According to non-reductive physicalism there are mental properties that are not type-identical with physical properties. Rather, they are determined and realised by them. On most versions the mental properties supervene on physical properties, on some versions, they are properties of one and the same event.
The most influential argument against non-reductive has been the causal exclusion argument. It shows, supposedly, that one cannot have both the intimate relationship between the mental and the physical as required by non-reductive physicalism and novel causal powers for mental events: the causal efficacy of the underlying physical events preempts or excludes the efficacy of the supervening mental events.
Some philosophers distinguished between two versions of the argument. One version is in terms of mental and physical events, the other in terms of instantiations of mental and physical properties. Little significance, though, has been given to this distinction. It has been argued that the argument is equally compelling in both versions. However, a comparison between them can be used to raise some important questions and to highlight some weaknesses of the argument. I shall argue that the weaknesses are considerably serious, and that the exclusion argument is less powerful than it is usually thought to be, and that it is far from being decisive against non-reductive physicalism.
In response to the first version of the argument the non-reductionists may appeal to token-identity. An event is a mental event, if it has mental properties. If it is one and the same event that has both the mental property and the underlying physical property, then the mental event has the same causal powers as the physical event. This move is not available in response to the second version of the exclusion argument, which goes hand in hand with the view that events are instantiations of properties. Identity is ruled out since the instantiation of the mental property and the instantiation of the underlying physical property constitute different events.
But the indicated move, the opponent insists, is of no help for non-reductive physicalism anyhow. Token-identity does not help, since the events in question have their causal powers solely in virtue of their physical properties. And to distinguish between the two versions is, therefore, of no significance.
What has been overlooked, however, is the following. The exclusion argument talks about causal preemption and exclusion. It talks about rival causes. What matters in the case of mental causation, we are told, is the causal role of mental properties. Only events can be causes, properties can only be causally relevant. The exclusion argument applies only to entities that can be rival causes. Properties cannot be rival causes, since they are not causes at all.
But what if a mental event just is the instantiation of a mental property? I shall argue that in that case the possibility that the mental and the underlying physical event are rival causes is ruled out by the close relationship between them. Since one determines and realises the other, they cannot be rivals.