Back to Chapter one References

Chapter 1 - What does it mean to be alive?

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

1.B A "science-friendly" teleological account of life

1.B.0 My account of life, in a nutshell
1.B.1 What distinguishes living things?
1.B.2 Empirical formal and finalistic criteria for being alive
1.B.3 Is my account of life reductionistic?
1.B.4 Strong emergence, downward and backward causation
1.B.5 Is the possession of intrinsic ends a sufficient and necessary condition for being alive?
1.B.6 Does my account of life explain and unify all of the necessary conditions for something's being alive?
1.B.7 What does it mean for a living thing to have a nature?
1.B.8 Do living things have a privileged ethical status?

1.B.0 My account of life, in a nutshell

The account of life which I am proposing here can be summarised in the following eight points (amplified in sections 1.B.1 to 1.B.8):

(i) two fundamental but inter-related causal features serve to distinguish living things from non-living things: their unique formal properties and their possession of intrinsic final ends. Additionally, I argue that on any "science-friendly" account of life, the formal features that characterise living things must be inseparable from the final features (intrinsic ends) which make them alive;

(ii) the formal and final properties which define life are at the same time empirical criteria, which have already been mentioned in the scientific literature. In other words, my account offers scientists a way of ascertaining which things are alive. The account I propose also explains why the association between the formal and final features that characterise life is not a contingent one: necessarily, any material entity instantiating the relevant formal features also possesses intrinsic final ends, and vice versa;

(iii) final causality supervenes upon other kinds of causality (especially formal causality), but the property of being directed towards an end cannot be reductively identified with any lower-level property, as the formal features of organisms cannot be adequately specified or described using non-teleological terminology. However, other, weaker forms of reduction are compatible with a teleological account of life;

(iv) teleological properties of organisms are strongly ontologically emergent properties, which enable organisms to instantiate both downward and backward causation. However, these forms of causation supervene upon micro-level, forward causal properties;

(v) the distinction between having intrinsic ends and merely extrinsic ends is a real one, which allows us to clearly distinguish organisms from contemporary artifacts, which are in no sense "alive". However, there is nothing to prevent future scientists from constructing an artifact that possessed genuine intrinsic ends. Such an artifact would be alive;

(vi) a teleological account that equates "being alive" with possessing intrinsic ends is a complete, all-inclusive account, insofar as it explains the unity of "life" as a natural category (thereby solving what Cameron (2000, p. 50) refers to as the problem of unity, and provides a unified theory of all of the necessary and sufficient conditions for being alive - i.e. the (efficient, material, formal and final) causal conditions for an entity's being alive, as well as their temporal (i.e. thermodynamic and evolutionary) features;

(vii) the teleological account of life that I am proposing also allows us to formulate a robust concept of the nature of a living organism, in a way that is fully compatible with Darwinism;

(viii) finally, the teleological account of life being defended here is able to generate biological and moral norms.

The account I am defending here has strong affinities with Cameron's (2000) teleological account of life, in which the property of being alive is defined as the possession of intrinsic ends. Cameron considers intrinsic finality to be both an ontologically primitive property and a strongly emergent one: its (final) causal features cannot be reductively identified with any properties of its structural components.

Unlike Cameron, I claim that we can specify formal criteria that are sufficient conditions for an entity's being alive. Scientists can thus ascertain whether something is alive by identifying these criteria. Cameron regards any attempt to specify non-teleological criteria which would enable scientists to ascertain whether an entity is alive, as tantamount to equating the property of having an intrinsic end with the properties that define the criteria for life (reductionism). I think Cameron is mistaken here. I also argue that defining life purely in terms of teleological criteria is scientifically unhelpful, as it does not enable scientists to determine which things are alive and which are not.

I also maintain that final causality supervenes upon the lower-level (efficient causal, formal and material) properties supporting it. Cameron rejects the word "supervenience", but I argue below that the difference between us is purely definitional.

While Cameron can justly claim that his account solves the problem of unity which he describes (2000, p. 50), I have gone one step further with my teleological account: in section 1.B.6 and the Appendix to part B, I attempt to show how intrinsic finality can unify all of the (efficient causal, formal, material and temporal) criteria for being alive, thereby satisfying the unity-of-criteria condition for life. Insofar as it recognises all these kinds of criteria, my account of life is "five-dimensional".

I also argue that the formal criteria I propose for intrinsic finality allow us to re-interpret the Aristotelian concept of nature in a way which is fully compatible with Darwinism.

Finally, the concept of an individual as having a certain nature can also generate biological norms regarding what it should and should not do, as well as moral norms for how we should and should not treat it.

The account of life I propose here attempts to define "life" only in the biological sense of the word. (Cameron (2000, p. 333 ff.) defends a general definition of "life", covering the various senses in which the term is applied to spontaneously generated organisms, plants, animals, humans and God.)

1.B.1 What distinguishes living things?

My first proposal, that form and finality distinguish living things, has strong Aristotelian roots. Aristotle (De Anima 2.1, 412a20-22, 28) defined the soul as "the form of a natural body which potentially has life, and since this substance is actuality, soul will be the actuality of such a body... [S]oul is the first actuality of a natural body which potentially has life" (1986, p. 157). Elsewhere (De Anima 2.4, 415b15, 20) he wrote that the soul was also the final cause, "that for whose sake" a living organism moves (1986, p. 165).

Specifically, I claim that:

(i) intrinsic finality is a unique feature of living things; and

(ii) intrinsic finality goes hand-in-hand with unique formal features which, characterise living things: i.e. a material object is alive and has intrinsic ends if, and only if, it possesses these formal features.

However, as I explain in section 1.B.2, the formal features cannot be understood without reference to finalistic concepts, which is why a purely formal characterisation of life could never succeed.

Intrinsic ends

My first claim is not new. For instance, Cameron (2000, p. 333) defends a simple definition of life: to be alive is to possess intrinsic ends. (Cameron also identifies a secondary, derived sense in which an organ may be said to be alive, by being naturally connected with a being that possesses intrinsic ends.) Indeed, Cameron argues that this was Aristotle's original position - a contentious claim, which he defends vigorously and convincingly (2000, pp. 86-135) against an array of scholars who have construed Aristotle as holding that not only biological processes, but also necessary and regular occurrences in the inanimate world required final causes, or that all things possessing a nature had a final cause. Finally, Cameron argues (2000, pp. 327-335) that Aristotle was in fact well aware of the distinction between intrinsic and derived ends, insofar as he taught that the body parts of an organism possess ends, but only in a secondary sense which derives its intelligibility from the nature and finality of the organism as a whole.

Form and finality are inseparable

My second claim is a robust one. I claim that any material object instantiating all of the requisite formal features for being alive cannot fail to possess intrinsic ends, and that no material object can be said to have intrinsic ends unless it satisfies all of the required formal conditions. Although I claim that an entity's having the right formal features can tell us that it has intrinsic ends, I do not claim that we can determine precisely what an organism's specific ends are, simply by examining its formal features. For instance, the various functions of a walrus' tusks (protection against predators; acquisition of mating partners; getting a grip when climbing out of the water onto ice; and anchoring themselves on the ocean bottom while digging for clams) could never be guessed by someone who was unfamiliar with their local environment.

The motivation for my second claim, that the life-conferring property of possessing intrinsic ends is inseparable from the unique formal properties of living things, is scientific. The first claim does not tell scientists how to identify living things. Only if the formal features of a material object determine whether it has an intrinsic end can scientists ascertain which things are alive.

My second claim is a controversial one, which Cameron (2000, pp. 145, 328) explicitly rejects. While granting the "coextensiveness in normal cases" of organisms' formal and final causal properties, he considers them separable in principle:

The formal causation of the growth of an organism is compatible with the absence of final causal influence; formal causation is therefore not sufficient for teleological causation (2000, p. 145).

I will henceforth refer to this claim as Cameron's "form-without-finality" thesis.

Scientific reasons for rejecting Cameron's form-without-finality thesis

If correct, Cameron's "form-without-finality" thesis calls into question the whole enterprise of identifying something - a new species, or a purported life-form on another planet, for instance - as alive. If the formal features of an object do not determine whether it has an intrinsic end (i.e. is alive), how are we supposed to tell whether it has one? Cameron could perhaps argue that extended observation of the entity over a period of time is required to ascertain whether it possessed teleological features such as a life cycle, reproduction and so on - although this seems to be at odds with his assertion elsewhere (2000, pp. 198-200, 207-208) that an instant duplicate of me would be alive, or that we could readily identify the function of eyes and hearts even in instantaneously created lions that suddenly popped into existence. But the real difficulty with this "temporal solution" is that in practice, biologists do not work like this: even though they may require an extended period of observation to determine what an organism's ends are (including its process of maturation), they do not stop there, but proceed to look for formal features of the organism (e.g. body clocks) that regulate these biological end-oriented processes.

Can an entity's formal features determine whether it has intrinsic ends?

