Ethics of Jury nullification and TDT?
I’ve been sort of banging my head on this issue (I have jury duty next week (first time)).
The obvious possibility is what if I get put on a drug use case? The obvious injustices of the anti-drug laws are well known, and I know of the concept of nullification, but I’m bouncing back and forth as to its validity.
Some of my thoughts on this:
Thought 1: Just decide if they did it or didn’t do it.
Thought 2: But can I ethically bring myself to declare guilty (and thus result in potential serious punishment) someone that really didn’t actually do anything wrong? ie, to support a seriously unjust law?
Thought 3: (and here’s where TDT style issues come in) On the other hand, the algorithm “if jury member, don’t convict if I don’t like a particular law” seems to be in general a potentially really really bad algorithm. (ie, one obvious failure mode for that algorithm would be homophobic juries that refuse to convict on hate crimes against gays)
Thought 4: Generally, those sorts of people tend to not be serious rationalists. Reasoning as if I can expect correlations among our decision algorithms seems questionable.
Thought 5: Really? Really? If I wanted to start making excuses like that, I could probably whenever I feel like construct a reference class for which I am the sole member. Thought 4 style reasoning seems itself to potentially be shaky.
So, basically I’m smart enough to have the above sequence of thoughts, but not smart enough to actually resolve it. What is a rationalist to do? (In other words, any help with untangling my thoughts on this so that I can figure out if I should go by the rule of “nullify if appropriate” or “nullification is bad, period, even if the law in question is hateful” would be greatly appreciated.)
No kidding.
If I were you, I would first concentrate on getting the other jurors to reason correctly about the facts of the case, which is, ahem, enough of a problem.
(Though apparently—and outrageously—one has to be careful here: I know a mathematician who once served on a jury and was accused of juror misconduct for presenting “facts not in evidence” to his fellow jurors, namely “the mathematical theory of probability”!)
With regard to the moral issue of nullification, it seems to me to be quite clearly a quantitative issue of how egregious the law is and how draconian the punishment is likely to be. By no means do I see any general policy of the form “always nullify when you don’t like the law” or “never nullify” as reasonable.
I’ll explain the case against an “always nullify” rule, since I expect that will be more counterintuitive to most folks here.
Getting a society’s rules to be codified into written law represents great progress, in the overall scheme of things. Back in the old days, whether you got to live in peace or thrown into a dungeon and your head cut off was determined by whether the overlords liked you or not—and the rules for pleasing them weren’t written down anywhere, certainly not in any form such that you could appeal to them and say “but you didn’t follow the rules!”
Well, that’s still fundamentally the case, except that we’ve gotten slightly better about it now. There are written laws, and, at least in theory, there are courts where you can go and say “but the government’s position is incoherent!” and have that actually, you know, matter. But much of this formidable-looking edifice of a legal system is still something of a fiction. Laws are still written without the intention of their being taken literally, or even necessarily enforced at all: they’re there so that the authorities have a “legitimate” pretext in case they decide they don’t like someone and want to go after them. And there are so many laws, of such obscurity, that essentially everyone is probably violating a bunch of them, and so at the end of the day we’re at the mercy of the people in power just like we always were. It’s not quite as bad as it was in the old days, but the difference is merely quantitative. We still have further progress to make toward the ideal of explicit rules that will protect us from the arbitrary caprices of The Man.
Now what does this have to do with jury nullification? Well, consider which direction on this axis you’re pushing in by adopting a policy of disregarding the written law and making policy ad-hoc while in a role not legally authorized to be a policy-making role. You may think you’re helping to correct an imperfect set of rules—and perhaps you are—but you’re also at the same time helping to enshrine the notion that the text of the written law doesn’t really matter. If people thought the law actually mattered, they’d be much more careful about which laws they passed. However, as long as the notion persists that laws need not be taken literally, legislatures will continue to pass symbolic laws that will be enforced unsystematically, making everyone—particularly society’s envelope-pushers—continually vulnerable to the whims of prosecutors and the like.
