The Contours of Being Causa Sui

I finished Smilansky’s Free Will and Illusion, which I recommend to anyone interested in the denial of free will.  Smilansky was supervised by Galen Strawson—author of perhaps the best argument against the existence of free will.  Smilansky is an eloquent spokesperson for free will denial.  As he says, “we are the unfolding of the given.”  In this book, Smilansky elaborates on Strawson’s ideas and reaches a more ambivalent conclusion. 

I don’t agree with this conclusion. Smilansky’s mistake, I think, comes early in Free Will and Illusion, when he rejects consequentialism.  Like Smilansky, Pereboom also takes issue with a prominent form of consequentialism, utilitianism, in his Living Without Free Will.  I want to challenge both Smilansky and Pereboom on this issue in (work in progress) article “Who’s Afraid of Creeping Exculpation?”

Smilansky distinguishes between two varieties of compatibilism, Effects Compatibilism and Control Compatibilism (this can be confusing because Tamler Sommers, another skeptic about free will, has referred to Consequentialist Compatibilism instead Effects Compatibilism).  On Smilansky’s view, Effects Compatibilism doesn’t do justice to the Core Conception, or the fundamental idea that responsibility requires up-to-us-ness.  But a consequentialist can reply: we are not ignoring the Core Conception; it is obsolete.  According to Strawson, in the long run, nothing is up to us.

Nevertheless, Smilansky raises a good question: does consequentialism precede or follow free will denial?  Consider this dilemma.  Many (including myself) have the intuition that being causa sui is more conducive to retributivism.  But why?  Suppose that punishing a causa sui agent does not maximize the consequentialist goals of one’s choice.  There is no obvious reason why one should practice retribution in this situation.  Consequentialism seems to precede free will denial.

The dilemma is even more problematic than that.  To the extent that punishment forms an agent’s character, then one might find it impossible to punish a causa sui agent.  Causa sui agents invent themselves from scratch.  An agent who incorporates punishment into its self-formation is not pure causa sui. 

But there are several problems with this objection. For one, punishment might not form an agent’s character so much as it forms an agent’s behavior.  Punishment does not seem to make bad people good so much as it makes bad people stop.  Secondly, perhaps, according to our natural conception, people are not pure causa sui.  Instead, perhaps we regard people as incorporating nothing foreign into the self except punishment and rewards.  Punishment and reward do seem to be essential features of our concept of personhood. This raises the question of to what extent an agent might be partially causa sui.  Can one entertain the thought, for example, of being causa sui with respect to friendliness and intelligence, but being “the unfolding of the given” with respect to taste in music or members of the opposite sex?

Just as we might divide one’s personhood into causa sui and “unfolding of the given” portions, so might we divide one’s lifespan into portions. A causa sui agent is (by definition?) causa sui at the moment of its creation. But can such an agent cease to be causa sui? This seems to be the case, for example, when even those people who believe humans are causa sui nevertheless insist that people should be punished and rewarded. So long as the agents continue to be causa sui, punishment and reward can have no effect upon their characters (but perhaps they punish and reward for different reasons). Would these ex causa sui agents nevertheless be responsible for their actions? Or alternatively, once causa sui, always causa sui?

Gardeners:

1. Is being causa sui more conducive to retributivism? Alternatively, is free will denial more      conducive to consequentialism? If so, why?
2. Does meta-ethics precede or follow the free will question?
3. Is it conceptually possible for an agent’s personhood or lifespan to be divided into causa sui and “unfolding of the given” portions? If so, how do these complexities affect an agent’s moral responsibility (or lack there of)?

Comments

[Mark Smeltzer]

How many times are you going to ask the same set of questions?

Granted, the context of the questions in this post is a *bit* different since it involves a cursory discussion of Smilansky, but it quickly degenerates into the same old "Why not consequentialism?" drivel that has been thoroughly trounced time and time again on this blog.

So, I suppose what I am wondering amounts to this: do you have any arguments to present or is it your intention to keep spamming the blog with rhetoric?

I, for one, am tired of the rhetoric...

If you are up to the challenge, please present the best argument you have for your position in a concise format and we'll be sure to give it a run for its money. I'm sure that would be a genuinely educational experience and a better use of our collective time and energy.

[Tamler Sommers]

Mark, why so bitter? You should become a free will/moral responsibility skeptic. It would do you a world of good.