Cameron puts forward three different reasons for rejecting the view that an entity's formal properties alone can guarantee that it has intrinsic ends and is alive. First, he argues (2000, p. 146) that "[t]he mere fact that a good is forever associated with a process is no indication that the process occurs for the sake of the good". Logically, this is unassailable, but it is irrelevant to my own proposal, which is that a set of (lower-level) processes whose very nature enables their subject to benefit in some way (on a holistic level) can legitimately be described as being "for the good of" their subject.

Second, Cameron argues that complex artifacts such as robots, which are not alive, nevertheless display "the same sorts of material and structural complexity that organisms and living things display" (2000, p. 328). Following Paul Taylor (1986), Cameron distinguishes between the derived ends of complex artifacts and the intrinsic ends of biological organisms.

I do not wish to question Taylor's distinction between intrinsic and derived ends. However, I would query Cameron's claim that contemporary artifacts possess the same formal features as organisms. I describe the formal features of organisms in section 1.B.2, and I argue in section 1.B.5 that those of contemporary artifacts are quite different. I contend that if scientists of the future constructed artifacts with the same "holistic" formal features as organisms, they would possess intrinsic ends and hence be alive.

A third argument of Cameron's against the view that the distinctive formal features of organisms are inseparable from their possession of intrinsic ends is that spontaneously formed inanimate objects exist in nature which instantiate the formal structures (and behaviour) of living organisms, while lacking their ends:

[F]or any sort of material structure that is claimed to ground teleology in biological systems, a materially identical counterpart can be found in systems which appear not to be teleological (2000, p. 154).

For instance, a wide range of mechanical systems possess feedback mechanisms which enable them to maintain a particular state, in the face of potentially destabilising environmental changes (2000, p. 174). Yet there is a vast gulf between the self-regulatory behaviour of these systems and the goal-directedness of organisms.

While I agree with Cameron that self-regulatory structures and/or behaviour per se do not confer intrinsic ends on an entity, I argue in section 1.B.2 that it is not the stability of a structure, but those formal features that allow its parts to be dedicated to the support of the whole they comprise, that make it teleologically directed. While I concede that the structures discussed by Cameron lack ends, I claim that other structures do possess them as an in-built feature.

Inconsistencies in Cameron's position

Cameron (2000) appears to contradict his "form-without-finality thesis" elsewhere in his work.

In his critique of the etiological account of functions, Cameron (2000, p. 185) approvingly cites the following remark by Richard Dawkins:

[A]ny engineer can recognize an object that has been designed, even poorly designed, for a purpose, and he can usually work out what that purpose is just by looking at the structure of the object (1986, p. 21, italics mine).

Dawkins later goes on to argue that Nature herself can be regarded as a "blind watchmaker". In any case, his point that the function of an object can be deduced from its structure is clearly at odds with Cameron's assertion (2000, pp. 145, 328) that an object could instantiate all of the structural features of an organism while lacking ends.

Additionally, Cameron's "form-without-finality" thesis leads to paradoxical conclusions. First, it would imply that there could exist a non-living duplicate of me, possessing my structural features yet lacking ends. Cameron himself contradicts this startling corollary of his "form-without-finality" thesis elsewhere in his work: he vigorously asserts (2000, pp. 198, 200) that a duplicate of me - even one created instantaneously - would exhibit "purposive behaviours" and have bona fide functions, making it genuinely alive.

One way in which Cameron might attempt to avoid the charge of inconsistency here would be to argue that the possession of intrinsic ends is determined not by form alone, but by a combination of form and matter. Indeed, Cameron approvingly quotes Aristotle in support of his assertion - with which I agree - that"[b]eing alive depends upon possessing material composition of a very definite kind" (2000, p. 342). However, the reason why material composition is so critical is a formal one - a point recognised by Cameron himself:

Life, indeed appears to be dependent upon the existence of particular structurings of matter in exactly the way indicated for emergence (2000, p. 341).

(For instance, it is idle to speculate about whether a duplicate of me, in which the carbon atoms were replaced with silicon atoms, would be alive, as the structural possibilities of carbon and silicon atoms are very different.)

It appears then, that the distinctive formal features of an organism cannot be separated from its aliveness. But that leaves us with two questions: how do we identify these features empirically, and why do we still need a teleological account of life if formal features enable us to distinguish living from non-living things?

1.B.2 Empirical formal and finalistic criteria for being alive

The empirical criteria which I list for something's being alive are not new ones; each of them is already known to scientists. Indeed, most of them can be found in the long cluster definitions listed above. Specifically, I suggest that the presence of both an Aristotelian formal cause and a final cause in a living body can be ascertained empirically by its possession of the following three biological properties:

The first two features would commonly be regarded as formal features of organisms, but as we shall see, they cannot be adequately characterised without reference to the telos of the organisms possessing them, so actually they could be described as both formal and finalistic; while the attribute of embedded functionality can be viewed as an empirical manifestation of both formal and final causality, insofar as it describes the manner in which the structured parts of an organism subserve the interests of the whole.

My claim, if correct, answers the first question in section 1.A.3.8: how can we tell if something is alive?

Why a master program?

The first of Koshland's "seven pillars of life" is a master program, i.e. "an organized plan that describes both the ingredients themselves and the kinetics of the interactions among ingredients as the living system persists through time" (2002, p. 2215). He goes on to say that for life on Earth, this program is implemented by DNA, which encodes genes, which in turn encode the chemicals that carry out the reactions for living systems. (Strictly speaking, our genome is programmed by the addition of chemical markers called methyl groups to the DNA, which can shut down genes (Motluk, 2004; Cohen, 2003).) (Recent research, reported in "New Scientist" (25 December 2002) has shown that RNA also controls key activities within cells.) I would suggest that the notion of a master program that governs the parts and their interactions is an appropriate way of expressing Aristotle's concept of a formal cause, in contemporary terminology.

The proposal that a living organism's master program is a manifestation of its formal cause is not a new one: Hugh Lawson-Tancred, in a footnote to his Penguin translation of Aristotle's De Anima (1986, p. 238), cites a remark by Professor Max Delbrueck, that "if the Nobel committee were able to award the prize for biology posthumously, they should consider giving it to Aristotle for the discovery of the principle of DNA". Incidentally, I have to disagree with Cameron's description of DNA as "a very specific sort of material cause" (2000, p. 152, italics mine), since it is the formal (structural) features of DNA that enable it to code for the development of an organism and the interactions between its parts. A "molecular program" (2000, p. 153) is not a material cause.

From a philosophical perspective, the significance of the master program is that it confers a unity of form on the organism.

The term "master program" should not be misunderstood here. We are not talking about a program that regulates the entire suite of the organism's behaviour (towards the outside world), or even a program that maintains the organism in a state of homeostasis, but rather a program that controls the organism's internal structure and the internal interactions between its components. This is our core definition of an organism's formal cause, but broader definitions are possible.

(In a broader sense, we might define an organism's formal cause as the entire suite of programs responsible for directing the formation and preservation of an organism's structural features at all levels - whether at the cellular, tissue, organic or organismic level. In the broadest possible sense, the formal cause would include the set of all programs and physical structures within an organism which assist its survival.)

I claim that the possession of a master program is a necessary (but not sufficient) condition of a material object's being a living organism. This is a strong claim, unlike the more modest claim that all organisms possess a program(s) of some sort, which could easily be satisfied if we defined "program" in Wolfram's broad sense (2002, p. 383): all it would mean is that the behaviour of living things is constrained by rules. The requirement for a master program says something more: that the formation and preservation of the organism's structure should be regulated by a single, unified set of instructions (such as those contained in an organism's DNA), even if these instructions are widely dispersed through the organism's body. My reason for advancing this requirement is that an assortment of low-level programs working independently of one another, in the absence of any kind of central co-ordination, would be unable to accomplish their respective tasks smoothly and harmoniously, as they would be liable to interfere with one another. In particular, an assortment of independent programs could not be relied upon to accomplish two vital tasks: first, directing the formation of the organism's bodily structures during its development; and second, preserving the organism as a single, viable entity and co-ordinating its activities.

However, the mere presence of a master program per se does not confer life on an entity, as it fails to explain why the entity should possess ends. I claim that in order to give rise to ends, the master program needs to be a very specific kind of program: one which generates a nested hierarchy of structure within the organism, and which also allows the functionality of the lowest levels to subserve that of the highest levels. Then and only then can we speak of the structure-as-a-whole as having ends of its own.

The word "subserve", when applied to an organism, can refer not only to its part-whole functionality, but also to the way in which developmentally immature states of the organism are directed towards the organism's mature state - a point made by Aristotle, who regarded maturation as one of the major holistic ends of an organism (see references in 1.4.7 below). However, the fact that an organism's master program governs its process of maturation does not mean that we can "reduce" this biological goal to the program that codes for it. Cameron (2000, p. 152-153) rightly criticises accounts that attempt to reduce the property of intrinsic finality to that of having an internal program; I make no such claim. Indeed, as I argue in section 1.B.3, we cannot properly understand the master program regulating an organism's development unless we first grasp what the development is for - a future state (maturation), which is a holistic goal of the organism.

Why a nested hierarchy?

