Scientists Aren't Smart Enough to Understand Democracy, Blogger Says

Some scientists have apparently concluded that people are too stupid to run a democracy. One would think the roughly 400-year history of democracy on the North American continent would have humbled the researchers at least a bit. But apparently not. Here's the first sentence from Yahoo! News:

The democratic process relies on the assumption that citizens (the majority of them, at least) can recognize the best political candidate, or best policy idea, when they see it.

Already we're into some thorny problems. What's the "best" candidate or idea? Do the scientists know? If not, how can they test their theory? Also, what does "best" mean in this context? Saying that there is a best candidate implies that there is some ordering of candidates from best to worst, and that a sufficiently knowledgeable person could choose the best among them. Does that seem remotely likely?

Speaking of testing the theory, here's what the scientist actually did:

Mato Nagel, a sociologist in Germany, recently implemented Dunning and Kruger's theories by computer-simulating a democratic election. In his mathematical model of the election, he assumed that voters' own leadership skills were distributed on a bell curve — some were really good leaders, some, really bad, but most were mediocre — and that each voter was incapable of recognizing the leadership skills of a political candidate as being better than his or her own. When such an election was simulated, candidates whose leadership skills were only slightly better than average always won.

Is it beyond the pale for me to point out the irony in a German researcher building in the assumption that the candidate with the best leadership skills is the best one? I seem to recall one inter-war German Chancellor who was widely hailed as a great leader. But I risk Godwinning my own post.

More to the point, was it necessary to test this by computer simulation? Garbage in, garbage out: once the various, and highly dubious, assumptions have been made, the conclusion is inevitable, but meaningless.

Nagel concluded that democracies rarely or never elect the best leaders. Their advantage over dictatorships or other forms of government is merely that they "effectively prevent lower-than-average candidates from becoming leaders."

This is perhaps the most tendentious part of the entire story. It may be the case that democracies rarely elect the "best" leaders. Heck, I'm willing to stipulate that it never happens. Maybe there's some mayor of Podunk who could this country like a finely oiled machine, and we'll never know. But so what?

Democratic governments are not run by a solitary individual. In the U.S., the President has lots of power, but he is checked by other power centers. Our democracy does not elect a dictator who rules over us without limit. Rather, many leaders are elected who each control some threads in a web of power. We individuals doing the electing have many outlets available to us: local school board members, city council members and mayors, state legislators, state governors, U.S. congressional representatives, U.S. Senators, and finally Presidents.

Taking the contention of the researcher to heart, we understand our neighborhood better than our whole town, our town better than our whole state, and our state better than the whole country. So it's probably true that we do a poorer job evaluating Presidents than school boards. But that's no reason to trash democracy. Rather, it's a good rationale for federalism: local decisions made at a local level.

Finally, the notion that the election of ideal leaders is the primary goal of any system of government is at best incomplete. Democracies, for instance, shine in at least two other areas. First, they provide an orderly means for one government to be replaced by another one. When a dictator dies (or is deposed), succession is often a problem. Perhaps there are multiple children who squabble over who gets to rule. Perhaps a military strongman grabs at the brass ring. One simple advantage of a democracy is that the new leader knows he at least has a majority behind him, and that support helps stabilize the transition. (It's notable that the least stable democracies, e.g. Italy, are parliamentary ones that are usually run by a coalition. Thus the leader does not have the unequivocal support of a majority.)

Second, democracy, unlike other forms of government, helps to promote civic responsibility. The subject of a dictator may need to be clever and know how to navigate the halls of power, but he has no stake in the process. The citizen of a democracy does, and can, at least theoretically, make a change. That mere fact can make an enormous difference, but again we return to the need for federalism, since change is easier at the local level than the national.

Environmental "Truth" in a Nutshell

Yesterday my family visited the Chabot Space and Science Center in Oakland. Pretty cool; definitely recommended. It's on a hilltop overlooking the bay, in a beautiful state park, and has many nice exhibits for ages 4-12 or so.

On the downside, a big chunk of the museum is given over to Bill Nye the Climate Guy. One attraction was a bicycle kids could ride while watching a video of Bill biking in front, and confronting a boy and his father for driving to the gym. "Why don't you bike or walk to the gym, or better yet, give up those gym memberships altogether?" says Bill. "Why don't you mind your own *&^# business?" I wanted them to respond.

But the worst one, and this really sums up something essential about the environmentalists, was a wind turbine demonstration. You were supposed to pull a handle, which started a flow of air going, that turned the wind turbines and made a light go on. Simple. Only, when I tried it, because of the angle of the turbine blades, the wind didn't make them turn. They just juddered and rocked back and forth, never actually making more than a few degrees of rotation in either direction.

