= 27d567bc2d48d9f40e9ccc20ea370b15)
Eliezer_Yudkowsky comments on The uniquely awful example of theism - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (169)
I think that marijuana legalization in particular can be used as an example we all agree about. (Does anyone reading this actually disagree about that? Leave a comment if so.)
Drug prohibition works rather well in Singapore. With efficient enforcement and minimal collateral damage, a ban on marijuana could fit in a class with bans on cigarettes and alcohol, all of which I could endorse for human societies.
Also, decriminalization (fines instead of felony convictions and jail) could reduce the collateral damage while retaining the norm-setting effects.
Do we have any randomized trials of the effects of weed legalization? I was considering proposing the near zero average health value of marginal medical spending as something nearly as clearly wishful thinking as atheism, but many have complaints about the randomized trial and other studies I point to. Yet the evidence against weed bans seems even less rigorous. My best guess is that we'd be better off without weed bans, but I find it very hard to be very confident of this judgment.
"Modern man is so committed to empirical knowledge, that he sets the standard for evidence higher than either side in his disputes can attain, thus suffering his disputes to be settled by philosophical arguments as to which party must be crushed under the burden of proof."
Thank you, that's the reply I was looking for. We know of a lot of harm caused by marijuana criminalization and very little harm caused by marijuana. Controlled experiments would be very hard to do. In such a state of evidence the sum is unambiguous: revoke the law.
Do we have any data supporting marijuana criminalization? It looks like a simple mistake that ought to be rewound. That summation of the current evidence seems unambiguous, even if we don't know the actual effects of decriminalizing down to the last decimal place - especially considering that decriminalization is in principle reversible after gathering more information.
That summation seems quite ambiguous to me. Only if you presumed that any law without clear supporting evidence should be rescinded could you conclude lack of clear evidence supporting weed bans implies that ban should be rescinded.
I would also take the position that any law without clear supporting evidence should be rescinded, on the basis that all else being equal a simpler legal system would be a better legal system. I don't believe it is necessary to take that position to conclude that drug prohibition should be rescinded however.
I'm actually pretty okay with that principle, at least under common law.
The stronger arguments for me are those that point to the manifest failure of prohibition to achieve its own stated ends. The experience of alcohol prohibition is strongly indicative of the general failure of prohibition. The failure of price controls also seems related - full out prohibition is in some sense a special case of price controls with the government attempting to set the price astronomically high (with jail time being the price the government attempts to set for consumption). Given the lack of evidence for prohibition being an effective strategy, the burden of proof should be on the maintenance of the existing regime. That's even without looking at the clear harm caused by the current approach.
The recent CATO report on the success of decriminalization in Portugal is a recent piece of good supporting evidence for the failure of criminilization. Given the political climate, asking for randomized trials is demanding unrealistically high standards of evidence. Isn't one of the goals of this whole rationality project to be able to make the best possible decisions under uncertainty?
IAWTC but as a general principle is it a good idea to reward people for making laws they don't like costlier to enforce? I'm sure some prohibitions are more harmful than others for no other reason than that the prohibitees behave worse.
Did alcohol prohibition in the US actually fail? We have much lower levels of alcohol use, and thus presumably of alcohol abuse, in the US than in Europe. Perhaps this is the legacy of Prohibition.
If memory serves me, the US actually has a higher rate of alcohol abuse vs. use relative to most other countries. Or possibly it was a higher rate of death/injury/illness due to alcohol.
At any rate, Europeans seem to drink more than Americans do (on average), but do not suffer appreciably higher rates of alcohol-related problems.
Did it succeed? Presumably the explicit intent of prohibition was to eliminate alcohol consumption in the population and the implicit intent was to do so without paying an unreasonably high cost due to intended or unintended consequences. The law clearly failed to eliminate alcohol consumption - it continued to be made, sold and consumed in illegal underground establishments. It also produced unreasonably high social costs (higher than those caused by alcohol when legal) through increased deaths and blindness due to high levels of methanol and through a huge increase in violent organized crime. It's hard to see any evidence of success there.
