Aristotle and Modern Biology
Journal of the History of Ideas, 33 (3)
July-Sept. 1972, 395-424.
HHC: titling added
Science had its origin, if not in opposition to Aristotle, at least in opposition to Aristotelianism. But science in its most authoritative form was what came to be called physics, not biology. Faced with the recent crisis in biology, in which the life sciences have been threatened with reduction to microversions of themselves and ultimately to chemistry and physics, one wonders if the besieged biologists, or at any rate their philosophical defenders, might not after all learn something to their advantage by reflecting on the one great philosopher who was also a great biologist. And we can learn from Aristotle; not, however, in a simple or straightforward fashion. There is no use just contrasting, as some have been tempted to do, Democritean with Aristotelian science and putting physicists in the former class, biologists in the latter. Even if we reject Simpson’s alleged reaffirmation of Roger Bacon and stoutly deny that “the study of Aristotle increases ignorance,”  we must nevertheless admit that in some important respects biology, like all modern science, really is, and must be, un-Aristotelian. This thesis could be defended in a number of ways; let me select four. 
First, the role of abstraction and the relation of scientific reasoning to everyday experience differ deeply in Aristotelian and in modern science. Only mathematics, Aristotle insists, abstracts from most of the ordinary perceptible properties (qualities, relations, states) of the things around us. Natural science, as he understands it, remains within the framework of everyday perception and makes more precise, within that framework, our formulation and understanding of the essential natures of quite ordinarily accessible entities. Modern science began and continues, on the contrary, precisely by closing its eyes to all but a few highly selected features of the world around us, abstracting from all but those variables which give promise of susceptibility to
1. G. G. Simpson, Principles of Animal Taxonomy (New York and London, 1961), 36n: “I tend to agree with Roger Bacon that the study of Aristotle increases ignorance.” In fact Bacon was objecting to the current translations of Aristotle, not to Aristotle’s teachings themselves. His statement reads as follows: “Si enim haberem potestatem super libros Aristotelis ego facerem omnes cremari qui non est nisi temporis amissio studere in illis et causa erroris et multiplicatio ignorantiae ultra id quod valeat explicari. Et quoniam labores Aristotelis sunt fundamenta totius sapientiae, ideo nemo potest aestimare quantum dispendium accidit Latinis quia malas translationes receperunt philosophi.” (“Compendium Studii Philosophiae,” Bacon Opera Inedita, Rolls Series number 15, 469). I am grateful to my colleague, Professor John Malcolm, for finding this passage.
2. See my Portrait of Aristotle (Chicago and London, 1964), esp. Ch. VII.
some sophisticated, usually quantitative, manipulation by the experimenter or the theorist.
Secondly, because he sticks so closely to the concrete, limited, and limiting physiognomy of things encountered in the everyday world, the Aristotelian scientist never confronts the teasing problem of induction which the modern scientist, or better, philosopher of science, necessarily has to face. The possible data of scientific calculations are infinite, the calculations themselves and their results are finite. However ingenious the arguments that have been used to comfort us before this gap, the gap remains. For Aristotle it does not exist. Admittedly even the most ingenious experimenter must rely, in Aristotelian fashion, on the stability of his surroundings: on his materials, his apparatus, as being reliably not simply this-here, but this-such.  Every time you pick up a handful of CuSO4 crystals, Norman Campbell argued, you are in effect acknowledging a law of nature: you confidently expect these blue crystals to have the same chemical properties as they did yesterday and will tomorrow.  Confidence in such stabilities does indeed depend, if you like, on Aristotelian induction, where we move from a rough and ready perception of the character of a thing to a more precise knowledge of just what kind of thing it is. But - pace Hume - it is only after this everyday induction that the problem of induction in science first begins: that is, the problem evoked by the necessary disproportion between data and hypothesis, between evidence and conclusion. Just because it remains within the horizon of everyday things-to-hand, Aristotelian science evades altogether the problem of induction in this its modern form.
A third way to emphasize the same contrast is to stress the role of productive imagination in scientific discovery. Aristotelian phantasia is powerless to excel what Kant called reproductive imagination. The productive power of that faculty, not a priori and once-for-all, as in the Schematism of the Critique of Pure Reason, but advancing hazardously to and beyond ever new frontiers: that is the moving force behind the scientific adventure, a force wholly beyond the ken of the much cosier Prinzipienforschung of Aristotle, cradled as it is within the comforting embrace of the familiar everyday world.
3. Wolfgang Wieland (Die Aristotelische Physik [Göttingen, 1962], 95n.) argues that the fashion in which modern physicists take a given experiment as general is still Aristotelian: “[HHC: German not reproduced]” It is, in his view, modern theories of induction that mislead us here. But those theories have recognized a logical gap which Aristotle failed, and for his purposes did not need, to recognize.
4. Norman Campbell, What is Science? (New York and London, 1952, 1921), Ch. II.
All this concerns Aristotelian methodology. Cosmologically, too, and especially for biology, there are features of Aristotelian science which the modern mind radically rejects. Aristotle’s world is finite, unique, eternal, consisting of a finite number of eternally existent species, “endeavoring” in their re-production to simulate the eternal circling of the celestial spheres (so he argues at the close of the De Generatione et Corruptione ). For modern biology, this eternal frame is shattered. All things flow. For many modern biologists, indeed, the theory of evolution is comprehensive for their science; all biological research, they feel, somehow derives from and contributes to it. That claim may be exaggerated. But even research seemingly unconnected with evolution is nevertheless related to it in some degree, as figure to ground. Some of Aristotle’s detailed biological work too was sound and still retains its validity despite the incorrectness of his cosmological theory, e.g., his description of chick development or his account of the life history of Parasilurus aristotelis or of Mustelus laevus. Much modern research, similarly, may reach correct conclusions independently of the general conception of life’s development from the non-living or of the transmutation of species. Yet the overall thrust of modern biology has been radically altered by the idea of evolution, much as the Aristotelian science of nature was guided by the contrary view of a static and finite cosmos. In their overall implications, the two are incompatible.
Despite these contrarieties, however, there is much to be learned from Aristotle in relation to the philosophical problems of biology. I want to discuss in this context three Aristotelian concepts: [HHC: Greek not reproduced] telos or the [HHC: Greek not reproduced], that for the sake of which; [HHC: Greek not reproduced] in contrast both to [HHC: Greek not reproduced… ], eidos as form and eidos as species, and finally that most puzzling of Aristotelian phrases, [HHC: Greek not reproduced], the “being-what-it-is” of each kind of thing.
First, telos. Again, it was Aristotelian teleology that seventeenth-century innovators were most emphatically determined to abolish from the study of nature. And again, as Aristotelian teleology had come to be understood, or misunderstood, this was a necessary move. Yet some sort of teleology, or teleonomy, as some modern biologists prefer to call it, keeps creeping back into biological language and thought. Let us consider, then, how and where something like Aristotle’s telos occurs in modern biology and how modern usage compares with his.
Two misconceptions must first be set aside. 1) Telos is not in the first instance - and in the study of nature is not at all – “purpose or plan.” In nature, “that for the sake of which” a series of events takes place is the intrinsic endpoint in which, if nothing fatally interferes,
5. Degeneratione et corruptione, II, 11, 338b7ff.
that series normally culminates. What usually happens to a fertilized robins’ egg, for instance? A baby robin hatches out of it; that is its telos. There is absolutely no question of any kind of “purpose” here, either man’s or God’s. To suppose otherwise is to introduce a Judaeo-Christian confusion of which Aristotle must be entirely acquitted. 2) Nor is Aristotle interested primarily in one over-all cosmic telos. Despite the passage from De Generatione e.t Corruptione already referred to, and despite the “teleological” causality of the unmoved mover, the kind of “ends” that usually interest Aristotle are the determinate end-points of particular processes within the natural world. True, the stability of the universe is, for him, a necessary condition of the orderly processes of its components; but this is by no means the target of his primary interest. On the contrary, it was, again, the Judaeo-Christian God who (with the help of neo-Platonism) imposed the dominance of a cosmic teleology upon Aristotelian nature. Such sweeping purpose is the very contrary of Aristotelian. 
The concept of telos in exactly Aristotle’s sense, however, does occur in exactly the area where he himself invokes it: in the study of ontogenesis. Here, indeed, the contrast of Aristotelian and Democritean science still appears valid, at least to a first approximation. The embryologist must put questions to the living embryo, in terms to which, and in which, it can respond.  And this is impossible on principle in terms of a thoroughgoing atomism, since in those terms there is nothing alive. What might happen to a really Democritean scientist faced with a biological problem was suggested by Frank Baker in a paper I have quoted elsewhere. Imagine, he says, an observer looking in the field of a microscope at the filaments of a fungus. “He witnesses,” says Baker,
“that at the tips of the filaments are disposed a number of radiating branches more frequently segmented than those of the stalk to which they are attached; which adjointed elliptical segments are easily set free by pressure in the surrounding medium. But, supposing that he decides to investigate these segments, what kind of ideas are going to control his choice of further observations; how will he proceed, loosely speaking, to discover ‘the nature’ of these structures; and, in brief, in such a context, what does his notion of ‘investigation’ already imply?” 