But the light came on anyway. Obviously, the demo had been rigged: when you pulled the handle, it made the light go on. The turbine had nothing to do with it. I get why a museum would do this, of course: you want the demos to work and not frustrate the kids. But the irony is thick: environmentalists constantly try to rig the numbers to make it look like converting to a more expensive, less efficient form of energy production will create jobs, or to make it look like wind plants produce more energy than they do, or what have you. Are here they are rigging a museum demo about one of their pet technologies. Priceless.

Break Out the Potassium Iodide!

Or rather, don't. XKCD explains relative radiation doses. It's interesting to read, and to compare the differences among Three-Mile Island, Fukushima, and Chernobyl.

Clearly, TMI was nearly a non-event by comparison. The maximum external dosage was only 1 mSv, one-third the dosage from a single mammogram. The lowest radiation dose clearly linked to increased cancer risk is 100 mSv, so if you got that maximum external dosage near TMI for 100 years, you'd just reach this threshold. It's pretty clear that containment worked at TMI.

Next take Fukushima. All the facts are not yet in, but it's clearly worse than TMI: a one-day dose 50km away was already measured at 3.6 mSv. But — probably — no actual reactor fuel has been exposed to the atmosphere, which implies that the radioactive materials causing this dose have a short half-life and will quickly decay. There is reason for caution here, but this is almost guaranteed to be no Chernobyl.

It has been over 20 years since Chernobyl. And according to XKCD's chart, you can still get a 6 mSv dose in one hour on the grounds of the stricken reactor. That means spending a single day there, unprotected, would increase your risk of cancer. Spending two weeks there could very well kill you with a 2.1 Sv dose.

You could safely go on a camping trip to TMI today. Not so Chernobyl. The jury is still out of Fukushima, but my bet is that in 20 years it'll be much closer to the safer end of this spectrum.

The Decline of Planetariums

Or is that "planetaria"?

When I was a kid, I was lucky enough to attend a school that had its own observatory and planetarium. As early as first grade, I was exposed to the wonders of the planetarium show, where the sky was projected on a hemispherical ceiling from this funny ant-like device. An astronomer (or maybe just a science teacher) would explain about the constellations, or the planets, or the Milky Way, or whatever, and could demonstrate on the ceiling what he was talking about. And it wasn't just in school that I enjoyed them - I visited several other planetariums over these years (the 70s and 80s) and got the same basic thing.

So when did they turn into overpriced movie theaters? My family and I visited the California Academy of Sciences in San Francisco a couple of weeks ago. One thing we were really looking forward to was the planetarium show. It's billed as a planetarium; that's what I expected. But when we entered the room, it was just an Omnimax theater. We didn't get a planetarium show at all; instead it was just a 30-minute movie.

I guess it's easier for them to run a movie every half hour than a live show. And trying to recreate childhood experiences almost always disappoints. But the content itself is so watered down now. There's no basic descriptive astronomy: what's a constellation? what's a star? where are the planets? The subject matter is too broad (this particular one was about the origin of life in the universe) and necessarily, therefore, too shallow.

So take this as a call for planetarium shows like they had in the old days. And if anyone knows where I can still find one today, leave it in the comments!

Why Bother Trying to Prove There Is No God?

Science and religion are like oil and water. Science is about devising theories about how the universe works; religion is about devising theories about how it *ought* to work. Granted, there are religious people who want to encroach on science, and I suppose they've made some small amount of headway with slightly clever ideas about Intelligent Design and so on. But these ideas are really no threat, in my opinion.

Which makes it all the stranger when real scientists take them seriously. In this case, cosmologist Don Page "says that a slightly negative value of the [cosmological] constant would maximise this process. And since life is some small fraction of the amount of matter in galaxies, then this is the value that an omnipotent being would choose." Whereas in the universe as we know it, the cosmological constant actually has a small positive value. Therefore, concludes Page, God must not have fine-tuned the constant.

What nonsense, and what a waste of time. There are any number of explanations why Page could be wrong: maybe greater density of stars and galaxies leads to more radiation, making it less likely for stable molecules to form. Maybe a higher constant leads to a universe whose duration is too short for life to evolve. Maybe anything. And moreover: who cares? This sort of theological debate really doesn't belong in science. Page may have a real discovery on his hands: that he has found a link between the cosmological constant and galaxy formation. That's useful knowledge. But to try to leverage it into a theological argument demeans it.

DNA Sequencing: Useless?

Adam Keiper, editor of the excellent journal of science and society The New Atlantis, posts to the Corner about an interview by Craig Venter in which he states that the Human Genome Project has yielded "zero" fruit.

