The Scorpion and The Frog

I’m surprised that, after studying free will for several years, I’ve never heard philosophers mention the story of The Scorpion and The Frog.

The story is simple.  A scorpion asks a frog to carry him across the river.  The frog is afraid of being stung.  But the scorpion reassures the frog that, if stung, the frog would drown, and therefore so would the scorpion.  The frog agrees.  Halfway through, the scorpion stings the frog.  They will both drown.  The frog asks the scorpion, “why?”  The scorpion says, “I’m a scorpion; it’s my nature.”

The Scorpion and the Frog illustrate a fundamental distinction: between defects in a person’s rationality (r-defects) and defects in a person’s desires and values (v-defects).  The scorpion has a v-defect, but not necessarily an r-defect.  The v-defect is clear: he wants to sting, and stings, innocents.  Is it the scorpion’s fault that he has the character he does?  No.  And then, in accordance with TNR-like principles, isn’t the scorpion excused?

To my eyes, the story of the Scorpion and the Frog exposes a blind spot in compatibilism.  Quite simply: compatibilists seem to never consider v-defects to be excusing conditions.  All of the standard accounts (Fischer’s, Dennett’s, etc.) seem blind to v-defects.  Fischer considers whether a person is reasons-responsive—regardless of what the reasons are.  Frankfurt considers whether a person acts in accordance with his highest desire—regardless of what that desire is.  Well, what if, through no fault of his own, that person’s highest desire is to torture and murder other people?

The scorpion does not necessarily have an r-defect.  Indeed, the scorpion can satisfy the compatibilist conditions of your choice.  He can be reasons-responsive and own his reasons-responsive mechanism.  He might act in accordance with his highest desire—to sting others (even if he drowns).  He might, in other words, be perfectly rational.

For example, the scorpion doesn’t say “well, it’s my character to be friendly and peaceful, but then this random, inexplicable urge to sting you gripped me.”  That r-defect would excuse the scorpion even in the eyes of compatibilists.  But he doesn’t say that.  He says, “I’m a scorpion; it’s my nature.”  Yet he still seems excused.

Instead of just considering defects that could impair a person’s rationality, we should also consider defects in a person’s desires and values.  They can be just as hurtful.  And they should be just as exculpating.

Comments

[Mark Smeltzer]

Not sure why you think these cases present a problem for compatibilists. An agent with a murdeous, torturing nature (the scorpion too), would surely deserve to be an object of scorn, especially on Frankfurt's account.

[Neil]

It's not true that Fischer ignores V-defects; an agent must be responsive to moral reasons, and in some kind of patterned way, to count as MR. I'm sure that's not going to satisfy you. But Wolf's explicitly compatibilist account might: she would hold the scorpion blameless bc morally insane.

[Michael Drake]

"An agent with a murdeous, torturing nature (the scorpion too), would surely deserve to be an object of scorn"

Why suppose that something must have freewill to be a proper object of scorn?

[Kip]

Mark,

The problem is not that they would be objects of scorn, the problem is that they would be held morally responsible. Yet, as the Scorpion and the Frog story suggests, it is difficult to hold someone responsible for just acting in accordance with their nature. That is, the facts that:

1. they didn't choose their character
2. they didn't choose that, given their character, they would act in a certain way

Seems to excuse their acting that way.

Neil,

I remember now that Wolf makes an explicit distinction between those who act rightly and wrongly.

I don't know about Fischer's view. I'm not sure that "it's morally right to torture and kill innocent people" is any less of a moral reason than "it's morally wrong to torture and kill innocent people."

Perhaps you mean: it's not a moral reason because it's false. But then we are getting into moral realism, and the content of morality. Do compatibilists really want to hinge their view on moral realism, and a particular view of what is moral? It always struck me that:

A. whether a person has free will?
B. whether a person exercises their free will to do good or evil?

Are two entirely separate questions.

Part of what seems to motivate skeptics like myself, Richard Double and Joshua Greene is that we believe moral realism is false.

[Kip]

Neil,

Here's another way of thinking about Fischer's view:

It seems to me that Fischer couldn't be limiting MR agents to agents without both r-defects and v-defects, because Fischer also believes that most people are, most of the time, MR. Yet, people act wrongly all of the time, and each of these acts seems to be the result of either an r-defect or v-defect.

Suppose that Joe robs a liquor store. Joe might have an r-defect: he actually believes it's wrong and undesirable to rob liquor stores. Still, his rationality is impaired, so that an urge to rob the store overpowers his cool rationality, and he robs the store anyway.

Or consider the alternative: Joe is perfectly rational. But, unfortunately, Joe has a v-defect: he believes it's good and desirable to rob liquor stores. So, he rationally acts in accordance with that v-defect, and robs the liquor store.

If I understand you correctly, Fischer's view would excuse Joe in either case. But can't we extrapolate from Joe's robbery to all wrongdoing? Can't we do the same for Hitler, or Osama bin Laden, or the guy who didn't tip a good waitress? It seems that every act of wrongdoing, major or minor, must be the result of an r-defect or v-defect. But then all wrongdoing would be excused. I'm pretty sure Fischer doesn't believe that.

[Badda Being]

Your problem is that compatibilists fail to show total fidelity to the consequences of their theory of moral responsibility, since otherwise they would excuse everyone for everything: they would find no one morally responsible for anything. This is a genuine problem if we assume a version of compatibilism which holds free will to be an essential component of moral responsibility -- a fair assumption, to be sure, since otherwise we are tasked to understand why compatibilists are motivated to reconcile free will and determinism in the first place. On the other hand, we might understand compatibilism as a reconciliation not between free will and determinism but between moral responsibility and determinism. Compatibilism would then be the view that people can be held morally responsible regardless of free will, adducing the brute fact that people are held morally responsible even while the question of free will is still disputed.