Really? Not a witness, defendant, lawyer, but a juror used their knowledge of probability and that counted as bringing in “facts not in evidence”??? What’s the point of having a jury if the jury members can’t, you know, actually make any use of their knowledge? What’s next, “basic reasoning skills are not part of the actual evidence, so you’re not allowed to think when deciding”? Excuse me, I have to go foam at the mouth now. :p
Anyways, as far as the rest, that essentially is part of my concern. (Though you detailed it much more completely than my own thoughts did, but fuzzy thinking along those lines was part of my concern)
Hence my quandary: If presented with a drug use case, should I nullify if defendant is “guilty”?
It seems to me that it might be possible to extract some of these intuitions without all of them.
For example, say that the “rules as written down” are something along the lines of “we get a group of your peers to determine whether you’ve done something wrong, and if you have what it is. Then we decide based on precedent what your sentence is.”
In that case, members of the jury are capable of deciding that defendants have not done anything wrong, even if there is a specific punishment listed for what the defendant did.
I don’t mean to argue that this is what is the case, for example I don’t believe that the rules I described are a good representation of US criminal law. But that is a fact about the specific system you are involved in; I don’t believe that it is a priori problematic to have “jury nullification” in the sense of jurors acquitting persons who committed “crimes.”
Beautiful. Very well said. I expressed similar intuitions, but you make the point much better. (Not that our arguments are in perfect agreement, but we’re both appealing to implications of nullification for the rule of law.)
People being convicted of violating unjust laws is a much, much larger problem than people being improperly acquitted. Both because one improper conviction is worse than one improper acquittal, and because I don’t actually think improper acquittals really happen much. Opposing jury nullification seems to me like a status move to maintain the power of the lawyers, at the expense of justice.
So if you’re on the jury for a case prosecuted under an unjust law, then yes, you do have a moral duty to acquit. Whether you should announce that it’s a jury nullification, or lie about how strong the evidence looks to you so others don’t get the same idea, is less clear. I lean towards the former.
Yes, a conviction for an unjust law is bad—that’s not in dispute. The problem is whether you are appropriately fighting this injustice by nullifying, if you reason by TDT, a decision theory that ranks very well on numerous desiderata that account for these intuitions. Appealing to specific moral duty doesn’t resolve the problem, for the same reason that appealing to greed doesn’t justify two-boxing on Newcomb’s problem.
For my part, I do have a big problem with drug laws (despite not planning to use them). And before thinking about this as a rationalist, I did favor jury nullification. But if I put my TDT hat on, here are the problems I see with it:
Like what loqi said, if jurors tended to nullify, then the system, anticipating this, would either not use juries, or resort to extremely stringent screening methods to make sure there are no nullifiers (potentially involving massive violations of privacy for jurors, which in turn has to make them very circumspect about talking about their political views… even today people complain about the intrusiveness of questions jurors get asked). Or reduce the conviction difficulty threshold so that they can allow a conviction to go through if jurors acquit “for a bad reason”.
(I think this is similar to the old debate about whether you should cut down the law to get to the devil—or Hitler, etc. Or whether a judge in a slavery-supporting society who “sees the light” should suddenly start ignoring the laws and make whatever decisions go against slavery.)
TDT asks you to consider the consequences, conditioning on yourself (and all similar instantiations of your algorithm) being the kind of person who would keep quiet about your opinion on nullification and the particular law, and then nullify if you didn’t like the law.
This seems to be equivalent to setting humans to take the policy of “using any discretion I have to bend application of laws in the direction of the legal regime I prefer”. But if humans deemed this optimal, it would not be possible to have a “rule of law” system, where there are definite laws, and people can know when they’re breaking them, and which disallows favoritism. It would probably not be possible to implement any law except those which are extremely popular (which may be a good thing).
So I’m forced to conclude that jury nullification is in the unfortunate position of “feeding off of its own existence”, like stealing or two-boxing. While it may appear that it’s a chance to shift utility toward people you like, your deciding to do so has broader implications.
