Natural history is the descriptive study of the natural world. The ultimate objective of science is to go beyond natural history to find generalizations, or explanatory theories, to account for our observations of nature. Theory enables us to explain a set of observations with fewer "bits" (a "bit" being equivalent to the answer to a single yes/no question) of information than are contained in the observations themselves.
The more dramatic the reduction in the amount of such information needed to account for observations and the more accurate the predictions we can make, the more explanatory power we credit to the theory. Predictive power is the gold standard for confidence in a scientific theory. The more sweeping and accurate the better, so long as the predictions are not vacuously vague. Newton's laws of motion, for example, apply broadly in the universe and are sufficiently accurate for many applications; Einstein's modifications are even more accurate and comprehensive.
Scientific theory involves many assumptions that may not always be stated. We assume that the facts of nature are objective and can be explained in natural terms, that is, without intervention of nonmaterial ("supernatural") factors.We also assume the universal validity of logical reasoning and mathematics. One of the most important assumptions that we make in building theory is that the fabric of causation in the cosmos is continuous and well-behaved, that facts are replicable—if we had the same conditions twice, we would have the same outcome. This may not be true in the ultimate sense (for example, if there is true randomness in the motion of atoms). More importantly for biology, our theory may assume replicability to a degree beyond what really applies, or, replicability may be the true state of Nature but our measurements too inaccurate. In fact, predictions and extrapolations can be almost completely inaccurate except in the short run, even for totally deterministic processes whose states or characteristics are not perfectly estimated (this phenomenon is sometimes characterized as "chaos" in the complexity literature).
The general belief among scientists is that we may not know the ultimate truth but that an ultimate truth does exist and that scientific methodology continually gets us closer to that truth. Philosophers of science debate whether this is actually so, noting that science is like other belief systems in resting on axioms—basic principles taken as givens and not to really be questioned. Indeed, science can be a kind of fundamentalism not unlike religion in its intolerance of challenges to its axioms. When, episodically, we become dissatisfied with the accuracy of this theoretical
Genetics and the Logic of Evolution, by Kenneth M. Weiss and Anne V. Buchanan. ISBN 0-471-23805-8 Copyright © 2004 John Wiley & Sons, Inc.
edifice and an alternative explanatory framework is suggested, we experience what Thomas Kuhn called a scientific "revolution" (Kuhn 1962).
One rather curious basic assumption, the principle of parsimony (sometimes called "Occam's Razor"), states that nature is no more complex than it has to be. In scientific practice, this means that we assume that the simplest explanation for an observation is the best one. We implicitly accept that this also means the truest. But of course we don't know how complex nature really is or, in information terms, the degree to which any new theory could explain our current observations with fewer bits of information. This is a special challenge in biology because the biosphere is continually recreated through birth, death, and mutation in ever-changing environments. Unlike chemistry, we cannot replicate observations precisely at our will. Each new organism is unique, and life, unlike theory, does not always behave in the most parsimonious way. In the extreme, if life really were just as complex as our observations, then biology could not go much beyond descriptive "natural history."
Evolutionary biology both describes and predicts. The history of life is generally assumed to have been a one-time affair, whose specific events are unique, contingent (that is, depend on unique circumstances), and hence not replicable. Yet, each individual is a new test of the challenges of survival, and in that and other ways the living world continually replays the general principles of evolution. We find regularities, and these have led to a formal theory of evolution. Nonetheless, this has limited power because specific events in the future cannot be predicted the way one can predict the nature of a chemical reaction, for example. What can be "predicted" (or if we look back in time, "retrodicted") are patterns we might expect to see among descendants, based on postulated processes that affected their ancestors. A central problem is that in inferring how evolution produced what we see today we already know the outcome, so that much of what we do is to fit observations to theory rather than make truly deductive predictions.
One example of a very general prediction is that if different species share a recent common ancestor they will share more characteristics with each other than with species of more remote shared ancestry. If we could specify the extent of the similarity—say, in percent of difference between them on some scale—that specification could reduce the need to enumerate all the traits of each species. Linnaeus developed his systematic classification of life using morphological traits that he believed were important. The same idea can be extended to genes: related species will share genetic (DNA sequence) similarities to an extent that corresponds in some way to their phylogenetic history. This kind of divergence from a common ancestor was the basic idea underlying Charles Darwin's metaphoric tree of life (Darwin 1859) (Figure 1-1), an image that Alfred Russel Wallace also used to express the diverging nature of life, and one similarly employed by evolution's advocate in Germany, Ernst Haeckel, to show the nature of life diverging from "some one primordial form." (In this book for their symbolic utility we will frequently mention specific prominent individuals, but historians of biology have shown clearly that most advances have come from the work of many, famous and less famous).
Relationships previously characterized by Linnaeus have generally held up to studies of genetic data; morphology is not a bad guide to taxonomy.There are exceptions, but they usually involve subtleties, very ancient splits, or traits that can change easily or rapidly with relationships that can only be resolved with extensive amounts of DNA data. Although Linnaeus knew about bacteria (they were first seen micro scopically in 1680 or so by Leeuwenhoek), he didn't understand them or their relationship to other living things and thus lumped them all into a category of miscellany that he called Vermes, in a class called Chaos (Magner 1994) (unrelated to the modern technical use of "chaos" referred to above). Sorting them out was left to future systematists. The complications are similar in nature to the complexity of nongenetic traits that have traditionally enabled debate among taxonomists.
In fact, genetic data are strikingly consistent with, and their characteristics were predicted by, darwinian principles, and it is significant that these findings were entirely independent of, and after, Darwin's formulation of his theory (in this book, we will use uncapitalized references, such as "darwinian," when discussing modified descendants of the original idea and capitalized references, such as "Darwinian," when discussing the specific notions of the person introducing them). Independent confirmation of theoretical ideas with new data is very important to the deductive aspects of science, and genetic taxonomy is an independent confirmation of Darwin. Of course, we know that morphological traits are affected by genes, so genetic data are not entirely independent; however, in a nonevolutionary world, for example, one made by a fixed creation event, there would not have to be any relationship between DNA sequence and morphological similarities.
If genes provide a kind of blueprint for life, genetic data should enable us to describe traits in different species or individuals with less information than is needed to describe each trait or individual separately. This is exactly the kind of reconstruction that Richard Owen and Georges Cuvier made famous in the early 1800s, when they used single bones to reconstruct whole animals, and why Thomas Huxley once exclaimed "A tooth! A tooth! My kingdom for a tooth!" (see Desmond 1994). Their theories were functional (not evolutionary): complex traits like a bone or
«' f ('
I" ■■ r-
Was this article helpful?