Biology and Philosophy 13: 587–595, 1998. c 1998 Kluwer Academic Publishers. Printed in the Netherlands.
A Clash of Paradigms or the Sound of One Hand Clapping DAVID L. HULL Department of Philosophy Northwestern University Evanston, IL 60208-1315 USA E-mail:
[email protected] A review of Gerry Webster and Brian Goodwin, Form and Transformation, Cambridge University Press, 1996, xiv + 271 pp. $74.95. In The Origin of Species (1859) Darwin took creationists as his foil, but he was well aware that a more powerful group of scientists was likely to oppose his theory of the transmutation of species – ideal morphologists. In a letter to T. H. Huxley in 1853, Darwin acknowledged that the discovery of the “type” for each great class of organisms is one of the “very highest ends of Natural History,” but “I detest the word as used by Owen, Agassiz & Co” (Correspondence of Charles Darwin 5:133). I have never thought that incommensurability is quite the problem that so many generations of Kuhnians have maintained. However, if there has ever been an instance of incommensurability, then it surely is between the sort of science that Darwin’s work exemplified and any form of “idealism.” I myself have not so much disagreed with explanations in terms of ideal types as failed to understand them. At best I have been able to muster fleeting images of what idealists must be after. In Form and Transformation, Webster and Goodwin have done the impossible. They have enabled me to understand what idealists are actually doing. One problem in discussing “idealism” is that it is at best a family of views, going under a plethora of names – rational morphology, structuralism, rational empiricism, transcendental idealism, etc. One explanation for this host of different names is that each of these groups arose independently. Some coined a new name for themselves before they realized that they had predecessors. Others chose a new name in order to distance themselves from their predecessors because idealists of all stripes have a fairly bad reputation among present-day scientists. In contrast, the Darwinians, neo-Darwinians
588 and modern Darwinians form a continuous lineage. No matter how much the content of “Darwinism” might change, the name stays the same – much to the frustration of its opponents. The Darwinian butterfly refuses to remain skewered. Webster and Goodwin list their predecessors from Driesch, Bateson and D’Arcy Thompson to Woodger, Waddington and Goldschmidt. What struck me about this list of predecessors is that each of these men was a maverick in his day. About the only thing that they had in common was opposition to the version of Darwinism prevalent at the time. However, Webster and Goodwin see a “kind of intellectual lineage here with a common theme centred on the problem of biological form, with diversity understood in terms of transformation” (p. xiii). This acknowledgment of an intellectual lineage is somewhat disconcerting to Webster and Goodwin because they champion a science of biology in which history and genealogy are supposed to play little, if any, role. In fact, one weakness of the Darwinian world view, as they see it, is that these notions are so central. Hence, Webster and Goodwin admit with some embarrassment that in the case of their own intellectual development “history does have a role to play after all” (p. xiii). Such honesty is characteristic of this book. Usually in jointly authored works, readers cannot tell for sure which parts belong to whom. However, in this case attribution is easy. Webster wrote the first part of the book on the problem of form, and Goodwin wrote the second part on fields and forms. Both parts concern fundamental reworkings of present-day ways of understanding the living world. In the first part Webster argues against a view that Michael Ghiselin finally forced biologists to notice; i.e., that species as the entities that evolve cannot be viewed as classes (or natural kinds) but must be reconceptualized as spatiotemporal genealogical individuals (or historical entities). The species category might have a form, but particular species do not. Webster argues several positions with respect to the species-as-individuals thesis. First, he distinguishes between species as genealogical units and the claim that these genealogical units are individuals. He acknowledges an important role for the genealogical species concept in evolutionary theory but insists that it is not the only legitimate species concept. A concept in terms of form and similarity is also required. Second, he presents a variety of arguments against treating genealogical species as individuals. Finally, he argues that the genealogical species concept is not entirely free of any reference to traits and similarity because the only access systematists have to genealogical relations is character distributions. In the second part, Goodwin is interested primarily in providing the broad outlines of a theory of forms as an alternative to the Darwinian theory. He
589 does not argue that Darwinian theory is totally false, that it applies nowhere, or that the genealogical perspective it entails is always inappropriate. Instead he intends a theory of forms to function on a par with Darwinian theory. The two supplement each other. The view of evolution that both Webster and Goodwin reject is very gene-centered. On this gene-centered view, genes play two important roles in evolution. First, they incorporate in their structure the information necessary to code for certain characteristics in development. Second, they transmit this information from one generation to the next. Through the course of development, organisms are assembled. They too play a pivotal role in selection. As a result of their interactions with the environment, they make replication differential. As a result of alternating replication and environmental interaction, selection occurs. Webster and Goodwin join the increasingly loud chorus of voices objecting to the preceding world view. Initially this world view helped simplify the relevant states of affairs so that they could be set out more clearly, but certain phenomena have remained recalcitrant. For example, if the traditional picture is adequate, why does it remain so difficult to say anything both sensible and determinant about the nature-nurture issue? One possible answer to such questions is that the basic way that present-day biologists divide up the living world needs reworking. Perhaps the time is ripe (possibly long overdo) for a fundamental restructuring of our biological world view. Goodwin not only attempts such a restructuring but in addition presents a sample of the sorts of generalizations that will be a part of this new world view. Thus, Webster and Goodwin reject one possible conceptual revision of biology (the one urged by Ghiselin and me) and argue for a quite different reformulation. In either case, they are dealing with extremely fundamental issues in biology – issues that in the past have been termed “metaphysical.”