[tnadelhoffer]

Mark,
I bet your tone would have been different had Pereboom, Smilansky, or Strawson posted Kip's thread rather than Kip. In any event, a word to the wise: if you are tired of someone's rhetoric, don't comment on their post. Better yet, don't read it all. If, on the other hand, you are going to participate, you ought to at least be civil. Or at least that's how I see it.

[Mark Smeltzer]

I'm not trying to be uncivil. Not am I bitter.

I'm simply commenting on the fact that a number of us have already interacted with Kip on nearly identical points in previous threads, and yet those interactions have not been fruitful due to the fact that Kip has not taken the risk of actually putting an argument out there.

Asking for an argument seems the *most* civil thing to do.

[Kip Werking]

Thanks, Mark. For a moment there I was afraid that nobody was going to comment on my post!

In response to your objections, I have several points to make:

1. Contrary to what you write, I think there were several original things in what I posted. For example, I suggested: that a causa sui agent cannot be punished, that an agent can be partially causa sui, and that a causa sui agent could cease to be one. I also asked whether meta-ethical questions should be answered before or after the free will question. These are, to my knowledge, entirely new observations. Indeed, I published that post *because* it contained these new, and hopefully interesting ideas. I would not have posted it otherwise.
2. I also challenged the intuition that consequentialism is more conducive to free will denial, and retributivism is more conduvice to orthodox views on free will (which assert that people are causa sui). This point might not be entirely original; Tom Clark from the Center for Naturalism consistently makes it. But the point is not often discussed in the literature. The two writings, to my knowledge, on a meta-ethical approach to the free will problem are Smilansky's discussion of Control Compatibilism versus Effects Compatibilism, and Pereboom's discussion of utilitarianism in Living Without Free Will. One might also include Michael Slote's article on the compatibility of free will denial and virtue ethics. Interestingly, all of these writings consider or endorse a more skeptical view on free will (retributivists do not seem to advertise their retributivism), and yet none of them embraces consequentialism. Smilansky explicitly rejects it, Pereboom rejects utilitarianism (but adopts a pro-quarantine/anti-punishment position which may be ultimately motivated by consequentialism, or not), and Slote argues for the compatibility of free will denial and virtue ethics. So, in this context, the point I made does seem somewhat new.
3. Even if the point is not new, it is of crucial importance! An issue as heavy as the relationship between meta-ethics and the free will question should not die after it is first raised.
4. Finally, you say that I have not given an argument for my position. But I did include an argument:

"Many (including myself) have the intuition that being causa sui is more conducive to retributivism. But why? Suppose that punishing a causa sui agent does not maximize the consequentialist goals of one’s choice. There is no obvious reason why one should practice retribution in this situation. Consequentialism seems to precede free will denial."

Note that this is the same argument that Tom Clark always gives against retributivism: he insists that there is simply no reason for it. I understand that this argument might not impress you much. The thought experiment can only work if one presupposes consequentialist intuitions. Without such intuitions, the thought experiment does nothing. In this respect, it is somewhat circular. But the thought experiment is the best sort of an argument an ethicist can give. Ultimately meta-ethics is a question of intuition and consensus. Nothing prevents a person from asserting that a certain meta-ethical position is right. And there is nothing to disprove that proposition except our intuition, and large consensus, to the contrary. Meta-ethics is not math.

[Mark Smeltzer]

Ever heard of Richard Double? You may want to check out his books and papers. Your view seems to have little distinction from his.

Anyway, it seems obvious that questions of applied meta-ethical theories are of a different sort than questions of applied agency theories. So, I don’t see any big controversy there. The ordering of the questions becomes important if the concerns are normative and practical. If the concerns are merely theoretical, then the ordering bears no significance.

For example, Fischer's account is simply an account of the mechanisms of responsibility. Fischer's account does not include a normative component! Without a presupposed meta-ethical account, nothing normative follows from the conjunction of the statements "A was responsible for X" and "X is morally reprehensible".

How we ought to act in virtue of (and with respect to) other agent's is a question of meta-ethics. If one's meta-ethic includes criteria for treating good people certain ways and evil people other ways or imports a responsibility-laden notion of justice, then the question of responsibility becomes important.

As far as I can tell, your central claim is that if consequentialism is true, then questions of responsibility are irrelevant. You may also have a secondary claim that responsibility is impossible. Furthermore, you may also be asserting the consequentialism is true. However, the successes of these latter claims obviously do not impact the success of the first claim, so I'm going to ignore them here.

