Causation

Alex Broadbent

Pros and cons

Causation features heavily on many reading lists and in many lecture courses. This piece therefore takes a more discursive approach. It is intended (i) to help a prospective dissertation-writer decide whether to study or involve causation in his or her project, and (ii) to provide someone already caught up in a research project on causation with a rough map of the area, including the habitation of at least some of the dragons.

I used to find causation a tedious topic, and then one day I suddenly saw the light, switched PhD topics, and have never looked back. I can thoroughly recommend it as a fruitful research area for a philosophical project of any level. To my mind, the main advantages of studying causation are:

• causation is an absolutely central philosophical topic, meaning you will become well-equipped to contribute to a wide range of debates across the philosophical spectrum;
• because it is central, there is a lot of excellent literature on it;
• causation poses deep and interesting philosophical questions, meaning not only that it is interesting to study, but also that there remain undeveloped areas, despite the quantity of literature;
• it is not a technical subject, and requires no special background in science, maths, logic, etc.

The main difficulties you are likely to encounter are:

• getting to grips with the point of philosophical work on causation – unlike, for example, the mind-body problem, the problem of causation does not always spring out and grab the imagination (this was a problem I had);
• defining a topic in such a large and busy field (another problem I had);
• although causation is not a technical topic, some parts of the recent literature have technical aspects, and negotiating these bits can be challenging (yet another of my problems).

The remainder of this guide is divided into two parts. The first is aimed at helping you with the first difficulty above – getting to grips with the point of working on causation, and getting a rough idea of the territory. The second lists some (overlapping) areas for research, some words on technical or other difficulties you might face, and a handful of references, along with a few personal thoughts about each area, which you should feel free to disregard as ill-informed, incoherent or just wrong, as the case may be.

Hume and his problem

Everyone always says that the modern philosophical study of causation starts with David Hume. Whether that is historically true doesn't matter much, practically speaking; the fact is that Hume's shadow, which lies across much of modern analytic philosophy, is particularly dense over the study of causation. This doesn't necessarily mean that you should start with Hume, however: retracing the steps of the last few hundred years of a particular branch of thought may not be the most efficient way to proceed, despite the fact that the Karate Kid approach is prevalent among philosophy syllabi (formal logic = wax on, wax off). So don't necessarily start with Hume. Nevertheless, you should make an effort to get to grips with some of his writing at some point. You might want to wait until you have done some reading in your chosen area (see below); then reading (or re-reading) Hume will come as a refreshing change, since Hume knocks the spots of most modern philosophical writers in terms of both intellectual vigour and accessible prose. It will also give you a fresh (and not necessarily flattering) perspective on what everyone in your literature is on about. – When you do read Hume, look at Sections III and IV of An Enquiry Concerning Human Understanding; and for some classic commentary, you could look at Chapter 1 of Mackie's book The Cement of the Universe (see Regularity Theories below).

Although I (heretically?) dispute the need to start with Hume, you do need to start with some idea of the problem which modern work on causation sets out to solve, and that problem is Hume's. In brief, and much simplified, the problem is this. Compare two sequences of events.

(1) I strike a match, and it lights.
(2) I strike a match, and my wife walks into the room.

What is the difference between (1) and (2)? – Obviously, in (1), the match being struck is the cause of the match lighting; whereas, in (2), it is a mere coincidence (I assure you) that my wife walked into the room at the moment I struck the match. Now Hume's problem is this: it is very hard to explain further what makes (1) and (2) different. That is, it is very hard to put your finger on the difference between a causal sequence and a mere coincidence. Hume's challenge, taken up by the modern philosophical tradition, is thus to say what that difference is, in a way which seems both informative and true.

Hume's own solution splits these two components. He offers a theory of what we think causation is, namely, some sort of necessity binding certain events together and not binding others. But he argues that our conviction is groundless: there is no such necessity. He offers a theory about the closest thing to it which really exists, namely, exceptionless regularities. Thus Hume distinguishes our concept of causation from causation 'in the objects'. You might see the challenge facing modern theorists as replacing Hume's two theories with just one. The aim is to come up with an informative analysis of our concept of causation which also truly describes some features of the world, or at least which is framed in non-mysterious terms (e.g. in terms of regularities among actual events).

