Are Emotions Teleological Constructs?

Knight Dunlap
The Johns Hopkins University

Nearly forty years ago William James married the emotions to the instincts so solidly that a divorce seemed to be impossible.[1] James' work was indeed but the giving of official sanction and due publicity to a union which had been tentatively announced by Charles Darwin in his Expression of the Emotions in Man and Animals. Apparently, the indissoluble bond has since been dissolved; but as a matter of fact, it is the death of one of the contracting parties that has broken the bond. Instincts have quietly passed away after a brief and feverish illness, and the widowed emotions have been left.

The detailed exposition of James resulted in a general acceptance of the scheme of emotions as organic results of responses. In so far as responses occur as unlearned patterns of specific and unique types, that is, as the manifestations of 'instincts,' we should expect 'emotions' of equally unlearned and unique types, capable of being named distinctively. In so far as responses are modified by the learning process, we should expect emotions correspondingly modified.

This analysis of the problem, indeed, seems as sound today as it did in 1882. If there is, for example, an instinct which may be, called flight which manifest itself in different ways in different circumstances, but always in such a way that it is recognizable as flight, and definitely discriminable from any other instinct, such as the fighting instinct; then there should be a corresponding emotion, which we might call fear, unique in its nature and generically distinct from any other emotions such as 'anger.'

This doctrine, prior to 1918, was so well established that social psychology was said to be 'founded' upon it. It was so thoroughly ingrained into genes psychology that the first behaviorism, although soft-pedalling on 'the emotions,' based itself upon 'the instincts' without an apprehensive question.

Then something happened. When I expressed before the American Psychological Association in 1919 my doubts concerning the reality of 'instincts,' my whole point seemed as fantastical to all but two or three of the members present as if I had expressed scepticism concerning the existence of spectroscopes. In a very few years, however, instincts' were banned from American psychology, and the revolt against instincts has even gone to the unwarranted extreme of putting a taboo on such perfectly good concepts as 'instinct' and 'instinctive,' which are really not bound up in the old 'instincts' at all.

Only a few psychologists, at the present time would base anything on 'instincts.' Apparently, the doctrine that 'an instinct' is nothing more than an arbitrary classification of responses in accordance with external results or external situations in which certain results may be accomplished, (i.e. a teleological class. name); and that even the results by which classification is made are arbitrarily selected by the classifier; is generally accepted except by those who hold to a spiritual doctrine of 'instincts' as energies put out by a substantial soul.

( 573)

After such a revolution, a logician might expect that emotions as uniquely different facts would have disappeared along with the 'instincts' to which they were obviously bound. This result, however, has not followed. Psychologists still talk of this 'emotion' and that 'emotion' as if they were sui generis; and the behaviorist has newly established behaviorism on the foundation of a few unique 'emotions' from which others are developed; the foundation' which had been demolished by psychology being thus replaced by more of the same material.

I have from time to time suggested that the situation is illogical. I want to insist that it is absurd. 'Emotions' can be named: we can speak for example, of anger, fear, and grief. But these names represent no psychological entities, no unique affective processes. They are names applied to varying complexes in the affective content, and they are teleologically chosen. The search for 'primary emotions' is as much an anachronism in psychology today as is the search for the soul: and it is a search of the same sort. We must face the fact that the 'emotions' are names to which correspond no concrete realities, but designate abstractions of the most arbitrary type.

Let us consider 'fear' as an illustration. What is 'fear'? Its outward expression, we already realize, runs the gamut of responses. It may be manifested in flight; in immobility; in fighting; in caressing; in anything. On the other hand, the same outward expressions are characteristic of many emotional conditions. One fights when angry; when afraid; when in love. The responses to different external situations may become somewhat differentiated with advancing age.: but the baby makes the 'same response to being dropped; to having a pin stuck in him; to being deprived of his bottle; to a violent shrill sound. No one but a behaviorist would be so rash as to conclude, with no evidence whatever, that the baby's 'emotion' is the same in all these cases. But, one will say with justice, these outward expressions are of slight consequence. It is true that fleeing 'fear' is different from fighting 'fear.' But the difference is unimportant as compared with the inner (bodily) states and processes. That, it will be said, is the same in all 'fear.'

