- Email: [email protected]
Can evolution be directional without being teleological?
Studies in History and Philosophy of Biological and Biomedical Sciences xxx (2015) 1e7
Contents lists available at ScienceDirect
Studies in History and Philosophy of Biological and Biomedical Sciences journal homepage: www.elsevier.com/locate/shpsc
Can evolution be directional without being teleological? George R. McGhee Jr. Department of Earth and Planetary Sciences, Wright-Rieman Laboratories, Rutgers University, 610 Taylor Road, Pistacaway, NJ 08854 USA
a r t i c l e i n f o
a b s t r a c t
Article history: Available online xxx
Convergent evolution reveals to us that the number of possibilities available for contingent events is limited, that historically contingent evolution is constrained to occur within a finite number of limited pathways, and that contingent evolution is thus probabilistic and predictable. That is, the phenomenon of convergence proves that truly contingent evolutionary processes can repeatedly produce the same, or very similar, organic designs in nature and that evolution is directional in these cases. For this reason it is argued in this paper that evolution can be directional without being teleological, and that the dichotomy that evolution must either be directionless and unpredictable or directional and predetermined (teleological) is false. Ó 2015 Published by Elsevier Ltd.
Keywords: Directional evolution Convergence Contingency Teleology
When citing this paper, please use the full journal title Studies in History and Philosophy of Biological and Biomedical Sciences
1. The false dichotomy of contingency versus teleology in evolution The York University biologist Jan Sapp (2012) argues that a false dichotomy has been erected in the construction of the evolutionary contingency-versus-directionality debate, namely “the false dichotomy between the factors of contingency and chance on one hand and directed evolution based on supernatural forces on the other. In the history of evolutionary biology, it is chance, the struggle for existence, and contingency that have always been on the side of evolution; goal-driven physiological change was always on the side of supernaturalism and the antievolution movement. This dichotomy has been maintained by an ongoing conflict between scientists and creationists” (Sapp, 2012, 694). Why is this dichotomy false? It is false because evolution can be directional without being teleological (McGhee, 2011, 272e73). The false dichotomy that evolution must either be directionless or unpredictable (a possibility that is mislabeled “contingent” in the false dichotomy) or directional and predetermined (teleological) repeatedly appears in the many arguments of Stephen Jay Gould, the Harvard paleontologist and best-known evolutionary essayist of the late twentieth century. For example, Gould (1989) frames this
E-mail address: [email protected].
false dichotomy as the “question of questions” in his argument against the views of the early twentieth-century paleontologist Charles Doolittle Walcott: “ultimately, the question of questions boils down to the placement of the boundary between predictability under invariant law and the multifarious possibilities of historical contingency. Traditionalists like Walcott would place the boundary so low that all major patterns of life’s history fall above the line into the realm of predictability (and, for him, direct manifestation of divine intentions). But I envision a boundary sitting so high that almost every interesting event in life’s history falls into the realm of contingency” (Gould, 1989, 290); that is, into the realm of unpredictability. The teleological endpoint in the false dichotomy certainly exists as such. That is, under a teleological view evolution is directional and predetermineddteleology is defined as “the fact or quality of being directed toward a definite end or having an ultimate purpose, especially as attributed to natural processes.”1 But what about the opposite endpoint in the false dichotomy, the evolutionary “factors of contingency and chance” referred to by Sapp (2012, 694)? In his article in this issue John Beatty differentiates between “contingent per se,” in which a given evolutionary event was a matter of chance (such as a genetic mutation), and
1
Webster’s New World Dictionary, 2nd college ed., s.v. “teleology.”
http://dx.doi.org/10.1016/j.shpsc.2015.12.006 1369-8486/Ó 2015 Published by Elsevier Ltd.
Please cite this article in press as: McGhee, G. R., Jr., Can evolution be directional without being teleological?, Studies in History and Philosophy of Biological and Biomedical Sciences (2015), http://dx.doi.org/10.1016/j.shpsc.2015.12.006
2
G.R. McGhee Jr. / Studies in History and Philosophy of Biological and Biomedical Sciences xxx (2015) 1e7
“contingent upon,” in which a given evolutionary event depended upon the occurrence of a prior event (such as the evolution of an ancestor). Does contingency in either of these two senses really require that evolution be directionless and unpredictable? At first glance it might seem so, as in our current understanding biological evolution is generally modeled to be the outcome of the interaction of chance (random genetic mutations) and uncertain conditions (changing environmental habitats in nature); that is, evolution is the product of genetic mutations and natural selection, where genetic mutations are seen as the source of new biological variation and natural selection is the mechanism sorting the biological variants in terms of their differential adaptations to different environmental conditions. In fact, both of these components of the standard model of evolution can produce directional trends. First, it is a well-known mathematical fact that random processes, chance processes, can product directional trends. The best known of these are Markov chains. A Markov process produces “a sequence of events such that each event is partly dependent on the outcomes of preceding events and partly dependent on a random process acting at the time of the event itself” (Raup, 1977b, 62). Computer simulations using Markov chains are thus an ideal way to explore temporal processes that are both contingent upon (Raup’s “each event is partly dependent on the outcomes of preceding events”) and contingent per se (Raup’s “partly dependent on a random process acting at the time of the event itself”), and can be used to simulate random walks in a single evolving phylogenetic lineage or random evolutionary branching in multiple phylogenetic lineages (Raup, 1977a, 1977b). In the field of evolutionary paleobiology itself, the University of Chicago theoretician David M. Raup ran computer simulations of chance (contingent per se) and historically contingent changes (contingent upon) in the morphologies of hypothetically evolving species. David Sepkoski has provided an account of these studies in his article in the issue. Raup’s early work (Raup & Gould, 1974) demonstrated that random simulations can produce: (1) apparent directional trends in morphological change in evolving lineages, (2) apparent correlation between combinations of morphological traits within an evolving lineage, (3) variation between rates of morphological change in different evolving lineages, (4) apparent “terminal” overspecialization within an evolving lineage, and (5) clumping of randomly evolved morphologies within a theoretical morphospace of the hypothetical spectrum of forms that potentially could be evolved. Prior to Raup’s random simulations, all of these phenomena were thought to be produced only by deterministic evolutionary processes. Thus Raup’s computer simulations demonstrated that chance processes (contingent per se) and historically contingent processes (contingent upon) could produce directional trends; that is, truly contingent evolution (in both senses of contingent) is not necessarily a directionless process. Second, natural selection is a deterministic process, not a random one. Wilson and Bossert (1971, 40) point out that natural selection is a deterministic process in that it involves “directions and rates that can be measured in populations and used to predict specific outcomes” and that only evolution via genetic drift is a stochastic process in which only probabilities of outcomes, not a specific outcome, may be predicted. The mechanism of natural selection differentially sorts biological variants in terms of their differential adaptations to environmental conditions, and differentially preserves the variants with better states of adaptation. Thus natural selection is a directional process, one in which the frequency of organisms with better states of adaptation increases with time in evolving populations (given constancy in the environmental conditions producing the selection). The fact that Charles Darwin’s concept of natural selection was both deterministic and directional has led Humboldt State
University biologist John Reiss (2009, 140) to argue that the very concept itself is teleological: “Darwin introduced a teleological determinism into the heart of his theory. This teleology is expressed in two related conceptions: (1) that evolution is a process going from a less-adapted to a better-adapted state and (2) that natural selection is a deterministic force, or agent, that directs the evolutionary process toward this better-adapted state.” That is, Reiss considers the very idea of adaptive improvement via natural selection in evolution to be teleological because improvement is a directional concept and any hint of directionality in evolution is teleological. Thus the false dichotomy that directionality in evolution must be teleological because contingent evolution is directionless lies at the heart of Reiss’s argument. Is it teleological that water flows in the downhill direction? That it goes from a state of higher potential energy to a state of lower potential energy under the influence of the deterministic force of gravity? Philosophers may argue whether that phenomenon is teleological or not, but in science it is an empirical observation, an established fact. Why then would it be teleological if organisms evolve in the direction of better states of adaptation under the deterministic force of natural selection? 2. How contingent evolution gives rise to predictable directionality: nature has limited choice Contingent evolution (in both senses) can be directional with being teleological, but can the direction of evolution be predictable without being teleological? This question addresses the second part of the false dichotomy that evolution is either directionless or unpredictable (a possibility that is mislabeled “contingent” in the false dichotomy) or directional and predetermined (teleological). Certainly any natural process that is “directed toward a definite end or having an ultimate purpose” can be considered to be teleological.1 Thus Reiss further objects to the concept of natural selection as a “deterministic force, or agent, that directs the evolutionary process” towards a better-adapted state because this “better-adapted state” appears to be a goal, and that a goal-directed process is teleological (Reiss, 2009, 140). Is it teleological that water flows in rivers toward the future “goal” of reaching a sea or lake? The flow of water is mindless, and water has no goal in sight, but nevertheless it will reach its lowest possible potential energy state under the deterministic force of gravity, which means that it will wind up in a sea or lake. And we can confidently predict that flowing water will wind up in a sea or lake although we may not be able to predict which sea or lake without a great deal of additional information about the geography and topography of the landscape across which the water is flowing. Natural selection is a deterministic process and thus, theoretically, one should be able to predict its outcome if one knows the magnitudes of all of the variables at work in the process (Wilson & Bossert, 1971). In actual practice that is a very difficult task indeed. In addition, the element of randomness does enter the equation of natural selection in that the state of the environmental conditions producing the natural selection mechanism (of sorting biological variants in terms of their differential adaptations to those environmental conditions) are uncertain. That is, we may be able to calculate the probabilities that environmental conditions will change in this direction or the other, but we cannot do so with absolute certainty. We are all familiar with the uncertainties involved with weather forecasting where predicted future weather conditions are given in probabilities, not certainties. It is at this point that the empirical pattern of actual evolution on Earth helps us out tremendously: much of evolution is convergent. Convergent evolution reveals to us that the number of evolutionary pathways available to life is not endless, but is instead quite limited.
