Doctor–parent–child interaction is analyzed in terms of Brown and Levinson's (1978) theory on facework and politeness. The determinism of the Brown and Levinson (B & L) model is discussed in terms of the present data, which show how discourse is a matter of continuous negotiation between participants. For instance, it is shown how the on record/off record distinction is best understood sequentially. When doctors seem to sense that respectful indirectness does not work, they tend to phrase their requests in an increasingly direct fashion. Hence, the full meaning of doctors' directives is revealed only via the outcome of social interaction, regulated by both parties (doctor and patient/spokesperson). This means that utterances must be analyzed sequentially (and not in a mechanistic, static fashion). Moreover, discourse cannot be understood in terms of any unidirectional social determinism. It is also shown here how doctors can talk to parents through children, as it were. Within a joking relationship format with the child, it is possible for a doctor to convey potentially offensive information to the child's parent. Doctors' moves can thus be seen as direct or indirect depending on type of addressee perspective (parent as participant or as side-participant). The pediatric multiparty setting thus highlights the impossibility of a more formalistic application of the B & L model. (Politeness, facework, negotiations, medical discourse, child discourse)

The present investigation examines parental regulation of child discourse in pediatric settings (in the face ofmedical authonty). On the basis of a close reading of 32 transcripts from pediatric consultations, it is shown how the patients (children) are offen excluded from direct interaction with the doctor.A contrastive analysis of different triadic patterns revealed that parents account for great Variation in children 's relative share of the discourse space. Low control parents typically act äs cultural brokers, mediating between doctors' medical framings and children 's everyday framings of consultation events. In contrast, high-control parents more typically appropriate child allocated turns in ways which have to be explained in terms of symbolic control (rather than medical reframings). Such parental control permeates the consultation in terms of control of children 's overall discourse space äs well äs in terms of control of child participation in the opening scqucnces.Parental control both assumes the outward naked form of dominance, äs in what we have called an executive type of control, and less direct forms of appropriation äs in validation types of parental regulation. In the latter case f parents can be seen to reassert their authonty over the child in being the ones who ultimately "ratify'or legitimize their children 's thinking.