This post comes in two parts, plus an appendix. Part 1 is mainly a revision of what’s already in the lecture, talking through Strawson’s Basic Argument and the “Hurley Response” to make sure that you are up to speed. Part 2 goes beyond and is meant for those with deeper interest. The Appendix explains why I structured the Hurley response the way I did. Ok, here goes.

  • Part 1

Boiled down to its core, the TL;DR restatement of Galen Strawson’s overall argument goes something like this:

Strawson’s Basic Argument (TL;DR version in L06 Slide #29)

  • (1) You are morally responsible for something only if you chose it. (Choice Principle)
  • (2) You are morally responsible for something only if you are also morally responsible for the way you are behind it. (Strawson’s Principle)
  • (3) It’s impossible for you to be morally responsible for the way you are behind your actions. (From (1), (2), by infinite regress.)
  • Therefore: (4) It’s impossible for you to be morally responsible for your actions. (From (2), (3))

As explained in the lecture, Premises (1) and (2) generates an infinite regress, justifying the intermediate conclusion (3). That is, given the two stated necessary conditions for moral responsibility (=(1), (2)), an infinite series of necessary conditions is generated. Since these are necessary conditions for your being morally responsible for doing something, they imply that you are morally responsible for doing it only if you have actually made those choices–the whole infinity of them. And that’s an impossibility–not just very hard, or practically impossible; just impossible.

It’s as if I’ve asked you to count to infinity–the problem there isn’t how fast you can count or how much time is given. If you claim that you’ve completed your task, I will simply ask you: So that last number you counted. Was it an even number or an odd number. (However you answered, I’ll point out that obviously there’s at least one other number you haven’t counted to–the next one. So please continue.) The entire idea of “successfully counted to infinity” is an incoherent idea. Like “I’ve successfully squared the circle“.

This means that holding fix Premises (1) and (2), it is impossible for the necessary conditions for our being morally responsible to be fulfilled. Hence, you can’t be morally responsible (ultimately) for the conditions within you that gave rise to your actions, hence, the intermediate conclusion (3).

Notice also that the conclusion is not just that you aren’t morally responsible for any of your actions–look at how I phrased the conclusion of the Standard Argument (W07 Slide #18). Yes, Strawson’s Basic Argument does imply that you aren’t morally responsible for your actions; but that’s because it says something even stronger–you can’t, it’s an impossibility. So it’s not just that moral responsibility doesn’t exist because the truth about the world refuses to cooperate (as in the Standard Argument), but that more shockingly, moral responsibility is essentially–by its very nature–impossible!

Should we accept Premise (1)? Note that here, the word “chose” is meant merely to be the stand-in for a range of possibilities. If you are a Libertarian, you can interpret “S chose to do X” as involving an indeterministic ability to do otherwise. If you are a Compatibilist, you can interpret it is a different way, i.e., as involving a decision that was itself part of the deterministic causal history of the universe. As long as it is something that involves an actual choice, Strawson’s argument will work.

What about Premise (2)? How is it motivated? Think of a scenario where “S chose to do X”–e.g., Jerome chose to go to a party–he made a decision in light of a set of preferences, values, ideals, etc. But what if Jerome was manipulated by Imrani using her brainwave altering device (secretly implanted in Jerome’s brain) to make Jerome make that choice? If you don’t think that Jerome is morally responsible for going to the party, you need a way to explain why–after all, he did choose to do so, and he did go to the party because he so chose! (2) offers a way out for you–well, Jerome wasn’t morally responsible for the vote because he wasn’t morally responsible for the way he is behind his action! So Premise (2) seems legit.