That said, I think your central claim is patently false, and I can show you why.

Suppose a certain consequentialist adopts this maxim: do that which will (1) maximize the distribution of pleasure to most number of people and (2) minimize the distribution of injustice to the most number of people. He adopts that maxim because he recognizes both pleasure and justice as goods.

Granted, this maxim has the same flaws that all consequential maxims have, but aside from that, it does seem be sufficient to invalidate your central claim because the second component calls into question the freedom and responsibility of agents at large.

Thus, at best, your central claim could be restated as, "that if X-consequentialism (to which questions of responsibility are irrelevant) is true then questions of responsibility are irrelevant", but that is an obvious tautology.

So, while it may be the case that responsibility is irrelevant to certain flavors of consequentialism, it is a trivially true claim – no one could, or would, argue against it!

Should you decide to present a defense one of those flavors of consequentialism, that would be worthy of further discussion. But, I’m not sure if debating meta-ethical systems is relevant to this blog’s purpose.

[Daniel Cohen]

My feeling is that ethics is completely orthogonal to the free will/responsibility question. The mere fact that philosophers who hold one kind of view (in ethics) tend also to hold another kind of view (on free will) is just philosophical sociology (no deeper than the observation that Australians tend to be realists, while the English tend to be antirealists). But the idea that ethics can settle questions of responsibility is definitely out there.

The most influential point of intersection that comes to mind is, of course, the utilitarian approach to responsibility (eg. Schlick). But, since Strawson, this approach has been criticised precisely for changing the subject. I.e. whether or not it is morally appropriate to punish people doesn’t settle the question whether or not they are responsible for doing wrong.

More recently, however, Jay Wallace (1996) has defended a “normative interpretation” of the conditions of responsibility. He argues that to be a responsible agent just is to be an agent whom it would be *fair* to hold responsible. In this way, it would appear that the truth or falsity of utilitarianism, for instance, would settle the question of responsibility, on Wallace’s view. Wallace, however, argues that it would be a mistake to derive the relevant principles of fairness from such higher-order moral principles, because doing so would seem too quickly to resolve the debate. The utilitarian approach, in particular, would seem thus to fail to do justice to our incompatibilist intuitions. Any plausible account of responsibility, however, should be able to explain why we are least *tempted* to deny free will.

Rather, Wallace argues, the principles of fairness relevant to settling the debate must be located by employing a method of reflective equilibrium. That is, we must derive lower-order moral principles that justify the intuitive application-conditions of responsibility – saying when people should be exempted or excused from responsibility. Wallace derives a condition saying that people should only be held responsible when they possess powers of reflective self-control.

So, even on his normative approach, it appears that Wallace still returns to a very familiar issue – whether people in fact possess powers of control. But it seems then that the normative approach doesn’t significantly alter the debate. So we might as well interpret the question of responsibility in straightforwardly *metaphysical* terms.

[Kip Werking]

Mark,

Of course I’ve heard of Richard Double. I’ve read his The Non-Reality of Free Will but not yet Metaphilosophy and Free Will. Professor Double was kind enough critique my early article on Kane’s view. He also sent me a pre-publication copy of The Ethical Advantages of Free Will Subjectivism. Although we’ve lost touch over the years, and I worry that I’ve offended him, I have a great amount of respect for Professor Double’s work.

I’m glad you mentioned Professor Double’s work, though. For one, Double’s The Non-Reality of Free Will presents *another* exploration of the relationship between meta-ethics and the free will problem. In particular, Double argues that free will denial follows from moral nonrealism. His position is more complicated and nuanced than that, however, and I hope I have not misrepresented his view.

I’m also glad you mentioned Double because, although you suggest that our views are similar, sufficient differences remained for there to be some tension between his position and mine. For example, Double seems more sympathetic to compatibilism, and less sympathetic to consequentialism, than I am. I also tend to think that the statement “people don’t have free will” is meaningful (because people are not causa-sui), regardless of whether Double’s moral nonrealism is correct (see Strawson’s The Bounds of Freedom in Kane’s Handbook for a discussion of this distinction). Nevertheless, like Double (as I understand him), I am largely skeptical of the existence of free will and moral facts.

Moving on, you say that “As far as I can tell, your central claim is that if consequentialism is true, then questions of responsibility are irrelevant.” First, I am not sure I was making any central claim. Instead, my post was intended to raise some novel issues and ask interesting questions (such as whether an agent might be partially causa sui).