Allen (1996) provides a useful summary of the key concepts in hierarchy theory, a selection of which are reproduced in Table 1.8. Briefly, what distinguishes nested from non-nested hierarchies is that the former involve levels which consist of, and contain, lower levels, while the latter do not. To use one of Allen's examples: an army is a nested hierarchy because it contains the soldiers which make it up, whereas the chain of military command is a non-nested hierarchy: a general does not "contain" the soldiers he/she commands.

A nested hierarchy of organisation is the second feature in Sarver's (1999) definition of life. I believe that this notion can be used to elucidate the meaning of Aristotle's concept of intrinsic finality. More precisely, a nested hierarchy is a necessary (but not sufficient) condition for intrinsic finality, because it is impossible to ascribe ends to an organism unless its parts work together for the benefit of the whole. To do this, the parts need to be hierarchically ordered. Finally, the hierarchy needs to be nested, in order for the parts to be intrinsically ordered towards the well-being of the whole. In a non-nested hierarchy, each part would be directed by some other part above and outside it; thus the part-whole finality would be merely extrinsic.

Table 1.8 - A Summary of the Principles of Hierarchy Theory (taken from Allen, 1996)
The Hierarchy theory is a dialect of general systems theory. It has emerged as part of a movement toward a general science of complexity. Rooted in the work of economist, Herbert Simon, chemist, Ilya Prigogine, and psychologist, Jean Piaget, hierarchy theory focuses upon levels of organization and issues of scale. There is significant emphasis upon the observer in the system.

Hierarchies occur in social systems, biological structures, and in the biological taxonomies. Since scholars and laypersons use hierarchy and hierarchical concepts commonly, it would seem reasonable to have a theory of hierarchies. Hierarchy theory uses a relatively small set of principles to keep track of the complex structure and a behavior of systems with multiple levels. A set of definitions and principles follows immediately:

Hierarchy: in mathematical terms, it is a partially ordered set. In less austere terms, a hierarchy is a collection of parts with ordered asymmetric relationships inside a whole. That is to say, upper levels are above lower levels, and the relationship upwards is asymmetric with the relationships downwards.

Hierarchical levels: levels are populated by entities whose properties characterize the level in question. A given entity may belong to any number of levels, depending on the criteria used to link levels above and below. For example, an individual human being may be a member of the level i) human, ii) primate, iii) organism or iv) host of a parasite, depending on the relationship of the level in question to those above and below.
Level of organization: this type of level fits into its hierarchy by virtue of set of definitions that lock the level in question to those above and below. For example, a biological population level is an aggregate of entities from the organism level of organization, but it is only so by definition. There is no particular scale involved in the population level of organization, in that some organisms are larger than some populations, as in the case of skin parasites.
The ordering of levels: there are several criteria whereby other levels reside above lower levels. These criteria often run in parallel, but sometimes only one or a few of them apply. Upper levels are above lower levels by virtue of: 1) being the context of, 2) offering constraint to, 3) behaving more slowly at a lower frequency than, 4) being populated by entities with greater integrity and higher bond strength than, and 5), containing and being made of - lower levels.
Nested and non-nested hierarchies: nested hierarchies involve levels which consist of, and contain, lower levels. Non-nested hierarchies are more general in that the requirement of containment of lower levels is relaxed. For example, an army consists of a collection of soldiers and is made up of them. Thus an army is a nested hierarchy. On the other hand, the general at the top of a military command does not consist of his soldiers and so the military command is a non-nested hierarchy with regard to the soldiers in the army. Pecking orders and a food chains are also non-nested hierarchies.
Constraint versus possibilities: when one looks at a system there are two separate reasons behind what one sees. First, it is not possible to see something if the parts of the system cannot do what is required of them to achieve the arrangement in the whole. These are the limits of physical possibility. The limits of possibility come from lower levels in the hierarchy. The second entirely separate reason for what one sees is to do with what is allowed by the upper level constraints. An example here would be that mammals have five digits. There is no physical reason for mammals having five digits on their hands and feet, because it comes not from physical limits, but from the constraints of having a mammal heritage. Any number of the digits is possible within the physical limits, but in mammals only five digits are allowed by the biological constraints. Constraints come from above, while the limits as to what is possible come from below. The concept of hierarchy becomes confused unless one makes the distinction between limits from below and limits from above. The distinction between mechanisms below and purposes above turn on the issue of constraint versus possibility. Forget the distinction, and biology becomes pointlessly confused, impossibly complicated chemistry, while chemistry becomes unwieldy physics.

The panexperientialist philosopher Charles Birch employs the notion of an internal relation to explain the individual unity of an organism. An internal relation is "a relation which is constitutive of the character and even the existence of something" whereby the "parts become qualitatively different by being parts of a whole" (2002, p. 6, italics mine). Because the relation between the parts is constitutive of their character, it is appropriate to speak of them as "organised into a hierarchy of compound individuals" (2002, p. 8). For Birch, it is this hierarchy of organisation which allows us to speak of the cell as a unity.

Birch also describes much smaller entities [molecules] as hierarchies of compound individuals (2002, pp. 6-7). Does this make them alive? On my account, the answer is no: molecules do not have master programs of the sort I described above.

We have not yet explained why organisms have intrinsic ends. A program that generates a hierarchically organised structure could be purely static. In an entity with intrinsic ends, the parts need to work for the good of the whole. To guarantee this, we need the added ingredient of embedded functionality.

Why embedded functionality?

The nested hierarchy of living things reflects a functionality in which the entire repertoire of the functionality of the parts is "dedicated" to supporting the functionality of the whole unit which they comprise. (By "the entire repertoire", I mean everything that the parts actually do, not everything they can do. The parts of a unit may have other potential uses: genes and even organelles can be exchanged between organisms, as shown by the phenomena of lateral gene transfer and endosymbiosis, respectively.)

Geer (2002) quotes the following explanation of the concept of embedded functionality by Dr James Tour, the Chao professor of chemistry and computer science at Rice University:

Dr. Tour explains: "[Let's say that] you see a tree [and] you want to make a table, [so] you chop down the tree [and] you make a table - that's [building] top down. But, the tree and I and everything else in nature are built from the bottom up. Molecules have certain embedded interactions between them and embedded functionality. Those come together to form higher-order structures called cells and those form higher-order structures and here we are." You might also envision this as building from the inside out, or by forming the required traits in the smallest conceivable building blocks first (Geer, 2002).

This "dedicated" functionality, the product of four billion years of evolution, can be seen at every level of organisation of a living thing, from the bottom up. Living things are built from the bottom up, by "dedicated", intrinsically adapted parts; today's human-built computers are designed from the top down, out of parts which have to be modified in some way, to suit the designers' ends.

Embedded functionality is by no means unique to living things: as we saw, it is found in molecules too. However, what I am suggesting is that all organisms, and only organisms, possess the combination of a master program that directs the generation of a nested hierarchical structure with the property of embedded functionality.

Entities with these specific formal features are guaranteed to possess intrinsic ends, because the parts work in a way that subserves the good-of-the-whole in a way that is built-in, and not merely accidental.

1.B.3 Is my account of life reductionistic?

What kind of reduction is unacceptable?

As a self-described emergentist, Cameron (2000, pp. 267-271) is happy to countenance a variety of different ways of reducing teleology to lower-level properties, including even the metaphysical version of Kemeny-Oppenheim reduction, which stipulates that only when we have a well-confirmed microcausal theory of base entities that explains their composition of higher-level properties can we say that the latter have been reduced to their base entities. The only kind of reduction which Cameron (2000, p. 271) rejects as incompatible with emergence is reduction by property identity, where being F is identical to instantiating some set of properties G1,....Gn in some relation R. In this sense, the property of temperature (for ideal gases) has been reduced to the mean kinetic energy of the constituent molecules: the two properties are the same.

To avoid a possible charge of reductionism by Cameron against my claim that an organism's instantiation of three formal features guarantee its possession of an intrinsic end, all I have to show is that final causality in an organism cannot be reductively identified with the formal features described in section 1.B.2 or any other non-teleological properties.

Why a reduction of final causality is impossible

The reason why any reductive identification of final ends with any lower-level properties is bound to fail is that the ends possessed by organisms are holistic ends, and no lower-level description of a biological process, however complete it may be, can be equated in meaning with a holistic description of the same process, even if the two descriptions are truth-functionally equivalent. It is for this reason that Mayr (1982) rejects the simplistic assertion by what he calls strong mechanists, that organisms are nothing but chemical machines, whose properties and parts are wholly reductively identifiable with and explicable in terms of the laws of physics and chemistry.

An additional reason why final explanations are not reducible to formal ones is that such a reduction incorrectly assumes that we can describe the formal features of a living organism using non-teleological terminology. But when we are talking about organisms, it is impossible to compartmentalise our language regarding formal and final causality in this way: the two forms of causality are inextricable. We cannot, for instance, properly describe an embedded functionality without reference to its ends. And although an inanimate object could contain a master program or nested hierarchy of organisation in the absence of embedded functionality, even these formal features, when realised in an organism, cannot be properly described without reference to the ends that they enable the organism to achieve.