Specific DNA sequencing efforts certainly have yielded fruit. Venter points out that finding a gene that adds only 1-3% to the likelihood of getting some disease is of no clinical value: also true. Keiper takes this further:

For better or for worse, our limited genomic knowledge is already influencing people's decisions about important matters. Some women who learn from DNA tests that they have a genetic susceptibility to breast cancer are electing to undergo mastectomies even though they don’t yet have, and might never develop, the disease.

It's not clear Keiper is applying Venter's numbers to his own conclusions, but if so he's trying to take his argument a bridge too far. The BRCA mutations that have been shown to correlate highly with breast cancer change a woman's likelihood from a baseline of about 6-8% by age 70, to a hefty 85%. That is, if you have the BRCA mutation, you have an 85% chance (better than 5 in 6) of getting breast cancer by age 70. That's a huge result, and it's not surprising that some women found to have the mutation would choose to circumvent the disease by having double mastectomies.

Before women reading this rush off to get genetic testing, I should point out that the BRCA mutation is quite rare - well under 1% of the population - and only a small percentage of breast cancer cases involve it. So in that sense, its clinical value may be small since the test is relatively expensive. But wouldn't a cheap BRCA test be of enormous value to, say, a woman of 30 who was considering starting a family?

The larger point Venter and Keiper make is quite right: the HGP was overhyped. But they go a bit too far to deflate the hype. It did have some value.

The Evolution of Morality

Dinesh D'Souza proposes to demonstrate (in a 3-part essay: Part I, Part II, Part III) the existence of the afterlife using the fact that humans are possessed with moral sensibility. I find the argument interesting, but highly flawed.

D'Souza outlines his argument in this way (from Part I):

Unlike material objects and all other living creatures, we humans inhabit two domains: the way things are, and the way things ought to be. In other words, we are moral animals who recognize that just as there are natural laws that govern every object in the universe, there are also moral laws that govern the behavior of one special set of objects in the universe, namely us. While the universe is externally moved by "facts," we are internally moved also by "values." Yet these values defy natural and scientific explanation, because the laws of nature, as discovered by science, concern only the way things are and not the way they ought to be. Moreover, the essence of morality is to curtail and contradict the powerful engine of human self-interest, giving morality an undeniable anti-evolutionary thrust. So how do we explain the existence of moral values that stand athwart our animal nature? The presupposition of cosmic justice, achieved not in this life but in another life beyond the grave, is by far the best and in some respects the only explanation. This presupposition fully explains why humans continue to espouse goodness and justice even when the world is evil and unjust.

One sentence in the middle is the most contentious, and unsurprisingly is the part D'Souza spends the most effort on: the idea that "morality [has] an undeniable anti-evolutionary thrust." Let us see if this idea has any merit.

The first argument raised in its defense is this (from Part II):

So if we are mere evolutionary primates, how to account for morality as a central and universal feature of our nature? Why would morality develop among creatures obsessively bent on survival and reproduction?... Darwin argued that "although a high standard of morality gives but a slight or no advantage to each individual man and his children over the other men of the same tribe, yet . . . an advancement in the standard of morality will rtainly give an immense advantage to one tribe over another." Darwin’s point is that a tribe of virtuous patriots, with each of its members willing to make sacrifices for the group, would prove more successful and thus be favored by natural selection over a tribe of self-serving individuals.

But as biologists now recognize, the argument has a fatal flaw. The question we have to ask is how a tribe of individuals would become self-sacrificing in the first place. Imagine a tribe where, for instance, many people shared their food with others or volunteered to defend the tribe from external attack. Now what would be the fate of individual cheaters who benefited from this arrangement but hoarded their own food and themselves refused to volunteer to fight? Clearly these scoundrels would have the best deal of all. In other words, cheaters could easily become free riders, benefiting from the sacrifices of others but making no sacrifices themselves, and they would be more likely to survive than their more altruistic fellow tribesmen.

D'Souza says this is a "fatal flaw" and dismisses it without further consideration. I think he should not be so quick on the draw, though. The most obvious rejoinder is that altruistic tribes do not suffer cheaters lightly: they punish them severely, by exile or death. The development of altruism need not have the same genetic basis as the development of punishment, but clearly the combination of the two is powerful. Once this combination, which we might dub "enforced altruism", develops, it would quickly outperform selfish tribes.

Note that this does not say that all members of the tribe are forced to act altruistically (which is not really altruism). What it says is that some individuals express the altruistic trait, and that those who do not either fake it or get punished. (A few probably cheat and manage to escape punishment, of course. But now we are down to arguments over numbers: the conditions under which enforced altruism proves genetically successful depend on the ratio of altruists to cheaters, and of undetected to detected cheaters.)