6. See the exposition of Wieland in the work referred to, note 3, above; see also his Science and Philosophy in Aristotle’s Biology (Darmstadt, 1972). For a close study of Aristotle’s use of telos in the explanation of generation, as well as, and in relation to, eidos and hyle, see the excellent paper by Anthony Preus, “Science and Philosophy in Aristotle’s Generation of Animals,” J. Hist. Biol. 3 (1970), 1-52.
7. Cf. M. Grene, The Knower and the Known (New York and London, 1966), 237.
8. Quoted ibid., 235-36, from A. E. Baker, “Purpose and Natural Selection: A Defense of Teleology,” Scientific Journal of the Royal Coil, of Science, 4 (1934), 106-19, 107-08.
An old-fashioned chemist, Baker suggests, might throw these segments (which, unknown to him, we may call spores) into concentrated H2S04. He would learn something; but would this “lay an effective basis to the study of mycology?” And why are these not as good facts to start with as any other? But suppose our investigator places the segments on jam (where he first found them) or sugar, and in the warmth, and watches what happens. He may then discover “their relation to the life cycle of the fungus.” Actually, Baker remarks, it would take a whole series of investigators “animated by a single scientific impulse or tradition” to lay the foundation which would so much as show him where to start. When he gets this far he can then, but not before, undertake his chemical analysis. Only, in other words, when the concept of germination is understood and its designatum assumed to exist, do the details “fall into order and acquire a significance,” such that detailed analysis of some parts of the process of germination can be undertaken. The orderly development of the organism under investigation must have been assumed, Baker concludes, before the right “facts” could be selected for further investigation and analysis.  An orderly development toward a normal end, therefore, is necessarily presupposed by the biological investigator before he can set out to make his investigation. A concept of “that for the sake of which” the development is occurring, of its natural telos, is contained in the very question asked.
Such considerations place teleological (or teleonomic) thinking in the position of at least a regulative idea (in the Kantian sense) at the beginning of biological research. From this point of view telos is a signpost to the study of nature: a “reflective concept” (em Reflexions-begriff) as Wieland argues.  Looking at the endpoint of the series helps us to start looking for its necessary antecedents; there is nothing “unscientific” about this, not even anything very un-Democritean. But is that all there is to it? A Kantian regulative idea - say, the infinite divisibility of matter, or indeed natural teleology as Kant coneived it - is a pure as-if. And many modern thinkers would be content with this, with “the appearance of end,” as Waddington calls it.  It’s a makeshift, they say, a crutch to lean on until we have mastered the necessary and sufficient conditions, or until we have constructed a machine to simulate an animal, or until we have synthesized life. Then we can throw away the crutch and walk alone.  Or perhaps, as Piaget argues, the very idea of a final cause, even in this seemingly harmless form, is based on a logical error and we don’t need it at all.
9. Baker, op. cit., 108-10.
10. W. Wieland, op. cit., 254-77.
11. C. H. Waddington, The Strategy of the Genes (London, 1957), 190.
12. See the (by now classic) interpretation of organicism in Ernest Nagel, The Structure of Science (New York, 1961), 428-46.
It rests, he maintains, on a confusion of three different notions: physical or physiological causality (cause a produces effect b), logical implication (the use of A implies the consequence B), and instrumentality (to get B “we must” use A). 
Both these views are mistaken. To assert that a robin’s egg hatches out a robin and not an oak tree is to state not a regulative idea but a fact of nature. Nor, a fortiori, does such an assertion represent any logical howler. At the least, it locates in the real world an orderly process, the details of which the biologist may undertake to study. It selects a certain segment of orderly temporal process in its orderliness as the locus of an inquiry. To this extent at least it locates real, not apparent, ends and suggests really, not seemingly or misguidedly, teleological questions.
But is that all? If there are real processes with natural endpoints, real telos in nature, are there not also teleological answers to the questions we put to nature? This seems to me a much more difficult problem. Professor C. P. Raven of Utrecht has written of the application of information theory to biology as “the formalization of finality,” and I had formerly taken him at his word.  Yet now I wonder. What is “finalistic” about information theory? Admittedly, both teleological explanation and cybernetical explanation are complementary to classical causal explanation; in this sense, both resist a one-level, Democritean approach. But that does not make them equivalent. I prefer to leave this question open here, and postpone a consideration of multilevelled explanation in general until I come to the concept of eidos. For the moment we may take it that in the study of individual development, the concept of a normal endpoint of development helps methodologically to locate the place of the inquiry and to locate it really in nature, whether or no the concept of telos is also embedded in the solution to the embryologist’s problem.
The most vexing problem with respect to Aristotelian telos and modern biology, however, concerns not ontogeny, but phylogeny. Granted, there is a definite, if perhaps limited, role for teleology (or teleonomy) in the study of ontogenesis; can one transfer the concept of
13. Jean Piaget, Biologie et connaissance (Paris, 1967), 225-26. Piaget refers here to an argument by J.-B. Grize: “HHC: French not reproduced]
14. The Knower and The Known, Ch. IX, 238.
end to the study of evolution, or can one at least discover in evolutionary explanation an analogue of teleonomic thinking? It has repeatedly been claimed both that Darwinian evolutionary theory rejects any cosmic telos and that it retains the concept of telos in some more acceptably “scientific” sense. Indeed, it has even been argued that it is precisely by virtue of its teleological structure that evolutionary theory, and only evolutionary theory, rescues biology from reduction to physics and makes it “an autonomous science.”  These claims may perhaps be clarified if we compare the “teleonomic” thinking of modern evolutionists with Aristotelian teleology.
The fixed endpoint of a natural process, for Aristotle, is the mature form of the adult individual (strictly, of the adult male!) of the species in question. “Nature is like a runner,” he says, “running her course from non-being to being and back again.  The being in the case is the developed adult of such and such a kind. In modern evolutionary biology, however, there is no such fixed form; the eidos itself, which is the telos of individual development, is transitory. What remains? For Darwinism, the telos that remains when the eternal species is removed is simply survival. Survival of what? The individual perishes; what survives? In Darwinian terms: the descendants of the slightly more “successful” members of a species. In neo-Darwinian terms: the alleles which made possible the development of phenotypes carrying the slightly more “successful” characters, or rather, statistically, a higher ratio of those alleles in the gene pool of the next generation. A robin’s egg is not, it seems, the way to make a robin, but a robin is a robin’s egg’s way to make more, and more probably surviving, robin’s eggs. The locus of the goal of biological process is not, as it appeared to Aristotle, in the mature individual, who is, as such, mortal and of no concern to evolution, but in the future gene pool of the population of individuals of a potentially interbreeding population.
Take, e.g., Kettlewell’s classic study of industrial melanism.  If tree trunks are blackened by industry, birds take more peppered moths, and hence more genes for peppered wings, than they do the carbonaria mutant of the species. Hence, fewer mutant genes are eaten and proportionally more survive. The telos of this process is the greater ratio of carbonaria genes in the next generation. Evolution-
15. F. Ayala, “Biology as an Autonomous Science,” Amer. Scientist, 56 (1968), 207-21.
16. DeGen. Anim.741b21ff.
17. H. B. D. Kettlewell, “Selection experiments on industrial melanism in the lepidoptera,” Heredity 9 (1955), 323ff. “A résumé of investigations on the evolution of melanism in the Lepidoptera,” Proc. Roy. Soc. Lond. B., 145 (1956), 297 ff.
arily speaking, that is what the whole business of being a moth is for; not, indeed, just “for” the survival of this gene, but if we had a complete count of all the genes in the population at time t0 and time t1, the differential ratio would give us the “end” of the story; the differential survival of some genes rather than others. But which genes? Whichever ones survive, of course. If we clean up industry and the tree trunks bleach again, more peppered and fewer carbonaria genes survive, and the endpoint goes the other way. Similarly in peacetime healthy human males are, other things being equal, better adapted - and that means of course in evolutionary terms more likely to leave descendants - than sickly ones, but in war time the contrary holds: the halt, the lame, and the blind are better adapted than the healthy. Biological process is first and last and always evolution; evolution is first and last and always a chronicle of survival, the survival of whatever survives.
A strange telos: we are told simply, what survives survives. But this, it has repeatedly been objected, looks like a mere tautology. And at first sight, at least, it has also been repeatedly objected, a tautology seems to have no explanatory power, let alone the explanatory power that would be characteristic of a teleological account. For a teleological account distinguishes, and sets out as aimed at, a goal to which it can then relate the antecedent steps. Here, however, goal and steps are collapsed into an empty identity.
