Back to Chapter one Back to Part B References

Chapter 1 - What does it mean to be alive?

Part B - A proposed solution to the "problem of life"

Appendix - What are the necessary conditions for something's being alive?

Material requirements

Concrete material realisation

My stipulation of a concrete material realisation as a necessary condition for being alive harks back to Aristotle, who, it will be recalled, refers (De Anima 2.1, 412a20) to the soul as "the form of a natural body" (1986, p. 157, italics mine). As he puts it (De Anima 2.1, 412a19):

For the body, far from being one of the things said of a subject, stands rather itself as subject and is matter (1986, p. 157).

The importance of a concrete material realisation is that it individuates an entity, making it a single subject. Without a material realisation, the entity, having no spatio-temporal location, could not properly be described as a single individual. An entity with purely formal properties would lack a specific "thisness" about it. Lacking individuality, it could not meaningfully be described as benefiting or as being harmed, as no specific (spatio-temporal) events could be said to affect it. Without the possibility of ascribing benefit or harm to such an entity, there is no way in which it could be said to possess a telos of its own. Such an entity would therefore lack intrinsic finality.

The implication of this is that a piece of computer software, even if it realised the properties of having a master program, nested organisation and embedded functionality, cannot be said to be alive, as its realisation is purely formal and not material. Hardware is an essential part of what it is to be alive. Living things are, after all, built from genes, not memes (pace Richard Dawkins).

(Some philosophers would argue that restricting the attribution of "life" to material entities is unduly narrow, as it would rule out the possibility of anything transcendent being alive - God, for instance. I do not propose to discuss this idea within my thesis, as my concern is with how the biological property of being alive is tied to the Aristotelian property of intrinsic finality and its attendant formal, material, efficient causal and temporal requirements. Aristotle himself employed a broader definition of life, which encompassed any kind of activity that could be regarded as having an intrinsic end - even a non-biological activity such as God's self-contemplation (Cameron, 2000, pp. 333-336.)

Spatial contiguity?

At this stage it might be asked whether an organism needs to be spatially contiguous. Could it have parts that were physically separated, but which were able to communicate with each other by transmitting chemical or electromagnetic signals? The idea is not a new one, as Holldobler and Wilson write:

The idea - the dream - of the superorganism was extremely popular in the early part of this [20th] century. William Morton Smith, like many of his contemporaries, returned to it repeatedly in his writings. In his celebrated essay, 'The Ant Colony as an Organism', he stated that the animal colony is really an organism and not merely an analog of one. It behaves, he said, as a unit. It possesses distinctive properties of size, behavior, and oragnization that are transmitted from one generation to the next. The queen is the reproductive organ, the workers the supporting brain, heart, gut, and other tissues. The exchange of liquid food among the colony members is the equivalent of the circulation of blood and lymph (1994, p. 110, italics mine).

Ants communicate with each other chemically, via pheromones, so the unity of the colony is maintained. The question of whether an ant colony is really one organism will be discussed in a later chapter, but the illustration has a point. There seems to be nothing inherently impossible in the idea of a spatially discontiguous organism. The conceivability of such an organism, however, in no way guarantees its real possibility. Moreover, there are two problems with the idea that need to be addressed: first, in what sense could a disconnected organism be described as one (the "unity" problem), and second, how could two such organisms be distinguished from one another (the "individuation" problem)?

I would suggest that the solution to both problems lies in Aristotle's notions of formal and final causation, recast as a master program with a nested hierarchy of dedicated functionality, possessing embedded functionality. I would propose that if:

(i) the parts of a superorganism, although physically separated, are controlled by the same master program, so that they are guaranteed to work together for the good of the whole;

(ii) there is a nested hierarchy of functionality;

(iii) the repertoire of the parts' functionality is dedicated to supporting the functionality of the whole to which they belong; and

(iv) none of these parts is capable of surviving on its own,

then there is no real reason to ascribe the parts a separate telos of their own, as their ends are completely subsumed within that of the extended organism to which they belong. The above conditions are, I believe, sufficient for the existence of a superorganism, and the first three conditions are certainly necessary.

