8/17/2007

VOX DEI BATE #5



















Time prevents me from a full comment on Vox's most recent set of observations. I look forward to his discussion of my previous post, which compares two predictions concerning whether intermediates to modern whales would ever be found in the fossil record.

In the meantime, Vox writes:

"....we should have no problem finding one or two and demonstrating that they are a new species that did not previously exist, assuming that this is, in fact, happening."

It's been done, by Dobzhansky's group, in the early 1960's: you can read about it here. There are other examples, though the literature in this case is not as deep as the pallid piles of papers in Keynesian macroeconomics which seem to trouble Vox. That's because it's clearly science, and real science is hard.

(UPDATE)

Now, again, (and Vox has acknowledged that this is valid, I think) I don't want to get sidetracked in a 'turf war' between economists and biologists, because it's in no way central to the topic. But Vox is at pains to explain, based upon his experience, the relevance of the comparison in his original argument:


From what I've seen of evolutionary projections, its margin of errors are even greater than those seen in macroeconomics. Since this level of inaccuracy causes me to dismiss macroeconomics as a science, it leads me to similarly dismiss evolution. Obviously, if it can be shown that the macroevolutionary margin of error is smaller than that of macroeconomics, I'd be happy to reconsider. (my emphasis)

I think there's a basic error here: the map is not the territory! Vox is not arguing here that the theory of evolution by natural selection (TENS) has been falsified generally. Rather, he's saying that 'evolutionary projections', models based upon TENS, have a relatively low precision, which presumably leads to a certain level of inaccuracy. Well, that's true for a lot of things in science, not just specific models based on TENS. A high margin of error is not sufficient to disqualify a research program as unscientific, however, unless it can be shown that a different research program based upon a different model has a lower margin of error.

An example from the history of science should make this clear:

Johannes Kepler, like those astronomers who came before him, began with the belief that planets moved in perfect circular orbits. Problem: this hypothesis didn't fit the data, notably the 'wanderers' (planetai) that seemed to change direction. Like Ptolemy before him, Kepler trusted in epicycles, 'wheels within wheels' that would preserve the hypothesis of perfect circular motion. These models worked with a certain margin of error and, for most heavenly bodies, the discrepancies were so small once the epicycles were added that most felt they could be ignored---except, in Kepler's case, for the orbit of Mars. An exhaustive series of calculations based upon the best data available (Tycho Brahe's observations) led to a closer fit to the Ptolemaic model, but still with a certain margin of error that Kepler found unacceptable:


If I had believed that we could ignore these eight minutes, I would have patched up my hypothesis accordingly. But, since it was not permissible to ignore, those eight minutes pointed the road to a complete reformation of astronomy.

So, by analogy, Vox is in a similar predicament to Kepler at this stage of analysis. He's unsatisfied with the error rate, and---feeling confident (perhaps overconfident) that he's got the best data available, concludes there must be something wrong with the model. But was that really enough to unseat Ptolemy, whose properly-tweaked model still gave better 'predictions' than any other model up till that time? Of course not! Kepler was obliged to propose another model and show that it gave better results, with a lower margin of error.

Which he did, one based upon ellipses rather than circles, and in the process Kepler discovered some generalizations about the behavior of heavenly bodies which presaged Newton and which have since been described as his 'Laws of Planetary Motion'.

But the story doesn't end there! Both Kepler's Laws and Newton's Law still maintain a certain margin of error, albeit one lower by several orders of magnitude than, say, geocentric solar systems or Archimedes' statics. But the existence of these discrepancies did not automatically lead either of the former constructs to be rejected as science! These 'Laws' came to be seen in the 20th century as limiting cases of Einsteinian mechanics, whose margin of error becomes significantly lower at certain scales. In fact, no one would've been willing to embrace Einstein's thought, with all of its counter-intuitive moves, had there not been a 'trade-up' in precision!

So, Vox, you can play the part of evolutionary skeptic all you want to, but you can't expect the scientific community to join you in rejecting TENS on the grounds that you've given (its high margin of error) unless you can do what Kepler did and propose a testable explanation whose margin of error is significantly lower.

Now, Vox, I am guessing that you don't reject Keynesian macroeconomics out of spite or prejudice, but that you honestly feel that there are other models that do a better job of predicting economic behavior. That's not my bag, and again I'm disinterested in provoking a 'turf war.' What I want to know, Vox, is if you or anyone else has a better explanation for the history of life than TENS, a model that yields falsifiable predictions that, when observations become available, have a lower margin of error than TENS. Because, if you can't, then your skepticism remains an intuition, rather than sufficient grounds to reject evolutionary theory as science.

8 comments:

Anonymous said...

Does this evidence of the generation of new species correlate with the fossil evidence of Drosophila evolution?

--emerod

Scott Hatfield . . . . said...

I'm not sure if I understand what you mean. This is a recent event, documented by Dobzhansky's team based upon work done in the field between approx. 1955-1965 by various teams of scientists under his direction. We don't need fossils to infer that a speciation event has occurred here, we have direct evidence of such an event in progress---I might add that we wouldn't call a 50-year-old artifact a fossil--we would call it 'remains'. Substitution of carbonates and other evaporites for organic tissue would be far from complete, etc.