1. Species as Forms Although Webster acknowledges that Ghiselin first outlined the position he discusses, he concentrates on my elaboration of the thesis. A common response to criticisms of one’s work is that the critic has completely misunderstood you. In general, I have been denied this avenue of escape. By and large, those people who have characterized my work, whether in support or in opposition, have gotten me dead to right. Webster sets out my views on the ontological status of species as sympathetically as I could want. The positions he argues against are the positions I happen to hold. We also share several basic assumptions about conceptual analysis both in and about science. For example, he agrees that theoretical terms such as “species” are theory laden.
590 How you define them is strongly influenced by the role that they play in the theory to which they belong. Webster also agrees with Ghiselin and me that, within evolutionary theory, species must be interpreted as genealogical units. However, in the theory of form that he and Goodwin are attempting to develop, species cannot be interpreted in this way because their theory is not genealogical. Species for them are fundamental units of form, not evolution. Webster argues that biological species should be classified and ordered the way that chemists classify and order the alkanes. Just as physicists and chemists treat the molecular structure of inorganic material in terms of electrons, atoms, and molecules, Webster proposes treating biological species as “morphogenetic fields, which share a common nature” (p. 38). Classifications of this sort have paid off in physics, chemistry and crystallography. He is betting that a similar classification will pay off in biology as well. In short, “biologist need to develop kind concepts which are more akin to those employed in the ‘exact’ physical sciences” (p. 25). Please, dear God, don’t let Ernst Mayr read this passage. Webster is candid in acknowledging the asymmetry that currently exists between these two species concepts: one is based on a well-articulated theory, while the other is being developed in the context of an emerging theory. As this theory of morphogenesis is articulated, it may well be the case that the species concept appropriate for it may also have to be modified – but not so modified that it is likely to converge on species as genealogical entities. Webster and Goodwin do not insist that their species definition in terms of form is the only legitimate species concept. They are pluralists in this respect. However, Webster does object to my treating the genealogical species concept as the only proper species concept, not just for evolutionary theory but for all biology. In doing so I am being too monistic. I must admit that I find myself reacting negatively to the current consensus that pluralism is the only acceptable philosophical thesis with respect to scientific concepts. No concept is adequate for anything. All the sorts of arguments that pluralists martial to support a constant variety of scientific concepts apply with equal (or greater) force to philosophical concepts. Hence, one would think that those workers who are relentless pluralists with respect to scientific concepts would become even more relentless when it comes to their own philosophical concepts. To the contrary, they are suddenly converted to intransigent monists when it comes to philosophical concepts. Pluralism is the only acceptable position in philosophy. My position is that a variety of serious alternatives are needed if conceptual change is to be a selection process, but these alternatives must be serious, and unrelenting winnowing is equally necessary. Variety for variety’s sake is not a very good recommendation in science or in philosophy. My current
591 view is that the genealogical species concept is the best bet for fulfilling most of the jobs required of a species concept in biology, but other candidates are also viable. For example, I think we need and might productively develop the notion of ecological units – call them species if you will. Webster and Goodwin have convinced me that we need units of morphogenetic development as well – again call them species if you will. If holding this position makes me a “pluralist,” then I will grit my teeth and bear it. 2. Dialectics Webster sees the relation between theory and practice in science as being “dialectical.” Better and wider sampling of data helps produce better theories, which improves how we collect and group data, which helps : : : etc. For some reason, Webster thinks that I disagree. I do think, for example, that from the genealogical perspective, descent must take priority to similarity when the two diverge. Two groups of organisms might be very similar to each other, but if they occur in different parts of the phylogenetic tree, then they must be treated as distinct no matter how similar they might be. Each species in a genealogical sense is individuated in terms of a type specimen (the holotype) and all organisms appropriately related genealogically to it. All sorts of evidence, including the distribution of homologous traits and geographic location, are used to infer the presence or absence of genealogical relations, but these relations are basic, no matter how they are inferred. Webster’s response by now is extremely