It needs to be stressed that the parts of a superorganism would need to be directed by a master program. Mere co-operation between distinct organisms, as is seen for example in symbiosis, does not suffice to make a superorganism. For example, some insects that feed on plants give sugary secretions to ants for food, and in return they are protected from enemies and sometimes even accepted as virtual members of the colony - a variety of symbiosis known as trophobiosis (Holldobler and Wilson, 1994, p. 143). Nevertheless, the behaviour of the adopted insects, as far as I am aware, is not chemically regulated by the colony which adopts them, so the insects cannot be said to be directed by a master program.

There are, to be sure, some cases of symbiosis where the degree of co-operation is far closer: the host organism acquires new genes from the symbiont by a process of lateral transfer, and the symbiont becomes so integrated into the host that it is effectively a part of the host cell (an organelle). In such cases, a genuinely new organism emerges, but as the parts are spatially contiguous (one is subsumed within the other), it cannot technically be described as a superorganism, where the parts are physically separated.

The "unity" and "individuation" problems can now be addressed. The unity of the superorganism is guaranteed by its master program, which controls all the parts. Two superorganisms are individuated by their possession of different master programs. To return to the ant colony analogy: each colony has its own distinctive pheromones, by which it discriminates between nestmates and strangers. Within a colony, pheromones comprehensively regulate the activities of the colony: they are used to signal the presence of food, recruit help when an ant is in danger, identify other castes, inhibit the laying of eggs by the queen's daughters, and fix the percentage of larvae that grow up to be soldiers - all for the benefit of the colony (Holldobler and Wilson, 1994, p. 55).

The idea that a superorganismic animal could have parts that functioned in different locations may seem counter-intuitive, but Varner has suggested a helpful analogy:

[S]uppose that, instead of being connected to my brain by long networks of nerves, the muscles in my hands were operated by a kind of natural radio signal. Then I could detach my arms and go down the hall to check my mailbox without leaving off typing (assuming, of course, that I could remove my arms without bleeding to death and that I am a touch typist!) (1998, p. 75).

Varner proposes a very inclusive criterion for individuation in asexually reproducing organisms: he suggests that "clonal reproduction never results in more than one individual, it just results in the one individual having noncontiguous parts" (1998, p. 74). He cites the example of an aspen grove, whose trees, although they appear distinct, are actually connected by the roots, underground. The grove is certainly one tree, and it is fundamentally no different from a live oak that splits into several branches just above the ground (p. 75). Next, he argues that if we sever the connections between the roots below the ground, we still have one tree, and finally, he avers that even if we remove one of the severed "parts" and plant it far away, it makes no ontological difference. I find this example unpersuasive. For my part, I can see no reason to call the removed tree (as I prefer to call it) a part of a larger whole, if its behaviour is no longer regulated by the whole to which it formerly belonged, and if it can survive and thrive without the whole.

Lower-level formal conditions for life

Possession of a boundary

What sort of concrete material realisation does a body need in order to qualify as being alive? Nicholas Humphrey (1993, p. 194) ties the possibility of consciousness to the possession of an intrinsic boundary by an individual. Without such a boundary, argues Humphrey, it could not feel anything happening to it and hence would have nothing to be conscious of. I would go further and suggest that nothing can be said to be alive without an intrinsic boundary.

