Blog Coordinator

« CEU Conference: The Manipulation Argument | Main | Interpretation, Embodiment and Responsibility Conference »



Feed You can follow this conversation by subscribing to the comment feed for this post.

Hi Tamler,

This is really cool – thanks! I share your intuition that there is nothing wrong or unfair about the fact that Lindsay and Lauren received different punishments, but I wonder whether this shows that ordinal proportionality doesn’t matter so much or whether it shows that I am implicitly assuming that, though they may vary in type, the two punishments are actually equal in severity. If it were clearly the case that Linday’s punishment was less severe on balance than Lauren’s, then I think I would think that either Sam was being objectionably merciful or that Tom (knowing the severity of the punishment his mother would dole out) was being objectionably cruel.

What if we make it easier to compare the punishments? Let’s revise the cases so that Tom and Sam have the following two options: tell dad, who they know is more lenient than mom and will only ground his daughter for one week, or tell mom, who they know will ground her daughter for two weeks. All else truly being equal, if Sam tells his dad and Tom tells his mom, I think I would ultimately think that either Sam was being objectionably merciful or that Tom was being objectionably cruel -- even if both punishments were in some sense reasonable.

As a matter of policy, perhaps it is best to give victims a prominent role in punishment decisions. Maybe doing so is the most efficient way to achieve justice over time. But that doesn’t change the fact that a decision-making procedure that relies so heavily on victim input might nevertheless deviate in particular instances from an ideal of retributive justice -- an ideal that includes ordinal proportionality.

Nice, Tamler (and Adam).

Joel Feinberg (I think in his Foundations of Philosophy [Prentice Hall Series) Political Philosophy book, although I'm not sure) made a distinction between relative and absolute justice. It seems that certain contexts involve one, but not the other. So if I'm going 80 in a 65 mile zone, it wouldn't be very reasonable for me to complain to the office that it is unfair that I be punished, because others also go 80 in 65 mile zones and don't get caught or punished. This is a context of "absolute" justice.

But there are obviously other kinds of context or situations in which relative justice is what matters. So if I say that anyone who scores more than 90 points on the exam will get an A, and Max and Melanie both get over 90 and I only give an A to Melanie and not Max (and nothing else is relevant), then this is clearly unjust.

How to distinguish the contexts--the ones in which relative justice matters or matters more and those in which absolute justice is hegemonic--I do not know.

I'm not sure that the observation above is relevant to your point exactly, but perhaps it is helpful. Also, does there have to be an actual victim, or is it pertinent that one imposes a risk of harm on others?

By the way, I just saw the American remake of The Girl with the Dragon Tattoo. (I'm a big fan of trilogies, you know!!! [Sorry!]) My wife and I really enjoy the moment when the girl's abuser gets what he deserves. I would even say: it is intrinsically good that he suffers in a way commensurate to the way he made the girl suffer. I think it is hard to argue for this kind of claim, but I do think retributivism taps into something important and deep about human beings.

Once again the desert-talk is just mystifying to me, but I do have some tho'ts on the cardinal/absolute vs. ordinal/relative issue.

What seems to be missing from the discussion is the fact that there are not just two separate "crimes" but also two separate "justice systems" involved. Suppose we add another character to the drama. Bill witnesses both attacks, but reports only one of them to the appropriate mother. That strikes me as a violation or ordinal proportionality. Given the equal harm and culpability, Bill should not be treating them differently. But neither Tom nor Sam is violating norms of ordinal proportionalty, since neither even has a chance to treat someone else differently than the way they treated their own sister.

Issues of relative fairness arise when we consider the punisher as opposed to the punishee. One prof may grade harder than another, and that's fine so long as they're both within the range of reasonable grading. But if one prof grades some students harder than others, then that's problematic -- it's something that requires a moral excuse.

Similarly, if Canada mandates two years for a fraud that would get five years in the US, that's fine so long as both are in the range of appropriateness. But if the US (and especially if one judge) gives one guy five years and another (equally harmful and culpable) guy only two, that's problematic. In some sense the US justice system is "an" actor, and so issues of ordinal proportionality arise.

