“The mature agent… will recognize his relation to his acts in their undeliberated, and also in their unforeseen and unintended aspects. He recognizes that his identity as an agent is constituted by more than his deliberative self.” (Williams 1995, 32)
Bernard Williams clearly thought that the unforeseen and unintended aspects of our actions were important. The child lays dead in the road, the driver killed him—but all the driver intended to do was get to the depot, and he was driving safely and attentively. I want to sketch an account of agency that, I hope, helps to make sense of the way in which the driver was an agent and was responsible for the child’s death.
It’s obvious that none of us have complete control over what we do, we can never set ourselves to do something and guarantee that it comes about. One response to this is the scepticism we get in Nagel’s “Moral Luck” (Nagel 1979), but I’m more drawn to Raz’s response: even though everything that we do depends to some extent on matters beyond our control, we need to distinguish between gambles and purposive action (Raz 2011). For a very inexperienced fisherperson, catching a trout is perhaps out of their control and a total matter of luck; but if they are a competent chef then cooking the trout is not out of their control—even though particular elements, like the oven working properly, are out of their control. Why? Because they are skilled as a chef, and the reason why they succeed is that they have these skills.
Raz thinks the skills that allow us to be responsible for actions or outcomes are skills that are almost-guaranteed to work, those which we can rely on to work out. So, I can be responsible for cooking the trout (alternatively, I can be responsible for the delicious meal). But I’m not so sure that this is the right approach. When David Beckham scores a free kick, he has, say, a 1-in-7 chance of scoring (and I’m being generous), but still he scored, and we need to recognise that. (In American Football, we might not want to say the QB is responsible for the success of a “Hail Mary” pass, but he is for an ambitious pass).
This gives us something of a fallibilist picture of intentional action: you are responsible for an outcome/action if it came about due to your intentional exercise of a skill where this skill has some decent chance of success but might be moderately fallible. I stated this in terms of intention, but really I think this is just an all-round good picture of authorship. In order to author anything, you need more than just an intention—authorship involves manifesting your intention (or values, character, reasoning, etc.)—and this is a picture of how we can manifest these intentions in the world.
There’s something important about the ways in which we reflect our values (reasoning, character… however we construe it) in the world. There is something important about the way in which actions “were someway or other intended [or] show some agreeable or disagreeable quality in the intention of the heart” (Adam Smith, The Theory of Moral Sentiments II.iii.intro.3). Yet a comprehensive account of agency also has to make sense of the other things we do—not just the times we manifest our values, but the times our abilities misfire or fail. Even if the only things that are attributable to us as moral agents are the things we author, there are other things that are attributable to us as agents that do not manifest our values.
When, for instance, David Beckham tries to score but the ball flies a mile wide of the target, we can’t just put that down to luck. It was because he tried to score that the ball flew wide. Likewise, it was because the driver set off for a drive that the child died. It wasn’t because the driver was speeding or drunk, but it was because he was driving. This is not mere causal responsibility, it’s agential: the driver is responsible in virtue of something he did as an agent, namely intentionally trying to achieve something. The driver can escape blame, he can escape any criticism of his conduct, but what the driver can’t escape is that he killed the child.
What I’m getting at is this: in order to be authors, people who can make an impact on the world in line with our values, we must use these fallible abilities. But how we act as agents doesn’t depend on what we intend to bring about; rather, it depends only on the fact that we intend to bring about something (see Davidson, but Williams also subscribes to this) and that something—perhaps something else entirely—gets brought about.
Williams doesn’t say much about the mature agent. But it seems that the mature agent has a proper understanding of her role in the world, and of the ways in which she might, sometimes unintentionally, impact others. If I’m right: we are agents when we make an unintended impact on the world and the mature agent realises this. And the mature agent also realises that this might leave her with a burden to bear, it might force her to make amends and blemish her life in a variety of ways (see John Gardner’s From Personal Life…). The importance of the things we cause as agents but not as authors needs more exploration. If we focus on morality and authorship we run the risk of ignoring the important ways that agential responsibility matters and of neglecting what it means to be responsible as an agent.
There’s much more to say, but I hope this is an interesting start and helps to shed some light on an area of Williams that has been underexplored—not moral judgment, but action. I’d like to hear what you think!
Hi Jake! Thanks for a thought-provoking post. I take the claim that we, as agents, engage in things we know may work out differently than we intended all the time; and that we are therefore responsible for the various ways they turn out. A driver knows there is a chance of an accident, a footballer knows the pass may or may not work out but deploys their skills to maximize the chances of success, and when I help my friends move house and sprain my back while carrying their couch, I can sincerely say I knew what I was getting into and there’s a sense of agential responsibility to the sprained back – indeed, I say “I sprained my back”, attesting my authorship of the sprain.
But this is where things get murky. Let’s adjust the child into, say, an adult cyclist just so things get a bit clearer, and ask, say a truck driver hits an adult cyclist. The cyclist is killed. The cyclist is killed because he was driving. But also, someone may say that the cyclist was killed because she was cycling. Both were following every precaution of safe driving / cycling, yet this accident happened. In such cases, we typically want to know, whose responsibility was this? Who killed the cyclist? (It seems to me that a similar argument can be pressed for the child.)
Another point I wonder about is that it seems to me that the range of foreseeable outcomes restricts this sense of agential responsibility. Does it to you? I’m talking about engaging in something, such as driving a truck, and ending up in a situation so outlandish one could not foresee happening. Or, engaging in a skill while being uninformed about the range of possible outcomes of that skill. Fishers typically know they may or may not catch a trout, ballers know they may or may not make that pass, drivers, cyclists and pedestrians alike know about the possibility of accidents, and (in my example) movers know that muscles may get sprained. But what of agents who are either atypical in the sense that they could not foresee a plausible outcome that is obvious to others, who engage in something about which we do not yet know everything (such as a scientific experiment), or who face outcomes so outlandish that these could not have been anticipated?
Hi Polaris, thanks for your comments!
I don’t think knowledge is central to my argument, though it bolsters it. My argument is that in order to make any mark on the world we must use these fallible ablities, but abilities–in virtue of being fallible–could always make some other mark on the world. The fact that we know that we are fallible perhaps suggests that we accept some sort of risk when we act… though I’m not entirely sure what role I want knowledge to play.
When it comes to foreseeability, I don’t think we need to be able to foresee that specific outcome. Although drivers might always be able to foresee killing someone by accident, I don’t think this condition always needs to be met: we know we are fallible and that we might fail, we need not be able to know the various ways we might fail. And if we are particularly ill-equipped to foresee things, we might still be responsible. Still, I do think there will be limits on what we might be responsible for. Hart and Honoré think that extraordinary natural events stop one from being responsible, and I suspect that this would cover lots of outlandish cases.
In the cyclist case I think we can say both that the cyclist and the driver are responsible for the cyclist’s death. Stephen Perry thinks this weighs against certain accounts of liability, those that say you are liable for anything that might follow from your action, because it is impossible to say *whose* action lead to the outcome. But I don’t think we need unique attributions, Perry’s problem is only a problem if you want to say that the crash was down to *one* person.
Not sure if this has answered all your comments, but would be glad to discuss further!