A variety of holistic ends

I would also like to emphasise that I do not claim that an organism's final cause is nothing more than the part-whole functionality of its bodily organs. As Cameron points out (2000, p. 149), organisms possess not one but a variety of holistic ends, such as reproduction, developing into mature adults, and their own individual flourishing. The way in which an individual's body parts are subordinated to the whole can tell us a lot about the individual's flourishing, but it is clearly inadequate to characterise the processes of maturation and reproduction, both of which involve reference to future states - whether of the organism or its progeny. On the other hand, I do believe that an organism's ability to realise all of these holistic ends supervenes upon a combination of the three formal features listed in section 1.B.2.


Mandik (2004) defines supervenience as follows: "[a] set of properties or facts M supervenes on a set of properties or facts P if and only if there can be no changes or differences in M without there being changes or differences in P". As applied to mental states, the supervenience thesis entails that "No two objects can differ in their mental properties without differing in their physical properties" (Mandik, 2004). I regard final causality as supervening upon the other forms of causality - especially formal causality. It follows that if two individuals have the same formal, material and causal properties, then on my account of final causality, either both of them have a telos (intrinsic end) or neither of them has one.

Cameron appears to reject this assertion: he clearly states that "[e]mergentists do not conceive of downward causation on the model of supervenience" (2000, p. 243). However, a closer look at the passage suggests that Cameron uses the term "supervenience" in a narrower sense than the definition cited above:

As the emergentists use the word, macro-to-micro determination is a causal relation; macro events are related as causes to micro events as effects. Often, however, micro-to-macro determination may be thought of not as a causal relation at all, but as a logical or supervenience relation, as when we say that the dots in a dot-matrix picture determine the picture's qualities... Emergentists do not conceive of downward causation on the model of supervenience, they understand it as an instance of ordinary causal relations between things and events in the world (2000, p. 243).

If I read Cameron rightly, he understands the term "supervenience" to mean non-causal upward determination. This interpretation seems to be confirmed by his assertion that that "claims of ontological emergence [Cameron's own position - V.T.] are even compatible with the claim that the behavior of wholes with emergent properties is wholly determined by the laws governing the behavior of the elements in the basal conditions" (2000, p. 258, italics mine), as well as his remark (2000, p. 270) that even the successful reduction by atomic theory of "water" to H2O is compatible with water's properties (its wetness, boiling point) being emergent properties of H2O molecules. In ordinary parlance, the properties of water supervene upon the properties of its constituent molecules.

I have argued that the teleological properties of organisms are determined by their formal properties. Clearly, on my account, the upward determination is causal: the master program directs the formation of a nested hierarchical structure, in which the entire repertoire of the functionality of the parts is "dedicated" to supporting the functionality of the whole unit which they comprise. The embedded functionality of a living organism qualifies as supervenience, according to Mandik's (2004) definition: no organism can suffer death or injury without changes occurring at a lower level.

Why we look for lower-level explanations

If final causality supervenes upon the other forms of causality, then we can answer the second question in section 1.A.3.8: we look for micro-level, non-finalistic explanations of death and injury because macro-level, finalistic descriptions supervene upon micro-level, non-finalistic descriptions. Of course, these explanations, while more basic, can never tell the whole story of what happens when an organism dies or is injured, as they are incapable of encompassing its built-in holistic ends.

1.B.4 Strong emergence, downward and backward causation

In this section, I argue that final causality - as I have described it in 1.B.2 - is an emergent property in a strong ontological sense, even though it supervenes upon lower-level properties. More specifically, the holistic intrinsic ends of an organism are capable of exerting their own causal influence on both the functioning of its lower-level parts (downward causation) and their future-oriented development (backward causation).

Table 1.9 Some definitions pertaining to strong and weak emergence (Cameron, 2000, pp. 278-279)
The core sense of emergence: A property P of a structure X with components a1, ... an is emergent if and only if (i) P depends causally for its existence on the interactions of a1, ... an in X and (ii) P augments the ontology of the world - P is not "contained in" the properties of a1, ... an in their interactions outside of structures of the same type as X.
Basal conditions: The basal conditions of a complex thing "include just the qualitative, intrinsic properties and relations of the parts, i.e., the properties and relations that these bear in and of themselves, without regard to any other objects, and irrespective of any further consequences of their bearing these properties for the properties of any wholes they might compose" (Healey 1991, 401).
Additivity: A property F is an additive property if it is a property of a complex system S composed of parts and properties P1...Pn, and none of P1...Pn has F.
Novelty*: A property P of a complex physical entity E of type T, where entities of type T possess basal conditions B composed of parts, properties and relations, is novel if and only if (i) P depends causally for its existence on the interactions of, and (ii) P is irreducibly different in kind from the kinds of properties and relations had by the component parts of the basal conditions B as they appear independent of their composing entities of type T; that is, (ii') P cannot be reductively identified with any of the kinds of properties and relations had by the component parts of the basal conditions B as they appear independent of their composing entities of type T.
Mere resultants: A property P of a complex physical entity E of type T, where entities of type T possess basal conditions B composed of parts, properties and relations, is merely resultant if and only if it is additive but not novel.
Weak ontological emergence: A property P of a structure X with components is weakly ontologically emergent if, and only if P is novel*.
Strong ontological emergence: A property P of a structure X with components is strongly ontologically emergent if and only if (i) P is novel* and (ii) P has causal powers which are absent from and their interactions independently of entities of the same type as X.

I argued in section 1.B.3 that lower-level forms of causation in an organism cannot render intelligible either its dedicated part-whole functionality or its future-oriented development towards maturity. We need to explain these processes from the top down - i.e. at a holistic level - because an explanation at lower levels cannot tell us what they are for: they occur for the sake of the good of the organism-as-a-whole and enable it to realise its intrinsic ends. Because a top-level perspective is required to properly explain what is going on, we can speak of the whole as having final (not efficient) causal powers over the parts.

Using Cameron's definitions in Table 1.9, final causality - as I construe the term in section 1.B.2 - remains a novel* property of organisms, because it depends on the interactions between the parts of an organism, but is irreducibly different in kind from the properties of the parts: as I argued above, finality cannot be reduced to some combination of efficient, formal and material causality. Thus it is at least a weakly emergent property. Additionally, it has causal powers which are absent from the parts of an organism, when these parts are not configured as an organism, making it strongly emergent. Since final causality supervenes (in the standard sense of the word) upon other forms of causality (as was shown in 1.B.3), it follows that supervenience is compatible with strong ontological emergence.

A metaphysical objection

The notion that macro-level events such as the good-of-the-whole might somehow determine micro-level processes remains a highly suspect one for many scientists and philosophers, and is often criticised for being "too mysterious". Cameron (2000, pp. 230-232, 240-243) addresses this objection at length, arguing that the same could be said for any causal relation:

No empirically discovered causal relation must pass before the bar of a priori reason's demand that it be made 'transparent' to the mind before we may accept it into our ontology. No causal relation, upwards, downwards, or horizontal is anything but opaque to our reasoning. The invocation of mystery ... is wholly out of place (or rather, time) in a post-Humean context (2000, p. 243).

Downwards causation

Downward causation explains what processes in the parts of an organism are for: promoting the present welfare of the whole. The embedded functionality of the organism (described above) manifests this finalistic feature, which can only be adequately described at the macro-level. Thus the fact that the macro-level causal powers of an organism are determined by the causal powers of its micro-level parts, does not entail that the macro-level causal powers can be explained from below, or that they are exhausted by those of the micro-level parts.

Backwards causation

At the holistic level, it is also quite legitimate to speak of biological processes within the organism as being directed towards its future ends (such as maturation or reproduction), even if the organism's ability to achieve these ends depends on its micro-level structural features (i.e. "supervenes upon" these features, according to the terminology I use). The end controls the process, not only in the counterfactual sense that if it were not there, the process would not occur (for if one accepts supervenience, the absence of an end would mean that the micro-level structural features would have to be different), but also in the stronger sense that the end (maturation) possesses a reality of its own that cannot be reductively identified with the efficient, formal and material causal features of the organism, and it is this reality which renders the process intelligible. In this sense, backward causation is quite real.

1.B.5 Is the possession of intrinsic ends a sufficient and necessary condition for being alive?

Intrinsic versus extrinsic finality: Is the distinction a clearcut one?

The relevance of the distinction between intrinsic and extrinsic finality has been contested by some philosophers. Leahy has argued that the distinction is blurred by human domestication of animals:

Furthermore it is not even clear that the distinction in telos is that marked. What is in the interests of animals is increasingly decided by human beings. A good guide-dog, sheep-dog, circus, zoo or farm animal, thoroughbred or pet siamese is treated on the basis of criteria proper to the different roles imposed upon them by human beings... (1994, p. 46).

The recent introduction of genetically modified organisms, which are expressly created for human ends, may seem to buttress Leahy's case. However, there is nothing to prevent an organism's having both intrinsic and extrinsic ends. My contention is that it is only in virtue of the former that it can be said to be alive.