Now just what are these inner states? I do not doubt that there are inner states: states of the muscles of the trunk, arms, limbs; states of the alimentary canal; states of the other tissues. I do not doubt that these states are experienced that they are stimulus-patterns which initiate responses. But are they the same, or measurably the same, in all 'fear?'

So far as objective experiment goes, the answer seems to be: No. Further; the states, or processes which do seem somewhat constant are also found in 'emotions' of quite different names. Shall we turn to introspection? Then the negative is still more emphatic. Different 'fears' do not feel the same. Not even 'fears' which arise in apparently similar circumstances.

So far as introspection goes, two 'emotions' of the same name are often more different from each other than they are from other 'emotions' of officially different names. This is not conclusive: introspection raises problems, never settles them. But the problem raised in this case will not down easily. If it is ever to be solved, experimental attacks must be devised to fit the problem, and not shaped along the lines of pseudo-problems. Emotion, however, is a fact quite aside from questions about 'the emotions.' We do get 'stirred up,' 'moved:' we feel. All introspection agrees on this point. What, then, do the names of 'the emotions' signify?

( 574)

Nothing more than the situations in which we are emotional. In one circumstance, there is a threat or danger. Something, some person, some occurrence, threatens to limit me, to injure me, to deprive me of something I value. This is a definite constellation and the emotion I experience in this situation, I call 'fear;' although the actual emotion may vary. But hold, wait a bit: I do not always call it 'fear.' Sometimes I call it 'anger.' Yes; but then certain features in my environment are emphasized, which are less conspicuous in the situation in which I call my emotion 'fear.' If a person threatens me, and I emphasize his part in the portended happening I may still 'fear' him, if I admit his greater power. Or I may be 'angry' at him. It all depends on the way I think about the situation, not on the actual state of feeling. In still other ways of apprehending the situation I may call my emotion 'resentment.' In still others I call it 'grief.' It is not for nothing that we may sob in any of these cases and even in 'joy;' for 'joy' is merely the name we give to any emotion which occurs when we are specifically conscious of additions made to life. The vital differences, so far as feeling itself is concerned, are those between calm 'joy' and ecstatic 'joy;' cold 'grief' and frantic 'grief;' chilly 'fear' and frantic 'fear.' The differences otherwise are minor. Frantic 'fear,' frantic 'joy,' frantic 'grief' are much more alike, as emotional states, than frantic 'grief' is like to depressed 'grief,' frantic 'joy' to quiet 'joy,' or excited 'fear' to cold 'fear.' That in which the various 'fears' resemble one another, the various 'griefs' resemble one another, and the various 'joys' resemble one another, is in the likeness of the situations in which these states arise, or rather, in the perceptual and ideational apprehension of these situations.

To recapitulate: The emotions as classified in daily life, and in our archaic psychology are classified and named in accordance with our estimation of the situations in which they arise. I may estimate a situation today as different from an extremely similar situation tomorrow. Then I call 'the emotion' in the two cases by different names, corresponding to the situations as estimated or apprehended.

I may apprehend my situation primarily as one in which my welfare is threatened. I call my inner responses fear. You may apprehend my situation as one in which I fight against the threat; you say I am angry. Which is correct? The answer may be: Both! You name my emotion correctly in so far as you connect it with any factor in the environment which is really efficacious in stirring me up, or any actual intellectual attitude of mine towards the environment, or on any demonstrable result of my emotion.

It is no wonder that experimental attacks on the problem of emotions have produced no results. Our problems have been formulated in teleological terms, although we have supposed we were studying emotion as such. Our methods have been determined by our misconception of the problems. We have eagerly sought, for example, for characteristic organic symptoms of the conventionally named 'emotions,' overlooking the fact that the differences for a single so called 'emotion' should be expected to be as great as between two differently named, 'emotions.' We have asked patient observers to report their 'emotions,' and have been misled by not knowing that the language in which the reports are given has mainly an environmental significance, and secondarily a cognitive reference.

( 575)

In studies of the 'personality' type, we gravely record such statements of the patient as that he feels 'more cheerful' today; that he feels 'depressed,' or 'hopeful,' etc., never suspecting that the reports have little bearing on the actual affective state, but signify primarily the patient's judgment of situations. As indications of the patient's intellectual attitudes, these reports are of great value. Even the answer to the question: "Do you feel depressed?" may be significant, in spite of the fact that the same person will answer "yes" this week and "no" next week, but the significance is intellectual and not emotional.