Please cite this article in press as: McGhee, G. R., Jr., Can evolution be directional without being teleological?, Studies in History and Philosophy of Biological and Biomedical Sciences (2015), http://dx.doi.org/10.1016/j.shpsc.2015.12.006
G.R. McGhee Jr. / Studies in History and Philosophy of Biological and Biomedical Sciences xxx (2015) 1e7
That is, “convergent evolution results from the finite number of ecological niches on our planet, each of which is exploited by many organisms that share a common origin” (Klimentidis, 2012, 1). Even given the many uncertainties in predicting the pathway direction of natural-selection-induced trends in evolution, those uncertainties are considerably reduced when it is realized that the actual number of pathways that evolution may take are limited. We can easily imagine a universe where that fact is not true; that is, a world in which the number of pathways available to evolution are limitless, are infinite. In such a universe there would be so many alternative pathways possible (the “multifarious possibilities” of Gould [1989, 120]) that it would never be possible to predict the trajectory of the evolution of life. In such a universe each individual species would have its own, unique niche or ecological role in nature and each species would have its own, unique adaptive form or ecomorphotype within the selectional parameters of its particular niche or ecological role. That universe does not exist. Instead, we live in a universe where convergence in evolution is rampant at every level, from the ecological roles of organisms and their adaptive forms down to the molecules from which those forms are constructed. Multiple species from disparate phyletic lineages have evolved to occupy the very same niche, to play the same ecological role in nature. And, as a selectional consequence, multiple species have convergently evolved the very same adaptive form, or ecomorphotype. What is astounding on our real Earth is that each of these convergent ecomorphotypes have evolved in a historically contingent process. For example, consider the evolution of the porpoise (or whales in general). Whales and porpoises are openocean dwelling mammals, but they evolved from land-dwelling mammals ecologically similar to the modern-day hippopotamus. A hippopotamus is an artiodactyl mammal adapted to the environmental conditions of a freshwater, semi-aquatic habitat. If we go back to the early Eocene in geologic time we find that other artiodactyl mammals had also evolved to live in freshwater, semi-aquatic habitats: animals like Pakicetus and Indohyus. Some of these animals evolved the adaptations needed to move into more brackish-water habitats, evolved the adaptations need to deal with osmotic problems associated with higher salt contents in the waters in which they lived: animals like Ambulocetus. Venturing further offshore, some of these animals evolved the adaptations needed to live in saltwater habitats and also to breathe while floating in water (instead of standing on land) by shifting their nostrils upwards on their skulls: animals like Kutchicetus and Rodhocetus. By the late Eocene some of these animals evolved tail flukes to better propel themselves through water in swimming, and their unused hind legs became smaller and smaller: animals like Dorudon and Basilosaurus. And finally in the Miocene we have the evolution of the open-ocean habitat mammals, such as the dolphin Kentriodon, which have no hind legs at all, whose forelimbs are fins, and which have tail fins and dorsal fins (Benton, 2005). A porpoise or dolphin looks like a fish: it has a fusiform, streamlined body like that of a swordfish or tuna, it has fins instead of legs, a large fin on its posterior end instead of a tail, and even has a fin on its back. The transformation of land-dwelling, four-legged mammals with tails and no fins on their backs into open-ocean dwelling mammals that look like fish is astonishing but empirically verifiable in the fossil record. And every step on that evolutionary transition was historically contingent (contingent upon), dependent upon chance genetic changes (contingent per se) in the phyletic lineage of these particular animals and dependent upon chance changes in the environmental habitats in which these animals found themselves living in the span of time from the early Eocene to the Miocene.
3
Surely such an astonishing evolutionary transition has to be unique, a contingent series of events (in both senses of contingent) that could never ever be repeated in the history of life on Earth. Yet, even more astonishing, is the empirically verifiable fact that that same evolutionary transition happened in the Mesozoic with the transformation of land-dwelling, four-legged reptiles with tails and no fins on their backs into open-ocean dwelling reptiles that look like fish: the ichthyosaurs. The evolution of ichthyosaurs has not been studied with the detail given that of whales, but the general trends can still be seen. In the late Permian a group of fully terrestrial basal neodiapsids, animals like Youngina with slender bodies and short limbs, gave rise to animals that moved into freshwater fluvial habitats and then evolved into marginal-marineadapted animals like Placodus, which had very short but still functioning legs with feet, and shallow-marine-adapted animals like Pachypleurosaurus, which had four limbs with paddle fins instead of feet, in the Triassic. The oldest-known open-ocean dwelling ichthyosaur is the early Triassis Utatsusaurus, an animal with four ventral fins and a tail fin but which had no dorsal fin like a porpoise. More derived ichthyosaurs, like the late Triassic Mixosaurus and the Jurassic Ichthyosaurus, evolved the dorsal fin and also much larger paddle-fins containing vastly increased numbers of phalangeal bones than those found in the feet of land-dwelling animals (Benton, 2005). And every step on that evolutionary transition was also historically contingent, also dependent upon chance genetic changes in the phyletic lineage of these particular animals and dependent upon chance changes in the environmental habitats in which these animals found themselves living during the span of time from the early Triassic to the Jurassic. To summarize, every step in the evolutionary transition of these two groups of land-dwelling, four-legged animals to marineadapted, fish-like animals was historically contingent, dependent upon chance and uncertain environmental conditions. The two groups of animals have radically different contingent histories: mammals are synapsid amniotes and reptiles are sauropsid amniotes. You have to go all the way back in time to the Carboniferous to find a common ancestor for these two clades of animals: their phylogenetic histories diverged over three hundred million years ago and are very different. The environmental conditions faced by the mammals in the span of time from the Eocene to the Miocene in the Cenozoic Era were certainly different from those faced by the reptiles in the span of time from the Triassic to the Jurassic in the Mesozoic Era. Yet in the contingent evolution (in both senses of contingent) of these two groups of animals the end result was the same. The convergent evolution of the ichthyosaur and porpoise reveals to us just how constrained the evolutionary process is on Earth, reveals to us just how limited the number of evolutionary pathways available to life are. And since the number of those pathways are limited, they are also predictable, but not with absolute certainty. For example, an ichthyosaur has four ventral fins (modified from the four limbs of its ancestors) whereas the porpoise has only two anterior ventral fins (its hind limbs have been lost). Their posterior fins are also different; the long-axis of the tail fin of the ichthyosaur is oriented vertically whereas that of the tail fin of the porpoise is oriented horizontally. However, one major morphological aspect is exactly the same in both animals: the streamlined, fusiform body, an adaptation for fast swimming. Thus I confidently predict that if any large, fast-swimming organisms exist in the oceans of Jupiter’s moon Europa then they will have streamlined, fusiform bodies. Their morphologies will be convergent upon those of the ichthyosaur or porpoise: nature has no other choice. In summary, I argue that convergent evolution provides empirical evidence that contingent evolution can be directional