Another critic of the distinction between intrinsic and extrinsic finality is Varner. Although he recognises the distinction between artifacts and living things as one which matters philosophically and ethically, Varner is leery of resorting to an extrinsic vs. intrinsic dichotomy to ground this distinction. Citing Nagel, he argues (1998, p. 66) that some artifacts can be regarded as "goal-directed systems" with ends that can be specified independently of the goals of their human producers. To illustrate his case, he offers the example of a Patriot missile, uncovered by an alien scientist long after a nuclear holocaust has wiped out all intelligent life on earth. The alien may be able to deduce that Patriot missiles are meant to intercept projectiles, without knowing a thing about late 20th century aerial warfare.

(We can strengthen Varner's case by imagining a 21st century new generation missile, with unforeseen military capabilities, designed and refined not by human beings but by a factory of robots, following the instructions of a computer, after all human life on earth has been wiped out in a nuclear conflict. Let us say that the computer was originally programmed by a long-dead mathematician, not to build a particular missile, but to follow a search algorithm for identifying and comparing possible designs for missiles. In this case, the design is actually "found" by the computer, and its capabilities were not foreseen by the program designer, so its built-in ends cannot be adequately specified in terms of the intentions of the human designer of the original search algorithm.)

A few comments are in order here. First, the Patriot missile does not have an internal master program that regulates its formation and the structure of its internal parts. Instead, what it has is a built-in program that co-ordinates its internal parts, after they have been assembled by some externally directed process. In a very real sense, then, its ends, though built-in, are not intrinsic to it. Living things, by contrast, are self-assembling.

Second, while the missile has a built-in goal, the parts appear to be linked by what Birch (2002) calls external relations. Unlike the biomolecules in a cell, the missile parts do not acquire any new physical properties by virtue of being put together; rather, the missile is an assemblage of pre-existing parts whose properties can be described in isolation from one another, and the workings of the ensemble can be subsequently deduced from an understanding of these external properties alone. To use one of Birch's similes (2002, p. 7), the relation between the parts is like that between the bricks in an office block. (It is likely that the parts of living things mesh together in a way that permits internal relations, for historical reasons: they evolved together, and acquired their features because they had to co-ordinate with each other.)

Third, even if there were internal relations between the missile parts, this would not be a sufficient condition for the missile to possess intrinsic finality. As I argued in section 1.B.2, they would also have to possess a nested hierarchy of organisation, in which the "components are ... organised into a hierarchy of compound individuals" (Birch, 2002, p. 8).

In a living thing, not only do the parts acquire new physical properties when they are organised into a whole, but they are also organised in a nested hierarchy, from macromolecules to organelles to cells to tissues to organs to organisms. It is this hierarchy which allows us to say that the telos of the parts is completely subsumed within that of the whole. Because human-built computers (and Patriot missiles) have no such hierarchy of internal organisation, they are, as Birch remarks, mere aggregates, not individuals.

Fourth, the functionality (or range of activities) of the parts of an entity possessing intrinsic finality needs to be dedicated, in such a way that it can only be understood by reference to the whole. For each layer in the nested hierarchy of a living thing, the repertoire of functions of the parts supports the next highest level of organisation. No such dedicated functionality exists in a Patriot missile, or in today's computers.

Finally, while it may be true that certain goal-directed features of artifacts can be appreciated without a knowledge of what they were designed for, the point remains that these artifacts are not self-directed. In Varner's example, the parts are not designed to maintain the missile, but to enable it to shoot down projectiles. The missile does not "benefit" in any sense from doing so - its mission is a "suicidal" one. Its finality is thus extrinsic, rather than intrinsic. (Likewise, a computer designed by human beings has a telos, but it is extrinsic: it is designed to perform computations.)

Why should intrinsic finality be a sufficient and necessary condition for life?

It has been argued that teleology is a defining feature of life, that the distinction between intrinsic and extrinsic finality is a valid one and that there are well-defined empirical criteria for identifying entities which instantiate intrinsic finality. However, we still need to defend the claim put forward in section 1.B.1, that intrinsic finality is a sufficient condition for being alive. We saw in section 1.A.3.1 that single-attribute definitions of life were vulnerable to counter-examples. Why, it may be asked, should this single-attribute definition be any different? And why should we look for a single defining attribute, anyway?

My answer is that only a single attribute definition of life can be a philosophically satisfying one. As we saw in section 1.A.3.2, the alternative is to accept some variety of cluster definition. However, cluster definitions are inherently unsatisfactory: they are incapable of satisfying the unity-of-criteria condition (that is, they cannot explain why the assorted criteria listed, and no others, are criteria for life) and they fail to solve Cameron's problem of unity (that is, they cannot tell us why the term "life" should apply to all living things, and only to those things). A unified definition of life is therefore vastly preferable to a looser cluster account, which Bedau describes as "a fall-back position that can be justified only after all candidate unified views [of life] have failed" (1996, p. 336).

In section 1.A.3.1, we concluded that of the three kinds of challenges faced by single attribute definitions of life - counter-examples, borderline cases and category challenges - only the first posed a real problem. The issue of whether intrinsic finality is sufficient to define "life" stands or falls on the question of whether there are any things that instantiate this feature and that we would not describe as alive, while the issue of its necessity hangs on the question of whether there are any things we would call alive, but which lack intrinsic finality.

There seem to be no valid counter-examples to the sufficiency of intrinsic finality. Any valid counter-example would either have to be a natural object or an artifact. It has been argued in this section that no contemporary artifact even comes close to instantiating the formal requirements for intrinsic finality (i.e. a master program, a nested hierarchy and embedded functionality). Crystals and flames seem to be the closest non-living natural objects to living things, in behavioural terms. However, neither crystals nor flames satisfy the formal requirements for intrinsic finality, described above.

It is certainly true (as I argue below) that future artifacts could indeed possess the property of intrinsic finality. However, we cannot regard these artifacts as valid counter-examples unless we make the question-begging assumption that anything built by human beings entirely out of non-living materials is inanimate. While it is commonly agreed that existing artifacts are not alive - a fact which can be explained on a teleological account, by their lack of intrinsic finality - there is no general agreement that artifacts constructed from non-living components cannot in principle be alive.

The sufficiency of intrinsic finality appears secure. We now have to ask whether there are any things lacking intrinsic finality, but which we would normally call alive. In fact, there do not appear to be any such cases. Teleonomic behaviour, described by Monod (1971) and Mayr (1982) as a feature of living things, appears to be a universal property of life: it can readily be discerned not only in plants and animals, but also in unicellular (one-celled) organisms. Even viruses possess this property, as I argue below in the Appendix to part B.

I conclude that no valid counter-examples can be adduced against the hypothesis that intrinsic finality - as I have described it - is a necessary and sufficient feature of living things. If (as I argue below in section 1.B.6), the property of intrinsic finality manages to satisfy the unity-of-criteria condition and solve Cameron's problem of unity, then we should regard the necessity and sufficiency of this property as philosophically established: there is no more philosophical "work" that it needs to do, and there is no other rival contender that can do the work it does.

Can artifacts instantiate intrinsic finality?

There is no reason in principle why a computational device could not be designed from the bottom up, out of chemical parts which acquire new physical properties when assembled (like the carbon atoms in a benzene molecule), and which exhibit the properties of a nested hierarchy of organisation and dedicated functionality. Indeed, a report (Geer, 2002) in "Techworthy" computer magazine quotes Dr. James Tour, Chao professor of chemistry and professor of computer science at Rice University, as claiming that the science of nanotechnology permits the possibility of constructing computers in this way.

"Nanotechnology," says Dr. Tour, "allows building from the bottom up... You put within the molecule, you program within the molecule, you embed within the molecule, [a] certain structure that gives you memory function, which gives you switching function. That will allow this molecule to hold onto electrons, to be your memory. That will allow this molecule to be in either one of two possible states, so you have your switch.

"Then you build little groups on the ends of the molecules, that we call alligator clips (conveniently), that will then hook these molecules onto probes. So, you do that all within molecules, but you make 1023 molecules at a time. Then you learn how to self-assemble them... That's bottom up as opposed to top down."

Additionally, the field of organic computing is likely to blur the boundaries between living things and human-built computers in the foreseeable future. Geer's (2002) report also cites research by Dr. Leonard Adelman, who proposed in 1994 that DNA molecules could be used to solve difficult computational problems, and demonstrated the usefulness of DNA in computing. For example, mathematical problems can be solved very rapidly by insinuating mathematical questions into the DNA code and embedding them in the chemical reactions that occur among DNA.

It should be borne in mind, however, that a DNA molecule, by itself, is not "alive" in the sense that I have defined above: although it is the key molecule for life on earth, it does not display the nesting hierarchy observed in living cells, let alone embedded functionality. (The related question of whether viruses, which are not composed of cells, but of DNA wrapped in a protein coating, can be said to be alive will be discussed below.)

Returning to the question of whether computers built by human beings could ever be said to be alive: it is certainly possible that a computational device could one day be designed with a master program regulating the internal parts that were produced and the interactions between them, where the parts of the device were linked by internal rather than external relations, with a nested hierarchy of dedicated functionality, supporting the continued existence of the system. Such a device could then be said to have a good of its own (i.e. intrinsic finality) by virtue of its form, and could therefore be described as alive, if it also possessed a concrete material realisation (i.e. a material cause). It needs hardly be said that such a device would be very different from today's human-built computers.

