Thursday, October 15, 2009

The Nature of Explanation

Before I get started on hierarchical explanations, here's a list of the previous substantive posts in this series:

The Truth about Evolution
The Nature of Science
The Nature of Evidence

Today, I want to look more carefully at the way scientists think about their own theories and arguments. Remember what you learned in school about the scientific method? How we start out making observations, from which we formulate hypotheses? Then hypotheses that are tested successfully become theories, and theories that go through rigorous testing become laws. Right. So that's pretty much bogus. I don't know anyone who really follows that pathway in doing science. Instead, what I find is that observations are interpreted within some kind of theoretical framework, a higher level model that informs the interpretation of data. The interaction of high level models and data lead to the formulation of theories or hypotheses that make sense of both. High level models can be held for any number of reasons that might be philosophical and/or quasi-theological as well as empirical. Low level theories must not contradict the high level model and must correspond to the data.

Take for example the idea that species evolved from a common ancestor. This is a very simple high level idea that lots of people thought of before Darwin. There are several observations that support this model: (1) the "progress" recorded in the fossil record, (2) comparative biology (e.g., comparative anatomy, comparative embryology, etc.), and (3) the biogeography of species of high affinity. If you like, you could add the philosophical "consilience of induction" of common ancestry to explain many types of data. In my estimation, and I think most evolutionary biologists would agree, these evidences are sufficient to support the common ancestry of all species.

In Origin, Darwin laid these evidences out in a very compelling fashion, but in doing so, he often drew comparisons between his own model and what he called "special creation." These comparisons were intended to illustrate the explanatory superiority of his model over the idea that God created species independently. Sometimes these comparisons were effective, and sometimes he got his logic a little muddled. The appearance of these arguments in Origin gives the impression that Darwin had formulated his theory as an anti-God hypothesis, which Darwin himself denied. In any case, it's important to realize that he had a compelling argument without the theological comments. Religious considerations might have influenced his thinking and arguments, but they certainly did not drive his science.

Now here's where it gets weird. By itself, the idea of common ancestry is somewhat useless scientifically. Common ancestry is a sort of nebulous idea without more empirically-driven hypotheses (low level theories) to connect it more directly to data. For example, how do species evolve? Do they simply adapt to their circumstances as Lamarck believed? Is there an internal drive to adapt? Do they evolve neutrally by genetic drift? Or is natural selection involved? All of these ideas have more or less specific empirical consequences. In other words, they're all fairly predictive, whereas the high level idea of common ancestry is not.

Likewise, we could look at the pattern of evolutionary inheritance. Do species evolve one into another wholesale, resulting in an unbranched linear pattern, like the great chain of being? Or do species diverge and branch, giving a pattern that looks more like a tree? Once they've diverged, do they ever come back together again, as in the case of hybridization? Again, these ideas are low level ideas close to the empirical data but separate from the higher level concept of common ancestry.

I think you've already figured out an important corollary to this line of thinking, that squabbling over low level theories has very little to do with the acceptance of the high level model. Scientists can argue over the effectiveness of natural selection without ever questioning the high level concept of common ancestry. Likewise, new fossils can be heralded as "changing our understanding of evoluion" in the sense that low level theories have to be modified. The strength of the high level model is relatively impervious to these changes.

This can be seen in history also. Darwin only partially triumphed in his lifetime. He convinced most scientists that common ancestry was true, but the majority of his friends and colleagues remained skeptical of natural selection. In other words, the evidence Darwin presented to argue for the high level model of common ancestry convinced most readers of Origin, but they generally disagreed about his low level theory of how new species originate. In this way, I find it unfortunate that some creationists today doggedly attack the neodarwinian synthesis, a particular low level theory to explain the origin of species, as if the acceptance of the high level model of common ancestry depended on it. I assure you that the evidence of common ancestry does not depend on any particular low level theory of the mechanism of evolution. The history of science between 1859 and the 1930s bears this out: Nearly everyone accepted evolution, but few agreed on how it happened. Even if the neodarwinian synthesis could be discredited (and that's a big if), acceptance of common ancestry would be unscathed.

I hope you can see better now how I think about creation. I accept the basic creationist dogma for reasons of faith (which will be discussed in a future post) and a some empirical observations. What is largely lacking from creationist biology are the low level theories that connect the model of creation to the empirical data. For example, Genesis 1:21 tells us that "God created the large sea-creatures and every living creature that moves and swarms in the water, according to their kinds." What does that mean? Like common ancestry, it's definitely making an empirical claim, but it's not one that is immediately relevant to science. What is needed are low level theories that are close to the empirical data and consistent with the high level model of creation. In this case, we might ask whether the modern species correspond to the same species created by God, or have the species changed in the time since God created them? If so, how much change has happened, and how have the changes occurred? As I see it, we need five low level theories in creationist biology: design, imperfection, speciation, systematics, and biogeography. I've covered these ad nauseum in this blog, so I won't bore you with more explanation. I think you get the point.

The same corollary about disputation applies here: Just because I reject certain creationist claims (even claims that are widely accepted and popular) does not mean that I reject creationism. We have to stop treating all creationist claims as being equally valuable. They're not. Creation, Fall, Flood, and Babel are essential ingredients to a young-age creationist. Acceptance or rejection of the "Canopy Theory" or speciation or the geologic column should not be litmus tests for creationist orthodoxy. That means all accusations of our brother and sister creationists suffering from compromise, brainwashing, and the like must stop.

Finally, please understand that this two-level hierarchy I've described here is a vast oversimplification. There are many different levels of explanation between purely theoretical ideas and the empirical data. Each biological discipline has its own models that drive the interpretation of data, some of which continue to be debated.

I think that's all I want to say on science. Next time, I need to deal with the nature of faith. Stay tuned!