[Brian Parks]

Great post Kip, as usual. What follows is long, but bear with me ;-)

“Instead of just considering defects that could impair a person’s rationality, we should also consider defects in a person’s desires and values. They can be just as hurtful. And they should be just as exculpating.”

I have to admit that on a primal level I feel the force of the compatibilist distinction between r-defects and v-defects. I would be much more inclined, for example, to want to berate and punish the scorpion for having a crappy-but-fully-rational character than for being mentally insane or otherwise incapable of rational function. I would imagine that most people share that inclination.

I think it would be highly beneficial at this juncture for both sides—compatibilist and anti-FW—to step back from their respective positions and explore the question of *why* we human beings might be inclined to consider v-defects exculpatory, but not r-defects.

(Now, I know that the traditionalists in this debate will want to ignore the point and jump right back into the futile question “But is the scorpion responsible or not?” I’m telling you, if we put that question aside for just a second, and take a careful, evolution-informed look at our own psychological tendencies, things will start to make *perfect* sense.)

Assume, for a moment, that our reactive attitudes—our tendency to blame, praise, punish, reward, and so on—have evolved out of millions of years of natural selection. We might ask, what adaptive functions do they fulfill? Clearly, the answer is that they fulfill behavior-shaping functions. They influence, maintain and alter future behaviors.

Our beating the piss out of the scorpion in retaliation for killing the frog influences his future decisional calculus and decreases the statistical probability that he will kill others—at least as long as we’re around to threaten him with future beatings.

In the words of Skinner, “the consequences of an act affect the probability of its occurring again.”

Now, a reactive attitude will not fulfill any behavior-shaping functions if the behavior in question is not rationally initiated. If I am incapable of rationally acting on the basis of my desires and beliefs, then *changing* my desires and beliefs will not accomplish anything. It represents a waste of time.

Likewise, if my behavior is involuntary, then punishing me for it will not change anything. If I have tourrette’s syndrome, and I keep twitching, and you beat the piss out of me every time I twitch, I’m still going to keep twitching. The cause of my twitching is not associated with my psychological traits, so your punitive attempts to alter those traits will not accomplish anything. They aren’t the root cause of the problem.

Now, generally speaking, it is maladaptive for organisms to waste their psychological and physical resources on futilities. Thus, to the extent that it really is a waste of time to punish someone for involuntary, irrational, insane behavior, we should expect human beings to be inclined against such punishment, or at least more inclined against such punishment than they would be in cases where the behavior is fully rational and emerges as a direct consequence of a morally deficient set of desires and beliefs.

In this way, we can make perfect sense of Eddy’s prior distinction between Bernie Madoff and a severely retarded, mentally insane person. We want to punish Madoff much more than we want to punish the retarded, insane person because Madoff is perfectly rational. The problem is *exactly* in his personality, his deficient set of psychological sensitivities, the internal psychological calculus that led him to steal from his clients. That calculus can be very effectively changed with pain and suffering. Hence our strong reactive desire to inflict pain and suffering.

In the case of the retarded, insane person, the problem is in a brain that just doesn’t function properly. Punishment essentially does nothing. And so the desire to punish isn’t nearly as strong.

If what is being said here is correct, then we should expect our reactive attitudes to *track* the rational closeness of a person’s behavior to the person’s core set of personality traits. The more that a person’s behavior is a rational, predictable, consequence of those traits—the desires, beliefs, values, sensitivities, motivations, and so on that guide the person’s decisions—the more we should expect ourselves to respond to those decisions with reactive, responsibility-themed attitudes.

And that’s exactly what we see. No surprises.

[Brian Parks]

Oops... when I wrote in paragraph 4,

"the question of *why* we human beings might be inclined to consider v-defects exculpatory, but not r-defects."

I meant the other way around, i.e., "why we consider r-defects to be exculpatory, but not v-defects.

[Neil]

Kip,

Fischer argues that reactivity is all of a piece - evidence that a person is capable of reacting to reasons (including moral reasons) in some kind of patterned way is evidence that they are capable of reactivity across the board. He would say something similar about responsiveness, I think. So if the agent is capable of responding to moral reasons in some kind of patterned way, failure to respond to a particular moral reason doesn't show that they couldn't have responded to that very reason: we are entitled to assume that the mechanism upon which they acted was properly responsive. So unless the person is really morally odd, we are entitled to assume that they satisfy the reasons-responsiveness part of MR. Now, this is a defeasible assumption: if it can be shown that wrt to some class of reasons that there is a failure of reactivity, all bets are off. So Fischer could excuse scorpion: if he is not globally responsive, or if he is compulsive wrt to a single class of reasons.

[Mark Smeltzer]

Michael and Kip,

To wit, for S to be a proper object of scorn, S must deserve to be treated scornfully; and scorn is certainly a form of (retributive) punishment.

For instance (taking a queue from the story), frogs ought to deny any request from a scorpion that needs help crossing a river. The ethical impetus implicit in that rejection constitutes a form of retributive punishment (formed on the basis of a scorpion's very nature), does it not?

The interesting thing in this story is that scorpions deserve this form of punishment from frogs not because they each have killed frogs while crossing rivers, but because the property will-kill-frog-while-crossing-river is part of the intrinsic nature of all scorpions (if the story is taken as fact) -- something which no scorpion has any degree of control over. (Although we might imagine another story that depicts a remorseful scorpion who decides to rid himself of his tail so that he never becomes a frog killer, and thus is able to make amends for that unfortunate quality.)

[Badda Being]

Neil, your defense of Fischer only goes to show that a distinction between rationality defects and value defects isn't necessary to formulate Kip's problem, which is that anything can be reduced to a failure of reactivity depending on how you schematize actions subsequent to some "reason" as either morally or compulsorily patterned, or as lacking any pattern whatsoever -- it's completely arbitrary. (Meanwhile, if Fischer only selectively recognizes such failure in spite of this fact, he is thereby implicated in moral realism or meta-ethical prescriptivism, as this implies that he nonetheless regards his own schematization to be the correct one.)

