Multicultural classrooms provide the intercultural pragmatist with a breeding ground to get firsthand information on intercultural communication. In such a context, if English is used as a lingua franca, these classrooms also become an ideal setting to observe the pragmatics of this relatively unknown variety. The present article aims to analyze qualitatively a sample of data produced by a multicultural group of masters students in which English is used as the lingua franca. More specifically, I intend to answer the following research question: In a multicultural class where English is the medium of instruction and students' peer to peer communication, whose pragmatic “rules” are followed? In other words, do these speakers stick to their own cultural pragmatic rules or follow nativelike ones? More specifically, I focus on the speech act of disagreement given its face-threatening nature and its disruptive potential if carried out in what interlocutors might perceive as the “wrong” way. Disagreement was also chosen given the relative paucity of studies on this speech act – as opposed, for example, others like requests or compliments. For this purpose, the 10 students – from very different cultural backgrounds – were asked to carry out a group assignment. Their negotiation and discussion process, however, was computer-mediated via the use of forums rather than face-to-face. This allowed the researcher to collect naturally occurring, spontaneous data in a relatively easy way (without the need for transcription). It also gave the students the opportunity to interact in a more democratic way. In fact, by being an asynchronous discussion online, they did not have to fight for the conversational turn, but all of them had the chance to contribute to the discussion at their own pace. The sample so collected consists of 15,598 words. The limitation of the sample calls for a qualitative rather than a quantitative approach but is valuable insofar as it represents naturally occurring data. Disagreement expressions were classified according to two main categories: strong and mitigated disagreement (following Kreutel 2007; Pomerantz 1984 and Rees-Miller 2000). Inspection of the data reveals that students on the whole show a tendency to avoid strong disagreement whilst favoring mitigated disagreement of different sorts (e.g., use of hedges, asking for clarification, giving explanations, etc.). Moreover, students with high linguistic proficiency displayed a wider range of strategies, following a more nativelike pattern – specifically; they seemed to follow the pragmatic rules of British English. On the other hand, students whose linguistic proficiency was lower also showed a tendency to avoid strong disagreement but were much more limited with regard to their mitigating strategies, favoring the nonnative overuse of expressions of regret and hedges. Other variables, such as familiarity with their partners and their linguistic proficiency, as well as the nature of the task at hand, also played a central role in the students' choice of the common rules of a native variety.