Yet that identity, we are told, presides over a rich and precise elaboration of “evolutionary mechanisms,” and hence of teleonomic patterns of structure or behavior. The case for this view is argued in a thought-provoking book by George Williams, Adaptation and Natural Selection. Although he accepts, and celebrates, the tautological character of the principle of natural selection, as the survival of the fitter in the sense of the more probable survival of what will more probably survive, he insists nevertheless that this principle, correctly used, can preside over a vast range of teleonomic investigations:
A frequently helpful, but not infallible rule is to recognize adaptation in organic systems that show a clear analogy with human implements. There are convincing analogies between bird wings and airship wings, between bridge suspensions and skeletal suspensions, between the vascularization of a leaf and the water supply of a city. In all such examples, conscious human goals have an analogy in the biological goal of survival, and similar problems are often resolved by similar mechanisms. Such analogies may forcefully occur to a physiologist at the beginning of an investigation of a structure or process and provide a continuing source of fruitful hypotheses. At other times the purpose of a mechanism may not be apparent initially, and the search for the goal becomes a motivation for further study. Adaptation is assumed in such cases, not on the basis of a demonstrable appropriateness of the means to the end, but on the direct evidence of complexity and constancy. 
18. George Williams, Adaptation and Natural Selection (Princeton, 1966), 10.
The study of the lateral line of fishes, he suggests, is a good example of this kind of reasoning:
The lateral line is a good illustration. This organ is a conspicuous morphological feature of the great majority of fishes. It shows a structural constancy within taxa and a high degree of histological complexity. In all these features it is analogous to clearly adaptive and demonstrably important structures. The only missing feature, to those who first concerned themselves with this organ, was a convincing story as to how it might make an efficient contribution to survival. Eventually painstaking morphological and physiological studies by many workers demonstrated that the lateral line is a sense organ related in basic mechanism to audition (Dijkgraaf, 1952, 1963). The fact that man does not have this sense organ himself, and had not perfected artificial receptors in any way analogous, was a handicap in the attempt to understand the organ. Its constancy and complexity, however, and the consequent conviction that it must be useful in some way, were incentives and guides in the studies that eventually elucidated the workings of an important sensory mechanism. 
How does this kind of teleonomic thinking compare with the use of telos in Aristotle? Aristotle presents his concept of “that for the sake of which” as a guide to the study of nature in opposition to the thinking of Empedocles, who would elicit the phenomena of the living world, without ordered ends, out of a combination of chance and necessity. At one stage in cosmic history, Empedocles imagines, there were heads and trunks and limbs rolling about the world. Those that happened to come together in a viable combination survived; the others perished. This was a very crude theory of natural selection, to be sure, but a theory of natural selection, nevertheless. Aristotle as a practising biologist objected: ox-headed man progeny and vine-bearing olives, such as Empedocles envisages in his transitory world, are an absurdity. What we always have in nature is the ordered passage to a definite endpoint: man to man, cattle to cattle, grape to grape, olive to olive. Only where there are such functioning, ordered series does the study of life begin. Williams would agree. Where we can use only the concepts of chance and necessity, he insists, we should. Thus the descent of flying fishes can be explained in terms of physics alone; their flight, however, which is “contrived,” in analogy to human contrivance, needs, he argues, another and teleonomic principle of explanation, as any piece of machinery does.
So, as we saw with the case of ontogenesis, we need, it seems, a teleological approach to locate a biological problem. But is the explanation, in the case of selection theory, teleological as well? Have we even found, as in individual development, a directed process to describe - however we may eventually explain it? I think not; for explanation in terms of orthodox evolutionary theory collapses pretty
19. Ibid., 10-11.
quickly into pure Empedoclean chance-times-necessity. Had fishes not had the sensory mechanism of the lateral line they would not have “heard” their predators coming and would not have survived. Or better: those whose “hearing” was slightly more acute left descendents in the gene pool, those not so gifted left fewer and finally none. Chance mutations necessarily sorted out by the compulsions of environmental circumstances: that is a pure Empedoclean, anti-teleological process. The peppered moth case is a striking example. We have here, we are told, “evolution at work”: now we see the whole process in little. Extrapolate this “mechanism” to the whole story of life and you have the vast panorama before you: no other principles are needed. But what is this story? The environment, for extraneous reasons (in this case the industrial revolution) changes; the gene pool is always changing; the changed environment necessitates changed predation (birds can’t as easily see black moths on black trees as peppered ones); changed predation necessitates differential survival of some genes rather than their alleles. So we necessarily get more black moths than peppered ones. Extrapolate this process to the whole of evolution and you see a vast sequence of necessities. True, the sequence is triggered, and kept going, by a set of curious chances. These, however, are “chances” only in the sense of being at a tangent to the “normal” sequence of development. They are to be explicated, on principle, in terms of natural, that is, physico-chemical laws. Thus, given the nature of bituminous coal, the tree trunks had to be blackened. Given the chemical nature of DNA, one supposes, the “errors” which occur in its replication will ultimately be explained also in physico-chemical terms. They happen by “chance” - just as in Aristotelian chance - only in the sense that they are outside the usual sequence of events to which one has been attending, in this case the “normal” development of the peppered moth. But they have, or will have, their physico-chemical explanation, which must ultimately exhibit their necessary occurrence. Again, extrapolate this reduction to the whole history of nature. Where is the teleology? It has served as a heuristic maxim to start us off on our inquiry, but in the sequence of survivals it does not survive even as a factor in the phenomena described, let alone as an explanatory principle.
Is there any other way to introduce teleology into evolution? Many philosophers and some biologists have tried to do this in terms of a theory of “emergence.”  But these theories, compared either with
20. Among biologists critical of the neo-Darwinian synthesis, see for example E. S. Russell, The diversity of animals, Acta Biotheor. 13 (Suppl. 1), 1962, 1-151, Cf. A. Vandel, L’Homme et L’Evolution (Paris, 1949). Among recent philosophers, see M. Polanyi, Personal Knowledge (Chicago and London, 1958 [Torchbook edition: New York, 1962], Ch. 13.
the precise and limited teleology of Aristotle or with the vanishing teleology of Darwinism, are vague, and empty of explanatory power. If one says, for example, with Vandel (following Bergson) that life has moved toward an increase of “le psychisme,” with two high points, the insects (generic inventiveness) and man (individual inventiveness), two objections at once arise.  First, some of the diverse branches of evolution have gone that way, but not by any means all. What of parasitism, what of long stable forms like Lingula, what of the vast variety of birds or “lower” mammals, what of the evolution of plants, etc., etc.? Second, even if we can see, very generally, some such tendency in the history of life on earth, how did it happen? What does such an assertion of the “emergence” of psychic powers explain, and how? This appears an even stranger extrapolation than the Darwinian. For one can imagine the melanism story stretched back to the beginning of time. I suspect (indeed, I have argued elsewhere) that this extrapolation entails untenable pseudo-reductions of richer to poorer concepts; but still one can see how it’s done. The “emergence” extrapolation, however, I for one simply cannot follow at all. There is a goal, mind or thought or inventiveness, we are told, for which evolution happened. It is the achievement of this goal that we are studying when we look at evolution’s course. But whose goal? Whose achievement? The giraffe, we know contra Lamarck, didn’t get a longer neck by trying; and are we to believe that the brachiopods tried to achieve thought and left it to us to succeed, or tried to achieve social rituals and left it to the ants to carry them through? Achievements must be some one’s achievements. A goal, even if it is an Aristotelian telos, not a conscious purpose, must be the endpoint of some entity’s becoming. Whose achievement is evolution? Whose goal, on an evolutionary scale, is thought? In any terms available to this writer at least the very question is nonsense. The concept of telos is intelligible and useful, I submit, only with reference to something already in existence. In the study of evolution, on the contrary, where we have no fixed individuals and therefore no fixed endpoint of process, we have, whether in the Darwinian view or in the efforts of “emergence” theorists to revise it, only the appearance of teleology, not its flesh and blood.
Still, that appearance keeps reappearing. Why? In emergence theory it is a case of metaphysical aspirations as yet unfulfilled. All honor to them; it may well be that this controversy will only come to rest once one has accepted a cosmology of some Whiteheadian kind. This, to most of us, has not yet happened, certainly not in such a fashion as to affect the practice or the thinking of biologists. But why does neo-Darwinism, as distinct from those broader and vaguer views, recurrently lay claim to being teleonomic in its structure?
21. A. Vandel, op. cii.
The answer is not far to seek. To give it will permit one more comparison with the teleology of Aristotelian science.
Darwinian evolutionary theory appears teleological because it is first and last a theory of adaptation. Deriving from Paley, it views all organisms as adaptation machines, aggregates of devices for the adjustment of the organism to its environment. On this view it is, as Williams insists, thinking in terms of adaptation, and this alone, that distinguishes biology as a science from physics and chemistry. Yet adaptation in evolutionary terms is for survival and survival only. Everything non-trivial in specifically biological processes reduces to this one phenomenon. But explanation in these terms, as we have already seen, either collapses into tautology or is reduced to necessity, and so in either case fails to retain its alleged teleonomic structure.