[Neil]

Um, you could try to recognize the patterns there really are? That seems a good way to proceed.

[Neil]

Less tersely, BB, your response seems to be that there is no way of cutting reactivity at its joints. That's an empirical hypothesis, and given the mass of data that bears on it, it looks like a shaky one. At least I'm going to need a good argument to be impressed.

[Kip]

Brian,

You make an excellent point. "Holding responsible" despite v-defects is rational in a way that "holding responsible" despite r-defects is not. This still leaves the question: to what extent should people with v-defects be excused, because their defects are not their fault.

Neil,

Surely, Fischer could excuse scorpion if scorpion failed to meet some of the rationality requirements in Fischer's account. The more interesting question is whether the scorpion could be excused, even if perfectly rational, because the scorpion has the v-defect of wanting to sting anthropomorphized frogs.

It's not clear to me that Fischer's account considers anything like v-defects. In fact, as I wrote in the original post, it seems to me that v-defects are largely ignored by compatibilists.

One ambiguity involves the term "moral reason." Is the belief that it is morally right to torture innocent people a "moral reason"? Or does Fischer only mean to include *true* moral beliefs, like "it is morally wrong to torture innocent people" (assuming that is wrong in an objective sense).

If Fischer means the former, then his account seems blind to v-defects, because it doesn't distinguish between "it's morally right to torture" and "it's morally wrong to torture." Both are moral reasons.

If Fischer only means the latter, then he is hinging his view on moral realism, and a particular content of morality. That seems to be a huge commitment, one that I'm not sure he wants to make.

There's one last point to make about Fischer's view (and compatibilist views in general): as you note, Fischer excuses people when certain r-defects result in bad behavior. But he doesn't excuse people when certain other r-defects result in bad behavior. Why? Fischer seems to be saying "well, the r-defect has to be particularly egregious before we will excuse you for acting in accordance with it."

That is like saying "you have a disease that makes you crippled, so you couldn't even complete the race, so we don't blame you for not winning the gold medal," to one person, while also saying "you had a cold, which resulted in you missing the gold medal by .5 seconds, but it was just a little cold, so we still hold you responsible."

[Eddy Nahmias]

(I haven't read all the posts closely to see if this has been covered but...)
Isn't it possible that having 'v-defects' as severe as the scorpion's indicates 'r-defects'? If one is simply unable to see that wanting to sting even at the cost of one's own death is irrational and/or to control one's behavior accordingly, it seems one is not reasons-responsive, even on Fischer and Ravizza's rather weak notion, but if not, certainly on a slightly more stingent notion of reasons-responsive like the one I hold.

Indeed, this is the sort of worry I have with Wolf's case of JoJo--if he would not give up his sadistic desires even in the face of good reasons to do so (e.g., being exposed to them by coming to America or reading philosophy!), then I'm not sure we should accept her stipulation that he is, on her definition, sane.

[Badda Being]

Neil, reactivity patterns are serialities. They can only be as real as, say, racial and gender categories and suchlike. That one can cut them at the joints is an essentialist hypothesis in excess of the mass of empirical data. I thought that was too obvious to bear arguing for, but I should remember that I'm coming at this from a different background.

[Kip]

Thanks for your comment, Eddy.

Your comment fascinates me because, even before you had written it, I had suspected that compatibilists tend to think that a lack of r-defects can *cure* v-defects.

Return to the scorpion. Suppose, as I suggested, that the scorpion is perfectly rational, but that he defectively values stinging innocents. Suppose that he values stinging innocents incredibly high (e.g. infinitely), so much so that he would be willing to die to sting an innocent frog.

This value of his is analogous to the value you might have for your family members: to be willing to die to protect them.

Now, you ask, does this v-defect suggest that the scorpion *also* has an r-defect? In other words, if we gave a perfectly rational scorpion this evil desire, shouldn't his perfectly rational brain be able to think its way into shedding that desire?

I absolutely believe the answer is no. This is what Hume meant, I think, when he wrote that "reason is the slave of the passions." There's nothing about rationality, per se, that suggests the content of morality. Psychopaths, serial killers, sadists, suicide bombers, etc., can all be perfectly rational, if they have the right utility function (e.g. if they hold these evil acts as being incredible valuable). Whether an act is good/right or bad/evil is a question above and beyond rationality.

I think compatibilists, more than non-compatibilists, think that healthy rationality can cure v-defects. They think people should be able to reason themselves out of beliefs like "torturing innocents for fun is morally good."

And because they think healthy rationality can cure v-defects, they don't worry about v-defects. Unfortunately, rationality is too weak to save them.

[Mark Smeltzer]

Kip,

Is there a criterion that we can use to identify v-defects?

R-defects seem relatively noncontroversial because the model for ideal rationality is built upon the idea of always making logically correct inferences, and from that we can assess the defectiveness of any particular rational process by counting the number of incorrect inferences it produces.

But... to even define a v-defect seems to presuppose axiological realism, does it not? Isn't that something you would shy away from?

[Brian Parks]

Kip,

"You make an excellent point. "Holding responsible" despite v-defects is rational in a way that "holding responsible" despite r-defects is not. This still leaves the question: to what extent should people with v-defects be excused, because their defects are not their fault."

Actually, my point was that the evolutionary purpose of reactive attitudes is to influence v-defects, not r-defects.

With the exception of cases where reactive attitudes lead to killings or incapacitations, they do not have the capacity to effectively alter the consequences of r-defects. Whipping a severely retarded, insane person who has no ability to act in accordance with reasons will do little to change the person's future behavior. It represents a waste of physical and emotional resources, and to that that extent it is evolutionary maladaptive.

