Evolution As A Theory

Your final question is interesting. It sounds like the Deist belief that Providence created a universe of laws and set it in motion, then withdrew from the scene. As Prince once said: "In this life, you're on your own."

This makes an interesting middle ground in the debate over evolution, one that allows for both a creator and the overwhelming scientific evidence that the universe is both old and exceedingly spooky. I find it interesting that Deism rang true for many of the Founding Fathers, but it's apparently baffling to many Americans today. It really needs a snappy jingle or something.

Jim,

Thank you for the insight, I agree.

Towards that, just after posting my comment here, I posted Creation verses Big Bang

I hope you appriciate it.

Explaining how life began is a very minor part of evolutionary research and actually belongs more in the field of biochemistry, from which most of the researchers come. As for evolution dealing with only historical changes, I hope that I have made it clear that evolution has been documented in progress - not only abundant cases of microevolution (animal locomotion traits, bacterial drug resistance, and more), but also macroevolution of new species, such as the apple maggot. Moreover, causation of historical change can still be elucidated in part. In evolution, we see the traces of past selection events. The interested reader can go back to my second posting (discussion) where I refer to the evidence in the ratio of DNA base changes or mutations that change a protein's function (nonsynonymous substitutions) to those changes that do not change the function (synonymous substitutions). The same ability to trace changes backward is true in many areas of physics, from such simple mechanics as ballistic analysis of where a bullet or ballistic missile was fired from, to, more dramatically, the many signatures - rays, particle fluxes, etc. - of which nuclear reactions have provided the sun's power. We are likely more certain of many evolutionary events such as the times and natures of species divergences than almost everyone on the planet is certain of the identity of his or her father. None of us were present at, conscious during, and aware of the act of our own conception, but most of us feel pretty confident about who is our father (too much so, for something on the order of 1-30% of all births, as DNA testing has revealed**).

Again, evolutionary theory as a sum of knowledge of biological complexity does not stand or fall on the basis of its ability to explain the origin of the first cells, any more than electromagnetic theory stands or falls on the basis of explaining the charge and mass of the electron. The latter deep intellectual 'mysteries' are still being explored in great extensions of theory - and experiment - that overlap and complement classical electromagnetic theory that makes all our radio communications, lasers, and PCs work so well.

A thread is developing in this discussion, and also in a comment, about the nature of the scientific process. First, I may address the comment by a mathematician, Bill Ott. In math, a hypothesis or conjecture must submit to complete logical proof in order to become a theory such as set theory. It need bear no relation to any external reality. In science, proof is based on observation of reality. Proof is never absolute, but consisting of a great preponderance of evidence. Intellectual schema that have overwhelming evidence and that can serve as a framework to deliberate all other evidence are called theories, such as quantum theory, electromagnetic theory, and, of course, for all of biology, evolutionary theory. Even then, a theory is subject to revision, which is generally an extension, not a replacement. Classical radiation theory was extended by quantum theory, by Max Planck and others, to explain the pattern of radiation by hot bodies and ultimately even to predict the existence of masers and lasers. The classical theory still works with great accuracy, unamended, in the optics of lenses, the design of communication towers, and much, much more. This theory has a large realm of validity (high accuracy), and it is overlain by a broader theory that is more demanding to apply to most inquiries or applications and for negligible gain. One does not use a sledge hammer to crack a nutshell, though one needs it for bigger jobs.

The discussion post by Mark Andersen opens up another topic, the nature of the scientific process in developing theories of how the real world operates. The classical model of the process is that of the controlled experiment. The scientist imposes a treatment on a system and observes the responses, comparing them to the behavior of a similar 'control' system (as nearly identical as possible) that does not receive the treatment. The differences in behavior of the treated and control systems are then compared to those predicted by the scientist's hypothesis. Another model is that of the observational experiment, such as Darwin's initial construction of the evolutionary process. An observational experiment is the choice when a treatment may be impossible or impractical. An example is the suite of LBA experiments by the FluxNet collaboration to measure the fluxes of CO₂, heat, water vapor, etc. from the Amazonian rainforest. Clearly, one cannot impose a treatment on any significant area of Amazonia in order to use the classical model of a controlled experiment. Rather, one might use a time series of observations of fluxes from various sub-regions and the simultaneous state of variables that are hypothesized to exert control over these fluxes - e.g., climate (water stress, or drought; temperature) or land clearance by settlers. A hypothesis is generated for the relation of the fluxes to the driving variables. The hazard here is that one may be observing correlations, not causation. For example, there is a reported correlation between the number of sheep in New Zealand and the number of television sets. However, the sheep are not demanding to watch TV and driving the purchase of TV sets. Rather, both numbers are related to the size of the human populations. There are several ways to improve, reaching toward the resolution of causation. One is to measure as many intermediate processes as is possible, such as leaf temperature that is partly driven by weather and also partly controls the rate of water release (leaf transpiration) or CO₂ uptake (photosynthesis). These measurements might be made in great detail on a small subsample of the forest, such as by measuring gas exchange by individual leaves under a great variety of conditions. One then uses validated models to scale up to the whole tree and then the whole forest. Another way is to impose a treatment on a small scale of time and space, then scaling up the responses to large regions for comparisons with large-scale observations from satellites and aircraft. A third way is to test the relations obtained from the analysis of past observations to predict the future relations of fluxes to climate, human actions, etc. If the relations are preserved, this supports the hypothesis of who controls what.