How does this situation compare with that of Aristotelian teleology? First, in Aristotle we find for each kind of thing a given normal endpoint of development, and relate to it a set of what Aristotle calls “hypothetical necessities.” Given, for instance, that a creature needs to hear - or be somehow sensitive to environmental vibrations - it will develop some kind of auditory organ, whether a vertebrate ear or a piscine lateral line. In the modern version, however, such hypothetical necessities become simple necessities. Since the endpoint to which one might refer them as means is not fixed in advance, it becomes simply the ineluctable issue of the preceding steps. Instead, therefore, of the necessary conditions being relative to the end, the end is the automatic product of its necessary (and sufficient) conditions. In Aristotle, secondly, the end being given, its achievement happens “always or for the most part” - but it may fail. All along the way, there is room for abnormality, for chance. In modern theory, however, there is no such leeway. Even though, in terms of our present knowledge, most mutations may be “chance” occurrences, that is so only in a sense analogous to Aristotelian tyche; that is, they are caused by some cause and effect sequence outside the usual pattern of development in the given case. But they are not - or cannot survive - as to automaton in Aristotle’s sense, as sheer random happenings. Did they not prove to be adaptive as the environment changes, natural selection would eliminate them. And if they do prove adaptive, on the other hand, they have to survive. Depending on how you look at it, in other words, everything is teleological (adaptive) or everything is necessary. There is no middle ground for the merely contingent in the interstices of an otherwise orderly sequence.
It is just this two-edged comprehensiveness, finally, combined with the authority of an algebraic formulation, that lends to modern natural selection theory its great explanatory power. First, there is a formal, mathematical instrument, the algebra of Fisher or Haldane, which may be used to measure natural selection and which lends
weight and precision to experimental results in this field. Such a formula, intrinsically tautological, is used to measure changing adaptive relations and therefore serves the appearance of teleology. But these relations in turn, when viewed as a series of organism-environment interactions, appear thoroughly necessitated, not teleological at all. When, however, such interactions are summed up, for long periods, in algebraic formulation, the results, neatly ordered, present apparent trends and thus once more give the appearance of teleology, an appearance once more reduced to necessity when we visualize the whole sequence of action/reaction from which they have eventuated. Thus if one accuses Darwinians of being “mechanistic” they point to the “trends” in evolution as they see it; if one accuses them of being “teleological” in their thinking, they point to the necessity of the whole show in terms of environmental pressure and the consequent changes in gene ratios. And if we try to bring this to-and-fro to rest, what have we? Once more, tautology: well, after all, what survives survives. If you look at the process a tergo it appears teleological; if a fronte, necessitated. And sub specie aeternitatis, when the theory is summed up in a formula for measuring differential gene ratios, you have a theorem universally applicable because empty, totally comprehensive because it expresses a simple identity. 
Why do we keep going round this merry-go-round? “Adaptation” is a means-end concept. Yet if all adaptations are for no specifiable end except survival, one keeps falling back into a universal necessity which is in turn reducible to the same old tautology. Stretch it out: it’s teleology. Collapse it one level: it’s necessity. Collapse it still further: it’s tautology. What is lacking to stabilize this endless vacillation? This brings me at last to my second major Aristotelian concept: eidos. For what is lacking in the modern concept of adaptation is precisely the definite telos, which in Aristotle is the mature form of the species, of the type, the [HHC: Greek not reproduced] of the [HHC: Greek not reproduced].
Nor, of course, do I mean by this some cosmic goal. Again, there are no such goals within Aristotelian biology. True, there are such in Aristotelian cosmology, as e.g. in respect to the proof of the unmoved mover. In the framework of biological investigation, however, we need not, indeed we may not, invoke such dialectical arguments. Modern biology, however, lacks even the more limited and concrete end-points of Aristotelian science. In short, when Darwinism evicted the watchmaker of Paley’s famous watch, it threw out as well the telos of the watch itself. But without a terminus ad quem of development, without a terminus ad quem for our understanding of the organization of a living system, of an organism, of an organ, of an organ-
22. Cf. my analysis of R. A. Fisher’s Genetical Theoretical Theory of Natural Selection, in The Knower and the Known, 253-66.
elle, one has no univocal concept of adaptation, of the adjustment of these means to that end. True, the organism, the organ, the organelle is continuously adapting itself to its environment, both internal and external; but what for? To what end? In ontogenesis, to the end of maturation and self-maintenance of the organism, the organ or the, organelle, that is, to the end of the origination and conservation of some form. The processes of adaptation, as distinct from their result and adaptedness, are thus related as dynamic processes to their goal, that is, to the actuality of the organized system which comes into being or maintains itself in being through those processes. And the result, adaptedness, as differentiated being-such-and-such of the parts of the organism, the organ or the organelle, is also subordinated, therefore, to the development or maintenance of the form, the eidos, of the whole. It is, then, precisely the Aristotelian concept of form, or some modern analogue thereof, which is lacking in the modern concept of adaptation, or better, of the organism as a pure aggregate of adaptive mechanisms. 
Let us look a bit more closely, then, at the relation between modern biology and Aristotle’s concept of eidos.
Perceptions of form - the shape of an oak leaf, the walk of a cat, the metamorphosis of a butterfly - the grasp of such configurations and changes of configuration, are among the basic insights by which the subject-matter of biology is singled out. A certain freedom of form within form, Buytendijk has shown, is the criterion by which we see a shape as “alive.”  Biological knowledge, the knowledge of men like Ray or Hooker, was a refinement of such elementary perceptions, a refinement to the point of genius, but not different in kind from its everyday counterpart. Modern biologists, however, at least the more theoretical, and more articulate, among them, sharply reject such old-fashioned connoisseurship. Form, and the recognition of form, are not only not (according to their own professions of faith) their central concern; they exhibit a positive dread of form. In the polemics that characterize contemporary taxonomy, for example, and in evolutionary controvery also, the favorite epithet of the combatants is “typology.” To call a man a typologist is the worst insult you can bestow.  It is hard sometimes to tell quite what is meant by the
23. Cf. T. Dobzhansky, “On Some Fundamental Concepts of Darwinian Biology,” Evol. Biol. 2 (1968), 1-33, where efforts are made, not wholly successfully, to disentangle some of these concepts.
24. See my account in Approaches to a Philosophical Biology (New York, 1969), Ch. 2, 74-75.
25. R. R. Sokal, (‘Typology and Empiricism in Taxonomy,” J. Theoret. Biol., 3 (1962), 230-67.
term; it is not necessarily Platonic realism, perhaps something like Aristotelian realism. In any case, it’s deadly. Perhaps, indeed, it is most of all this eidophobia, if one may so christen it, that makes biologists shudder at the very name of Aristotle. Yet the Aristotelian concept of eidos could teach reflective biologists much about the foundations of their discipline.
Eidos in Aristotle is used differently in different contexts, but when used technically it seems to represent a single concept, although it is already rendered in Latin (as in modern European languages) by two separate terms: forma and species. I have not found, however, any indication that Aristotle took the term eidos to be in any formal sense equivocal. Thus he includes genos, for example, in the philosophical dictionary (Delta 28), but uses eidos in that discussion as if there were no problem about its meaning.  Since he deals so carefully with the several meanings of equivocal terms, I can only conclude, therefore, that he thought of eidos as one comprehensive concept, with different applications in respect to different problems, and perhaps with two major applications which correspond, for us, to “form” and “species.” Or, if even this partial separation is incorrect, we may separate, for ourselves, the two aspects of his one concept when we try to see what it can tell us in the context of modern biological methods. 
First, then, form in contrast to matter. Eidos in this context functions in a number of striking respects in the same way as the concept of organization (or information) in modern biology.
a. For one thing, form and matter in Aristotle constitute a pair of concepts used relatively to one another and relatively to the problem at hand in a great variety of different contexts. The eidos of an entity or process is its organizing principle, the way it works to organize some substrate capable of such control. Though it is sometimes equiva-
26. For a detailed account of Aristotle’s usage in his own biological writings the reader should consult Professor David Balme’s definitive treatment, as well as A. L. Peck’s notes in the Introduction to his edition of the Hist. Anim. D. M. Balme, “Telos and Eidos in Aristotle’s Biology,” Class. Quart. 12 (1962), 81-98; A. L. Peck, Introduction, in Aristotle, Hist. Anim., volume I, (Cambridge, Mass., 1965) esp. notes 5-11. Cf. also D. M. Balme, “Aristotle’s use of differentiae in zoology” in Aristote etles Probièmes deMéthode(Louvain, 1960) 195-212.
