We must recognize that single events have multiple causes

A few years ago now, I went a bit mad. A relationship in which I was deeply invested had just come to a skidding halt, and I was desperate to understand why. Was it my fault? Was it hers? What, precisely, had been the cause?

Racking my brain in search of some kind of coherent answer, I found myself lost in a thicket of competing ‘What if…?’ scenarios. What if I’d been a better listener? What if she’d never met that other guy? What if we’d had a comfier mattress and hadn’t been so tired all the time – could that have been it?

I tortured myself attempting to reduce the frivolous factors to the deeper ones – but the harder I tried, the further the ‘What ifs’ seemed to proliferate. Eventually, I achieved some peace in the recognition that perhaps these accounts did not have to be reconciled in any linear fashion. Of course a single event – such as a relationship, or its demise – could have multiple causes.

The notion of multiple causes finds a varied expression in the history of philosophy. In A System of Logic (1843), John Stuart Mill despairs at the impossibility of picking out a single ‘cause’ from the background ‘conditions’ of an event. Here we might imagine a ball flying through a pane of glass. What was the cause of the breakage? Perhaps the thrower? Perhaps the ball? Perhaps the vulnerability of the glass? Perhaps the effect of gravity? Each of these counts as a condition, but picking out any particular condition as the cause seems arbitrary.

In his essay ‘On the Notion of Cause’ (1912-1913), Bertrand Russell likewise notes that progress in science consists in recognising ‘a continually wider circle of antecedents’ as necessary for the precise calculation of any event. For Russell, however, the scientific pursuit of exactitude soon bottoms out in the ‘mature’ science of physics, according to which differential equations specify the changes in position and velocity of fundamental particles, and ‘causes’ play no role at all.

But to backtrack from Russell’s conclusion, it has to be pointed out that in every other area of science – from biology, to psychology, to sociology – causal enquiry remains alive and well. No one ever stopped looking for the causes of cancer, or the First World War, because of particle physics. To my mind, this rupture between physics and all the rest is best explained by ‘agency’ theories of causation, according to which human beings define causes in terms of ‘handles’ for manipulating events. From the perspective of the physicists’ block universe, in which all human activity lies spread out at once, there can be no such handles for change. From our perspective, however, the notion of causality continues to flourish as we push and pull and prod at the world in search of regularities to exploit.

The agency theory of causation also provides a neat solution to Mill’s dilemma; what separates ‘causes’ from mere ‘conditions’ is our ability to control them. To return to the ball flying through the pane of glass, it is easy to see that some of its conditions are more susceptible to human influence than others. If we wish to avoid the breakage, we can decide not to throw the ball, or to use a softer ball, or to install thicker windows. We cannot, however, manipulate the strength of gravity. The former conditions are therefore causal in a way that the latter is not. These are the ‘handles’ of agency theory. Contra Mill, however, we need not require that such handles be singular. There are always multiple means of getting a grip.

The recognition of multiple causal handles – of a choice about where we place the source of our problems – has the pleasing upshot of pouring cold water on the single-factor fundamentalists who dominate political discourse. What was the cause of the 2008 financial crash? Reckless deregulation! say the socialists. Reckless overregulation! say the libertarians. What is the cause of any given strain on public services? Lack of government spending! says the left. Immigration! says the right. And so forth.

Taking a step back from such controversies, it should be obvious that complex social phenomena always have multiple causes, and we should be suspicious of anyone who claims otherwise. I am not, however, promoting fence-sitting – nothing of the sort. In fact, I believe the change in perspective that renders causes in terms of handles offers up two practical heuristics for navigating causal disputes. First, when it is apparent that there exists a choice of causal handles, advocates of a particular handle must go beyond merely demonstrating that it exists. They will be forced to say why their handle is fairer, or otherwise more desirable, to lean upon. And second, advocates of a particular causal handle will be forced to speak of practicalities. If a cause isn’t tractable, then it is not strictly speaking a cause at all.

For the first heuristic, consider the dispute between conservatives and feminists about the causes of sexual assault. A conservative might content themselves with a counterfactual such as: ‘If women dressed appropriately, there would be fewer sexual assaults – therefore clothing is the cause.’ There might or might not be some factual basis to that claim but, by the lights of my argument, bald facts are beside the point. Feminists have an equally compelling counterfactual (‘If boys were taught to treat women with respect, there would be fewer sexual assaults’) plus an additional argument about why it is fairer to place the burden of responsibility where they do. This should be enough to break the deadlock – or at least to deny the single-factor fundamentalist their rhetoric of objectivity.

Concerning the second heuristic, I have in mind the increasingly clamorous demand that political leaders must name ‘Islam’ (or ‘radical Islam’) as the cause of so many geopolitical ills. I do not deny that religion is clearly a condition of religious terrorism – but is it really a cause? It is trivially easy to assert the counterfactual: ‘If it weren’t for religion, there would be no religious terrorism’ – but since there exists no straightforward means of eliminating religion from one person, let alone billions, the hypothetical is rendered causally impotent. The progressive alternative – to treat religious fundamentalism as an effect of geopolitics – serves to locate the cause where policymakers can keep some traction. This might seem like sheer pragmatism, but as such it exhausts the rationality of cause seeking.

What can be said, then, about my breakup? If each of my ‘What if…?’ scenarios points to a condition of the relationship, then perhaps some subset might have been causal handles. If I had only known at which handle to grasp in order to keep us together, I would have pulled with all my might. But of course there were two of us at the controls, and not every manoeuvre lay open to me. I could blame myself or my ex for any one of those moves we failed to make – but unlike those political disputes, and where a choice among causes is critical to practical decision-making, my relationship is long dead. I can know that its demise had multiple causes, but I needn’t pick between them now. There are no handles on the past.

Joe Boswell

This article was originally published at Aeon and has been republished under Creative Commons.