Please cite this article in press as: McGhee, G. R., Jr., Can evolution be directional without being teleological?, Studies in History and Philosophy of Biological and Biomedical Sciences (2015), http://dx.doi.org/10.1016/j.shpsc.2015.12.006
4
G.R. McGhee Jr. / Studies in History and Philosophy of Biological and Biomedical Sciences xxx (2015) 1e7
(predictable within limits with finite possibilities), contrary to Gould’s argument that “contingent” evolution is a totally directionless and unpredictable process. On the other hand, I also argue that evolution can be directional without being teleological (totally predictable with a predetermined goal). 3. The actual spectrum of evolutionary possibilities In the previous sections it was argued that the following dichotomy is false: evolution is either “contingent” (directionless, unpredictable) or teleological (directional, predetermined). Contingent evolution is not directionless and unpredictable, and is thus not the antithesis of teleological evolution as framed in the false dichotomy. I argue that not only is the word “contingent” misused in the false dichotomy, but also that the actual spectrum of evolutionary possibilities goes beyond a simple dichotomy of two extremes. Biological evolution has a broader range of directional possibilities: the spectrum from total unpredictability (an infinite number of possible evolutionary pathways), to predictability within limits (a finite number of evolutionary pathways), to total predictability (one predetermined evolutionary pathway). The actual spectrum of evolutionary possibilities is better described by the following terminology, in which the word “contingent” is correctly used: the spectrum from unconstrained evolution (totally indeterminate, disordered), to contingent evolution (probabilistic, constrained), to teleological evolution (totally determinate, ordered). Gould’s view of the mechanism of evolution as directionless and unpredictable is more descriptive of an unconstrained, indeterminate process rather than a chance-dependent probabilistic or contingent process. Convergent evolution reveals to us that the number of possibilities available for contingent events is limited, and that contingent evolution is constrained to occur within a finite number of possible pathways, as will be discussed in the next section. 4. Convergent evolution: proof of constraint in contingent evolution Evidence that Gould’s view of evolution is in fact that of unconstrained evolution can be seen in his own favorite thoughtexperiment of “replaying life’s tape” in the evolutionary process (Gould, 1989). Here he imagined the history of the evolution of life on Earth to be similar to a videotape of a popular movie. He then tried to imagine what would happen if you could rewind the videotape to a point early in the movie, erasing everything that had happened in the movie after that point, and then rerun the videotape forward to see what would happen in the movie as it unfolded for the second time. Would the historical sequence of events in the evolution of life in the rerun of the movie be the same, or very similar, to the original movie? Or would evolution take radically different pathways such that the rerun of the movie would be totally different from the original movie? Gould (1989, 51) argued strongly for the second scenario: “Any replay of the tape would lead evolution down a pathway radically different from the road actually taken .. The diversity of possible itineraries does demonstrate that eventual results cannot be predicted at the outset. Each step proceeds for cause, but no finale can be specified at the start, and none would ever occur a second time in the same way, because any pathway proceeds through thousands of improbable stages.” Hence Gould’s view is that evolution is directionless and unpredictable, which is what one would expect from a process of unlimited, unconstrained evolution. However, Gould (1989, 284e 85) instead argued that his view of evolution was that of
“contingent” evolution, not unconstrained evolution: “Contingency is the affirmation of control by immediate events over destiny . Our own evolution is a joy and a wonder because such a curious chain of events would probably never happen again.” That is, he concluded that an evolutionary process that is a chain of contingent historical events would be unrepeatable and thus unpredictable. This conclusion leads directly to the false dichotomy of “contingency” versus teleology as two antithetical extremes rather than unconstrained evolution versus teleology as two antithetical extremes. Convergent evolution proves that Gould’s conclusion from his “replaying life’s tape” argument is demonstrably false. The fossil record preserves for us actual examples of such a contingent process at work: in the Mesozoic Era one run of the evolutionaryecological movie produced fish-like reptiles (ichthyosaurs) and a second run of the evolutionary-ecological movie in the Cenozoic Era produced fish-like mammals (porpoises), as previously discussed. That is, Gould’s conclusion that in “contingent” evolution, “Each step proceeds for cause, but no finale can be specified at the start, and none would ever occur a second time in the same way” (Gould, 1989, 51), is demonstrably false. Each step in the evolution of the ichthyosaur and the porpoise was historically contingent (both contingent per se genetic mutations and a contingent upon sequence of ecomorphotype ancestors), yet the finale (the evolution of a fish-like marine animal from a four-legged terrestrial animal) did occur a second time in the same way, contrary to Gould’s hypothetical conclusion. Why did Gould make this mistake? At the very outset of his argumentation, Gould states that the “diversity of possible itineraries does demonstrate that eventual results cannot be predicted” (1989, 51). The fossil record does not “demonstrate” this conclusion at all. The empirical pattern of convergent evolution actually reveals to us that the “diversity of possible itineraries” is finite and quite limited, as in the example regarding the evolution of the ichthyosaur and the porpoise. Thus Gould’s conclusion that “eventual results cannot be predicted” is also false, as is also demonstrated by the ichthyosaurs and porpoises. Why does life evolve convergently? Convergent evolution is produced by evolutionary constraint, particularly functional constraint. Functional constraints are determined by the laws of physics and geometry, and are extrinsic to the organisms affected by these constraints (McGhee, 2007, 2015). That is, from a theoretical perspective, functional constraints exist in the universe regardless of whether any biological organisms exist to experience them or not. Convergent evolution is the result of the fact that there are limited numbers of ways to solve a functional problem within the constraints imposed by the laws of physics and geometry. Fastswimming organisms in the Earth’s seas all have convergently evolved streamlined, fusiform bodies that minimize drag in swimming. We have considered in detail the convergent evolution of fusiform bodies by the ichthyosaurs (reptiles) and porpoises (mammals) and to this list we could add the fast-swimming great white sharks (cartilaginous fishes), swordfish (bony fishes), and penguins (birds). We believe that the laws of physics and geometry are the same throughout the universe, thus the functional constraints imposed by these laws should also be the same throughout the universe. Thus we can predict that the evolution of alien forms of life on Earth-like worlds should converge on the same functionally constrained forms that have been produced in the evolution of life on Earth: fast-swimming alien organisms will have fusiform bodies; organisms with powered flight will have wings; landdwelling, sessile, photoautotrophic organisms will have tree forms; and those tree forms will have leaf-like structures, and so on (McGhee, 2011).
