In the Eye of the Beholder
Physics World features a list of the "Ten Most Beautiful Experiments in Physics" this month, with some interesting commentary about what constitutes "beauty" in this context (Though the author shows a disturbing lack of pop-culture savvy when he cites one particular comment from SlashDot: "One of the contributors described watching small plastic bags circulating in wind pockets, commenting that 'sometimes there's so much beauty in the world, I just can't take it'."). The New York Times also picks up on the story, offering the top ten with additional annotations about the history of the experiments.
The "winner" in the survey was the double-slit interference experiment carried out with electrons, with Young's original double slit experiment with light making the list as well. They're hard to argue with:
The double-slit experiment exemplifies the wave-particle duality of light, as well as quantum physics itself. It demonstrates that light interferes with itself in passing through a pair of slits. It also shows that even single electrons - proceeding one by one - interfere. Richard Feynman is said to have remarked that it contains everything you need to know about quantum mechanics.
There really is something magical about the way that single electron counts pile up in just the right way to make an interference pattern. You can do the same trick with light (given a sufficiently sensitive camera), but it's not as impressive-- everybody already expects light to interfere, and the quantization somehow isn't quite as surprising, either.
For the most part, the Physics World survey respondants went deep into physics history for their choices: only three of the ten were performed in the 20th century, while Galileo alone gets two mentions in the top ten, and one experiment dates to the third century BC (Eratosthenes's measurement of the Earth's circumference based on the lengths of shadows at noon at different latitudes). They were also strongly inclined toward simplicity, with the Milikan oil-drop experiment and Rutherford scattering being the only really complicated ones on the list.
As for my own picks, "beauty" is an awfully difficult thing to judge, though, and it's not a term I tend to apply to experimental physics. This is largely because I'm an experimentalist by trade, and tend to be less impressed with simplicity and elegance than I am with experiments that find extremely clever ways to do astonishingly difficult measurements. Cavendish's torsion pendulum is a simple and elegant demonstration of the inverse-square law of gravitation, but the updated torsion pendulum experiments at the University of Washington absolutely blow me away. The ingenuity they've shown in finding elegant solutions for all the problems that crop up in torsion balance experiments, while keeping the Cavendish concept as the core of the device, is simply amazing.
Similarly, while the double-slit experiment with electrons is perhaps the most elegant demonstration of wave-particle duality you'll find with matter waves, I'm rather fond of the recent fullerene diffraction experiments-- everybody already knows that electrons and atoms behave like waves, but these guys have set out to observe diffraction in the largest systems possible. They've seen clear diffraction with C_60 and C_70-- those soccer-ball-shaped molecules consisting of 60 or 70 carbon atoms-- and are planning to try even larger systems (one rumor said they were planning to try diffraction of viruses).
And as long as we're talking about matter-wave intrference, I think the Davisson-Germer experiment (showing the diffraction of electrons by nickel crystals) gets somewhat slighted. It's not what you'd call conventionally beautiful-- the experiment worked only because they broke part of their apparatus-- but they get extra bonus points for serendipity and sheer determination. They blundered into one of the great demonstrations of the wave nature of matter, and helped confirm the daring (and seeming daft) predictions of the nascent quantum theory. They're sort of the physics equivalent of Alexander Fleming and his moldy Petri dishes.
Rutherford scattering (which did make the list) is another great story. The classic experiment demonstrating the structure of the atom grew out of what was basically a make-work project for a grad student. Rutherford was bombarding thin gold foils with alpha particles, which according to then-current understanding, should've passed clean through the foil. He set a student to checking whether any were actually reflected, expecting a null result, and instead found that lots of particles were scattered backwards-- he famously described it as "quite the most incredible event that ever happened to me in my life. It was almost as incredible as if you had fired a 15-inch shell at a piece of tissue paper and it came back and hit you."