Compartmentalisation, or the existence of an intrinsic boundary is Koshland's third "pillar of life". There are two reasons why having a boundary is important. First, an organism that lacked a boundary could not distinguish itself from other individuals, and would thereby be rendered incapable of advancing or defending its own intrinsic ends. For instance, it could not "defend itself against injury" (Koshland, 2002, p. 2215). Second, an organism needs a boundary to contain whatever parts and systems it needs to maintain itself. Koshland spells out the rationale in detail:

All the organisms that we consider living are confined to a limited volume, surrounded by a surface that we call a membrane or skin that keeps the ingredients in a defined volume and keeps deleterious chemicals - toxic or diluting - on the outside. Moreover, as organisms become large, they are divided into smaller compartments, which we call cells (or organs, that is, groups of cells), in order to centralize and specialize certain functions within the larger organism. The reason for compartmentalization is that life depends on the reaction kinetics of its ingredients, the substrates and catalysts (enzymes) of the living system. Those kinetics depend on the concentrations of the ingredients. Simple dilution of the contents of a cell kills it because of the decrease in concentration of the contents, even though all the chemicals remain as active as before dilution. So a container is essential to maintain the concentrations and arrangement of the interior of the living organism and to provide protection from the outside (2002, pp. 2215-2216).

While endorsing the thrust of Koshland's arguments, I should point out that he is assuming that living things are held together by chemical reactions. I would urge caution here: we cannot exclude the possibility that there may be an organism whose internal organisation is maintained by the transmission of, say, electromagnetic signals between its components. Such an organism could survive as long as its parts were suitably configured for sending and receiving signals. Indeed, in different environments, the organism might change its underlying hardware completely, taking in all kinds of "nutrients" whose sole common property was that they were suitable for being configured into the electrical circuitry required to maintain the organism. Tipler's (1982)example of the von Neumann probe, described in section 1.A.3.6 above, could operate in such a fashion, as it would be capable of taking in raw materials from the planet it landed on, and replicating itself. If such a machine could be constructed in such a way that its parts had internal relations, a nested hierarchy of organisation and dedicated functionality (i.e. all the requirements for intrinsic finality) then it could have multiple chemical "realisations", but would still qualify as being alive. However, even an organism constructed in this way would still need an internal boundary of some sort, to protect itself from the elements and to contain its vital circuitry.

The boundary condition rules out certain speculative proposals for what kinds of things could be described as organisms. For instance, gas clouds and light beams lack boundaries and therefore cannot qualify as organisms.


Koshland's seventh pillar, seclusion, or some way of preventing one set of chemical reactions from interfering with another, within a cell, is guaranteed in terrestrial organisms by specific enzymes that work only on the molecules for which they were designed. One might think that only a cellular life-form or an organism that was based on specific chemicals would have such a need, but Koshland offers an analogy which could equally apply to hypothetical organisms (discussed above) whose parts communicated by electromagnetic rather than chemical signals. He likens seclusion to insulating an electrically conducting wire so that it is not short-circuited by contact with another wire. The gist of the argument seems to be that organisms, having a large number of parts and an even larger number of interactions, require a certain degree of regulation to ensure that the interactions work properly - or, in Aristotelian terms, to ensure that organisms can achieve their intrinsic ends. The precise workings of this internal regulation would depend on the formal properties of the organism.

Cellular structure?

So far, the only terrestrial organisms we have examined have been cellular life-forms. Opinion is divided on whether acellular life-forms such as viruses should be considered to be alive. I discuss this question below. I conclude that despite the inadequate rationale ("They reproduce") offered by most scientists who believe that viruses are true life forms, there are sound Aristotelian reasons for viewing them as bona fide organisms. They appear to satisfy the formal, final and material requirements for being alive, as well as instantiating most of Koshland's "seven pillars of life". I also argue that although a virus is a life-form, it is only alive in a secondary sense of the word: it participates in the life of its host, to use a Platonic metaphor.

Opinion is divided on whether acellular life-forms such as viruses should be considered to be alive. On the account that I have defended, certain arguments against viruses being alive - the fact that viruses do not respire, display irritability, move, grow or excrete - are irrelevant, as these features are not constitutive of life as such. However, if one looks at the arguments advanced by defenders of the status of viruses, they tend to focus on viruses' ability to replicate, which, I have argued, is at most a necessary criterion for life, not a sufficient one.

