It claims that it is possible for the intelligent adaptations, or any sort of "accommodations," made by the individuals of one generation, to set the direction of subsequent evolution, even though there be no direct inheritance of acquired characters from father to son.

It proceeds in the case of instinct somewhat thus: Suppose we say, with the first theory given above, that the organism has certain reflexes which show some degree of adaptation to the environment; then suppose we admit the point, urged by the advocates of the lapsed intelligence theory, that the gradual improvement of these reflexes by variations in the endowment of successive generations would not suffice for the origin of instinct, seeing that partial instincts would not be useful; and, further, suppose we agree that many of the complex instincts really involved intelligent adaptation in their acquisition.

These points carefully understood, then one new and further principle will enable us to complete a theory which will avoid the objections to both the others.

This principle is nothing else than what we have seen already--namely, that the intelligence supplements the partial instincts in each generation and makes them useful in the respects in which they are inadequate, and so keeps the young alive in successive generations as long as the instinct is imperfect.

This gives the species time gradually to supplement its instinctive endowment, in the course of many generations each of which uses its intelligence in the same way: time to accumulate, by the occurrence of variations among the offspring, the changes in the nervous system which the perfect instinct requires.
Thus as time goes on the dependence of each generation upon the aid of intelligence is less and less, until the nervous system becomes capable of performing the function quite alone.