How Much Certainty Do We Need To Punish?

A Reply to Kolber

Scene: Outside Brooklyn Law School, mid-morning. CHADF, a law professor, reads something on his phone while waiting for an Uber. He is holding a cup of coffee in his other hand. KOLBERT, also a law professor, is walking quickly, deep into editing his latest law review article, mostly oblivious to the outside world. KOLBERT collides with CHADF, causing him to spill coffee all over his shirt.

C: Hey!

K: Oh, hey. Wow. Oops.

C: That’s it?

K: Well, I am in a hurry.

C (recognizing who it is): Kolbert, is that you?

K: Oh, hi ChadF!

C: I was just reading your article on my phone.1 You know, I have to write a response to it, but now this.

K: Yes, this. I did say I was in a hurry.

C: And no apology? No offer to buy me another cup of coffee? To help cleaning this mess up? No nothing?

K: Hey, hold on. Why are you getting so upset?

C: Need I rehearse to you exactly what’s just happened here? You slammed into me, and now I’ve got—thankfully not too hot—coffee all over me. A “sorry” would be nice.

K: Aren’t you assuming a lot?

C: What?

K: I said, I think you’re making a lot of assumptions. About what I owe you, about whether I’m responsible.

C (groans): You sound like your article.

K: And why shouldn’t I? I believed what I wrote, and as I said, you’re making a lot of assumptions.

C: Like what?

K: That I have free will; that I had the power to do otherwise; that I’m really responsible for running into you; that I deserve your resentment.2

C: If not you, then who?

K: Nobody. You know, stuff just happens. This was dictated by the universe—there was, really, nothing I could do about it.

C: And you’re sure about that?

K: Not necessarily. But are you sure about it?

C: I seem pretty sure. I find it hard to believe that you’re not responsible. Hold on, let me cancel my Uber. (Looks down at his phone, and taps a few icons.)

K: But this is only the tip of the iceberg. You must believe I have free will, that I’m responsible, and that I deserve some sort of reproach; that you’re the one who’s entitled to give it to me, that I may even owe you a new cup of coffee. You’ve asserted all those things, and assumed a lot in the process. Your confidence, even arrogance, is astounding.3

C: Is it? I find all of these things coming rather naturally to me.

K: You know as well as I do that it’s no argument that these things “come naturally.”

C: I do wonder, however, who has the burden of proof. It seems to me that you are responsible for running into me and that because of this you owe me at least an apology, and probably another cup of coffee to replace the one I’m now wearing on my shirt. You probably ought to feel bad as well.

K: But again, the assumptions! The certainty with which you must hold them! I would add: the unjustified certainty!

C: That was a problem I had with your article. You ask all these abstract philosophical questions, and I get it—I have some doubts. But then I look at the particular case—take this one, involving you, for instance—and my certainty returns. You are responsible. I have more confidence in my particular judgments than I do in any of your abstract questions and the doubts they mean to sow in me. And I think that counts for something.4

K: But you do have an obligation to reconcile these conflicts—your abstract uncertainty, your concrete certainty.

C: I agree; it’s just that I don’t know that the abstract doubts necessarily win out. For consider: Why shouldn’t the presumption be that you are responsible, absent a good reason to think otherwise?5 I take it you weren’t forced to run into me, right?

K: No.

C: You aren’t insane?

K: No

C: You didn’t act in self-defense?

K: No

C: In fact, you had really no excuse or justification for running into me?

K: No. I was just careless. I wasn’t paying attention.

C: So, I look at that and say, well, why shouldn’t I hold you responsible for what you did?

K: But determinism might be true! We might be all mindless robots, fated to act in all these ways!

C: And I agree that might be the case, but that doesn’t seem to change my sense that I am justified in my reaction and to assume that you are responsible and blameworthy if you dont have an excuse or justification for your actions.6

K: OK, look, fine. Sure. I’m not conceding anything. Let’s get inside and I’ll buy you a cup of coffee. I don’t think we’re not going to solve the free-will determinism problem here. But I think I can run my argument in a different way, on a different level, and I want to try it on you.


They duck into a nearby Starbucks.