If this seems difficult, don't be alarmed – it is. One of the tough parts of studying causation is that it forces you to confront the relation between conceptual analysis and metaphysics. If you secretly don't know what either conceptual analysis or metaphysics is, then again, don't worry; if you had to know what philosophy was in order to start doing it then there would be no philosophers.

Roughly speaking, 20th century efforts to analyse causation up until around 1970 were mostly regularity theories. They took their cue from Hume and sought to analyse causation in terms of patterns among actual events. The simplest regularity analyses suffer from obvious defects: some exceptionless regularities appear to be mere coincidences (every time this match is struck, my wife walks into the room – granted, it only happened once, but still there were no exceptions); and some causal sequences fail to be exceptionless regularities (everyone knows that matches in general sometimes fail to light). To deal with various problems, analyses became increasingly complex for little obvious gain in explanatory power.

Then there was an explosion of interest in counterfactual conditionals, sometimes referred to as subjunctive conditionals – statements of the form 'If I had been smarter, I would have picked a different topic'. They refer not just to what is the case but to what would be the case. Work on counterfactuals gave rise to the what is probably the dominant contemporary sort of theory of causation – counterfactual theories. Classically, David Lewis suggested that an event c causes another event e if, had c not happened, e would not have happened either. This view ties with a common-sense way to argue for the presence of causal connections in particular cases, which has also been enshrined in law (the sine qua non or but for test for causation). There are some serious difficulties with this view, however. Suppose I have drunk a lot of beer and it has made me drunk. I muse, 'If I hadn't drunk all that beer, I wouldn't be drunk'. But that might well be false: I am attending a friend's birthday party where all sorts of drink is on offer, and if I hadn't drunk beer then there's a high chance I would have drunk at least an equivalent quantity of something else. So the counterfactual is false, but the beer causes me to get drunk. This is an instance of causal redundancy – the beer causes me to get drunk, but it is redundant, in that without the beer I would still have got drunk. Redundancy (particularly a kind called preemption) is often regarded as an insurmountable problem for counterfactual theories. Another kind of problem arises when we consider all the other conditions without which I would not have been drunk, but which we do not normally consider causes. For instance, if I had not been born, I would not have got drunk; but we can hardly blame my mother for my current drunken state.

There has been a great deal of literature about counterfactual theories of causation, and this has led some to call for everything to be thrown out the window in favour of some radical new approach. Four such are contrastive, probabilistic, 'mark' and agency theories. Each of these is discussed briefly below.

Research areas

Here are some areas which might yield fruitful research topics. There are of course many others. I'll just list a few references for each – there are plenty of reading lists on causation. You can also check out the entry in the Stanford Encyclopedia of Philosophy. The following areas are certainly not exclusive.

Regularity theories

Regularity theories are not popular at the moment. This means you might have a difficult time finding comtemporary literature to engage with. However it might also give you an opportunity to come up with something fresh. Be aware, however, that you will have to engage with some of the huge but now rather dusty literature which built up when regularity theories were all the rage. You should start with:

• Hume, David. 1748. An Enquiry Concerning Human Understanding. Sections III and IV.
• Mackie, John. 1974. The Cement of the Universe, esp. Chapter 1. Oxford: Oxford University Press.

Counterfactual theories

People have known there is a relation between counterfactuals and causation for a long time but counterfactual theories really took off with David Lewis, c. 1970. The reason for this is that Lewis developed a technical framework for counterfactuals. So where before, counterfactual analysis of causation looked like substituting one mystery for another, suddenly it looked like reducing one concept to another one. There is no escaping the fact that you will need to get a grip on some of the details of Lewis's semantics for counterfactuals if you want to understand his theory of causation, since it plays an active part solving some of the most obvious problems which face counterfactual analyses. This takes some work, but it is interesting, and the technicalities are not too daunting – it's more just that the ideas are large and difficult.

Start with:

• Lewis, David. 1973. 'Causation'. Journal of Philosophy, 70, pp 556-567. Also in his Philosophical Papers – Volume II (see below).

