Guest Post: Response to Follow-up
by H. Kern Reeve
A response to Mark Psiaki’s follow-up, posted here
A. The first flaw, if I understand what was said last night, concerns the relationship between the predictions that Prof. Reeve makes in his work and the validity of the theory of macro evolution. He implied that the two were intimately connected. In other words, he seemed to be saying that he could not have made the predictions that he made without the model that all of the complex forms of life, as we know them, arose from a single cell or set of single cells through much repeated application of the types of processes that he is studying. This is a non sequitur. It seems obvious that Prof. Reeve is studying and making predictions in the field of micro-evolution or even, one might say, in the field population dynamics. To say that, given a certain genome or given a certain predisposition to a certain type of behavior, we can predict a certain relative frequency of the sexes in a certain species is not to say that we can explain how that species came to be what it is, to have the genome that it has, to have the behavioral predispositions that it has.
Kern: As Brian Barringer noted in a rebuttal he forwarded to me, I simply wanted to show that the theory of evolutionary change through natural selection makes precise predictions about the phenotypes that organisms exhibit.
ID does not make such specific predictions, because there is no logical apparatus for generating them. Moreover, their appears to be no hope for generating this logical apparatus, as Brand and Hunter and others have argued that we cannot know the mind of the Creator. If we can’t know the mind of the Creator, then we can’t know the design criteria, and so we have no idea what counts as an observational match or mismatch to predicted designs. However, I felt in your and my brief interaction after the panel discussion that you actually agreed with me on the latter points (against Hunter), so I want to return to your criticisms in this email message below.
Again, to make an analogy, one could call the science of apples falling "Micro-Newtonian mechanics" and the science of the planets orbiting the Sun "Macro-Newtonian mechanics." For Prof. Reeve to argue the way he argues about his research supporting macro-evolution would be the same as arguing for the acceptance of Newton’s explanation of the motion of the planets by doing predictions and experiments about the falling of peaches, bananas, and cherries. Suppose Newton had proceeded in this fashion. Suppose that Newton had merely reasoned from the falling apples to other types of falling fruit. He would never have bothered to develop calculus and to use it to solve the difficult 1/r^2 nonlinear differential equation that was needed in order to prove that the apple falling and the planets orbiting were one and the same thing. Suppose that he claimed that his successes with the other fruit constituted sufficient reason to accept his proposition that he had explained the orbiting of the planets around the Sun. If that had happened, then Newton’s name would not be a household word and I would not have the job that I have today.
Prof. Reeve may believe in macro-evolution, he may do good micro-evolutionary science or population dynamics science, and he may believe that his success in the lab or at the computer somehow support his macro-evolutionary beliefs, but his beliefs about this connection do not constitute an actual connection. Prof. Reeve correctly disputed Prof. Hunter’s argument that Newton’s success with F=m*a lent credence to Newton’s belief that F=m*a is true because there was an intelligent designer behind who decided that F should equal m*a. Prof. Reeve’s excellent point is that one can use F= m*a quite successfully without needing to believe that it arose because of intelligent design.
Kern: The first preliminary point is that there is not a precise distinction between micro and macro evolution– exactly how many generations must elapse in the divergence of two biological forms before we have made a transition between "microevolution" and "macroevolution"? The terms are used heuristically to denote "shorter" versus "longer" time scales, but there is no well-defined break between them that corresponds to the observational difference between objects falling vertically from trees and objects revolving in ellipses around the sun. Thus, there is no reason to demand a different set of equations for the two (ill-defined) realms of microevolution and macroevolution, at least with regard to the question of why species look and act the way they do.
Now I suspect that you are hinting that there might be such a "break" at the moment that two species lineages split off from each other, where speciation involves the evolution of some kind of reproductive-isolating mechanism between two populations or two forms within a population. But all that happens here is that one arena for selective competition between alternative genes splits off into two arenas in which the mixtures of competing genes may now be different. The formal equations of gene frequency change due to selection are unchanged– it is just that now these equations must be separately applied to the two newly disjunct populations (and the parameter values within the equations may be different). (By the way, speciation has been observed both in nature and in the laboratory, so it is not that speciation requires long time scales and thus is purely hypothetical.)
