[PENULTIMATE DRAFT of a paper now published in Aristotle and Contemporary Science,  Vol. I,  ed. Demetra Sfendoni-Mentzou, Intro. by Hilary Putnam. New York, Berlin:  Peter Lang,  2000]

© Ronald de Sousa
University of Toronto.     (sousa@chass.utoronto.ca)


I have learned with some degree of wonder, in the course of the present conference, that Aristotle knew everything worth knowing. By a happy coincidence, this knowledge coheres perfectly with what we now believe we know about everything worth knowing, including matter, mind, space and time, social science, and even quantum mechanics and DNA. The present contribution, I fear, will strike a dissonant note, in suggesting that there was actually something Aristotle didn't get right.

My plan is as follows:

I shall first explain why Aristotle has an unsolved problem about the teleology of decay and death.

Next, I shall describe two modern attempts at a solution to this problem, or rather one attempt to reject the problem altogether and one abortive solution, and explain why both of these fail.

Third, I shall sketch the surprising correct answer, which is that there appears to be a degree of evolutionary necessity connecting the following four characteristics of human beings and other animals:

(i). Differentiated multi-cellular organization;

(ii). The segregation of sexual and somatic cells;

(iii). Sexual reproduction; and

(iv). Death.

In conclusion, I shall raise a question about the nature of the "evolutionary necessity" involved in the claim just summarized.

Aristotle's Unsolved Problem.

Death and decay are part of nature. They are therefore natural phenomena.[1] Aristotle seldom denies obvious commonplaces, and he certainly nowhere denies this one. Yet this commonplace poses an insurmountable problem for his general theory of nature. How could this be?

The reason relates to Aristotle's theory of natural objects or substances, which reiterates in many places that substances, which is to say naturalobjects, are structured by an inherent teleology. (Met. Z.9:1034b11) Thus the oak exists potentially in the acorn, and becomes actualized when the acorn grows into the tree; but the actual oak is prior to the acorn in virtue of the general priority of actuality over potentiality. (Met. Z.13:1038b27) So the nature of any substance involves a natural development in the course of which potentiality flowers into actuality.

Why not say fruits into actuality rather than flowers? To raise this apparently trivial question about the appropriate metaphor is to set oneself the problem that Aristotle can't solve. For what is, from the teleological point of view, the phase in the life-cycle of a plant when it has reached its full actuality? Whether the answer lies in the flower or in the fruit is impossible to decide without committing oneself to interests, in flowers, fruit, or seeds, that seem to be external to the individual substance concerned. But even supposing that there are objective grounds for arriving at a correct answer, that answer clearly cannot be one which locates the final cause of an individual's development in decay and death. Although these are the plant's end in the temporal sense, they are not its end in the teleological sense. But why not?

A possible solution to this problem might come from the expansion of our perspective from the contemplation of a single natural object to the contemplation of the universe as a whole. We can then think of the succession of individual organisms that are born and die as forming only a partial picture of a whole which attains to perfection only in its entirety. Such a view seemed attractive to Descartes and to Leibniz. It is, at least, in the spirit of Descartes's last defense of God against the assault of the problem of evil: "there may in some way be more perfection in the universe as a whole because some of its parts are not immune from error, while others are immune, than there would be if all the parts were exactly alike." (Descartes 1984, p. 61). In the Monadology, Leibniz makes much the same move: the universe is so created as to produce "as much variety as possible, but with the greatest order possible, that is, ... as much perfection as possible." (Leibniz 1991, sec. 58). But the consensus of Aristotelian scholars seems to be that it cannot be attributed to Aristotle, for whom the teleology of individual substances is not subservient to some grander scheme in which the universe as a whole is a substance.[2]

Solutions and non-solutions

In a post-Darwinian perspective, one way of dismissing the problem as well as one bogus solution suggest themselves.

The bogus solution is that death is an adaptation, in the sense that if organisms did not die, they wouldn't leave room for evolution to take place. This solution, suprisingly, appears to have been seriously advanced in some quarters.[3] It is bogus, because it imputes to evolution a completely impossible sort of global teleology, as if evolution were itself an agent bent on eliminating any obstacles in its path. The underlying assumption seems to be that evolution is a process that leads to ever "higher" forms, for which "lower forms" have to make way, as a craftsman might need to clear his workspace of older débris to make way for newer works.

