A Physiological-Genetic Theory of Feeling and Emotion

Floyd Henry Allport

Harvard University

Since the important work of Dr. Cannon, published in 1915, there has been little progress in the study of emotion. The James-Lange theory, which has retained the merited esteem of psychologists, has waited long for confirmation at the hands of physiologists. Yet the latter have been inclined to interpret their discoveries as hostile rather than favorable to it. Perhaps the main defect of the James-Lange statement is one of omission. It fails to differentiate the patterns of visceral or somatic response whose return afferents might give a basis for the different emotions felt. It is the purpose of this paper to propose a theory for differentiating the emotions physiologically, and to clarify if possible the obscure basis of pleasure and unpleasantness.

Although introspection upon emotional states is difficult, there is probably a general agreement that all emotions subjectively comprise (1) an affective element, that is they are either pleasant or unpleasant, and (2) a differentiating factor (x) which serves to distinguish between emotions which are alike in respect to the affective component. On the pleasant side the x factor might distinguish joy from love, or conjugal from consanguineal love. On the unpleasant side it would differentiate such states as anger, fear, and grief. We shall take the stand that both these components (the affective and the differential) are made up, according to the James-Lange principle, of bodily sensations resulting from diffuse patterns of response to the stimulating object. The response patterns for the affective and the x components will now be considered in turn.


I. If we search the body for a neuromuscular mechanism suitably correlated with the antagonistic poles of pleasant and unpleasant, we shall find such a mechanism in the autonomic nervous system and the viscera which it supplies. The physiological antagonism lies between the two divisions of the autonomic, namely, the sympathetic (sometimes called the thoracicolumbar) and the craniosacral. As is well known from the work of Cannon and earlier investigators, the sympathetic motor effects include checking of digestion and sex activity, accelerating the heart beat, constricting the blood vessels, and reinforcing its own activity by the liberation of autacoids. All of these changes are in opposition to those produced by the craniosacral efferent fibers. Now all the emotions studied by Dr. Cannon and reported as involving sympathetic impulses are clearly unpleasant in their affective value. This fact did not receive the emphasis it deserved in interpreting these brilliant investigations. Dr. Cannon himself laid stress upon the apparent identity of the bodily changes in all the major emotions as an argument against the James-Lange theory.[2] We, however, may justly emphasize this very identity as a support of the theory, since it provides a common mechanism for precisely that element which anger, fear, and pain do have in common, namely unpleasantness. The question of a physiological basis for the differentiating factor is, of course, still left open. We return to it presently.

Since the sympathetic division appears thus intimately concerned in unpleasantness, we may regard its antagonist, the craniosacral, as the foundation of the pleasantly toned pattern responses. The craniosacral division innervates those responses which follow immediately upon the attainment of objects of food and sex. Digestion and sex and breeding activities depend upon its unimpeded efferent impulses. The resident sensations arising from the performance of these activities constitute the two primordial pleasures of mankind. Both the craniosacral and the sympathetic divisions are subject to a vast amount of conditioning, so that with in-

(134) -creasing age the number of affectively toned objects and situations (that is, those evoking affective responses) is enormously increased.

There are, of course, other pleasures beside those of food and sex. We might mention bodily exercise and habit, and also the excitement of games (without anger or worry), as sources of pleasure not clearly involving. the craniosacral division. We may be obliged therefore to include in our pleasure-giving mechanism the free-functioning of the somatic effectors through the innervation of the cerebrospinal system itself. It should be noted also that the craniosacral is more closely related to the cerebrospinal system than is the sympathetic. Its fibers do not enter the ganglion chain system, but comprise really cerebral and spinal nerves whose fibers turn inward and supply smooth muscle.

The first part of our theory then states that emotions are fundamentally distinguishable as pleasant and unpleasant, and that these two contrasted qualities are related to two mutually antagonistic processes whose effector movements in the viscera set up characteristic organic sensations which we call affective quality. The craniosacral division of the autonomic, supplemented under certain conditions by the cerebrospinal system, innervates those responses whose return afferent impulses give rise to pleasantness. The sympathetic division produces visceral responses which are represented in consciousness as unpleasantness.

Aside from experimental knowledge of the autonomic functions, this portion of the theory receives further support from introspective observation.