Another argument for the evolution of altruism is based on the idea that individuals can be thought of as "survival machines — robot vehicles blindly programmed to preserve the selfish molecules known as genes," in the words of Richard Dawkins. In D'Souza's words:

This idea helps us understand why certain insects, birds, and animals endanger their own welfare to promote that of their fellow creatures. Vervet monkeys and prairie dogs, for instance, give warning calls that signal approaching predators, sometimes at the cost of becoming the target of those predators. Why would they risk their lives in this way? Kin selection holds that it is because they are genetically related to those they are helping. So there is an evolutionary payoff: The risk-takers are maximizing not their individual chance for survival but the chance for their genes to make it into future generations. From the gene’s point of view, helping one’s kin is simply a form of helping oneself.

But of course kin selection is a very limited explanation, in that it only accounts for why animals and people behave altruistically toward relatives. In life, however, humans and even some animals behave that way toward innumerable others who don’t share their genes.

It's worth pausing at this point to mention that this last bit is a major stolen base. All of us share the vast majority of our genes; only a tiny residue of genetic difference separates Swedes from Australian aborigines. So the selfish gene hypothesis could easily explain all altruism among humans.

But of course kin selection is a very limited explanation, in that it only accounts for why animals and people behave altruistically toward relatives. In life, however, humans and even some animals behave that way toward innumerable others who don’t share their genes. Robert Trivers argued that this is because of "reciprocal altruism." A better term would be reciprocal bargaining: What Trivers means is that creatures behave generously toward others in the expectation that they will get something in return....

Even reciprocal altruism, however, cannot explain the good things that we do that offer no actual return. A fellow gets up to give his seat on the bus to an 80-year-old woman. No, she isn’t grandma, nor is it reasonable to say that he’s doing it so that next week she will give him her seat. So neither kin selection nor reciprocal altruism provides any solution in this case....

Some biologists concede that evolution is at a loss here. "Altruism toward strangers," writes biologist Ernst Mayr, "is a behavior not supported by natural selection." Still, some diehard champions of evolution do try to accommodate such behavior within their evolutionary framework. Their best attempt is to argue that seemingly disinterested altruism toward strangers has a well-hidden personal motive. Essentially it is performed in order to enhance one’s social reputation.

Again, though, there are easy responses to what D'Souza characterizes as intractable problems. First, as already mentioned, all humans are kin. Given a random gene in a random human being, the odds are extremely good that that gene also exists in any other human. Second, the description of reciprocal altruism is too limited. While we may not expect that particular individual to reciprocate, we are members of a social species. A fellow who gives up his seat on the bus may expect that someone else will give him a seat when he is 80. The best way to perpetuate the social behavior of "giving up bus seats to the elderly" is to practice them. This is a different and more subtle argument than the simplistic one of enhancing one's personal reputation.

But all these responses are not really necessary, because D'Souza's arguments are weakened by his insistance that every instance of altruism have a genetic basis. Nowhere in evolutionary theory is it required that genes evolve unerring behavior. Many insects are fooled by flowers into "mating" with them; D'Souza would presumably argue that this behavior could not have evolved because insects that were not fooled would outcompete the others. But in reality, it's just not worth it not to be fooled: the cost of being fooled is low. Similarly, the survival cost of giving up one's seat on a bus is low. Some of his other examples (e.g., "assisting AID victims in Africa") are better, but note how few people engage in these forms of altruism.

He also ignores the fact that humans have brains. Once a capacity for altruism (even just kin altruism) and a capacity for abstract thought have both evolved, it is not a stretch to infer that those capacities might start to be connected. Human evolution is not only a genetic story, but also a social one. This is obvious to anyone who has raised children: in their state of nature, they are not particular moral. A young child will lie, steal and cheat if he can get away with it. It is only through normative socialization that these tendencies are curbed; clearly they are learned, not genetic. (This is not to say that genetic tendencies toward moral behavior are stronger in some than others. But no one is born a saint any more than they are born an NBA All-Star.)

One other point D'Souza makes about morality (in Part III) is this:

All evolutionary attempts to explain morality ultimately miss the point. They seek to explain morality, but even at their best what they explain is not morality at all. Imagine a shopkeeper who routinely increases his profits by cheating his customers. So smoothly does he do this that he is never exposed and his reputation remains unimpeached. Even though the man is successful in the game of survival, if he has a conscience it will be nagging at him from the inside. It may not be strong enough to make him change his ways, but it will at least make him feel bad and perhaps ultimately despise himself. Now where have our evolutionary explanations accounted for morality in this sense?