This is, of course, doubly absurd. First, many so-called lower forms are actually the most enduring of all living things. In fact it has been estimated that unicellular organisms comprise more than half of the earth's total biomass. Unicellular organisms can be regarded, in a sense to be elucidated in a moment, as immortal; and if one can imagine the earth having been filled all at once with multicellular organisms that were in the same sense immortal, then there might indeed have been no interesting process of evolution. But -- and this is the more important point -- there was no prior necessity for evolution to take place, though one may concede the modest tautology that if existing organisms had left no space for new forms, there would have been no new forms.

So much, then, for the bogus solution. The more tough-minded neo-Darwinian might be tempted to reject the problem out of hand. For a Darwinian, the problem of the teleology of death does not exist, since from that point of view all teleology is a kind of illusion. Though teleological talk in biology is indeed common, it is constrained by the assumption that wherever it is legitimate to see a teleological element, it can be accounted for in terms of natural selection. Conversely, much of what we find in nature may well be due not to natural selection or adaptation, but to chance, drift, or even "order for free" (Kauffman 1995); but none of that can be thought of as literally goal-oriented. Consequently, the issue raised by the Aristotelian metaphysic no longer arises. Such is the non-solution to what may now seem to be a non-problem.

This non-solution can be elaborated a little. It is known that -- with some specific exceptions which will concern us in a moment -- cells set to divide freely in vitro will continue to do so for only a few dozen generations. Thereafter, the copying process peters out, as if copies had degraded to the point where they are incapable of further reproduction. If we remember that any process of copying, however careful, involves a certain probability of introducing copying errors, this suggests what might be called the xerox-copy model. Faithful reproduction to any given number of copying generations can be secured by a suitable increase in the redundancy of the information to be copied. (Thus when telegraphing a number, it is advisable to send it in letters as well as in figures). But the introduction of redundancy is not cost-free. In the case of sexually reproducing organisms, natural selection is likely to prefer those organisms that have enough redundancy in their cell-reproducing mechanisms to endure long enough to produce a new generation. But once the organism has actually reached sexual maturity and produced offspring, the genes it has passed on are more efficiently conserved and reproduced by those offspring than by the continuation of the parents' own reproductive activity. The reason is that the latter option would demand a higher level of redundancy in the parents' own somatic cells. Think of two ways of producing 50 copies with a xerox machine. If the same original is copied each time, the 50th copy (or the 1000th, for that matter) will be as clean as the first. But if each time what is copied is the last copy made, the 50th will doubtless be quite unreadable. Thus the probability is high that by the time we have made 50 successive copies (or lived about 120 years of a human life) the next copy will contain too many errors to breed true.

The biological surmise which this suggests, crudely put, is that the sex cells provide the intact archival copy that is referred to when new runs of copies are to be made, and death is a side-effect of the limited redundancy of genetic information encoded in each somatic cell. At every generation, copies must be made from the original -- the sex cells -- because the copies already made of somatic cells in this generation are no longer capable of producing more. Death is in no way an adaptation; rather it is merely the absence of any "adaptation" that would entail perpetual life.

The xerox-copy model can be regarded as the null hypothesis regarding the existence of death. Oddly enough, however, there are facts that appear to undermine it. I mention just two:

One is the existence of a process of apoptosis, or programmed cell death. While there might be some measure of truth to the xerox-copy hypothesis, the extraordinarily orderly process of cell suicide refutes the hypothesis that the only cause of cell dissolution and death is their inability to breed true (see Lavin and Watters 1993).

The second reason to reject the null hypothesis is that it is in fact possible for cells to reproduce themselves more or less forever. But only three sorts of cells are known to do this: unicellular organisms such as bacteria, sex cells, and cancer cells. Recent work has localized the problems in copying in the ends of the chromosomes, the telomeres, and identified an enzyme, telomerase, capable of regulating the telomeres and through them the longevity of the cells to which they belong. These two facts together suggest that without actually rehabilitating the bogus adaptiveness of death, there may be an element of teleology involved in the ineluctability of death for metazoan individuals such as ourselves.

The Sex and Death Syndrome.

Very briefly, the correct answer to Aristotle's problem seems to be that a tight relation of mutual dependency exists between the four following characteristics of certain sorts of individual:

1. Being a metazoan (i.e. a multicellular, differentiated, individual animal);

2. Exhibiting a relatively strict segregation between sexual cells and somatic cells ("Weismann's barrier");

3. Being a sexually reproducing animal; and

4. Being an individual destined for death.

Before I say a little more about these features, two qualifications are in order.