Edward Rybicki, a professor of virology (University of Cape Town, 1998) rejects traditional textbook definitions of life, based on what he refers to as the "classical" properties of living organisms: reproduction, nutrition, respiration, irritability, movement, growth and excretion). He objects that such definitions are biased in favour of animals and plants, and argues that "the only real criterion for life is: [t]he ability to replicate".

Rybicki quotes two definitions of life to justify his characterisation of viruses as living things:

An organism is the unit element of a continuous lineage with an individual evolutionary history (Quoted by Rybicki from Luria S. E., Darnell J. E., Baltimore D. and Campbell A. 1978. General Virology, 3rd Edn. New York: John Wiley & Sons. p.4).

In other words,

an organism is merely the current slice in a continuous lineage; the individual evolutionary history denotes the independence of the organism over time (Rybicki, 1998).

According to this definition, viruses qualify as alive because they replicate, have an evolutionary history and do not depend on any particular host (or even species of host) in order to replicate. However, by the same token, a computer virus would also qualify as being alive.

Another definition, quoted by Rybicki, suggests that viruses are temporally non-contiguous organisms, which are alive when their inbuilt program, contained within their nucleic acid, is activated, and dead while the program is inactive:

Life can be viewed as a complex set of processes resulting from the actuation of the instructions encoded in nucleic acids. In the nucleic acid of living cells these are actuated all the time; in contrast, in a virus they are actuated only when the viral nucleic acid, upon entering a host cell, causes the synthesis of virus-specific proteins. Viruses are thus alive when they replicate in cells, while outside cells viral particles are metabolically inert and are no more alive than fragments of DNA (quoted by Rybicki from Dulbecco R. and Ginsberg H.S. 1980. Virology, pp.854-855).

This definition sharpens the rationale for regarding a virus as a living organism: a virus possesses a program of sorts, which is activated when it invades a host. This sounds like Aristotle's formal cause. A virus may even be said to be "designed" in a way to enable it to easily penetrate its host, take over the host cells and force them to make more copies of itself - in other words, it appears to possess intrinsic finality. In addition, a virus, unlike its computer analogue, possesses a material substrate: the nucleic acid strand of which it is composed.

It may be objected that the simple structure of a virus precludes it from possessing a nested hierarchy of organisation, one of the defining properties of living things. However, from the following description by Spencer, Nibert and Sgro (1994), it would appear that the proteins enclosing the nucleic acid in a virus interact in a co-ordinated fashion to promote the well-being and replication of the "whole" to which they belong:

A virus is a submicroscopic parasite that must infect a host cell in order to replicate, i.e. make copies of itself. The genetic information - the viral genome - is encoded by nucleic acid, either DNA or RNA. The genome is enclosed in one or more layers of protein and, if the virus is enveloped, lipid as well. In many cases, the protein layers are highly symmetrical and are composed of many copies of a few viral proteins, arranged in shells or "capsids," in which repetitious protein-protein interactions are found....

Each virus encodes its own collection of viral proteins, acting in concert, each with specific roles that enable or enhance viral replication. The function of each viral protein is inherent in its tertiary structure (three-dimensional conformation). Some viral proteins function as components of the virus capsid. Others act as enzyme catalysts of chemical reactions that are essential to viral replication, such as RNA synthesis or proteolysis. Some viral proteins may actually do both, participating in forming the capsid and acting as an enzyme catalyst (italics mine).

From the foregoing description, it seems unreasonable to deny that viruses are genuinely alive, as they appear to satisfy the formal, final and material requirements for being alive, as well as instantiating most of Koshland's seven pillars: a program (DNA or RNA), improvisation (they mutate), compartmentalisation (a coating), energy (supplied by the host cell), regeneration (they reproduce), and adaptability (they can hide inside their host until conditions become more favourable). (The satisfaction of the seclusion criterion is more doubtful.) Despite the inadequate rationale ("They reproduce") offered by most people who believe that viruses are true life forms, there are sound Aristotelian reasons for viewing them as bona fide organisms, unlike, say, computer viruses.