One also hears of 'thought experiments,' in which one proposes relations that might exist in the real world, generating a self-consistent system or theory that has not yet been tested for consistency with the behavior of the external world. The most famous set of thought experiments are probably Einstein's, in his development of the theory of special relativity. He used extant observations, such as from the Michelson-Morley experiment, but otherwise constructed a new framework for describing how motion is perceived in different frames of reference. His strokes of genius included making the constancy of the speed of light in all frames the key constraint. Subsequent to this thought experiment, his ideas were tested (after some time!) and found extremely accurate, as well as capable of dramatic extension in the general theory of relativity. The original 'theory' was only a provisional one, truly a hypothesis. After the successful tests, relativity could truly be said to be a full-blown theory, that is, a quite comprehensive, verified framework for discussing how the world works.

Evolution was long contested as being an untested hypothesis, or even an untestable hypothesis. Since evolution was difficult to observe under natural conditions, evolution languished as a post hoc analysis, little capable of making predictions. In recent decades, this barrier has fallen. Richard Lenski pioneered directed evolution of bacterial populations under imposed stresses or selection pressures. Now, such directed evolution is used to generate bacterial strains or even individual enzymes capable of metabolizing substrates into desired products on industrial scales. Inadvertent directed evolution is unfortunately common, in the evolution of drug resistance in bacteria from our action in (mis)using antiobiotics. Evolution in predicted directions is routinely observed in macroscopic traits of readily visible organisms, such as fruit flies (Burke and Rose, 2009). With the tools of molecular biology, we can go to the other extreme of size, resolving changes at individual sites in proteins and in the genes that code for them. We can quite confidently state which loci in the genes are under selection pressure; they show a higher incidence of establishing mutations that change the protein (thus, its function) than of establishing mutations that change the DNA but code for the same amino acid (not changing the protein's function). A good review, on the sobering topic of the evolution of the HIV virus, is given by R. Nielsen. We still find it difficult to predict exactly which genetic sites will change, but this abates with increasingly good tools for predicting the form and the catalytic activity of proteins of arbitrary sequences of amino acids. (Interesting sidelight here: recent studies show that the latter 'synonymous mutations' in DNA can still affect the protein product (Parmley and Hurst, 2007), by altering the way that messenger RNA is spliced together from separate pieces - each made as a transcription from sequences of DNA - to direct the final protein synthesis.)

All well and good, but can we ever see whole new species arise, as one of the final tests of evolution? Up to here, I have really been talking about microevolution, the evolution of traits within a single species. Since species may take many years to branch from other species, this has been difficult to document. One example observed by humans in real time is the development of a new species of apple maggot (Filchak et al., 2000), when a new host plant was introduced into the US. This case, and a healthy set of other examples in plants and animals is discussed by Joseph Boxhorn on the Web. He also discusses the subtleties in defining species as being distinct from one another, a not insubstantial challenge.

The methods of science are thus seen as more complex than as commonly taught in schools. Even the scientific literature erects what we may call useful fictions, in presenting studies as if all the steps occurred in linear, logical order. In fact, we stumble upon some results before generating hypotheses, we make wrong turns, we have to combine observational experiments with controlled experiments. A chronological retelling of any study's progress would verge on stream-of-consciousness and be as difficult to comprehend. In the end, other scientists care about what we found, not our idiosyncratic route to get there. The standards for acceptance of results, however, are upheld (alas, with varying degrees of rigor) by insistence upon adequate experimental design, parsimony of explanation, clear logic, reproducibility (rarely as simply repeating the same work) and statistical tests to estimate the probability that the linkages observed were real and not fortuitous, or random chance. We check each other's work by reviewing manuscripts before publication, publishing critiques, and using techniques published by others, verifying that they work. Science is a group enterprise in a wide sense. As Claude Bernard said, "Art is I; science is we."

References

Burke MK, Rose, MR. 2009. Experimental evolution with Drosophila.Am. J. Physiol. Regul. Integr. Comp. Physiol.296: R1847-R1854.

Filchak KE, Roethele JB, Feder JL. 2000. Natural selection and sympatric divergence in the apple maggot Rhagoletis pomonella. Nature 407: 739-742.

Parmley JL, Hurst LD. 2007. How do synonymous mutations affect fitness? Bioessays 29: 515-519.