“Why presuming innocence has nothing to do with assigning low prior probabilities to the proposition that defendant didn’t commit the crime”
by Professor Larry Laudan
Philosopher of Science*
Several of the comments to the July 17 post about the presumption of innocence suppose that jurors are asked to believe, at the outset of a trial, that the defendant did not commit the crime and that they can legitimately convict him if and only if they are eventually persuaded that it is highly likely (pursuant to the prevailing standard of proof) that he did in fact commit it. Failing that, they must find him not guilty. Many contributors here are conjecturing how confident jurors should be at the outset about defendant’s material innocence.
That is a natural enough Bayesian way of formulating the issue but I think it drastically misstates what the presumption of innocence amounts to. In my view, the presumption is not (or at least should not be) an instruction about whether jurors believe defendant did or did not commit the crime. It is, rather, an instruction about their probative attitudes.
There are three reasons for thinking this:
a). asking a juror to begin a trial believing that defendant did not commit a crime requires a doxastic act that is probably outside the jurors’ control. It would involve asking jurors to strongly believe an empirical assertion for which they have no evidence whatsoever. It is wholly unclear that any of us has the ability to talk ourselves into resolutely believing x if we have no empirical grounds for asserting x. By contrast, asking juries to believe that they have seen as yet no proof of defendant’s guilt is an easy belief to acquiesce in since it is obviously true.
b). asking jurors to believe that defendant did not commit the crime seems a rather strange and gratuitous request to make since at no point in the trial will jurors be asked to make a judgment whether defendant is materially innocent. The key decision they must make at the end of the trial does not require a determination of factual innocence. On the contrary, jurors must make a probative judgment: has it been proved beyond a reasonable doubt that defendant committed the crime? If they believe that the proof standard has been satisfied, they issue a verdict of guilty. If not, they acquit him. It is crucial to grasp that an acquittal entails nothing about whether defendant committed the crime, What it focuses on is how strong or weak is the proof that he did so. Because their verdict decision is entirely a question about whether guilt has been proven or not, the guilt-not-proven verdict leaves wholly unresolved the issue whether the defendant did or did not commit the crime. Boastful claims to the press from defense attorneys about how their newly acquitted clients have been ‘exonerated’ or ‘vindicated’ are patently misleading. What they should be proclaiming on the courthouse steps is something like: “There’s at least a 5-10% chance that my client didn’t commit the crime.” (Except in Scotland, ‘innocence’ simply does not figure among the verdict options open to Anglo-Saxon jurors.)
c). Legal jurisprudence itself makes clear that the presumption of innocence must be glossed in probatory terms. Consider this model federal jury instruction:
The law presumes defendant to be innocent of all the charges against him. I therefore instruct you that the defendant is to be presumed by you to be innocent throughout your deliberations until such time, if ever, you as a jury are satisfied that the government has proven him guilty beyond a reasonable doubt.” US v. Walker (1988)
Bayesians will of course be understandably appalled at the suggestion here that, as the jury comes to see and consider more and more evidence, they must continue assuming that defendant did not commit the crime until they make a quantum leap and suddenly decide that his guilt has been proven to a very high standard. This instruction makes sense if and only if we suppose that the court is not referring to belief in the likelihood of material innocence (which will presumably gradually decline with the accumulation of more and more inculpatory evidence) but rather to a belief that guilt has been proved.
As I see it, the presumption of innocence is nothing more than an instruction to jurors to avoid factoring into their calculations the fact that he is on trial because some people in the legal system believe him to be guilty. Such an instruction may be reasonable or not (after all, roughly 80% of those who go to trial are convicted and, given what we know about false conviction rates, that clearly means that the majority of defendants are guilty). But I’m quite prepared to have jurors urged to ignore what they know about conviction rates at trial and simply go into a trial acknowledging that, to date, they have seen no proof of defendant’s culpability.
