The Panda's Thumb: More Reflections in Natural History

by Stephen Jay Gould

  Physically incapacitated now (I am 96), I can only cheer the work of others in a field where I was a pathfinder. Again I thank you.

  J Harlen Bretz

  In November 1979, at the annual meeting of the Geological Society of America, the Penrose Medal (the profession’s premier award) was given to J Harlen Bretz.

  20 | A Quahog Is a Quahog

  THOMAS HENRY HUXLEY once defined science as “organized common sense.” Other contemporaries, including the great geologist Charles Lyell, urged an opposing view—science, they said, must probe behind appearance, often to combat the “obvious” interpretation of phenomena.

  I cannot offer any general rules for the resolution of conflicts between common sense and the dictates of a favored theory. Each camp has won its battles and received its lumps. But I do want to tell a story of common sense triumphant—an interesting story because the theory that seemed to oppose ordinary observation is also correct, for it is the theory of evolution itself. The error that brought evolution into conflict with common sense lies in a false implication commonly drawn from evolutionary theory, not with the theory itself.

  Common sense dictates that the world of familiar, macroscopic organisms presents itself to us in “packages” called species. All bird watchers and butterfly netters know that they can divide the specimens of any local area into discrete units blessed with those Latin binomials that befuddle the uninitiated. Occasionally, to be sure, a package may become unraveled and even seem to coalesce with another. But such cases are noted for their rarity. The birds of Massachusetts and the bugs in my backyard are unambiguous members of species recognized in the same way by all experienced observers.

  This notion of species as “natural kinds” fit splendidly with creationist tenets of a pre-Darwinian age. Louis Agassiz even argued that species are God’s individual thoughts, made incarnate so that we might perceive both His majesty and His message. Species, Agassiz wrote, are “instituted by the Divine Intelligence as the categories of his mode of thinking.”

  But how could a division of the organic world into discrete entities be justified by an evolutionary theory that proclaimed ceaseless change as the fundamental fact of nature? Both Darwin and Lamarck struggled with this question and did not resolve it to their satisfaction. Both denied to the species any status as a natural kind.

  Darwin lamented: “We shall have to treat species as…merely artificial combinations made for convenience. This may not be a cheering prospect; but we shall at least be freed from the vain search for the undiscovered and undiscoverable essence of the term species.” Lamarck complained: “In vain do naturalists consume their time in describing new species, in seizing upon every nuance and slight peculiarity to enlarge the immense list of described species.”

  Yet—and this is the irony—both Darwin and Lamarck were respected systematists who named hundreds of species. Darwin wrote a four-volume taxonomic treatise on barnacles, while Lamarck produced more than three times as many volumes on fossil invertebrates. Faced with the practicum of their daily work, both recognized entities where theory denied their reality.

  There is a traditional escape from this dilemma: one can argue that our world of ceaseless flux alters so slowly that configurations of the moment may be treated as static. The coherence of modern species disappears through time as they transform slowly into their descendants. One can only remember Job’s lament about “man that is born of a woman”—“He cometh forth like a flower…he fleeth also as a shadow, and continueth not.” But Lamarck and Darwin could not even enjoy this resolution, for they both worked extensively with fossils and were as successful in dividing evolving sequences into species as they were in parsing the modern world.

  Other biologists have even forsworn this traditional escape and denied the reality of species in any context. J.B.S. Haldane, perhaps the most brilliant evolutionist of this century, wrote: “The concept of a species is a concession to our linguistic habits and neurological mechanisms.” A paleontological colleague proclaimed in 1949 that “a species…is a fiction, a mental construct without objective existence.”

  Yet common sense continues to proclaim that, with few exceptions, species can be clearly identified in local areas of our modern world. Most biologists, although they may deny the reality of species through geologic time, do affirm their status for the modern moment. As Ernst Mayr, our leading student of species and speciation, writes: “Species are the product of evolution and not of the human mind.” Mayr argues that species are “real” units in nature as a result both of their history and the current interaction among their members.