My natural bias toward complicated and audacious experiments leads me to agree in large part with the author of the Physics World piece when he says:
I was a little disturbed, however, by the ease with which many people seemed to think that the experiments that they were proposing had been conceived, or could be carried out and understood. This seemed a function, in part, of the way that these experiments are often taught. Demonstrations can vastly simplify the experimental process through the use of modern equipment constructed with the "right answer" in view. Textbooks and Web simulations - which exist for most of the experiments on the top 10 list - can involve far greater simplifications.
One of the worst victims of this effect is the Michelson-Morley experiment (last one, I promise), which I've mentioned before. It's a wonderful experiment (says the professional Optics Guy...), which lasers have rendered a trivial demo-- I knocked together a table-top Michelson interferometer in about fifteen minutes one afternoon a few weeks ago. Michelson and Morley did it without the benefit of lasers, though-- they used a sodium lamp, and arm lengths of several meters. I hate to think of the amount of time they spent squinting at the damn thing while they tried to align it. And then they floated the whole thing on a vat of mercury and spun it around-- all for a null measurement. It's easy to underrate what was really a tour de force experiment (and, indeed, it limps into the survey midway through the "also ran" list of the original article). Michelson and Morley get slighted because people now tend to think of their experiment as being easier than it really was, and because the theory they disproved (the "luminiferous aether") seems a quaint anachronism.
It's interesting to note that, with one or two exceptions, all the experiments on the list are used as class demonstrations or very simple labs these days. Those really are the things that stick with people from physics classes (which is really the only place you hear much about the great historical experiments). That also probably accounts for the inclination toward very simple experiments, rather than the more complicated and clever ones that impress me for their technique-- you can't do fullerene difraction or the real Michelson-Morley experiment in class, but two-slit interference of electrons is possible.
(And again, I'm always struck by the way that even really simple demos can be so amazingly effective-- I broke out the Liquid Nitrogen demos to introduce laser cooling to our freshman seminar class yesterday, to good effect. But even the senior science majors in the Relativity and Quantum Mechanics class afterwards (which I also taught-- two diferent classes, back to back. I was surprised at just how difficult and draining that turned out to be...) were impressed by just seeing nitrogen dumped out on the table.)
I could ramble on about this stuff for hours, but I'll stop here. It's an interesting article, even if the actual results (as with all such surveys) reveal more about the people who responded than the TRVTH of what the greatest experiments in physics really are. It's fun to look at what's on and what's off the list, in the same way that it's fun to argue about whether the 1927 Yankees could've beaten the 1972 Dolphins in roller hockey, and other such sports absurdities. Basically, it's a good excuse to look back and celebrate some of the most clever experiments in the history of science, and remember just how cool it is that the world works the way it does.
Question 2 (90 pts.): Which Tire?
Having twice called them "humorless dorks" for their Insultingly Stupid Movie Physics page (here in some detail and more recently in passing), I feel like I ought to say something nice about the people behind the Intuitor collection of web pages. Not just because I'm a sap, but because there's some really good stuff there (mixed with some less wonderful material).
The Amazing Applications of Probability and Statistics page has a couple of very nice pieces, one explaining how the problem of false positives in tests with very low failure rates can be much bigger than you might think, and another explaining Benford's Law, which describes the probability of finding one of the digits 1-9 as the first digit in a number drawn from real data (another explanation is in this New York Times article).
They also have a pretty good Basic Physics Savvy Quiz. A couple of the questions are of slightly dubious value (a couple are basically questions of semantics (#3 and #34), while some others are only vaguely related to physics (#29, for example)), but for the most part they do a good job of addressing common misconceptions, or at least as good a job as a true/false test can do.
(For those who care, I aced it, but that was as much because I have some familiarity with how the people who make these things up think as because of my great knowledge of physics...)
