Causality is one of the central notions in our conception of the world. We think of the things and events we experience as connected, and causal relations between them is perhaps the most important connection. Thoughts of causation are central to how we think about our own actions, thoughts, responsibilities and relationships. Yet, however common the notion of causation is in our thoughts, it proves to be very mysterious once we focus on it in an attempt to gain a better understanding of it.
As a number of philosophers (including Nicolas Malebranche and David Hume) emphasized, it appears that we do not have any direct experience of causation. We see events follow one another, and often conclude that one caused the other, but it is not as though we perceive some sort of force passing from one to the other. Even in the case of our own actions, we are strikingly unable to explain how a decision to ask a question brings about motion in our bodies.
Even setting such epistemological problems aside, we might wonder what it is we are saying when we say that one thing causes another. We seem to be saying more than that one thing follows the other, but what is that something more? It appears to involve some thought about other events of a similar type (for instance, in saying that a thrown rock broke a window, we appear to commit ourselves to the belief that there is some sort of general connection between throwing rocks and breaking windows), but it is not clear what that involvement amounts to.
This article will present three major historical discussions of causation, some of the major contemporary theories about how to understand the notion, and conclude by briefly describing some philosophical puzzles that directly bear on which theory of causation we should accept. Issues about causation appear in every area of philosophy except logic, however, so a vast number of related issues will not be touched on here.
Among contemporary philosophical discussions, few make more references to historical accounts than does that concerning causation. In particular, the contrast between Kant and Hume is often the starting point for investigations into the general nature of our concept of causation, and our knowledge of causal relations.
Aristotle's well-known treatment of causation in his Physics set many of the terms (often literally) for arguments about causation over the next two millenia. To a modern reader, however, many of his claims about causation seem to fit poorly with our current use of the notion. It is useful to think of his claims about different 'causes' as claims about different bases of explanation of a thing's being the way that it is, of which our contemporary notion is a species.
Aristotle distinguished four types of cause:
The great Scottish philosopher David Hume discussed the notion of causation at length in his Treatise on Human Nature and Enquiry Concerning Human Understanding. Beginning with the empiricist assumption that the content of all of our ideas had to be drawn from experience, Hume set out to determine what the content of our idea of causation is. One thing we normally take to be a central aspect of the idea, Hume claimed (drawing on the work of Malebranche), is the notion of necessary connection. If we believe that something A causes an effect B, we take B to have been a necessary consequence of A - given that A happened in the way it did, it was necessary the B should occur. This necessity is taken to be of a comparable strength to the connection between, for instance, 3+5 and 8.
Yet, turning to experience, Hume was unable to find any such necessary connection. In both the outer and inner worlds, all we experience is a succession of things - nowhere do we sense anything stronger than temporal connections between things. Where then does the thought of necessary connection come from? Hume claimed that our apparent experience of necessary connection was nothing other than the experience of a tendency of our own minds to anticipate consequences based on past associations. For instance, once we have experienced lightning followed by thunder a number of times, our minds begin to expect thunder every time we see lightning. We then simply confuse the inner sensation of our own expectation with an experience of connection - we effectively project a feature of our mind onto the objects.
Hume's analysis has been used as an argument against metaphysics, ideology and attempts to find theories for everything. A.J. Ayer and Karl Popper both claimed that their respective principles of verification and falsifiability fitted Hume's ideas on causality.
The most famous response to Hume's revisionist/skeptical views on our notion of causation comes from the German philosopher Immanuel Kant. In his Critique of Pure Reason and Prolegomena to Any Future Metaphysics, Kant accepted Hume's claim that we could not draw the notion of causation from outer experience, but drew the opposite conclusion from Hume. Whereas Hume combined this claim with the claim that all the content of our ideas had to come from experience, Kant combined the claim with the claim that it is unquestionable that our notion of causation involves genuine necessary connection. From that, Kant concluded that the empiricist claim is false in this case, and that the concept of causation is a priori - not drawn from experience.
