The problem of perception, especially in cases of hallucination and illusion, asks what constitutes phenomenal character when the object we take ourselves to perceive is absent. Intentionalism is often presented as a fourth alternative to sense-datum theory, naïve realism, and adverbialism. This article argues that intentionalism does not offer a genuinely distinct solution to the problem of perception. Any solution it can provide reduces to options already available to non-intentional theories. The argument follows the distinction between approaches that accept a direct object (or an entity constituting phenomenal character) and those that reject such an object.
Arguments that Predictive Processing and 4E (embodied, embedded, extended, enactive) approaches to cognition are compatible often claim either that Predictive Processing is nonrepresentational or that 4E approaches are representational. Such arguments are inadequate because they overlook the variety of representational posits in Predictive Processing, the diverse attitudes of 4E practitioners toward representations, and the unconstrained use of the term 'representation' in cognitive science. No single argument depending on representational 4E approaches or nonrepresentational Predictive Processing can encompass this heterogeneity. The paper illustrates this criticism by analyzing similar arguments from Jacob Hohwy and Michael Kirchhoff.