Please cite this article in press as: McGhee, G. R., Jr., Can evolution be directional without being teleological?, Studies in History and Philosophy of Biological and Biomedical Sciences (2015), http://dx.doi.org/10.1016/j.shpsc.2015.12.006
G.R. McGhee Jr. / Studies in History and Philosophy of Biological and Biomedical Sciences xxx (2015) 1e7
5. Towards a predictive science of evolution The false dichotomy of “contingent” evolution versus teleological evolution is a major impediment to developing a predictive science of evolution. This false dichotomy underlies the pervasive negative reaction of the majority of evolutionary biologists to any statement that aspects of evolution are predictable: all claims of predictability in evolution must be teleological because “contingent” evolution is unpredictable. This false dichotomy must be exposed for what it is: false. Contingent evolution is constrained to operate within finite limits, within limited pathways, and is thus probabilistic and predictable. It is the very goal of science itself to seek general laws of nature. Yet Gould argues that this is not possible in “contingent” evolution, and that attempts to do so are teleological (see the quote at the beginning of this paper; Gould, 1989, 290). Gould ultimately argued in his final book, The Structure of Evolutionary Theory (2002, 1335), that in the analysis of evolution “answers must be sought in the particular and contingent prior histories of individual lineages, and not in general laws of nature that must affect all taxa in a coordinated and identical way.” Contrary to Gould’s vision for the structure of evolutionary theory, the phenomenon of convergent evolution is hard empirical evidence for forces or mechanisms in nature that in fact do affect all taxa in a coordinated and identical way. Cartilaginous fishes, bony fishes, reptiles, mammals, and birds were all constrained “in a coordinated and identical way” by selection for fluid-drag minimization to evolve swordfish, great white sharks, ichthyosaurs, porpoises, and penguins with fusiform streamlined body morphologies. I take it as an example of the pervasive negative influence of the false dichotomy of “contingency” versus teleology in evolutionary biology that the recent proposal of a general law of evolution, the constructal law, was made by an engineer, Adrian Bejan of Duke University, rather than by a biologist. In the constructal law (or principle of flow-efficiency improvement), it is argued that many biological configurations evolve in order to facilitate flow within the configuration. Bejan quickly discovered the negative influence of the false dichotomy in evolutionary biology when he used the word “design” in discussing the constructal law: “Design may be the foundation of the built world, but it is anathema when the conversation turns to nature .. If you want to send a chill across a lecture hall full of scientists, just mention design in nature” (Bejan & Zane, 2012, 29). At the same time, he clearly argues that there are no teleological goals in the evolutionary process: “Nature does not produce optima, or ‘end designs’ or ‘destiny.’ Nature is governed by the tendency to generate shapes and design that evolve in time to reduce imperfection. Design evolution never ends” (Bejan & Zane, 2012, 25). Evolution following the constructal law is both directional and predictable, yet not teleological: Everything that moves is a flow system that evolves over time; design generation and evolution are universal phenomena. The changes we witness in animals, plants . represent a clear improvement over the configuration that had been flowing before. This the direction of evolution, creating flows that move more easily, better, farther, etc. The design we see in nature . is a manifestation of this tendency in nature to generate shape and structure to facilitate flow access. This is direction and evolution without intention. Flow systems do not want to move more easily; they do not seek greater access for the currents that flow through them.” (Bejan & Zane, 2012, 31; italics in original) The constructal law is a “general law of nature” that affects all biological configurations “in a coordinated and identical way,” a
5
prime example of what Gould argued was not to be the goal of evolutionary biology. We need to study the empirical patterns of convergent evolution in our search for general laws or principles in evolutionary theory. Evolution is constrained to follow a limited number of pathways and we must map not only these existent pathways, but also map the nonexistent potential pathways that evolution cannot follow (at least on Earth). The analysis of both existent and nonexistent evolutionary pathways is necessary for the development of a fully predictive science of evolution. One way to map both existent and nonexistent evolutionary pathways by using adaptive landscapes (McGhee, 2007, 2015). Adaptive landscapes allow us to take a spatial approach to the concepts of natural selection and evolutionary constraint. Just as the deterministic force of gravity will eventually move water down to a sea or lake, where water has its lowest possible potential energy state, the deterministic process of natural selection will move evolving organisms up adaptive peaks, where they have their highest possible potential states of adaption (Fig. 1). It is the empirical observation from convergent evolution that the number of these higher states of adaptation, or adaptive peaks, is limited that gives evolution its direction. If we know the shape of a geographic landscapedwhere the topographic highs and lows are, and thus where the watersheds and lakes aredwe can predict in which direction water will flow and we can predict where it will eventually wind up. If there is only one lake in the geographic area of a particular watershed, we know with complete certainty where water from rainfall within the watershed area will wind up. If there is more than one lake in the watershed area, we can predict the probability that water will wind up in a particular given lake. By analogy, if we know the shape of the adaptive landscapedwhere the adaptive peaks and valleys aredthen we can predict in which direction a group of organisms will evolve and where that group will eventually wind up in the landscape. The challenge we face is the task of actually mapping that adaptive landscape, of spatially locating the adaptive peaks and valleys relative to each other within a grid metric of distances (Fig. 2). The task is similar to that faced in creating a topographic map. We start with a grid metric of distances, namely the grid of longitudinal and latitudinal distances. Then we measure the position of the mountains (topographic highs) and lakes (topographic lows) with reference to their latitude and longitude within the grid. In order to create a map of an adaptive landscape we must create a grid metric of distances analogous to the latitudeelongitude grid used in creating a topographic map. In morphological evolution, such a grid metric can be created by using the theoretical morphospace analytic technique (McGhee, 2007, 2015). Convergent evolution is of tremendous help in this process in that it reveals to us the adaptive peaks that exist in a given area on the adaptive landscape. But we must proceed further and locate the adaptive valleys as well, and then we must spatially arrange the adaptive peaks and valleys within a grid metric of distances (Fig. 2). Then we will be able to map the limited pathways that contingent evolution has taken in the past, and to predict the direction that contingent evolution may take in the future. 6. Reclaiming our proper evolutionary terminology In summary, we must be able to speak once again of direction, predictability, and design in contingent evolution without being accused of teleology. We must expose Gould’s dichotomy of “contingent” evolution versus teleological evolution (Gould, 1989, 2002; Sapp, 2012) as false in order to reengage evolutionary biologists in the analysis of directionality and design in evolution. For
Please cite this article in press as: McGhee, G. R., Jr., Can evolution be directional without being teleological?, Studies in History and Philosophy of Biological and Biomedical Sciences (2015), http://dx.doi.org/10.1016/j.shpsc.2015.12.006
6
G.R. McGhee Jr. / Studies in History and Philosophy of Biological and Biomedical Sciences xxx (2015) 1e7
Fig. 1. Modeling the effect of natural selection in an adaptive landscape with vectors. Natural selection favors organisms with morphological traits (xey axes) that possess higher (better) degrees of adaption (z-axis); evolution under natural selection will always proceed in the upslope direction within the landscape. Therefore, under natural selection theory, evolution is directional. (Figure from McGhee, 2007).
Fig. 2. Two adaptive peaks of functional possible forms (FPF) located in a flat plain of nonfunctional possible forms (NPF) in an adaptive landscape. The spatial arrangement of the adaptive peaks within the flat plain is specified with reference to the grid metric of measured morphological traits (xey axes). (Figure from McGhee, 2007).
the general public we must expose the mischaracterization of contingent evolution as totally random, directionless, and unpredictable as the product not only of the anti-evolutionists but also of some of our own evolutionary biologists. For that same general public we must rekindle the sense of wonder in the natural evolution of life on Earth by wresting the terms design and direction in evolution away from the intelligentdesign creationists. Contingent biological evolution can produce both design and direction through natural processes, contrary to the intelligent-design creationists’ claim that complex organic design and directional evolution can only be produced by interventionist supernatural processes. The empirical pattern of convergent evolution in the fossil record can also be a tremendous
help to us in this terminological struggle. The phenomenon of convergence proves that truly contingent evolutionary processes can repeatedly produce the same, or very similar, organic designs in nature and that evolution is directional. To paraphrase Charles Darwin’s famous last sentence in On the Origin of Species (1859, 490), the “limited forms most beautiful and most wonderful” that surround us on Earth are a simple natural product of the finite number of evolutionary pathways that exist in our universe. Acknowledgments I thank Peter Harrison for the invitation to the Oxford workshop on contingency in evolution, Ian Hesketh for his logistical help at
Please cite this article in press as: McGhee, G. R., Jr., Can evolution be directional without being teleological?, Studies in History and Philosophy of Biological and Biomedical Sciences (2015), http://dx.doi.org/10.1016/j.shpsc.2015.12.006
G.R. McGhee Jr. / Studies in History and Philosophy of Biological and Biomedical Sciences xxx (2015) 1e7
the workshop and in manuscript editing, and John Beatty for valuable discussions at the workshop.
References Bejan, A., & Zane, J. (2012). Design in nature: How the constructal law governs evolution in Biology, Physics, Technology, and Social Organization. New York: Anchor Books. Benton, M. J. (2005). Vertebrate palaeontology. Oxford: Blackwell Publishing. Darwin, C. (1859). On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life. London: John Murray. Gould, S. J. (1989). Wonderful life: The Burgess Shale and the nature of history. New York: W. W. Norton. Gould, S. J. (2002). The structure of evolutionary theory. Cambridge, MA: Belknap Press of Harvard University Press. Klimentidis, Y. C. (2012). On the limits of diversity. Review of convergent Evolution: Limited forms most beautiful, by George McGhee Frontiers in Genetics, 3. article 136, pp. 1e2.
7
McGhee, G. R., Jr. (2007). The geometry of evolution: Adaptive landscapes and theoretical morphospaces. Cambridge: Cambridge University Press. McGhee, G. R., Jr. (2011). Convergent evolution: Limited forms most beautiful. Cambridge: Massachusetts Institute of Technology Press. McGhee, G. R., Jr. (2015). Limits in the evolution of biological form: A theoretical morphologic perspective. Interface Focus, 5(1e6), 20150034. http://dx.doi.org/ 10.1098/rsfs.2015.0034. Raup, D. M. (1977a). Probabilistic models in evolutionary paleobiology. American Scientist, 65(1), 50e57. Raup, D. M. (1977b). Stochastic models in evolutionary palaeontology. In A. Hallam (Ed.), Vols. 59e78. Patterns of evolution. Amsterdam: Elsevier Scientific Publishing Company. Raup, D. M., & Gould, S. J. (1974). Stochastic simulation and evolution of morphology: towards a nomothetic paleontology. Systematic Zoology, 23(3), 305e322. Reiss, J. O. (2009). Not by design: Retiring Darwin’s watchmaker. Berkeley: University of California Press. Sapp, J. (2012). Evolution replayed. BioScience, 62(7), 693e694. Wilson, E. O., & Bossert, W. H. (1971). A primer of population biology. Sunderland, MA: Sinauer Associates.
