On our failure to disconfirm what nobody ever said.

We express pleasure that our critics have clarified their theoretical positions, provide additional methodological detail demonstrating that most of the methodological criticisms are mistaken or irrelevant to the hypothesis we intended to test, and convey our regrets that our work provoked a rhetorical rather than an empirical response. The objections to our research on facial response and emotion fall into two broad categories. The first includes the hypothesis itself and the credit for that hypothesis, which no one is eager to accept. The second, and more complicated, includes a collection of methodologica l criticisms, supposedly indicating that this authorless hypothesis may be perfectly true, but that it has been inadequately tested by our experiment. Tomkins (1981) falls mainly into the first category, denying and then ridiculing the hypothesis, with only an occasional swipe at the method; Hager and Ekman (1981) fall mainly into the second category, shying away from statements about the plausibility of the hypothesis, but raising numerous methodological questions; Izard (1981) takes both positions. Our experiment tested whether voluntary facial expressions were sufficient to create or increase the corresponding emotional experience by posing subjects' faces while they watched films that corresponded or did not correspond to the posed emotion and measuring physiological arousal and self-reported emotion. It tested whether the facial expressions were necessary to create the corresponding emotion by posing subjects' faces into expressions that were incompatible with the emotional properties of the film. Our results discontinued both of these hypotheses: Subjects watching a sad film were equally sad whether their facial expression was fearful, sad, nonemotional, or unmanipulated. These results were replicated across