Also, I am not sure that the central claim you suggest captures what I was trying to say. This is probably my fault. Tamler, in a different thread, characterized my claim as “MR skepticism entails consequentialism,” which is definitely not what I was trying to say. Instead, I would characterize my central claim as follows. Consider the related claim R that “people’s being causa sui would be more conducive to retributivism.” It seems (at least to me) that people are naturally inclined towards asserting R. My claim is that R is false.

That said, I would defend the (quite different) claim you suggest (consequentialism->irrelevance of responsibility), in the right context. But one must ask: are question of responsibility irrelevant *to what*? A consequentialist, as I understand it, does not hold that questions of responsibility are never relevant, just that they are not ultimately relevant. Let me cash that out. To say that responsibility is irrelevant to a consequentialist is to say that, if doing so would not maximize the consequentialist’s goals, then the consequentialist should punish the innocent and let the guilty go free. It is not to say, however, that when evaluating how possible courses of action will maximize one’s goals, a consequentialist cannot consider whether one is responsible (on a weak understanding of responsibility) for a given action. On the contrary, such evidence of responsibility is almost always very relevant. Situations where it might be better to punish the innocent (kill one person to save ten people) and let the guilty go free (perhaps Nazis on deserted islands) are rare. Typically, one maximizes relatively non-controversial goals (like happiness) by punishing, and therefore deterring, bad behavior.

Your argument against my central claim is an argument against the interpretation according to which responsibility can never be relevant to a consequentialist’s choice of how to act. As I have shown above, if this is your interpretation of that claim, then I do not defend it, and I agree it is false.

You also write:

“Should you decide to present a defense one of those flavors of consequentialism, that would be worthy of further discussion. But, I’m not sure if debating meta-ethical systems is relevant to this blog’s purpose.”

Seriously? The blog says “devoted to Agency Theory, including related issues in Philosophy of Action, Moral Psychology, Metaphysics and *Ethics*” right at the top.

[Tamler Sommers]

Kip,

For the record, my post did not characterize your claim as “MR skepticism entails consequentialism.” I only noted that if Rawls is an MR skeptic (and his claims are not inconsistent), then it's possible to be an MR skeptic but not a consequentialist. This obviously relates to the theme of your post, but I didn't mean to imply anything about your position.

[Kip Werking]

My apologies, I misunderstood. :)

[Mark Smeltzer]

Kip,

"It seems (at least to me) that people are naturally inclined towards asserting R. My claim is that R is false."

I'm on board with that -- as would be any other compatibilist.

"Your argument against my central claim is an argument against the interpretation according to which responsibility can never be relevant to a consequentialist’s choice of how to act. As I have shown above, if this is your interpretation of that claim, then I do not defend it, and I agree it is false."

Okay, but what is your response to the MR-sensitive consequentialism that I described in my previous post?

This consequentialist has two high order goals: to maximize pleasure and to minimize injustice. He is going to act in ways that are jointly conducive to those goals. So, to him, responsibility is ultimately relevant.

That's why I said this case is a defeater to the broad application of your claim -- because the only sense in which your claim will be true is when it is stated as a tautology.

[Kip Werking]

Mark,

Your post made me pause and think. The definitions of consequentialism and deontology (are these two the negations of each other?) seem fuzzy to me. Much writing about consequentialism seems motivated by the sort of counter-examples I described (such as letting the guilty go unpunished and punishing the innocent). Obviously, these examples describe a different sort of consequentialism than the one you consider. In particular, they prioritize utility or happiness over justice.

If the consequentialist’s goal is to maximize justice, however, then that consequentialism seems to be just retributivism (putting aside, for the moment, the complication that your consequentialist also seeks to maximize pleasure). So, on this view, consequentialism and retributivism aren’t opposed; rather, retributivism is just a species of consequentialism.

Now, the original claim was: consequentialism -> blind to blame. You suggest that this is not always true. Instead, call consequentialism which doesn’t seek to maximize ultimate justice nj-consequentialism. Then nj-consequentialism -> blind to blame. You claim this is a tautology. That might be imprecise, but I think that captures the general point.

I think you’re right about this. In response, I can only make a few points:

1. It seems that the term consequentialism is usually associated with nj-consequentialism (one that doesn’t imply retributivism). But technically, I suppose consequentialism refers to both nj-consequentialism and j-consequentialism. I’m not sure though?