Now for Susan Hurley’s response. As mentioned in the lecture, what I presented is much truncated and simplified from her actual writings. Let’s start again with the TL;DR version of Strawson’s argument as stated above. Hurley concedes that the argument is valid–a regress condition like Premise (2), plus an actual choice condition like Premise (1)–will invariably generate an regress that can’t be completed. The lesson she drew is a straightforward one: If you think that you can be morally responsible for your actions, then you have to find a way to neutralize the regress. But to do that properly, you can’t just go about rejecting premises willy-nilly–keep in mind that Premises (1) and (2) of are not arbitrarily laid down; they respond to considerations in their favor.

The most charitable way forward is thus to see how little we need to tweak Strawson’s argument before we break the regress. So here’s what we can think of Hurley as doing–let’s see if we can come up with an alternative version that differs as little as possible from Strawson’s starting point, and that captures the spirit of why we might find whatever acceptable in it in the first place; but one that doesn’t generate the regress. (Why do things this way? Scroll down and see the Appendix at the end of this post.)

This is where Hurley observes that a relatively small tweak to is enough to break Strawson’s argument, while preserving the spirit of the motivations behind Strawson’s original. So, instead of (1), you have (1a) and (1b):

  • (1A) You are morally responsible for an action only if you (knowingly) chose to do it. (Actual Choice)
  • (1B) You are morally responsible for the way you are behind your action only if you would have chosen to be that way, if you could do it. (Hypothetical choice)
  • (2) You are morally responsible for something only if you are morally responsible for the way you are behind it. (Strawson’s Principle)

From (1A), (1B) and (2), you can still generate an infinite series of necessary conditions for your being morally responsible for doing anything. But unlike the original argument, those conditions can be met since they don’t require you to have actually made an infinite number of actual choices–because an infinite series of hypothetical ‘choices’ actually can all be so. To appreciate this point, you need to grasp an important contrast between these two scenarios:

Scenario 1: I have considered/taken/chosen each and every even number, and divided it by 2 (An impossibility)

Scenario 2: For each and every even number—I would have been able to divide it by 2, if I could consider/take/choose it for my calculation. (Not an impossibility)

Scenario 1 sets up an impossible task. For the statement describing Scenario 1 to be true, you have to have actually divided each (i.e., every) even number in the even number series. I’ve already explained how you couldn’t even count all the numbers, let alone do an actual arithmetic operation on each of all the even numbers.

In contrast, for Scenario 2 does not set up an impossible task. For the statement describing Scenario 2 to be true, you don’t need to do anything (counting, diving)–a series of statements about the even numbers just needs to be true. Specifically, it has to be true that if you had considered the number, you would have been able to divide it by 2 without remainder. This is analogous to how the unconscious soldier in W06 doesn’t have to actually consent to have his gangrenous limb amputated (he’s unconscious, remember) for the statement “he couldn’t reasonably not consent” to be true, and so, hypothetical consent to hold.

This means that with (1A), (1B) and (2), it’s no longer impossible for you to be morally responsible for your actions, moral responsibility is not essentially impossible. This doesn’t prove that we actually are morally responsible for our actions, but it does break Strawson’s argument. Perceptive students ask, at this point, why should we prefer (1A) plus (1B) to Strawson’s original (1)? This is an important question, but now we need to go into slightly deeper waters…


  • Part 2

To begin with, let’s recall what I said earlier about coming up with an alternative version of your philosophical opponent’s argument that differs from it as little as possible and also captures the spirit of why we might find that acceptable in it in the first place (see also the Appendix below for the motivation for doing that).

To appreciate this, let’s not go straight to (1A) and (1B) but first considered a “first try” in which we replace (1) with:

(1C) You are morally responsible for something only if you would have chosen it, if you could.

Now, (1C) plus (2) will also not generate the infinite regress. But it’s not a good response because this proposal fails to do justice to the considerations motivating the original (1) and (2). If what we have are (1C) and (2), then a person who would have chosen to do something if he could, but didn’t actually chose to do it–for example, someone who really wanted to go to a party so much so that he would have chosen to go if he could, but nonetheless got to the party because he was carried there by unannounced strangers–such a person would have met the necessary conditions for moral responsibility.