If you mean, is this the sort of change that we would expect documented by the fossil record? Yes, if you include stasis---no, if you only want to include fossil sequences that demonstrate morphological change. The 'Llanos A' population in the studies I cited were well on their way to hybrid sterility with some, but not all populations of Orinocan fruit flies. Once isolated, they are free to go their own way and in the presence of different environmental effects are expected to eventually display not just behavioral differences (which is what triggered the initial isolation and hybrid sterility) but also morphological differences which, if present in a fossil sequence, could be used to infer evolution.

It's important to realize that this modest correlation in one example would inspire only interest. It is the cumulative effect of similar correlations in the hundreds of thousands of lineages preserved in the fossil record that inspire not merely interest, but confidence that evolution has occurred and that living things share common descent, etc.

Anonymous said...

I meant, "Does this type of speciation event correlate with the types of changes shown in the fossil record?"

I guess I was hoping to see documentation of a speciation event of the type evidenced in the fossil record.

--emerod

Scott Hatfield . . . . said...

OK, I think I understand what you're saying now. Help me out. When you write that you were "hoping to see documentation of a speciation event of the type evidenced in the fossil record", I think you're expecting to see a sudden dramatic change in body form, as when we line up two fossils with similar features, but obvious morphological differences.

You've been misled, if that's the case. When we line up fossils A, B and C as a sequence showing an obvious series of morphological changes, we are not claiming that A became B, then B became C. What we are claiming is that we see transitional features within the sequence. B might not be descended from A. It could be descended from a relative of A's!

Now, creationists typically emphasize the fact that we haven't actually demonstrated any direct connection between fossils A and B, or they talk about the gaps between A and B. If we later find another fossil specimen that seems to have transitional features between the two (call it A*B), they'll still talk about gaps: "Where's the linke between A and A*B, between A*B and B?" It never ends.

Anonymous said...

Goddidit!

Sorry, I don't believe that model. You have an immense amount of patience. I guess Vox' argument is one of the key components of creation magicians: picking out at flaws inadequacies yet-to-be-discovered-aspects margins-of-error rather than bringing the evidence for their own models. Vox seems very evasive about his own theory that would Explain Everything.

I love your example; a scientist back when the only model available had a margin of error would use that model as the best one there, not that it should be ditched for an even worse one for only being a partial explanation.

Anonymous said...

Your update is an impressive defense of abductive reasoning, Scott. I predict that Vox will falter on this point. Not due to your example, necessarily, but simply because I think the process of abductive reasoning has not been worked out fully enough to enable a simplistic dismissal of it; therefore Vox will be required to wade in without quite knowing how deep the water is.

When we line up fossils A, B and C as a sequence showing an obvious series of morphological changes, we are not claiming that A became B, then B became C.

That would sound so mechanistic and predictable, like a children's fairy tale. I know that no rational explication of TENS would ever present it like some kind of definite sequence of discrete events endowing progressively increasing complexity on the organism. How silly!

What we are claiming is that we see transitional features within the sequence. B might not be descended from A. It could be descended from a relative of A's!

Surely, life is like a fractal bush, not an infinitesimally narrow one-dimensional vector: it never ends. You seem to leave little hope for ever nailing down what "really" happened! Now my positivistic self is totally depressed.

--emerod

Cubist said...

Your "margin of error" discussion reminds me of a joke...

There were these two campers sleeping in a tent in the woods. Suddenly, they were awakened by a noise -- the growl of a ferociously hungry bear! Both of the campers knew it was time to get the heck out of there, pronto. So one camper began to frantically gather his stuff... and the other one started putting his boots on, very carefully.
The frantic-gathering camper said, "What are you, crazy? You're never gonna outrun that bear!"
The boot-wearing camper replied, "I don't need to outrun the bear. I just need to outrun you."

Vox obviously thinks a valid scientific theory has to outrun the bear; in reality, a valid scientific theory just has to outrun its competition. And if there isn't any competition...

Drew said...

Half of his questions and observations don't even really make sense: they don't demonstrate any real attempt to learn what the key terms and issues are so that he can criticize their usage.

The question about the average rate of evolution is particularly silly. It's like asking what the average rate of organ growth is. Uh, what? That question is too vague to be useful. We COULD get an answer, by arbitrarily defining all sorts of ambiguities (in what species? In just adolescence or a lifetime? A normal lifetime? What do we count as an organ: can we count skin?), but the answer would be an utterly meaningless number, telling us nothing much about the specifics of organ growth.

I think what he meant to ask was the commonly observed ranges of rates of morphological change, which has in fact been measured (in "Darwins," goofily enough), and is many times faster than any morphological transition seen in the fossil record (which is why natural selection seems in practice to primarily slow down rates of change rather than speed them up). THAT question at least has some meaning and has an answer that is useful for determining the plausibility of what the fossil record shows.

It's also amusing that he simply stutters in anger that you would suggest that the evidence for evolution is more complicated than fossils.