Animal intelligence: Experimental studies

by Edward L. Thorndike

Provisional Laws of Acquired Behavior or Learning

The Law of Effect is that: Of several responses made to the same situation, those which are accompanied or closely followed by satisfaction to the animal will, other things being equal, be more firmly connected with the situation, so that, when it recurs, they will be more likely to recur; those which are accompanied or closely followed by discomfort to the animal will, other things being equal, have their connections with that situation weakened, so that, when it recurs, they will be less likely to occur. The greater the satisfaction or discomfort, the greater the strengthening or weakening of the bond.

The Law of Exercise is that: Any response to a situation will, other things being equal, be more strongly connected with the situation in proportion to the number of times it has been connected with that situation and to the average vigor and duration of the connections.

These two laws stand out clearly in every series of experiments on animal learning and in the entire history of the management of human affairs. They give an account of learning that is satisfactory over a wide range of experience, so long as all that is demanded is a rough and general means of prophecy. We can, as a rule, get an animal to learn a given accomplishment by getting him to accomplish it, rewarding him when he does, and punishing him when he does not; or, if reward or punishment are kept indifferent, by getting him to accomplish it much oftener than he does any other response to the situation in question.

For more detailed and perfect prophecy, the phrases ‘result in satisfaction’ and ‘result in discomfort’ need further definition, and the other things that are to be equal need comment.

By a satisfying state of affairs is meant one which the animal does nothing to avoid, often doing such things as attain and preserve it. By a discomforting or annoying state of affairs is meant one which the animal commonly avoids and abandons.

The satisfiers for any animal in any given condition cannot be determined with precision and surety save by observation. Food when hungry, society when lonesome, sleep when fatigued, relief from pain, are samples of the common occurrence that what favors the life of the species satisfies its individual members. But this does not furnish a completely valid rule.

The satisfying and annoying are not synonymous with favorable and unfavorable to the life of either the individual or the species. Many animals are satisfied by deleterious conditions. Excitement, overeating, and alcoholic intoxication are, for instance, three very common and very potent satisfiers of man. Conditions useful to the life of the species in moderation are often satisfying far beyond their useful point: many conditions of great utility to the life of the species do not satisfy and may even annoy its members.

The annoyers for any animal follow the rough rule that alterations of the animal’s ‘natural’ or ‘normal’ structure—as by cuts, bruises, blows, and the like,—and deprivations of or interference with its ‘natural’ or ‘normal’ activities,—as by capture, starvation, solitude, or indigestion,—are intolerable. But interference with the structure and functions by which the species is perpetuated is not a sufficient criterion for discomfort. Nature’s adaptations are too crude.

Upon examination it appears that the pernicious states of affairs which an animal welcomes are not pernicious at the time, to the neurones. We learn many bad habits, such as morphinism, because there is incomplete adaptation of all the interests of the body-state to the temporary interest of its ruling class, the neurones. So also the unsatisfying goods are not goods to the neurones at the time. We neglect many benefits because the neurones choose their immediate advantage. The neurones must be tricked into permitting the animal to take exercise when freezing or quinine when in a fever, or to free the stomach from certain poisons.

Satisfaction and discomfort, welcoming and avoiding, thus seem to be related to the maintenance and hindrance of the life processes of the neurones rather than of the animal as a whole, and to temporary rather than permanent maintenance and hindrance.

The chief life processes of a neurone concerned in learning are absorption of food, excretion of waste, reception and conduction of the nerve impulse, and modifiability or change of connections. Of these only the latter demands comment.

The connections formed between situation and response are represented by connections between neurones and neurones, whereby the disturbance or neural current arising in the former is conducted to the latter across their synapses. The strength or weakness of a connection means the greater or less likelihood that the same current will be conducted from the former to the latter rather than to some other place. The strength or weakness of the connection is a condition of the synapse. What condition of the synapse it is remains a matter for hypothesis. Close connection might mean protoplasmic union, or proximity of the neurones in space, or a greater permeability of a membrane, or a lowered electrical resistance, or a favorable chemical condition of some other sort. Let us call this undefined condition which parallels the strength of a connection between situation and response the intimacy of the synapse. Then the modifiability or connection changing of a neurone equals its power to alter the intimacy of its synapses.