The argument here is that evolutionary explanations for morality reduce it to just a greater form of self-interest, and that this cannot be explaining by a "nagging conscience." But this is far from clear. Why couldn't the mechanism for inducing altruistic behavior be a mental tug that we call a "conscience"? D'Souza seems to imagine that evolution must express self-interest by a little red devil in our minds encouraging us to act deviously. But consider hunger, a sense that we can hopefully agree developed via evolution. When we are hungry, we are not induced to merely ape the motions of others who are eating. Rather, we feel sensations akin to pain, which we have learned we can assuage by actually eating. Similarly, our conscience is our mental conduit for feeling our moral sense, and it is expressed by a sort of pain when we do not act appropriately. Our conscience is designed to be triggered by things which even D'Souza would recognize as being in our gene's self-interest but, being an imperfect mechanism, is sometimes triggered by things which are not. This should not be any more surprising than the discovery that we sometimes eat unhealthy things.

Having made his various points about evolution, D'Souza hits us with his conclusion:

Now let’s make the supposition that there is cosmic justice after death and ask, Does this help to explain the great mystery of human morality? It seems clear that it does. Humans recognize that there is no ultimate goodness and justice in this world, but they continue to uphold those ideals.

By now it should be obvious why I think this is flawed: the premise - that human morality is mysterious and has no otherwise good explanation - is not reasonable. Human morality can be explained perfectly well by evolutionary means; D'Souza does them a disservice by his overly-simplistic explication of them.

Full disclosure: even if I did not think evolution provided a good explanation for human morality, it would take more than D'Souza's argument to convince me of the existence of an afterlife in which we are judged for our actions in this life. His argument makes the common error of confusing metaphysics with science. What evolution can predict is a scientific question: it can be tested, verified or found wanting, modified and tested again, etc. Such is the scientific method. His hypothesis, however, is untestable, and his argument, metaphysical. It requires us to assume at a minimum that there is no physical explanation for morality. As a somewhat silly example: suppose humans did not evolve from primates at all, but - as in 2001: A Space Odyssey - were created by aliens, who inserted some moral sense in our genetic code. That is certainly an alternative explanation for the origin of morality, and it does not require us to posit an alternative universe to which we are connected (by no means detectable to science). It would require us to posit alien genetically engineers with an interest in humanity, but it strikes me at least as an open question which hypothesis is the more unlikely. In the meantime, the evolutionary explanation seems to work better than most.

Einstein Vindicated Yet Again

One of the responsibilities - and joys - of doing science is in discovering something new, something that doesn't fit with existing theories. For that reason, it's important to continue to test those theories, and to test their limits.

NASA's Fermi Gamma Ray Space Telescope has been operating over the past year in this capacity. In 1905, Albert Einstein developed the theory of Special Relativity, which is based on the assumption that light travels the same speed under all conditions. This is hard to test thoroughly on Earth, because light travels so fast. The Fermi telescope, though, can record Gamma Ray Bursts (GRBs) - very short-term events (a few seconds at most) emitting a wide range of photon energies. GRBs sometimes occur very far away. The photons emitted race toward Earth for millions, even billions, of years, and we can record their arrival and calculate their relative speed.

In May, 2009, a GRB lasting 2.1 seconds detonated in a galaxy about 7.3 billion light years away. Fermi detected photons with energies about a factor of a million apart, which arrived only 0.9 seconds apart. Thus, the maximum difference in speed between the two would have resulted in a 3-second disparity in arrival time over 7.3 billion years (and possibly no disparity at all, depending on when the photons were actually emitted). This corresponds to an accuracy about 1 part in 100 million billion, or put another way, a difference of 3 nanometers/second in a speed of light of 300,000 km/second.

This is "good news" in the sense that it eliminates a class of theories about the structure of the universe. Some theories held that the speed of light actually was not quite constant under certain conditions, and would have predicted a much higher discrepancy that was measured here. Those theories must now be rejected or, at least, modified.

But it's sort of "bad news" (to the extent that any scientific fact can be given such value labels as "good" or "bad") in that it means we didn't discover a new discrepancy in our understanding of the universe. Such discrepancies are the source of all new theories; Einstein himself based Special Relativity on the Michelson-Morley experiment (which demonstrated that 19th-century theories about the propagation of light were fatally flawed). It would have been more "helpful" to find a large discrepancy in speed, which would enable us to reject all theories except those that admitted such a discrepancy. But, facts being facts, there's no use crying over it.

We know that the two great theories of 20th century physics, Relativity and Quantum Theory, are incompatible. The grand project of determining how they fit together just took another baby step.