Please cite this article in press as: McGhee, G. R., Jr., Can evolution be directional without being teleological?, Studies in History and Philosophy of Biological and Biomedical Sciences (2015), http://dx.doi.org/10.1016/j.shpsc.2015.12.006
Contents lists available at ScienceDirect
Studies in History and Philosophy of Biological and Biomedical Sciences journal homepage: www.elsevier.com/locate/shpsc
Can evolution be directional without being teleological? George R. McGhee Jr. Department of Earth and Planetary Sciences, Wright-Rieman Laboratories, Rutgers University, 610 Taylor Road, Pistacaway, NJ 08854 USA
a r t i c l e i n f o
a b s t r a c t
Article history: Available online xxx
Convergent evolution reveals to us that the number of possibilities available for contingent events is limited, that historically contingent evolution is constrained to occur within a finite number of limited pathways, and that contingent evolution is thus probabilistic and predictable. That is, the phenomenon of convergence proves that truly contingent evolutionary processes can repeatedly produce the same, or very similar, organic designs in nature and that evolution is directional in these cases. For this reason it is argued in this paper that evolution can be directional without being teleological, and that the dichotomy that evolution must either be directionless and unpredictable or directional and predetermined (teleological) is false. Ó 2015 Published by Elsevier Ltd.
Keywords: Directional evolution Convergence Contingency Teleology
When citing this paper, please use the full journal title Studies in History and Philosophy of Biological and Biomedical Sciences
1. The false dichotomy of contingency versus teleology in evolution The York University biologist Jan Sapp (2012) argues that a false dichotomy has been erected in the construction of the evolutionary contingency-versus-directionality debate, namely “the false dichotomy between the factors of contingency and chance on one hand and directed evolution based on supernatural forces on the other. In the history of evolutionary biology, it is chance, the struggle for existence, and contingency that have always been on the side of evolution; goal-driven physiological change was always on the side of supernaturalism and the antievolution movement. This dichotomy has been maintained by an ongoing conflict between scientists and creationists” (Sapp, 2012, 694). Why is this dichotomy false? It is false because evolution can be directional without being teleological (McGhee, 2011, 272e73). The false dichotomy that evolution must either be directionless or unpredictable (a possibility that is mislabeled “contingent” in the false dichotomy) or directional and predetermined (teleological) repeatedly appears in the many arguments of Stephen Jay Gould, the Harvard paleontologist and best-known evolutionary essayist of the late twentieth century. For example, Gould (1989) frames this
E-mail address: [email protected].
false dichotomy as the “question of questions” in his argument against the views of the early twentieth-century paleontologist Charles Doolittle Walcott: “ultimately, the question of questions boils down to the placement of the boundary between predictability under invariant law and the multifarious possibilities of historical contingency. Traditionalists like Walcott would place the boundary so low that all major patterns of life’s history fall above the line into the realm of predictability (and, for him, direct manifestation of divine intentions). But I envision a boundary sitting so high that almost every interesting event in life’s history falls into the realm of contingency” (Gould, 1989, 290); that is, into the realm of unpredictability. The teleological endpoint in the false dichotomy certainly exists as such. That is, under a teleological view evolution is directional and predetermineddteleology is defined as “the fact or quality of being directed toward a definite end or having an ultimate purpose, especially as attributed to natural processes.”1 But what about the opposite endpoint in the false dichotomy, the evolutionary “factors of contingency and chance” referred to by Sapp (2012, 694)? In his article in this issue John Beatty differentiates between “contingent per se,” in which a given evolutionary event was a matter of chance (such as a genetic mutation), and
1
Webster’s New World Dictionary, 2nd college ed., s.v. “teleology.”
http://dx.doi.org/10.1016/j.shpsc.2015.12.006 1369-8486/Ó 2015 Published by Elsevier Ltd.
Please cite this article in press as: McGhee, G. R., Jr., Can evolution be directional without being teleological?, Studies in History and Philosophy of Biological and Biomedical Sciences (2015), http://dx.doi.org/10.1016/j.shpsc.2015.12.006
2
G.R. McGhee Jr. / Studies in History and Philosophy of Biological and Biomedical Sciences xxx (2015) 1e7
“contingent upon,” in which a given evolutionary event depended upon the occurrence of a prior event (such as the evolution of an ancestor). Does contingency in either of these two senses really require that evolution be directionless and unpredictable? At first glance it might seem so, as in our current understanding biological evolution is generally modeled to be the outcome of the interaction of chance (random genetic mutations) and uncertain conditions (changing environmental habitats in nature); that is, evolution is the product of genetic mutations and natural selection, where genetic mutations are seen as the source of new biological variation and natural selection is the mechanism sorting the biological variants in terms of their differential adaptations to different environmental conditions. In fact, both of these components of the standard model of evolution can produce directional trends. First, it is a well-known mathematical fact that random processes, chance processes, can product directional trends. The best known of these are Markov chains. A Markov process produces “a sequence of events such that each event is partly dependent on the outcomes of preceding events and partly dependent on a random process acting at the time of the event itself” (Raup, 1977b, 62). Computer simulations using Markov chains are thus an ideal way to explore temporal processes that are both contingent upon (Raup’s “each event is partly dependent on the outcomes of preceding events”) and contingent per se (Raup’s “partly dependent on a random process acting at the time of the event itself”), and can be used to simulate random walks in a single evolving phylogenetic lineage or random evolutionary branching in multiple phylogenetic lineages (Raup, 1977a, 1977b). In the field of evolutionary paleobiology itself, the University of Chicago theoretician David M. Raup ran computer simulations of chance (contingent per se) and historically contingent changes (contingent upon) in the morphologies of hypothetically evolving species. David Sepkoski has provided an account of these studies in his article in the issue. Raup’s early work (Raup & Gould, 1974) demonstrated that random simulations can produce: (1) apparent directional trends in morphological change in evolving lineages, (2) apparent correlation between combinations of morphological traits within an evolving lineage, (3) variation between rates of morphological change in different evolving lineages, (4) apparent “terminal” overspecialization within an evolving lineage, and (5) clumping of randomly evolved morphologies within a theoretical morphospace of the hypothetical spectrum of forms that potentially could be evolved. Prior to Raup’s random simulations, all of these phenomena were thought to be produced only by deterministic evolutionary processes. Thus Raup’s computer simulations demonstrated that chance processes (contingent per se) and historically contingent processes (contingent upon) could produce directional trends; that is, truly contingent evolution (in both senses of contingent) is not necessarily a directionless process. Second, natural selection is a deterministic process, not a random one. Wilson and Bossert (1971, 40) point out that natural selection is a deterministic process in that it involves “directions and rates that can be measured in populations and used to predict specific outcomes” and that only evolution via genetic drift is a stochastic process in which only probabilities of outcomes, not a specific outcome, may be predicted. The mechanism of natural selection differentially sorts biological variants in terms of their differential adaptations to environmental conditions, and differentially preserves the variants with better states of adaptation. Thus natural selection is a directional process, one in which the frequency of organisms with better states of adaptation increases with time in evolving populations (given constancy in the environmental conditions producing the selection). The fact that Charles Darwin’s concept of natural selection was both deterministic and directional has led Humboldt State
University biologist John Reiss (2009, 140) to argue that the very concept itself is teleological: “Darwin introduced a teleological determinism into the heart of his theory. This teleology is expressed in two related conceptions: (1) that evolution is a process going from a less-adapted to a better-adapted state and (2) that natural selection is a deterministic force, or agent, that directs the evolutionary process toward this better-adapted state.” That is, Reiss considers the very idea of adaptive improvement via natural selection in evolution to be teleological because improvement is a directional concept and any hint of directionality in evolution is teleological. Thus the false dichotomy that directionality in evolution must be teleological because contingent evolution is directionless lies at the heart of Reiss’s argument. Is it teleological that water flows in the downhill direction? That it goes from a state of higher potential energy to a state of lower potential energy under the influence of the deterministic force of gravity? Philosophers may argue whether that phenomenon is teleological or not, but in science it is an empirical observation, an established fact. Why then would it be teleological if organisms evolve in the direction of better states of adaptation under the deterministic force of natural selection? 2. How contingent evolution gives rise to predictable directionality: nature has limited choice Contingent evolution (in both senses) can be directional with being teleological, but can the direction of evolution be predictable without being teleological? This question addresses the second part of the false dichotomy that evolution is either directionless or unpredictable (a possibility that is mislabeled “contingent” in the false dichotomy) or directional and predetermined (teleological). Certainly any natural process that is “directed toward a definite end or having an ultimate purpose” can be considered to be teleological.1 Thus Reiss further objects to the concept of natural selection as a “deterministic force, or agent, that directs the evolutionary process” towards a better-adapted state because this “better-adapted state” appears to be a goal, and that a goal-directed process is teleological (Reiss, 2009, 140). Is it teleological that water flows in rivers toward the future “goal” of reaching a sea or lake? The flow of water is mindless, and water has no goal in sight, but nevertheless it will reach its lowest possible potential energy state under the deterministic force of gravity, which means that it will wind up in a sea or lake. And we can confidently predict that flowing water will wind up in a sea or lake although we may not be able to predict which sea or lake without a great deal of additional information about the geography and topography of the landscape across which the water is flowing. Natural selection is a deterministic process and thus, theoretically, one should be able to predict its outcome if one knows the magnitudes of all of the variables at work in the process (Wilson & Bossert, 1971). In actual practice that is a very difficult task indeed. In addition, the element of randomness does enter the equation of natural selection in that the state of the environmental conditions producing the natural selection mechanism (of sorting biological variants in terms of their differential adaptations to those environmental conditions) are uncertain. That is, we may be able to calculate the probabilities that environmental conditions will change in this direction or the other, but we cannot do so with absolute certainty. We are all familiar with the uncertainties involved with weather forecasting where predicted future weather conditions are given in probabilities, not certainties. It is at this point that the empirical pattern of actual evolution on Earth helps us out tremendously: much of evolution is convergent. Convergent evolution reveals to us that the number of evolutionary pathways available to life is not endless, but is instead quite limited.