That was what came to my mind, anyway.


Interesting topic! I have a question about how the presence of a victim diminishes or eliminates the demand for ordinal proportionality. If the presence of a victim makes range-only, cardinal proportionality sufficient, does this entail that it would be fair for a *single* victim to mete out different punishments to two people who are equally blameworthy for the same crime? We readily accept the difference in Lindsey and Lauren's punishment because the victim and punisher are different people. But, would we also think that these different punishments are fair if they involved the same victim and punisher? I hate to hijack your post and introduce another new variant of your original case, but here goes:

Imagine that Lindsey, Lauren, and Sam are all siblings. Lindsey and Lauren are both trying to be cool and both end up egging Sam. When the time comes, Sam gives both of his sisters a contemptuous look and they both feel really terrible and embarrassed about what they did. The next day, Lindsey leaves for camp, with the pangs of guilt still weighing on her. Sam then decides, after reflecting on the previous night's events and feeling newly angry about it, to tell only on Lauren, who is then grounded. We can imagine that Sam doesn't tell on Lindsey because she's away, and nothing will come of it. On this variant of your cases, both punishments are cardinally proportional, but Sam plays a role in punishing one more severely than the other.

This seems really unfair to me, but I'm not sure what relevant difference there is between it and your original two cases. As far as I can, tell your view would entail that Sam's differential treatment is not unfair, since cardinal proportionality is satisfied in both punishments. But, I have a strong sense that something is unfair here, and that Lauren would have some grounds on which to complain. Am I alone? One way of responding to this case is to claim that Sam is acting unfairly because he simply has no reason for punishing Lauren more severely than Lindsey. But, this is already to give ground to ordinal proportionality proponents who maintain that fairness requires reacting to such reasons. A hardcore cardinal proportionalist will say that the only relevant reasons involve the proportionality of the punishment to agent's blameworthiness. If these are the only considerations that matter for fairness, then Sam's actions seem on the up and up.

Maybe I'm alone in thinking there is something unfair about the variant I proposed? If I'm not, is there some way to make sense of the unfairness that doesn't give up ground to the ordinal proportionalist?

What would you think if a person convicted of a crime was allowed to choose different punishments laid out for them by those they harmed. For instance they could choose to spend 3 months in prison or have to endure several severely painful, but not permanently harmful, beatings (I would go for the beatings.) Would the punished person's ability to choose the most agreeable punishment be problematic because the offender might theoretically be punished less with which the choice he finds more agreeable? I don't think so - so long as the victim is satisfied and the offender is sufficiently deterred. Also I imagine in both punishments the offender might think back and wish he had chosen prison over the beating and vise versa.

Thanks everyone!

Adam, it's funny, I don't share your intuition about the revised case since both punishments seem reasonable for what happened. Are you assuming that there's an ideal punishment that falls between the two. Ten days? Eleven days? Maybe I'm not understanding you right. Nor do I think an "ideal" form of retributive punishment has a strict ordinal requirement because of the presence of the victim. I think that what's driving the intuition about fairness in my case is that Sam (and Tom) are humiliated by the act and by getting the opportunity to have some say in the punishment, they reclaim a little of their self-respect (especially Sam in this case).

John, your Melanie and Max example is victimless so I agree. That's just the kind of case Von Hirsch brings up but it's disanalogous to most cases of criminal punishment where there are victims. So that point supports what I'm saying, I think. You ask:

"Also, does there have to be an actual victim, or is it pertinent that one imposes a risk of harm on others?"

Great question, not sure. My hunch is that there needs to be an actual victim. Drunk driving cases where no one is harmed seems to me to have more of an ordinal requirement certainly than cases where someone is harmed.


For someone who finds desert talk so mystifying you sure have a lot to say on the subject! Anyhow, you make a lot of good points.

