Looking at the comment thread here, it seems to me that many commenters are unaware, or at least fail to remember, that in the U.S. and Canada, and presumably many other Anglospheric common law jurisdictions, well over 90% of criminal cases are these days resolved by plea-bargaining and never reach trial.

This is an essential feature of how the criminal law system works today -- the whole infrastructure of criminal courts would be overwhelmed to the point of collapse if it actually had to provide jury trials for more than a tiny minority of all defendants who get convicted. In the contemporary socially atomized and thoroughly bureaucratized society, in which law has become complex, vast, and abstruse to the point where it's completely outside the intellectual grasp of the common person, trial by jury is little more than an ancient historical relic. Focusing on juries as a central issue of the modern criminal law makes little more sense than focusing on the monarchy as a central factor in the modern U.K. government and politics. Of course, jury trials provide good material for movies and TV shows, and like in many other things, folks who haven't had any personal experience with the ... (read more)

Before discussing justice from the standpoint of rationality, we need to agree on the goals of justice. Are they to prevent crimes, to punish those convicted, to provide relief to the victims (monetary, by injunction, or by feeling vengeance has been done), to ensure laws are implemented, or to create other second-order effects on society by being seen to enforce moral standards? And how should these goals be balanced?

I can think of a few things that I might change about how trials are conducted in the United States.

Suggestion 1: Don't have the jury in the courtroom as the trial is taking place. Instead, make video recordings of all testimony and show them to the jury after each witness's testimony is concluded. And if a lawyer or witness makes a statement that the jury is not supposed to hear, it can be removed from the tape instead of the jury hearing it then being asked to ignore it. Also, allow the jury to review the tapes during deliberations, so they don't have to rely on memory alone.

Suggestion 2:

Allow jurors to question witnesses, but only indirectly. After the attorneys are finished questioning a witness, ask the jurors to submit in writing any questions they would like to have the witness answer, have the attorneys from both sides read the questions, and then give both attorneys the opportunity to question the witness further. If a question is relevant and legal to ask, it will likely be in the interest of one of the sides to ask it, and using the attorneys as intermediaries seems like it would make juror-directed questioning less likely to lead to problems with inappropriate questions. (I don't know if this would actually work or not, though.)

Here's a bit of information on juror-submitted questions.

As a related observation, one apparent bias that reflects the status considerations in the modern Anglospheric world in an interesting way is that while drunk driving has come to be considered as a heinous crime, the even more dangerous practice of driving sleep-deprived attracts no significant attention. It's not even that much of an enforcement problem -- if "sleepy driving" became as much of a public attention-grabbing buzzword as "drunk driving," I'm sure our best technical minds would be up to the task of devising effective field tests for sleep deprivation, to be administered along with breathalysers.

I think you can get some useful insights into the reasons why punishments might differ based on moral luck if you take an ex ante rather than an ex post view. I.e. consider what effect the punishment has in expectation at the time that Alice and Yelena are deciding whether to drive home drunk or not, and how recklessly to drive if they do.

Absent an extremely large and pervasive surveillence system, most incidences of drunk driving will go undetected. In order to acheive optimal deterence of drunk driving, those that do get caught have to be punished much more. While drunk drivers will face different punishments ex post, the expected punishment they face ex ante will be the same. If there are in fact factors that make your drunk driving less dangerous (less drunk, more skilled driver, slower speed, etc.), these will decrease the expected punishment.

So basically, the ex ante expected punishment for a particular dangerous act does not differ based on moral luck. Ex post punishment does differ, and that is a cost, but the countervailing benefit of not having a costly and intrusive surveillence system outweighs it I think.

Notes:

1. This is the second time I've linked this recently, but

Trial by Jury; Trial by Judge

Juries were originally taken to know something about the relevant events. The modern form of jury trials is a weird hybrid of inherited practices and contemporary political ideals. Like most legal phenomena :-)

That modern juries are inexperienced in criminal matters could be a positive feature. Judges may be jaded by constant exposure to narratives of crime. In wealthy countries, serious crimes are exceptional events. Jurors have reason to pay attention to narratives of such events.

Further, legal experts worry about naive juries being swayed by irrational considerations. This might draw expert attention to cognitive biases, and result in institutional steps to avoid them. Legal experts may be less willing to concede that they themselves are biased. Eliminating juries might mean less expert consideration of hazards to good reasoning in criminal trials.