This does not mean that we can put viruses on a par with other living things. The life they have is a borrowed one, as they rely on their host to provide all of their metabolic functions and raw materials:

Without a host cell, viruses cannot carry out their life-sustaining functions or reproduce. They cannot synthesize proteins, because they lack ribosomes and must use the ribosomes of their host cells to translate viral messenger RNA into viral proteins. Viruses cannot generate or store energy in the form of adenosine triphosphate (ATP), but have to derive their energy, and all other metabolic functions, from the host cell. They also parasitize the cell for basic building materials, such as amino acids, nucleotides, and lipids (fats) (Davidson, M. and Florida State University, 2002).

Using Platonic terminology, we may say that a virus participates in the life of its host, in which it "lives, and moves and has its being" (Acts 17:28, originally from Epimenides' poem, Cretica, in honour of Zeus). In other words, a virus is a life-form, but only in a secondary sense of the word.

Some viruses, according to Rybicki (1998), are associated with satellite viruses: for instance, the tobacco necrosis satellite virus, which depends for its replication on the presence of the tobacco necrosis virus. These satellite viruses could be described as "third-tier" or tertiary life-forms.

Efficient causal conditions for life


The extraction of chemical energy from nutrients, or metabolism is Sarver's (1999) fifth condition for life. As we have seen, viruses do not metabolise: they derive their energy, and all other metabolic functions, from their host cell. However, this arrangement only serves to highlight the fact that life on earth would not possible without the extraction by some organisms of chemical energy from nutrients. However, it would be unwise to exclude the possibility of a form of life elsewhere in the universe which did not require chemical energy but was able to use some other form of energy.

There is, however, a more fundamental reason for regarding some form of metabolism as a necessary feature of life. As we saw in section 1.A.3.1, the biological property of nutrition, defined as an organism's "ability to take disorganized material and spontaneously organize it" into its own structure (Wolfram, 2002, p. 824), can be regarded as not only an efficient causal condition for life, but also a thermodynamic condition, insofar as the exercise of this function results in a local entropy decrease. Because the laws of thermodynamics apply everywhere in the cosmos, we can assume that no life-form could thrive anywhere without the ability to incorporate new material into itself.

Interaction with the environment

Interaction with the environment is Sarver's (1999) seventh condition for life. Since it is a law of nature that the entropy of any closed system increases over time, the conclusion that living things necessarily require energy to combat entropy seems inescapable. To obtain this energy, they must therefore interact with their surroundings. Without this interaction, any organism would rapidly perish and be unable to realise any of its intrinsic ends.

Ecology can be defined as the study of the interaction between organisms and their environment.


According to Winder's (1993) definition of life, a living thing must be able to live independently of other organisms. This requirement appears empirically dubious: there are some forms of symbiosis and parasitism in which one organism is heavily dependent upon another. However, Winder might argue that the dependent organism could always find another host, and therefore does not need the one in which it currently resides.

Animal embryos constitute another exception to Winder's requirement. Winder might reply that an embryo can be viewed as part of its mother. But according to each of the three formal conditions I have developed for being alive listed above, an animal embryo, although totally dependent on its mother, has to be considered as a distinct organism, since it possesses its own master program, nested hierarchy of organisation and embedded functionality, giving it a telos of its own, which may at times even conflict with that of its mother (e.g. when it competes with its mother's body for access to nutrients). (An animal embryo is a "self-assembling" organism, in the sense that although it cannot develop properly outside its mother's body, it already possesses a complete internal program for putting itself together. No extra "assembly instructions" are required once it has inherited its full complement of genetic material from its parents.)

From a philosophical perspective, even the notion that an entity A might have a good of its own, while remaining dependent throughout its life on another entity B for its subsistence, appears to make perfect sense. The worry that A's ends would be wholly subsumed within B's is misplaced: this would occur only if A's functionality were dedicated to supporting that of B, in the way described in section 1.B.2. In any case, competition between A and B for access for resources (food or oxygen) would suffice to establish a disparity between A's and B's ends.