2. Not all tautologies are uninteresting. Indeed, many tautologies are not obvious. For example, the claim that people are not causa sui is a tautology. But many people find it both interesting and counter-intuitive.

3. The claim was not offered as an amazing discovery. It was not even offered as a major claim in my post (these would be the denial of R and other claims, such as whether an agent can be partially causa sui). Instead, the notion “(nj-)consequentialism -> blind to blame” was noted because then nj-consequentialism provides an ethical framework for the post-free-will world.

Just as a consequentialist might seek to maximize justice, a deontologist might endorse exceptionless rules about utility or happiness (ones which I might prefer to some varieties of consequentialism, such as j-consequentialism). So, perhaps I should have made a substantive claim about values instead claimed that consequentialism, in general, is preferable to retributivism (although, as I showed above, these may not be mutually exclusive). Instead of “consequentialism -> blind to blame”, or the (tautological?) claim “nj-consequentialism -> blind to blame”, I might make the following substantive claim. A consequentialist’s goal should not be justice for its own sake (where that means always punishing the guilty and letting the innocent go free), but rather whatever the consequentialist ultimately values (such as personal or global utility), and justice should just be regarded as a means towards these ends. Indeed, from this perspective, it might be unjust to punish the guilty, or let the innocent go free, if doing so minimized the consequentialist’s goals!

[Mark Smeltzer]

Kip,

The distinction between consequentialist and deontological meta-ethical systems has to do with whether norms are formulated a posteriori or a priori (respectively).

In rule-consequentialism prescriptive rules are set out with the intend of maximizing the distribution of certain goods. These rules are subject to revision over time. (c.f. http://plato.stanford.edu/entries/consequentialism-rule/ for an overview)

In act-consequentialism there are no prescribed rules, only the prescribed notion that one should *always* act in ways that maximize the distribution of certain goods, and this entails that the course of action that should be taken will be relative to the context of specific situations. (c.f. http://plato.stanford.edu/entries/consequentialism for an overview)

Consequentialistic systems are meta-ethical because they do not prescribe any certain ends. Just the ways in which those ends ought to be met, and how to define norms for action. Only an ethical system, such as Mill's utilitarianism, can prescribe specific ends.

The system I describe is a consequentialist system that could be cached out as either a act or rule based system (it cannot be a deontolical system because the maxim contains the idea of distribution, which can only be applied a posteriori). Futhermore, it is an ethical system because it prescribes actual ends.

What I'm trying to get you to do is to set forth an ethical system! (If there are no ends that we ought to value, then the topic of meta-ethics is irrelevant.) Until you do, you don't have grounds to say anything about which ends are valid and which are not.

[Paul Torek]

Daniel,

You mention Jay Wallace's “normative interpretation” of the conditions of responsibility: that to be a responsible agent just is to be an agent whom it would be *fair* to hold responsible. To which I say, that's obvious - and an excellent starting point for productive discussion.

Reflective equilibrium may well return us to pondering the very familiar issue of whether people in fact possess powers of control. But this doesn't mean that ethics loses its punch. It's reflective equilibrium, not reflective metaphysics-has-its-way-with-ethics, doncha know.

[Mark Smeltzer]

That statement may be more contentious than you realize: the clause "to be" signifies a metaphysical reality, but the later clause "it would be fair" signifies an implied reality, and these two contexts do not obviously match up. Moreover, one could interpret "fairness" in terms of the notion of ultimacy, which could stall the discussion.

A statement that might make a better starting point for productive discussion is: to be treated as a responsible agent just is to be an agent whom it would be warranted to hold responsible.

Some hard incompatibilists would argue that we aren't warranted in holding anyone responsible since no one *is* responsible, but others will establish alternative sources of warrant. For example, they could be utilitarians and believe that the world would be a better place if people are held responsible (like Smilansky).

[Kip Werking]

Mark:

Again, I’m well aware of what rule and act-consequentialism are.

You write:

“What I'm trying to get you to do is to set forth an ethical system! (If there are no ends that we ought to value, then the topic of meta-ethics is irrelevant.) Until you do, you don't have grounds to say anything about which ends are valid and which are not.”