This is actually a tricky business, and there's a cottage industry of sorts in making up tests for conceptual understanding in physics. It's difficult to come up with straightforward multiple-choice questions (let alone true-false questions) that actually test a student's understanding of the important concepts, as opposed to their ability to work out the meaning of confusingly-worded test items, or just plain test-taking skills. (In terms where I give multiple-choice quizzes, I always make one where every answer is "C," just to see if anybody will change a correct answer just because they think that can't possibly be right...)
I can't quite decide whether it's easier or harder to come up with questions for an actual exam. On the one hand, it's easier to miscalculate, and come up with questions that are way too hard, or too time-consuming for a timed test. It gets worse as you move on through the curriculum, too-- there are maybe a half-dozen problems in quantum mechanics with the kind of exact analytical solutions you like to have in an exam context. The vast majority of the interesting problems in physics require approximate solutions of one sort or another, or are just too complicated to ask students to grind through in the short time periods available for most exams. (When I was in college, they got around this by having 24-hour take-home final exams, which were absolutely brutal, but did contain some interesting problems... Williams runs on an honor code system, though-- we don't have an honor code, so exams need to be scheduled and proctored. This is a source of mild irritation which may lead to another post later on...)
On the other hand, knowing that students will be working the whole thing out in detail allows some more freedom in the choice of problems. Multiple-choice questions are necessarily graded on a binary system (though I've often wished I could fail people who choose "Heavier objects fall faster" on multiple-choice tests in mechanics...), with no partial credit (unless you're a shmuck and write a question with two correct answers). Full problems are graded on a broader scale, and allow opportunities for partial credit. This gives students the opportunity to bail themselves out with the pleading "If I knew how to do part a, I'd do this for parts b, c, and d..." essay, and also allows the instructor to save a bit of face by being very generous if the test turns out to be harder than intended.
Still, it's a dicey business. You have to walk a fine line in exam writing-- if the test is too easy, the students will love it, but you don't get any real information out of it; if it's too hard, they'll resent it, and again, you won't get any real information. If everybody's clustered in a ten-percentage-point range at one end of the scale or another, you can't tell who really understands the material, and who's completely lost. The ideal is something with around a 75 average and a good spread of grades-- the best students should be over 90%, and the worst down around 60. I think I managed this once in the ten exams I gave last year.
Why am I going on about this? I'm giving an exam on Special Relativity next Wednesday, so I'm trying to come up with good test questions...
(And, just for the record, I haven't forgotten that I ended my big intro-to-Relativity post on a cliff-hanger. I'll get to the rest of it eventually-- it's hard to do without pictures and equations, though. I'm tempted to Google up somebody else's explanation and just link to it, but I'm a little too vain for that...)
All Knowledge Is Contained in... Blogdom?
While I'm not generally a big fan of "do my homework for me" posts on Usenet or elsewhere, it occurs to me that there's an outside chance that somebody who reads this might know something useful about this topic, so:
Anybody out there have any experience with the introductory physics textbooks Matter & Interactions (two volumes, subtitled Modern Mechanics and Electric & Magnetic Interactions) by Ruth Chabay and Bruce Sherwood, either from the teaching side, or the student side? We're in the process of revising the introductory curriculum here, and some aspects of this look very attractive. It's big on computer-based modeling, and works lots of "modern physics" topics in early on in the sequence, which is one of the things we hope to accomplish.
I'd be interested in hearing from anyone who has experience with the text, though. The order of topics is a little unusual, and switching to this from the more traditional Halliday, Resnick, and Walker would require quite a bit of work. If anybody out there has taught from these books, or taken a class that used them, let me know what you thought, either by email or in a comment. Is it a pain to teach out of? Is it hard to learn from? Is it simply a joy to use, and we should switch at once?
The "blogosphere" has so far failed to get anybody a job, but maybe it can be helpful in doing the job I already have...
A couple of people have asked me to comment on the big physics fraud story in Salon. I've held off talking about it for a variety of reasons-- the jury's still out, I don't know the details that well-- but mostly because I'm not quite sure what to say about the whole mess.
