36. THE SURVIVAL GAME

by Martin Wells

From the moment that an animal is born it has to make decisions. It has to decide which of the things around it are for eating, and which are to be avoided; when to attack and when to run away. The animal is, in effect, playing a complicated and potentially very dangerous game with its environment, a game in which it must make decisions for which the reward is survival and the penalty for a mistake discomfort or obliteration.

This is a difficult and unpleasant business and few animals would survive if they had to start from scratch and learn about the world wholly by trial and error, for there are too many possible decisions which would prove fatal. So we find, in practice, that the game is always rigged' in favour of the young animal in one way or another. All species cheat in some way. Either the animal is protected during the early stages of its learning about the world around it, or the knowledge of which way to respond is built into its nervous system from the start.

Let us consider two extreme cases: first, our own species, as an example of an animal in which responses are wholly, or almost wholly, determined by individual experience. I have a small son. The fact that he has survived to the age of two, despite an inquiring mind and for all practical purposes no knowledge whatever of the properties of things in the world around him, is largely due to considerable care on the part of his normally untidy parents. Since he came, we clear up.' We hide saws and chisels; we put away bottles of detergent and the stuff that is sold to kill the slugs; we lay ladders flat instead of leaving them propped against things; and we shut the garden gate. In short, the boy's opportunities for experiment are kept within limits, so that while he is free to learn by experience, we can be fairly sure that he is not going to hurt himself really badly. As he learns more about the world, we can relax the limits.

That is one way of rigging the survival game. The other extreme way of fixing the result is to have the right answers to all potentially disastrous experiments fitted in at the outset. Some animals are indeed built like this: sea anemones, for example. Any sea anemone knows what is edible and what is not. It will grasp food with its tentacles and cram it into the mouth. It will reject inedible objects and close up when poked. A sea anemone does not learn to do these things: they are a property of the construction of its nervous system.

The fact that animals behave sensibly can be attributed partly to what we might call "genetic learning", to distinguish it from the individual learning that an animal does in the course of its own lifetime. Genetic learning is learning by a species as a whole, and it is achieved by selection of those members of each generation that happen to behave in the right way. However, genetic learning depends upon a prediction that the future will more or less exactly resemble the past. The more variable individual experience is likely to be, the less efficient is genetic learning as a means of getting over the problems of the survival game. And because most animals live in non-uniform environments, the details of which are liable to vary from one generation to the next, it is not surprising to find that very few species indeed depend wholly upon genetic learning. In the great majority of animals, behaviour is a compound of individual experience superimposed upon an inherited bias to behave in particular ways.

(from The Listener, 5th January, 1961)