High school biology textbooks usually present a list of properties which collectively constitute a definition of life. A typical list, from [Day 85] (p.34), is: Movement, Feeding (obtaining energy and raw materials), Respiration, Excretion, Growth (increase in size and complexity), Reproduction, and Sensitivity.
Many similar lists have also been proposed, each combining some aspects of organisms from the evolutionary perspective, and some from the ecological perspective. One of the major problems with check-list definitions, however, is that it is invariably possible to find an example of a real organism which fails to meet all of the requirements. A common problem case is the mule, which is certainly alive, although incapable of reproduction.
Considering the difficulties of such an endeavour, it is reasonable to ask whether we could ever arrive at a concise definition of life. Chris Langton, regarded by many as the founder of the scientific discipline of `artificial life' (at least in its present incarnation), is skeptical that we will ever reach a consensus on this issue. From his experience of creating computer programs which exhibit many of the properties associated with living organisms, he remarks ``every time we succeed in synthetically satisfying the definition of life, the definition is lengthened or changed.''2.15
Attempts to fuse the ecological and evolutionary pictures in an illuminating and precise manner are complicated by the intricate, non-linear interactions involved. One of the major problems in developing a coherent picture of living systems is the description of the two-way interactions between processes occurring at vastly different time scales. Waddington says:
``Perhaps the main respect in which the biological picture is more complex than the physical one, is the way in which time is involved in it ... [T]o provide anything like an adequate picture of a living thing, one has to consider it as affected by at least three different types of temporal change, all going on simultaneously and continuously. These three time-elements in the biological picture differ in scale. On the largest scale is evolution ... On the medium scale ... life history ... Finally, on the shortest time-scale ... rapid turnover of energy or chemical change.'' [Waddington 57] (pp.5-6).
In the 40 years since Waddington wrote this, little progress has been made in developing models which satisfactorily capture the interactions of processes happening at these different time-scales. Recent attempts to achieve such a synthesis are discussed by John Holland in his book Hidden Order [Holland 95] (esp. Chapter 5). Further efforts from the theoretical biology community at modelling the interface between ecology and evolution are discussed in [Travis & Mueller 89], [Stanley 89] and [Feldman 89].
In Section 2.1.2 some objections were raised to defining life in terms of evolution, as evolution is just a mechanism for change. Margulis and Sagan accept this, but claim that both autopoiesis and reproduction are distinguishing processes of living matter [Margulis & Sagan 86]. They argue that, while autopoiesis defines the organisation of an individual living entity, we still need to retain the idea that organisms reproduce in order to appreciate the larger picture of life:
``Autopoiesis occurs, then, to maintain an organism during its own life, but by itself autopoiesis does not guarantee that an organism will show genetic continuity or that the characteristics of any given organism will persist faithfully through time. The process that ensures genetic continuity is reproduction. But autopoiesis remains the primary process. On the one hand, without it the organism would not survive to reach the stage at which reproduction becomes feasible. On the other hand, autopoiesis does not depend on reproduction, at least within a single generation. An infertile but healthy person with muscle, circulatory, excretory, and other organ systems in excellent running order is autopoietic even though he is unable to reproduce. In an evolutionary sense, however, such an individual has forfeited genetic continuity; he is already dead.'' [Margulis & Sagan 86] (p.13).
Any approach that redefines life using terms such as reproduction and autopoiesis has the great advantage that such terms can be precisely defined. Nils Barricelli, possibly the first person to implement and experiment with an artificial evolutionary system ([Barricelli 57], [Barricelli 62], [Barricelli 63]) clearly realised the problems with using words such as `life'. In his work, Barricelli gave a precise definition of what he called a `symbioorganism' (the work was based upon the concept of symbiogenesis, discussed earlier). Rather than asking whether the organisations that evolved in his system were alive, he turned the question upon its head:
``As a matter of fact [the question of whether the evolved organizations are alive] has no meaning as long as there is no agreement on a definition of `living being'. However, the reciprocal question `whether the objects we are used to call [sic] living beings are a particular class of symbioorganisms' has a meaning. This is the question we have been trying to answer in this paper ...'' [Barricelli 63] (p.99).
This tactic of reversing the question, and using terms which can be precisely defined in preference to terms such as `life', will be the one taken throughout most of this work. Clarification of some of the terms to be used is given in Section 2.5. We will return to the problem of providing an adequate characterisation of the concept of life in Section 7.3.3.