You seem to be suggesting that perhaps the intent of the ban was also to send an official message of disapproval, and thus influence society to disapprove of alcohol more and so reduce consumption through self limiting behaviours. You also suggest without any evidence that this may have lead to lower levels of alcohol use (only a benefit if you hold the opinion that alcohol consumption is inherently a negative even in the absence of negative side effects) and make a further unsupported leap to suppose that lower levels of use correlate with lower levels of abuse. Even if it is the case that the US has lower levels of alcohol use and abuse than Europe, how do you propose to establish direction of causality from a country with a sufficient tradition of religious disapproval of alcohol to allow a ban on alcohol to pass? Is it not equally plausible that the same social attitudes that made a ban feasible in the first place also explain lower levels of use before and after the ban?
What evidence makes you think the US has lower levels of alcohol abuse than Europe? The US has relatively high rates of alcohol-caused liver disease and has more teenage drinking than many countries in which alcohol is more freely available and partaken of. Contrast the US with the United Kingdom on these two charts:
current national alcohol consumption per capita alcoholic liver disease per capital
Prohibition worked.
It doesn't seem to be an unreasonable assumption that a lack of supporting evidence (of both necessity and effectiveness) should in fact be reason to rescind a law. Do you disagree with this, or are you just looking for hidden assumptions in Eliezer's comment?
If you do disagree I would be genuinely interested to hear your reasoning for it.
The word "clear" is important. The vast majority of laws lack clear evidence for or against them. Perhaps it would be good to rescind all such laws, but the confidence one could reasonably have in such a position isn't remotely like the confidence one could reasonably have in mild atheism.
My confidence in the claim that it would be good to rescind all laws that do not have clear evidence in favour of them stems from my own personal mental analogy of the legal system (something I have a layman's knowledge of) with software engineering (something I have a professional knowledge of).
I see both as a complex system of interacting rules that can often produce unintended consequences (bugs in software). It is a general principle of good software engineering that all else being equal, simpler is better. Duplication of functionality is bad and should be removed. Code that exists for no apparent purpose is often best removed. By analogy, laws that duplicate functionality should be removed and laws that do not seem to have benefits should be removed.
This is clearly an imperfect analogy, and analogical reasoning in general is something to be wary of, but I find it instructive to see the ever-growing complexity and inefficiency of the legal system (where laws are far more often added than removed) with the very common phenomenon in software engineering of systems over time becoming unmaintainable and inefficient due to a tendency to add and never remove code.
The flawed nature of analogical reasoning clearly sets the level of confidence I can derive from this way of thinking well below confidence in atheism but there are other supporting arguments for the position to be made from other directions and overall I have a moderately high level of confidence that the legal system would be improved by a rebalancing of the rate of accretion of new laws and the rate of pruning of old laws.
That's an unreasonably high bar - there is very little I am as confident of as atheism.
Can you name any evidence supporting the necessity of, to pick a moderately troublesome example off the top of my head, copyright? I'm not aware of any alternatives being tried (successfully or otherwise) in modern countries, so there's no actual evidence for its necessity. Shall we abolish governmental protection of intellectual property? That's a somewhat tenable position (donation-based profit, etc.), but I'm guessing most people here don't hold it.
I suspect that if your suggestion's consequences were carefully inspected, it would turn out to be more or less indistinguishable from a very extreme form of libertarianism. I'm aware of no clear evidence that prohibiting civilian possession of assault helicopters and anti-tank missiles is beneficial. Are you? Perhaps they'd be primarily used to resist oppressive governments.
It's also worth observing that plenty of professed rationalists take the exact opposite approach to you. They follow the precautionary principle: ban anything unless we have evidence it's not harmful.
I don't think much of either approach. Suggesting that we should have a hard-and-fast rule of what we have to do in the absence of clear evidence is a bad idea. Humans have capacities for intuition and logic in addition to our capacity to gather empirical evidence. If evidence is lacking, we need to take a best guess, not just say "let's permit/ban it".
You seem to be expecting a much higher standard of evidence than I had in mind. Perhaps necessity was too strong of a word. Utility? Benefit? Something like that.
All I ask is that laws have 1) a clearly defined goal of solving a problem that society wants to solve, and 2) empirical evidence (gathered after the fact, if needed) that they are doing what they were intended to do with acceptable side-effects. Marijuana criminalization seems to badly fail at least the latter, and the former depending on what problem you think it's solving.