So at the very least you should say upfront that you regard it as optimal to nullify unjust laws, which will get you tossed out of the pool but otherwise unhurt. (I’ve heard that if you even exhibit familiarity with related keywords, that’s enough to get dismissed.)
Thanks, though part of my question is, due in part to the sorts of issues you bring up… should I consider it optimal for me to run the algorithm of nullify laws that seem unjust to me?
The consequent here is independent of the antecedent—I don’t think the system you describe is possible under either circumstance.
I’m not seeing how this follows without some additional value judgments. You’re basically saying “Widespread nullification would fuck up the legal system, so don’t do it”, instead of ”… so beware of the trade-offs involved”.
Not perfectly, no, but any decent approximation has the norm I described (that you shouldn’t use your discretion to favor lawbreakers simply because that would make the law closer to what you personally desire) as a pre-requisite—I don’t see how it would be otherwise.
I thought that’s what what I was doing with:
I don’t have the link handy, but I recall reading that judges are completely within their rights to dismiss a juror who is openly planning to nullify. If this is true, then honesty is not a viable strategy.
I was sent home twice for that very reason. They weren’t rude about it; they just sent me home once I said that I won’t convict someone of a “crime” that didn’t actually harm anyone or their property.
Surely posting about it here makes it a matter of ordinary repeated games, TDT or not.
Now there’s an interesting thought. :)
Though, realistically, given the fraction of the population that would see this and to what extent anyone here could track how each of us decides and our reasons for doing so when we’re jury members, probably not. However, reducing from the sorta acausal abstractions of TDT to simple iterated games may at least help to think about it, may help provide some useful intuition pumps at least.
So, given that we’re now in an iterated game with respect to nullification style issues… What’s the right answer? (Actually, legal system stuff is complicated, so even figuring out how to partition the players, not to mention the payoff matrix, is going to be tricky.)
Thought 3 is where you make a mistake. You’re not choosing for jury members who won’t convict on hate crimes. You’re only choosing for future jury members who have a similar choosing alorithm to yours. The haters obviously don’t. Thus the only thing you need to ask yourself is: Do I want future jury members with similar ethics to mine to convict on drug laws or not? And that’s your desicion.
Edit: But as Unnamed wrote, the difficult choice is what to say during jury selection.
But, the algorithm itself isn’t the same as the values I input into it.
What I mean by that is that I’m going “I think drug laws are unjust… THEREFORE I should refuse to convict in the case of a violation of those laws”
So is the “If I, personally, think a set of laws are seriously unjust, should I essentially ignore them when I’m a jury member?”
(note, my natural initial inclination was “yes, nullification of drug laws is a good idea”, with the only question being “do I lie when asked in the first place if I’d be willing to convict on a drug offense or if I’d heard of nullification” etc… But once I got to thinking about it, to the actual algorithm I’m effectively running, well, hence my eventual asking here for advice)
Another question is what you should say during jury selection. What if they ask your opinion about drug laws, or if you will be able to faithfully & impartially apply those laws regardless of your opinion?
Well, first question is “should I consider nullification type stuff a good idea at all?”
Second, is if I consider the answer to be yes, should I be willing to lie about it? That seems even more shaky. I’m thinking ethical injunctions here and such. (ie, “don’t go lying about relevant stuff in the context of a court/justice system (at least one that is vaguely reasonably fair) ”
So yeah, that’s a second question.
Minor update: called up this evening to check and they said that I don’t have to go after all.
Humans don’t run TDT. Do the right thing for the case you are presented with.
And “the Right Thing” being “respect the principle of the rule of law even if I find the law in question repugnant” or “respect the principles of justice as best as I individually determine them ignoring any considerations of the value of rule of law”?
