The difference between systems and networks is, to my mind, that the former are goal-directed and the latter are not. But what does it mean to say that something is goal-directed? "Common sense tells us that biological systems are goal-directed," write the authors (14 page PDF), but "goal-directed actions are initiated and terminated not by environmental features and goals themselves, but by markers for them." That motivates us to want to say that a system embodies a representational state in a way that a network does not. Most of the article looks at the biological process of goal-directedness, but the fun begins about three quarters of the way through as the authors describe the philosophical implications of their findings. In particular, it makes me wonder whether goal-directedness is, in many cases, an epiphenomenon, that is, something that results from the behaviour (that we describe after the fact) rather than a cause of the behaviour. If not, "the correspondence of a representation to what it represents must be a cause of the usefulness of the representation." If it isn't - if it is, say, an innate response - it can hardly be said to be goal-directed.
Today: Total: [Direct link] [Share]