The examples you use both have straightforward utility (compensating positive externalities, reducing death rates), and mixed evidence of effectiveness (lots of art created but copyright terms of infinity minus epsilon inhibit building shared culture, misuse of legal firearms suggests more powerful weapons would also get misused for greater potential damage but firearm crime correlates poorly to ownership rates).
Prohibiting by default strikes me as untenable on practical grounds, as well as being morally dubious in the extreme. As an aside, however, I actually would support abolishing intellectual property as weakly superior to the current scheme, but I doubt either is optimal.
How can you gather the evidence after the fact without experimentation? You have to try out alternative copyright schemes, for instance, to test whether it's actually working well. Otherwise I don't know what you'd consider empirical evidence for success.
How can you tell? What would the actual effects of decriminalizing it be? What would widespread marijuana use do to traffic accidents, the intelligence of the general public, etc.? You can argue that it's surely better than alcohol and tobacco, but the obvious counterargument is that those are too entrenched to do away with (especially alcohol) and therefore have to be grandfathered in for pragmatic reasons.
Who's right? Maybe you're right, but the only way to tell is to experiment. I'd be all in favor of more experiments in things like criminal law, to be sure, but I don't think the evidence in favor of a marijuana ban at present is much worse than that in favor of copyright.
As mentioned previously, decriminalization in Portugal serves as a pretty good experiment, and the outcome was much less harm.
I don't think such a trial is possible; laws have to be uniform over large geographic areas to function, so noise from demographic effects drown out the data, and no government would allow its laws to be determined randomly for study purposes, anyways.
I have often thought new laws at the federal level should be treated much like new drugs are treated by the FDA. They should go through rigorous testing according to clear criteria. The proposer of a new law should specify in advance what effect they hope this law would have on the world and how we might measure this effect. Then when the law gets passed, a few small states (or perhaps counties) are initially chosen as the test group - the law only takes effect in those areas. To control for the placebo effect, we should also have a few regions in which the law is announced to take effect but is not, in practice, enforced. Then after ten years we look at the data and see whether the states with the new law are better off than they were before and better off than the states without it - better off specifically according to the previously-specified metric to a statistically significant degree.
Only after it passes that test, is the law extended to the entire nation.
Remember: "if it saves one life, it's worth it."
Not all things can be measured practically, and testing of politically-loaded topics (including most laws that have serious opposition) would be inherently biased. Also, as mentioned upthread, many laws would have no effect or drastically different effect when true at only the local level from the effect if national; for example, dry counties (counties with local prohibition) do not resemble the national-scale effects of Prohibition.
What do you mean by "marijuana legalization"?
Marijuana should be freely obtainable from licensed merchants. It would probably be taxed at a high rate, and there should be laws against operating machinery while under its influence. There would also probably be some licensing mechanism to guarantee its purity.
If possible, all individuals currently in prison for marijuana-related offenses should immediately be released.
The resulting savings in law enforcement resources and prison space, and the accompanying increase in tax revenue would be tremendous. In addition, patients in extreme pain would be able to manage that pain at minimal expense using plants which are easy to cultivate, rather than paying exorbitant amounts to pharmaceuticals.
Biologists would be free to research the effects of THC on the human body, which they are currently forbidden from doing. There have been some indications that THC my have tumor-destroying properties -- not nearly strong enough that I should immediately recommend cancer patients start lighting up as often as possible, but certainly strong enough that if not for the legal issues, follow-up studies would have been performed by now.
That is what I, at least, mean by "marijuana legalization". From my epistemic vantage-point it is an easy win and it is astounding that our government continues to get it so blatantly wrong.
Also, legalizing marijuana would help stabilize the Mexican government. Legalizing opium would cripple Al-Qaeda and help us get out of Afghanistan. Afghanistan's economy depends on opium-growing. In our plans to pull out of Afghanistan, we currently have the preconditions both of building and of destroying their economy.
Note: I don't advocate legalizing opium.
As long as tobacco is legal, anything less harmful should also be legal. (This is more of an argument in favor of banning tobacco than legalizing other substances, though.)