Whipping Madoff, however, will most definitely bring about changes in his behavior, at least as long as the threat of future whippings remains present to influence his future decisional calculus. He may have a crappy set of character traits, but he is nonetheless a rational being. He is fully sensitive to his own suffering, and will do what is necessary to avoid it.

Now, I agree wholeheartedly that the compatiblist distinction between a v-defect and an r-defect is arbitrary. Both kinds of defects have deterministic correlates in the brain, and both have profound effects on behavior that are *not* person-specific. If you or I had Madoff's v-defects, we would make the same sheister choices that he makes. Likewise, if you or I had the brain of some severely retarded, insane criminal, we would commit same crazy crimes stuff that the criminal commits.

The beauty of physics and biology is that they don't discriminate between my 'self' and your 'self' and his 'self' and her 'self.' To them, we're all just one universe unfolding according to the same causal laws and principles.

I do, however, think that it would be helpful for us to explain to the realists *why* deviant behaviors that follow deterministically from v-defects make them feel stronger reactive emotions than deviant behaviors that follow deterministically from r-defects. There's a reason, and the reason has *nothing* do with the tenuous conjecture that one kind of defect somehow 'really' is exculpatory while the other not.

The reason lies in the evolutionary processes that shaped our brains to feel reactive emotions--anger, blame, praise, gratitude, and so on--in the first place. To ignore such processes and to take the literal truth of our reactive emotions for granted is to miss the *entirety* of what's actually going on.

[Brian Parks]

Mark,

"But... to even define a v-defect seems to presuppose axiological realism, does it not? Isn't that something you would shy away from?"

To avoid the implication of axiological realism, I think it might help to change our terminology from "v-defect" to "v-trait."

We can roughly define a "v-trait" as a disposition in an individual to have a certain desire or motivation. Kip's point is that the ability to rationally process information in accordance with logical principles does not, in itself, generate any specific desires or motivations.

[Neil]

Kip, you're not getting the account. Fischer is not concerned with moral reasons one by one; he is concerned with patterns of responses. So the first question is whether the scorpion's v-defect is isolated or not. If it is not, then we have (prima facie) evidence for global moral responsiveness, and therefore for a positive answer to the semi-compatibilist descendant to the 'could have done otherwise' question. As I said, the evidence is defeasible: we will want to know not only about whether scorpion responds to other moral reasons, but also whether he is capable of responding to this one in other contexts.

Fischer does not owe us an account of what moral reasons are. So far as I can tell, his account is compatible with all the leading contenders. The way of progress is specialisation; if you try to do everything, you will achieve nothing. Finally, you say this is a problem for compatibilism. But all accounts of MR aim to give an account of action which is guided by reasons; I'm not seeing that the problem has anything to do with causal determinism.

BB, we infer real regularities from data by abduction. This is how science works. Is it essentialist? Well, yes . Science is, inter alia, about discovering essences.

[Kip]

Neil,

We're talking past each other.

I understand the point you are trying to make. I've read Responsibility and Control. And I understand the distinction between a mechanism that fails to respond to just one reason in one case, and a mechanism that systemically fails to respond to reasons in many cases.

My previous comment had two points:

1. that Fischer's account either doesn't consider v-defects (if "moral reason" is interpreted broadly) or commits itself to moral realism, etc., (if "moral reason" is interpreted narrowly)
2. that Fischer's distinction between moderate reasons-responsiveness and weak or strong reasons-responsiveness is ad hoc, because any flaw in the agent's mechanism is (generally) not the agent's fault, and therefore the agent should be excused for actions that result from these defects--whether the defects are of the weak reasons-responsiveness or the moderate reasons-responsiveness type.

Regarding 1, you say that Fischer doesn't owe us an account of "moral reasoning." I'm not sure if I would say that Fischer owes us anything. The point 1 above is a fork, and I think the consequences are not good for Fischer's view regardless of whether he adopts the narrower or broader definition of "moral reason."

Regarding 2, you say that Fischer's view can account for situations like Scorpion by distinguishing between weak and moderate reasons-responsiveness. I agree that, if the distinction is justified, the account handles the situations nicely. My point in 2, however, argues that the distinction is not justified: whether the defect is relatively minor (moderate reasons-responsiveness) or relatively egregious (weak reasons-responsiveness), the defect is not the agent's fault, and we shouldn't hold the agent accountable for actions that issue from that defect.

But perhaps I've somehow become terribly misguided. I get nervous arguing these subtleties with Oxford philosophy professors...

[Kip]

Mark,

If "v-defect" bothers you, we can certainly go the "v-trait" route, as Brian proposes.

"V-defect" comes from considering a particular situation: agent X has committed a wrongdoing. Something caused him to commit this wrong doing. The psychological reasons for why he committed the wrongdoing can be divided into two kinds:

R-defects: errors in logical reasoning, given certain evaluative beliefs (values, desires, wants, preferences, etc.)
V-defects: errors in the evaluative beliefs themselves, regardless of whether the person is otherwise rational

In this context, everyone is *already* agreeing that the agent committed wrongdoing. A moral anti-realist like myself might say "well, I don't know if it was wrong for him to kill those people in some absolute, platonic sense, but I certainly know that I don't want to get shot, and so we need to do something about this guy." We might call this practical moral realism.

From that practical moral realist sense, from the sense in which society agrees that we should do something about people who cause harm to others, we can ask the question "why did the agent commit the wrongdoing?" And, if we identify an evaluative belief (e.g. value, desire, want, preference) that caused the wrongdoing, then we can call that a v-defect.

[Neil]

Kip, by "moral reasons" Fischer means the reason that we either take or recognize (depending on your meta-ethics) to be moral reasons; we have to apply the theory, so it is our views that count. That doesn't mean that Fischer can't recognize v-defects as v-defects: Scorpion's taking himself to have the moral reason to sting frogs comes out on our interpretation as scorpion's being blind to the moral reasons against stinging frogs. So I don't see the problem.