In fact, we can easily think of even more outrageous, detective mystery style examples. Imagine that A really hates B and wanted him dead. A is on the balcony on the third level when she noticed that B is just below on the ground floor. She figures that if the large flower pot on the ledge were to fall on him, he would be killed. Since she really wanted him dead, she would have chosen to push the flower pot off the ledge if she could, and in fact, started to go towards the pot. But before he could do anything, the edge of the balcony’s ledge gave way–it’s always been a bit rickety and the wind just happened to blow a little bit harder–causing the pot to fall on B, killing him. If (1C) is our necessary condition for moral responsibility, then A would have satisfied it–and this seems way too easy. In contrast, the original (1) would have ruled out A’s responsibility. Consequently, opting for (1A) instead of (1C) better preserves the spirit of why we might find the original premise acceptable in the first place.

The long and short is that at the very least, any alternate proposed in place of (1) and (2) should be at least as good as the original. And by this measure, (1A), (1B) and (2) measures decently. This doesn’t mean that we should prefer (1A), (1B), (2) to the original (1), (2), but we are part way there–at least pending further objections. One student from a previous semester presented a worry.

I completely see your point. The regress argument is important because so long as you accept Premise 2, it would demand the completion of an infinite series… I see that Hurley’s hypothetical choice condition is probably the most satisfactory solution… But I am a little unsatisfied with it. For example a criminal could argue that “I didn’t chose to be born in a broken family, and I wouldn’t have chosen to be born in a broken family, if only I could have made that choice… But since I was born in a broken family, that led me to be the kind of person I am, and hence the things I did. I’m not morally responsible for the things I do.” Because we cannot claim omniscience, it is hard for us to find out objectively whether or not A would have chosen the condition within him that made him choose X, since we are not A, and even A would probably not be a reliable source, since it is in his interest to say he would not have chosen the condition in him that made him do X…

Excellent questions. First, a note about a problem that I didn’t go into in the lecture, but that the student alluded to–the indefiniteness of “I would have chosen X, if I could”, especially the part in bold. Who is the chooser? What are his motivations and beliefs? What is the position from which he is choosing? Let’s take earlier example where Jerome went to the party because Imrani’s device made him choose to go. If that chooser is “Jerome exactly as he was when he made the choice to go to the party”, then (1B) is going to be worthless as a condition on moral responsibility distinct from (1). If, on the other hand, the answer is: “Jerome as he ought to be–possessing true virtue and complete knowledge”, then conceivably, no one is ever going to be morally responsible for doing anything bad. So presumably, it has to be something in between both possibilities–but exactly how? Spelling these things out is no trivial matter. Here, I can only wave my hands wildly to say that hardworking philosophers are working hard to sort out the details, am I am somewhat confident that a suitable answer exist, though it will likely be quite complicated. (Ok, it’s in Hurley and others, but it’s too much work to get into the details now.)

The other main worry expressed, however, concerns whether we can know if all those hypotheticals are true. Here, we need to distinguish between a question about how we can know whether a person is morally responsible, and a different question about what makes the person morally responsible. (Recall a similar issue with Consequentialism.) Both Strawson and Hurley are, strictly speaking, only talking about the latter. So even if Hurley’s response against Strawson is correct, it’s entirely possible that there are many grey cases where we can’t tell (from the outside) whether the person is morally responsible. But this doesn’t mean the account is false.

More importantly, it would be unfair to think of the Hurley response as providing a complete picture of what moral responsibility requires. In one sense, it’s a little bit ad hoc–it’s main virtue is that it blocks Strawson’s regress while preserving the spirit of where Strawson was coming from. As a move in a debate to confute your opponent, that’s a powerful move, But if we are going to adopt a positive account of moral responsibility, I really hope it’s not just because it doesn’t get us into a regress! A good account will need to have more than that in its favor.