familiar. The pheneticists argued for it in the 1960s and 1970s, and it was taken up again by pattern cladists in the 1970s and 1980s. Theoretically, descent may be important, but “relations of ‘descent’ cannot serve as a practical basis for the extension of the specific name and therefore the delimitation of the taxon, for in the overwhelming majority of cases where the holotype individual is a specimen from the wild it is impossible to ascertain directly the relevant relations between the holotype individual and any of its ancestors, descendants, collateral relatives and their descendants which together comprise the extension of the name” (p. 46). Furthermore, since such theoretically significant relations as descent are almost always inferred and not observed, Webster concludes that on my view “taxonomy is impossible” (p. 46). Systematists are lucky. Human beings are roughly the same size as the organisms that they want to classify, microorganisms excepted. Thus, in practice, systematists can see organisms and their traits. Because of this peculiar situation, so Webster argues, a definition of the species category in terms of observable traits is preferable to definitions that depend on factors that we don’t happen to be able to observe so directly. Webster argues that
592 “some concept of type or kind is epistemologically indispensable for picking out relevant individuals as well as for subsequently describing them” (p. 29). I don’t know what Webster would recommend for microorganisms or for those scientists whose subject matter is entirely beyond the limits of unaided human observation. Physicists seem to know much more about subatomic particles than biologists know about observable organisms. I know of no way to reconcile this disagreement between Webster and me. I have argued about it too often with too many people. For some people, concepts defined in terms of those characteristics that people happen to be able to observe fairly directly are preferable to those that are not. The application of all other concepts depends on those concepts that are epistemologically prior. For other people, scientific theories are sufficiently important that they can override practical considerations. Theoretically we know that gene exchange and descent are important in the living world. Even though we must infer the presence of these relations on the basis of evidence that is frequently both indirect and scarce, that is what we must do. Although I don’t know what to say with respect to the differences that exist between Webster and me about epistemological priority, I don’t see what this dispute has to do with the contrast between the practical and the theoretical. His definition of species in terms of morphogenetic fields is just as theoretical as mine and just as far removed from anything that might count as “direct” observation. I have seen a lot of morphological traits in my day, but I have never seen a morphogenetic field, and neither has Webster. The presence of such theoretical entities must be inferred. I suspect that, as Webster continues to develop his species concept in the context of such abstract and abstruse entities as morphogenetic fields, his look-see operationism will cease to appear so attractive to him, especially when some young Turk begins to complain about how far he has strayed from the observational basis of all science. Webster also finds the genealogical species concept inadequate for experimental purposes. I agree that for many experimental purposes, genealogy is not relevant. For example, anyone studying the benefits of giving birth alive would need to study species of organisms scattered throughout the phylogenetic tree. Similarly, anyone studying the mechanics of flight would not care if the males of a particular genealogical species were able to fly but not the females. However, Webster claims that genealogical species are never the appropriate entities for experimental study (pp. 51–52), but here I must disagree. For some experiments, genealogical species are exactly the right entities to study; e.g., ascertaining the amount of gene flow necessary among the various populations of a single genealogical species to keep that species cohesive.
593 Webster also discusses a variety of other issues with respect to species concepts and their definition, e.g., an interesting examination of rigid designation and its correspondence to the type-specimen method; but I cannot discuss all these various issues here. However, I must take note of one more important point upon which Webster and I agree. For both of us, the distinction between spatiotemporal individuals and spatiotemporally unrestricted kinds is important in our understanding of science. Spatiotemporally unrestricted concepts are necessary if we are ever to have any spatiotemporally unrestricted laws. Others find this distinction of no significance whatsoever. For example, if natural phenomena are analyzed in terms of sets, then species are sets because everything and anything can be construed as a set, whether it is Richard Nixon, the tooth I had pulled last week, all the nuts and bolts in my car, not to mention the empty set, the set of all sets, and possibly even a bare particular. In this night, all cows really are black, or as my druggie friends in the 1970s were fond of informing me – all is one.
