by Jarkko Oikarinen (Harris, 1995). IRC provides a simple text-based virtual environment for synchronous (real time) typed computer-mediated communication between multiple users. An individual connects to an IRC server using a software client. Multiple clients can connect to an IRC server simultaneously. Servers in turn are linked together to provide an IRC network (Randall, 1997). In January 2004, there were more than 1,000 IRC networks in existence with frequently more than a million individuals simultaneously connected across the globe (see http://searchirc.com/networks.php for current statistics). IRC is predominantly used for social interaction within a social setting. Conversation on IRC is often light, flirtatious, and sometimes sexually explicit, supporting socio-emotional communication and the formation of personal relationships (Byrne, 1994; Koh, 2002; Reid, 1991; Surratt, 1996). IRC provides a party-like atmosphere for social interaction (Randall, 1997). Social interaction on IRC can occur in public or private. Public communication occurs within “channels.” Talking to people within the channel simply entails typing a message and pressing the enter key. The message appears on the computer screen of all channel occupants. Appropriate behaviours differ between IRC channels according to the context and norms of each channel (Cheung, 1995). Private communication can occur between any two people connected to the IRC network. Private messages, seen only by the individual to whom they are directed, can be sent to any other IRC user whether or not they are present in the same channel. IRC users are not restricted in the number of channels they can join, or the number of individuals with whom they exchange private messages. In addition to the use of channels and private messaging, IRC users can establish a direct client-to-client (DCC) link to interact directly with each other using a more secure connection. A further way of ensuring private communication is for an individual to set up his/her own channel, automatically conferring the individual with channel operator status. Channel operators have a level of control over both channel and channel users. A channel operator can make the channel invite only, secret or private; limit the number of users in a channel; allow only channel operators to talk; change the channel topic; prohibit messages from outside the channel; bestow channel operator status on others; and remove and ban specified individuals from the channel (Mardam-Bey, 2000). Access to these commands mean that channel operators effectively control social interaction within the channel, with other channel members’ rights and privileges dependent upon the strength of the relationships they develop with channel operators (Nocera, 2002; Paolillo, 1999). These commands work as social control mechanisms within the channel (Koh, 2002). While all communication on IRC is typed, users refer to this typed communication as “chatting” and use typographical features (e.g., use of ellipsis, multiple punctuation, spelling, emoticons, capitalisation for emphasis) as paralinguistic and prosodic cues to imbue oral qualities to typed messages. Overtime, linguistic conventions have developed (e.g., ‘r’ for are and ‘u’ for you) that may be used consistently across channels or be specific to a particular channel or language group (Hard, 2000; Paolillo, 1999; Werry, 1996). Messages are typically short and abbreviated. Werry (1996) estimated the average message on IRC contained only six words, with the length of message and number of channel participants inversely related. In addition to typed messages, “actions” may be performed to provide context and a sense of presence (Bays, 1998; Gelleri, 1998; Werry, 1996). “Actions” are performed using a command and a typed description of an action or expression of emotion narrated in the third-person singular. For example, if an individual using the nickname “slowboat” typed “/me keels over in surprise”, every person in the channel would see on their computer screen “slowboat keels over in surprise”. Newcomers (“newbies”) to IRC do not initially use these typographical and action features, and the process of “pretended orality” needs to be learned (Gelleri, 1998). As multiple users are able to type messages and perform actions within a channel without the requirement for turn-taking, the text that appears on screen frequently contains an intertwining of messages on different topics and actions. All messages and actions are displayed in the chronological order in which they are received by the IRC server. Newcomers have difficulty following and contributing to the intertwined messages in channels (Gelleri, 1998) and may experience difficulty in engaging others in conversation (Rintel & Pittam, 1997). With experience, IRC users develop the ability to follow the conversational sequences and use conventions (e.g., starting messages with the nickname of the person to whom the message is addressed; abbreviating messages in order to reduce the time between conversational turns and maintain attention) to aid in the management of this (Werry, 1996). In order to communicate within this frenetic environment, each individual selects the nickname (commonly referred to as a “nick”) they wish to be known by. The choice of nickname is the primary means of self-presentation, from which others may infer gender, character attributes, and an indication of the role the individual wishes to adopt (Byrne, 1994; Rintel & Pittam, 1997). In addition to the initial construction of gender through nickname, shared cultural conceptions enable the performance of gender through the use of speech, self-referencing pronouns, and behaviour (Leaning, 1998; Rodino, 1997). Female nicknames on IRC attract high levels of attention from users with both male and female nicknames (Rintel & Pittam, 1997), including sexual harassment (Herring, 1999). Over time, the nickname chosen on IRC becomes part of the user’s identity (Bechar-Israeli, 1995). Reid (1991) described IRC as providing a psychological environment for the construction, exploration, and deconstruction of identity, supporting the practice of new behaviours. Similarly, Koh (2002) emphasised IRC as a “play-space for the disembodied self” (p. 225), where users colluded to engage in play, with online identity representing “the electronic ID unleashed” (p. 228). To further explore the dynamics of social interaction on IRC, we conducted qualitative research in IRC using Grounded Theory methodology (Glaser & Strauss, 1967) as part of a larger project on social interaction in virtual communities (Roberts, 2001; Roberts, Smith, & Pollock, 2000, 2002). The results presented here are those specifically pertaining to social interaction on IRC. We examine the interplay between the IRC context, the representation of the individual and the disinhibited social interactions that occur.