The Priority of Attention

Abstract

IT (intentionality) is the major stumbling block to the claim that machines could one day possess true intelligence. The question is not whether machines would be able to produce outputs indistinguishable from those of a person, as proponents of intelligence have traditional ly maintained.1 Searle has shown, rather, that the real question is whether machines could ever be conscious of objects in the way we know ourselves to be.2 That would seem to make it, at least in part, a phenomenological problem. By means of microphenomenology, then, I shall try to show that the basic structure of IT, as detected in reflection, can be analyzed in terms of an even more primitive factor, FA (focal attention). Since attentional processes can be correlated with neurological selectivi y, it will be argued that any machine constructed to perform such processes would be capable of IT. I shall, in short, attempt to support the possibility of creating automata capable of intentional activities by trying to describe how they oc cur in ourselves.3 The term 'intentionality' has many meanings. But the sense of IT in which we ought to be interested, given the question of artificial intelligence, is the ofness or aboutness that thought processes are said to bear to their objects.4 When I reflect on such processes in myself I find them to be a grasping or understanding of something that somehow the process. To say the object transcends the process is only to say it presents itself as independent of the process, not that it necessarily exists, since it could be anything from a dream to an abstraction.5 The task before us, then, is to explain how processes of FA can both a) perform the office of rendering their contents intelligible, and b) present themselves as correlated with objects entirely distinct from those processes. I shall try to make the case that certain attentional strategies are sufficient for man or machine to accomplish both of those requisites, intelligibility and transcendence.