For more on counterfactuals, have a look at:

• Lewis, David. 1973. 'Counterfactual Dependence and Time's Arrow'. Nous, 13, p 455. Also in his Philosophical Papers – Volume II (see below).
• Lewis – Counterfactuals. Cambridge, Mass: Harvard University Press. Have a look at the first few pages but stop when the going gets tough!
• Fine, Kit. 1975. Review: Critical Notice. Mind, 84, pp 451-458.
• Elga, Adam. 2000. 'Statistical Mechanics and the Asymmetry of Counterfactual Dependence.' Philosophy of Science (Proceedings), Vol. 68, pp S313-S324.
• Bennett, Jonathan. 1984. 'Counterfactuals and Temporal Direction.' The Philosophical Review, Vol. 93, pp 57-91.
• Bennett, Jonathan. 2003. A Philosophical Guide to Conditionals. Oxford: Oxford University Press. A useful guide to a technical area by an excellent philosopher who does not relish the technicalities. In particular see the chapters 12 and 13.
• Goodman, Nelson. 1983. Fact, Fiction and Forecast (4th edition). Cambridge: Cambridge University Press. See Chapter 1 for his 'inference' theory of counterfactuals – a less popular approach than the Lewis-Stalnakar 'worlds-based' theories. Discussed by Bennett (see previous entry).

For discussions of varieties redundant causation, including 'pre-emption', start with:

• Lewis, David. 2000. 'Causation as Influence.' 2004, in Collins, Hall and Paul (eds.) – see below. A revised version of a paper of the same name originally published in The Journal of Philosophy, 2000.
• McDermott, Michael. 1995. 'Redundant Causation.' The British Journal for the Philosophy of Science, Vol. 46, pp 523-544.
• Schaffer, Jonathan. 2004. 'Trumping Preemption.' In Collins, Hall and Paul (eds.) – see below.
• Broadbent, Alex. 2007. 'Reversing the Counterfactual Analysis of Causation.' International Journal of Philosophical Studies, Vol. 15, pp 169-189.

For interest, you might look at another counterfactual theory which failed to gain popularity because it was not backed by a theory of counterfactuals which many found compelling:

• Mackie, John. 1974. The Cement of the Universe. Oxford: Oxford University Press. See Chapter 2.

'Non-Humean' theories

This seems an appropriate place to mention some writers who have argued for a rejection of Hume's approach, in favour of the view that causation cannot be analysed in terms of regularities, counterfactuals, or anything else. They claim that it does exist, though, thus adopting a realist attitude towards it, and then going on to offer various theories about what it is. I am not aware that such views offer much scope for development since, by their very nature, they are prone to lack detail. Nevertheless they pose an interesting challenge to the general project which other, more detailed theories are engaged in. More recent and, to my mind, promising departures from the classic regularity and counterfactual approaches are probabilistic, 'mark' and agency approaches (see below).

• Armstrong, David. 1983. What Is A Law of Nature? Cambridge: Cambridge University Press.
• Tooley, Michael. 1990. 'Reductionism versus Realism.' Philosophy and Phenomenological Research, vol. 50, Supplement, pp 215-56. Reprinted in Sosa and Tooley (eds.) – see below.

Causal selection

This is my pet topic – an area which has received far less attention than I think it deserves. Therefore I highly recommend it. The problem of causal selection is this. I strike a match, and it lights. If there had been no oxygen present, the match would not have lit. But normally we do not say that the presence of oxygen caused the match to light. This is more obvious when we take morally or legally loaded examples: a car swerves onto the pavement, injuring a pedestrian; had the pedestrian not been there, she would not have been injured; but say that the driver caused the injury, and may blame her and punish her as a result. The problem is dismissed by Lewis, who thinks that all causes are equal, and that we discriminate invidiously between them depending on our interests. However several people have questioned whether this attitude really does justice to the fact that our applications of causal concepts are so ubiquitously selective. The problem of selection is particularly vicious for Lewis's counterfactual theory but it is a general problem, arising for regularity accounts too.

