Causation, further themes

5. Causation as a relation

Built into the functionalist theory is the idea that causation is an intrinsic binary relation between events. But some philosophers have disputed that it is a binary relation in the first place. If they are correct, any relational analysis of causation like the functionalist one must be mistaken.

Christopher Hitchcock (1995, 1996), for instance, has argued that the structure of a causal statement is more complicated than that of a simple statement of a binary relation. Causal statements implicitly imply, he argues, a contrast between the situation in which the cause occurs and a situation in which it does not. The precise meaning of a causal statement depends on the specific contrast situation that is implied. According to this understanding, every causal statement should be read as having the form ‘c caused e relative to the contrast situation c’. In terms of this relativity to a contrast, he explains a number of interesting features of causal statements, such as the role of contrastive stress in causal statements. For example, a statement of ‘Jim was arrested because he stole a bike from the store’ has a different meaning from the statement of ‘Jim was arrested because he stole a bike from the store’. Hitchcock analyses the first as saying that Jim was arrested because he stole, rather than did something else to, the bike, whereas the second says that he was arrested because he stole the bike, rather than some other item. The full import of Hitchcock’s observations has yet to be determined, but it might reasonably be thought that they should be accommodated in the pragmatics rather than the semantics of causal discourse (see Pragmatics).

David Lewis (2004) and D. H. Mellor (1995) have gone further in criticizing the view that causation is a relation. Their arguments revolve around true statements in which absences and omissions figure as causes and effects, for example ‘The absence of a fire extinguisher caused the fire to take hold’. They argue that absences and omissions are not events or positive entities of any kind, and consequently, since relations require existing relata, causal statements involving them cannot be construed as describing relations of any kind. Again, it is unclear how compelling this point is. First, it might be argued that causal statements describe relations between facts rather than events, and since there can be negative facts such as the fact that there was no fire extinguisher present, the statement above about the causal effects of the absence of a fire extinguisher does not pose a problem for this version of the relational analysis (see Negative facts). Second, it might be argued that causal statements involving absences and omissions belong to a discourse that is parasitic on causal discourse about events. Indeed, the meanings of causal statements about absences and omissions seem to be adequately captured by counterfactuals about causal relations between positive occurrences. For example, the statement above seems to mean ‘If a fire extinguisher had been present, its use would have caused the extinguishing of the fire’. If so, it could be argued that no great significance should be attached to causal statements involving absences and omissions in view of their secondary status.