21 August 2010

A Strong Criticism of Monarchy

From the comments to the World Cup piece:

Monarchy is rule by a single individual. It works on this wise. Immediately after his succession, the new monarch enthusiastically attempts to rule the country. For a certain period, shall we call it a year. As there is only so much time between breakfast and supper, this is largely impossible. The next year, he carries on out of a sense of duty. The third year, he announces that he does not want to be bothered with this ruling crap, but if there are any fit women around would you please send them up. Monarchy then gives way to pornocracy: porne is Greek for prostitute.

Previous objections to Monarchic rule which I have rejected rest on the possibility that the monarch might be an idiot or a psychopath. In my estimation, mere idiocy or psychopathy are less damaging to good government than politics is.

The commenter brings up the more fearsome possibility that the monarch could be a normal sane bloke, more concerned about what his girlfriend thinks of him than about whether GDP next year will be 2.5 trillion or 3 trillion.

If decisions simply end up being made by some random attractive woman (or boy) instead of the hereditary ruler, that's not a problem in itself. But the reason this situation is so much more dangerous than mere insanity is that it produces politics, (meaning a struggle for power), based this time not around arming supporters or controlling journalists, but around forming close personal relationships with the monarch. This was often the main form of "politics" in historical monarchies.

I'm not sure that it is a worse form of politics than exists in a democracy or a military Junta, but my aim in proposing monarchy is to remove politics altogether, which is obviously more difficult than I thought.

11 August 2010

Devaluation of Significance

I referred in my last post to a lost writing of mine on the subject of abuse of statistics in economics. I've sort of found it - I sent it as an email in response to this blog post by Noel Campbell at Division of Labour. (Read it - it's short).

He quoted from my response, but I can't find the actual email I sent him. I do have a draft of it, so it would have been very much like this:

That's a superb question, and I think the answer will surprise (and disturb) many.

Your paper will include a calculation of significance. This is essentially an estimate of the probability that a correlation as strong as the one you found would exist purely as a result of randomness in the data, even if your theory is false.

This calculation assumes the "proper" sequence of events. You have a theory, and you test the data for a correlation. Since you in fact poked around for correlations, then came up with a theory, the significance calculation is not valid. The true significance depends on the probability that, having found a randomly-caused correlation somewhere, you can then invent a theory to explain it. That probability is very difficult to estimate, but is probably much greater - meaning that the significance of the correlation is much smaller.

It is very counterintuitive that the order of your actions affects the validity of your findings, and indeed it is a close relative of the famous Monty Hall problem - the poster child for counterintuitive probability. When you reveal the correlation that you already knew of, you are revealing no information about the chance of your theory being correct, much as when the quizmaster opens the door that he already knows doesn't have the car, he reveals no information about the chance that the door you first picked has the car. Conversely if you pick a door and find that it doesn't have the car, that does change the probability that the first door had it, and if you had no prior knowledge of the data, the correlation does change the probability of your theory being true.

Back to science. As you say, theories aren't formed in a vacuum, and so there is not such an clear division between the "right" way of doing it and the "wrong" way of doing it. Nobody is completely ignorant of the data when they start to theorize. That is a real problem with nearly all statistics-based results that are published today. They are all presented with significance calculations based on the assumption that the forming of the theory was independent of the data - an assumption that is very unlikely to be completely true. Therefore nearly every significance published is an overestimate.

This was much less of a problem when collecting data and analysing it was difficult and laborious. Now that large data sets fly around the internet, and every researcher has the capability of running analyses at the click of a mouse, it is a problem that has already got out of hand.

I didn't want to be rude at the time, but I found Campbell's response shocking. He seemed to fully accept my argument, but wasn't bothered by the implication that pretty much all published research relying on analysing pre-existing statistics was wrong. Rather, his conclusion was that since everybody else was doing what he was doing, nobody should complain and demand "purity" (his scare quotes). That came to mind particularly reading Bruce Charlton's discussion of the state of honesty in science.

The story of real science

Bruce Charlton has published what he calls an "mini-e-booklet": http://thestoryofscience.blogspot.com/

I think he is saying, in greater detail and at much more length, and with the point of view of an insider, what I was saying in the last few days: that science has declined, because science has become an industry which no longer allows for the extraordinary honesty that real science requires.

This is the problem of science today – it has been bloated by decades of exponential growth into a bureaucratically dominated heavy industry soviet factory characterized by vastly inefficient mass production of shoddy goods. And it is trundling along, hour by hour, day by day; masses of people going to work, doing things, saying things, writing things…

Science is hopelessly and utterly un-reformable while it continues to be so big, continues to grow-and-grow, and continues uselessly to churn out ever-more of its sub-standard and unwanted goods.

Switch it off: stop making the defective glasses: now...

There are some very general arguments he makes which I have been meaning to spell out for a while. He suggests that the peak of science was in the mid-20th century, and it was a transitional state.

this transitional state of classic science was an early phase of professional science, which came between what might be called medieval science and modern science (which is not real science at all - but merely a generic bureaucratic organization which happened to have evolved from classic science). But classic science was never a steady state, and never reproduced itself; but was continually evolving by increasing growth, specialization and professionalization/ bureaucratization.

