last revised November 3rd, 2004.Other reading: Lowe deals with these issues mostly in his Part III, chapters 8-11. The SEP has Jonathan Schaffer on 'The Metaphysics of Causation', Peter Menzies on 'Counterfactual Theories of Casusation', Phil Dowe on 'Causal Processes', and much else. One important link is to James Woodward 'Scientific Explanation'. Another is Andrew Brennan 'Necessary and Sufficient Conditions'; Norman Schwartz has a useful site on this same topic. Another enormous and disputed area tied up with causation is the analysis of conditional sentences.
From the beginning philosophers have been interested in what we can recognise as causation. Plato provides discussions of what kinds of explanation are possible; Aristotle is still known for his 'four causes', though only one of them is unambiguously in the area of causation as we discuss it today (though the fourth or 'final' cause is also the subject of intense and continuing debate in the philosophy of biology and of human action). But we can follow tradition and start our survey of its many problems with Hume, since his broadly sceptical account provided the foil for most subsequent work, certainly for most of what has been done in recent analytic philosophy.
Hume recognised that, for us, causation is the "cement of the universe": things hang together in what we take to be fairly intelligible ways because they are causally related. But despite this central organising role, what do we actually discover when we observe causal connections? He thinks we would like to discover some special sort of connection or link between cause and effect, something that might underwrite our inferences from one to the other, and make the sequence intelligible. Other philosophers had certainly traded in notions like these, and had offered principles to govern causal relations (no more in the effect than in the cause, etc.). But Hume tells us that we do not, and cannot find such links. All we find is, you might say, one damn thing after another. There is no specially luminous link, there is only the constancy of the various conjunctions of cause and effect: when we have C we always then have E.
Hume offers a psychological theory to explain why we feign more of a connection than dispassionate enquiry can reveal, but for us the main claim is the negative one, that causation is simply a matter of constant conjunction.
Notice also that Hume's account uses a contrast between what we intend to convey by using certain bits of language (that there is a special link) and what is actually going on in the world when we use those bits of language correctly (some sort of constant conjunction). A third issue we might sometimes want to distinguish as well is that of the cues we rely on to tell when to use the language correctly and the logical structure and relations that the language brings with it, neither of which we may be explicitly aware of.
Theories that agree with Hume are designated 'regularity' theories of causation. Nowadays they tend to be expressed in terms of necessary and sufficient conditions, which are themselves usually taken to involve an appeal to the laws of nature. To get a feel for the ideas here, let us consider the fact that oxygen is necessary for human life. That is, if a human is alive he or she is breathing oxygen, or all living humans are breathing oxygen. These claims reflect the laws of nature that govern life in our bit of the universe. Consider next the fact that flying from Port of Spain to Bridgetown is sufficient for getting from Trinidad to Barbados. That is, if you fly from Port of Spain to Bridgetown then you get from Trinidad to Barbados; or all people who fly from Port of Spain to Bridgetown get from Trinidad to Barbados. Notice that, roughly speaking, when x is necessary for y then we have if y then x, or all y are x; when x is sufficient for y we have if x then y, or all x are y.
Using these notions, people have said that a cause is both necessary and sufficient for its effect. But what we find is rather more complicated. Using Mackie's early account, what we find can be summarised as saying that C causes E when C is at least an inus condition of E, where an inus condition is taken to be an insufficient but necessary part of an unnecessary but sufficient condition. To take Mackie's example, a short-circuit may cause a fire, but a short-circuit in the presence of a fire-extinguishing system might not; equally the fire may have been caused by a bomb, or any number of other factors. However, in what happened, the short-circuit is an essential part of that set of conditions for a fire.
One point that is worth noting here is that this account, like most in the literature, plays down our everyday tendency to distinguish the cause from various attendant circumstances that are equally necessary.
The traditional objection to regularity accounts is that they fail to distinguish causal from merely coincidental regularities. Transposed to the nomological level, this becomes the demand for what distinguishes laws of nature from merely accidental universally true claims. What one is inclined to say, and what Hume did say himself, is that when all C are followed by E, then if C had not occurred E wouldn't have either. And thus we are set upon what is known as the counterfactual analysis of causation.
The dominant counterfactual account of causation was that originally advocated by David Lewis, together with his theory of the meaning of counterfactual conditions construed in terms of possible worlds. As Menzies explains,
The central notion of a possible world semantics for counterfactuals is a relation of comparative similarity between worlds. One world is said to be closer to actuality than another if the first resembles the actual world more than the second does.... For now we simply note two formal constraints he [Lewis] imposes on this similarity relation. First, the relation of similarity produces a weak ordering of worlds so that any two worlds can be ordered with respect to their closeness to the actual world, with allowance being made for ties in closeness. Secondly, the actual world is closest to actuality, resembling itself more than any other world resembles it.
In terms of this similarity relation, the truth condition for the counterfactual "If A were (or had been) the case, C would be (or have been) the case" is stated as follows:(1) "If A were (or had been) the case, C would be (or have been) the case" is true in the actual world if and only if (i) there are no possible A-worlds; or (ii) some A-world where C holds is closer to the actual world than is any A-world where C does not hold.We shall ignore the first case in which the counterfactual is vacuously true. The fundamental idea of this analysis is that the counterfactual is true just in case it takes less of a departure from actuality to make the antecedent true along with the consequent than to make the antecedent true without the consequent.
Lewis goes on to define causal dependence. For actual, distinct events, this amounts to Hume's idea that if C doesn't occur then E doesn't occur either. As construed by Lewis, this notion is not transitive, so since he wants causation to be transitive, he defines the causal relation in terms of its ancestral, the idea of a causal chain linking C and E.
While a lot of the discussion has revolved around problematic cases which we will look at shortly, it is worth while noting a serious objection to counterfactual analyses (urged by Lowe, pp. 186-189). The problem arises from the notion of similarity between possible worlds. On the one hand, it seems that in evaluating degrees of similarity we shall often have to depend on assumptions about causal connections, so it would seem to be circular to offer an analysis of causation itself in those terms. On the other hand, we have to decide how to rank different possible worlds, and there seems to be no obvious objective measure available in many cases (Lowe stresses in particular the choice between worlds with different initial conditions and those with different laws of nature). But causation is thought to be an objective matter, not something that depends upon our choices.
One might note also that unless one buys Lewis' realism about possible worlds, there is a remaining problem of what "the counterfactual structure of the world" actually amounts to.
All philosophical analyses provoke counterexamples, and Lewis' is no exception. I shall list some of the main ones, and also some of the main issues that an adequate account of causation needs to deal with.
Other issues that you can find discussed in the SEP articles and elsewhere include:
© Ed Brandon, 2004.
last revised November 3rd, 2004.
Return to 2004-2005 Homepage.