David's Blague

So, my Dad is riding his bike from Fairbanks to Spokane right now. Dad and Mom and a bunch of other supporters of the ride are keeping a bike blog here: http://ridefair.blogspot.com to send encouragement and to share updates. Basically, Dad posts when he can, and everyone else fills in around the edges. It's a fun read, and fun writing, too.

Seed Magazine ran a contest, for which I belatedly wrote this entry (missing the deadline by over a week). The question they asked was, "What is the most significant force acting against science in society today? How can it be overcome?" This is my answer:

Dumb Questions

Ten years ago, a friend of mine graduating from Harvard accepted a job teaching Chemistry in a public high school outside Boston. Career-wise, this put him on a different path than mine, which made conversations with him a welcome break from my life in the cubicle.

"One of my kids asked the dumbest question today."

Me, "I thought the only dumb question was the question you don't ask."

Him, "I said to the class, 'Write this down, it's going to be on the test,' and one of my kids raised his hand and asked, 'Will we need something to write with?' Tell me how that's not a dumb question."

He made a good point. It was a dumb question.

I have though about that conversation a lot since then. I have young children now, who ask a lot of questions of their own. This, in turn, gets me thinking about how my own father dealt with my questions- good ones and bad ones.

One of his approaches, which I've re-purposed for the next generation of Bruntons, is humor. "Are we there yet?" "Yeah, I'm inside right now, playing the piano. What are you doing?" "No, Daddy, that's silly. We're still in the car." Lots of times, kids ask questions just to get a conversation going. Using humor is a way to oblige that urge, without pandering to dumb questions.

I can remember getting frustrated with my own dad doing this, and telling him that his answer was dumb. "Well, Dave," he'd say, "If you ask a dumb question, you'll probably get a dumb answer."

Imaginary Enemies

It's easy to imagine all manner of threats to science. Religious fundamentalism, postmodernism, funding cuts, decay of education, competing methodologies, and politicization all come to mind. But these have always been present, and science has flourished. What has changed? Why does science find itself, at the dawn of a new millennium, still casting around for modest improvements to hundred-year-old hypotheses?

During the same hundred years, computation has been reinvented at least a half-dozen times, with multi-core processors and cloud computing threatening to do it again. It has even been suggested that computation has hobbled science— the ubiquitous availability of raw data and cheaper computing power have rendered the scientific method obsolete.

This notion is as absurd as suggesting that the voyage of the Beagle (the trip, not the book) rendered further inquiry about the origin of species obsolete. More data plus more capacity to create models for comparison should result in scientific breakthroughs.

What about politics? Religion? The rise of fundamentalism? While it's possible to dredge up some conflict between empiricism and policy (or religion), these a more unlikely culprits than even computers. Science flourishes under scrutiny and competition. If the pope could not make the sun revolve around the Earth five hundred years ago, why should now be any different?

Besides, the underpinnings of fundamentalism and traditionalism are not threatening to science. Everything happens for a reason. Someday we will understand it all better. This was meant to be.

Perhaps there is a better model than the empiricism encoded in the scientific method? Axiomatic approaches- philosophy, logic, mathematics- have always been friends to science, but this is no assurance that other approaches would be. Is it possible a competing approach has rendered prior art obsolete? I think not.

Science is nothing but the method of empiricism. Ask a question, hypothesize an answer, conduct an experiment. Compare the results to the hypothesis. We have a framework of falsification and repeatability, but these are not the core of science- they are merely ways of distinguishing good experiments from bad ones. Suggesting we have nothing more to learn is the very height of hubris, where even we humans have not yet reached.

Shall we blame our educational systems? Classrooms have never been the place where science thrived, and they still aren't. We have a better educated populace than any time in our history- classrooms are not to blame. Culture? Perhaps scientists have grown fat and lazy along with the rest the population, relying on handouts and give-aways.

No, not even this has a ring of truth, as fine a sense of Schadenfreude as that would give. Despite all the talk of welfare moms and free riders, Americans are working longer and harder than ever. For scientists, this is doubly true- decades of training yield low-paying jobs without any real hope of a breakthrough.