First, unpleasantness is both more readily identifiable by introspection and more imperative than pleasantness. The unpleasant emotions are also more numerous and more characteristically emotional than the pleasant. This is readily explained by the fact that the sympathetic motor impulses (owing to the resistance of the ganglion bodies) are necessarily stronger than those of the craniosacral division whose action they inhibit. They are also more widely diffused through the viscera, and they support somatic motor activities of a violent and characteristic sort.


Secondly, if unpleasantness results from the effects of sympathetic motor impulses, we should expect that it would be slower of arousal than pleasantness. The delay in the ganglionic synapses, the long stretch of unmedullated postganglionic fibers, and the diffuse character of the sympathetic innervation (as contrasted with the opposite conditions of the craniosacral) would require a considerably longer transmission time for the sympathetic impulses. Our common experience attests the longer latency of unpleasant feelings as compared with the immediate thrill of pleasure derived, for example, from gustatory or erotic stimuli. The case of stumbling in going down stairs is a good example. In the writer's experience there is a sudden reflex recovery of balance; and then, when several steps have been descended, there wells up gradually a mass of unpleasant organic sensations. Annoyance and anger also have a long latent period. A characteristic non-emotive 'fore-period' has been found in extensive collections of introspection upon anger.[3]

Thirdly, recent careful investigation indicates that pleasantness and unpleasantness never occur simultaneously, that is, that true `mixed feelings' do not exist.[4] This fact is in harmony with the sharp antagonism of autonomic functions which we have assumed to be the basis of these feelings. The more complex states also, such as fear and rage, are in strict antagonism with the love emotions. The sex drive, reciprocally, is one of the strongest agencies in the dispelling of anger in family quarrels. To sum up, we find our first proposition supported on the introspective side by definiteness, imperativeness, localization,[3] latent period, and unmixed quality of the affective state.

II. We may return now to the x factor which serves to differentiate the various emotions within a single affective

(136) class. Accepting Dr. Cannon's dictum of homogeneity of the visceral changes, we propose that the sensations which comprise the differentiating factor arise from stimulation of the proprioceptors in the muscles, tendons, and joints of the somatic part of the organism. Kinaesthetic impulses resulting from the emotional response deserve as true a place as organic impulses in the conscious fusion. Different, and in many respects antagonistic, somatic effectors are brought into play according to whether the individual attacks or flees. Facial expressions as well as bodily movements are strongly differential. The view proposed is that afferent impulses from these somatic response patterns add the characteristic complex of sensations through which an emotion of fear is different in its consciousness from an emotion of anger. The cerebrospinal or projicient activity of adjustment to the situation thus supplements in consciousness the sympathetic core of pure feeling.

Let us appeal again to familiar experiences for elucidation of our second proposition. Attitudes are generally taken, or overt responses quickly made in nearly all emotional situations. In repression, where all somatic response is inhibited, the emotion is simplified and reduced to its foundation of visceral unpleasantness. The unpleasantness is probably increased by the very blocking of the somatic outlets. We may consider also the temporal relations of the two components in emotion. In the case of stumbling on the stair the starting response was completed before the sympathetic affective component was felt. The emotion therefore was not true fear, but simply an intense unpleasantness. To take another example, when the objective situation causing anger or embarrassment has been removed, the visceral component, being more sluggish than the somatic, outlasts the latter in the form of a purely unpleasant affective state which delays the full recovery of composure. If a wild animal is held in the hand, or if a child is pursued and brought to bay, there will follow a quick alternation or succession of attempts at escape (whose emotion is fear) and desperate attack (with the emotion of anger). The shift from intense

(137) fear to intense rage is too sudden to admit of a complete change in the visceral pattern. We may quite plausibly attribute it to the quick change in the response pattern of the striped muscle, superimposed upon the constant visceral undercurrent of unpleasant affectivity.

If some reminiscence of James's arguments be permitted, we may recall the impossibility of separating introspectively the somatic factors, the frown, clenched fist, etc., from the emotional experience. When we abstract from these only affectivity remains. As to the pleasant emotions, we may assert that the differentiating factors, for example in the various types of love, reside in the habits of adjustment toward the loved object. To love a baby is to have the tendency, at least, to fondle it in a lover-like fashion. This is an abridgment of the whole behavior repertory which affords the full emotional value of sexual love. In friendship the somatic component may be reduced to a touch of the hand or a half embrace. Some facilitation of the sacral and allied mechanisms probably forms the pleasant affective core of all these experiences.[6]

