An argument for the possibility of conscious robots would have to show that the brain is neither necessary nor sufficient for the possession of consciousness. I will set about giving just such an argument. Proponents of the enactive theory of perception have argued that neural activity doesn t always suffice for the having of conscious experience. They have argued that the body and environment can also play a constitutive role in enabling conscious experience.
In this paper I will argue for the stronger claim that neural activity isn t necessary for conscious experience either. A robot could, I will argue, enjoy phenomenal consciousness. This has been denied by at least one prominent proponent of the enactive theory of perception (see Alva Noe (2005. 230)) who has argued that a robot wouldn t count as a subject of experience. In the absence of a subject of experience, Noe thinks it makes no sense to attribute phenomenal consciousness.
I will argue that on the contrary a robot could be a subject of experience. My argument will proceed in three stages. The first stage argues that a creature is a subject of experience if it has a first-person perspective. I set out some conditions a creature must satisfy if we are to attribute to that creature a first-person perspective. The most important of these conditions is that the representations the creature produces must have reflexive content they must, in a sense I explain, be representations that refer to themselves.
The second stage of my argument uses a variation on Andy Clark s (2000) argument for the conclusion that access implies qualia. I claim that any representation that has reflexive content will be one to which we have access. Clark has argued that access implies qualia, so it follows that a representation with reflexive content will also have qualia.
The final step in my argument will be to show that action-oriented representations (see Clark 1997 for an account of this type of representation) have reflexive content. Many robots that are capable of producing adaptive behaviour do so by means of action-oriented representations. These robots, I will argue, already meet the conditions for having a first-person perspective. Thus robots with a low-degree of phenomenal consciousness I will claim already exist.
My paper will finish by attempting to motivate this conclusion through a reflection on the connection between consciousness and life. Robots that produce adaptive behaviour are models of life. I will argue that because of the connection between consciousness and life these robots are also models of consciousness.