  Species branch off from ancestral stocks, usually as small, discrete populations inhabiting a definite geographical area. They establish their uniqueness by evolving a genetic program sufficiently distinct that members of the species breed with each other, but not with members of other species. Their members share a common ecological niche and continue to interact through interbreeding.

  Higher units of the Linnaean hierarchy cannot be objectively defined, for they are collections of species and have no separate existence in nature—they neither interbreed nor necessarily interact at all. These higher units—genera, families, orders, and on up—are not arbitrary. They must not be inconsistent with evolutionary genealogy (you cannot put people and dolphins in one order and chimps in another). But ranking is, in part, a matter of custom with no “correct” solution. Chimps are our closest relatives by genealogy, but do we belong in the same genus or in different genera within the same family? Species are nature’s only objective taxonomic units.

  Shall we then follow Mayr or Haldane? I am a partisan of Mayr’s view and I wish to defend it with an offbeat but, to my mind, persuasive line of evidence. The repeated experiment is a cornerstone of scientific methods—although evolutionists, dealing with nature’s uniqueness, do not often have an opportunity to practice it. But in this case, we have a way to obtain valuable information about whether species are mental abstractions embedded in cultural practice or packages in nature. We can study how different peoples, in complete independence, divide the organisms of their local areas into units. We can contrast Western classifications into Linnaean species with the “folk taxonomies” of non-Western peoples.

  The literature on non-Western taxonomies is not extensive, but it is persuasive. We usually find a remarkable correspondence between Linnaean species and non-Western plant and animal names. In short, the same packages are recognized by independent cultures. I do not argue that folk taxonomies invariably include the entire Linnaean catalog. People usually do not classify exhaustively unless organisms are important or conspicuous. The Fore of New Guinea have a single word for all butterflies, although species are as distinct as the birds they do classify in Linnaean detail. Similarly, most of the bugs in my backyard have no common name in our folk taxonomy, but all the birds in Massachusetts do. The Linnaean correspondences only arise when folk taxonomies attempt an exhaustive division.

  Several biologists have noted these remarkable correspondences in the course of their fieldwork. Ernst Mayr himself describes his experience in New Guinea: “Forty years ago, I lived all alone with a tribe of Papuans in the mountains of New Guinea. These superb woodsmen had 136 names for the 137 species of birds I distinguished (confusing only two nondescript species of warblers). That…Stone Age man recognizes the same entitites of nature as Western university-trained scientists refutes rather decisively the claim that species are nothing but a product of the human imagination.” In 1966, Jared Diamond published a more extensive study on the Fore people of New Guinea. They have names for all the Linnaean bird species in their area. Moreover, when Diamond brought seven Fore men into a new area populated by birds they had never seen, and asked them to give the closest Fore equivalent for each new bird, they placed 91 of 103 species into the Fore group closest to the new species in our Western Linnaean classification. Diamond relates an interesting tale:

  One of my Fore assistants collected a huge,
black, short-winged, ground-dwelling bird, which neither he nor I had seen before. While I was puzzled by its affinities, the Fore man promptly proclaimed it to be a peteobeye, the name for a graceful little brown cuckoo which frequents trees in Fore gardens. The new bird eventually proved to be Menbek’s coucal, an aberrant member of the cuckoo family, to which some features of body form and leg and bill shape betray its affinity.

  These informal studies by biologists have been supplemented in recent years with two exhaustive treatments by anthropologists who are also competent natural historians—Ralph Bulmer’s work on vertebrate taxonomies of the Kalam people of New Guinea, and Brent Berlin’s study (with botanists Dennis Breedlove and Peter Raven) of plant classification by the Tzeltal Indians of highland Chiapas, Mexico. (I thank Ernst Mayr for introducing me to Bulmer’s work and for urging this line of argument for many years.)