(1) Two separate justice systems. I thought of that as well, but I think there are two ways of handling that. First, it parallels at least the U.S. system in that (a) there are different judges and juries for different cases, and (b) different sentencing laws within each state. But recognize that you might call that unfair or a violation of ordinal requirements. Second and more important, you can imagine that the two parents would handle the punishments the same way-two weeks grounding. So the issue is just whether the brother is going to report it or not. I think the intuition isn't affected.

I'm not moved by the Bill example since Bill was not a victim of the offense. So you wonder why Bill is reporting one instead of the other. But if he reports on Lindsday so that she ends up getting grounded anyway, I don't find that unfair.

Your prof example again is victimless so I agree.

Philip, thanks, this is great. You write:

"If the presence of a victim makes range-only, cardinal proportionality sufficient, does this entail that it would be fair for a *single* victim to mete out different punishments to two people who are equally blameworthy for the same crime?"

I agree with you about your case too. So what I can I say? Here's one possible response--the injustice of the outcome has its source in Sam's differential treatment of his two sisters rather than in the different outcomes of the punishment. We have a strong sense that people should treat their family members equally, which is why no parent is allowed to say that they love one child more than another even if they do. (Recently, a mother confessed to me that she loves one of her children significantly more than the other, but felt really guilty about it. When I joked that she should just tell them, she was mortified.)

But your point still holds: it's not as simple as "the presence of the victim diminishes or eliminates ordinality requirements." Other conditions have to be in place as well.

Guy, I'd be fine with that, especially if the victim was satisfied with either punishment. That's the thesis of "A Defense of Flogging" by the way--that criminals should be able to choose floggings over jail-time.

I partly agree with Mark that "Issues of relative fairness arise when we consider the punisher as opposed to the punishee." And I find the illustrative cases spot-on. I'd like to offer some further analysis, which (if it's crazy) should not be held against Mark's general point.

First let's look at the sense in which the US justice system is "an" actor. Suppose two different judges hand out two-year and five-year sentences, respectively, for fraud. And suppose that the racial makeup of their jurisdictions is very different, with the five-year sentencer having been chosen by the President for her "law and order" approach that the Prez thinks is just right for the highly urban jurisdiction.

Not only do we have a problem, but I think that the five-year prisoners can legitimately complain that they are being treated unfairly. It's wrong *to* the punishees, not just *of* the punisher. So instead of Mark's "when we consider the punisher as opposed to the punishee" I would substitute "in addition to the punishee".

Why doesn't this extend across the Canadian border? Because that's a different moral conversation. Potential punishers and punishees in the US must justify policies to each other and come to some sort of approximate consensus, but they don't need to justify those policies to Canadians, or vice versa. The Canadian way of dealing with fraud will be decided in Canada, not in North America generally.

Thanks for your response, Tamler. What am I assuming? I’m just assuming that there *is* an ideal punishment. Maybe it’s one week, maybe it’s two weeks, or maybe it’s somewhere in between. So, I’m not only accepting ordinal proportionality, but I’m also denying the range-only interpretation of cardinal proportionality. (I did say that I thought both a one-week and a two-week punishment would both be reasonable, but the sense in which I think both punishments would be reasonable is similar to the sense in which I think that two competing philosophical theories might both be reasonable—the reasonableness has more to do with my lack of access to conclusive evidence than it does with denying that there might in fact be one true theory.)

I think the intuition I’m driving at can be most easily brought out by first considering Philip’s (very nice!) revision of the Freaks and Geeks case. So, say that Sam metes out two different punishments to his equally blameworthy sisters. What makes it wrong for him to do this? You say it’s wrong because we think people ought to treat their family members the same, assuming they are equally blameworthy. On the contrary, I think it's wrong simply *because* both wrongdoers in this case are equally blameworthy. As far as I’m concerned, the fact that they’re both Sam's sisters just highlights that there are probably no instrumental reasons for treating them differently.