General treatments of causal selection:

• Mackie, John. 1965. 'Causes and Conditions.' American Philosophical Quarterly, Vol. 2, pp 245-264. Reprinted in Sosa and Tooley (eds.) and in Sosa (ed.) – see below.
• Mackie, John. 1974... A fuller development of the ideas in the 1965 paper.
• Lipton, Peter. 2004. Inference to the Best Explanation (2nd Edition). London and New York: Routledge. See Chapter 3 for Lipton's theory of contrastive explanation. Most of the material is also produced in his classic 1991 paper 'Contrastive Explanation', which is also worth examining as an exemplary philosophical essay.
• Lipton, Peter. 1993. 'Making a Difference.' Philosophica, Vol. 51, pp 39-54.
• Hart, H.L.A., and Honore, A. 1985. Causation in the Law (2nd edition). Oxford: Clarendon Press. See the first part, on philosophical matters. In particular look at the introductory pages and at the section concerning circumstance in which conditions sine qua non are not causes.
• Lipton, Peter. 1992. 'Causation Outside the Law.' In Gross, H., and Harrison, T.R. (eds.), Jurisprudence: Cambridge Essays. Oxford: Oxford University Press. pp 127-148.

These two pieces engage more closely with the Lewis tradition:

• Menzies, Peter. 2004. 'Difference Making in Context.' In Collins, Hall and Paul (eds.) – see below.
• Schaffer, Jonathan. 2005. 'Contrastive Causation.' Philosophical Review, Vol. 114, pp 297-328.

Transitivity

There has been some discussion of whether causation is transitive – whether, when a causes b and b causes c, a causes c. If it is, then your birth is among the causes of your reading these words; that is not an intuitive view. There is a link between transitivity and causal selection – you might see the view that causation is transitive as a natural extension of the unselective view held by the likes of Lewis.

Lewis-focussed work:

• Hall, Ned. 2004. 'Causation and the Price of Transitivity.' In Collins, Hall and Paul (eds.) – see below.
• Paul, L.A. 2004. 'Aspect Causation.' In Collins, Hall and Paul (eds.) – see below.

For more general discussion:

• Hart and Honore – same sections as mentioned under Causal Selection above.

Absences

Do absences cause? Can my failure to water a plant cause it to wilt? There are broadly two considerations here. The first concerns what it would mean for the nature of the causal relata (see below) if we allowed that absences caused. How can a nothing do anything? The second concerns the problem of causal selection (see above), which is particularly virulent among absences. If you had watered my plant, it would not have wilted: does that mean your failure to water my plant is equally a cause of its wilting? How about the Queen's failure to water my plant?

• Beebee, Helen. 2004. 'Causing and Nothingness.' In Collins, Hall and Paul (eds.) – see below.
• Lewis, David. 2004. 'Void and Object.' In Collins, Hall and Paul (eds.) – see below.
• Mellor, Hugh. 2004. 'For Facts as Causes and Effects.' In Collins, Hall and Paul (eds.) – see below.

The nature of the causal relata

What sorts of things are causes and effects? Events? Facts? Property instances? Many writers take the view that there are more important things to worry about, and adopt a working hypothesis (usually, that causation is among events). However some have argued that the question should be taken more seriously, and have attempted to solve some of the problems arising for theories of causation by arguing for particular views of the causal relata.

• Mellor, Hugh. 2004. 'For Facts as Causes and Effects.' In Collins, Hall and Paul (eds.) – see below.
• Mellor, Hugh. 1995. The Facts of Causation. London and New York: Routledge.
• Paul, L.A. 2004. 'Aspect Causation.' In Collins, Hall and Paul (eds.) – see below.
• Lewis, David. 2004. 'Void and Object.' In Collins, Hall and Paul (eds.) – see below.

Contrastive approaches

Contrastive theories of causal explanation have been around for a while, claiming that we employ contrasts in our explanations (explicitly or implicitly) in order to distinguish explanatory causes from irrelevant ones. Recently there has been an upsurge of interest in the role of contrasts in our ordinary causal talk, though not always terribly well informed by the discussions arising from theories of contrastive explanation. This is a pity, and in my view there is plenty of scope for work here.

Start by reading the Lipton, Menzies and Schaffer entries under Causal Selection.

Probabilistic causation

I should say that probabilistic causation is not a speciality of mine. But it probably should be. The starting point of probabilistic theories is that many causal judgements are probabilistic. They go on to note that the world itself may be fundamentally probabilistic anyway, for all we know, given the currently popular interpretations of quantum mechanics. Probability theories propose to analyse causation in terms of the way the occurrence of a given event affects the probability that another event occurs. The obvious place to start is with Salmon's work, which is excellent. It is not my impression that you need a serious mathematical to study this area fruitfully, only some basic notion of probability theory.