1.B.6. Does my account of life explain and unify all of the necessary conditions for something's being alive?

1.B.6(a) Does my account of life unify the necessary conditions for something's being alive?

As we saw above, cluster definitions of life fail to satisfy the unity-of-criteria condition, and they fail to solve what Cameron calls the problem of unity (2000, p. 50). In other words, they fail to provide a unified explanation of why some criteria, but not others, are part of the definition of life, and why some entities are classified as alive, while others are excluded. The demand for such an explanation appears to be reasonable, as "life" is supposed to designate a natural category.

Problem of unity

If the Aristotelian account of life being defended here is superior to the cluster accounts we examined above, then it must satisfy these two definitional requirements as well. At first, Aristotle's account appears vulnerable to Cameron's problem of unity, since it defines as alive any being possessing one or more of the following powers, corresponding to various faculties of the principle of life (soul): "thought; perception; motion and rest with respect to place; and further motion with respect to nourishment, decay and growth" (De Anima ii.2 413a20-26). The selection of powers appears arbitrary. One might ask: what makes entities instantiating these powers so special? However, Cameron argues that Aristotle, despite regarding the property of life as multivocal, was able to account for the unity of life in teleological terms: to be alive is to possess intrinsic ends (2000, p. 333). What the different faculties of the soul have in common is that they enable living things to realise their own good.

Now that Aristotle's account has been shown to solve the problem of unity, we can address some of the counter-examples raised earlier against two criteria invoked by Aristotle (De Anima 2.4) as distinguishing features of life - namely, nutrition or reproduction. We saw earlier that these features were unsuitable for formulating single-attribute definitions of life: for instance, even flames can reproduce and crystals can grow. However, inanimate objects that "grow" and "reproduce" cannot meaningfully be said to do so for their own good, as they fail to satisfy the three formal requirements discussed above for having a good of one's own (i.e. for possessing intrinsic ends). Thus nutrition and reproduction, considered in isolation, cannot serve to define life, but when they are re-defined as teleonomic activities occurring in physical entities possessing intrinsic ends, they can be regarded as hallmarks of life.

Unity-of-criteria condition - why only a teleological account will do

Although the problem of unity has been addressed by providing a unifying teleonomic account of the various sorts of activities that occur in different kinds of living things, we are still left with the unity-of-criteria requirement described earlier. What we need is an account that unifies all of the criteria we use to determine whether something is alive or not - i.e., the necessary and sufficient conditions for life. Without some account that encompasses all of these conditions, scientists and philosophers are left once more with an untidy cluster of requirements for life, which suggests to some thinkers (Wolfram, 2002) that the term "life" designates an arbitrary category rather than a natural one.

Before I attempt to provide a unifying account of these conditions along teleological lines, I need to explain why an alternative, non-teleological account would not be viable. The most obvious rival contender would be a formal account, as the form of a living thing confers on it a kind of unity. The problem with a formal unification of the conditions for life is that there is no single formal condition which defines "life". Instead, as we saw earlier, there are three distinct formal criteria which, taken together, constitute sufficient conditions for being alive: a master program that directs the formation of and interactions between the parts; a nested hierarchy of organisation; and embedded functionality. We need to know what these formal requirements for life have in common: why these three, and only these three?

The prospect of developing a unifying account of life in terms of any of the other "dimensions" (material, efficient causal or temporal) appears even more remote. None of these dimensions deals with organisms at the holistic level, as living individuals; instead, scientists employ a reductionist methodology to describe the properties of life along these dimensions.

I conclude that only a teleological account offers a reasonable prospect of being able to encompass all of the criteria we use to identify something as being "alive".

The teleological account I wish to defend here is that all of the formal, material, efficient causal and temporal conditions for something's being alive are simply requirements that any material entity must satisfy in order to be able to realise intrinsic ends (of whatever sort), given the laws of nature in our universe. I have included a proviso regarding the laws of nature because in a world with different laws, the necessary conditions for being alive would be very different.

On my proposal, the formal conditions for being alive are simply informational and structural requirements that any physical entity must satisfy before it can be regarded as having an end or telos of its own, namely: (i) a master program to direct its formation and co-ordinate its activities; (ii) a nested hierarchy of organisation that subordinates the parts to the whole they belong to; and (iii) embedded functionality, which guarantees that the activity of the parts is dedicated to the good of the whole. The "common thread" uniting these conditions is that if any of them were not realised, the entity would no longer be capable of teleonomic behaviour, and hence would be unable to realise any ends of its own.

If, however, all of the above formal conditions are realised by a physical entity, then the behaviour of the entity cannot fail to be teleonomic: the entity is inherently capable of doing things for itself. Thus no "extra" formal conditions are needed.

There may, however, be lower-level formal requirements that an entity must satisfy in order to meet conditions (i), (ii) and (iii). Indeed, I shall argue below that some of the seven "pillars of life" stipulated by Koshland (2002) are actually low-level formal requirements that serve to guarantee the integrity of living cells.

On the account I am developing, the material conditions for being alive are simply the physical requirements an entity must satisfy in order to possess the right kind of form - i.e. one which meets the three conditions described in section 1.B.2.

The efficient causal conditions for life can be defined as the set of interactions that must take place between a physical entity and its environment in order for it to achieve those intrinsic ends (such as nutrition and reproduction) which are common to all organisms.

Finally, the temporal conditions for life are simply those which an individual organism or lineage of organisms must satisfy, in order to continue achieving its own ends over a prolonged period of time. Given the laws of nature in our universe, an organism (or lineage of organisms) is likely to encounter certain potentially destructive changes over the course of time, so it needs to be able to combat or cope with them. Specifically, organisms and lineages of organisms need to be able to withstand the onslaught of entropy and to adjust to ongoing environmental changes. These requirements generate the thermodynamic conditions and the evolutionary conditions for life, respectively.

Our account of life is thus "five-dimensional", insofar as it stipulates necessary conditions for life along each of the five axes identified in section 1.A.2.

1.B.6(b) What are the necessary conditions for something's being alive?

I discuss the necessary conditions for life in detail in the Appendix, where I focus particularly on former "Science" editor-in-chief Daniel Koshland's "seven pillars of life" (2002), which I classify according to the "five-dimensional" scheme described earlier. The alleged requirements of life which I discuss include: material requirements (concrete material realisation, spatial contiguity), lower-level formal requirements (possession of a boundary, seclusion and cellular structure), efficient causal conditions (metabolism, interactions with the environment, independence), and temporal conditions (temporal contiguity, thermodynamic requirements such as energy, reproduction and a life cycle; the ability to evolve; and adaptability).

Some of the more surprising results of my investigation are as follows:

1.B.7 What does it mean for a living thing to have a nature?

The Darwinian and Aristotelian conceptions of life are usually thought of as diametrically opposed. In particular, Aristotle's doctrine that every species is eternal and that its members share a common nature is widely regarded as utterly antithetical to Darwin's theory that species are in continual flux and evolve over time. I shall argue that on the contrary, the Aristotelian and Darwinian accounts of life have much in common, and that although Aristotle's definition of a species is in need of some revision, a neo-Darwinian concept of nature remains viable.

A central feature of Aristotle's account of nature was his observation (Parts of Animals i.1 640a25-6, 640b1-4, ii.1 646a35; Generation of Animals iv.3 767b35; Physics ii.1 193b8, ii.2 194b13, ii.7 198a25; Metaphysics vii.7 1032a22, xii.3 1070a29, xii.5 1071a20, 1071a25, xiii.10 1087a21) that living things normally beget other living things after their own kind - or, as Cameron (2000, p. 106) puts it, "breed true". This fact alone was sufficient to convince him that the opinion of his philosophical opponent Empedocles, that everything which happens to plants and animals is due to chance, could not possibly be correct. Empedocles' account has been described as proto-evolutionary: he claimed that new life-forms fortuitously appear (and old ones disappear) over the course of time. However, Aristotle argued that Empedocles' biological theories were utterly unable to account for the simple fact that creatures breed true. Chance, by definition, cannot account for regular and reliable occurrences in nature - such as the fact that creatures breed true. If living things breed true, Aristotle maintained, it must be because they are the sorts of things they are. As Cameron succinctly puts it:

[T]he fact that species breed true can only be explained by postulating animal and plant natures. Nature, not chance, is the cause of breeding true (2000, p. 107).

It may surprise some readers to realise that Darwinists come down on Aristotle's side of this ancient philosophical quarrel. Although Darwin taught that species evolve over time, what is usually overlooked is the glacial pace at which they do so. The average lifespan of a species is measured in millions of years. Novel variations arising within a new generation are a rarity. That being so, parents require some natural mechanism for propagating their genetic traits to their offspring. Many of the definitions of life listed above specifically mention this point. (Thus Monod (1971) includes reproductive invariance as a characteristic of life: the source of information expressed in a living organism is another structurally identical object, whose information corresponds to its own structure. Maynard Smith, 1975, mentions heredity as a characteristic of life; Winder's 1993 definition mentions "self-replication with each offspring slightly different" (italics mine); Sarver, 1999, stresses that reproduction "enables the transmission of traits from parents to offspring".)

