Action without attention

Must an agent use attention in order to guide its purposive bodily movements? Or, to put it more simply, is attention necessary for action? Wayne Wu presents an argument for the claim that attention is necessary for action in Wu 2008, 2011 and 2014. The argument is this: since action requires a solution to the Many–Many Problem, and since only attention can solve the Many–Many Problem, attention is necessary for action. Our aim is to question the first step of this argument: to argue that not all actions require a solution to the Many–Many Problem. To understand this argument, one must appreciate Wu’s ‘Many–Many Problem,’ which is inspired by Alan Allport’s concept of ‘many-to-many possible mappings’ (Allport 1987: 397, see also Neumann 1987). As Wu introduces it, ‘The Many–Many Problem poses a fundamental challenge to bodily agency. Here is one way of raising it: how is coherent action possible in the face of an overabundance of both sensory input and possible behavioural output?Much of the input is irrelevant to the agent’s current goal, much of the output incompatible. Action arises only if the agent reduces this many–many set of options to a one–one map’ (Wu 2008: 552). Wu claims that the problem comes about not because of an abundance of perceptual input and behavioural output, per se, but because of an abundance of these for the agent: ‘the behaviour space used to explicate the problem is a psychological space . . . the inputs that structure the behaviour space are items to which the agent stands in psychological relations’ (Wu 2014: 81). This is an important addition that helps us to understand how to assess the subsequent claims that action requires a solution to the Many–Many Problem and that only attention can provide that solution. For Wu, action is only present in the case that there are multiple targets and responses from the perspective of the agent (see alsoWu 2011: 60). In order to justify his claim that action requires a solution to the Many– Many Problem, Wu uses a thought experiment: he imagines that creatures without the Many–Many Problem would have preset responses to stimuli and thus would only be able to respond via reflex. Wu calls this ‘pure reflex’ to distinguish it from ‘normal human reflex’ (Wu 2014: 89). Wu claims that pure reflex does not fulfill the minimum requirements of agency, and so these creatures do not act: