Listening to authorities claiming settled science

Scott Aaronson posts Kuperberg’s parable on why we should accept non-expert authority figures on subjects like global warming.

In the parable, the physician cannot explain to the patient the risks of cigarette smoking, and can only say that the science is settled that it is bad. This is not enuf info for the patient, and asks more questions about the actual risk. The physician does not have the expertise to answer the questions, and just wants the patient to do as he is told.

This reveals an authoritarian mindset. I have met people who believe in always doing whatever the physician or other professional recommends. They seem baffled if I say that I like to make my own decisions. The fact is that when a physician makes a recommendation on some subject like eating saturated fats, he is probably not following the settled science.

I realize that there are a lot of people with irrational objections to scientific evidence, and that it may be a waste of time for the physician to learn the details of smoking risks. But he could refer the patient to a reliable source for the patient who wants to be informed.

I follow the hard science, and that is almost never wrong. But the science authorities are often wrong. They join fads and over-hype results.

I see physicists jumping on stupid fads all the time. Climate science is probably ten times worse.

There is hard science showing that CO2 concentrations are increasing, that the increases are attributable to humans burning fossil fuels, and that CO2 absorbs infrared heat. I do not question any of that. But the climate authorities want us to accept all sorts of other things, like the recent USA-China agreement where China makes a non-binding commitment to stick to projections of emission reductions starting in 2030. It seems like just a public relations stunt to me.

There are huge uncertainties in the temperature projections, and considerable doubt about the policy implications. A slightly warmer world might do more good than harm. It is a crazy idea that we should all drive electric cars because some authority figure says the climate science is settled.

Greg Kuperberg explains his parable:

If you look carefully, the patient in effect demanded certitude from the doctor, only to then use it against her. The doctor was willing to provide the general correlation between smoking and fatal illness, and to discuss causal models. The patient brushed that away as well, first by citing that correlation does not imply causation, then by implying that a general theoretical model might not apply to him specifically.

These are all clever debate positions, but what’s behind them is a misinterpretation of the role of uncertainty in science; and a misreading of the purpose of talking to scientists or doctors. The role of scientists is not to win debates — maybe debates with each other sometimes, but not debates with people in general. The role of scientists is to explain and advise to their best abilities. Part of a good explanation is admitting to uncertainties. But then the clever debater can always say, “Aha! Until you know everything, you don’t really know anything.” (As Scott already pointed out.)

No, this is wrongheaded for multiple reasons. First, the physician is not a scientist. He does not do experiments on patients. He is not trained to do or evaluate scientific research. One might occasionally publish his findings, but they nearly always treat patients according to accepted textbook wisdom.

Physicians need to explain and advise, but that is not the role of scientists. Scientists need to demonstrate evidence showing how their hypotheses are superior to the alternatives. And yes, their role includes estimating the uncertainty in their assertions. But the doctor in the parable fails to do all those things.

It really is true that a lot of medical studies show a correlation without causation. And there are population risks that do not apply to particular individuals. Table salt is an example. The advice is commonly given that cutting back on salt will make people healthier, but salt has no known adverse affects for most people.

A central point in Al Gore's movie is a chart showing a correlation between CO2 concentrations and warming. What he does not explain is that the CO2 increase seems to occur 800 years after the warming. Yes, people are right to ask these questions. Aaronson and Kuperberg show their authoritarian mindset when they say that we should just accept official advice without question.

Scott adds:

I think it’s unfortunate that people focus so much on the computer climate models, and in retrospect, it seems clear that climate researchers made a mistake when they decided to focus so heavily on them—it was like announcing, “as soon as these models do a poor job at predicting any small-scale, lower-order phenomenon, you’re free to ignore everything we say.” ...

On reflection, I’d like to amend what I said: building detailed computer models of the climate, and then improving them when they’re wrong, is exactly the right thing to do scientifically. It’s only a mistake politically, to let anyone get the idea that the case for action on climate hinges on the outputs of these models. ...

In my own life, when people talk nonsense about quantum computing, I often feel the urge to give them a complex, multilayered reply that draws heavily on my own recent research, even if a much simpler reply would suffice. After all, that way I get to showcase the relevance of my research! Yet whenever I succumb to that urge, I almost always decide right afterward that it was a mistake.

He is currently writing a book whose main purpose is addressing the quantum computing nonsense that people still had after reading his previous book. I wonder whether he will address my arguments against QC.

I guess he is saying here that he has does interesting research into the computational complexity classes that arise from QC, and he thinks that such research is worthwhile, but it all collapses if QC turns out to be impossible. This is similar to the arguments for string theory -- mainly that lots of very smart people have discovered some fascinating mathematical models, and it would be shame if it had nothing to do with the real world.

It is argument from authority, in spite of all real world evidence being to the contrary.

Update: Here is Scott's latest, giving another example of the leftist-authoritarian mind at work:

But doctors have always gone beyond laying out the implications of various options, to tell and even exhort their patients about what they should do. That’s socially accepted as part of the job of a doctor.

The reason for this, I’d say, is that humans are not well-modeled as decision-theoretic agents with a single utility function that they’re trying to maximize. Rather, humans are afflicted by laziness, akrasia, temptation, and countless things that they “want to want, but don’t want.” Part of the role of the doctor is to try to align the patient’s health choices with his or her long-term preferences, rather than short-term ones.

In the specific case of marijuana, I’d think it reasonable for the doctor to say something like this: “well, marijuana is much safer and less addictive than tobacco, so if you’re going to smoke something, then I’m happy it’s the former. But please consider getting your marijuana through a vaporizer, or mixing it in smoothies or baking it in cakes, rather than smoking it and ingesting carcinogenic ash into your lungs.”

Akrasia is an obscure Greek word meaning "the state of acting against one's better judgment" out of a lack of self-control or weakness of will.

In other words, the people are sheep who need authorities to tell them what is good for them. But such advice should be modified to not conflict with popular leftist causes, like dope-smoking. He also says:

I haven’t studied the IPCC reports enough to say, but my guess is that they massively understate the long-term danger — for example, by cutting of all forecasts at the year 2100.

He also says that he is rejecting policies that cause: "no extra suffering today, but massive suffering 100 years from now that causes the complete extinction of the human race".

The climate experts are only predicting a sea level rise of a couple of feet in the next century, not complete extinction of the human race. So it seems clear that he is not concerned with the hard evidence for what is really going to happen. He wants some like-minded leftist authoritarians to dictate policy to the masses, and have them accept it as progress towards a better world.

Update: Lubos Motl piles on. He frequently criticizes global warming alarmism. (That is, he says there may be some human-induced warming but it is not economically significant.)

However, the doctor wasn't able to offer any numbers to the smoker. I will do it for you momentarily. More importantly, the doctor was a Fachidiot who considers the recommendations she hears from fellow doctors as a holy word – and seems to completely overlook the human dimension of the problem, especially the fact that people have some positive reasons to smoke, reasons she is completely overlooking. (In the same way, people may have very good reasons to veto a surgery or some treatment, too.) She looks at the patient as if she were looking at a collection of tissues and the only goal were to keep these tissues alive for a maximum amount of time. We sometimes hear that doctors are inhuman and look at their patients as if they were inanimate objects; however, it's rarely admitted that the physicians' recommendations "you have to stop A, B, C" are important examples of this inhuman attitude!

This parable does seem to illustrate different thinking styles. I am really surprised that smart people like Kuperberg and Aaronson so aggressively defend such an anti-factual persuasion method.