I conclude that there are no good reasons for regarding independence as a necessary feature of life.

Temporal conditions for life

Temporal contiguity?

Perhaps the most fundamental question relating to the temporal conditions for life is whether an organism needs to be temporally contiguous. Could there be an organism which lived for an interval of time, only to have its life stopped and subsequently re-started again when conditions were favourable? The answer appears to be in the affirmative. Close approximations can be found in the real world - for instance, bacterial spores, a resting phase displayed by some types of bacteria in response to adverse environmental conditions. Certainly, philosophical problems of identity would arise if the organism were to completely disintegrate and subsequently be re-constituted. In that case, it could be argued that it was no longer the same organism. However, if the information that constitutes the organism's "formal cause" is preserved within the "body" (or "matter") of the organism, even while its life-functions have ceased, then the issue of identity over time becomes fairly unproblematic. In Aristotelian terms, what seems to be required for identity is that the form be preserved over time within the organism's matter, even when every organismic function (ergon) has been switched off.

Thermodynamic requirements: energy, reproduction and a life cycle

The fourth and fifth "pillars of life" listed by Koshland (2002, pp. 2215-2216) are energy (to keep living systems metabolising) and regeneration (including reproduction) to compensate for wear and tear on the system. Since the intrinsic end or "purpose" of these features is to hold back the inexorable march of time, they may be considered as temporal conditions for life - more specifically, as thermodynamic requirements.

As Wolfram (2002, pp. 824, 1178) correctly observes, these conditions are not unique to living things. The question we have to address is: are they necessary conditions for life? Given that all living things require some degree of functionality in order to achieve their intrinsic ends, Koshland's energy requirement - that a living thing must be a thermodynamically open system - seems indispensable, given the laws of nature which obtain in our universe. The physical necessity of regeneration (given the laws of thermodynamics) also seems inescapable, to counter what Koshland calls the thermodynamic losses of a metabolising system.

Although he classifies it as an aspect of regeneration, Koshland also considers reproduction to be a necessary feature of living systems. (As we saw earlier, reproduction cannot serve as a sufficient condition to define life, as it is found in a host of systems, including abstract computational systems, that "bear no other resemblance to ordinary living systems" (Wolfram, 2002, p. 824).)

Koshland is careful to point out that the capacity to reproduce cannot be considered a defining property of living individuals (for then, as he remarks, two rabbits - a male and a female - would be alive but one, by itself, would be dead), but argues that reproduction is a necessity for any kind of living system, for thermodynamic reasons. According to Koshland, reproduction is necessary to counter the accumulation of slight imperfections in the constant resynthesis of bodily constituents during an individual's lifetime (in other words, ageing). Reproduction gives a living system the opportunity to start over.

One might object that for all we know, there might be some open system on some planet that is capable of regenerating itself almost perfectly, over a very long period of time - say, a million or a billion years. Such a system might exhibit the finality, form and functionality of a living individual, even without the capacity to reproduce - and then, just die out. But this objection begs the question of how and whether such a system could arise in the first place.

Moreover, there are two other weighty reasons for considering reproduction seriously as a necessary feature of life. The first is the fact that every kind of living thing with which we are familiar, reproduces. As Aristotle remarks (De Anima 2.4, 415a28-29):

For this is the most natural of the functions of such living creatures as are complete and not mutilated and do not have spontaneous generation, namely to make another living thing like themselves, an animal an animal, a plant a plant ... (1986, p. 165).