27. The two aspects I am separating are, admittedly, brought very close together by Aristotle himself, not only in Post Anal. (94 a 20ff.), where genos is given a place corresponding to that of “material cause” in the Physics, but also in Metaphysics 1024b 8, Z 1038 a 6, H 1045a 23 f. and 11058 a 23, where genos is identified with hyle. I am grateful to Professor Balme for calling my attention to the latter passages, and confess that I might not have made the distinction between the two pairs of terms as flatly as I tried to do, had I read his papers (just referred to) before writing this essay. No one interested in Aristotle’s biology and its relation to his philosophy of science can afford to neglect Professor Balme’s careful and illuminating work.
lent to morphe or shape, that is by no means always the case. Nor is form in nature a separate, self-subsistent “absolute”; on the contrary, it must once more be emphatically affirmed, it exists in, and only in, that which it informs. In the context of the entity or process in question it exists as the organizing principle of that process, just as its matter exists as the potentiality of such (or other) organization. Thus in noses (one of Aristotle’s favorite examples) snub is the form of the matter flesh and bone. In bone, however, boniness is the form of earth and fire or whatever elements compose it. Biological systems lend themselves par excellence to this dual - but not dualistic - analysis. It depends on the particular system one is studying what will be form and what matter in a given case; but the two-level analysis is always apposite. On the one hand, this matter as the matter of this form is by no means to be ignored, since natural form exists only as actualized in an appropriate matter. As the matter of this form, indeed, it can be studied with profit for itself - as modern biology studies, in much greater exactitude than Aristotle could dream of, the physico-chemical substrate of living systems. But the form too can be studied for its own sake, as it is by modern systems-theorists, even though it exists only as enmattered and depends for its existence and continuance on the laws governing the matter of which it is the form. And again, be it noted, if Platonists and scholastics generalized the Aristotelian concepts to form a cosmic hierarchy from the abstraction of prime matter to the Divine Mind, this was not Aristotle’s primary intent. Eidos and hyle were for him a pair of analytical tools, to be applied in the study of nature relatively to one another and relatively to the particular inquiry.
Despite the simplicity of his examples and the crudity of his “chemistry,” Aristotle’s methodological thesis is an important one. Eidos in the sense of organizing principle is indeed a definitive concept for biological method. True, in view of the advance of scientific knowledge since Aristotle’s time, its modern counterpart is couched in very different terms. Thus, for example, G. L. Stebbins in his Basis of Progressive Evolution describes the principle characterizing living systems as that of relational order. “In living organisms,” he writes, “the ordered arrangement of the basic parts or units of any compound structure is related to similar orders in other comparable structures of the same rank in the hierarchy, permitting the structures to co-operate in performing one or more specific functions.”  This is a much more precise statement, indeed, but it plays the same role relative to the chemistry and physics of living things as does Aristotle’s concept of form in relation to matter. The colinearity of the DNA chain is a relatively simple example of such order. The concept of information may also play a similar part. In fact, it is even closer to
28. G. L. Stebbins, Basis of Progressive Evolution (Chapel Hill, N.C., 1969), 5-6.
Aristotelian form. For information can be found in any sort of system (and any system, in Aristotle’s view, can, and should, be studied in form-matter terms), yet in living things the quantity of information is vastly greater - and more interesting - than in non-living systems.  From this point of view, indeed, Raven’s “Formalization of Finality” might better be called “The Formalization of Form.” Just as, from the Aristotelian scientist’s point of view, matter is the possibility of taking on one form or another, so the elements of an information-bearing system can assume any one of many equiprobable states, one of which, by virtue of its very improbability and in proportion to that improbability, becomes a bearer of information. In short, the relation between entropy and negentropy in biological processes expresses a quantitative equivalent of Aristotle’s qualitative distinction between the material and formal aspects of a given system or subsystem, of an organism, organ system, tissue, etc.
b. As either of the instances just cited indicates, moreover, (that is, Stebbins on relational order or Raven on the information-theoretical aspect of biology), the role in biology of a concept of form, organization or information should demonstrate once for all the irreducibility of biology to physics and chemistry (at least in their classical, reduced, and one-level form). This was Aristotle’s thesis also, against Democritus. Organized systems cannot be understood in terms of their least parts alone, but only in terms of those parts as organized in such systems. Organized systems are doubly determinate; they exist on at least two levels at once. True, the form-matter pair of concepts do not of themselves generate a stratified cosmos; but they do show us how to resist reduction to one single cosmic level, whether of Democritean atoms or of the fundamental particles of modern physics. There may or may not be a cosmic hierarchy; the fact remains that whenever we study living systems we are studying particular, limited systems that are hierarchically organized, organized on at least two levels. We are studying systems composed of elements obeying their own laws, but constrained at the same time by arrangements of those very elements which constitute, as such, laws of a higher level. The
29. The locus classicus for the application of information theory in science is Science and Information Theory, Leon Brillouin, (New York, 1956). The application of information theory to biology has been discussed in a number of places, notably by Henry Quastler, see The Emergence of Biological Organization (New Haven, 1964). The point that the distinction between living and non-living systems with respect to information is quantitative - though great enough to appear qualitative - I owe to Dr. Thomas Ragland of the University of California, Davis, who has lectured on this subject to my class in the philosophy of biology. Michael Polanyi, in “Life’s Irreducible Structure” (in Knowing and Being [Chicago and London, 1969], 225-39), argues, on the contrary, that the distinction between the two kinds of systems is logical and qualitative; yet he admits, in terms of evolution, a continuous transition from one to the other.
higher level - form, organizing principle, code, fixed action pattern or what you will - exists only in its elements, and depends on them for its continuance, yet the laws of the elements in themselves, corresponding to Aristotelian hyle, permitting any number of informing arrangements, do not as such account for the principle which in this case happens to constrain them.
If, moreover, the question of teleological explanation left us puzzled, here, it seems to me, the case for non-“mechanistic” explanation becomes much clearer. The concept of eidos, like that of telos, is indispensable to help locate a biological problem: it has heuristic value. If used with good judgment it locates, as telos does in limited areas, a real phenomenon, a structure or process, in nature; it has descriptive value. But much more clearly than telos, it also has explanatory power. To discover the working principle of an organized system, as in the specification of the DNA code, or in the functional explanation of the lateral line (apart from the question of its origin), is to explain the system just as truly as one explains it by analyzing its physico-chemical parts. To learn how an organized system operates is just as conducive to the understanding of it - the scientific understanding of it - as is the analysis of the same system into its elementary components. Indeed, the teleonomic study of biological systems is probably reducible, I would suggest, to the diachronic rather than the synchronic study of their form. Teleonomy sits uneasily on evolutionary theory, one would then suspect, because, since the sequence of living systems that have inhabited this planet does not itself constitute a living system, it has no eidos, and therefore no telos, to which the study of its necessary conditions could be referred.
c. Biological explanation, then, works in terms of form or matter, systems-theoretic study of wholes or part-analysis (ultimately physico-chemical) of their constituents, with the two kinds of explanation complementing one another. That complementarity, thirdly, however, is not symmetrical. Again, there is a striking resemblance here to the Aristotelian form/matter pair. All natural things, organized parts of such things and processes exhibited by them, are inherently informed matter and can be studied on both levels; but form is prior. In non-Aristotelian language: although the upper level of a doubly determinate system depends on the lower level, and the laws of the lower level, for its existence, and is inseparable from it, it is the upper level that makes the system the kind of system it is. We have to refer to the upper level, we have seen, to generate a problem, to describe the system we are studying, and in certain cases at least to explain the operation of the system as such. Thus in the code case or in physiological explanation, the problem-location, the description and the explanation all refer to form: information. In some cases perhaps only the first and second obtain as explanatory principles. Research
on the lower level, physics in relation to molecular biology, or physiology in relation to behavior, or chemistry in relation to metabolism, may indeed go on indefinitely without explicit reference to the higher level, the structure of an enzyme, the typical course of a nesting behavior, the normal growth process of an embryo; but it will not be study of this system without some implicit reference at least to the organizing principle concerned. The higher level, though dependent on the lower, is both epistemologically and ontologically prior to it. For Aristotle, of course, it is also prior in time: man begets man eternally. That is what evolution has altered: for us potency is, in nature as a whole, prior to actuality. But given the existence of an organized being or process, form is then prior to matter, not only cognitively but onto-logically as well. Prior cognitively: to know the system is to identify, describe, and understand it in terms of its operating principles, of the way it uniquely constrains its components to make this system of this kind. Prior ontologically: since the principles of “matter” on their own could logically, and in terms of the laws of probability, take on any “form,” it is the existence of this form that makes the system what it is.
Once more, of course, we are talking about relative levels of a system or subsystem e.g., organic bases vs. their arrangement in a code, or the reactions entered into by chemical trace elements vs. the structure of the metabolic pathways in which they serve. Whatever system or subsystem we happen to be attending to in a given inquiry, however, this relative priority of the higher level obtains. One might call this relation, as I have suggested elsewhere, a principle of ordinal complementarity.  Aristotle understood it well.
d. Finally, there is a special area where the Aristotelian concept of form has proved strikingly parallel to some aspects of modern thought: that is, in his doctrine of soul, the form of organized bodies. We may leave the vexed question of active reason aside and note briefly two theses in Aristotle’s “psychology” to which modern reflection about biology seems to be slowly and painfully returning.