(Note: my natural inclination is to favor nullification of the drug laws. But once I got thinking about it, well, hence my asking my fellow aspiring rationalists for advice)
Because you are asking for concrete advice about a real life problem, I am going address your question in the somewhat convenient world we happen to live in, and note that the Supreme Court has upheld jury nullification, and so if you decide to acquit because the law is invalid, you are acting within the rule of law.
I am trying to stay away from mind killing territory as much as possible, and so won’t comment on the less convenient world where jury nullification is explicitly illegal.
There have been more recent rulings, from what I understand, that only sorta upheld it. (including stuff along the lines of how juries have the “power” but not the “right”, which best as I can figure, means they have the de facto power, nothing can be done about it, but they shouldn’t do it.)
Anyways, I wasn’t so much asking about the legality of it itself as much as if it was the Right Thing to do, taking into account the various other considerations of the importance of rule of law/predictability of the court system, TDT style considerations of the relevant algorithm I’d be effectively running, etc.
Thank you, though.
What if TDT is optimally ethical?
You know, its major selling point to begin with?
TDT is very effective for agents with the computational power to run it on top of a powerful epistemology that accurately notices other TDT-like agents whose decisions will correlate with their own.
This does not describe humans on juries worried about the decisions of other unspecified humans on other juries.
Every formal, well-justified decision theory is that way. That’s no reason to avoid tractable approximations of them.
If you have no problem basing your decision on “greater good” (utilitarian) considerations, or Golden Rule/Categorical Imperative type reasoning, all of which are computationally intractable and do not perfectly describe humans, then you shouldn’t make this objection to TDT either.
I am not saying to avoid tractable approximations. But it is not likely that other people on other juries are using or know about TDT or any tractable approximations.
TDT’s optimality does not hinge on whether others use it; it’s just that TDT can handle other TDTers as a special case. It is still optimal for interacting with non-TDTers, so this doesn’t show why he shouldn’t be trying to approximate it here.
Do you think that any reasonable approximation of TDT would say anything other than ignore those unspecified other people who are not TDT-like and may as well be pieces of cardboard with whatever decision they will make written on them, and just do the right thing for the case you are presented with?
Wait—so you think tractable approximations of TDT will still lead to the right conclusion, but that Psy-Kosh still shouldn’t use it, even for practice?
In any case, yes, at the appropriate level of granularity, there are relevant correlations between yours and others decision processes—specifically, at the level of whether you deem it optimal to use discretion to bend application of laws in ways that favor your values. See komponisto’s and my long comments on this discussion for more detail.
But FWIW, yes, I do find this reasoning from acausal correlations (that TDT relies on) to be problematic and worthy of more scrutiny.
No. I advised him of the output of me running my tractable approximation of TDT: focus on the right thing for the presented case.
This is incorrect. Suppose you and other humans start out using variations CHDT (Crazy Human Decision Theory), which have some level of correlation with each other, but do not take this correlation into account when making decisions. Then you learn about TDT, think it is a good idea, and incorporate some of its ideas into some variation THDT (Timeless Human Decision Theory), which ends up being something like run CHDT but override some decisions if principles derived from TDT make it seem like a good idea. So you look at this jury nullification problem, and start running CHDT which returns that it seems like a good policy. But TDT says that if your decisions with others are correlated with other agents, you should take that into account. And you don’t like what you predict would happen if everyone embraced jury nullification, so you change your decision. At this point you have destroyed the correlation between your decision and the normal CHDT using humans because you overrode the component of your decision theory that works the same way.
This could work if the other humans are also using THDT, because then they could override their CHDT answers for the same reasons you do. But we are talking about juries drawn from a population of mostly non-rationalists who have never heard of TDT. The decisions you make differently because you know about TDT do not correlate with their decisions.
Most of the interesting effects of TDT do seem to relate to the case of other agents also using TDT, afaik. (Although there’s that other post about the TDT paradox, of course...)
Jury nullification is an act of civil disobedience. If you want to strongly signal your opinion about drug laws, I think that’s a great statement to make, but jury nullification is practiced so incredibly rarely that it ends up being more of an emotional than rational decision.