C: Alright, what’s the argument?

K: So maybe the whole thing about blaming and holding me responsible for the coffee—I’ll take it because the stakes or so low (although geesh, coffee is sure getting expensive these days). But my article is really about the criminal law and about the standards of justification for that. And in the criminal law, the stakes are really high—we can take away someone’s liberty and even their life.7 That’s why we have the beyond-a-reasonable-doubt standard at trial. My point is that we should apply this to every part of the criminal justice system, whether or not we apply it in our day-to-day interactions. What do you make of that?

C: Hmm, so in our encounter, you’ll allow a lower degree of confidence in things—that you are responsible; that you owe me an apology; that you owe me a coffee—because after all, it’s just a coffee. It’s not your life; you’re not going to prison for carelessly running into me.

K: Right.

C: But with the criminal law, we’d better be really sure—not just about guilt, but about other things too. Did you have free will? Was the law just? Is the punishment deserved and is it proportionate? When we put things into that context, we expect a greater certainty. It’s like when I ask you if the bank is open, and you say, “yeah.” But then I say, “Are you sure? I’ve really got to deposit this check before it closes.”8 And then maybe you’re not so sure—I’ve changed the context so that I really need more certainty than you can give. Your casual confidence just won’t do.

K: Yes.

C: We might even contrast this with the civil part of the law. In a civil trial, we do have a lower standard of proof—preponderance of the evidence—because the stakes can be high, but aren’t as high as in a criminal trial. We’re shuffling money around, not putting people’s liberty at risk.

K: Yeah, sure, that’s interesting. I’d have to think about that. I might want to run a similar argument there as well, but you’ve got me mostly right.

C: But I want to bring up a concrete example again. Remember Bernie Madoff?

K: Of course.

C: He almost went to Brooklyn Law School!9

K: Really? Didn’t know that.

C: Anyway. He’s in prison now for a long time. He pleaded guilty to securities fraud and will probably die behind bars. He cost a lot of people a lot of money, and he (let’s just stipulate this, although I think it’s true) did it out of greed. He wasn’t just careless, or mentally ill; he wanted a kind of lifestyle that required a lot of money and he used crooked means to get it. My sense is that if anyones responsible he is, and he probably deserves not only to pay back the people he bilked out of their life’s savings, he deserves some kind of punishment, some kind of suffering. And so he’s not just responsible, he’s criminally responsible.

K: Look, you can pick a case where you have that certainty, but that doesn’t change my larger point. Maybe there’s a few cases where we have enough confidence in punishment, etc., but on the whole, we don’t have that kind of confidence, or shouldn’t have that kind of confidence, in our criminal justice system.

C: Well, I guess that’s sort of an empirical question, isn’t it?

K: C’mon. You really think we have Madoff-level confidence in all our criminal convictions? That’s absurd. And that’s just as to substantive guilt and punishment. I’m not even talking about procedural rights, which you surely believe in. You think we give everyone a fair trial?

C: Not at all.10

K: So I’m saying, we’re putting people away for a long time, or even for a short time. We’re making them suffer, not just making them feel bad or apologize. So we need to be sure that they are actually responsible, that they actually deserve this punishment, and that they’ve had a fair trial, decent representation, etc. The odds of that being the case, the odds of that being the case beyond a reasonable doubt, seem just, well, impossible.

C: But wait a minute. If we concede that maybe some cases meet that, then I think that does change things. First, we have the question that I’ve just mentioned about whether this is true as an empirical matter, that is, how often we are really super confident that we’ve got the right guy, gave him a fair trial, and punished him the right amount. I agree with you that, man, that’s hard to fathom that this really is regularly the case, although maybe it happens more often than not. But second, as a conceptual matter, it does make it possible to see your argument as an argument about what we should make our criminal justice system like. You’re setting out the conditions for a just retributive system, an ideal. And so we can say: “This is what we’re aiming for, and we’ve got a long way to go.” You haven’t defeated retribution. You’re just saying, “You know if you want to be a retributivist, here’s what your criminal justice system has to look like, and the kind of certainty it’s got to guarantee.”