(Aristotle's true "polar opposite" in modern times would have to be Richard Goldschmidt, who proposed that living things do not always reproduce after their kind: on very rare occasions, "hopeful monsters" are born which are radically different from their parents, enabling new biological functions to arise in nature. Darwin maintained the contrary view that if nature is capable of generating complex structures in organisms, then they must arise gradually. The sudden appearance of such structures would be a miracle; hence Darwin's invocation of the old aphorism that "Nature does not make leaps".)

The real difference between Aristotle and modern Darwinists lies not in the latter's supposed denial of the objective reality of species, but in their differing conception of a species. Plato and Aristotle employed a typological or morphological concept:

The word 'species' conveyed the idea of a class of objects, members of whom shared certain defining properties... Such a class is constant, it does not change in time, all deviations from the definitions of that class are merely "accidents", that is, imperfect manifestations of the essence (Mayr, 1996, p. 267).

The biological concept of a species is quite different. The term "species" may refer to either a category in the Linnaean hierarchy (kingdom, phylum, class, order, family, genus, species) or to a taxon - "Species are groups of interbreeding natural populations that are reproductively isolated from other such groups" (Mayr, 1996, p. 264). The biological purpose of the species is "the protection of a harmonious gene pool" (Mayr, 1996, p. 264). Reproductive isolation is achieved through isolating mechanisms within individuals. Isolating mechanisms are not 100% efficient; thus related species may sometimes hybridize. Nevertheless, "hybrids between species, particularly animals, are almost always of inferior viability and more extreme hybrids are usually even sterile" - although occasionally in plants they may give rise to new species (Mayr, 1996, p. 263). The biological species concept is only applicable to organisms that reproduce sexually; it does not apply to asexual organisms.

Historically, the reason why scientists began to abandon the morphological conception of species in the second half of the nineteenth century was that morphological characters proved unreliable for the recognition of biological species: members of closely related sibling species may have few or no morphological differences, while numerous different morphological types may occur within a biological species, either because of individual genetic variation or different life history categories (males, females, immatures) (Mayr, 1996, p. 268).

I would suggest that the modern biological concept of a species, far from being at odds with Aristotle's philosophy, actually provides a better explanation for Aristotle's observation that living things "breed true": on the modern concept, this is true by definition. That does not, however, make it trivial: the fact that all populations of organisms do in fact breed true remains an a posteriori truth, which is grounded in the nature of each species of organism. I shall say more about these natures below.

Mayr (1996) describes biological species taxa as concrete phenomena: he even describes them as "particulars, 'individuals', biopopulations". This brings us to the most significant difference between the traditional morphological and modern biological concepts of a species: according to the former way of thinking, an individual is a member of a species by virtue of its own characteristics; whereas on the latter view, the relevant characteristics are those of the population as a whole, which is said to comprise a gene pool. An individual is said to belong to a species - even if it is sterile or denied the opportunity to interbreed with other individuals -if its genes are the product of the same gene pool as that of the population.

Aristotle described three major holistic ends of an organism:

individual flourishing (Physics ii.2 194a28-33, ii.3 195a23-5, ii.7 198b8-9; Politics i.2 1252b34-5; Eudemian Ethics i.8 1218b9-11, ii.1 1219a9-11; Metaphysics i.3 983a31-2);

maturation (Generation of Animals ii.3 736b4-5; Physics ii.2 194a29-33; Movement of Animals 6 700b15-16); and

reproduction (Politics i.2 1252a28-30, De Anima ii.4 416b23-5, Generation of Animals ii.23 731a25-b7, History of Animals viii.1 588b25-6).

It is clear that individual members of the same species do not share identical ends: the two sexes, for instance, possess different internal organs, so their individual flourishing is realised differently. Different morphological types within the same species also have different ends.

On the other hand, if we look at the holistic ends that are realised by all of the various members of a population as a whole, it becomes readily apparent that these ends remain stable within a population over long periods of time. For any individual organism X we select from any biological population, there are individuals (of the same sex and morphological type) among its ancestors thousands of years ago, whose organs possess the same ensemble of functionality as those of X; thus the individual flourishing (on a purely biological level) of the organism X is the same as that of its ancestors. Additionally, X's biological body clocks will be the same as those of its "near" ancestors who lived "only" a few thousand years ago, meaning that the goal or end point of its process of maturation is the same. Finally, the individual X belongs to a gene pool which is sufficiently stable, even over a period of thousands of years, that any accumulated changes in the population's gene pool will not yet be sufficient to constitute a reproductive barrier; thus X's reproductive ends are the same as those of its ancestors. This suggests one way in which a neo-Aristotelian who accepts Darwinism might build the concept of "nature" into the definition of a chronospecies - a biological term for the temporal extension of a species.

If for any individual X selected from a population at time t2, some ancestral individuals can be found in the same population at an earlier time t1, such that:

(i) the part-whole functionality of their organs (and hence their bodily morphology) matches that of their descendant X;

(ii) the end points of their biological body clocks are the same as those of their descendant X; and

(iii) there are no reproductive barriers such as would affect viability of offspring if - counterfactually - these ancestors were to mate with their descendant X,

then we can say that:

(a) these ancestral individuals have the same holistic ends as their descendant X;

(b) the nature of the population has remained the same over the time interval from t1 to t2; and therefore

(c) the population at t1 belongs to the same chronospecies as the population at t2.

Standard definitions of "chronospecies" contain an ineliminable reference to counterfactuals, as my proposed definition does. The difference between my definition and standard definitions lies in the explicit reference to ends, and the fact that species identity is made consequent upon the fact that for any individual selected from a population, some of its ancestors possess the same ends and hence the same nature.

(For instance, Hyperdictionary (2003, Web address defines a chronospecies as follows: A chronospecies is a species which changes physically, morphologically, genetically and/or behaviourally over time on an evolutionary scale (experiences a phyletic shift) such that the species from the early point in time and the species it becomes at the later point in time could not be classified as the same species had they existed at the same point in time.)

Cameron has argued persuasively (2000, pp. 138-166) that Aristotle regarded the property of directedness upon an object as an ontological primitive. That being the case, Aristotle's view might appear to be in conflict with the Darwinian tenet that evolutionary processes in nature are "blind" and not "directed" at anything. However, I would suggest that the contradiction is only apparent: the randomness of evolutionary processes is quite compatible with the occurrence of non-random teleonomic processes within the body of an individual organism. If evolution had an end, it could only be extrinsic - a notion which Darwinism rejects - whereas the ends of an individual organism are intrinsic. The fact that evolution has no extrinsic "goal" or "purpose" does not mean that an individual's body parts possess no functions or intrinsic ends.

The final obstacle for a neo-Aristotelian account of nature relates to transitivity of identity. An individual X may have the same nature (according to the teleological criteria listed above) as its ancestors Y1, Y2, ... Ym, and these individuals may have the same nature as their ancestors Z1, Z2, ... Zn, yet it may be the case that X and Z1, Z2, ... Zn possess different natures by the same criteria (e.g. due to the accumulation of genetic reproductive barriers over a long period of time) - generating a paradox.

The correct resolution of this paradox is to recognise that the notion of "sameness" used here differs from most standard cases. First, there is no external yardstick with which we can measure the feature being compared between individuals: an individual's species, like its neighbourhood, can only be defined relative to the individual itself. In this respect, an individual's species is unlike its color (which can be measured on a scale of wavelengths and intensities) or the number of hairs on its head. Hence the comparative property of "belonging to the same species as an individual X" can only be defined relative to X, like the property of "being in the same neighbourhood as X".

Second, a species is an inherently imprecise concept because the notion of a reproductive barrier is imprecise. This fact makes it impossible for scientists to fix precise boundaries for species. The reason for the absence of reproductive barriers between an individual and its proximate ancestors is that the genetic changes that have accumulated in the population since these ancestors died are as yet too tiny to constitute a barrier: viewed against the variability of the gene pool as a whole, they could be described as background "noise". Likewise, new functions are evolving in populations all the time, but because it takes millions of years for a new function to evolve, the appearance of incipient functions is not noticeable over relatively short intervals of thousands of years. Only after a very large number of generations have elapsed can we speak of a change of nature in the population. In Aristotelian terms, we can say that the intrinsic ends of a lineage of organisms change over very long periods of time.

There are thus two good reasons why transitivity of the sameness relation might not hold. The exact point at which a lineage changes its nature and evolves into a new species may elude us, but this generates transitivity paradoxes only if we assume that "nature" is both an externally defined concept and a precisely defined concept. In any case, the fact that species sameness is not a transitive relation should not occasion any perplexity. The sceptical query, "Where do you draw the line between a species and the species that succeeds it?" presupposes that a clear line has to be drawn in the first place.

Likewise, the fact that two distinct species S1 and S2 could both conceivably share the same nature as the common ancestral population from which they both arose, does not vitiate the concept of nature as such. Indeed, the very question, "Which species does the ancestral population belong to - S1 or S2?", is anachronistic: at the earlier time t1, it cannot be meaningfully posed. What we should say here is that small differences within two recently segregated populations which do not yet constitute a reproductive barrier may, over the course of time, accumulate to the point where the two populations no longer possess the same nature.