*Currently a Professor of Philosophy & Law, The University of Texas School of Law
Among Laudan’s books:
1977. Progress and its Problems: Towards a Theory of Scientific Growth
1981. Science and Hypothesis
1984. Science and Values
1990. Science and Relativism: Dialogues on the Philosophy of Science
1996. Beyond Positivism and Relativism
2006. Truth, Error and Criminal Law: An Essay in Legal Epistemology
Larry: Thanks so much for this post! It greatly illuminates what I tried to say in my earlier comments on Schachtman, but more than that, it may suggest a better way to state a general point that I am constantly trying to convey, with mixed success; namely that assessing probativeness differs from assessing probabilities. The former is a matter of how good a job was done. The ‘rating’, if there is to be such a thing, has entirely to do with what was shown, how effective the performance was at ruling out the denial of a claim H (in this case, innocence). [One might separately ask about the warrant for H, above and beyond what was or wasn’t shown by a given body of evidence.]
However, it isn’t so clear to me how to relate concepts from statistical tests to your remark that “There’s at least a 5-10% chance that my client didn’t commit the crime.” Perhaps something like this: the evidence presented, E1, E2,…En is not so inconsistent with innocence as to be practically impossible, under the assumption of innocence. Or: the set of coincidences that would have to be the case in order to have amassed the evidence they’ve presented, when in fact the person is innocent, is not so very tiny (not less than .05 say). A different kind of claim, which I don’t think you’d mean, is: at least 5% of the time, when this is the best evidence of guilt they can muster, the person is innocent. Or perhaps, “if we continually deemed persons innocent on the basis of this evidence, or evidence no better than this, we’d be wrong about innocence less than 5% of the time.
Do any of these fit?
The standard thinking among legal scholars which generated the words I put in the mouth of my archetypal lawyer is this: a). exoneration studies (of which there is now an abundance) reveal that there is a false conviction rate at trial of ~5%; b). that, in turn, is routinely construed as indicating that the de facto standard of proof is about 95% likelihood of guilt; c). leading to the supposition that a defendant will/should be acquitted whenever his apparent guilt is <95%. I don't endorse this reasoning but I believe it represents how a typical lawyer would approach the issue.
Larry: First of all, sorry your comment didn’t show up til now: you hadn’t commented before; but now your comments will show up immediately.
I’m wondering how the false conviction rate would actually be known, in general–aside from those cases garnering sufficient interest to reexamine (I will look into exoneration studies). But, in any event, that assessment would differ, would it not, from the criteria actually applied by a juror in summing up the evidence (to judge if guilt has been “proven” to the required standard). Nothing you’ve said is at odds with that in the least; in that sentence, you were describing what a defense lawyer could reasonably say,as opposed to the juror assessment of not guilty.
Reblogged this on Not Knowing Things and commented:
A philosophy of science approach to bayesian inference required by the justice system.
Reblogged this on Епанечников блог and commented:
“…the presumption (of innocence) is not (or at least should not be) an instruction about whether jurors believe defendant did or did not commit the crime. It is, rather, an instruction about their probative attitudes.”
I don’t see any incompatibility with what my Texas colleague Prof. Laudan writes and with what I wrote. The Bayesian prior of probable innocence and then the evidence that will be used to change that prior into a posterior is precisely to determine what the probative attitude should be when all the evidence is taken into account. The evidence that caused the authorities to arrest and then indict the person on trial is precisely what would be used by a jury that started from a prior that assumed probable innocence to its final posterior. A 1/N prior where N is the population is appropriate when you know nothing of the evidence (e.g., the police find a dead person whose identity has not been made known at that point). But this will quickly be changed towards guilt by nothing more than the fact that crimes require motive and motive usually resides with a much smaller population, e.g., people known to the victim, people who have already been in trouble with the law, and so forth. A 1/N prior will not survive for long in the face of such considerations.
The problem I see is that the jury (trained as we all are in Bayesian thinking 🙂 has to avoid using the same evidence twice. But a jury that started with a belief that the accused is more likely guilty than innocent simply because of an indictment, and then is presented with evidence that brought the probability of guilt to greater than 50% in the eyes of the authorities (as will certainly happen) cannot use that evidence again to further raise the probability of guilt, for then they would be violating the rule that P(H|E,E)=P(H|E).
Maybe there are ways around this problem, but the problem is there unless specific steps are taken in the process to prevent this rule from being violated.
BTW, some of the comments on the previous thread were off-point. Prof. Schachtman specifically asked about the Bayesian situation, and comments to the effect that some frequentist approach is better, for example, do not answer his question.