Fortunately, all of what I say above is testable: If evolution though natural selection has continually sculpted the macroscopic features of organisms, then we should in principle be able to predict and explain all macroscopic features of all organisms, because all such features should be those that promote the long-run competitive survival of the genes encoding those features (a very specific criterion that follows deductively from the population-genetic theory of evolution by natural selection). (I pause to note here that there are competing theories from within evolutionary biology, namely theories which emphasize other evolutionary forces such as genetic drift- these are the theories that are pitted against selection-based theories in empirical tests within evolutionary biology. These tests are part of what makes evolutionary biology so satisfying and exciting!)
Thus, the selection theory doesn’t just predict why species A looks or acts the way it does versus some alternative imagined way (ie., that, in its environmental context, selection favored genes with specifiable phenotypic attributes). The theory automatically also predicts or explains why species A looks and acts the way it does whereas phylogenetically distant species B looks or and acts quite differently. That is, the theory predicts or explains differences that must have involved macroevolutionary time scale divergences. (It also predicts or explains why two distant species might nevertheless look the same or act the same in some respect!)
Now, suppose that we have missed something that causes the way selection acts to change fundamentally over long evolutionary time scales. That is, suppose the evolution through selection that we can observe directly over short time scales (bacteria in chemostats, Drosophila in artificial selection experiments, etc.) and which supports evolutionary selection models, somehow misses something about how selection operates (or is inhibited) over longer time scales. Then we will detect this from the repeated failure of an ability to predict or explain the current features of organisms. Moreover, the fit will be worse for species which have been around for longer periods of evolutionary time (e.g., as judged by fossil or genetic evidence). Thus, if current selection theory is WRONG because it omits this mysterious time-scale-dependent, selection-altering process, then our predictions will not only fail but also fail in predictable ways. In other words, it is testable! If the predictions frequently do fail, then we would have to seek the equivalent of Newton’s higher order differential equations for long time scale evolution. So far they haven’t failed. These kinds of tests are nowhere in sight for Intelligent Design, by the way.
Isn’t it true, however, that one can believe in intelligent design rather than macro-evolution and still do all of the same kind of laboratory and computer work that Prof. Reeve does with equal success?
Kern: Yes, you can believe anything that has no empirical consequences and still do all of the same kind of laboratory and computer work that I (or any other scientist in any other field) does with equal success. Of course, that you can do this provides no evidence for the belief nor does it transform that belief into a scientific theory (but I think you agree with this, right?).
Based on what I heard from Prof. Reeve last night, I believe that I could do his kind of work, should I learn the necessary background material, without ever having to accept his supposition that all of the different biological forms came into existence by purely naturalistic processes.
Kern: I am not sure what a "naturalistic process" is. How are they defined and how do we detect them and distinguish them from "supernatural" processes? (Hunter couldn’t define them.) Those terms play no role either in my deduction of predictions from theory or in my empirical tests.
On the other hand, Prof. Reeve could never do the type of rocket science that I do if he did not accept that F = m*a. Thus, my ability to make successful predictions in the lab and at my computer constitutes real proof that F= m*a (that is, in the low-speed, low-mass limit that does require Einstein’s theory of relativity),
Kern: Agreed! Well, it constitutes strong support, but not really "proof."
but Prof. Reeve’s success at prediction does not constitute proof of macro-evolution.
Kern: We have to be precise about what is meant by "proof of macroevolution". Referring to the argument above, I am arguing that testing the selectionist theory of organismal design (and finding support) across taxa (both closely and distantly related) does provide support for the hypothesis that selection within lineages shaped organismal phenotypes on both short and long time scales. Separate streams of evidence from paleontology and evolutionary genetics provide support for the historical hypotheses of phylogenetic connection and that the lineages branched in specific ways. (Note that evolutionary theories of history and evolutionary theories of organismal design are directed at explaining different things. They are of course coupled because the former supplies independent evidence the genetic continuity assumed by the latter. However, the selectionist design theory could be tested even without this independent evidence.)
B. The second flaw in Prof. Reeve’s presentation about prediction in science is that he demands too much of prediction. A theory has the right to define what it will predict. Its critics do not have to right to demand that it predict whatever they want to predict.
