Trial and error

In line with empiricist theory, the tendency has been to explain behaviour in terms of the influence of the environment, i.e. as responses dictated by external stimuli, and there has been a corresponding neglect of the influence of the inherited set of instructions or innate releasing mechanisms (I.R.M.), as Lorenz calls them, that determine the generalities of an anergic pattern.

An organism is not born with a blank mind—the tabula rasa of the empiricists—but with a general set of instructions and a corresponding general model reflecting the experience of its unit of phylogeny as a whole, which will be differentiated as a result of its experience, so as to adapt them to the needs of its specific environment.

This fact is apparent from a number of experiments on different types of animals. Fantz experimented on a sample of 1,000 chicks. He found that they pecked 10 times more often at sphere-shaped objects than at pyramidical-shaped ones. The chicks were hatched in darkness and tested on their first exposure to light, from which it must follow that this behaviour can only be explained on the basis of innate tendencies. Tinbergen found that newly hatched herring gulls preferred pecking at objects which resembled the bill of the parent from which they were fed. Marked preferences for certain specific objects were also established among baby chimps. Fantz conducted similar experiments with children. In one of these, 49 children aged from 4 days to 6 months, were presented with,

“. . . three flat objects, the size and shape of a head. On one we painted a stylized face in black on a pink background, on the second we re-arranged the features in a scrambled pattern, and in the third we painted a solid patch of black at one end, with an area equal to that covered by the features . . . The results were about the same for all age levels: the infants looked most often at the real face, looked slightly less often at the scrambled face, and largely ignored the control pattern.”

Fantz concluded:

“Lowly chicks as well as lofty primates perceive and respond to form without experience if given the opportunity at the appropriate age of development. Innate knowledge of the environment is demonstrated by the preference of newly-hatched chicks for forms likely to be edible, and by the interest of young infants in kinds of form that will later aid in object recognition, social responsiveness and spatial orientation.” ¹

The modifications brought about to this general model during behaviour are normally referred to as “learning”. Its role is not to alter innate behavioural tendencies so much as to enable them to be satisfied with ever-greater precision.

Thus an embryo will develop not merely because its environmental conditions are correct, but also because it contains a complex set of general instructions that, as a result of interaction with its environment, will be slowly differentiated. The same is true of all behavioural processes.

One of the consequences of adopting this model of behaviour is that the notion of learning by random trial and error becomes untenable.

If a rat is put into a maze, we know that it can be taught to find its way out. However, before hitting upon the correct route, it will have to make a series of unsuccessful trials. Now, if these trials were arranged to form a series, would it be possible to put order into this series, or must each trial be considered merely a random one? The trial and error theory appears to assume the latter hypothesis. Strictly random behaviour does not occur in a system whether it be an organism, a society, or an eco-system. These all display varying degrees of order, and hence of limitation of choice, or of non-randomness.

Looked at slightly differently, if a rat finds its way through a maze as a result of a series of trials, what will determine its first trial? If a hundred possible moves are open to it, why should it make one rather than any other?

All the actions of an animal, whether it be an earthworm, a rat or a man, as Craik was the first to show, must be regarded as based on hypotheses as to the nature of their environment, or, rather, of their relationship with their environment.

Thus, supposing that the first move made by the rat, rather than lead it towards the opening, on the contrary led it into a further cul-de-sac, the model responsible for this error would have to be modified.

Thus, when the rat made its next move, the situation would be interpreted in the light of a new model—one that had taken into account the failure of its predecessor to interpret correctly the short-term systemic situation in which the rat had found itself, and one furnishing an evermore precise representation of the system for the purpose of finding it way out of the maze.

In other words, each action would be regarded as a correction of an error, and if these were taken as forming a series then the latter would be ‘damped’, in the sense that the errors would be progressively reduced, i.e. the sub-system would be tending towards four-dimensional equilibrium or along a ‘creode’, i.e. towards ever-higher stability. The rate at which this process would be occurring would depend on the sub-system’s ability to postulate hypotheses leading to adaptative behaviour.

Reference

(1) Fantz, Robert L. ‘The Origin of Form Perception’ in S.A. May 1962

Back to top