I'd explain, but there's too much, so let me just sum up: a scientist at Bell Labs, Jan Hendrik Schoen (which I'll use as the ASCII rendition of "Schön" (o-with-umlaut)) published a number of papers with some really surprising results-- they managed to make big organic molecules behave like semi-conductors. This was shocking and ground-breaking work, or so it seemed. Other groups trying the same experiments have failed to duplicate the important results, and now people have noticed that key graphs in the various papers Schoen published appear to be disturbingly identical. The data are simply too good to be true, and it's looking more and more like they're simply not true, that the whole thing was fabricated, in whole or in part.
As yet, there's no solid evidence as to what sort of chicanery went on, but it doesn't look good. And this raises the usual slew of questions: "How did this happen?", "What was he thinking?", "Why didn't anybody catch this sooner?" Salon cranks the angst to 11 (as is their wont) by invoking the hordes of young researchers who had planned to build their careers in the new field Schoen opened, but now find themselves stranded (they also annoyingly harp on the fact that "taxpayer dollars" were spent on the work in question). Even backing down a bit from that breathless level, though, it's a troubling situation. We can expect this to be added to the list of scientific failings trotted out by everyone from anti-environmentalists to "Intelligent Design" theorists, to post-whateverist professors of literature still smarting from the Sokal hoax.
It's tempting to say "Hey, the system works..." After all, if Schoen's work was fabricated, the fraud was discovered by other scientists who set out to reproduce his results. This is what science is all about, after all-- new results aren't accepted until they're verified, and the lack of independent verification is what led to the discovery of the problem.
There's some truth to that, but really, it's about as credible as the claims that the Enron and Worldcom crashes were a great success of the market system-- after all, their fraudulent business practices didn't save them from collapse. The market works, no regulation is needed. The fact that Ken Lay still has a nice house and money in the bank while his employees will be working at McDonald's when they're eighty thanks to the looting of their retirement savings is a minor detail.
The truth is, in physics as in the market, the system that's in place relies on the good will of the people who use the system. We count on businessmen to play fair, and not cook the books to inflate profits, and we count on scientists to honestly report on their experiments, and not make data up out of whole cloth. There are some safeguards in both systems-- auditors in business, referees in science-- but by and large, if somebody wants to cheat, or game the system, they'll be able to do it. There's no limit to the ingenuity of the venal, whether their goal is filthy lucre or a directorship at the Max Planck Institute. Schoen's done less material damage, in terms of wasted money and ruined lives, than Lay and Skilling and Ebbers did, but the damage to the credibility of science in general is pretty severe. (Assuming that the charges of fraud are borne out, of course...)
As for what should be done about the problem of scientific fraud, I'm not sure there's anything that can be done. Most of what needs doing will happen naturally-- people will check the work of their collaborators and co-authors more carefully, not wanting to suffer the same fate as the Bell Labs authors, referees will read papers more closely, and check the references more carefully, journal editors will scrutinize all new articles, and scientists will generally walk on eggshells where ethical issues are concerned. And it's not like Schoen exploited a legislative loophole that could be closed-- there aren't that many formal rules governing this sort of thing, and there's no overall scientific governing body to set formal standards. You can't even really expect referees to catch fraud in papers that are as revolutionary as Schoen's appeared to be-- it's a chicken-and-egg problem, as you need to publish the description of the work before people can try to duplicate it, but you'd need to duplicate the work before publishing in order to really prevent fraud.
As I've said before, it's a miracle that the system we have works as well as it does, but it's also better than just about any other system you could come up with. This is just one of the cases where the system failed.
UPDATE: The committee appointed to investigate the matter decided it was fraud, and Bell Labs has fired Schoen. Feel free to mentally remove the various "if it is fraud..." disclaimers from the above. Elsewhere in blogdom, David Harris has a nice summary of the committee's findings, and Charles Murtaugh has some comments (BlogSpot is down as I type this, or I'd link to them directly).
