American Scientist

In Search of Deep Time: Beyond the Fossil Record to a New History of Life. Henry Gee. Free Press, 1999. $26.

Henry Gee is a senior science writer for Nature, but he was trained as a paleontologist and is thus in an ideal position to write a popular account of a major revolution in scientists' approach to the fossil record. But this is not an impartial outsider's account: It is a passionate advocacy of the new perspective by someone who actually participated in the revolution. The new paleontology is based on cladistics, the rigorous use of morphological (structural) characters to work out the degrees of relationship between species on the assumption that evolutionary novelties only appear once and thus define a group or clade of all the species that inherit the new character by descent from the hypothetical common ancestor. But cladistics imposes rigid limits on how far one can trace such relationships. In particular it insists that one can never know if one form is directly ancestral to another—all we can do is work out degrees of relationship. This means that fossils have to be treated in exactly the same way as living species. Just because a fossil member of a group appears in an earlier geological stratum than any of the other members, we cannot assume that it was the common ancestor.

Gee explains this by appealing to the notion of "deep time," the vast extent of geological time that dwarfs our normal human perception of historical relationships. The fossil record is often so sparse that we have no idea when a group first appeared, because there are plenty of cases where groups survived for tens of millions of years without leaving a single fossil. (The coelacanth is the classic example: This fish was once thought to be extinct because it had disappeared from the fossil record, then contemporary specimens were taken from the Indian Ocean). The first known fossil is thus no reliable guide to the date or the structure of the first member of its group. Nor, Gee insists, does the record tell us anything about the adaptations of the species we study. Apart from the difficulty of reconstructing the ancient ecologies, and the impossibility of knowing the effect of, for instance, ancient parasites, there are many cases where it has become plain that a structure for which there is now a clearly defined adaptive purpose was first evolved in forms that could not have used the structure in the modern way—so how can we possibly tell what they did use it for?

The consequence of all this, the cladists insist, is that all the old paraphernalia of evolutionary explanations must be dismissed as unscientific speculation. All we can do is assess degrees of relationship. We cannot identify ancestors or "missing links," and we cannot devise testable theories to explain how particular episodes of evolution came about. Gee is adamant that all the popular stories about how the first amphibians conquered the dry land, how the birds developed wings and feathers for flying, how the dinosaurs went extinct and how humans evolved from apes are just products of our imagination driven by prejudices and preconceptions. They reflect our modern ideas about the purposes of adaptive structures and about the progressive trend we think we see in the history of life up to humankind. They cannot be part of science because they cannot be tested against the fossil record. This is not an anti-evolution diatribe because cladism depends on the notion of common ancestry, and Gee admits that Darwin's theory can be applied in the modern world where we can actually see ecological relationships at work. But cladism places severe restrictions on what kind of questions can be asked about the distant past, in the cause of raising paleontology to the ranks of a hard science.

The two great strengths of Gee's account are its iconoclastic destruction of many popular evolutionary scenarios and the author's intimate knowledge of the personalities and events surrounding the revolution. We learn how new fossils, and cladistic interpretations of existing fossils, have exploded the old myths about the great steps in evolution. Limbs evolved in fish that could clearly never have used them to venture onto dry land. The whole complex of bird-like characters, feathers and all, evolved in dinosaurs that could not fly or even glide. We are also introduced to the paleontologists who pioneered the revolution, including the "Gang of Four" in the fossil fish department at the Natural History Museum in London, led by Colin Patterson, with whom Gee himself studied. The outraged complaints of old-fashioned paleontologists such as the late Bev Halstead—who tried to dismiss cladism as a Marxist plot—are described with relish. At this level, Gee's book works really well, giving the reader both an introduction to the new lines of evidence and a real sense of the excitement generated by scientists promoting a new approach to their discipline.

Gee is surely right to claim that many of the old evolutionary scenarios were based on prejudice and outdated thinking, although his account misses some significant points. He criticizes the American paleontologist A. S. Romer for telling a "fairy story" about how the earliest fish developed armor to protect themselves from the "dragons" of the time, the eurypterids, or sea scorpions. How, he asks, do we know that the eurypterids were responsible? He does not mention that it was Romer who first questioned the popular myth of the amphibians developing limbs to conquer the dry land—he suggested they used the limbs to crawl to the next pool in times of drought.

Gee also attacks the modern assumption that humans evolved through a sequence defined by fossil species. (There were howls of anguish when this point of view was included in a display at the Natural History Museum.) Again, he fails to mention that it was commonplace in the early 20th century to sideline the known fossils from human ancestry. This was because paleontologists then thought that many lines of evolution would independently develop human characters, exactly the opposite of the cladists' position.

Gee's assault on the old stories is effective, if selective, but we need to think carefully about the cladists' insistence that all such historical narratives must be banned from science. It is surely contradictory to dismiss them as untestable and then provide good evidence to disprove several of them. Unless I am missing something, a hypothesis that has been falsified must have been falsifiable—so what does Gee mean when he says that all such narratives are untestable? He certainly does not mean that they are unverifiable, because his own philosophy of science insists that all hypotheses are tentative and open to falsification. Perhaps he means that, having destroyed the old narratives, it will be difficult to come up with new ones that can be similarly tested. But this is a practical claim, not one based on a methodological principle.

Others have argued that narratives are unscientific—Misia Landau made this point about paleoanthropology some years ago, noting how stories about human origins parallel the structure of many creation myths (Narratives of Human Evolution, 1991). But the fossil record has demolished several "myths" about human origins, including the once-popular idea that our ancestors developed big brains before they became bipedal. Narratives are not in principle untestable, for the obvious reason that some have been refuted in the course of the past hundred years.

In the end, I remain profoundly unhappy about any attempt to claim that certain types of questions about the past should be banished from science on methodological grounds. The history of science shows us that this is often a mistake. Charles Lyell, the founder of uniformitarian geology, applied his methodology so rigidly that he dismissed as unscientific all efforts to understand the earth's origin. Yet nowadays we routinely discuss theories about the formation of our planet, because lines of evidence have opened up that were simply unavailable in Lyell's time.

I welcome the critical side of the cladists' assault on the popular myths about evolution, but I suspect that their effort to dismiss all historical narratives from science is another product of the physics envy that occasionally tempts paleontologists and geologists to adopt impossibly harsh methodological criteria. As Ernst Mayr, among others, has argued, we need to recognize that studying the past requires a different kind of relation between theory and evidence. There is not, as Gee implies, a single kind of science, nor does the use of the imagination in itself make a hypothesis unscientific.

There is another reason for concern, however, which derives from the public perception of science. Paleontologists may want to impress their scientific colleagues with the "hardness" of their new methodology—but if they tell us that all efforts to explain the origin of new forms in the course of evolution are unscientific, they leave up for grabs the whole area of how the natural world came into its present form. If science admits that it cannot in principle explain origins, the creationists will be only too willing to tell the public that alternative sources of information are available. In these circumstances, do we really want to concede that science cannot even ask the questions everyone would like to have answered?