This touches on a very important subject in science, which I will call the humility of science. Good science has a sense of humility. It says that science can answer some important questions about the physical world in which we live. It does not say that it can necessarily answer any and every question that someone might like to ask about our physical world. Science has been very successful at answering some questions, and the number of questions that it can answer has grown steadily. Many people have mistakenly (sometimes arrogantly) misconstrued this progress to imply that science can answer any question than anyone might want to pose. The truth is that we do not know the limits of science, but it would be wisest to suspect that it may have definite limits. It appears that cosmology has certain limits when it looks at the early stages of the big bang, and certainly it can say nothing about anything that might have happened before the big bang. Behavioral sciences seem to have clear limits too. Why have we made huge progress on treating heart disease and minimal progress on treating mental disorders? One can’t help but wonder whether we are coming up against a limit of science. Of course, the people with a vested interest in the research and clinical treatment dollars associated with mental disorders will never admit that there are fundamental limits to their science, but the rest of the public suspects that they are not to be totally trusted on this matter.
Kern: As Brian Barringer pointed out, and I hoped to make clear, I was seeking ANY prediction from ID theory. I heard one that was quickly retracted when it failed. Much worse, we weren’t given any hint of the logical apparatus for extracting predictions, even in principle. (There is no accepted scientific theory that does not make predictions, contrary to what Hunter wanted to say.)
I have a very different view of the limits to scientific knowledge. It would seem a bad strategy to ever assume that we are ever at the limits (given that we can never really know where they are or if they even exist). Suppose we sensed that we are near the limits: what do we do– slow down our inquiries? If we are wrong, we will cut ourselves off from future insights prematurely. I think that scientists have often claimed that we are near the limits, either because we know everything or because there is no way to proceed forward, and in all cases they have been proved wrong. Personally, I think we are just beginning to learn about the biological world, and I discern no looming limits to knowledge about it.
As an example of the limitation of science, consider quantum mechanics. One of its fundamental tenants, known as the Heisenberg uncertainty principle, is that certain questions of classical physics are unanswerable at the atomic and sub-atomic level: one cannot know exactly the position and velocity of a particle. This principle drove classical physicists, including Einstein, mad. They hated it. They were used to assuming that such questions could be answered, and the new quantum mechanics folks came along and told them that they had to stop asking the usual questions. This was purely a negative result, yet it was a key to making advances in this area.The theory of intelligent design, or put better, the assertion that there exists irreducible complexity in certain biological mechanisms or biochemical processes, is similar. It makes few predictions.
Kern: The "theory" of intelligent design does not predict irreducible complexity. It doesn’t predict anything. How do we know that a creator wouldn’t prefer to produce life without irreducible complexity?. (After all, it is vulnerable to catastrophic failure!) The hypothesis that irreducible complexity cannot be arrived at through step-wise natural selection isn’t derived from anything.
It should be admitted this hypothesis is having severe difficulties already, given that many biochemists have shown how irreducibly complex biochemical cascades (a prime example) could be built up by gene duplication of autocatalytic biochemical pathways.
Its principle prediction is that there will never be found a naturalistic descent-with-modification (i.e., natural selection) explanation for how these irreducably complex systems came to be. This is a negative prediction, and many evolutionary biologists don’t like its negativity.
Kern: It’s not that they don’t like the hypothesis because the hypothesis is negative; it’s that they don’t like it because they think it is false (see above).
It is a prediction, nonetheless.
Kern: From what?
It does not give power to predict about the sex ratios in certain populations, as Prof. Reeve would like it to, but that is not a problem, because it did not claim that it would make such predictions.
Kern: I never said that the "theory" of irreducible complexity should be able to predict sex ratios; I said that the "theory of intelligent design" should be able to predict sex ratios, since presumably the creator designed everything in biology and the universe as a whole, for that matter. If ID has limits to what it can predict, then these limits should be described and explained.
________________________
H. Kern Reeve is a professor in the Department of Neurobiology and Behavior at Cornell University
It looks like we need a fair bit more discussion of this. One point which I think is relevant is the one which Wiglaf makes here.
Though he never does it explicitly, how is Reeve defining intelligent design theory in this exchange?
Comment by Freawaru — April 11, 2006 @ 4:18 pm