Please cite this article in press as: McGhee, G. R., Jr., Can evolution be directional without being teleological?, Studies in History and Philosophy of Biological and Biomedical Sciences (2015), http://dx.doi.org/10.1016/j.shpsc.2015.12.006
G.R. McGhee Jr. / Studies in History and Philosophy of Biological and Biomedical Sciences xxx (2015) 1e7
That is, “convergent evolution results from the finite number of ecological niches on our planet, each of which is exploited by many organisms that share a common origin” (Klimentidis, 2012, 1). Even given the many uncertainties in predicting the pathway direction of natural-selection-induced trends in evolution, those uncertainties are considerably reduced when it is realized that the actual number of pathways that evolution may take are limited. We can easily imagine a universe where that fact is not true; that is, a world in which the number of pathways available to evolution are limitless, are infinite. In such a universe there would be so many alternative pathways possible (the “multifarious possibilities” of Gould [1989, 120]) that it would never be possible to predict the trajectory of the evolution of life. In such a universe each individual species would have its own, unique niche or ecological role in nature and each species would have its own, unique adaptive form or ecomorphotype within the selectional parameters of its particular niche or ecological role. That universe does not exist. Instead, we live in a universe where convergence in evolution is rampant at every level, from the ecological roles of organisms and their adaptive forms down to the molecules from which those forms are constructed. Multiple species from disparate phyletic lineages have evolved to occupy the very same niche, to play the same ecological role in nature. And, as a selectional consequence, multiple species have convergently evolved the very same adaptive form, or ecomorphotype. What is astounding on our real Earth is that each of these convergent ecomorphotypes have evolved in a historically contingent process. For example, consider the evolution of the porpoise (or whales in general). Whales and porpoises are openocean dwelling mammals, but they evolved from land-dwelling mammals ecologically similar to the modern-day hippopotamus. A hippopotamus is an artiodactyl mammal adapted to the environmental conditions of a freshwater, semi-aquatic habitat. If we go back to the early Eocene in geologic time we find that other artiodactyl mammals had also evolved to live in freshwater, semi-aquatic habitats: animals like Pakicetus and Indohyus. Some of these animals evolved the adaptations needed to move into more brackish-water habitats, evolved the adaptations need to deal with osmotic problems associated with higher salt contents in the waters in which they lived: animals like Ambulocetus. Venturing further offshore, some of these animals evolved the adaptations needed to live in saltwater habitats and also to breathe while floating in water (instead of standing on land) by shifting their nostrils upwards on their skulls: animals like Kutchicetus and Rodhocetus. By the late Eocene some of these animals evolved tail flukes to better propel themselves through water in swimming, and their unused hind legs became smaller and smaller: animals like Dorudon and Basilosaurus. And finally in the Miocene we have the evolution of the open-ocean habitat mammals, such as the dolphin Kentriodon, which have no hind legs at all, whose forelimbs are fins, and which have tail fins and dorsal fins (Benton, 2005). A porpoise or dolphin looks like a fish: it has a fusiform, streamlined body like that of a swordfish or tuna, it has fins instead of legs, a large fin on its posterior end instead of a tail, and even has a fin on its back. The transformation of land-dwelling, four-legged mammals with tails and no fins on their backs into open-ocean dwelling mammals that look like fish is astonishing but empirically verifiable in the fossil record. And every step on that evolutionary transition was historically contingent (contingent upon), dependent upon chance genetic changes (contingent per se) in the phyletic lineage of these particular animals and dependent upon chance changes in the environmental habitats in which these animals found themselves living in the span of time from the early Eocene to the Miocene.
3
Surely such an astonishing evolutionary transition has to be unique, a contingent series of events (in both senses of contingent) that could never ever be repeated in the history of life on Earth. Yet, even more astonishing, is the empirically verifiable fact that that same evolutionary transition happened in the Mesozoic with the transformation of land-dwelling, four-legged reptiles with tails and no fins on their backs into open-ocean dwelling reptiles that look like fish: the ichthyosaurs. The evolution of ichthyosaurs has not been studied with the detail given that of whales, but the general trends can still be seen. In the late Permian a group of fully terrestrial basal neodiapsids, animals like Youngina with slender bodies and short limbs, gave rise to animals that moved into freshwater fluvial habitats and then evolved into marginal-marineadapted animals like Placodus, which had very short but still functioning legs with feet, and shallow-marine-adapted animals like Pachypleurosaurus, which had four limbs with paddle fins instead of feet, in the Triassic. The oldest-known open-ocean dwelling ichthyosaur is the early Triassis Utatsusaurus, an animal with four ventral fins and a tail fin but which had no dorsal fin like a porpoise. More derived ichthyosaurs, like the late Triassic Mixosaurus and the Jurassic Ichthyosaurus, evolved the dorsal fin and also much larger paddle-fins containing vastly increased numbers of phalangeal bones than those found in the feet of land-dwelling animals (Benton, 2005). And every step on that evolutionary transition was also historically contingent, also dependent upon chance genetic changes in the phyletic lineage of these particular animals and dependent upon chance changes in the environmental habitats in which these animals found themselves living during the span of time from the early Triassic to the Jurassic. To summarize, every step in the evolutionary transition of these two groups of land-dwelling, four-legged animals to marineadapted, fish-like animals was historically contingent, dependent upon chance and uncertain environmental conditions. The two groups of animals have radically different contingent histories: mammals are synapsid amniotes and reptiles are sauropsid amniotes. You have to go all the way back in time to the Carboniferous to find a common ancestor for these two clades of animals: their phylogenetic histories diverged over three hundred million years ago and are very different. The environmental conditions faced by the mammals in the span of time from the Eocene to the Miocene in the Cenozoic Era were certainly different from those faced by the reptiles in the span of time from the Triassic to the Jurassic in the Mesozoic Era. Yet in the contingent evolution (in both senses of contingent) of these two groups of animals the end result was the same. The convergent evolution of the ichthyosaur and porpoise reveals to us just how constrained the evolutionary process is on Earth, reveals to us just how limited the number of evolutionary pathways available to life are. And since the number of those pathways are limited, they are also predictable, but not with absolute certainty. For example, an ichthyosaur has four ventral fins (modified from the four limbs of its ancestors) whereas the porpoise has only two anterior ventral fins (its hind limbs have been lost). Their posterior fins are also different; the long-axis of the tail fin of the ichthyosaur is oriented vertically whereas that of the tail fin of the porpoise is oriented horizontally. However, one major morphological aspect is exactly the same in both animals: the streamlined, fusiform body, an adaptation for fast swimming. Thus I confidently predict that if any large, fast-swimming organisms exist in the oceans of Jupiter’s moon Europa then they will have streamlined, fusiform bodies. Their morphologies will be convergent upon those of the ichthyosaur or porpoise: nature has no other choice. In summary, I argue that convergent evolution provides empirical evidence that contingent evolution can be directional
Please cite this article in press as: McGhee, G. R., Jr., Can evolution be directional without being teleological?, Studies in History and Philosophy of Biological and Biomedical Sciences (2015), http://dx.doi.org/10.1016/j.shpsc.2015.12.006
4
G.R. McGhee Jr. / Studies in History and Philosophy of Biological and Biomedical Sciences xxx (2015) 1e7
