RESUMO

The present analysis proposes a non-mediational approach to the study of affective phenomena. It starts off with the common recognition that "emotion" is not a technical term. Even so, researchers often treat it as if it were, confusing ordinary language with technical language. This leads to two problems: first, a referentialist bias, according to which we assume emotions to be something unapparent that one must infer and describe; and second, the nominalist fallacy, according to which we assume that emotions have causal effects on actions by the fact of naming them. I review some proposals to solve the problem, among which are some behavioral alternatives. Although these alternatives overcome many of the problems mentioned, they do not completely avoid them. I conclude that a strict non-mediational approach is possible and necessary. It supports the analytical separation of ordinary and technical language. Technical language abstracts relevant properties of ordinary language that become relevant parameters to model certain emotions, as they are referred to in ordinary language. I present some possible parameters and examples for consideration and conclude that the non-mediational approach is a plausible alternative that can stimulate research programs to find natural regularities in affective phenomena.