• Salmon, Wesley. 1980. 'Probabilistic Causality.' Pacific Philosophical Quarterly, vol. 61, pp 50-74. Also in his collection Causation and in Sosa and Tooley (eds.) – see below.
• Anscombe, Elizabeth. 1971. Causality and Determination. Cambridge: Cambridge University Press. Inaugral lecture; reprinted in Sosa and Tooley (eds.) and in Sosa (ed.) – see below.
• Mellor, Hugh. 1995. The Facts of Causation. London and New York: Routledge.

Mark theories

As it happens, proponents of 'mark' of 'transfer' theories of causation seem also to be interested in probabilistic causation. Again, this is not my area of expertise, but it is an interesting and developing area.

Mark theorists:

• Salmon, Wesley. 1980. 'Causality: Production and Propogation.' Proceedings of the 1980 Biennial Meeting of the Philosophy of Science Association, vol. 2, ed. Peter Asquith and Ronald Giere, pp 49-69. Reprinted in his collection Causation and in Sosa and Tooley (eds.) – see below.
• Dowe, Phil. 2000. Physical Causation. Cambridge: Cambridge University Press.

On the general project of 'empirical analysis':

• Bontly, Thomas. 2006. 'What Is An Empirical Analysis of Causation?' Synthese, Vol. 151, pp 177-200.

Agency (or manipulation) theories

Another radical departure from the traditional lines of work: agency theories see causation as something we contribute to the world. (Broadly speaking, they are Kantian rather than Humean, if that helps.) Roughly speaking, a cause is something which we consider to be a useful handle by which to manipulate the world, or – by extension – which we can imagine some suitably placed agent manipulating. Interestingly, notable proponents of these views are driven by a background in philosophy of physics; but the literature itself is fairly small, accessible, and non-technical. – A fruitful area, in my view, but you should have some passing familiarity with Lewis's (early) counterfactual theory of causation, just so you have a sense of what these writers are reacting against.

• Menzies, Peter, and Price, Huw. 1993. 'Causation as a Secondary Quality.' British Journal for the Philosophy of Science, Vol. 44, pp 187-203.
• Price, Huw. 1992. 'Agency and Causal Symmetry.' Mind, Vol. 101, pp 501-520.

Causal asymmetries

Not a theory, but a topic of great interest. What is the difference between cause and effect? Some obvious differences seem wrong – for example, you might think that causes always come before their effects, but on consideration that seems less plausible. If I pick up a coffee mug, the cause of its rising is my lifting it; but my lifting and the mug's rising are surely simultaneous. – Even if not, surely they could be, and if you think causes always precede their effects then you think that is impossible. Other answers seem more likely to be true, but less illuminating. Causes make their effects happen, whereas effects do not make their causes happen; but this is not really an advance on the what we already knew – that causes cause their effects. There is some interesting literature on this stuff, also crossing into the direction of time and the thermodynamic asymmetries, which might be interesting if you have a vaguely physics-related background (not compulsory however).

• Price, Huw. 1992. 'The Direction of Causation: Ramsey's Ultimate Contingency.' PSA: Proceedings of the Biennial Meeting of the Philosophy of Science Association, 1992, pp 253-267.
• Dowe, Phil. 1996. 'Backwards Causation and the Direction of Causal Processes.' Mind, Vol. 105, pp 227-248.

There is a series of further papers between Dowe and Price replying to each other which might make an interesting starting point for a project, especially if you have some interest in the philosophy of physics.

Useful collections

• Collins, J., Hall, N., and Paul, L.A. (eds.). 2004. Causation and Counterfactuals. Cambridge, Massachusetts: MIT Press.
• Lewis, David. 1986. Philosophical Papers – Volume II. Oxford: Oxford University Press.

The following two collections have a lot of overlap but each also contains some interesting papers omitted from the other.

• Sosa, Ernest (ed.). 1975. Causation and Conditionals. Oxford: Oxford University Press.
• Sosa, Ernest and Tooley. Michael (eds.). 1993. Oxford: Oxford University Press.