First, it's well to note that there exist multicellular differentiated organisms that are not strictly metazoans, but plants. Plants, and some fungi, have followed a different evolutionary path and present different problems. The present sketch leaves them out of account.

Second, there is a sense in which a tight and straightforward connection exists between sex and death even in unicellular organisms. Not all sex is reproductive (even setting aside the purely social or recreational functions of sex in primates). Sex, in the relevant sense, may be defined as the exchange of genetic material (Margulis and Sagan 1986). In the typical case of sexually reproducing species, this involves the continued existence of the sexually reproducing individuals whose sex cells are merged, and the recombination of some of those genes into a new genome. The total number of resulting individuals is increased, but here the connection with death remains to be established, since there does not seem to be any a priori reason that the parental individuals might not continue to live forever. But an exchange of genetic material sometimes occurs between bacteria without any modification in the number of existing cells. (This is known as conjugation, and plays an important role in the story I am about to tell.) All such genetic exchanges result in the extinction of the original genomes, which is a kind of death of the individuals instantiating that genome.

The logic binding the four factors[4]

While it is wise to eschew a naive adaptationism which assumes that all characteristics of "higher" organisms must have entailed a competitive advantage, it still makes sense to ask what differential advantages organisms combining these four features might have enjoyed, and what precisely is the logic that binds the four together. The answer, much simplified, seems to be something like the following.

The existence of multicellular differentiated organisms is obviously contingent: they might never have evolved. Two accidents appear to have been necessary to bring them into existence.

One accident was the occurrence -- preserved well enough by natural selection to spread -- of the process of conjugation just mentioned. The exchanges in question -- lethal though they be to at least one of the parties, as I have just suggested -- may have had the happy effect of allowing one cell to use the genetic material of another in order to repair its own genome, where that had been damaged by copying errors. (Margulis and Sagan 1993). The existence of such a repair mechanism would have led to a greater stability in the genomes endowed with it. This might have been enough to keep a primitive gene-exchange recipe in circulation until it could be adapted to sexual reproduction. Such a mechanism would have the double effect of producing a greater variety of forms, while eliminating non-viable forms. The second accident required for the process to begin was the "discovery" of cooperation between cells, involving some form of specialization. Cooperation at various levels may have involved processes of group selection,[5] since cooperation between different cells (or different species, for that matter, in more exotic forms of symbiosis) entails a corresponding increase in mutual dependency. Cooperation is likely to be more effective insofar as it is based on differentiation of functions, making possible a division of labour.

One form of division of labour is of particular interest here: the distribution of roles (and indeed an increasingly strict segregation) between somatic cells, which divide only by mitosis, producing two identical copies of the "mother" cell, and the gametic cells, which sometimes divide by meiosis, of which the resulting cells are equipped with only half the mother's genetic information.

At the end of the last century, August Weismann posited the continuity of "germ plasma," which entailed the impossibility for any modification of the somatic cells to be communicated to the germ cells. On this hypothesis, since any modification of the somatic cells of an individual, whether brought about by striving, by environmental luck, or by any other factor, is precluded from being communicated to the germ line, no acquired characteristics can be passed on to the next generation. Weismann's guess proved correct, and the sequence of events leading to this segregation is now known: after just a few divisions, some cells, the ancestors of somatic cells, lose their capacity for meiosis, while others, the ancestors of gametic cells, lose their potential for differentiation into specialized organic cells. The impassable barrier thus erected is by no means necessitated; indeed, its impermeability is variable and admits of degrees. The strict constraints which it imposes on metazoan cells therefore requires an explanation in terms of natural selection.

In fact, it is precisely this segregation of gametic from somatic cells which cements the link between sexual reproduction and the death of individual organisms.

Generally speaking, the segregation in question makes sense only where there is sexual reproduction.[6] As for sexual reproduction itself, it is notoriously both costly and dangerous. Costly, in that it requires two individuals instead of a single one to produce another, and dangerous in that every successful genetic combination is destroyed in favour of a new model, unique and untried. For both these reasons, asexual reproduction would prima facie seem the better bet, on the principle that what ain't broke is better not fixed. Nevertheless, though there are asexually reproducing metazoans, all of them have ancestors that reproduced sexually. This suggests that natural selection favoured sexual reproduction despite its drawbacks, probably, among other reasons, because the shuffling of the genetic cards at every generation allows metazoans to stay one step ahead of their parasites, whose own success depends on a stable environment (Ridley 1993).