Just curious: Why did at least 2 people vote CronoDAS' comment down?
Presumably because of the implicit assumption that banning harmful things is inherently a good idea. We haven't really discussed that issue because it is not necessary to hold the position that banning people from making choices harmful to themselves is not justified to argue against drug prohibition. Drug prohibition clearly fails on its own terms without needing to convince proponents of the fact that even if it worked it would still be immoral. Many people here would probably take that fact for granted however and so people were presumably voting down the implicit assumption in the statement. They would expect further justification for the claim.
Isn't that pretty much what CronoDAS said, though? The stated premise of banning certain drugs is that they are harmful (either individually or socially). So, again, drug prohibition fails on its own terms, because they are not even choosing the most harmful substances to ban.
A serious attempt to ban harmful substances would start with things like rat poison and cocaine, and work its way down as far as things like tobacco and alcohol and cheesecake, but probably leave out things like psilocybin and THC.
But of course I'm presuming there is a serious attempt to ban harmful substances. In reality there is a serious attempt to ban getting high, which is not the same thing at all.
(Responding to old post.)
A serious attempt to ban harmful substances would not necessarily start with rat poison because while rat poison is more harmful than other substances, it's less harmful relative to how people use it. The fact that eating rat poison causes more damage to you than smoking weed is irrelevant; eating rat poison is an unusual use case for rat poison, but smoking is a typical use case for marijuana.
Well, I would guess that most politicians know perfectly well that the illegality of marijuana is ridiculous, but they can't say so.
I fear you give politicians too much credit. The implication of this is that politicians are more rational on average than voters. I don't see much evidence to support that theory. The facts seem more easily explained by politicians being on average no more rational than the population as a whole.
I suspect that for large numbers of politicians -- probably the maqjority -- the question of whether a proposition is true doesn't really interest them all that much. They are more interested in whether a proposition will win or lose them votes. If they think it'll lose votes, they won't agree with it, and most lack the intellectual curiousity to care whether it is true.
More rational, no. More informed, yes, necessarily; they receive a constant supply of information relevant to their political decisions, largely from lobbyists and think tanks. It is biased information, but so is most of what the general public receives.
I wouldn't say they can't say so - just that they (correctly) estimate that if they came out in favor of legalisation, there are more people who would stop voting for them (not only people who are against legalization, but also those who would subconsciously tag a politician like that as a "weirdo" or "pothead"), then there are people who would start voting for them.
So it's up to us non-politicians to do the job of repeatedly and loudly pointing out that marijuana doesn't deserve harsher treatment than alcohol and tobacco - not because we're morally superior to politicians, but because we aren't subject to the same constraints.
By "can't say so", I mean "can't say so and retain any power", which to a politician's ears is pretty much the same thing.
It's another example of anti-silly bias. Cannabis legalisation has roughly the same support as banning abortion, but anti-abortionists are treated as proper participants in a live debate, but people who oppose drug prohibition are still treated as silly stoners. No-one seriously tries to answer the arguments against these laws; they just fulminate about "irresponsibility" and "our children" and play the silly card.
If the very slow shift of demographics hits a tipping point on this issue, we could see some fairly rapid change.
Yes. Sometime in the 1990s.
Whoops! Didn't happen!
Support for legalizing marijuana is negatively correlated with age. I could not find statistics by age, but I would imagine that older Americans, being generally more conservative, are more likely to oppose abortion. Voter turnout increases with age, so although legalizing marijuana and banning abortion have similar overall levels of support, banning abortion has higher support among people whose opinions politicians care about (voters).
Does it really decline with age, or did older people form their values in a different culture? It's possible people's values are stable over time but people born a long time ago were more likely to form different values from the ones formed by people born more recently. Has anyone tried to distinguish between these possibilities?
http://lesswrong.com/lw/aw6/global_warming_is_a_better_test_of_irrationality/61ff
Pardon my sloppy phrasing. I did not intend to imply causality one way or another, merely the correlation. I edited the original comment to reflect my intent.
Robert Altemeyer reports a correlation between a likely-related attitude (support for 'traditional' authority) and -- not age, but having children. Continued education has a stronger apparent effect in the opposite direction. But I don't think he directly addresses this question.