Some of the arguments offered by Tsal (1985) to challenge the possible independence of affective processes have been made previously (e.g., Lazarus 1982; Watts 1983), and they have been extensively addressed (Ellis 1985; Kleinginna and Kleinginna 1985; Plutchik 1985; Rachman 1981; Zajonc 1984). We will, therefore, respond to them only briefly. Tsal implies that affective reactions are always mediated by some form of representation, whereas we propose that there exist instances where affect is independent of influences (Zajonc 1978; Zajonc and Markus 1982). Tsal's argument against the independence of affect rests on the assumption that in those affective experiences where an antecedent process cannot be directly demonstrated, an unconscious process must have taken place. An unconscious process must have taken place because, by definition, affect is always mediated by some form of representation. But unless Tsal can show just how-and in what form-an unconscious antecedent was involved in a given affective reaction, his proposition is unfalsifiable. Otherwise, if just one case is found where affect is independent of cognition, and Tsal cannot supply even indirect evidence proving the participation of unconscious processes, then his contention loses ground. Several instances that have been cited previously (Zajonc 1978, 1984) satisfy these criteria. These instances do not seem to implicate conscious cognition, and if they do implicate unconscious cognition, it is totally uncertain how such mediation might take place. Tsal himself speaks of innate aversions in the second paragraph of his article. What are the cognitive of innate aversions and what is their form? According to Tsal, innate aversions imply innate cognitions, since by his definition all aversions are mediated by some form of representation. What are these innate cognitions and how can we prove their existence? How can we explain taste aversion in bait shyness established when any possible association be. tween food and the delayed nauseous reaction is oblit. erated by anesthesia (Garcia and Rusiniak 1980)? Ho, can we explain the fact that facial musculature, unac. companied by subjective elements, produces distinctivc autonomic responses that correspond to distinctivc emotional reactions (Ekman, Levenson, and Frieser 1983), and facial muscles may produce subjective effects as well (Zajonc 1985)? These cases might eventually be shown to involvc unconscious cognitions, but thus far no shred of evidence exists that they do. It does not suffice to asserl that these instances necessarily involve cognition. One must show in what way unconscious cognition might be implicated and bring at least some evidence that it is. Similarly, it would not suffice to simply assert thai affect is independent of cognition. We have gone, therefore, well beyond mere assertions. The independence of affect has been substantiated in a number of ways. For example, the existence of neuroanatomical structures that allow for the independent arousal of affect (Zajonc 1984) was among several specific facts supporting the affective independence hypothesis. Thus, the retino-hypothalamic tract is a direct pathway from the retina to the hypothalamus. It allows for an emotional arousal triggered directly by sensory stimulation that requires no higher-order mediation (Nauta and Haymaker 1969). The fact that aggression and rage can be elicited by a direct stimulation of the hypothalamus (Flynn, Edwards, and Bandler 1971) also suggests the possibility of affective reactions without the participation of cognition. And the same holds for the existence of separate centers that control affective and lexical aspects of speech (Ross and Mesulam 1979). Typical of Tsal's argument is his analysis of the WellsPetty (1980) data. He treats these results by the flat assertion that Evidently, (sic) the head movements could have influenced subjects' only (sic) by their association to agreement and disagreement. Thus, these representations must have been (unconsciously) activated by the head movements thereby affecting the subjects' attitudes (p. 359; italics ours). Evidently? The problem with this sort of reasoning is obvious. First, it is asserted that affect must entail cognition, providing no logical or empirical grounds. Then, if an instance appears where the presence of me*R. B. Zajonc is Charles Horton Cooley Distinguished Professor of Social Sciences and Director, and Hazel Markus is Associate Professor of Psychology and Senior Research Scientist, both at the Research Center for Group Dynamics, University of Michigan, Ann Arbor, MI 48106. This work was supported by National Science Foundation Grant BS-81 17977.