K: That’s clever.

C: Thank you, but it’s not really my own idea. Jeffrie Murphy made a similar point in his article on Marxism.11 Given material inequality in opportunity, the Marxist can say, it would be unjust to punish people on retributive grounds. It would be unfair. We can’t hold them responsible in the status quo. But that doesn’t diminish the appeal of retribution as an ideal. It just says we must make material conditions a certain way so that retributive punishment becomes justified.

K: But that goal seems so far away. There are so many cases that punishment turns out to be unjustified on my view, where the certainty just isn’t there. We aren’t even close.

C: Again, I agree. The point, again, is conceptual. Are you giving us a reason to reject retribution or just to try that much harder to get there?

K: Hmmm.

C: Anyway, let’s go outside. I’ve got to get to the airport. (Taps on his phone.) My Uber will be here in 7 minutes. And amazingly, it looks like my shirt has dried out.

They head back outside.

K: One last point. Even if you’re right about retribution as an ideal, and I’m not sure you are, you miss that my article is not just against retribution, but also a case for consequentialism.12 Retribution seems to require this really high standard for punishment, and so we just can’t punish unless we meet that standard. So maybe we work to meet that standard as an “ideal.” But maybe we instead look at consequentialism, which sensibly in my view weighs at both the risks of bad punishment, but also the risk of nonpunishment. Retribution seems to leave us paralyzed. Consequentialism can say: “Look, it’s bad to punish wrongly, but sometimes we really have to punish. There are things we lose by not punishing.”

C: Yeah, I had a problem with that too.

K: Really?

C: You seem to acknowledge the retributivist response in your reply to Larry Solum, but you miss the force of the argument.13 You stick retributivists with the idea that it’s better to let people go free than to punish them. I’ll concede this, even though I don’t think it’s entailed by their view.

K: Fine.

C: But saying that we should be really worried about punishing the innocent doesn’t mean the retributivist has to say that there is no cost to not punishing the guilty. In fact, there is a real cost, and that cost is justice not being done! So, there’s more of a symmetry between the consequentialist and the retributivist here than you let on.

K: What do you mean? For consequentialists, there’s a real cost for not punishing people who commit crimes: in public safety or in deterrence. For retributivists, the cost seems mostly on the side of punishing the innocent.

C: Mostly, but not entirely. Consider this recent quote from the wife of a slain police officer in St. Louis when she found out that the city wasn’t going to pursue the death penalty. She disagreed. Let me read it to you; a student in my death penalty seminar just sent it to me. Hold on (CHADF fumbles with his phone and pulls up the e-mail):

‘This is one of those times that you go for the death penalty. You allow a jury to decide instead of just making a decision arbitrarily not to go for it,’ Elizabeth Snyder said. ‘[The suspect’s] going to get three square meals a day. He’s going to get a comfortable bed, a TV, gym time, friends . . . everything that my husband does not get. This killer’s family will one day be able to visit him in prison. I have to visit my husband in a grave.’14

Now, one doesn’t have to agree with this sentiment, and in fact, I don’t. Not only is the death penalty deeply morally problematic, the person who shot the officer was only 19. But leave that to one side. If we believe there is such a thing as justice, then the retributivist will be worried that not punishing can have costs, just as punishing can—the injustice of someone getting away with doing wrong without paying the price for it. There are risks to letting this happen, too, and so they may tip the balance in some cases—to overcome some of our doubts about punishing the innocent. Not always, but sometimes.

K: Is this your Uber?

C: Yeah, it is. Look, nice talking to you. It was fun. The strange thing is, I actually agree with your conclusion, but I think you get there the wrong way. You give the retributivist too much room to make his point.

K: Really?

C: Yeah, I’ll shoot you an e-mail about it when I get home.

K: Cool.

C: And thanks for the coffee, assuming I can hold you responsible for getting it for me and that you deserve my thanks.

K: Very funny.

Later that day, CHADF sends KOLBERT the following e-mail.