Kant believed that more needed to be said in order to respond to Hume, however. For an advantage of knowing that some concept was directly derived from experience was that this yielded the knowledge that we were justified in applying that concept to experience. As an example, we can be sure that we are not misguided in applying our concept 'cat' to the world, for the world is the place from which that concept came. However, if the concept of causation was not drawn from the world, then one might legitimately worry what bases we had for thinking it had anything to do with the world at all. We have not drawn our concept 'magic' directly from experience, for instance, and that is part of the reason why we are not justified in applying it in our experience of the world.
In response, Kant appealed to a different way in which we might be justified in applying a concept in experience; namely, if experience itself is only possible when that concept is employed. As an analogy, consider someone asking how, as a police officer, one is justified in enforcing the law. The answer is that enforcing the law is constitutive of being a police officer, so that there can be no question of the justification of doing such insofar as one is a police officer. Of course, there is the further question of whether one is justified in being a police officer, but that is a separate question. Carrying the analogy over, if application of the concept of causation is necessary in order to have experience, then one can only be unjustified in applying the concept if one is unjustified in having experience. Yet no one, not even Hume, ever questioned the justification of that.
The major challenge for defenders of the Kantian line is then to show how application of the concept of causation is necessary for experience. Kant himself argued for this via the notion of an objective temporal sequence. More specifically, he claimed that experience requires experiencing things has having some temporal order other than the subjective order of one's own perceptions (for instance, the fact that I see one thing after another doesn't automatically entail that one thing happened after the other), but that the distinction between objective and subjective temporal orders requires the concept of causation.
Hume is often taken as a proponent of an analysis of the notion of causation in terms of regularities. On such a view, to say that A causes B is just to say that A-type events are typically followed by B-type events.
Such a view does not obviously fit with many paradigm instances of causation. For instance, it is often said that a fallen lamp caused the Chicago fire, but it is hardly the case that falling lamps typically cause major fires, or even fires at all. In other words, the regularity analysis would seem to discount some events as instances of causation that seemed to be clear instances. Conversely, there seem to be many regular conjunctions of types of events that we do not count as causation. For instance, the event of having a first birthday party might be regularly followed by taking one's first steps, but this is not because the party causes one to learn to walk.
An influential amendment to the regularity theory was suggested by John Mackie. Mackie suggested that all causes must be 'inus condition's of their effects. 'Inus' is an acronym for "insufficient but necessary part of an unnecessary but sufficient condition." Say that a throwing a rock causes the window to break. Now, strictly speaking, throwing a rock is not sufficient to break a window - for instance, there has to be a lack of intermediate walls that would block the rock. But, combined with a set of background conditions (such as a lack of intermediate walls), throwing a rock is sufficient to break a window. Yet this whole set of conditions is not necessary for breaking a window - after all, one could also break a window by crushing it with an anvil from above.
We can see how this amendment would help with the birthday party example. Of the various sufficient conditions for learning to walk, having a birthday party is just not a necessary component of any of them.
The philosopher David Lewis notably suggested that all statements about causality can be understood as counterfactual statements. So, for instance, the statement that John's smoking caused his premature death is equivalent to saying that had John not smoked he would not have prematurely died. (In addition, it need also be true that John did smoke and did prematurely die, although this requirement is not unique to Lewis' theory.) Cast in terms of possible worlds (a notion which Lewis did much to develop), we could also cast the claim that John's smoking caused his premature death as the claim that, in the nearest possible worlds where John smokes, he dies prematurely, and in the nearest possible worlds where he does not smoke, he does not die prematurely.
One problem Lewis' theory confronts is causal preemption. Suppose that John did smoke and did in fact die as a result of that smoking. However, there was a murderer who was bent on killing John, and would have killed him a second later had he not first died from smoking. Here we still want to say that smoking caused John's death. This presents a problem for Lewis' theory since, had John not smoked, he still would have died prematurely. In terms of possible worlds, this means that it is false that in the nearest possible worlds where John doesn't smoke, he doesn't die prematurely. Lewis himself discusses this example, and it has received substantial discussion..