I conclude that the concept of nature, while needing considerable revision in the light of what scientists now know about evolution of new species, remains a viable and useful one, as it situates an individual within a community where there are no reproductive barriers between members, and where some other individuals can be found who share the same teleological ends.

1.B.8. Do living things have a privileged ethical status?

Problems with the desire-centred theory of welfare

The notion that a non-sentient organism (a plant, let us say) could be said to have interests sounds strange to many philosophers. Feinberg (1974, pp. 49-50) has argued that only a being with conations (e.g. wishes, desires, hopes, drives, aims and goals) can have interests, and that only conscious conations can define interests. Varner (1998, pp. 57-61) criticises this view because it does not do a good job in accounting for cases where an individual's desires and interests conflict.

Firstly, an individual may have desires (or conscious conations) that go against his/her best interests. Varner asks us to imagine the case of Nanci, a cat who desires access to the outdoors, but does not and cannot understand the risks involved, such as contracting the feline leukemia virus (or, in the area where I live, being run over by a car). Nanci may want to go outside, even though it is manifestly obvious (to us) that her desire is against her best interests. While a biologically grounded theory of interests can readily explain why this is so (e.g. viruses are bad for her health), a theory which grounds interests solely in mental states requires awkward counterfactuals to support the same conclusion (e.g. if, per impossibile, she could master a language and understand what we do, she would no longer desire to go out).

Secondly, an individual may have interests that (owing to ignorance on his/her part) he/she does not desire. It was certainly in the interests of nineteenth century mariners to take 10 milligrams of ascorbic acid a day, to avoid scurvy, but because they knew nothing about ascorbic acid, they had no desire for it. Such a case is unproblematic if we allow interests to be grounded in biological facts, but a psychological, desire-based account of interests can only construe the mariners' interests counterfactually: the mariners would have desired ascorbic acid, had they been sufficiently informed (in other words, had they known what we now know).

The psycho-biological account of welfare

Varner proposes what he calls a

psycho-biological account of welfare, according to which to say X is in A's interest means that

(1) A actually desires X, [or]

(2) A would desire X if A were sufficiently informed and impartial across phases of A's life, or

(3) X serves some biologically based need that A has in virtue of being the kind of organism A is (1998, p. 62).

Varner makes it clear that the "or" is meant to be inclusive. His account can easily explain the anomalous cases cited above, by reference to condition (3). The only significant criticism I would wish to make of Varner's account is that it does not prioritise desires, informed desires and biological needs. Surely, in the case of Nanci, if there is a feline leukemia epidemic, then going outside is not in her interests, even though she may take an interest in it (i.e. desire to go outside) - to cite a distinction that Varner himself makes (p. 57). On Varner's psycho-biological account, a strange conclusion follows: both going outside and staying indoors are in Nanci's interests: the former satisfies condition (1), while the latter satisfies condition (3). One might allow the co-occurrence of conflicting interests satisfying conditions (2) and (3) - for instance, it may be in a firefighter's interest to remain in a burning building in order to satisfy her impartial interest in saving as many lives as possible, while getting out as quickly as possible may serve some biological need of hers (especially if she is predisposed to getting bronchial problems) - but it seems unreasonable to put mere desires, especially uninformed ones, on a par with biological needs.

One merit of Varner's account is that it would allow us to say that plants have interests, as they can satisfy condition (3). However, the account is open to an obvious riposte: artifacts (e.g. cars and human-built computers) have needs too, so why do we preclude them from having interests? (Varner considers the idea that a can opener has interests to be a reductio ad absurdum.) First, he addresses the empirical question: how do the needs of plants (or more generally, organisms lacking desires) differ from those of artifacts?

Varner solves this problem by proposing an etiological account of life: he proposes that "biological functions, rather than goals or end-states" are required to "draw a sharp distinction between all artifacts, on the other hand, and all living organisms on the other" (1998, p. 67). According to Varner, what is good for an organism cannot be adequately defined in terms of its end-states, as artifacts (such as a Patriot missile) may also have "built-in goals". On Varner's account, the function of an organ or subsystem is explained in terms of the selective advantage it provided its owner's ancestors over other individuals lacking this function. I have already discussed the problems with Varner's etiological account of life in section 1.A.3.6, so I shall not repeat my criticisms here. Varner's claim that the Patriot missile has built-in goals was evaluated in section 1.B.5, where I argued that the missile lacked intrinsic finality.

Varner then argues that organisms, unlike artifacts, are the result of natural selection (1998, p. 69), from which it follows that (according to his etiological definition) biological functions are only found in organisms. Varner then re-writes condition (3) of his psycho-biological account of interests as follows:

(3) X would fulfill some biological function of some organ or subsystem of [an individual] A ["biological function" is then defined in terms of its selective advantage, as above] (1998, p. 68).

The fact that organs or subsystems have a function gives their possessors a "good of their own" in an objective, biological sense - a feature which artifacts lack:

Cars, can openers and intrinsically valuable things [e.g. works of art - V.T.] all can be coherently said to have needs, but only things with a good of their own can coherently be said to have interests (1998, p. 56, italics mine).

The problem with Varner's account is that it uses the concept of natural selection to make a sharp distinction between organisms and artifacts - a distinction which is both empirical and ethical. Although the property of having a history of natural selection may track the organism-artifact distinction pretty well (if we choose to ignore the troubling cases of abstract computational systems that evolve, and future artifacts that may qualify as alive), it is incapable of defining it. To define this distinction, we require the concept of intrinsic finality, which is manifested in the formal properties of possessing internal relations, a master program, a nested hierarchy of organisation and dedicated functionality.

A more serious problem with Varner's distinction is that although provides us (in practice) with an excellent yardstick for objectively determining what is good for an organism, it does not properly define what "good" means. "Being the product of Darwinian evolution" does not appear to be a morally relevant property per se.

The concept of natural selection appears inadequate to explain why organisms, and not artifacts, "qualify ... for direct moral consideration" (1998, p. 69). The nub of Varner's argument (1998, pp. 73 - 74) is that (i) the fulfilment of biological functions in sentient organisms is in their interests irrespective of their even being able to take an interest in their fulfilment (as is shown by the cases of Nanci the cat, who wants to go out during an epidemic of feline leukemia virus, and the nineteenth century mariners, who had no desire for ascorbic acid); so (ii) the fulfilment of biological functions (which we can enumerate in a non-arbitrary fashion, using the concept of natural selection) in non-sentient organisms, should be considered as being in their interests, even though they cannot take an interest in their fulfilment. I do not wish to criticise the logic of Varner's move from (i) to (ii), which seems reasonable. I would simply like to point out that the role served by natural selection in this argument is merely to identify what the biological functions of organisms are, rather than why they are morally significant. That question is left unanswered.

Putting the problem another way: Varner has failed to answer the question of why a biological interest, per se, matters morally, while a mere need (e.g. a car's need for oil) does not. Granting that the needs of plants are the product of Darwinian evolution while the needs of cars are not, I might say, "So what?"

Outline of an alternative view

It has been argued that there is a real, non-arbitrary distinction between living and non-living entities, and that this difference can be explained in terms of the form and finality of living things. What is good for an organism can be ascertained from understanding the organism's master program, its hierarchy of organisation, and how its functionality, on each level, supports that of the next highest level.

We have seen that defining a thing's "good" as "the set of functions which helped its ancestors" does not suffice to make them morally relevant. But if instead "good" is defined (roughly) as "the set of functions to which its parts are dedicated", then at least its benefit to the thing itself is clear.

The next question to be addressed is: does this difference matter ethically for us? Is there something about being alive that warrants special ethical treatment?

I would like to offer answer this question by highlighting a deficiency in the desire-centred theory of welfare, which ignores purely biological interests. If this theory is correct, then "doing good" has to be cashed out in terms of "satisfying someone's desires" - perhaps, if one is generous, the greatest number of desires in the greatest number of interested parties, along the lines proposed by some utilitarians. Yet it is not at all clear why the satisfaction of anyone's desires is even a prima facie good thing, without further need of justification. While it is factually true that most of the things people want are good for them, most of the time - otherwise they would probably be dead - this does not tell us why X's being wanted by A should make X good for A.

If instead we consider the satisfaction of biological interests, then in this context, "doing good" means (on my proposal) "promoting some biological function(s) in an organism, which its body parts are dedicated to supporting" - roughly, promoting its health. Promoting the health of an organism, as a doctor, a vet or the parent of a child might do, does sound like a good that needs no further justification. If this is not acting ethically, then it is difficult to see what might be.

An understanding of an organism's nature can tell us what is good for it, insofar as it will have some counterparts of the same species who share the same intrinsic ends as itself.

Of course, the good of a moral agent - or even a sentient animal - is a multi-faceted thing, which certainly cannot be reduced to the notion of health. In chapter 5, I explore the notion of "basic goods" and argue that their scope encompasses not only the human arena but the entire spectrum of goods that can be realised by organisms in the world around us.