I think such transitional phases occur in different fields quite frequently. Part of my disillusionment with libertarianism is that it is an attempt to recapture a transitional state in government that was never sustainable - the state where a new class is taking over power and opens up freedom for everybody because it has not yet thrown off its self-identification as an underdog that benefits from freedom.

The failure of science is also an aspect of the widely-recognised but ill-understood problem of trying too hard: some things can only be achieved by trying to do something else.

The scientists of the past, like the individuals making up the governments of the past, were privileged. They ruled or researched not in order that they optimise some output, but because they could - they had reached positions of genuine personal responsibility, and had to make their own judgement.

If these "very general arguments" sound rather woolly, do not adjust your set. That's why I haven't published on them already - nevertheless, I bring them up now because they're bugging me and I think Charlton's writing is relevant to them.

Back to the specifics, Part 3 quotes an earlier post of Charlton's that chimes very closely with what I was saying yesterday:

Charlton BG. Are you an honest scientist? Truthfulness in science should be an iron law, not a vague aspiration. Medical Hypotheses. 2009; Volume 73: 633-635


Anyone who has been a scientist for more than a couple of decades will realize that there has been a progressive and pervasive decline in the honesty of scientific communications. Yet real science simply must be an arena where truth is the rule; or else the activity simply stops being science and becomes something else: Zombie science. Although all humans ought to be truthful at all times; science is the one area of social functioning in which truth is the primary value, and truthfulness the core evaluation. Truth-telling and truth-seeking should not, therefore, be regarded as unattainable aspirations for scientists, but as iron laws, continually and universally operative. Yet such is the endemic state of corruption that an insistence on truthfulness in science seems perverse, aggressive, dangerous, or simply utopian.


There are points I disagree with: Charlton tells the orthodox story of Lysenko - he was a gangster, he brought politics and political arguments into science. As I said yesterday, that lacks the understanding that he believed he was not the first to do so, that he believed he was only trying to correct the political influence that had already occurred. We are distracted by the fact that Lysenko's enemies were not merely removed from influence, but actually imprisoned - that is incidental, just part of the difference between Stalin's Russia and our world. The dissenting scientist today is as much an enemy of the state as Vavilov was, the only difference is that our establishment is secure enough to leave its enemies at large, while Stalin wasn't.

The reason I insist on this is that the orthodox story makes the problem seem too easy: don't allow monsters like Lysenko, keep politicians out of science. It isn't that easy - the politics that matters is the "office politics" of science itself, not the real politics of the government.

Charlton does not suggest a solution like mine of yesterday - de-emphasising the quest for originality in favour of more checking and reproduction - but it's clearly a prerequisite for the sort of changes he does advocate. To restore the primacy of truth to science a necessary step would be to ensure that only truth-seekers were recruited to the key scientific positions, and to exclude from leadership those who are untruthful or exhibit insufficient devotion to the pursuit of truth. Obviously, before you can do that you have to have a way to find out who is truthful and who isn't - you have to check.

Certainly there needs to be a slowing-down of science - Charlton and I are as one on that.

There's another point that Charlton gets close to: Real achievement in science requires a great deal of luck - the thing you are looking for has to really be there. However, when someone is in a career, it is unjust to value them by whether they are lucky. That is one of main forces that has driven a wedge between the practice of science and any real product - every research project has to produce something publishable (failing incompetence by the scientists), whereas in reality most research of the most valuable kind finds nothing, producing only a few jackpots for the lucky. The only solutions within the structure of science as bureaucracy is to either know what you are going to find in advance (which is useless), or publish results which are in fact devoid of real content, drowning any real results in the noise. This is largely achieved by abuse of statistics - something I thought I'd addressed in relation to economics, but I can't find. Perhaps I'll post something later.

09 August 2010

Originality and Science

One probably-final point to come out of the Lysenkoism discussion of the previous two posts:

Yesterday I admiringly referred to Richard Feynman's quote

I'm talking about a specific, extra type of integrity that is not lying, but bending over backwards to show how you're maybe wrong, [an integrity] that you ought to have when acting as a scientist. And this is our responsibility as scientists, certainly to other scientists

As I said, that is a key part of cutting out the cascade of distrust that can occur when science becomes politically sensitive.

The problem is that that was an easy thing for Feynman to say, because Feynman was a flamboyantly insane genius, and the last thing he ever had to fear was being ignored. The situation is rather different for one graduate student or new PhD among twenty aiming for the same grant-funded research post. In that position, playing down the significance or certainty of one's own work is a ticket to the dole queue.

And the density of competition is astonishing. There was a piece in Nature a couple of years back, on the limitations of fMRI, that pointed out that from 2007-2008 there have been eight peer-reviewed papers published involving fMRI per day - 19,000 since 1991. "About 43% of papers expore functional localization and/or cognitive anatomy associated with some cognitive task or stimulus". Thousands upon thousands of papers, each searching for the little piece of originality that will give them importance.