I still thinking you're missing the point about what work moderate reasons responsiveness is supposed to do for Fischer. The point of the test is that it is supposed to reveal whether the mechanism upon which the agent acts is in fact responsive and reactive to the very reasons for failing to respond to we are holding him responsible. The idea is that capacities of agents are multi-track, and are revealed in patterns of behavior. Just as I know whether a particular basketball player had the capacity to make a shot by looking at their record from the free throw line, so I discover the properties of a mechanism by looking at a range of relevant counterfactuals.

From April to November I am mere Australian philosopher, and therefore 10% less smart. So I could be wrong.

[Badda Being]

Neil, there are always "other moral reasons" behind actions that diverge from a given pattern, but they may not be included among the patterns you recognize. Thus your entire system of patterns could be disrupted by events which, in order to preserve the truth of that system, are therefore relegated to the category of blameworthy actions. Interestingly, this means that for anyone who is afflicted with a sense of society's moral decay, his inferences are inadequate to the data in front of him. He is behind the curve in terms of the future genealogy of morals, trapped in ideology. For: any action which is in one sense morally blameworthy is in another sense germinal to a new pattern. History is so replete with examples that I dare not insult your intelligence by giving you one. But would Fischer confess that Ted Bundy was an ethical pioneer whose time never came? I have my doubts. So: there is nothing at all essential about the patterns he discerns other than the fact that he simply posits them and subsequently protects them through a series of exclusionary practices disguised as philosophy.

And yet, and yet: in the recently emerging spirit of conciliation, I would like to qualify my statements by asserting that I am by far the most unschooled person here in both philosophy and basic manners, so not only could I be wrong, I could be making a complete ass of myself.

[Neil]

Well, I don't see any evidence of lack of manners, and if you're unschooled in philosophy you're doing a decent job of holding your own. But I don't think the problem you've identified is really a problem. There are two ways of spelling out the scenaio. On one, the divergence between the moral reasons someone recognizes - correctly - and our moral views is localized. That can happen, and if it does then of course we can make a mistake in the way you suggest, holding someone blameworthy for violating a moral norm when they did the right thing. But that's not a problem for an account of moral responsibility; that's just a basic fact of life in a world in which the folk morality is not perfect. The fault would lie not with the account of moral responsibilty, but with the substantive view of morality. Again, Fischer is offering the former and not the latter.

What if the divergence is really great? In that case, the theory might be too lenient: we would fail to recognize a pattern of reactivity that is really there, and excuse inappropriately. First comment: that's not a great cost (for my money anyway). Second comment: really radical divergences between our morality and someone else's are unlikely and may be literally inconceivable, for a range of reasons (to do with radical translation, with the way in which the natural properties upon which moral properties supervene causally regulate moral terms, and for good evolutionary reasons).

[Paul Torek]

Badda brought up meta-ethical prescriptivism, but unfortunately the point seems to have been ignored. Prescriptivism is one non-realist meta-ethics that allows a distinction between "what the agent considers moral reasons" and simply "moral reasons"; there are others - or at least proposals under different labels. If Fischer is forced to choose between moral realism or prescriptivism or projectivism or ... - that doesn't seem a terribly restrictive choice.

Brian,

You tell a good evolutionary story, one that I think is at least roughly correct. But is this supposed to have any implications for FW or MR realism?

Compare: a good neuropsychologist might be able to trace a chain of causes from light reflected off an object on your table, to the activation of many red cones in your retinas, to the activation of various circuits in your brain, to the typing of the word "apple" on your keyboard. And she could do this without ever mentioning the purported fact that there is an apple on your table. Does that mean there is no apple, or that it played no role in what you typed?

Or compare: we could re-tell the evolutionary story omitting all reference to "desire", "motive", and the like. We could focus on patterns of behavior, such as maintaining a certain minimum distance from large carnivores, and the genes that promote such behavior, and never mention fear of carnivores. Would that justify motive skepticism?

[Brian Parks]

Paul,

Thanks for the comment, I look forward to what will surely be an interesting philosophical discussion ;-)

“You tell a good evolutionary story, one that I think is at least roughly correct. But is this supposed to have any implications for FW or MR realism?”

I think so.

Let me start by saying this. If we assume, in accordance with traditional theism, that an eternal divine mind created human minds, either by a direct biblical act or by ‘guiding’ the process of evolution in a certain way, then moral realism—including realism about moral responsibility—becomes an internally consistent position. On such an assumption, the moral realist can make sense of the fact that our moral sensitivities match the objective truth about morality. They match *precisely* because God designed them to match. He gave us a mind with certain moral sensitivities so that we would be able to recognize his moral law and follow it in our actions.

But if we abandon the traditional theistic assumption, as I think we have to do at this point if we want to be intellectually honest, then serious problems emerge for the moral realist. The position of moral realism ceases to be internally consistent.

Why does moral realism cease to be internally consistent on the rejection of traditional theism?

The answer is subtle but powerful, so let me make it carefully, starting with a thought experiment.

Imagine a universe, call it P-Universe, structurally identical to our universe except for the fact that (P) below constitutes an objective moral truth,

(P): Every mother should kill her first born child.

Granted, (P) may not fit well with our intuitions (for obvious evolutionary reasons), but its truth represents a possibility, at least on the assumption of moral realism.

Suppose that P-Universe, like our universe, has rational, conscious human beings with a physical and psychological nature that emerged out of a process of millions of years of evolution by natural selection.

Ask yourself. What are the odds that the human beings in P-Universe would have evolved to be genuinely sensitive to the truth of (P)? What are the chances that the women in (P) would have evolved to *feel* the moral goodness of killing their first born babies? What are the chances that the men in (P) would have evolved to be content with the killings, indeed to admire them as instances of exemplary moral behavior on the part of the mother?

