Abstract

I present a computational theory of moods that identifies moods with biases in the operations of our cognitive functional architecture. Moods, I argue, are not representational states. They are biases in the operations of functional architecture-level processes such as memory and attention. Changes at the level of functional architecture have global effects on representation-level states and processes in ways that are independent of the semantic content of those states. My theory accounts for the intuition held by most philosophers that moods are 'objectless'. Moods are objectless because they are not intentional, representational states at all. It is also consistent with the empirical data on moods that demonstrate mood-related effects on processes such as memory, attention, problem-solving strategies and categorization. Moods have global effects on our mental lives in virtue of the fact that they are biases in the operations of those processes that shape and influence our thoughts, motivations and conscious experiences. ;I critically examine the traditional, philosophical 'cognitive theories' of affect and argue that this explanatory approach cannot yield a satisfactory account of moods. The failures of the cognitive approach suggest that moods are not representational states, but are processes that underlie and affect our representation level states and processes. Moods affect how we think, not what in particular we think about. This motivates my argument that moods are operations at the level of functional architecture. One of the implications of this view is that moods are cognitively impenetrable, according to Zenon Pylyshyn's criteria. ;I defend my theory of moods against alternative analyses, including Paul Griffiths' higher order functional state account, Robert Thayer's claim that moods are bodily states of arousal, and the view---held by psychologists such as William N. Morris and philosophers such as Michael Tye ---that moods are essentially feelings or sensory experiences. I argue that these analyses do not provide satisfactory accounts of mood. ;The computational analysis of mood I present in my dissertation offers a naturalistic, functional account of moods and places them within the context of other cognitive processes of interest to philosophers and cognitive scientists