(predictable within limits with finite possibilities), contrary to Gould’s argument that “contingent” evolution is a totally directionless and unpredictable process. On the other hand, I also argue that evolution can be directional without being teleological (totally predictable with a predetermined goal). 3. The actual spectrum of evolutionary possibilities In the previous sections it was argued that the following dichotomy is false: evolution is either “contingent” (directionless, unpredictable) or teleological (directional, predetermined). Contingent evolution is not directionless and unpredictable, and is thus not the antithesis of teleological evolution as framed in the false dichotomy. I argue that not only is the word “contingent” misused in the false dichotomy, but also that the actual spectrum of evolutionary possibilities goes beyond a simple dichotomy of two extremes. Biological evolution has a broader range of directional possibilities: the spectrum from total unpredictability (an infinite number of possible evolutionary pathways), to predictability within limits (a finite number of evolutionary pathways), to total predictability (one predetermined evolutionary pathway). The actual spectrum of evolutionary possibilities is better described by the following terminology, in which the word “contingent” is correctly used: the spectrum from unconstrained evolution (totally indeterminate, disordered), to contingent evolution (probabilistic, constrained), to teleological evolution (totally determinate, ordered). Gould’s view of the mechanism of evolution as directionless and unpredictable is more descriptive of an unconstrained, indeterminate process rather than a chance-dependent probabilistic or contingent process. Convergent evolution reveals to us that the number of possibilities available for contingent events is limited, and that contingent evolution is constrained to occur within a finite number of possible pathways, as will be discussed in the next section. 4. Convergent evolution: proof of constraint in contingent evolution Evidence that Gould’s view of evolution is in fact that of unconstrained evolution can be seen in his own favorite thoughtexperiment of “replaying life’s tape” in the evolutionary process (Gould, 1989). Here he imagined the history of the evolution of life on Earth to be similar to a videotape of a popular movie. He then tried to imagine what would happen if you could rewind the videotape to a point early in the movie, erasing everything that had happened in the movie after that point, and then rerun the videotape forward to see what would happen in the movie as it unfolded for the second time. Would the historical sequence of events in the evolution of life in the rerun of the movie be the same, or very similar, to the original movie? Or would evolution take radically different pathways such that the rerun of the movie would be totally different from the original movie? Gould (1989, 51) argued strongly for the second scenario: “Any replay of the tape would lead evolution down a pathway radically different from the road actually taken .. The diversity of possible itineraries does demonstrate that eventual results cannot be predicted at the outset. Each step proceeds for cause, but no finale can be specified at the start, and none would ever occur a second time in the same way, because any pathway proceeds through thousands of improbable stages.” Hence Gould’s view is that evolution is directionless and unpredictable, which is what one would expect from a process of unlimited, unconstrained evolution. However, Gould (1989, 284e 85) instead argued that his view of evolution was that of
“contingent” evolution, not unconstrained evolution: “Contingency is the affirmation of control by immediate events over destiny . Our own evolution is a joy and a wonder because such a curious chain of events would probably never happen again.” That is, he concluded that an evolutionary process that is a chain of contingent historical events would be unrepeatable and thus unpredictable. This conclusion leads directly to the false dichotomy of “contingency” versus teleology as two antithetical extremes rather than unconstrained evolution versus teleology as two antithetical extremes. Convergent evolution proves that Gould’s conclusion from his “replaying life’s tape” argument is demonstrably false. The fossil record preserves for us actual examples of such a contingent process at work: in the Mesozoic Era one run of the evolutionaryecological movie produced fish-like reptiles (ichthyosaurs) and a second run of the evolutionary-ecological movie in the Cenozoic Era produced fish-like mammals (porpoises), as previously discussed. That is, Gould’s conclusion that in “contingent” evolution, “Each step proceeds for cause, but no finale can be specified at the start, and none would ever occur a second time in the same way” (Gould, 1989, 51), is demonstrably false. Each step in the evolution of the ichthyosaur and the porpoise was historically contingent (both contingent per se genetic mutations and a contingent upon sequence of ecomorphotype ancestors), yet the finale (the evolution of a fish-like marine animal from a four-legged terrestrial animal) did occur a second time in the same way, contrary to Gould’s hypothetical conclusion. Why did Gould make this mistake? At the very outset of his argumentation, Gould states that the “diversity of possible itineraries does demonstrate that eventual results cannot be predicted” (1989, 51). The fossil record does not “demonstrate” this conclusion at all. The empirical pattern of convergent evolution actually reveals to us that the “diversity of possible itineraries” is finite and quite limited, as in the example regarding the evolution of the ichthyosaur and the porpoise. Thus Gould’s conclusion that “eventual results cannot be predicted” is also false, as is also demonstrated by the ichthyosaurs and porpoises. Why does life evolve convergently? Convergent evolution is produced by evolutionary constraint, particularly functional constraint. Functional constraints are determined by the laws of physics and geometry, and are extrinsic to the organisms affected by these constraints (McGhee, 2007, 2015). That is, from a theoretical perspective, functional constraints exist in the universe regardless of whether any biological organisms exist to experience them or not. Convergent evolution is the result of the fact that there are limited numbers of ways to solve a functional problem within the constraints imposed by the laws of physics and geometry. Fastswimming organisms in the Earth’s seas all have convergently evolved streamlined, fusiform bodies that minimize drag in swimming. We have considered in detail the convergent evolution of fusiform bodies by the ichthyosaurs (reptiles) and porpoises (mammals) and to this list we could add the fast-swimming great white sharks (cartilaginous fishes), swordfish (bony fishes), and penguins (birds). We believe that the laws of physics and geometry are the same throughout the universe, thus the functional constraints imposed by these laws should also be the same throughout the universe. Thus we can predict that the evolution of alien forms of life on Earth-like worlds should converge on the same functionally constrained forms that have been produced in the evolution of life on Earth: fast-swimming alien organisms will have fusiform bodies; organisms with powered flight will have wings; landdwelling, sessile, photoautotrophic organisms will have tree forms; and those tree forms will have leaf-like structures, and so on (McGhee, 2011).
Please cite this article in press as: McGhee, G. R., Jr., Can evolution be directional without being teleological?, Studies in History and Philosophy of Biological and Biomedical Sciences (2015), http://dx.doi.org/10.1016/j.shpsc.2015.12.006
G.R. McGhee Jr. / Studies in History and Philosophy of Biological and Biomedical Sciences xxx (2015) 1e7
5. Towards a predictive science of evolution The false dichotomy of “contingent” evolution versus teleological evolution is a major impediment to developing a predictive science of evolution. This false dichotomy underlies the pervasive negative reaction of the majority of evolutionary biologists to any statement that aspects of evolution are predictable: all claims of predictability in evolution must be teleological because “contingent” evolution is unpredictable. This false dichotomy must be exposed for what it is: false. Contingent evolution is constrained to operate within finite limits, within limited pathways, and is thus probabilistic and predictable. It is the very goal of science itself to seek general laws of nature. Yet Gould argues that this is not possible in “contingent” evolution, and that attempts to do so are teleological (see the quote at the beginning of this paper; Gould, 1989, 290). Gould ultimately argued in his final book, The Structure of Evolutionary Theory (2002, 1335), that in the analysis of evolution “answers must be sought in the particular and contingent prior histories of individual lineages, and not in general laws of nature that must affect all taxa in a coordinated and identical way.” Contrary to Gould’s vision for the structure of evolutionary theory, the phenomenon of convergent evolution is hard empirical evidence for forces or mechanisms in nature that in fact do affect all taxa in a coordinated and identical way. Cartilaginous fishes, bony fishes, reptiles, mammals, and birds were all constrained “in a coordinated and identical way” by selection for fluid-drag minimization to evolve swordfish, great white sharks, ichthyosaurs, porpoises, and penguins with fusiform streamlined body morphologies. I take it as an example of the pervasive negative influence of the false dichotomy of “contingency” versus teleology in evolutionary biology that the recent proposal of a general law of evolution, the constructal law, was made by an engineer, Adrian Bejan of Duke University, rather than by a biologist. In the constructal law (or principle of flow-efficiency improvement), it is argued that many biological configurations evolve in order to facilitate flow within the configuration. Bejan quickly discovered the negative influence of the false dichotomy in evolutionary biology when he used the word “design” in discussing the constructal law: “Design may be the foundation of the built world, but it is anathema when the conversation turns to nature .. If you want to send a chill across a lecture hall full of scientists, just mention design in nature” (Bejan & Zane, 2012, 29). At the same time, he clearly argues that there are no teleological goals in the evolutionary process: “Nature does not produce optima, or ‘end designs’ or ‘destiny.’ Nature is governed by the tendency to generate shapes and design that evolve in time to reduce imperfection. Design evolution never ends” (Bejan & Zane, 2012, 25). Evolution following the constructal law is both directional and predictable, yet not teleological: Everything that moves is a flow system that evolves over time; design generation and evolution are universal phenomena. The changes we witness in animals, plants . represent a clear improvement over the configuration that had been flowing before. This the direction of evolution, creating flows that move more easily, better, farther, etc. The design we see in nature . is a manifestation of this tendency in nature to generate shape and structure to facilitate flow access. This is direction and evolution without intention. Flow systems do not want to move more easily; they do not seek greater access for the currents that flow through them.” (Bejan & Zane, 2012, 31; italics in original) The constructal law is a “general law of nature” that affects all biological configurations “in a coordinated and identical way,” a