However, this torrent of research demonstrates a solution as well as a problem. I wrote yesterday that "it is ... impractical to replicate every experiment, confirm every observation, check every calculation". Clearly, I was wrong. There is ample manpower in the science industry to double- and triple-check important results, but the system does not value the work highly enough for it to actually get done. Only original work actually merits funding.

That is a widespread problem in non-commercial fields, most obvious in the arts. In commercial arts, most artists make small variations or combinations of existing products, just trying to be a little more attractive and entertainment. The minority who are truly original are highly valued, because they are providing material for the rest to refine or perfect. Indeed, I can think of no other distinction between "high" and "popular" art, but that high art always seeks to be original, and popular art isn't too bothered. In academic arts, the only valid work is to do something really new. The end result is a product that is always different, but never very good. In science, every new paper is original, but most of them are wrong.

I would assume that in the cases of both art and science, the original assumption was that the market worked well enough to perfect existing work, but that originality required help and subsidy. However, the subsidised sectors at length became isolated from the commercial, to the point that now there is no commercial sector relevant to the academic work being done, and the new stuff is being pumped out into a vacuum.

It seems obvious that it would be beneficial for science to move more slowly and carefully, but the academic system has evolved in a way that does not permit it. It would take a major shakeup to get the science establishment to start to value that caution.

08 August 2010

What they want to hear

There was another interesting point in the discussion at Hans von Storch's that I brought up - an interesting comment by "Toby" on the earlier piece:

Lysenko told the politicians what they wanted to hear - a "short cut" to socialism. Which side of the current "debate" is telling politicians what they want to hear? The ones arguing that money must be spent and sacrifices made? Or the ones advocating that nothing be done?

That is a good question, and is the root of much of the political polarization of climate science.

Toby implies that politicians want to hear that nothing need be done - money need not be spent.

A right-winger - like myself - believes that what politicians want to hear is that their departments and budgets must be enlarged.

As I explained, the distrust of motives is enough by itself - without any actual dishonesty or malpractice - to mess up the scientific process in a field where unequivocal confirmation or rejection of theories is difficult to come by.


There has been some interesting discussion at Hans von Storch's blog about Lysenkoism. Nils Roll Hansen wrote some posts.

I don't agree with the conclusions reached.

Lysenko was not a politician, he was not a fraud, he was not an ideologue. Lysenko was a scientist.

Lysenko, like the majority of scientists today, worked for the government.

In the scientific controversy that involved Lysenko, he reported to his superiors (the government). That was his job as a senior member of the scientific establishment.

The scientific controversy was politically sensitive. Lysenko claimed that his scientific opponents were politically motivated: their science was based on bourgeois ideas of inherited superiority. That claim was not implausible, and Lysenko had no reasonable alternative but to draw the attention of his superiors to the possibility.

The politicians did their job - they reached a conclusion about how to run a government department based on the advice they received and their judgement of that advice.

When we tell the story of Lysenkoism, we tell it in the knowledge that Lysenko was wrong. What we look for are the indications that the process was bad - that the wrong conclusion was being reached.

My opinion is that there are no such indications. Yes, it was "politicized science", but the main political force on the science was the belief that orthodox genetics was itself the product of the political assumptions of the Western scientists that developed it. That perception was probably exaggerated, if not totally erroneous, but it was a genuine belief honestly held.

The point is that for politics to mess up science, it is not necessary for anyone to let the political implications of a theory take precedence over the evidence. All that is necessary is for some participants to believe that other scientists are doing that. That is enough to cause theories to be suppressed, and thereby for the science to be systematically skewed.

It is not enough, either, to say that at the end of the day the evidence should speak for itself, and the trustworthiness of its spokesmen not be relevant - nullius in verba, and all that. That is all very fine, but it denies the fact that some science is difficult. It is so impractical to replicate every experiment, confirm every observation, check every calculation, that nullus in verba is the next thing to radical scepticism in the philosophical sense. You have to trust some scientists, and that means you have to choose who to trust, and that means you have to take into account politics.

In the very long run, you can learn who is actually trustworthy and who is not. But that is a painful bootstrapping process - you need a little trust to give you some facts, and then you use those facts to evaluate the trustworthiness of those who addressed them. That gives you a little more trust, to gather a few more facts, and so on.

To call, as Hansen does, for "independence" for science does not address the problem. It just means that scientists will be punished for scientific dissent rather than political dissent - which makes the situation worse. If science is run by politicians, you can probably advance whatever theories you like so long as you support the right policies. If science is run by scientists, you must support the authorized theories to succeed.

There is, then, no silver bullet to depoliticise science. There are, however, treatments that can make science work better. Since a small amount of distrust has such a catastrophic effect, the least dishonesty cannot be tolerated. This is behind the now dying attitude that Feynman talked about, of bending over backwards to draw attention to everything that tells against you theory, so that you cannot possibly be accused of concealing any of it.