Re: Our conversation today, following up

Hey Kolbert,

So I said I was with you in the end, and here’s why:

I think where the retributivist really goes wrong is taking our personal relationships as somehow the basis for our system of criminal justice system. That’s why in our conversation earlier today, it was important that I started out with our little “run-in” as it were. We seem naturally to fall into patterns of praise and blame; of punishment and reward. I think it would be really odd to get rid of these things; and it would be really hard. Can I look at my friends and see them simply as robots? Should I be thinking twice before I get angry at them or express gratitude? It would be hard to live life that way.15 How, for example, would love work? So: You ran into me, I got angry, I expect that you apologize to me. In the regular course of things, you would apologize (as you eventually seemed to do). In a weird way, if I didnt get mad at you, that would be treating you as something less than human—if I treated you just like a child, or like someone who couldn’t control his behavior. I think you would expect me to get mad at you, and would expect yourself to apologize (again, in the ordinary course of things).

The retributivist sees these sorts of interactions as providing a template for the criminal justice system. The retributivist says that when people do something really wrong, we should get angry at them and punish them! That’s the direction our sentiments go, and it’s part of being a human being. And in fact, if we didnt punish them, we might even be doing them (the wrongdoers) a disservice by treating them as less than human.16 Of course, the stakes are higher when we punish people as opposed to simply resent them, so we should hold juries and even the entire criminal justice system to a higher standard. But that doesn’t (the retributivist will go on) mean we shouldn’t shoot for that ideal. Justice is at stake! When we don’t punish the deserving, we in a way harm them, and more importantly, leave an injustice unremedied.

The retributivist sees a continuity between our personal relations and our legal ones, with the legal system in a way mirroring our personal attitudes of praise and blame, and institutionalizing them. Now, what I want to say is that the retributivist is right about our personal lives, but wrong about our legal ones. The legal system shouldn’t be viewed as an institutionalization of our attitudes about praise and blame. The legal system should be about public safety, and in pursuing that goal, we should be worried about things like punishing the innocent, but we should above all be worried about keeping people free from harm. It shouldn’t be about realizing “justice.” I’ve made this argument in other articles, in various ways.17 I won’t bore you with it here. But the key point is that there is a break between responsibility in the personal realm and responsibility in the criminal justice realm.

So what I do with you running into me and what the criminal justice system does with Bernie Madoff are really totally different things because they are about realizing different values. What I am doing with you and the coffee is in the personal realm; it’s about relating to you as a colleague and as a friend and as an object of the emotions of love and hate and resentment and gratitude. With Bernie Madoff, we are really asking about: What do we need to do to deter others from committing this crime? How do we send that message? It’s important that Madoff was guilty, but it’s not important that we worry about metaphysical things like desert and justice, or even about our emotions like hate and resentment. These are things I should worry about when I’m getting mad at you (and you should worry about, too), but with the criminal justice system they are mostly irrelevant.18

I think there are real questions—questions you are surely aware of—when we go consequentialist about criminal justice in this way. It may be hard to make sense of a lot of our procedural protections, or even the beyond-a-reasonable-doubt standard. We may have to work hard to reconstruct these on consequentialist grounds or just, as I have urged, see them as important side constraints in their own right19 (in other words, as part of a “hybrid” view). The trick is not to fall into the trap that the goal of criminal justice is to respond to the personal feelings of injustice, like the ones raised by Snyder, above, that people doubtless feel. In my view, that really is not the state’s job.20

There is more to be said, of course. I’ll leave you with this thought. Retributivists get at some real, and really important, intuitions. I think the retributivist response to your paper may be to agree with most of it, and just say (as I did in our conversation earlier today) that it’s an ideal we have to strive for, even though there is no guarantee we will ever attain it. To them, this doesn’t diminish its force as an ideal, and they may even grudgingly admit that they must supply second best theories for our non-ideal world. But they will hang on to retribution, nonetheless. That’s why I have tried to go in another direction. To acknowledge the force of retributivist intuitions, but to confine them to their appropriate sphere.

That’s all I have now; hope to talk soon.



*  Professor of Law, Saint Louis University School of Law.