5
prime example of what Gould argued was not to be the goal of evolutionary biology. We need to study the empirical patterns of convergent evolution in our search for general laws or principles in evolutionary theory. Evolution is constrained to follow a limited number of pathways and we must map not only these existent pathways, but also map the nonexistent potential pathways that evolution cannot follow (at least on Earth). The analysis of both existent and nonexistent evolutionary pathways is necessary for the development of a fully predictive science of evolution. One way to map both existent and nonexistent evolutionary pathways by using adaptive landscapes (McGhee, 2007, 2015). Adaptive landscapes allow us to take a spatial approach to the concepts of natural selection and evolutionary constraint. Just as the deterministic force of gravity will eventually move water down to a sea or lake, where water has its lowest possible potential energy state, the deterministic process of natural selection will move evolving organisms up adaptive peaks, where they have their highest possible potential states of adaption (Fig. 1). It is the empirical observation from convergent evolution that the number of these higher states of adaptation, or adaptive peaks, is limited that gives evolution its direction. If we know the shape of a geographic landscapedwhere the topographic highs and lows are, and thus where the watersheds and lakes aredwe can predict in which direction water will flow and we can predict where it will eventually wind up. If there is only one lake in the geographic area of a particular watershed, we know with complete certainty where water from rainfall within the watershed area will wind up. If there is more than one lake in the watershed area, we can predict the probability that water will wind up in a particular given lake. By analogy, if we know the shape of the adaptive landscapedwhere the adaptive peaks and valleys aredthen we can predict in which direction a group of organisms will evolve and where that group will eventually wind up in the landscape. The challenge we face is the task of actually mapping that adaptive landscape, of spatially locating the adaptive peaks and valleys relative to each other within a grid metric of distances (Fig. 2). The task is similar to that faced in creating a topographic map. We start with a grid metric of distances, namely the grid of longitudinal and latitudinal distances. Then we measure the position of the mountains (topographic highs) and lakes (topographic lows) with reference to their latitude and longitude within the grid. In order to create a map of an adaptive landscape we must create a grid metric of distances analogous to the latitudeelongitude grid used in creating a topographic map. In morphological evolution, such a grid metric can be created by using the theoretical morphospace analytic technique (McGhee, 2007, 2015). Convergent evolution is of tremendous help in this process in that it reveals to us the adaptive peaks that exist in a given area on the adaptive landscape. But we must proceed further and locate the adaptive valleys as well, and then we must spatially arrange the adaptive peaks and valleys within a grid metric of distances (Fig. 2). Then we will be able to map the limited pathways that contingent evolution has taken in the past, and to predict the direction that contingent evolution may take in the future. 6. Reclaiming our proper evolutionary terminology In summary, we must be able to speak once again of direction, predictability, and design in contingent evolution without being accused of teleology. We must expose Gould’s dichotomy of “contingent” evolution versus teleological evolution (Gould, 1989, 2002; Sapp, 2012) as false in order to reengage evolutionary biologists in the analysis of directionality and design in evolution. For
Please cite this article in press as: McGhee, G. R., Jr., Can evolution be directional without being teleological?, Studies in History and Philosophy of Biological and Biomedical Sciences (2015), http://dx.doi.org/10.1016/j.shpsc.2015.12.006
6
G.R. McGhee Jr. / Studies in History and Philosophy of Biological and Biomedical Sciences xxx (2015) 1e7
Fig. 1. Modeling the effect of natural selection in an adaptive landscape with vectors. Natural selection favors organisms with morphological traits (xey axes) that possess higher (better) degrees of adaption (z-axis); evolution under natural selection will always proceed in the upslope direction within the landscape. Therefore, under natural selection theory, evolution is directional. (Figure from McGhee, 2007).
Fig. 2. Two adaptive peaks of functional possible forms (FPF) located in a flat plain of nonfunctional possible forms (NPF) in an adaptive landscape. The spatial arrangement of the adaptive peaks within the flat plain is specified with reference to the grid metric of measured morphological traits (xey axes). (Figure from McGhee, 2007).
the general public we must expose the mischaracterization of contingent evolution as totally random, directionless, and unpredictable as the product not only of the anti-evolutionists but also of some of our own evolutionary biologists. For that same general public we must rekindle the sense of wonder in the natural evolution of life on Earth by wresting the terms design and direction in evolution away from the intelligentdesign creationists. Contingent biological evolution can produce both design and direction through natural processes, contrary to the intelligent-design creationists’ claim that complex organic design and directional evolution can only be produced by interventionist supernatural processes. The empirical pattern of convergent evolution in the fossil record can also be a tremendous
help to us in this terminological struggle. The phenomenon of convergence proves that truly contingent evolutionary processes can repeatedly produce the same, or very similar, organic designs in nature and that evolution is directional. To paraphrase Charles Darwin’s famous last sentence in On the Origin of Species (1859, 490), the “limited forms most beautiful and most wonderful” that surround us on Earth are a simple natural product of the finite number of evolutionary pathways that exist in our universe. Acknowledgments I thank Peter Harrison for the invitation to the Oxford workshop on contingency in evolution, Ian Hesketh for his logistical help at
Please cite this article in press as: McGhee, G. R., Jr., Can evolution be directional without being teleological?, Studies in History and Philosophy of Biological and Biomedical Sciences (2015), http://dx.doi.org/10.1016/j.shpsc.2015.12.006
G.R. McGhee Jr. / Studies in History and Philosophy of Biological and Biomedical Sciences xxx (2015) 1e7
the workshop and in manuscript editing, and John Beatty for valuable discussions at the workshop.
References Bejan, A., & Zane, J. (2012). Design in nature: How the constructal law governs evolution in Biology, Physics, Technology, and Social Organization. New York: Anchor Books. Benton, M. J. (2005). Vertebrate palaeontology. Oxford: Blackwell Publishing. Darwin, C. (1859). On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life. London: John Murray. Gould, S. J. (1989). Wonderful life: The Burgess Shale and the nature of history. New York: W. W. Norton. Gould, S. J. (2002). The structure of evolutionary theory. Cambridge, MA: Belknap Press of Harvard University Press. Klimentidis, Y. C. (2012). On the limits of diversity. Review of convergent Evolution: Limited forms most beautiful, by George McGhee Frontiers in Genetics, 3. article 136, pp. 1e2.
7
McGhee, G. R., Jr. (2007). The geometry of evolution: Adaptive landscapes and theoretical morphospaces. Cambridge: Cambridge University Press. McGhee, G. R., Jr. (2011). Convergent evolution: Limited forms most beautiful. Cambridge: Massachusetts Institute of Technology Press. McGhee, G. R., Jr. (2015). Limits in the evolution of biological form: A theoretical morphologic perspective. Interface Focus, 5(1e6), 20150034. http://dx.doi.org/ 10.1098/rsfs.2015.0034. Raup, D. M. (1977a). Probabilistic models in evolutionary paleobiology. American Scientist, 65(1), 50e57. Raup, D. M. (1977b). Stochastic models in evolutionary palaeontology. In A. Hallam (Ed.), Vols. 59e78. Patterns of evolution. Amsterdam: Elsevier Scientific Publishing Company. Raup, D. M., & Gould, S. J. (1974). Stochastic simulation and evolution of morphology: towards a nomothetic paleontology. Systematic Zoology, 23(3), 305e322. Reiss, J. O. (2009). Not by design: Retiring Darwin’s watchmaker. Berkeley: University of California Press. Sapp, J. (2012). Evolution replayed. BioScience, 62(7), 693e694. Wilson, E. O., & Bossert, W. H. (1971). A primer of population biology. Sunderland, MA: Sinauer Associates.
Please cite this article in press as: McGhee, G. R., Jr., Can evolution be directional without being teleological?, Studies in History and Philosophy of Biological and Biomedical Sciences (2015), http://dx.doi.org/10.1016/j.shpsc.2015.12.006