  The Kalam people, for example, use frogs extensively as food. Most of their frog names have a one-to-one correspondence with Linnaean species. In some cases they apply the same name to more than one species, but still recognize the difference: Kalam informants could readily identify two different kinds of gunm, distinguished both by appearance and habitat, even though they had no standard names for them. Sometimes, the Kalam do better than we. They recognize, as kasoj and wyt, two species that had been lumped incorrectly under the single Western name Hyla becki.

  Bulmer has recently teamed up with Ian Saem Majnep, a Kalam, to produce a remarkable book, Birds of My Kalam Country. More than 70 percent of Saem’s names have one-to-one correspondence with Western species. In most other cases, he either lumps two or more Linnaean species under the same Kalam name but recognizes the Western distinction, or else he makes divisions within a Western species but recognizes the unity (in some birds of paradise, for example, he names the sexes separately because only males carry the prized plumage). In only one case does Saem follow a practice inconsistent with Linnaean nomenclature—he uses the same name for drab females in two species of birds of paradise, but awards different names to the showy males of each species. In fact, Bulmer could only find four cases (2 percent) of inconsistency in the entire Kalam catalog of 174 vertebrate species, spanning mammals, birds, reptiles, frogs, and fishes.

  Berlin, Breedlove, and Raven published their first study in 1966, explicitly to challenge Diamond’s claim for the generality of extensive one-to-one correspondence between folk names and Linnaean species. They held initially that only 34 percent of Tzeltal plant names matched Linnaean species and that a large variety of “misclassifications” reflected cultural uses and practices. But a few years later, in a frank article, they reversed their opinion and affirmed the uncannily close correspondence of folk and Linnaean names. They had, in the earlier study, not fully understood the Tzeltal system of hierarchical ordering and had mixed names from several levels in establishing the basic folk groups. In addition, Berlin admitted that he had been led astray by a standard anthropological bias for cultural relativism. I cite his recantation, not to show him up, but as a token of my admiration for an act all too rarely performed by scientists (although any scientist worth his salt has changed his mind about fundamental issues):

  Many anthropologists, whose traditional bias is to see the total relativity of man’s variant classifications of reality, have generally been hesitant to accept such findings…. My colleagues and I, in an earlier paper, have presented arguments in favor of the “relativist” view. Since the publication of that report more data have been made available, and it now appears that this position must be seriously reconsidered. There is at present a growing body of evidence that suggests that the fundamental taxa recognized in folk systematics correspond fairly closely with scientifically known species.

  Berlin, Breedlove, and Raven have now published an exhaustive book on Tzeltal taxonomy, Principles of Tzeltal Plant Classification. Their complete catalog contains 471 Tzeltal names. Of these, 281, or 61 percent, stand in one-to-one correspondence with Linnaean names. All but 17 of the rest are, in the authors’ terms, “underdifferentiated”—that is, the Tzeltal names refer to more than one Linnaean species. But, in more than two-thirds of these cases, the Tzeltal use a subsidiary system of naming to make distinctions within the primary groups, and all these subsidiaries correspond with Linnaean species. Only 17 names, or 3.6 percent, are “overdifferentiated” by referring to part of a Linnaean species. Seven Linaean species have two Tzeltal names, and only one has three—the bottle gourd Lagenaria siceraria. The Tzeltal distinguish bottle gourd plants by the utility of their fruits—one name for large, round fruits used as containers for tortillas; another for long-necked gourds well suited for carrying liquids; and a third for small, oval fruits that are not used at all.

  A second, equally interesting generality emerges from studies of folk classification. Biologists argue that only species are real units in nature, and that names at higher levels of the taxonomic hierarchy represent human decisions about how species should be grouped (under the constraint, of course, that such grouping be consistent with evolutionary genealogy). Thus, for names applied to groups of species, we should not expect one-to-one correspondence with Linnaean designations but should anticipate a variety of schemes based upon local uses and culture. Such variety has been a consistent finding in studies of folk taxonomy. Groups of species often include basic forms attained independently by several evolutionary lines. The Tzeltal, for example, have four broader names for groups of species, roughly corresponding to trees, vines, grasses, and broad-leafed herbaceous plants. These names apply to about 75 percent of their plant species, while others, like corn, bamboo, and agave, are “unaffiliated.”