Consequently, I don’t think anything important would change if we made the unlucky sister into an unlucky non-relative. If you tried to tell the unlucky non-relative that it was okay for Sam’s sister to get a less severe punishment, she *might* agree with you, but I think she would only do so with good reason insofar as she thought there were *instrumental* reasons for Sam to punish his sister less severely. Maybe punishing relatives requires more self-sacrifice on the part of victims than punishing non-relatives. But if punishing his sister and punishing the non-relative were equally costly, I don’t think Sam could provide any morally legitimate reason to treat his sister less harshly than the equally blameworthy non-relative.

To push this just one step further, I don’t think anything important would change even if we revised the case so that there were once again two different victims meting out the two different punishments to the two different wrongdoers. Again, if you tried to tell the unlucky wrongdoer that it was okay for the other wrongdoer to get a less severe punishment, she *might* agree with you, but I think she would only do so with good reason insofar as she thought either (a) that there were *instrumental* reasons for the other victim to punish the other wrongdoer less severely (such as those described above), or (b) that there were instrumental reasons for *allowing* the other victim to punish the other wrongdoer less severely.

With respect to this second option, you yourself seem to suggest that allowing victims to make their own decisions about punishment might be a good way to help restore victims’ self-respect. I agree! But this doesn’t necessarily show that ordinal proportionality doesn’t matter (or even that cardinal proportionality should be given a range-only interpretation); it might just show that it’s often less important to give criminals exactly what they deserve than it is to restore victims’ self-respect. So this is the sense in which I think an ideal form of retributive justice requires ordinal proportionality: if these goals of punishment (i.e., to give wrongdoers what they deserve and to restore victims’ self-respect) were not competing, then equally blameworthy wrongdoers would always receive equally severe punishments.

Are we any closer to sharing the same intuition?

Paul, thanks. Regarding your case, it seems to me that the unfairness is rooted in the seemingly racist motivations that led to the 5 year sentence. And that's independent of any considerations of ordinal proportionality. As far as U.S. being a single 'actor', states have significantly different sentencing guidelines. And unless state laws are considered unconstitutional, they don't have to be justified beyond state borders.


Not sure if we are getting closer. A couple things. First, I wasn't trying to defend a range only view of cardinal proportionality in this post, since my primary target (Von Hirsch) accepts it already. His view, along with Morris and many others, is that your assumption of a precise punishment that (cardinally) proportionate to the blameworthiness of the criminal presupposes a "particularly heroic kind of intuitionism." I think that's right, it's hard to conceive how a theory could determine precise punishments even in principle. So I think you're making a pretty big assumption, even if you're not making the further (ludicrous) claim that we have epistemic access to what those punishments would be. One question before moving to your second point: if cardinal proportionality is not 'range only,' then doesn't ordinal proportionality become redundant? Or is the idea that we need ordinal proportionality to deal with our lack of access to the 'ideal' punishment?

Second, and this is really interesting, you suggest that my cases may might "just show that it’s often less important to give criminals exactly what they deserve than it is to restore victims’ self-respect." The idea is that there are competing goals of punishment (1) give the criminal the punishment he or she deserves (according to the principles of cardinal and ordinal proportionality) and (2) restore victim's self respect.

I'm not sure if this is a semantic or substantive disagreement but I don't think (1) and (2) can be untangled in the way that (1) can be untangled from, say, utilitarian or rehabilitative goals. IN other words, you can't determine (1) without having information about (2). But you could determine (1) without having information about the deterrent effect of the punishment etc.

I think it's helpful to think of non-criminal offenses (cheating on spouses, say)--where the fair or just outcome seems to depend intimately on how the victim feels.

Thanks for these insightful criticisms everyone! This is very helpful. Keep em coming if you have them.

Cool post! Phil and Paul have already essentially made the point that occurred to me, but here it is again via a slightly different example. Suppose Lindsay and Lauren and Sam are all siblings. Suppose the girls both egg Sam and he tells Mom on both of them. Both Lindsay and Lauren are equally chagrined about what they've done to Sam. Mom then grounds Lindsay for a month, while sparing Lauren the grounding.

