ON CONDITIONAL REFLEXES
that the higher abstractions of man, which are due to the more developed complexities of his nervous system, would often make such simple experiments impossible. Once a conditional reaction is established with an animal, no amount of any sort of 'intellectual' persuasion, or the like, would disturb his glandular secretions, as the animal's range of 'meanings' is very limited. These secretions can be diminished or even abolished by other means, but not by 'intellectual' means alone. In the 'normal' man, his 'knowing' that the sound of the metronome or bell is part of an experiment and not a signal for actual food, would, or should, alter his nervous reactions and glandular secretions and make the experiments much more complex. The conditional reactions of the animals have still the element of unconditionally. In man, they may become fully conditional and depend on a much larger number of semantic factors called 'mental', 'psychic'., than we find in any animal.
On the human level, outside of the experiments with the salivary glands, we have in the psychogalvanic reaction a most subtle semantic means of experimenting with the effect of words as connected with some secretions, probably at least the sweat glands. Humans react to different events or words by minute electrical currents (among others) which can be registered by a very sensitive galvanometer and the curves photographed. It is interesting to notice that so-called 'self-consciousness' disturbs the success of the experiments, or makes them impossible, at least with some individuals. It should be remembered that general statements are invalidated if there are any exceptions.
In experiments, we are usually interested in their success. When analysing the oo-valued degrees of conditionally, we are equally interested in their failures, which suggest a far-reaching revision of the interpretation of our experimental data in this field. Although some writers say that the reactions registered are 'beyond control' (unconditional), this statement, in general, is not correct, and should be amended to 'often beyond control' (conditional of different degrees). It is impossible to go into details here, as the problems are extremely complex. In addition to this, the testing of degrees of conditionality presents an extremely wide new semantic Held for experimentation which has not yet been attempted. It should be noted, however, in passing, that in these experiments different types of 'mentally' ill, as well as the 'healthy' persons, exhibit different types of curves.8
When psycho-logical events or s.r are interpreted, the difficulties become particularly acute. Thus, we seldom discriminate between the average and the 'normal' person. In the animal world, under natural conditions - by which is meant entirely without human interference -