I have been much impressed by the statement, in a recent report, of the method employed and results obtained in ascertaining employees' emotional states. The investigator first guessed the individual's state on the basis of his behavior. Then the individual was asked to report his state. If the two characterizations agreed, the rating was accepted as valid. If they disagreed, the individual was questioned later in the day, when it was usually found that his report agreed with the investigator's judgment!

In various ways, we find individuals tending to report their emotional states in terms of the ways they think they should feel in specific circumstances, as well as ways which they deem appropriate to the external conditions (which do not always amount to the same thing). Even when verbal report is ruled out, and other types of responses are interpreted, the interpretations are confused by the fact that the same response means different things at different times.

When a person is whistling lively tunes at his work, is he intrinsically 'cheerful'? Unless by 'cheerful' we mean merely that he whistles, the significance of his behavior is ambiguous. Often he is depressed. But usually, he thinks he should be cheerful, in a more inner sense, or that he should be optimistic, bucked up or looking on the brighter side. The determinants of classification seem in the predominant number of cases to be not mainly the actual inner feeling, but the external circumstance, and the individual's judgmental attitude toward the circumstances.

In another direction our confusion has been more deadly. We have, during the last two decades, more and more emphasized the emotional factor in life, at the expense of the cognitive. We have insisted, for example, that the perceptual and ideational qualifications of the child are not paramount: that his emotional attitudes in the home, in the school, among his mates, are of profound importance in determining his development, his learning, his adaptations. We have made a great pother about this discovery, in education and child guidance, and have created over night a lot of 'experts' in directing the child's emotional development. But when the results are checked, we find that little progress has been made, and that little in defiance of the theories. We are really dealing in this field with matters of understanding, judgment, and thinking generally. Solomon is justified in his maxim: practical success has been attained by getting the child to understanding his situation, to think accurately, and adopt healthy ideals. As we drop the ballyhoo about 'emotional attitudes' and face the problems as they actually are, we shall make still more progress.

There are emotional states or processes (whichever you choose to call them). These emotional items, moreover, actually do vary. Neither of these statements can be proved: but they are good working hypotheses. That there are different 'emotions,' discretely distinguishable in kind from one another, aside from their

( 576) cognitive associations, is a different hypothesis, which time has shown to be sterile. The more plausible hypothesis is that the variation is of the type which would occur in a complex in which there are n elements, any one of which may vary in a graded way, so that the total variation is poly-dimensional but without discriminable steps or jumps. The value of n, and the nature of the elements, if any, remain to be determined. It is safe to say, however, that if affective elements are ever discovered, they will not be items to which we would apply the conventional names of 'the emotions.'

It would seem that the best attack on the problem of emotion (not 'the emotions'), should be through the study of actual variations, regardless of emotional names. As a matter of fact, beginnings have been made in this line, It is unfortunate that we have to make these beginnings through classification; but if we hold resolutely to the fact that classifications are only classifications, we may even so make classification a useful servant.

We have noted that certain emotions, regardless of names, are pleasant, others unpleasant; some are exciting, others are the reverse. There are obviously dimensions of emotional variation which are by no means independent of external conditions and results, or of ideational processes; but which we can actually study as psychological features, collating them with cognitive factors, and not confusing them. Perhaps if we forget the conventional teleological classifications and actually study their obvious dimensions, we shall learn something about them and discover other dimensions.

Such study as I urge must of course be psycho-physiological, rather than neurological. If we change the terms slightly, the James-Lange theory of 'the emotions' vanishes; but the modified James-Lange theory of emotion remains a permanent contribution to the armament of psychology. There is neither hope nor good sport in investigating emotions except as organic or bodily states; products of response to the environment, not intermediaries in response. The sooner we recognize that the traditional 'instincts' and traditional 'emotions' were really yoked together and that the passing of 'instincts' left 'emotions' so forlorn as to be hardly serviceable even for behaviorism and other reactionary psychologies, the sooner will our study of emotion become productive.

Johns Hopkins University  KNIGHT DUNLAP


  1. Presented before the Southern Society for Philosophy and Psychology, New Orleans, December 30, 1931.

Valid HTML 4.01 Strict Valid CSS2