Can you imagine going up to a woman in P-Universe and asking her whether she believes that it’s moral to kill her first born, and then having her say, “Obviously it’s moral! Can’t you see that? Have you no appreciation for good and bad, right and wrong? How could I not kill my first born? Ugggh… the thought…”

Obviously, you aren’t going to get that answer, not even in P-Universe. You won’t get that answer in any universe where evolution operates on its own.

If a gene emerged in the population of P-Universe that inclined individuals to be genuinely sensitive to the truth of (P), that gene would not be able to reliably select through a single generation, much less through the long-term. Indeed, what would select through the long-term would be the opposite of that gene, a gene that inclined parents to wholeheartedly reject (P), to feel *intense* love for their children and a strong moral obligation to care for them, not an obligation to kill them.

Now, contrast the moral claim of (P), with the factual claim,

(G): Sugar is poisonous to all life forms.

Let’s look at the implications of the difference between (G) and (P).

If a woman in P-Universe lacks an inclination to perceive the truth of (P), and fails to kill her first born child, nothing happens to her. No consequences. It’s like (P) was never true in the first place, it has no way to force its presence on anyone, no way to exert a causal impact, no capacity to push back. There is no reason for evolution to select for a brain that can perceive it.

But what about (G)? If an individual in G-Universe lacks an inclination to perceive the truth of (G), and proceeds to eat sugar, something is *definitely* going to happen. The individual is going to die, and so will her genes. (G) pushes back on the organism, it has causal impact, and therefore there is a way—indeed, an urgent need—for evolution to select for a brain that can perceive and internally represent it.

The point, then, is this. If we drop traditional theism, and accept the fact that our mental faculties are products of natural selection, then moral realism becomes an untenable position. Even if there were objective moral truths, we would have absolutely no way to reliably know what they are, no reason to believe that our moral sensitivities in any way map to them. Indeed, we would have strong reason to believe that our moral sensitivities do *not* map to them; they are invisible to the evolutionary processes through which our moral sensitivities emerged.

“Compare: a good neuropsychologist might be able to trace a chain of causes from light reflected off an object on your table, to the activation of many red cones in your retinas, to the activation of various circuits in your brain, to the typing of the word "apple" on your keyboard. And she could do this without ever mentioning the purported fact that there is an apple on your table. Does that mean there is no apple, or that it played no role in what you typed?”

But on the assumption that there really is an apple—or something—out there, she can explain how it is possible for my organism to have evolved the ability to accurately represent its information. There is a well-defined evolutionary value to my having that ability. I am more likely to survive and reproduce if I can internally map and process the objects that exist in my environment.

In contrast, if there really is some normative, non-factual, moral truth associated with a given behavior, she would not be able to explain how it is possible for my organism to have evolved the ability to *accurately* and *veridically* sense that truth. My moral sensitivities have not evolved because they map to moral truth per se, they have evolved because they promote behaviors that are net-conducive to my survival and reproduction.

Do the behaviors that are net-conducive to my survival and reproduction just *happen* to be the same behaviors that are consistent with objective moral truth? What a wonderful, convenient coincidence that would be, no?

Again, I think if we stop clinging to what we *want* to be true, and just take an honest, rational, scientifically-informed look at reality, we will see quite clearly that our moral sensitivities are just that: feelings that have selected their way into the human psyche in virtue of the adaptive behaviors that they give rise to. Absent the assumption of divine intervention, it is *inscrutably* unlikely that they represent anything more than that.

[Badda Being]

Neil, thank you kindly for the clarification. However, thinking past what you have written, I would like to suggest that Fischer misplaces the locus of moral responsibility relative to these two scenarios. In the first scenario, moral responsibility is a mere specter of what it is in the second. Ted Bundy is then to be considered more morally responsible than anyone generally thought to be a model citizen. But so also is Rosa Parks or Gandhi, who also were not generally thought to be model citizens. So even -- and especially -- is the subject which is otherwise inconceivable as a moral agent, and which we therefore conceive of as a mere event or, if you will, an act of God. Meanwhile, the model citizen is then to be considered a moral agent on autopilot.

I'll leave my comments at that and take your feedback off the air, lest this discussion spins further off track.

[Paul Torek]

Brian,

Your specification of the P-universe violates the widely accepted meta-ethical principle of supervenience. You posit a naturalistically identical universe with different moral properties. No ethical naturalist will follow you there, and most non-naturalists will balk as well.

Additionally, I doubt that moral-responsibility realism requires moral realism. At least, it's not obvious. I prefer a non-realist account of "humorous", but I suspect we might be able to come up with some interesting conceptual truths about humor, or reliable generalizations that are independent of one's particular sense of humor.

[Brian Parks]

“Your specification of the P-universe violates the widely accepted meta-ethical principle of supervenience. You posit a naturalistically identical universe with different moral properties. No ethical naturalist will follow you there, and most non-naturalists will balk as well.”

The assumption that P-Universe is naturalistically identical to our universe is not needed. You can take P-Universe to be whatever kind of universe you want it to be, with whatever initial conditions and physical laws you find appropriate, so long as those conditions and laws allow for the existence of conscious organisms that behave based on their mental states. The point is that if such organisms emerged in P-Universe through an evolutionary process, then they will *not* be sensitive to the truth of (P).

“I prefer a non-realist account of "humorous", but I suspect we might be able to come up with some interesting conceptual truths about humor, or reliable generalizations that are independent of one's particular sense of humor.”

We can identify truths about what makes me laugh, what makes you laugh, what laughing is like internally, what it maps to in terms of brain processes, how it emerged evolutionarily, and so on. But we cannot identify truths about whether something is really, truly, objectively “laugh-worthy”, or whether it “should” make us laugh, or whether it “deserves to be laughed at.” The universe does not have a sense of humor, it cannot manifest those sorts of truths.