Leaving aside Aristotle's quaint belief in spontaneous generation, his point remains valid. Reproduction is found in every species of living thing on earth. In Aristotelian terminology, we might say that an organism's capacity for reproduction is a "proper accident" of its soul - a necessary by-product of the more fundamental property of being alive. (However, Aristotle (De Anima 2.4, 415a24-25) also classifies nutrition as a function of the "first and most general faculty of the soul, in virtue of which all creatures have life" (1986, p. 165). We therefore have to consider at least the theoretical possibility that there may exist some life-form possessing the faculty of nutrition without reproduction.)

The second reason for regarding the ability to reproduce as a hallmark of life is the Darwinian or historical paradigm of biology: living things cannot be considered apart from their genes, and the reason why living things have the genes that they possess is that those genes have out-competed other genes in a four billion-year evolutionary race to replicate themselves. If all life-forms are the product of natural selection, then reproduction is an essential characteristic of all species. (I say "species" rather than individuals, because in many species, there are biologically normal individuals whose morphological type - which I shall discuss in section 1.B.7 - renders them incapable of reproducing.)

It should, however, be borne in mind that we know of only one planet so far which supports life, so it seems unwise to generalise about any life-forms on other planets. In the meantime, we cannot dispense with reproduction as a necessary condition of life until we discover some mechanism whereby a life-form could originate on a planet, without some process of self-replication being involved. (Interestingly, a new theory by Martin and Russell (2003) suggests that cells, whose walls were originally made of iron sulphide deposited by hot springs, originated first and served as incubators for organic molecules, which eventually acquired the ability to self-replicate. In other words, the cell may be a more basic feature of life than reproduction.)

Sarver's (1999) sixth defining feature of life is the occurrence of characteristic phases of development in a life cycle. A life cycle of some sort seems unavoidable if we accept that reproduction is a necessary feature of life. Another reason for regarding a life cycle as a sine qua non of life in our universe is the fact that living things inevitably age and wear down.

Evolutionary requirements for life: ability to evolve

Koshland's second pillar, improvisation, or a way in which an organism can change its master program (achieved on Earth through mutation), appears to be a necessary condition for life in a universe where environmental change is the norm. Without such a mechanism, Koshland argues, species would rapidly die out. It would of course be a mistake to say that evolutionary change occurs "for the sake of" the species; Darwinian evolution is a random process which does not occur "for the sake of" anything. We may, however, legitimately say that a capacity to evolve enables lineages to survive: without it, they would soon perish.

This leaves open the question of whether non-evolving life might have originated on some other planet and rapidly died out - perhaps because its replication mechanisms were too perfect to allow for the occasional genetic copying error. There is, however, another reason for regarding the ability to evolve as a necessary feature of any life-forms that emerge through natural processes. The high degree of organisation of even the simplest organism makes its emergence by purely random processes (i.e. molecules fortuitously coalescing together) vanishingly improbable. The only natural alternative is a non-random one, and the most obvious candidate is the non-random winnowing mechanism of natural selection. In other words, some kind of (abiotic) chemical evolution, constrained by natural selection, generated the first structures that fulfilled all the criteria for being "alive". (Another possibility is that there are as-yet-unknown laws of the universe which account for the emergence and self-organising properties of life (Kauffman, 1993, 2000).)

But even if life can only emerge naturally as a result of natural selection, there seems to be nothing to prevent human beings from manufacturing artificial life-forms that possess the vital property of intrinsic finality (and the attendant formal properties), but lack the capacity to evolve. On my account, these structures would have to be considered alive.

Even if the capacity to evolve should prove to be a necessary temporal feature of life in our cosmos, but its necessity would only be knowable a posteriori. It would therefore be very unwise to build the concept of evolution into the very definition of life.


Koshland's sixth pillar (behavioural adaptability to environmental hazards) seems unexceptionable, if rather vague. Koshland elaborates:

behavioral manifestations of adaptability are a development of feedback and feedforward responses at the molecular level and are responses of living systems that allow survival in quickly changing environments (2002, p. 2216).

The obvious qualification is that adaptability, like rapidity of environmental change, is a matter of degree. Some organisms live in very stable environments; others, like ourselves, are in a continual process of adjustment.