I think you’ll find the beginning of my answer to your new question in my last comment. There I wrote:

“Just as a consequentialist might seek to maximize justice, a deontologist might endorse exceptionless rules about utility or happiness (ones which I might prefer to some varieties of consequentialism, such as j-consequentialism). So, perhaps I should have made a substantive claim about values instead claimed that consequentialism, in general, is preferable to retributivism (although, as I showed above, these may not be mutually exclusive). Instead of “consequentialism -> blind to blame”, or the (tautological?) claim “nj-consequentialism -> blind to blame”, I might make the following substantive claim. A consequentialist’s goal should not be justice for its own sake (where that means always punishing the guilty and letting the innocent go free), but rather whatever the consequentialist ultimately values (such as personal or global utility), and justice should just be regarded as a means towards these ends. Indeed, from this perspective, it might be unjust to punish the guilty, or let the innocent go free, if doing so minimized the consequentialist’s goals!”

Here I show sympathy for ends such as “utility or happiness” yet I also hesitate to assert any specific end for my own preferred ethical system (“whatever the consequentialist ultimately values”). It is also obvious, from my previous writings, that I agree with the notion that one should punish the innocent and let the guilty go free *if* doing so maximizes one’s goals. This brings me to my point: whatever my goals, as a consequentialist, are, they are not the justice-related sort of goals you cite as a counterexample to “the general application” of the principle that consequentialism is blind to blame. Instead, they would be non-justice related goals such that one can (hypothetically) imagine situations in which punishing the innocent or letting the guilty go free maximizes them.

Beyond that negative answer, I can only say this. Imagine an “objectivity/realism” dial or knob (Dennett talks about these) which marks how much objectivity or realism we are going to regard our moral norms as having. By this, I mean, the more objectivity/realism a norm has, then the more that norm is factually correct not *just* because people think so, but independent of anybody’s thoughts. So a maximally real norm would be one that is true even if there are no minds in the universe to think it is true. Ultimately, and more rigorously (as I think I should be in philosophical discussions, as on this message board), I think the knob should be towards the “objectivity/realism” side; furthermore, as one turns the knob in this direction, the more moral norm disappear and moral nonrealism obtains. As Hamlet said “"there is nothing either good or bad, but thinking makes it so" (2.2.249-250). In other words, the more of an ontological burden we place upon moral facts, the less likely they are to exist. I tend to place more of a burden on them, just as I place more of a burden on the term “free will” (which, on my view, entails being causa sui), because I think that ordinary people’s unreflective intuitions *do* burden these terms in this way (and that some compatibilists, for example, use revisionist definitions of “free will”).

However, as one turns the dial in the other direction, I recognize that certain things exist which one might call moral facts, in the sense that there is a large consensus amongst human beings that X is right/wrong. The larger the consensus, the more I would be inclined to consider them moral facts in this weaker (Humean) sense. This is the context in which I live my ordinary life. In this context, my ethical system would roughly be to promote my happiness and the happiness of those I love.

A good question is: as a moral nonrealist, how can I endorse nj-consequentialism over more justice-related ethical systems? I’m not sure of the answer to that question. All I can say is that, if anything is valuable for its own sake, ensuring that the guilty are punished, and that the innocent are not punished, is *not* it. Instead, that has all the appearance of being a means to an end (such as a happier society, or deterring future harm). Indeed, I think our inclinations towards just punishment evolved for precisely those teleological reasons, and not for their own sake. If all just punishment did was make the wrongdoer suffer, without deterring future harm, it would not have evolved.

[Mark Smeltzer]

Kip,

The references and summaries were provided in response to this statement from your post on the 9th: "The definitions of consequentialism and deontology (are these two the negations of each other?) seem fuzzy to me."

My intent was simply to provide relevant, operating definitions so we could keep the discussion moving. As well as for the benefit of hobbyists who frequent the blog who might be interested in checking out the references.

Having said all that, do you concede the point that responsibility and retributive desert can be ultimately relevant to consequentialists (even if they believe these notions to ultimately false, such as Smilansky does)?

Rule-consequentialists, especially, might adopt hard rules such as never-punishing-the-innocent for consequential reasons.

[Kip Werking]

Mark,

Oh, my comment "The definitions of consequentialism and deontology (are these two the negations of each other?) seem fuzzy to me," was meant to indicate my confusions about the differences between the two. For example, people often regard consequentialism and deontology as opposed, but as others have shown (and I've tried to show above), deontological systems can be a species of consequentialism. Again, I'm not very knowledgeable in this area. And it is still unclear to me what the precise criteria for a deontological system is (unfortunately to say "The distinction between consequentialist and deontological meta-ethical systems has to do with whether norms are formulated a posteriori or a priori (respectively)" doesn't help me much). But I do understand the general difference between act and rule consequentialism.