Likewise for moral responsibility. We can identify truths about what qualities an agent’s choices would need to have in order to successfully provoke judgments of praise, blame, or neither from specific individuals. But we cannot identify truths about what qualities an agent’s choices would need to have in order for those judgments to be really, truly, objectively accurate or deserved.

“Additionally, I doubt that moral-responsibility realism requires moral realism. At least, it's not obvious.”

Moral-responsibility realism and moral realism are both instances of normative realism. Therein lies the problem.
Let me explain.

Suppose there is a dispute about whether electrons attract or repel each other. The object of the dispute is a fact, and so the universe would take a side. It would make its side readily known in the form of what actually happens in the experiment.

Now, suppose there is a dispute about whether an agent is morally responsible for a choice. Does the universe take a side in that dispute? If so, how does it reveal its side?

What would it mean for the universe to say “Yes, Fischer, to deserve judgment for a choice, an agent must be moderately reasons-responsive. No, Kane, to deserve judgment for a choice, an agent does not need to have been able to choose otherwise. No, Strawson, to deserve judgment for a choice, an agent does not need to be the ultimate cause of that choice.”

????

Moral responsibility is a normative concept, not a factual concept. When Fischer says “Madoff is morally responsible”, he doesn’t just mean to say “Madoff is reasons-responsive.” If that were all he meant to say, then he would just say that: “Madoff is reasons-responsive.” A dry fact. No one will disagree.

“Madoff is reasons-responsive” doesn’t get the point across. Fischer is trying to make a normative point: “Madoff is reasons-responsive AND reasons-responsiveness is THE condition that makes people deserve judgment for their choices.”

So, though there may be a subtle distinction between moral realism and moral-responsibility realism, they are both untenable in light of evolution for the same reasons: they target norms, not facts.

Suppose that it is an objective truth that no one is morally responsible for anything. No one ever deserves to be the object of reward or punishment, not under any circumstances. Why would human beings evolve to be sensitive to that truth? What would such a sensitivity accomplish in terms of gene survival, other than to obstruct the valuable behavior-shaping function that reward and punishment fulfill in our species?

Or suppose that people are responsible for their r-defects, but not their v-defects. Would something that was previously a waste of time, i.e., expending resources in an effort to make those with r-defects suffer retributively for their choices, suddenly become an effective way of changing the future behavior of those with r-defects? Would something that was previously an effective way of changing the future behavior of those with v-defects, i.e., expending resources in an effort to make those with v-defects suffer retributively for their choices, suddenly become a waste of time?

The fundamental question remains to be answered, hopefully someone can try to answer it: If there are objective truths about non-factual norms in our universe—i.e, truths about what is objectively right and wrong, or about the extent to which people genuinely deserve to be praised and blamed for their choices—why would an evolutionary process select for brains that are specifically sensitive to them? What benefit would that provide to a conscious organism in terms of its survival and reproduction?

I’ll grant that it’s *possible* that the normative truths of our universe just happen—by *pure* chance—to match the psychological sensitivities that an entirely naturalistic evolutionary process seeking to maximize gene survival would select for. But the probability is *inscrutably* low.

[Mark Smeltzer]

Brian,

Actually, all Fischer is committed to saying is that Madoff is a suitable candidate for the reactive attitudes -- assuming he takes responsibility for acting on his own MRR mechanism (you have read some of his papers, right?). He is not committed to saying anything in particular about what Madoff deserves.

Richard Double follows a path similar to the one you've sketched in his book The Nonreality of Free Will : if moral realism is true, then moral responsibility is possible; moral realism is false; therefore, moral responsibility is impossible.

It is a rather boring argument: who is he trying to convince?

If it is taken for granted that moral responsibility requires moral realism to be true, then shouldn't it be obvious to Double that defenders of moral responsibility are conscious advocates of moral realism? And if that is the case, how many of those defenders of MR are going to give such an argument even an iota of consideration? "Okay, Mr. Parks.. ahhem, err, I mean, Mr. Double, since I reject premise #2, now what?" they might ask.

Double at least has the sense to climb down off his horse to engage in meaningful debate by scrutinizing the specifics of his opponents' accounts in the remainder of his book (one wonders how else he would have filled its pages!). That, at least, has the potential of being a productive path.

[Brian Parks]

Mark,

“Richard Double follows a path similar to the one you've sketched in his book The Nonreality of Free Will : if moral realism is true, then moral responsibility is possible; moral realism is false; therefore, moral responsibility is impossible.”

Hmm… I think you mean : if moral responsibility is possible, then moral realism is true. Moral realism is false, therefore moral responsibility is impossible.

“It is a rather boring argument: who is he trying to convince? If it is taken for granted that moral responsibility requires moral realism to be true, then shouldn't it be obvious to Double that defenders of moral responsibility are conscious advocates of moral realism?”

No. Why should that be obvious? There are defenders of moral responsibility—academic and lay—who have yet to rigorously consider the matter on that level. He is probably writing for them—in addition to whoever else wants to know the truth.

Putting Double’s argument aside for the moment, *my* argument focuses specifically on the internal consistency of the pro-moral-responsibility position. There are scientifically-minded, non-theistic philosophers who assert, or at least implicitly accept, that our psychological sensitivities match with a set of real, objective normative truths about morality and moral responsibility. That assertion is incompatible—or at least in *serious* tension with—the widely-accepted scientific claim that our psychological sensitivities emerged through an impersonal process of evolution by natural selection.

This incompatibility results from the fact that normative truths, in contrast with factual truths, do not “impact” or “push back” on biological processes. Every normative truth in our universe could revert to its polar opposite, but that would not in any way affect the practical considerations that make this kind of brain that I have, with a tendency towards these kinds of psychological sensitivities, net-conducive to my survival and reproduction. The kind of brain that I have would still have been more adaptive than its alternatives, and therefore it would still have selected in evolution, regardless of the fact that it would not have been appropriately calibrated to sense normative truths.