Jobs I Couldn't Do, #299,792,458 in a Series
Spent the morning at the "retirement benefit orientation" here at work, listening to salespeople from Fidelity Investments and TIAA-CREF try to convince me to throw money their way, and especially trying to talk up the idea of sinking a bunch of my retirement investments into the stock market.
That had to be the easiest job in the world three years ago, but I almost felt sorry for them today... Almost.
(Yeah, I know, in the long term (and I'm certainly in this for the long term...), the stock market always increases, blah, blah, blah... Still, that's got to be kind of a hard sell with the Dow under 8,000 and still falling...)
Essential Football Strategy Tips
A quick strategy tip for the New York Giants: When you get down to first and goal from the eight, immediately take a holding penalty, making it first and goal from the eighteen. Then you can run your regular offense, instead of feeling obliged to break out the ridiculous and worthless short power run plays. Whatever rocket scientist is behind calling the slow-developing sweep play on fourth and goal from the one (and third and one later in the game, because it worked so well the first time) should be horse-whipped. The announcers praised the field-goal kicker for being 9-for-9 on the season, neglecting to point out that only two of those have been from outside extra-point range, owing to the complete offensive breakdown inside the twenty.
I'm cursed in life to root for sports teams composed of chuckleheads...
At least they won. And the Pats won as well, while the Redskins and Cowboys lost. A good football day, in the end. But damn, that was a frustrating game to watch.
Substantive blogging will be light this week-- Kate's parents came up to visit over the weekend, which meant that time I might otherwise have used to get a lecture or two ahead at work (allowing me time to blogroll and post) was spent playing "chase" and "tug-of-war" with a ridiculous little dog...
We need pets.
Fads, Fallacies, and Amusing Television
SciTech Daily links a Washington Post article (from a month ago, oddly-- I'm surprised it hasn't vanished into the pay archives) about the fuzzy line between science and pseudo-science, particularly the way pseudo-science creeps into science forums like the Discovery Channel.
A Discovery Channel Store spokeswoman, Pamela Rucker, explains the situation by noting that certain store products present a more "whimsical take" on the company's broad, science-nourishing values. But the Discovery Communications empire harbors many such tensions. On the one hand, the corporation touts a public responsibility initiative to help young viewers "critically analyze" information they get from the media. On the other, its network Animal Planet carries a program called "The Pet Psychic," which disavows on its Web site any intention to convey "medical advice or other factual information."
I'll admit to a certain fascination with the "science" programming on the Discovery Channel and its various offshoots, particularly the "crank archaeology" subgenre, where vaguely credentialed "researchers" hold forth on how there's no way even modern technology could build the Pyramids, so they must've been the work of benevolent godlike aliens from the Orion Nebula. The contemporary UFO pieces are pretty amusing, too (and all seem to feature the same "expert"-- the show's not complete without an appearance by the Scary Neck Beard).
These shows are really something of a guilty pleasure-- sort of like watching professional wrestling. It's an entertaining spectacle, but ultimately, sort of degrading to everyone involved. The occasional nights when I channel-surf between "Smackdown" and "Mysteries of the Ancient World" are especially bad-- I sometimes start to wonder whether I wouldn't be better off spending my time going on the web and downloading something with more socially redeeming value, like porn.
On some level, I feel like I ought to come out against this stuff. It's entertaining, but it's crap-- and, as James Randi points out in one of his older books (which I got off Ralph Alpher's book shelves, and am reading in bits at work), it's insulting crap. Erich von Daniken and his ilk are essentially saying "Those primitive swarthy types never could've managed an advanced civilization without help," which is demeaning. As Randi notes, they never speculate that aliens must've helped build the cathedrals of western Europe, or the Colosseum in Rome even though those would seem at least as improbable as a giant pile of rocks on the banks of the Nile. It's far more impressive, to my mind, to consider the level of faith and dedication and sheer determination that would've been needed to build these things with primitive tools than to imagine them as lightly tossed-off by-products of alien civilization. Indeed, if aliens were responsible for the Pyramids, I have to wonder whether they're really all that advanced-- I mean, there's not even an eight-track player, let alone a DVD home theater system, let alone a holodeck... Were we colonized by grey-skinned, bug-eyed, back-to-nature types, or something?