First, Aristotle’s concept of psyche as such is functional. This is a concept of mind suggested by thinkers as different as Putnam, Ryle, and Polanyi.  Polanyi puts this thesis in his essay, “Logic and Psychology,” in terms of his distinction between focal and subsidiary awareness. We can formulate the distinction between mind and body, he says,
as the disparity between the experience of a subject observing an external object like a cat, and a neurophysiologist observing the bodily mechanisms by
30. The Knower and the Known, 233.
31. Hilary Putnam, “The Mental Life of Some Machines,” Intentionality, Minds and Perception (Detroit, 1967), 177-200; Gilbert Ryle, The Concept of Mind (London, 1949); Michael Polanyi, op. cit.
which the subject sees the cat. The experience of the two is very different. The subject sees the cat, but does not see the mechanism he uses in seeing the cat, while, on the other hand, the neurophysiologist sees the mechanism used by the subject, but does not share the subject’s sight of the cat... to see a cat differs sharply from a knowledge of the mechanism of seeing a cat. They are a knowledge of quite different things. The perception of an external thing is a from-to knowledge. It is a subsidiary awareness of bodily responses evoked by external stimuli, seen with a bearing on their meaning situated at the focus of our attention. The neurophysiologist has no experience of this integration, he has an at-knowledge of the body with its bodily responses at the focus of his attention. These two experiences have a sharply different content, which represents the viable core of the traditional mind-body dualism. “Dualism” thus becomes merely an instance of the change of subject matter due to shifting one’s attention from the direction on which the subsidiaries bear and focusing instead on the subsidiaries themselves. 
Thus, mind is not a separate something but is what Ryle calls “minding.” It is the higher-level, operating principle of a complex system:
Some principles, - for example, those of physics - apply in a variety of circumstances. These circumstances are not determined by the principles in question; they are its boundary conditions, and no principle can determine its own boundary conditions. When there is a principle controlling the boundary conditions of another principle, the two operate jointly. In this relation the first can be called the higher, the second the lower principle.
Mental principles and the principles of physiology form a pair of jointly operating principles. The mind relies for its working on the continued operation of physiological principles, but it controls the boundary conditions left undetermined by physiology. 
This approach is generalizable, moreover, to living things as such. In general the “soul” of any living thing is its style of operating on and in its environment, no more, but also no less.
Secondly, the “kinds” of “soul” distinguished by Aristotle appear to correspond in general to the major divisions which, however we see the problem of “higher” and “lower” among organisms, do in fact seem to obtain. If we look, independently of any special theory of the how of evolution, at its general course, we find, I think, three and only three really “surprising” “advances.” First, there is the origin of life. Here we get what Aristotle calls “nutritive soul.” Living things grow and reproduce. These are the minimal functions of all life. Secondly, living things acquire the principle of sentience and self-locomotion. This is “sensitive soul.” We get organisms capable of behavior, centers of irritability, appetite, and self-motion. And thirdly we have the origin of man, of culture, of the human social world, of what Aristotle calls “passive reason.” Surely these steps,
32. M. Polanyi, “Logic and Psychology,” Amer. Psychologist, 23, (1968), 39-40.
distinguished by Aristotle, are just those that may well puzzle the evolutionary theorist.  These are very crude distinctions if you like, but it may be worth noting that with painful deviousness we are coming back to the simple divisions in the levels of organization in the world around us which Aristotle had recognized long ago, divisions which he made, as we are trying to do, not dualistically, like Plato, but in terms of function, of the inherent organizing and operating principles that mark off kinds of complex systems as unique.
So much for eidos in contrast to hyle. What about eidos in contrast to genos? It is here that the opposition to Aristotle is centered. For if modern biologists in fact use concepts like organization and information in ways that resemble Aristotle’s use of “form” in the form-matter pair, they strongly object to the Aristotelian species concept, even though in Aristotle this appears to be, if not the very same concept, at least the same concept under another aspect. I want to make two points in this connection: first, to indicate how the use of the species concept in modern biology does still resemble the Aristotelian, and secondly, to explain, or at least to locate clearly, the modern resistance to Aristotelian thinking on this score. Again, however, let me reiterate briefly what I said at the outset about the different cosmologies associated with the ancient and the modern view. Aristotelian species are certainly eternal, modern species certainly are not. This meta-scientific contrast should not be underrated. A modern biologist can no more be a complete Aristotelian than he can be a complete Cartesian. Yet in the routine use of the species concept there is nevertheless a residual, though not a merely vestigial, similarity, and at the same time, in the epistemological foundation of that use, a very deep-seated contrast.
First, the similarity. Eidos and hyle form, we have seen, throughout the range of Aristotelian sciences, an analytical pair to be used relatively to one another and to the subject matter in question in a particular investigation. When it comes to the eidos/genos contrast, however, eidos assumes a different and less relational aspect. (This holds, I think, even if we admit that genos means hyle, as it sometimes does; see note 27.) Only individuals are real for Aristotle - the modern biologist would agree - but they are individuals of such and such a kind. The infima species, like any form, exists only in, and as form of, the individuals who exemplify it. The species is the sum total of its specimens, past, present, and future, and they are the individuals they are in virtue of their membership in that species. But there is nothing relative about this. Eidos interpreted as species takes on an absolute character which, in the natural world at least, eludes eidos as paired with hyle. This of course, in its explicit enunciation at least, is just
34. Cf. for example, the obscure but thought-provoking argument of David Hawkins in The Language of Nature (Garden City, N.Y., 1967).
what modern taxonomists, whether evolutionists or pheneticists, so stoutly object to. And yet practising taxonomists of whatever school do in fact continue to treat “species” as having a special role, a role in some way less conventional and closer to the real ways of nature than the concepts designating “higher categories.” True, as Simpson points out, all categories are “objective” in that all taxa are collections of real organisms. And they are all “subjective” insofar as they are all concepts in the minds of taxonomists. Nevertheless, he admits, “species” is more clearly “non-arbitrary” than other categories. 
Admittedly, some biologists, from Darwin himself to Ehrlich and Holm, have predicted that the species concept would wither away and we should be left with a classless aggregate of biological particulars.  Yet biologists still classify and argue about the foundations of such activities; and in the view of most of them the species concept has been refined, indeed, even transformed out of recognition, but not abolished. Perhaps this is correct. Certainly, the “biological species concept,” defined in terms of potentially interbreeding Mendelian populations, or the “multi-dimensional species concept,” tailored for the inclusion of non-sexually reproducing as well as of Mendelian populations, looks at first sight very unlike the traditional originally Aristotelian, concept.  I want to point out only that it is not as wholly unlike as it is usually painted.
Aristotle is usually accused of tagging species by means of one single character selected a priori and abstractly.  This is unfair. He is certainly no a priorist. Again, modern science had to reject him at its outset because he was not “a prioristic” enough: he stayed too close to the concrete pronouncements of everyday experience, he was too good an empiricist.  And even though he writes in the Topics of, and is perpetuated in the tradition as insisting on, definition per genus et differentiam, he himself suggests both in the Post. Anal. and in De Part. Anim. that the “substance” of a thing (ousia) or its nature cannot be captured by specifying any one differentia alone.  Indeed, his opposition to Platonic “division” is based at least in part on the insistence that we divide up natural things as nature demands, not by one character and its contrary, but by the cluster of characters which helps us to single out a natural kind within a larger group. 
35. G. G. Simpson, op. cit., 114.
36. Darwin, Origin of Species, Ch. XV (“species are only well-marked varieties”); Paul R. Ehrlich and Richard W. Holm, “Patterns & Populations,” Science, 137 (1962), 652-57.
37. See Ernst Mayr, Animal Species and Evolution (Cambridge, Mass., 1969), 18-20.
38. See e.g., Simpson, op. cit.
39. See Portrait of Aristotle, esp. Ch. III, and Wieland, op. cit.
40. Post. Anal. II, 13, 96a33 if. and Part. Anim. 634b29ff.
41. Part. Anim., loc. cit.
In this piecemeal and empirical approach to classification he is not so different from modern taxonomists as it is now fashionable to consider him. And again, like modern biologists, he is driven, despite his insistence that only individuals are real, to grant to the species concept some kind of uniqueness, as the least and most “real” of universals. In short, it is in eidos as species that the relativity of form is somehow or other anchored in reality.
Just how this anchoring comes about is hard to say. Just how is the relational concept which we render “form” to be identified with the non-relational concept called “species” in the inclusive but univocal concept eidos? I can give no satisfactory answer, but only suggest the location of an answer in the even more puzzling conception of the [HHC: Greek not reoproduced], which I shall discuss briefly in the concluding sections of this paper. Even more difficult would be the task of specifying the lesson to be derived for modern biology from the plurality-in-unity of Aristotelian eidos. I can only register tentatively the suggestion that one might profitably reflect on the link, whatever it turns out to be, between organization or information on the one hand and species on the other.