I think it might be useful to focus on how juries decide cases. For example, juries could be told to hold off on proposing solutions. The jury's group dynamics could be tweaked to promote rational discussion, perhaps by allocating jurors specific roles in the discussion (eg "your job is to tell the ... (read more)

Regarding the determination of punishments, fans of game theory should find interesting the system of classical Athens. Here's a description from a lecture transcript:

If a penalty was called for, and it was not one that was described by law (and very few penalties were described by law), the following procedure was used: the plaintiff who had won the case proposed a penalty, [and] the defendant then had the opportunity to propose a different penalty. The jury then -- again no deliberation -- just voted to choose one or the other, but they could not propo

I'm currently researching the rationality of criminal trials, though my focus is on evidence law. The current adversarial system does have some advantages:

• An adversary system creates incentives to bring information before the court. Parties want to win. In particular, the defendant has a direct interest in avoiding punishment. Defendants can help themselves by pointing out flaws in the prosecution case. In general, an adversary system rewards parties who put arguments and evidence before the court.

• However, this can include bad arguments and emotive but

Ceteris paribus, I would think that the lower gullibility of judges would be entirely overwhelmed by the effects of increased corruption. Take the corrupt judges in Pennsylvania that were all over the news last year, for example. The difference in accuracy between a jury and a judge pales in comparison to this sort of thing; that's fine if corruption is proportionally more rare than that accuracy gap, which is probably true if most cases of corruption are uncovered.

But if you look at the story of those Pennsylvanian judges, they did a miserably bad job of ... (read more)

I was in an exchange here a while back about this question. Komponisto and I basically agree that a trial, whether the guilt is decided by a judge or jury, should work like this:

• For a conviction, they (jury or judge) must find that the odds of guilt meet some specific, high threshold, such as 100:1 odds.

• They should be allowed to incorporate any relevant Bayesian evidence (such as "probability that a random person committed this crime")

• But there should be a list of exceptions to the above for cases where cognitive biases significantly disto

You don't get a lesser drunk driving charge if you can prove you're pretty good at driving drunk.

Uh... what? You do get a lesser charge if you avoid killing anybody. What's more, you actually called attention to that fact before, and are trying to argue that it's morally wrong! So, basically, one of your arguments against X relies on an incorrect assertion that not-X. I wanted to call this "begging the question", but it's probably the wrong nomenclature; the pattern of argument is just extremely weird.

In Japan, the conviction rate is about 99%. (Many criminal defense attorneys go their entire career without ever having a client being found innocent.) This is widely attributed to the facts that Japan uses bench trials and that prosecutors are expected not to bring cases to trial unless they are certain to win.

There's one easy step I'd add for jury selection that would likely go a long way to helping. Add in tests to see if potential jurors are more likely to convict someone who is ugly. And have those people thrown out of the jury pool (part of me wants to have them shot but obviously that's not actually a good result). There's a lot of evidence that many jurors are more likely to convict people they find unattractive. See this article (I don't unfortunately have access to the study in question at the moment). The good news is that the evidence suggests that th... (read more)

Suppose that Alice and Yelena, on opposite ends of town, drive home drunk from >the bar, and both dazedly speed through a red light, unaware of their surroundings. >Yelena gets through nonetheless, but Alice hits a young pedestrian, killing him >instantly. Alice is liable to be tried for manslaughter or some similar charge; >Yelena, if she is caught, will only receive the drunk driving charge and lose her >license.

This is a fairly classic problem, well-presented. There's a credible argument to be made for treating them the same, but I don... (read more)

How would the justice system look if LessWrong had to rewrite it from scratch?

Probably much worse. Too many here fit the description in Hayek's last book The Fatal Conceit.

How would the justice system look if LessWrong had to rewrite it from scratch?

I'm not really sure that we would change it much. The justice system looks relatively good to me, compared to the paucity of our other institutions.

Fun crazy ideas that come to mind:

1) Punishments get scaled by the judged likelihood of guilt, i.e. judge says there's a 65% chance Bill is the killer, Bill gets 65% of the punishment.

2) All punishments become monetary fines varying by judged negative utility, i.e. Judge says murdering Joe was worth x negative utilons, Bill is fined to outweigh damage done with good.

Potential problems/thoughts: Bankruptcy? Lower bound on fines/guilt likelihood? Diminishing percieved utility of money/punishment in large amounts? How to measure negative utility of crime, p... (read more)

One common theme I see in the other comments is that the justice system is just way too slow and underequipped to handle every case properly. It might be fruitful to focus on some way of making the justice system more efficient while maintaining roughly the same output as before.

I'm not a lawyer, but one possibility occurs to me: how about a time limit on each section of a trial? Similar to some debate formats: prosecution gets X minutes to introduce the case, defense gets Y minutes to respond, and so on, up to a limited number of segments which can only b... (read more)