I'm inclined to think that this case involves significant injustice and largely for the reason that Paul adduced: when it's the same family it's the same 'legal system' and so differential treatment must be justified by reference to the norms of that system. If Mom can't point to a relevant difference between what Lindsay and Lauren did, then the justice system acted arbitrarily. However, another family could have norms that dictate different punishment that can be fairly applied in that separate 'legal system,' while generating different severities of punishment that would be unobjectionable when compared only to how other families punish.

Anyway, the idea is that cardinal proportionality is always morally relevant, but ordinal is only relevant to comparing cases within a legal system.

Zac, thanks. I disagree though and here's why. The injustice in your case is based on the total arbitrariness of the mother's decision. There doesn't seem to be any reason for her to ground Lindsay but not Lauren. So the problem isn't that Lindsay isn't getting what she deserves. It's that the mom is treating her two daughters differently for no reason.

I don't want this to get too convoluted but let me try a variation of your case. This time there are two brothers and the egging happens to Tom one year, Sam the next. The first year Tom is grounded for two weeks. The next year, Sam either doesn't tell his Mom or tells her but asks her not to ground Lindsay. He wants to handle it his own way and doesn't feel the same way Tom did about what should happen and wants to handle it his own way. So Lindsay isn't grounded. That seems OK to me. Now the Mom has a reason for treating the sisters differently. And it seems like both sisters are getting what they deserve.

So in this case: same justice system, ordinal proportionality violated, but both punishments are fair. What do you think?

Thanks, Tamler. I know you weren’t trying to defend a range-only view of cardinal proportionality here--I was just trying to make it clear that I knew I was both making a big assumption and denying something you were assuming in order to get to the point of your post. (And I actually agree that it would be hard to conceive of a theory that could determine precise punishments even in principle, but I take that as a reason to reject retributivism rather than as a reason to revise it.)

In response to your question, I think you’ve nailed it. I think that if we had epistemic access to the ‘ideal’ punishment, then cardinal proportionality would just imply ordinal proportionality. However, since we lack that access, I think we ought to at least achieve ordinal proportionality by consistently applying our best guess as to what the ideal punishment might be.

As for untangling (1) from (2), you won’t get much serious argument from me here, since I’m not sure what anyone means by ‘desert’ or if there is even a fact of the matter about what ‘desert’ means. In what I said in my previous post, though, I was intending to use it in a such a way that you could in principle determine (1) without having information about (2). But if you’re using it another way, then maybe we’re just talking past each other here.

//What do you think?//

I think you've come around to my point of view!

To be clear, you're now saying the mom needs a moral excuse for treating her dau'ters differently -- it's not enuf to say "Well, they're both in the appropriate range."

I may be mistaken (I often am), but my impression was that you were originally saying that the latter would be a sufficient excuse for the difference (at least when there was a victim): //When there is a victim, then, the ‘range-only’ view of proportionality and desert is sufficient.//

Right, to be clear, the range-only view of proportionality entails that blameworthiness (which includes culpability and harmfulness of offense) does not yield a precise punishment, just a range of punishments, all of which would be proportionate or what the person deserves. There still have to be reasons for choosing certain punishments within that spectrum. Norval Morris' view was that we should should the least punishment in the spectrum that would meet utilitarian benefits of the punishment. This will often lead to violations of ordinal proportionality. But the Judges who do this still have moral reasons for choosing punishments in that range, just not reasons that are fundamentally connected to the desert or blameworthiness of the offender.

Adam, we may be talking past each other but there's a substantive question in play (I think): are violations of ordinal proportionality fundamentally unjust?

If you believe they are, then, like von Hirsch, you'll want to reduce them as much as possible, even if doing so means disregarding the wishes of the victim or sacrificing utilitarian benefits. If you think they're not, then you can freely take these other factors into account as long as you stay within the cardinal range.

Great, I run a case by my daughter where two kids do the same bad thing in school to two other kids but in one case the victim asks the principle not to punish the student so harshly and the other doesn't. I ask my daughter if they should get different punishments. She says: " They did the same thing. Why should they get different punishments?" Then she sees my crestfallen face and says: "but that's just my opinion. I'm not sure that would actually happen." How sharper than a serpent's tooth...