In discussions of the regularity theory of causation, one issue that is brought to the front is how strict the regularities must be. One might tempted by thought that the regularities must be strict - that if A causes B, then all A-type events must be followed by some B-type event. But this is deeply problematic. In this sense, war does not cause deaths, nor does smoking cause cancer. As a result, many turn to a notion of probabilistic causation. Informally, A probabilistically causes B if A's occurrence increases the probability of B. This is sometimes interpreted to reflect imperfect knowledge of a deterministic system but other times interpreted to mean that the causal system under study has an inherently chancy nature. Philosophers such as Hugh Mellor have analyzed the notion of causation in terms of a cause preceding and increasing the probability of the effect. One advantage of this approach is that it allows the full machinery of modern statistics to be brought in, allowing an impressive level of rigor.
The probabilistic approach is not without problems, however. The birthday party example still works as a counterexample: a person's having a first birthday party raises the probability that he will learn to walk soon, but the party is not a cause of the learning to walk. Much of the work done on probabilistic analyses of causation has therefore been directed at finding ways to 'narrow on' in the relevant features.
Another controversial feature of this analysis is that it makes no claims about temporal direction, for the notion of 'increased probability' is by itself insensitive to temporal facts. Some entity's having a funeral dramatically increases the likelihood that it was alive at some point, but in no sense can a funeral be a cause of being alive.
Some theorists have equated causality with manipulability. Under these theories, x causes y just in case one can change x in order to change y. This coincides with commonsense notions of causations, since often we ask causal questions in order to change some feature of the world. For instance, we are interested in knowing the causes of crime so that we might find ways of reducing it.
These theories have been criticized on two primary grounds. First, theorists complain that these accounts are circular. Attempting to reduce causal claims to manipulation requires that the notion of manipulation is more basic than that of causal interaction. But describing manipulations in non-causal terms has provided a substantial difficulty.
The second criticism centers around the many apparent instances of causation that are beyond any possible manipulation by us. For instance, we might claim that the Big Bang caused the spread of matter throughout the universe, even though this is something completely and necessarily beyond any human practical concerns.
Among philosophers, the response to these criticisms has been to see the appeal to manipulation as something other than an analysis of the notion. Instead, the appeal might rather show us something illuminating about the role that the concept of causation plays in our mental lives. These account use manipulation as a sign or feature in causation without claiming that manipulation is more fundamental than causation.
Almost every major philosophical problem has something to do with causation. There are, however, a smaller set of problems that bear a tight relationship to how we should understand causation.
In his discussion of causation, Hume stated that it was an independent necessary condition of causation that the cause temporally preceded the effect. As we saw, Kant also held there to be a very tight connection between causation and time (though, unlike Hume, he held that temporality was not built into the concept of causation). However, most contemporary theories are formulated in ways where this is not a necessary consequence - the notions of counterfactuals, regularities and probabilities can all be applied to pairs of events without regard to which came first.
The question is whether this is a problem for the modern theories, or a virtue. Intuitively, it might seem that all our thoughts about causation involve the cause preceding the effect. However, we seem to have no deep difficulty in making sense out of stories of time-travel, in which something in the future changes something in the past. Further, recent developments in physics have begun to describe the structure of the universe in very strange ways, and it may be philosophically arrogant to claim that any physicist who proposes the possibility of backwards causation is speaking unintelligibly. (For more, see Time).
We naturally assume that our minds and bodies causally interact: decisions cause our legs to move, and hitting our toe causes the feeling of pain. However, these mental and physical events seem, at least at first glance, to be radically different kinds of entities. Because of that, many philosophers have doubted whether it is possible to give laws describing their connection. Other philosophers (such as Spinoza, however, have claimed that it makes no sense to talk of causal connections that do not admit of some general intelligibility. (For more, see Philosophy of Mind).
Stanford Encyclopedia of Philosophy Retrieved December 22, 2007.
New World Encyclopedia writers and editors rewrote and completed the Wikipedia article in accordance with New World Encyclopedia standards. This article abides by terms of the Creative Commons CC-by-sa 3.0 License (CC-by-sa), which may be used and disseminated with proper attribution. Credit is due under the terms of this license that can reference both the New World Encyclopedia contributors and the selfless volunteer contributors of the Wikimedia Foundation. To cite this article click here for a list of acceptable citing formats.The history of earlier contributions by wikipedians is accessible to researchers here:
Note: Some restrictions may apply to use of individual images which are separately licensed.