« Prejudice, Legal Realism, and the Right/Remedy Relationship | Main | Law schools marketing to law professors »

Thursday, May 10, 2018

Dialogical Reply to Flanders's "How Much Certainty Do We Need to Punish?"

And now for something a little different. Chad Flanders has written a reply to my Punishment and Moral Risk article in the form of a dialogue between fictional versions of ourselves. He uses "ChadF" (or just "C") to refer to fictional Chad Flanders and "Kolbert" (or just "K") to refer to fictional Adam Kolber. It seems only appropriate to respond with a dialogue as well.

[SETTING: Outside an independent bookstore in Brooklyn.]

C: Hey Kolbert, fancy seeing you two days in a row. Stay back, this is a new shirt!

K: Yeah, sorry I caused you to spill your coffee yesterday. And sorry that when I went to buy you a new one, I passive-aggressively complained about the high price of coffee these days. I guess I just wasn't myself. I even felt like some of the words I said weren't really my own.

C: Weird! I feel that way right now. There's probably some expression in German that perfectly captures the feeling. Anyway, no sweat about the coffee. In fact, overall, we probably agree about more than we disagree.

K: True, most importantly perhaps, we agree that how people treat uncertainty about desert can sensibly vary from situation to situation. I believe typical retributivists are far more willing to allow the deserving to go free than the undeserving to be punished. So they have higher credence requirements for state punishment than for, say . . .

C: . . . getting annoyed that someone spilled coffee on you.

K: Exactly. And all of this fits with what you wrote in your email, even if I wouldn't put matters exactly the same way.

C: Anything we talked about yesterday you disagree with or want to clarify?

K: Well, since you asked, in my paper, I don't defend full-blown free will skepticism. I simply raise doubts about free will. So some people might go about their ordinary lives believing in free will, apologizing for spilling coffee and so on. It's only when the risk of moral error get seriousness enough that they need to worry about being wrong about free will and related issues. 

C: What else?

K: I like your point about Bernie Madoff. 

C: Right, to oversimplify, I implied that one might feel more confident that Bernie Madoff deserves punishment than confident in all of your abstract philosophical propositions required to show that he deserves it.

K: I'm sure people may have that "feeling," but it does seem problematic. In a way, the main point of the paper is to reveal that retributivists have more confidence in their punishment of particular offenders than is warranted. Maybe there's some back and forth equilibrium here such that one's confidence in the guilt of particular offenders might somehow bolster confidence in the individual propositions required to show that punishment is just. But my approach is pushing for a kind of consistency that most retributivists have ignored.

C: I see. You make it sound like retributivists are committing the conjunction fallacy.

K: I don't know, but your Madoff example might point to the possibility. I'd put it this way: Suppose someone said, "I realize my chances of picking the right numbers for any one of the six lottery number draws is rather small, but I still think my chance of winning the whole thing is pretty good." You'd have solid ground to challenge that person's reasoning.

C: But you admit that there may be some cases where retributivists actually have enough confidence to punish?

K: I'm not so sure about that. It depends on how one estimates confidence levels in the propositions. If a retributivist had very high confidence in the four general propositions (e.g., 99% each), and the particular case seemed rather easy as to each of the five other propositions (99% confidence each), the retributivist would have a maximum of 92% confidence in the conjunction. Is that sufficiently high to punish consistent with retributivist values? Maybe. But even if it were sufficient, I have my doubts that a reasonable person could have such high levels of confidence in free will and related matters.

C: I see. So maybe when I said that retributivism might be something we're committed to as an idealization of the real world, you didn't completely agree. 

K: Right. There are idealizations and there are idealizations. Depending on how one fills in the probabilities, justified retributivism may simply be too far removed from the land of plausible aspirations. (If you're idealization allows for the elimination of moral risk, though, then sure, my criticism wouldn't apply anymore. So it all depends on what we're allowed to idealize.)

C: What about my claim that you underplay the important retributivist value of giving people the punishment they do deserve?