As a provisional hypothesis to account for what satisfies and what annoys an animal, I suggest the following:—

A neurone modifies the intimacy of its synapses so as to keep intimate those by whose intimacy its other life processes are favored and to weaken the intimacy of those whereby its other life processes are hindered. The animal’s action-system as a whole consequently does nothing to avoid that response whereby the life processes of the neurones other than connection-changing are maintained, but does cease those responses whereby such life processes of the neurones are hindered.

This hypothesis has two important consequences. First: Learning by the law of effect is then more fully adaptive for the neurones in the changing intimacy of whose synapses learning consists, than for the animal as a whole. It is adaptive for the animal as a whole only in so far as his organization makes the neurones concerned in the learning welcome states of affairs that are favorable to his life and that of his species and reject those that are harmful.

Second: A mechanism in the neurones gives results in the behavior of the animal as a whole that seem beyond mechanism. By their unmodifiable abandonment of certain specific conditions and retention of others, the animal as a whole can modify its behavior. Their one rule of conduct causes in him a countless complexity of habits. The learning of an animal is an instinct of its neurones.

I have limited the discussion to animals in whom the connection-system is a differentiated organ, the neurones. In so far as the law of effect operates in an animal whose connection-system is not anatomically distinguishable and is favored and hindered in its life by the same conditions that favor and hinder the life of the animal as a whole, the satisfying and annoying will be those states of affairs which the connection-system, whatever it be, maintains and abandons.

The other things that have to be equal in the case of the law of effect are: First, the frequency, energy and duration of the connection,—that is, the action of the law of exercise; second, the closeness with which the satisfaction is associated with the response; and, third, the readiness of the response to be connected with the situation.

The first of these accessory conditions requires no comment. A slightly satisfying or indifferent response made often may win a closer connection than a more satisfying response made only rarely.

The second is most clearly seen in the effect of increasing the interval between the response and the satisfaction or discomfort. Such an increase diminishes the rate of learning. If, for example, four boxes were arranged so that turning a button caused a door to open (and permit a cat to get freedom and food) in one, five, fifty and five hundred seconds, respectively, a cat would form the habit of prompt escape from the first box most rapidly and would almost certainly never form that habit in the case of the fourth. The electric shock administered just as an animal starts on the wrong path or touches the wrong mechanism, is potent, but the same punishment administered ten or twenty seconds after an act will have little or no effect upon that act.

Close temporal sequence is not the only means of insuring the connection of the satisfaction with the response producing it. What is called attention to the response counts also. If a cat pushes a button around with its nose, while its main occupation, the act to which its general ‘set’ impels it, to which, we say, it is chiefly attentive, is that of clawing at an opening, it will be less aided in the formation of the habit than if it had been chiefly concerned in what its nose was doing. The successful response is as a rule only a part of all that the animal is doing at the time. In proportion as it is an eminent, emphatic part of it, learning is aided. Similarly discomfort eliminates most the eminent, emphatic features of the total response which it accompanies or shortly follows.

The third factor, the susceptibility of the response and situation to connection, is harder to illustrate. But, apparently, of those responses which are equally strongly connected with a situation by nature and equally attended to, some are more susceptible than others to a more intimate connection.

The things which have to be equal in the case of the law of exercise are the force of satisfyingness; that is, the action of the law of effect, and again the readiness of the response to be connected with the situation.

The operation of the laws of instinct, exercise and effect is conditioned further by (1) what may be called the law of assimilation or analogy,—that a situation, especially one to which no particular response is connected by original nature or previous experience, may connect with whatever response is bound to some situation much like it,—and (2) by the law of partial activity—that more or less of the total situation may be specially active in determining the response.

The first of these laws is a result of the facts that conduction in the neurones follows the line of least resistance or closest connection, that the action-system is so organized that certain responses tend to be made in their totality if at all, and that slightly different situations may, therefore, produce some one response, the effects of their differences being in the accessories of that response.

The second law is a result of the facts that the situation, itself a compound, produces a compound action in the neurones, and that by reason of inner conditions, the relative intensities of different parts of the compound may vary. The commonest response will be that due to the modal condition of the neural compound, but every condition of the compound will have its response.