The question remains, though–can we do more than go half way? What would it take to make a positive case for (1A), (1B), (2) against the original (1), (2)? This is a much longer story than I really want to tell here (the bulk of Hurley’s paper is actually about the background methodological issues involved in resolving exactly this question). To come to grips with the question, we need a discussion of the general philosophical considerations regarding how we should choose between competing proposed conditions on moral responsibility, and what we should do if it turns out that on a given proposal, moral responsibility is impossible. This will involve us having a suitable account of what we are really, at bottom, looking for when talking about moral responsibility.

So is moral responsibility more like (putatively) objective entities and feature of the world (like gold, the reflective properties of light, or for that matter, phlogiston), or is it something more deeply embedded in our practices (like movie stars, and bankruptcy, or for that matter, witch doctors)? Relatedly, when are are asking who is moral responsible when, are we asking–what are the necessary and sufficient conditions for some fact being the case or for something to exist? Or are we really asking–what the proper norms governing when it is appropriate for us to hold people to certain expectations, or think well or badly of them, or praise or censure them? How you answer these questions will have great impact on how you think the initial question–What would it take to make a positive case for (1A), (1B), (2) against the original (1), (2)?–should be answered.

(For those who are interested far beyond what this module can offer (and are suitably hardcore), you can look for S. L. Hurley, “Is Moral Responsibility Essentially Impossible”, Philosophical Studies 99 (2000): 229-268. A version of the discussion can also be found in her book Justice, Luck, and Knowledge (Harvard, 2005). For some of the issues raised in the previous paragraph, see the chapter titled “Methodological Interlude” in R. Jay Wallace, Responsibility and the Moral Sentiments (Harvard, 1998).)


  • Appendix: On Modesty in Philosophical Debate

Imagine that you are in a philosophical debate and your opponent puts forward an argument P, Q, therefore, R. You accept that the argument is valid, which means that should you accept the premises as true as well, you would also be committed to accepting the conclusion as true. But you don’t accept the conclusion. You believe that R is false, that not-R is true. So you need to reject at least one premise. You notice that rejecting the premise P looks promising. But it’s not as if P was just arbitrarily assumed–your opponent has some respectable background considerations for putting forward P. In that case, a good way to proceed is to see if you can come up with an alternative to P (let’s call it P’) that also satisfies those same considerations, but either render the argument invalid–already enough to confute your opponent–or even better, that allows you to actually conclude not-R.

And you want to find an alternative that is as close by as possible–this is why you only target P and not worry about Q as well. Maybe you don’t even agree with Q, but there are advantages to assuming it for the sake of the argument. By presenting your alternative version of your opponent’s argument as but a small tweak, you put a lot more pressure on him–it’s as if you are saying to him “hey look, even when we assume pretty much almost everything you assume, we still don’t get to draw the conclusion you draw…” (Compare that to: “hey look, by starting from assumptions very different from yours, I get to show that your conclusion doesn’t follow”, where the obvious reply would have been–“yes, obviously… your point being?”)

Hence, what I called “Modesty in Philosophical Debate”–think of it as a version of the ancient Philosophical Daoist doctrine that the sage does everything by doing nothing (無為而無不為; Daodejing 48). Well, not literally nothing, but you get the drift–as we sometimes say, less is more. I’m not saying that this is the only way to proceed in a debate, or even that it’s the best in every case–it is not. Sometimes, you opponent’s position is so bad that you can readily argue for the falsity of several premises or directly show that his argument just isn’t valid. In those cases, you might as well skip the modesty. In other cases, your opponent begins from such different starting points that no amount of tweaking will allow the desired result. Your best outcome here while remaining modest is to find a way to show that his account entails a contradiction or a conclusion that even he doesn’t accept. But what happens if that’s not available, what if both his starting point and your very different starting point lead validly and self-consistently to your respective opposed conclusions? Well, let’s wait until W09 Knowledge and its Discontents to talk about that conundrum.