1. Adam J. Kolber, Punishment and Moral Risk, 2018 U. Ill. L. Rev. 487. I should emphasize that the words of “Kolbert” represent my best efforts at approximating what Adam Kolber says in his article and what he would say in response to some challenges. Adam Kolber should not be taken to necessarily agree with everything that “Kolbert” says.

2. Id. at 489–90.

3. At one point in his article, Kolber accuses the resolute retributivist as displaying an almost awesome arrogance. See Kolber, supra note 1, at 528

4. See P.F. Strawson, Freedom and Resentment, who makes a similar argument about abstract philosophical questions and concrete cases. Strawson thought it highly unlikely that we could ever be moved to give up our ordinary responses to people—what Strawson called the “reactive attitudes”—in the face of a general philosophical proposition to the effect that “determinism is true.” In this, Strawson was surely following David Hume, who thought that our “animal habits” would be, on the whole, unmoved by philosophical reasoning, and that this was a good thing. For a general defense of Strawson’s approach, see my Strawson and Buddha on the Nature of Persons (manuscript on file with author). A copy of Strawson’s great essay can be found at http://people.brandeis.

5. Charles Larmore suggests this as a general way to think about justification. It is not belief as such, but changes in belief that need justification, and we need a good reason to change our beliefs. See Charles Larmore, The Morals Of Modernity 11 (1996).

6. Again, this borrows from Strawson, who distinguishes two ways of attacking someone’s responsibility for an act—an abstract philosophical way (“I was determined, because determinism is true”) and a concrete, practical way (“I was pushed, I was insane,” etc.). Strawson thinks only the latter way really has purchase given our practices of praise and blame, punishment and reward. See generally Strawson, supra note 4.

7. For this point made in the context of imposing the death penalty, see Sheila Jane Simon, In Sentence of Death, There Can Be No Doubt, Chicago Sun-Times (Dec. 9, 2001) at 26.

8. I take the example from Keith DeRose, in his discussion of contextualist epistemology. Keith DeRose, Contextualism and Knowledge Attributions, 52 Phil & Phenomenological Res. 913, 913 (1992).

9. Diana B. Henriques, The Wizard Of Lies: Bernie Madoff and the Death of Trust 34 (2011).

10. For problems plaguing the public defender system in Missouri, see for example Robert Patrick, Public Defenders in Missouri Say Caseloads Have Them Overworked, and Discipline Has Some Scared, St. Louis Post-Dispatch (Nov. 27, 2012),

11. Jeffrie G. Murphy, Marxism and Retribution, in Why Punish, How Much? 127 (Michael Tonry ed., 2011).

12. See generally Kolber, supra note 1.

13. Id. at 521–22.

14. Andy Banker, Suspect in Blake Synder Killing Won’t Face Death Penalty, Fox2now St. Louis (Dec. 8, 2017, 4:10 PM),

15. See Chad Flanders, Punishment, Liberalism, and Public Reason, 36 Crim. J. Ethics 61 (2017); see also Strawson and Buddha on the Nature of Persons.

16. This gives rise to the idea of people having a possible “right to be punished.” On the “right to be punished,” see John Deigh, On the Right to Be Punished: Some Doubts, 94 Ethics 191, 195–97 (1984).

17. See Flanders, Punishment, Liberalism, and Public Reason, supra note 15; see also Chad Flanders, Public Reason and Public Wrongs, 55 Dialogue 45 (2016); Chad Flanders, Can Retributivism Be Saved?, 2014 B.Y.U. L. Rev. 309 (2014). I’ve developed the idea most fully in Political Philosophy as if Punishment Mattered (unpublished manuscript) (on file with author).

18. I discuss one way in which they might be relevant in Adam Smiths Jurisprudence: Resentment, Punishment, and Justice, in Adam Smith: His Life, His Thought, His Legacy (Ryan Hanley ed., 2016).

19. Flanders, Can Retribuitivism Be Saved?, supra note 17.

20. Id. at 362 (“Indeed, if we appeal to desert to ground our punishment of individuals . . . we are left appealing to our personal convictions—however philosophical they may be—and so in a very real sense we are on our own. We are trying to make the state a vehicle for our own comprehensive doctrine.”)

The full text of this Article is available to download as a PDF.