That is quite funny about your daughter. Regarding the substantive question at stake, I think a lot hinges on what you mean by "fundamentally unjust". If violations of ordinal proportionality are fundamentally unjust, does that just mean we have a pro tanto reason to avoid violating ordinal proportionality? Or does it mean that we always have a sufficient reason to avoid violating ordinal proportionality? I think there is a sense in which you can believe violations of ordinal proportionality are fundamentally unjust in the former, weaker sense while also thinking that we should take into account the wishes of the victim or utilitarian benefits. But in that case, it would be misleading to claim that the presence of the victim (or potential utilitarian benefits) strictly weakens the demand for ordinal proportionality. This is because the (pro tanto) reasons to avoid violations of ordinal proportionality in these cases are just as strong as ever--it's just that the reasons to defer to the victim are sometimes even stronger.

Tamler: I've probably come to the dance too late, but nevertheless, I wanted to say that this is interesting stuff. My thought as I read the thread was similar to what Zac and others have suggested, that you need identity of jurisdictional authority to measure intuitions, and that once you have that, violations of ordinal proportionality will seem unfair. Your response to Zac was that what's unjust in his scenario is the arbitrariness of the disparate punishments. But I would have thought there are many possible sources of ordinal disproportionality, one of which could well be (and likely is, in most cases) arbitrariness in sentencing (e.g., different sentences explained, but not justified by, factors about race/gender/age, etc.).

But at any rate, the examples you give, both in the original post and in the response to Zac, don't strike me as violations of ordinal proportionality in two cases of punishment, precisely because they involve only one case of punishment, namely, the case in which mom finds out what Lindsay did. In cases where she doesn't, you massage this by saying Lindsay sees the hurt in Sam's face and that's her punishment (she feels guilt, etc.). But that's not a sentence that's been imposed by an authority that renders it analogous to the first case; instead, it's (at most) self-imposed.

Now you might say that what's been established here is just a disparity of ordinal proportionality w/r/t desert. But then we attribute all sorts of good or bad events to people in a way that just complicates the matter and doesn't reveal anything clear or uniform about the nature of (our intuitions about) desert, at least as it's supposed to pertain to retributive punishment. Karma's a bitch, as we say.

Dave, thanks, In my response to Zac though, the Mom can find out in both variations. But she reduces Lindsay's punishment at Sam's request. So you do have a judicial body of sorts issuing different punishments and violating ordinal proportionality, no?

Here's a real life version of the same case: a victim of a drunk driving accident whose legs were shattered asked the judge to reduce the sentence of the driver. As a result, prosecutors dismissed the most serious charges and the driver received a far less severe sentence than she might have otherwise, and certainly than Does that seem terrible unfair to you? (or others?)

All-too-brief description of the story is here:

OK, good, that's clearer. Here's my mixed intuition on your present case: it's not unjust that this driver didn't get what he deserved. This preserves ordinal proportionality for desert, but it makes justice (fairness?) a function of (at least) desert plus some ceteris paribus clause (where other things might include mercy in light of the victim's pleas). Is this just wacky, though?

//Norval Morris' view was that we should [choose?] the least punishment in the spectrum that would meet utilitarian benefits of the punishment.//

I'm wondering how he would describe cases where the judge chooses a greater punishment (but one that's still in the spectrum). Would that be unjust? Because then we could have examples where the criminal's punishment was simultaneously what he deserved and an injustice.

Dave, not wacky at all. I think that's what Adam is driving at as well. But if you accept that cardinal proportionality is range-only for desert, what's the motivation for dividing justice into desert plus other (non-consequentialist?) factors? Even if you think ordinal proportionality is required, where the ordinal punishment lies on the range allowed by cardinal prop. is determined by other factors that have nothing to do with blameworthiness...

Mark, Morris believes that parsimony--giving the lowest sentence that will achieve your ends--is a proper goal of punishment. So, I think the answer to your question is 'yes.'

I think my two responses here may be contradictory.


The comments to this entry are closed.

Books about Agency

3QD Prize 2014: Marcus Arvan