Biologists study throughout the widely varying phenomena of living nature the organization of systems or subsystems at any number of levels. Living things, as information-bearing systems, have arisen gradually from non-living systems much poorer in information content, and, once evolved, they continued to vary continuously, to throw up, in correlation with their changing environments, myriad new patterns in every conceivable direction of novelty. At the same time, there are cuts in this continuum, not only the infinite number which we might make anywhere, but a few (in relation to the infinite possibilities) which present themselves as preeminently “natural” or, in Simpson’s term, “non-arbitrary.” These we designate as cuts between “species” that is, between carriers, for a time, of distinctive patterns of information. One can study the organization of any organized system or subsystem, of chloroplasts, cell membranes, muscle cells, populations of genes, populations of whole organisms, communities of populations of many species, etc., etc., but there are also some points at which the transfer of a stable pattern of organization (or information) from one living individual to another stops - stops “really,” not because we decide to stop analyzing just there, but because there is a gap, a real discontinuity. To populations confined by these plain discontinuities (plain at least in sexually reproducing organisms with a relatively long generation span),  we give the same
42. For the complexities, e.g., of bacterial taxonomy, see Mortimer P. Starr and Helen Heise, “Discussion,” Systematic Biology, Nat. Acad. Sci. publication 1962 (1969), 92-99.
name that Aristotle gave them; eidos, species, the very name he gave to organization, eidos, forma. But here we are not, as in multi-levelled analysis of forms and their matter, singling out such patterns as we discover by applying “form” and “matter” as shifting and relational tools for our own study of nature, tools that locate form-in-matter here, there, and everywhere. We are finding certain forms singled out for discontinuity within the continuity of the phenomena, as it were, by nature itself. It may of course be objected, as Simpson remarks about higher categories, that in a sense all forms discovered everywhere are equally objective. Every cell is really bounded by a real membrane: the cytologist studies in the parts of one kind of organism the structures of cells as such. Geneticists study drosophila not because the species of that genus are themselves of greater intrinsic interest than elephants or antelopes, but because they are good experimental subjects and so from them much can be learned about the organizing principles of all heredity. Yet there is something unique - even uniquely obtrusive - about species. The developmental biologist - as distinct from the old-fashioned zoologist or botanist - studies organized processes that range much farther than any given species, or even phylum or kingdom. Yet even he has to select, and learn to know, some species in order to study in it the universal life pattern that interests him.  Much as he would like to, he cannot evade these fundamental gaps. In practice he is still an Aristotelian in spite of himself.
Yet in their attitude to taxonomy, ancients and moderns are very different. If forms are really pinned down into discontinuous species in a fashion not so very unlike that recognized by Aristotle, why should modern biologists so emphatically deny that any shadow of Aristotelian thinking lingers in their own methods? Partly because they take a crude and truncated “Aristotelianism” as identical with the thought of Aristotle himself. But there is another and more deep-seated reason, and that concerns, in Aristotle, the relation of knowledge, especially the knowledge of species, to perception. Modern science, let us recall once more, began by rejecting the Aristotelian approach to nature, in part at least because it was too directly tied to everyday perception of natural entities and processes and so prevented the flights of abstractive thought and creative imagination on which, as we can now see, the development of science largely depends.  The ideal of “scientific method” for many philosophers and
43. Professor Dennis Barrett of the Davis zoology department, while denying that he is a “zoologist,” admits sadly that he has to know “his” organism, the sea urchin, in order to study in it the development of the fertilization membrane.
44. Such founders of modern science as Harvey and Newton, indeed, thought they derived their great discoveries very directly from experience; we, with three centuries of hindsight, know they were more daringly imaginative than they believed.
scientists has become the correlation, not of perceptions, usually directed as they are to complex, concrete individuals, but of sheer particulars, of “hard data,” with abstract laws, whether universal or statistical: correlations peculiarly susceptible, it seems, to quantitative manipulation and experimental control. For Aristotle, on the contrary, knowledge, however theoretical, is rooted in the full, concrete, perceptual world; it analyzes that concrete world and gains new insight into it, but never leaves it as ultimate, as well as initia1, dwelling place. Now such perceptual insight is indeed essential to certain kinds of biological practice: from the macroscopic recognition of specimens in the field to the recognition of structures in electronmicrographs. The late C. F. A. Pantin called such biological connoisseurship “aesthetic recognition.”  The double meaning of “aesthetic”: informal or connoisseurlike on the one hand, and having to do with “aesthesis,” perception, on the other, should be kept in mind in considering what this means. Pantin recalls, for example, seeing a worm in the field and saying, “Why, that’s a Rhynchodemus, but it’s not bilineatus, it’s an entirely new species.” This, he points out, is not the yes-no procedure of the museum taxonomist, nor does it resemble at first sight the generalizing procedures of the exact scientist.  It is, precisely as for Aristotle, a case of seeing a this-here as a such-and-such - or in this case, not quite a such-and-such but a somewhat-different. But just such perceptual recognition of real kinds is what modern theorists profess to abhor. They seek to produce scientific knowledge in an abstracter and more completely specifiable way. Scientific knowledge has its pedigree, they claim, by mathematical thinking out of bare particulars, rather like Love in Diotima’s story, who was begotten by Resource on Poverty.
This distrust of anything but bare particulars on the one hand and high flights of theory on the other comes clearly into view in the contemporary taxonomic controversy between the phenetic and phylogenetic schools.  The pheneticists profess to take all and any particulars, without prior weighting derived from taxonomic skill and experience, feed them into computers (those praiseworthy inorganic animals) and come out with classifications better (for what purpose?) than those derived from less restricted starting points and less quantitative manipulation. If they sometimes admit sadly to producing by this method something justifiably called “types,” they are at any rate, they claim, “empirical typologists,” with no initial predilection
45. C. F. A. Pantin, “The Recognition of Species,” Science Progress, 42 (1954), 587-98; cf. his posthumous Tamer lectures, The Relations between the Sciences (Cambridge, 1965).
46. “The Recognition of Species,” 587.
47. D. Hull, “Contemporary Systematic Philosophies,” Annual Review of Ecology and Systematics, 1(1970), 19-54 and M. Starr and H. Heise, op. cit.
for one cluster of characters rather than another. Phylogeneticists, on the other hand, hasten to cover their undoubted, but theoretically suspect, taxonomic insights, derived from the aesthetic recognitions of field experience, under the convenient bushel of evolutionary descent. Darwinian-Mendelian theory, in other words, serves them as an abstract and therefore scientifically respectable cover for their delicate perceptual discriminations: for the .heritage of Ray, Hooker, and even, on the side of biological practice as against theory, of Darwin himself. Particulars tied together by computer techniques, says the one side; particulars tied together, says the other, by lines of descent inferred though necessarily unobserved. Some writers, notably Gilmour, try to go between the horns of the phenetic-phylogenetic dilemma by espousing a pleasant pragmatism.  We classify as we need to for our uses, says he; as the use shifts, so does the classification. This easy way out, however, while correct in a way, since there are of course many possible classifications of anything for many possible uses, fails to still the controversy. For it neglects the fact that some classifications do seem to be, quite apart from our wants and uses, less arbitrary than others. To give due weight to the role of aesthetic recognition in taxonomy, I submit, and to acknowledge its rootedness in the perception of real this-suches, would permit taxonomists to make more sense than they have recently done of the real nature of their calling.
It should be duly noted, in passing, that the grounding of scientific knowledge, and especially of scientific discovery, in perception (rather than in sensation or the bare observation of bare particulars) is beginning at last to be acknowledged by philosophers of science. In a general way, perception as the paradigm of discovery was the Leitmotif of Hanson’s writing.  The “primacy of perception” as our chief path of access to reality was the central theme of Merleau-Ponty’s work.  A similar theme dominates Straus’s phenomenology.  And in Polanyi’s Personal Knowledge, Tacit Dimension, and other essays, both earlier and more recent, one has, as distinct from those more general intuitions, a carefully articulated epistemology which explicitly makes of perception, understood in a Gestalt-cum-transactional fashion (not unlike Aristotelian aesthesis), the primordial and paradigm
48. J. S. L. Gilmour, “Taxonomy,” Modern Botanical Thinking (Edinburgh, 1961), 27-45.
49. N. R. Hanson, Patterns of Discovery (Cambridge, 1965). Cf. also his posthumously published Perception and Discovery (San Francisco, 1969).
50. M. Merleau-Ponty, La Phénoménologie de la Perception (Paris, 1945), trans. by Cohn Smith, The Phenomenology of Perception (New York and London, 1962).
51. See E. W. Straus, The Primary World of Senses, trans. by J. Needleman (New York, 1963) and Phenomenological Psychology, trans. in part by Erhing Eng (New York, 1966).
case of knowing, and explicitly makes the achievement of perception the primordial and paradigm case of discovery.  These lessons are beginning to have some impact on philosophers of science, especially on those who base their philosophy largely on physics. But biology, from which, through Aristotle’s biological practice, the acknowledgement of the primacy of perception took its start, still (with a few honorable exceptions like that of Pantin) stubbornly resists this fundamental insight.
