Our characterisation of the living state seems to apply to all kingdoms of organisms, though some readers might balk at the exclusion of viruses. Viruses aside, therefore, the characterisation seems sufficiently general. But is it also sufficiently restrictive?
A crystal of any substance in contact with a saturated solution of the same substance shares many of the properties of organisms. It exchanges material with its surroundings; molecules are exchanged between the crystal surface and the solution. It might therefore be said to "eat" and "excrete". It also grows. And it can replicate itself by stimulating the formation of similar crystals from the solution. Several authors have drawn attention to these similarities between crystals and organisms and have suggested that crystals, which are plainly not alive, emphasise the difficulty of trying to distinguish sharply between the living and the non-living.
However, our characterisation of "livingness" unequivocally excludes crystals. Crystals do not metabolise and do not have internal transport processes (though under some conditions the units of a crystal lattice change places slowly, and apparently randomly). Like viruses, therefore, crystals can have elaborate structures but they have no internal states. They have no genomes and therefore nothing analogous to a pattern of gene expression. They do not respond to stimuli through signalling pathways. Therefore they are not alive and have nothing approaching the complexity21 of organisms.
Self-regulating objects such as robots are sometimes considered "living". Their electronic circuitry is certainly complex and even their mechanical parts might have much higher information contents than crystals. Robots detect specific stimuli and respond to them in organised and (all being well) appropriate ways. They contain detailed coded information that defines and controls their responses. External stimuli cause specific parts of this information to be expressed in the robot's actions. However, a robot has no
21 A prokaryotic DNA of one million bases contains two million bits of information. (Given that there is a choice of four bases for each position, there are two bits of information per base.) For human DNA, multiply this figure by about 6,000. Crystals are repetitive structures. Even if the unit crystal is very complicated the information content never remotely approaches these figures.
internal state. Its energy supply is not self-regulated; it does not "metabolise". The spatial organisation of its internal parts is fixed. It does not assemble and maintain itself, or exchange materials with its surroundings in order to do so. Therefore, robots are not wholly autonomous; they do not exhibit the three-way dialogue among internal state, pattern of gene expression and stimulus-response that is characteristic of life.
It would be interesting to know whether any non-living entity fits our characterisation of "livingness". We have not been able to identify one. Provisionally, therefore, we conclude that our characterisation is sufficiently restrictive as well as sufficiently general. It applies to all living things but to nothing else.
Was this article helpful?