  Often, the grouping of species reflects more subtle and pervasive aspects of culture. The Kalam of New Guinea, for example, divide their nonreptilian four-footed vertebrates into three classes: kopyak, or rats; kmn for an evolutionary heterogeneous collection of larger game mammals, mostly marsupials and rodents; and as for an even more heterogeneous collection of frogs and small rodents. (Under repeated questioning by Bulmer, the Kalam deny any subdivision between frogs and rodents within as, although they do acknowledge [and dismiss as unimportant] the morphological similarity between small furry as and rodents among kmn. They also recognize that some kmn have pouches and others do not.) The divisions reflect fundamental facts of Kalam culture. Kopyak, associated with excrement and unclean food around homesteads, are not eaten at all. As are collected primarily by women and children and, although eaten by most men and collected by some, are forbidden foods for boys during their rites of passage and for adult men who practice sorcery. Kmn are hunted primarily by men.

  Likewise, birds and bats are all yakt, with the single exception of the large, flightless cassowary called kobty. The distinction is made for deeper and more complex reasons than mere appearance—for the Kalam do recognize avian characters in kobty. Cassowaries, Bulmer argues, are the prime game of the forest and the Kalam maintain an elaborate cultural antithesis between cultivation (represented by taro and pigs) and the forest (represented by pandanus nuts and cassowaries). Cassowaries are also the mythological sisters of men.

  We maintain similar practices in our own folk taxonomy. Edible mollusks are “shellfish,” but Linnaean species all have common names. I well remember the reprimand I received from a New England shipmate when I applied the informal scientific term “clam” to all bivalved mollusks (to him a clam is only the steamer, Mya arenaria): “A quahog is a quahog, a clam is a clam, and a scallop is a scallop.”

  The evidence of folk taxonomy is persuasive for the modern world. Unless the tendency to divide organisms into Linnaean species reflects a neurological style wired into all of us (an interesting proposition, but one that I doubt), the world of nature is, in some fundamental sense, really divided into reasonably discrete packages of creatures as a result of evolution. (I do not, of course, deny that our propensity for classifying in the first place reflects something about our brains, their inherited capacities, and th
e limited ways in which complexity may be ordered and made sensible. I merely doubt that such a definite procedure as classification into Linnaean species could reflect the constraints of our mind alone, and not of nature.)

  But are these Linnaean species, recognized by independent cultures, merely temporary configurations of the moment, mere way stations on evolutionary lineages in continual flux? I argue in essays 17 and 18 that, contrary to popular belief, evolution does not work this way, and that species have a “reality” through time to match their distinctness at a moment. An average species of fossil invertebrates lives five to ten million years (terrestrial vertebrates have shorter average durations). During this time, they rarely change in any fundamental way. They become extinct, without issue, looking much as they did when they first appeared.

  New species usually arise, not by the slow and steady transformation of entire ancestral populations, but by the splitting off of small isolates from an unaltered parental stock. The frequency and speed of such speciation is among the hottest topics in evolutionary theory today, but I think that most of my colleagues would advocate ranges of hundreds of thousands of years for the origin of most species by splitting. This may seem like a long time in the framework of our lives, but it is a geological instant, usually represented in the fossil record by a single bedding plane, not a long stratigraphic sequence. If species arise in hundreds or thousands of years and then persist, largely unchanged, for several million, the period of their origin is a tiny fraction of one percent of their total duration. Therefore, they may be treated as discrete entities even through time. Evolution at higher levels is fundamentally a story of the differential success of species not the slow transformation of lineages.