Propositional attitudes

DOI: 10.4324/9780415249126-V028-1
Version: v1,  Published online: 1998
Retrieved March 07, 2021, from

2. Propositional attitude states

If beliefs, desires, intentions, and the like, are states of subjects, what kind of states are they? A plausible view is that they are functional states – states defined or determined or specified by their functional roles. According to functionalist theories, propositional attitude states are fully determined by their causal relations to one another, to perceptual inputs, and to behavioural outputs (see Functionalism). So, for example, what makes a certain state a state of believing that there are Cheerios in the cupboard turns on its being typically caused by packets of Cheerios, or more precisely, on the role it plays in the processing of perceptual input – say, guessing the contents of that visually presented cardboard box in the cupboard; on the way it leads to other states – say, believing that there is cereal in the cupboard; and on the way it combines with hunger and desire to eat some cereal to produce behaviour that leads to eating.

There are, however, two different ways of thinking of functional states: as role states or as realizer states. Realizer states are first-order states that stand in the functional roles; they are the particular states which play the roles. Role states are second-order states: the state of being in, some first-order state playing a certain kind of functional role. So, in the case of belief, say, there is the first-order realizer state (the state which actually fills the belief role – the state which mediates between perceptual inputs and behavioural responses in the way distinctive of belief); and there is the second-order role state (the state of being in some state or other which realizes the belief role – the state of being in a state that mediates in the way distinctive of belief between perceptual inputs and behavioural outputs). Should we think of belief in particular and propositional attitudes in general as realizer states or as role states?

The obvious reason for holding that propositional attitude states are realizer states rather than role states is that the propositional attitudes are centrally involved in the causation of behaviour, and it is the realizer states but not the role states that are centrally involved in the causation of behaviour. It is, for instance, the particular internal state which fills the role we associate with believing that snow is white that causes the behaviour distinctive of this belief – for example, uttering the words ‘Snow is white’.

There are, however, also considerations that point the other way, that favour holding, for example, that my belief that snow is white is the role state which is realized by some particular internal state of mine. In particular, it seems that taking propositional attitude states to be role states allows one to formulate psychological generalizations which one would miss if one took propositional attitudes to be realizer states. Consider, for example, the generalization: everyone who believes the world is going to end in ten minutes gets anxious. Suppose that there are actually M realizer states for the belief: N1,N2,…,NM. These states, we may suppose, each play the appropriate causally intermediate roles distinctive of the belief that the world is going to end in ten minutes in somewhat the way that transistors and valves, despite being very different, can both play the roles distinctive of amplification. If the belief state is identical with the realizer states, then the only way to generalize is to treat the realizer state as a disjunction N1 ∨N2 ∨…∨NM. But this disjunction is ugly, and not suitable for the formulation of psychological law. The generalization can be saved by recasting it as one about role states, but then the only way the initial generalization can be a generalization about the belief that the world is going to end in ten minutes is if that belief is a role state. The general point is that the patterns we capture in psychological generalizations relate to role states, so the cost of holding that psychological states are realizer states is that psychological generalizations cease to relate psychological states.

Whether propositional attitude states are best thought of as role states or realizer states, it has seemed plausible to many philosophers that the realizer states are physical states of the brain – or, at any rate, states which supervene upon physical states of the brain (see Mind, identity theory of; Supervenience of the mental). This is put forward as a plausible scientific hypothesis. Only in the brain are there states of sufficient complexity to play the needed roles. It is, though, allowed as a logical possibility that there could have been other (perhaps even non-physical) realizer states which filled the same role.

Some philosophers, however, have worried about whether neural states could be the kinds of things which can be adverted to, even inter alia, in the explanation of rational action. The connections between neural states are merely causal, they argue – playing out the universal laws of physics, or of neuroscience – whereas the connections between states of propositional attitude are rational (and serve to rationalize behaviour). One belief makes another the rational one to have; what one believes and desires may make one action more rational than another; and so on (see Reasons and causes).

However, there are serious problems with this sort of objection. Would the cause of rationality be better served by having the attitudes occurring at random? And surely a decent account of the propositional attitudes places them in the physical world – they are not otherworldly mysteries – and given this, there seems little alternative to placing them in the brain: we know that the propositional attitudes play complex causal roles in the production of our behaviour, and only the brain has states capable of the needed complexity. Finally, it is unclear why one might suppose that functionally defined states cannot have the rationalizing properties which are alleged to vanish from the scientific picture (see Loar 1981 for further discussion).

Citing this article:
Oppy, Graham. Propositional attitude states. Propositional attitudes, 1998, doi:10.4324/9780415249126-V028-1. Routledge Encyclopedia of Philosophy, Taylor and Francis,
Copyright © 1998-2021 Routledge.

Related Searches


Related Articles