3. A Science of Forms Because I have spent a good part of thirty-five years working in theoretical systematics, I feel relatively confident that I have understood what Webster is after. For example, I had little trouble in understanding why Webster and Goodwin need a species concept quite different from the genealogical concept so central to evolutionary biology for their own work. However, nearly everything in Goodwin’s second part is new to me. I do not know enough about the science that he describes to evaluate the accuracy of his descriptions, but I do see what he is after. He is trying to develop a theory of form. The second part of this book is devoted to Goodwin’s criticizing the hold that the evolutionary paradigm has on biology and listing regularities and observations that are likely to be elements in the new synthesis that he is trying to construct. For example, apparently at one time such biologists as Monod held an atomistic reductionist view of development according to which the structures generated at any one hierarchical level are uniquely specified by the entities at the next lowest level, all the way down to genes. Goodwin insists that the “generation of a particular structure in an organism depends not only on the properties of the elements making up the structure (cells in this case) but upon additional influences affecting the spatial order that emerges from cell-cell interactions” (p. 135). In support of this view, Goodwin cites a whole series of experiments in which the mesenchyme that normally produces body limbs is extracted, disaggregated and then returned to the proper place under the ectoderm. Normal limbs never develop under these conditions. Hence,
594 something more than genes is necessary for the production of normal tetrapod limbs. But Goodwin is quick to admit that few biologists subscribe to such strict atomism anymore (p. 136). In fact, Darwinians are well aware of all the phenomena that Goodwin discusses. The difference lies in emphasis. Darwinians acknowledge the importance of development, but it enters into their explanations in the role of “constraints” to the more general and important processes of gene action and transmission. Goodwin urges the opposite weighting. Development is central and the genetic background is demoted to being merely a constraint on development. Once all the relevant constraints are specified in both systems, the two sets of factors may well turn out to be equivalent, but as far as heuristics is concerned, differences in emphasis can have significant effects on how science develops. Besides this equivalence is only in principle possible. Central to the view of science within which Webster and Goodwin are functioning is the notion of “transformation.” If ever a distinction begged to be misunderstood, it is the evolutionary versus the idealist notions of transformation. According to evolutionists, the development of limbs in tetrapods can be followed from the origin of fins to different structures in the tetrapod chunk of the phylogenetic tree. Following fins down one branching path, they stay basically fins until the present. Down other paths they become legs. In parts of this chunk of the phylogenetic tree, some legs stay legs up to the present, while others become arms, wings and even “fins” again. In still other cases, they disappear altogether. These structures are being transformed literally in space and time. One of the tasks for systematists of the phylogenetic persuasion is to discern these diverging transformation series. Webster and Goodwin have quite a different notion of transformation in mind. For them characters are not evolutionary homologies but homologies in the pre-Darwinian sense. They model their notion on the concept of equivalence under transformation in mathematics (p. 143). Just as a circle can be transformed into ellipses of increasing eccentricity by increasing the distance between the foci, biological forms can be arranged in transformation series that need not correspond to phylogenetic transformation series. Because Goodwin wants to discover universal regularities among biological phenomena, he must reject the genealogical notion of homology. Phylogenetic transformation series are spatial, temporal and highly contingent. Given a slightly different set of conditions, the series might have turned out very differently from the way it did. Idealist transformation series are much more abstract and rational. If any laws are to be found with respect to homologous characters, the relevant notion of “homology” cannot be evolutionary.1
595 Another universal in animal development concerns the zootype. A comparative study of epigenetic gene expression reveals a stage in the development of all animal phyla at which a similar spatial pattern of genes is expressed. This expression pattern, in turn, is correlated with a morphological stage termed the phylogenetic stage of development. An evolutionist might explain such a regularity in terms of an early, very successful adaptation which has not changed significantly since it first developed. Goodwin explains it in terms of a morphogenetic field which has only a very limited set of possible trajectories. As in the case of Webster’s first part of the book, I cannot touch on all the content of Goodwin’s contributions. They are too many, too varied, and too foreign to me for me to discuss. Throughout the recent history of evolutionary theory, conflicts between it and its opponents have had all the characteristics of one hand clapping. Finally, with the work of Goodwin, Webster and their group, Oyama and the developmentalists as well as Kauffman’s Santa Fe group, the basic outlines of a serious competitor are materializing. If ever there is a time that the intimations of a science of organic form sensed by certain biologists from Aristotle and Owen to the present are to be realized, it is now.
Note 1
Incidentally, D’Arcy Thompson’s most aesthetically appealing example of a topological transformation – the transformation of a porcupine-fish into a sunfish – is anything but accurate. In topological transformations, all elements must remain in their respective cells, but even after D’Arcy Thompson reworked this illustration, it continues to be largely impressionistic and not strictly a topological transformation.
