Phenomenal Experience and the Perceptual Binding State loscha Bach Harvard Program for Evolutionary Dynamics, Cambridge, MA 02138, USA Abstract. How can a computational model of cognition account for the hard problem of consciousness? This contribution addresses some of our intuitions about the nature of phenomenal experience and the first person perspective, and suggests avenues for their realization in a cognitive architecture. Keywords: phenomenal consciousness, hard problem, cortical conductor theo- ry, attention, binding, cognitive architectures. Dealing with the Hard Problem It seems that if Artificial Intelligence is pursued as a cognitive science, it cannot avoid to account for the arguably most elusive and mercurial property of the human mind: the conscious experience of phenomenal states. While I have tried to account for some of the functional properties of machine consciousness elsewhere (Bach 2018a, 2018b, 2009), this contribution avoids technical and formal arguments and instead tries to offer a brief introduction into some of the most relevant conceptual intuitions with regard to understanding consciousness as a property of an intelligent system. Modeling perception, memory, decision making, reward based motivation provide challenges to cognitive science, yet nothing about these faculties seems mysterious. The same applies to extending Al systems with reflexive and metacognitive capabili- ties. But how could an Al model ever hope to explain the feeling of what-it's-like? David Chalmers (1995) characterizes this as the Hard Problem of Consciousness: The ability of an organism to be the "subject of experience". To further specify what that means, Giulio Tononi and Christof Koch (2015) have offered five axioms, which I would briefly summarize as follows: I. Consciousness is real and actual (for the same reason that compelled Descartes to his famous dictum "cogito, ergo sum"). 2. Consciousness is compositional (for instance, visua