Journal of Mental Deficiency Research, 1991, 35, 374-383 Paradox, reprimand and extinction in adults with mental handicap V. E. WOOD & A. S. CHAMOVE Department of Psychological Medicine, University of Glasgow, Glasgow, Scotland ABSTRACT. To assess the efficacy of paradoxical directives, levels of challenging behaviour during 2 weeks of paradox, reprimand and extinction were compared with baseline levels in four adults with mental handicaps attending a day centre. Paradox was the most effective procedure for reducing the frequency and severity of challenging behaviour by an average of over 70% by the end of 2 weeks and up lo 90% m certain subjects; extinction was least effective. Paradox was most effective with more defiant subjects, when staff rated treatment success as low., when improvement using extinction and reprimand was poorest, and in reducing aggressive behaviour. The present authors suggest the overjustification effect offers an explsnation for the effects of paradox. INTRODUCTION The essence of paradox involves prescribing the very behaviour that the therapist wishes to reduce. Paradox may also involve reframing and restraining, but these elements are not essential to it. Reframing reinterprets the behaviour in a positive way; in restraining, the therapist pretends to counter any change to more positive behaviour. There are compliance- or defiance-based paradoxes (Rohrbaugh et al., 1981). In the former, change occurs when the client tries to obey a paradoxical prescription but cannot, or tries and realises its absurdity; the latter is based on the assumption that the client will oppose the paradoxical instruction. While symptom prescription appears to be the essence of paradox, it is not clear why paradox works. Explanations of other techniques that also use symptom prescription do not seem to offer as convincing an explanation for different types of paradox. (1) Theorists using massed practice suggest that such practice generates reactive inhibition which interferes with the learned habit, such as a tic (Teichman & Eliahu, 1986). (2) Those using implosion would suggest that imagining the fearful event until extinction occurs to all cues in the avoidance serial cue hierarchy would make associations between the stimulus and the response less likely (Greenberg & Pies, 1983; Weeks & L'Abate, 1982) and would increase a sense of self-mastery (Chamove, 1986; McPhail & Chamove, 1989). Implosion is similar to that of stimulus satiation (Ayllan, 1963) and Gestalt therapy's exaggeration technique (Enright, 1970). Whereas, with both of the above, some forms of paradox require no practice at all. (3) Erickson's directive Correspondence: Valerie E. Wood, Departmem of Psychological Medicine, University of Glasgow, 6 Whiitingehame Gardens, Great Western Road, Glasgow GI2 OAA, Scotland. 374