I am not, of course, alleging that perception in its newly-discovered role plays the same part as it did in Aristotelian science. It was the basis, and the home, of discovery and of knowledge; it is the primordial case of discovery and of knowledge, and all discovery and all knowledge are structured as it is. In Polanyi’s terms: in all knowledge, as in perception, we rely on subsidiary clues within our bodies to attend focally to something in the real world outside. However “abstract” that something be, both the bodily base and the from-to structure characteristic of sense perception persist. In the present context: there is no disgrace, therefore, in acknowledging the perceptual skill of field naturalists and taxonomists as part of science. It is not a “primitive” survival, but a visible analogue of the achievement of knowledge, the paradigm case of our way of gaining contact with reality.
Let me return now to the question raised earlier: how does eidos as the correlate of genos escape the relativity of the eidos/hyle pair? What is unique about species that makes it the paradigm case of natural form? Here we come to that most idiosyncratic of Aristotelian terms: the [HHC: Greek not reproduced], what it is for a such-and-such to be a such-and-such. “Essence,” with its age-old accretions of misleading connotations, is a poor translation of Aristotle’s phrase; perhaps “being-what-it-is” is the best one can do.  When one speaks of “form,” this is what one’s discourse is aiming at; the form of a given kind of thing is just what it is for that thing to be the kind of thing it is.
According to the Topics, the “being-what-it-is” of a kind of thing is what its definition designates. Indeed, a definition is there defined as “a phrase indicating the being-what-it-is.” To some interpreters, therefore, [HHC: Greek not reproduced] appears to be primarily a logical term. Aristotelian science starts from first principles, including real definitions. These specify, among the properties of a thing, certain characters which are “essential” to it, and that means characters which
52. M. Polanyi, op. cit., also The Tacit Dimension (Garden City, N.Y, 1966). Cf. William T. Scott, “Tacit Knowing and The Concept of Mind,” Phil. Quart. 21(1971), 22-35.
53. Cf. “What-is-being,” the rendering of Joseph Owens in his Doctrine of Being in the Aristotelian Metaphysics (Toronto, 1957).
can be deduced from the initial predicating statement. Definitions, in other words, are premises of scientific demonstrations, and the [HHC: Greek not reproduced] is of interest as the reservoir, so to speak, from which the predicate of such a premise can be drawn.  Such a rendering seems to bring the Aristotelian approach close, in form at least, to the so-called “hypothetico-deductive method.” It must of course be noted, however, that Aristotelian demonstration is anchored, through nous, in the direct, “intuitive” knowledge of first principles. It is not, like dialectical syllogism, merely hypothetical. Yet there is a parallel. For in both cases it is the deduction of some properties from others that is chiefly of interest, and the [HHC: Greek not reproduced] is thought of in this context as ancillary to this primarily logical game. So far as the importance of deductive method goes, Aristotle himself, at least in his logical writing, certainly reinforces this impression.
Yet if one searches the corpus for scientific demonstration, one finds relatively little of it. Most of the arguments of most of the treatises do not look like assertions of defining phrases followed by deductions from these. Most of them appear to be not strictly demonstrative, but inductive, dialectical, or aporetic. They move from common experience or common opinions, weighing the views of others, analyzing difficulties, in the hope, it seems, of arriving at (not starting from) an insight into some specific nature. This searching nature of much of Aristotle’s writing leads W. Wieland to a different interpretation. For him the [HHC: Greek not reproduced] is not so much a guide to definition, and therefore to demonstration, as it is a heuristic tool, a topos or path along which the thinker may seek insight into some special problem. 
For form and matter this heuristic or methodological interpretation is indeed fruitful, as I have emphasized. Yet via form as species we have seen that there is also a resting place for form-matter analysis, a place at which form becomes uniquely non-relational. And it is here that form exhibits its ontological foundation in “the being-what-it-is” of each kind of thing, the very foundation which, according to
54. The “logical” is one of the aspects of the [HHC: Greek not reproduced] distinguished by C. Arpe in his dissertation, Das [HHC: Greek not reproduced] Aristoteles (Hamburg, 1937). Cf. E. Tugendhat, TA KATA TINΣ (Freiburg/M’ünchen, 1958), 18, n. 18: Erst [HHC: German not reproduced]. But cf. also Wieland, op. cit., 174; the “logical” here is perhaps closer to what I am calling the “methodological” interpretation. It is closer to heuristics - the search for principles - than to demonstration from them. The deductive aspect is stressed by Prof. Moravcsik in his reading of [HHC: Greek not reproduced] (oral communication).
55. Wieland, op. cit., pp. 174-75. Wieland emphasizes the methodological function of the [HHC: Greek not reproduced] (which he identifies, by implication, with Arpe’s “logical” function) to the exclusion of its other aspects, esp. what Arpe calls the “physical” or “teleological” and the “ontological.”
Aristotle, his predecessors lacked, the ignorance of which prevented them from discovering the right method for the investigation of nature. Neither the logical nor the methodological approach to Aristotelian science makes sense without this frankly ontological foundation. The [HHC: Greek not reproduced] is expressly what definition is about, both its target and its presupposition. Indeed, without the [HHC: Greek not reproduced] as the referent of definition, the real being-what-it-is of this kind of thing which the defining phrase designates, the demonstrations that follow on definition would be simply “hypothetico-deductive” in the sense of positivism or phenomenalism. They would not be rooted, as Aristotelian science was certainly meant to be, in the natures of the things themselves, and in our understanding of these natures. Without the [HHC: Greek not reproduced] as endpoint and foundation of inquiry, moreover, the inquirer would be confined to an endless and directionless groping; but that is not the case. On the contrary, the real being-what-it-is of the kind of entity under investigation is constantly guiding his search, and directs it to its successful issue. Despite the apparent formal correctness of the logical interpretation, therefore, and despite the importance of the methodological aspect, the traditional ontological interpretation of the [HHC: Greek not reproduced] is still fundamentally, sound. 
What does it teach us? We would not link the being-what-it-is of things, as Aristotle did, to eternal kinds, nor would we restrict these “kinds” to kinds of substance. Everything becomes, including species; and what becomes may as well be events or processes as more literally “things.” But despite these deep differences, there are, I believe, two important lessons to be learned from the Aristotelian concept of being-what-it-is as the designatum of definition and the target of inquiry. Aristotle understood, as most modern philosophers of science until recently have not, that the investigation of nature arises out of puzzlement about some particular problem in some limited area: nobody investigates, or can investigate, everything at once. And in such an investigation, further, it is the real nature of the real entity or process that the investigator seeks, and sometimes finds. Science is pluralistic and realistic, not uniform and phenomenalist, as modern orthodoxy has supposed.
If we put these brief remarks together with our earlier reflections about form and matter, we find, in conclusion, three important methodological lessons to be derived from the study of Aristotle. Through the concept of form as an analytical tool, correlative with matter, Aristotle can remind us of the many-levelled structure both of inquiries into complex systems and of the systems themselves, and thus of the inadequacy of a one-levelled atomism for the understand-
56. See Arpe, op. cit.. For a clear summary of the traditional view, see Bernard J. Lonergan, S. J., Verbum - Word and Idea in Aquinas (Notre Dame, 1967), 16-25.
ing of such systems. In conjunction with the grounding of form in the [HHC: Greek not reproduced] of each kind of thing, further, he can remind us of the falsity of two other modern misconceptions: the unity of science concept on the one hand, the claim that the subject-matter and method of science are everywhere the same, and, on the other, the insistence that science must renounce any claim to seeking contact with reality: that theories float, as pure constructs, on the surface of the phenomena, with no mooring in the real nature of the real events or things.
The first of these reminders is plainly related to the concept of organization or information and hence (as I have already argued) to the subject-matter of biology, and to the question of its reducibility or irreducibility to chemistry and physics. The second reminder, of the plurality of science, a reminder of the good Kantian principle  that we can have no systematic knowledge of the whole of nature, should help also to liberate biology, or thinking about biology, from the overabstract and reductive demands imposed by taking one science, classical physics, as the ideal of all. And lastly, the acknowledgement of scientific realism should release the biologist to admit the insights into the concrete manifold of his subject-matter, from which his work originates and in which, however abstract and sophisticated it may become, it still is anchored. Nor, finally, as I emphasized at the outset, is this a plea for a return to Aristotelianism. It is a plea for us to listen, despite our fundamental differences of metaphysic and of method, to some of the tenets that Aristotle, as a biologist-philosopher, advocated long ago, and to try to interpret them in ways that could be useful to us as we attempt to articulate and revise our conception of what the investigation and the knowledge of living nature are.
University of California, Davis.
57. In a discussion of a similar argument, Professor Günther Patzig has pointed out to me that the principle of the plurality of science is non-, even anti-Kantian, if by it we mean to espouse a plurality of scientific methods. For Kant the method of science was indeed one. What is Kantian, however, is the denial that we can have one unified, finished system of knowledge for the whole of nature. If we could have such a system, we could not have diversified sciences with diversified methods. If we cannot have such a system, on the other hand, then a plurality of fields, and of methods, is at least logically possible, and on Kantian grounds.