The Greatest Threat to Science is From Science

The real threat to science comes from the scientific method, coupled with that incontrovertible and universal rule that dumb questions yield dumb answers. Science is beholden to the questions asked by scientists. Unfortunately, science has spent a great deal of the past hundred years accepting dumb questions from outside the scientific community and trying to answer them scientifically and defending scientific answers from non-scientific objections.

Science has very little to offer on topics such as God, morality, society, politics, policy, belief, or (dare I say it?) truth. These are, from the standpoint of science, topics full of dumb questions.

Where science is not engaged with these non-scientific questions, scientists are often falling victim to that pernicious scientific crime of defending scientific orthodoxy. The highest form of scientific accomplishment is a grudging acceptance of our inability to prove a hypothesis false. An inability to accept this limitation is an inability to tap into the power of science. Orthodoxy (of any kind) is the enemy of empiricism. Yes, it is "just" a theory.

This is not relativism or nihilism- there may very well be good answers to these questions. Just not in science.

Mitigating the Threat

Overcoming the insidious threat of a century of dumb questions will not be easy or fast. It will require far greater humility than offered in this critique, and a focus much narrower than any grand challenge. It will require scientists to ask good questions, and ignore bad ones.

A good question, scientifically speaking, is one that can be answered empirically. It is part of an academic ecology, which is in turn part of an even broader social ecology, but it does not take their place.

A good scientific question begs for more than a "yes" or a "no" answer. A good scientific question should give rise to a hypothesis. The hypothesis should be a model capable of predicting a data set analogous to the data set produced in an experiment. The experiment should be capable of producing, and reproducing answers. Perhaps most of all, a good scientific question must be relevant to the scientist.

This is not to suggest that scientists, or the scientifically minded, must not ask unscientific questions. "What kind of ice cream should I buy for dinner?" Or, "Is there a god?" But no right thinking person should not try to answer these questions with science.

Science s roundly reject those questions which parade as science, but are not, whether they come from within or without the scientific community. "Is there a God?" is not a scientific question, whether it is asked by Richard Dawkins or by the Creation Institute.

Finally, science must re-engage society. Scientists educates the public about the role of science, and the public provides science with a nicely formatted, rank-ordered list of important scientific questions. Maybe if the scientists actually answer a few of them, they'll stop getting beaten up.

Labels: culture, science, seed

Buffon's Needle is a fun little experiment that relates to physics, math, statistics, and the Monte Carlo method. Basically, you draw lines on the floor, spaced an equal distance apart, and you throw needles (of the same length as the distance between the lines) on the lines. The ratio of needles on a line to needles off of a line allegedly converges on twice-Pi.

I didn't have the patience to do this with actual needles, but I did make a little Processing Applet that simulates the experiment with reasonable success. There was some math involved in the writing of the applet, which (in my case) leaves a lot of room for errors. But the calculation does seem to head toward Pi most the time (in the display, I divide the ratio by two for visual effect).

Plus it's kind of pretty.

Buffon's Needle, in Processing.

A while back, I got roped into filming some stuff about robots for a television show. I don't even know if the show aired, but I do know they left me with a very nice Robotis Bioloid comprehensive kit, with which my kids and I have made a few robots. This video is from one of our first attempts, and shows my daughter inspecting the robot we named Shirley, and making a few adjustments. She's about eighteen months old in the video.



Both kids taught me pretty early that they can take part in all manner of geekery, as long as I'm willing to be patient and to include them. We're working on beefing up the processing capabilities of the Robotis kit, which will hopefully be featured here again soon.

A while back, the Dorks at Dorkbot DC had a little soldering party. For some of us, this resulted in cool 3x3x3 LED Cube. For all of us, this resulted in a shiny new Arduino Dicemila. The Arduino Diecimila is a nifty prototyping board (http://www.arduino.cc/en/Main/ArduinoBoardDiecimila), and stringing together 27 LEDs couldn't have been easier.

Since there are plenty of good photos on Flickr, I'll spare the Internet the indignity of yet another LED cube. However, I did write a little code module for mine, which had the two benefits of easily re-arranging the order of the output pins (so I didn't have to pay attention to this when I was soldering), and of running a cute little test pattern.

So, if anyone has stumbled across this page looking for code to modify for an Arduino LED Cube, stumble no further. If anyone wants to see my LED Cube in action, or needs help soldering one of your own, I'm always game for dorking around on Capitol Hill in DC. Email me!