To answer your question:

"Having said all that, do you concede the point that responsibility and retributive desert can be ultimately relevant to consequentialists (even if they believe these notions to ultimately false, such as Smilansky does)?"

Yes, I happily concede that point, and I thank you for making me think through the complications. As I've already said here, I was hastily using the term "consequentialism" in reference to just "nj-consequentialism." J-consequentialisms, such as the one you describe, are not typically associated with consequentialism (I think?), but technically belong under that heading. Once this is understood, one sees that consequentialism does not necessarily entail "blindness to blame."

[Mark Smeltzer]

Deontology does not mean "rules" -- rule-consequentialism is not a deontological system. In a deontological system, the ethical norms are knowable a priori. That is the distinguishing factor -- on deontological systems it is impossible to have experiences cause one to refine the norms because the norms are grounded a priori (without respect to actual consequences and/or experiences), but that is not true of rule-consequentialism: here norms are formulated a posteriori using quantified results and are subject to potential revision based on new data and/or significant changes in the relevant variables.

Unfortunately Stanford's site does not have a primer article on deontology, or I'd provide a link.

[Mark Smeltzer]

Sorry if that last post sounded a bit condescending. I'm very tired and didn't put much effort into the quality of the presentation and I appologize for that.

Regardless, I hope that explanation will be helpful.

I'm glad you raised the question because these distinctions are important to the discussion :)

As I said earlier, until you commit to an ethical system and defend it, statements such as "All I can say is that, if anything is valuable for its own sake, ensuring that the guilty are punished, and that the innocent are not punished, is *not* it" (from your prior post) have no dialetical value.

If you're going to stick to a firm ethical relativism or ethical egoism, you lack grounds to assert generalized norms such as the one quoted here because others operating within the same meta-ethical framework may value things such as retributivism given that neither of those positions allow for "higher order" values that could migitate which ethical norms are universally valid and which are not.

[Daniel Cohen]

Mark,

I’m a bit confused about your definition of deontology. I thought deontological theories were distinguished, simply, by analysing rightness in terms of the *duties* one must obey. The view that the relevant duties must be derived a priori sounds, rather, like a distinctively Kantian variant of deontology.

Anyway, the relevant issue for this debate appears to concern the distinction between consequentialism and *non-consequentialism*. Deontology, after all, is only one species of non-consequentialism. It is arguable, however, that rule-consequentialism is, indeed, a non-consequentialist theory, depending on how one defines consequentialism. If, for instance, consequentialism is the view that the rightness of an action is a function of the consequences of that action, then ‘rule-consequentialism’ will turn out to be a non-consequentialist theory. (For example, rule-consequentialists might argue that one ought to keep one’s promise because better consequences would be produced if everyone were to keep their promises than otherwise. In this case, it is clear that the rightness of keeping one’s promise is *not* a function of the consequences of one’s keeping one’s promise. Rather, it is a function of the consequences of *everyone* keeping their promises.)

[Mark Smeltzer]

Daniel,

I disagree that the distinction between rule-consequentialism and deontology is superfluous to the discussion because it seems that there is a temptation for some to conflate the two.

I submit your own example norm as evidence for this claim, which was that "one ought to keep one’s promise" grounded by the fact that "better consequences would be produced if everyone were to keep their promises than otherwise".

That's not an example of rule-consequentialism because the example norm is grounded a priori. Thus, the basis for that norm is deontological in nature -- to see this, simply ask yourself what kind of empirical data could invalidate the grounds for this norm (as stated); the answer is none.

A rule-consequentilist could still assert that particular norm, but it would be done on the basis of empirical research -- upon the basis of experience. That is why I said earlier that the primary distinction is that consequentilist norms are necessarily grounded a posteriori.

So, if you want to assert a distinction between rule-consequentilism and act-consequentialism, I would of course agree. But, rule-consequentialism is essentially a consequentialist theory because it is aimed at maximizing the distribution of certain ends.

[Paul Torek]

Mark,

I see your point about the epistemic dimension of "would be *fair* to hold responsible". But one can abstract away from that epistemic dimension, and IMHO often should, in free will and moral responsibility debates. Sometimes it's good to confine our attention to cases where those who do the holding-responsible know, and take themselves to know, all the morally relevant facts about the agent's motivation, etc.