So I feel like I ought to be out there fighting the good fight, like Randi and Bob Park, denouncing junk science wherever it rears its ugly little head. I am a trained scientist, after all, and ought to use some of that training and education to combat idiocy for the general good of society. On the other hand, though, it's hard to really do that sort of thing without coming off as a humorless dork (something I try not to do). You could spend every waking moment hunting down misuses of physics in the media, and writing cranky letters to the editor about them, but what would be the point? Most of this stuff is fairly harmless, and it's better to choose your battles and fight only the actively harmful stuff-- the medical quackery that leads people to gamble their lives on placebos, or the cases where anti-intellectualism creeps into policy-making.
The counter-argument, of course, is that even the "harmless" stuff contributes to a culture of pseudo-science, and provides the anti-intellectual background in which deplorable frauds like "Intelligent Design Theory" grow and thrive. To which I can only say: maybe. It might be true that the steady diet of pseudo-scientific gibberish pushed by the "alternative medicine" crowd, and the various "-ologists" of the fringe movements contributes to the policy successes of the more theocratic of our politicians. On the other hand, though, this sounds a lot like the "violent tv shows cause violent behavior" argument, which I think of as a classic example of dubious science.
The other question that has to be asked is "What can we really do about this?" It's frequently said that the success of quackery in its various forms represents a failure of the scientific community to educate the general public sufficiently in what science does. There's a little of that, to be sure, and also some willful ignorance on the part of the victims, but I think the root causes are deeper and probably impossible to eradicate.
The ultimate cause of most of the great examples of junk science is a combination of naive optimism and wishful thinking. It's not so much a matter of people being too poorly educated to understand that the massive dilution theory of homeopathy is a bunch of crap, or that the laws of thermodynamics forbid perpetual motion machines, as it is a case of people desperately wanting free energy and a cure for what ails them. The victims of pseudo-science aren't being gulled because they're naturally foolish, they're buying into the scams because they want to be fooled.
We'll never get rid of the desire for impossible rewards-- it's a part of human nature. It's not even clear that we should really want to get rid of it, as it's closely related to whatever it is within us that makes a man look at a soaring bird and say "I bet I could make a machine to let me do that..." Our greatest strength as a species is our ability to look at the world, want something more than what it offers, and make it the way we'd like it to be. In some cases, that desire to re-shape the world along more pleasing lines leads to the great advances of human civilization; in other cases, it leads people to invest in "free energy" schemes that run counter to all the laws of physics. Deep down, the vast majority of people understand that there's no such thing as a free lunch-- the essence of the laws of thermodynamics in one glib phrase-- but that doesn't stop people from wanting there to be a free lunch, or trying to get one when they think they see a chance.
I'm glad that there are people like Park and Randi out there fighting the good fight, and doing their part to see that the worst of the scam artists get what's coming to them. And I'm happy to stand with them to oppose the big and harmful attempts to legislate nonsense-- be it "intelligent design" or "alternative medicine" or the attempt to buy science to fit a particular ideology. But at bottom, it's not worth sweating the small stuff, and it's probably futile to fight it-- Randi and Park, for all their efforts, stand about the same chance of eliminating gullibility through education as Richard Dawkins does of eliminating religion through annoying rants. I'll denouce crap when I run across it, and it seems harmful enough to do some damage, but I'll stay away from bashing stupid movie physics, and continue to chuckle mildly at the Discovery Channel when it veers off into the hinterlands of crank science.
I still feel guilty about watching pro wrestling, though...