Weismann's barrier allows the sex cells to devote themselves entirely to the preservation of their genetic message. They can be expected to be best at doing this, and therefore to spread, simply in virtue of the fact that they do nothing else. Such is the quasi-tautological logic of natural selection: those cells that reproduce most faithfully will leave the most faithful copies. This tautology, in fact, is at the heart of the almost irresistible illusion of teleology which result from natural selection: since a presently existing formula is there only because of the existence of certain necessary antecedents, it is tempting to conclude that to produce the former is literally the goal of the latter. It's as if the cells carrying the genetic heritage of an organism manipulated the somatic cells into taking care of them, so as to maximize the extent and accuracy of their own reproductive capacity.

According to the logic of this process, then, we must expect that the genes of metazoans will favour the production of cell assemblies -- individual bodies -- whose job it is to look after their daily subsistence, but without any power to modify their genetic heritage itself. What's more, the indefinite capacity for reproduction that belongs to the sex cells must be actively suppressed in the somatic cells which, in Dawkins's notorious phrase, constitute but the "vehicle" of that genetic heritage. Those somatic cells need to continue just long enough to maximize their chance of passing on that heritage; but that mission accomplished, the somatic cells have outlived their usefulness. Why not then leave them to their own fate, to reproduce if they can and fail when they must? (That would presumably make individual death probable, but not certain.) The phenomenon of apoptosis, demonstrating as it does that cells have programs for tidy self-destruction, shows that they have not been thus left to their own devices. The reason for this is suggested by the simple empirical observation that the one type of cell that does appear capable to reproduce itself forever, apart from sex cells and unicellular organisms, are cancer cells. Cancer cells, we might say, go on reproducing for their own sake, freed of any regard for the cooperative arrangements to which they were originally a party. So by equipping somatic cells with the capability of apoptosis, the genetic program, which is itself protected from contamination by Weismann's barrier, ensures that the somatic cells do not undermine the integrity of their genetic message by degenerating into cancerous cells, reproducing indefinitely in defiance of the interests of the whole organism in which they were once full partners.

In sum, the logic that binds the four factors characteristic of metazoans is this: the chance discovery of conjugation allowed genetic exchange and genetic repair. This in turn allowed for the symbiosis of different cell lines, which made for cellular differentiation. The other characteristics all followed from the necessity for the genetic material to protect itself against the accumulation of copying errors in the reproduction of DNA: sexual reproduction, the segregation of the somatic cell lines, and the death of each individual body. The advantages of this arrangement would seem to be these:

1. The germ line can specialize in replication, free from the worries of provisioning its vehicle.

2. The soma, on the other hand, is kept busy doing just that, and so is unable to compete effectively with the sexual cell line.

3. Just to make sure, it isn't allowed to try, as it is programmed for self-destruction.

Concluding reflections

The existence of metazoans, as I've already stressed, is a historical accident. But without them, there would not exist Rational Animals.[7] All the rest (sex, death, and segregation of soma) happens, as Aristotle might have said, "always or for the most part," though there is a range of variation in the number of sexes, and the strictness of the segregation. There are intriguing but unanswered question is about the nature of the necessity involved. Could there be, on some planet, organisms that fulfilled the first of our four conditions -- differentiated multi-cellular organisms of the degree of complexity exhibited by earthly metazoans -- where not all of the three other factors were present? If not, that may be due not to any inherently biological necessity, but rather to laws that are ultimately of the order of mathematics or economics. Pace those enemies of sociobiology for whom competition among organisms is just a reflection of vulgar nineteenth-century laissez-faire economics rather than a biological fact, it seems that not just genes, but individual cells, cell lineages, as well as groups, are what they are in part as the result of the competitions in which their ancestors found themselves.

Death, then, while an accident of life, is of the essence of the life of metazoan individuals if anything is. It is more tightly integrated into the modern successor to Aristotelian teleology than it is in Aristotle's own theory of nature. Oddly enough, though Aristotle accepted a comprehensively teleological conception of the world, whereas the teleology of modern biology is derivative or metaphorical rather than real,[8] his teleology had no place for death, whereas the modern conception sees death as having a function in the context of a complex of interdependent factors, in which sexual reproduction plays a crucial part. Sex and death are bound together in a tight web which even Aristotle could not have guessed at.


