Too much thinking, too little writing

[marcusmarcusrc]

When do we know something for certain? Who gets to decide?


When it comes to evolution (see several recent articles in the New York Times), I feel like all the important parts are known well enough that I am comfortable requiring schools to teach natural selection as the only credible mechanism for explaining the origin of species. Sure, we can’t prove that the Flying Spaghetti Monster did not tweak a mutation here or there with His Noodly Appendage, but there is as yet little need to call upon him for explanations. If someday we find a “babelfish” equivalent, or 100 years from now we have yet to find convincing paths leading to the so-called “irreducibly complex mechanisms”, then there may be a reason to introduce this stuff into classrooms. But for now, the evidence for anyone who knows even a little molecular biology is pretty overwhelming. But given that I have not yet become God-Emperor, who gets to decide that the evidence is overwhelming? Our elected officials? Clearly (Bush, Frist, Santorum, Inhofe) not. A panel of scientists or teachers, like the NAS, AAAS, NATB? Well, they would likely arrive at the “right” answer, but this sort of decision making process seems prone to leading to perceptions of an “elite” class separate from the populace and therefore a general distrust. In addition, leaving out stakeholders often leads to solutions that implement very poorly. Do we let the states decide, and hope that the shift of biotech companies to pro-evolution states will eventually cause the ID states to relent? But then are we abandoning thousands of students to poor education systems just because they were born in the wrong place? Maybe the court system can help, when parents sue to take God out or put Her back into the classroom? But my time on a debate team has always made me doubt the adversarial system, though at times professors and lawyerly friends have almost convinced me that it is still better than most of the alternatives.

So if this much controversy can be raised about evolution, which is close to a slam dunk case, what hope do we have for reasonable solutions to climate change, secondary cigarette smoke impacts, health effects of fine particles, asbestos, Vioxx, missile defense, DDT, vaccines, breast implants, ?

Climate change is obviously the one I know the most about. Yet my name is on a paper stating that our model shows that the 95% bounds on the warming in 2100 are 1.0 degrees C (pretty acceptable, as long as you aren’t an alpine ecosystem or an especially vulnerable coral reef) to 4.9 degrees C (a pretty disastrous outcome), and being intimately involved in the process I know that there are many potential sources of uncertainty that are not included in our model as well as things we just don’t understand/can’t predict. This makes it easy for politicians to stand up and fight against hard constraints until ‘sound science’ (oooh, how I hate how that phrase is used) comes to a consensus. And there are always a couple of maverick scientists you can quote for any dispute (Behe for evolution, Singer/Christy/Lindzen/Michaels/Mcintyre/ for climate change, etc.) to ensure that ‘consensus’ is never reached.

Theoretically, I went into technology and policy because I felt like we needed more scientifically literate people in the policy arena in order for good decisions to be made, but the longer I study how this all works (or doesn’t work), the more I feel like I’m sinking into a morass of confusion, weasel-words, and uncertainty. Though perhaps for now I should just focus on writing a couple hundred pages and getting a degree, and leave solving the world’s ills for next year...

Comments

[galadhelsul.livejournal.com]

I haven't seen many predictions that actually come from the idea that all life comes from a small number of common ancestors. There is a lot of circular reasoning that basically amounts to "similar species are similar to each other", but few actual predictions. Currently, the family tree of species is constructed based on observed similarities. This prevents that family tree from having any actual predictive value, because the predictions that you could make given the family tree are the observations that were used to construct it. An example of a prediction that could be made, but as far as I know, has not been made and verified would be "given the existence of this set of similar species, there should be this other species we have not yet observed with these characteristics (either now or in the past)." Similar to how some elements of the periodic table were described before they were ever found (or made).

Creation and evolution both have the same fundamental problem, which is that they allow anything to happen. Neither one is really science, because neither one can be disproven. In order for an idea to be science, it has to be possible to describe a body of evidence that could theoretically be produced that you would accept as disproving the theory. The easier way of stating this is that science is all about making testable predictions. If the predictions are correct, the science is correct. If they are incorrect, the science is wrong and needs to be modified. If the predictions are either untestable or so vague as to be meaningless, it isn't science. Predicting that some species will be similar to other species is too vague to count as science. Predicting the exact nature of how they will be similar would count as science, but only if the prediction is made before it is checked.

Evolution has a further complication towards counting as science because much of the data that would be used to prove/disprove minor differences in competing evolutionary theories (the fossil record) is very incomplete. As such, nobody would accept an unverified prediction as actually disproving evolution. This means that a small number of correct predictions is not sufficient to prove evolution. For evolution to be accepted as science, a large percentage of its predictions must be correct.

[marcusmarcusrc.livejournal.com]

Well, it isn't circular because the majority of the family tree was drawn up before PCR made molecular biology easy. So the fact that the family tree derived from DNA sequences almost exactly matches up with that derived from observed phenotypes is really a stunning confirmation of common ancestry. (There are some exceptions. Panda bears, for example, were considered to be more related to raccoons than bears based on skull shape. This is what I was taught in elementary school. In the mid-90s papers were published showing that in fact, panda bears are more bear-like than raccoon-like, barely. So a minor modification to the tree is made such that panda bears branch off from the bear line just after the raccoon line branches off).

Furthermore, based on the fossil record and dating techniques, you can get a general time estimate of divergence of two species. Then with an estimate of what the mutation rate is, you can predict how different non-coding or non-conserved regions in the two species would be. If you want to quibble about whether or not those regions are really "non-coding" or "non-conserved", then you can look at the "redundant" coding where two different codon sequences code for the same amino acid. When scientists have done this, the fossil record and the quantity of DNA divergence always (to the best of my knowledge) match up.

If you prefer non-DNA based predictions, from what I've read (and this gets further away from my area of expertise as a former biochemist) they have found whale fossils matching predictions that whales derived from a more land-going animal in the distant past. Similarly, I think fossils have been found matching predictions in horse lines and in bird lines. But really, I think molecular biological proof is where it is at.