Normative truths are *invisible* to evolution. Evolution doesn’t care about them. Thus, the only way that evolution could produce a brain with psychological sensitivities that accurately represent them would be if it did so entirely through the luck of the draw. Not likely.

“Okay, Mr. Parks.. ahhem, err, I mean, Mr. Double, since I reject premise #2, now what?”

Well, you’ve identified the problem, haven’t you? Until you fix it, the actual truth will remain out of reach for you ;-)

Seriously, though, I agree that the debate about moral responsibility represents an entertaining intellectual exercise, and I’m perfectly willing to assume moral realism in order to engage in it.

I assume that most of us, however, will eventually want to get to the bottom of things, to the actual truth. To do that, we need to critically explore the assumption of moral realism. If we are going to hold onto that assumption, we need to confront the significant problems that well-established science poses for its viability.

If moral responsibility advocates want to ignore those problems, to dismiss them as “not interesting”, or as “not consistent with my world view”, that’s fine. But it’s their loss. They are depriving themselves of the actual truth: the whole point of doing philosophy in the first place.

[Mark Smeltzer]

Brian,

I agree that the question of moral realism is very interesting in its own right. Moreover, I think we can agree that the question of moral realism is a debate unto itself.

In the context of this discussion, however, I think there is a way to make your worry somewhat interesting: have you thought about analyzing Fischer's account from a Strawsonian perspective? PF Strawson followed in the tradition of Humean Sentimentalism and held more to the pragmatic side of things.

On the pragmatic side (which natural selection certainly would be sensitive to), we can assert the realism of moral communities insofar as we can assert the realism of communities of agents who have (many) common reactive attitudes. In the context of the moral communities, the question can be asked, "What qualifies as a suitable candidate for these reactive attitudes to be applied to?"

Surely it could be decided that it is an exercise in futility to apply the reactive attitudes to things like rocks, trees and other mundane objects. From there, things get more complex... But, surely you could see how even from an evolutionary perspective that things might be explained in that way, and in this scenario it makes sense for moral communities to talk about a common standard for moral responsibility even though some of the communities disagree about the moral standards. If that is the case, does moral non-realism really make moral responsibility irrelevant?

Fischer doesn't seem to think so. It was part of his game plan to formulate his account in a manner that addresses even a critic like Double's concerns. Sure, a moral realist will be able to take Fischer's account much further than a moral non-realist, but I'm not sure if I'm ready to accept your call to arms that moral non-realists have to reject Fischer's account outright.

(I may have demonized Double's argument a bit in my prior summary, but it seems I got the basic idea across.)

[Whitney Segura]

Great post! It was very well researched and I enjoyed it very much. I bookmarked your site and will be back very soon, I look forward to reading some interesting posts! Thanks, Whitney

[Paul Torek]

Brian,

Suppose that comedians divide between "compatibilists" who think that humor is compatible with the audience's expecting a joke, and "incompatibilists" who think that it is not. Couldn't we imagine this disagreement being settled objectively? Suppose humor-researchers eventually discover a natural law to the effect that no one finds something funny unless surprised by it. Game over, "incompatibilists" win. Note that you cannot build a funny-o-meter based on this knowledge. Humor remains subjective.

This is not intended to be fully analogous to the ethical realism and moral responsibility question. You can tell, because in the analogue the incompatibilists win. :-) But I hope it shows that even if a topic is subjective, we might be able to state objective conditional truths about it.

Regarding the P-universe, things would have to be more than just a little different, naturalistically, for P to be true. Now suppose for the sake of argument that in the nearest possible worlds in which P is true, the organisms think it's false. That doesn't prove anything about our truth-tracking abilities, because there's just too much distance between us and them. Our cognition tracks truth in some domain iff (roughly) in the vast majority of nearby possible worlds containing truths in the domain, we get them right more often than wrong. Since the P-universe is arguably a distant possible world, it's not germane.

Suppose that in the Q-universe,
(Q) sex is always torturous for both partners.
Would an intelligent, evolved species that reproduced sexually, believe that Q is true? Probably not; it seems more likely that they'd embrace some comforting illusions, and develop selective memory loss, which would prevent the truth of (Q) from deterring sex. But does this show that we don't really know that sex can be rewarding? No, because that universe is just way too far out there to bear on our situation.

If there are naturalistic ethical facts, what kind of facts could they be? Naturalists differ, but all place ethics squarely in a High Traffic Zone of evolutionary pressures. If ethics is about, say, the flourishing of sapient beings, there isn't room for a complete disconnect between evolution and ethics. Suppose "flourishing" is further cashed out in terms of the (evolved) sense of well-being: the promotion of well-being can be expected to align well with the promotion of survival and reproduction.

For the above sort of ethical naturalist, there is an imperfect but very non-coincidental congruence between ethical value and evolutionary fitness. But some (e.g., Stephen R. Brown and William Casebeer) go even further. They view ethics as "functional" in a sense roughly comparable to that in which the function of the heart is to pump blood; i.e. that is what it was selected for. On such a view, it seems likely that your P-universe argument is a non-starter. Possible worlds in which P is true would tend to be precisely those in which killing one's first-born enhanced evolutionary fitness. So, where P was true, they would believe it.

You might find such "evolutionary ethical naturalism" (to give it a label) implausible. But you need to argue against it, not just presuppose its falsity. Else, you could be just like the neuropsychologist in my previous post, explaining "away" the perception of an apple while ignoring the fact that the very content of the explanation reveals the presence of an actual apple.

[Paul Torek]

Correction: actually, I think either kind of ethical naturalism potentially poses the "actual apple" problem for you. It's just more obvious in the case of evolutionary ethical naturalism.