1.   As stressed in R.A.H. King, "Aristotle's Theory of Aging and Modern Gerontology," presented at the conference Aristotle and Modern Science, Thessaloniki,  1997.
2.   See Broadie (1990); Cooper (1982).  For a discussion of some recent uses of the organismic model applied to larger wholes, see Mitchell (1995).
3.   For example, by the eminent French biologist and philosopher of biology Jean Ruffié, who avers:  " La reproduction sexuée crée sans cesse de nouveaux types ... mais ceux-ci ne peuvent diffuser leurs combinaisons ... que si les anciens leur laissent la place.... " (Ruffié, 1986, p. 12).
4. In this section I have drawn extensively from Clark (1996) and Buss (1987).
5. The mechanisms of group selection, and the kinds of conditions that are required for them to operate, have been greatly clarified by Sober and Wilson (1998), so that group selection need no longer give rise to an automatic raised eyebrow on the part of any self-respecting Darwinian.
6. There are, however, exceptions: the humble paramecium, for example, illustrates a primitive version of the tight connection between sexual reproduction and death.  It is equipped with two nuclei. The larger one, which is in charge of all metabolic functions, is doomed to perish; while the job of the smaller job is exclusively to organize the process of cellular division.  Once that division is complete, the principal nucleus divides one more time, and one of its successors forms the large, secondary nucleus, whereas the smaller one takes up its vocation of preserving of the integrity of the genetic material to be transmitted to the next generation.
7. The phrase must, needless to say, be understood correctly. What it actually means is capable of being irrational.  If this is not obvious, I have attempted to make it so in de Sousa (1971).
8.  For a view of teleology as essentially analogical, see Matthen (1997).



Aristotle. 1994. Metaphysics. Trans and commentary David Bostock. Oxford: Oxford University Press.

Broadie, S. 1990. Nature and craft in Aristotelian teleology. In Biologie, logique et métaphysique chez Aristote, eds D. Devereux and P. Pellegrin. Paris: CNRS.

Buss, L. W. 1987. The Evolution of Individuality. Princeton: Princeton University Press.

Clark, W. R. 1996. Sex and the Origins of Death. New York: Oxford University Press.

Cooper, J. 1982. Aristotle on natural teleology. In Language and Logos: Studies in Ancient Greek Philosophy, eds M. Schonfield and M. Nussbaum. Cambridge: Cambrdge University Press.

de Sousa, R. 1971. How to give a piece of your mind, or the logic of belief and assent. Review of Metaphysics 25:51-79.

Descartes, René. 1984. <1641> Meditations, in The philosophical writings of René Descartes. John Cottingham and Robert Stoothof, eds. vol. 1. Cambridge: Cambridge University Press.

Kauffman, S. 1995. At Home in the Universe: The Search for the Laws of Self-Organization and Complexity. Oxford, New York: Oxford University Press.

Lavin, M., and D. Watters. 1993. Programmed Cell Death: The Cellular and Molecular Biology of Apoptosis. Philadelphia: Harwood Academic Publishers.

Leibniz, Gottfried Wilhelm. 1991 Leibnizís Monadology: an edition for students.  N. Rescher, ed. Pittsburgh: Pittsburgh University Press.

Margulis, L., and D. Sagan. 1986. Origins of Sex: Three Billion Years of Genetic Recombination. Yale University Press: Summit Books.

Matthen, M. 1997. Teleology and the product analogy. Australasian Journal of Philosophy 75:21-37.

Mitchell, S. D. 1995. The superorganism metaphor: Then and now. In Biology as Society, Society as Biology: Metaphors, eds S. Maasen, E. Mendelsohn, and P. Weingart, 231-47. Dordrecht; Boston; London: Kluwer.

Ridley, M. 1993. The Red Queen: Sex and the Evolution of Human Nature. New York: Macmillan Viking.

Ruffié, J. 1986. Le Sexe et la mort. Paris: Odile Jacob.

Sober, E., and D. S. Wilson. 1998. Unto Others: The Evolution and Psychology of Unselfish Behavior. Cambridge, MA: Harvard University Press.