K: Retributivists strike a balance between the good of punishing those who deserve it and the bad of punishing those who don't. They will, indeed, vary in how they strike that balance, so you're right that nothing is logically entailed about this balance just from retributivism itself. But I speak specifically of those retributivists who subscribe to the values that seem to underlie the Blackstone ratio. And I find it unlikely that retributivists have good grounds for treating errors about facts all that differently than they treat other errors of deservingness. But short answer: My consistency argument against retributivists only applies to those that think it substantially worse to punish those who don't deserve it than to punish those who do (and I think this relative weighting is rather common among traditional retributivists).

C: I see. So have you found our conversations helpful?

K: Absolutely! You haven't felt picked on, have you?

C: Not at all! Besides, growing up with the name ChadF, I'm used to it. Nobody expects my name to have that final "f" sound. And people are always writing "ph" instead of "f," and it's a pain when I give out my name over the phone. 

K: I could imagine. By the way, you said yesterday that you were working on a reply to my paper. How's it going?

C: It's going well. Do you think you'll blog about the replies when you do your Prawfsblawg guest stint? 

K:  Probably not. I worry readers may get sick of the topic if I first blog about it and then post responses to all five replies. We'll see.  But if I do, I hope people know that I'm grateful for our conversations these last couple of days. They've really clarified several matters. And thanks for putting together the group that wrote the replies. It's a treasure trove!    

Posted by Adam Kolber on May 10, 2018 at 08:24 AM | Permalink


Thanks, Chad, for many reasons! I ask the reader to come up with approximate confidence levels for each proposition to try to impress upon the reader how hard it is to be both reasonable in one's approximations and highly confident in the conjunction. Granted, the reader might say, "I can't give you numbers," but presumably we could get ranges of numbers. And even if we can't get that, the reader can consider some possible answers: If the reader is 90% confident in each of the first four propositions, total confidence drops to about 65%, and that's before we even allow for the offender specific propositions, such as whether this person actually committed the crime he's charged with.

Saying "we have some confidence" is not enough. We certainly wouldn't allow a jury to say that. We require the jury--if not to quantify--to at least satisfy a very high standard. Saying "confidence enough to punish in most cases" is hard to argue with, but it assumes the very matter we're trying to determine.

As for striking the right balance of errors, if someone said that the seriousness of failing to punish those who do deserve it is equal to the seriousness of punishing those who don't deserve, then you might only require something like 50% confidence in your justification. And if you're able to fill in the proposition confidence levels in a way that gets to 50% or higher, then the paper has no beef with you. But I don't think most retributivists view those kinds of harms equally. At least, the traditional explanation of the BARD standard seems to treat those harms differently.

So something's got to give: It could be confidence in the values that seem to underlie Blackstone/BARD or it could be retributivist justifications of the punishment of particular offenders. As for whether the goal of getting retributivist confidence levels up is promising, I suspect not. But I encourage you to estimate your own confidence levels in each proposition to see if you can get something viable.

Re: "I think the attack on retribution needs to go deeper, to question why it's a worthwhile goal or a permissible goal for a state to go after." Nothing I say is meant to put to rest the substantive debate about the goals of punishment. But that debate's been going on a long time. So I'm making an effort to put on a new spin on things (as are those like Vilhauer, Hanna, Tomlin, Gross and presumably others who have each examined parts of the equation).

Posted by: Adam Kolber | May 10, 2018 2:30:23 PM

*************Posted on behalf of Chad Flanders******************************

Thanks for this! I now think I see my response as combining with Mary's. We both have worries about what it means to have "confidence" in a judgment; Mary has a more general version, and I want to say our confidence can vary. But I think we both see that retributivists could simply say, "Well, we have some confidence, or confidence enough to punish in most cases, and giving people deserved punishment is a good enough thing that maybe we should do it even if we don't have absolute certainty that our punishing in justified. We can come close enough, by our lights. Nothing in life is all that certain, after all. But we do our best." Can the consequentialist in punishing say she is certain about outcomes? Not really. She does her best.

And this in part is what I was getting at near the end of my reply. If we make the attack you make, the goal of retributivism still stands as a goal we could aim at, and at least try to hit sometimes. I think the attack on retribution needs to go deeper, to question why it's a worthwhile goal or a permissible goal for a state to go after.

Posted by: Adam Kolber | May 10, 2018 2:10:30 PM

The comments to this entry are closed.