III. The theory which has just been advanced seems to be sustained by certain facts of genetic development which may be briefly summarized. The emotional states of the new-born babe appear to be undifferentiated. Judging of course from behavior alone they have no further character than pure unpleasant affectivity. The somatic responses are the same whether the state be caused by hunger, internal pain, or intense visual or cutaneous stimuli. The autonomic apparatus is prepared for its work at an earlier date than the

(138) cerebrospinal reactions which contribute the differentia of emotions. We may call this simple unpleasant experience of the new-born the 'protopathic' emotion.[7] The affective component which we have recognized as the fundamental basis of classification is thus the earliest and most primitive ingredient of human emotion. Within a short time the child gradually brings into play the various somatic responses of the pre-potent, or instinct type, such as rejecting, struggling, and withdrawing. When the afferent impulses, for example, from struggling against a blanket or an impediment to nursing are added to the protopathic core of unpleasantness we may assume that the true anger emotion begins to be felt. With overt movements of attack and defence the differentiating factor is thus added to the sympathetic pattern.

IV. A fuller comprehension of the theory may be obtained by reviewing the neural conditions for the arousal of unpleasant emotion. These conditions are necessarily those which help in breaking through the high resistance of the sympathetic and sending inhibitory impulses to the smooth muscle. (i) The first is that of the intensity of the stimulus. Almost any sensation becomes unpleasant if it is made sufficiently intense for the energy of the impulse to cross the sympathetic threshold. The peal of thunder often continues to arouse fear throughout adult life. Our theory offers a good basis for distinguishing physiologically between pains which are unpleasant and those which are not. Unpleasant pains are severe ones: their afferent impulses are powerful enough to break through into the sympathetic. The same consideration explains the pseudo-emotional quality often ascribed to intense bodily pain. (2) Repetition or insistence, such as the neural summation of petty annoyances in producing anger, is another condition favoring the arousal of unpleasant emotion. (3) Suddenness of the stimulus, or lack of proper somatic adjustment of the cerebrospinal system often causes the impulse to receive its outlet through sympathetic efferents. The fear aroused by the strange, the uncanny, or the extra-

(139) ordinarily large, belongs in this class. (4) Blocking of the usual somatic responses to the powerful drives, such as those of food and sex, usually through social agencies, is a potent factor in bringing about an invasion of the sympathetic. (5) Finally, the state of visceral tonus or preparation may be an important factor in lowering the sympathetic threshold and increasing emotionality. Cynicism, irritability, and other emotional attitudes indicate a permanent lowering of the resistance. Transitory effects, or moods, also increase susceptibility to fear or anger.

V. A theory of the kind we have outlined must not be considered as excluding the possibility of characteristic cortical processes in emotional and affective states. The interoceptive and proprioceptive impulses may of course be elaborated in specific ways upon reaching the thalamus or cortex. Differences in conduction such as facilitation and acceleration may also be characteristic.[8] We are here simply pointing out certain receptor and afferent processes, resulting from bodily changes, which furnish the data for emotional consciousness in the same way that the retina, affords the raw materials of visual experience. The facts of behavior and physiology demand, however, that we lay due stress upon the reaction patterns which characterize both the fundamental affective component and the somatic differentia of the various emotions.


  1. Read before the American Psychological Association, December 28, 1921. 132
  2. Cannon, W. B., `Bodily Changes in Pain, Hunger, Fear, and Rage,' 275-283.
  3. Richardson, R. F., `The Psychology and Pedagogy of Anger,' Educ. Monographs, no. 19.
  4. Young, P. T., ‘An Exper. Study of Mixed Feelings,' Amer. J. of Psychol., 1918, 19, 237-271
  5. Young, P. T., `The Localization of Feeling,' Amer. J. of Psychol., 1918, 119, 420-430. Though Dr. Young seems to consider the feeling quality unlocalizable, the reports of his subjects indicate a considerable consciousness of visceral reaction in U., particularly in the region of the stomach.
  6. There is a similarity in regard to the use made of antagonistic mechanisms between the theory presented here and that of Watson ('Psychology from the Standpoint of a Behaviorist,' pp. 250-251). Positive reaction tendencies, which we may regard as a ' Watsonian euphemism' for pleasure, are said to be due to the tumescence and preparation of the sex effectors; negative reaction tendencies (unpleasantness) are ascribed to the opposite effects (detumescence, etc.) in these organs. The physiological antagonism in these two processes is obvious though not particularly stressed by Watson. The present writer believes that this theory, though correct in principle, stops short of its goal in neglecting the rôle of the remainder of the autonomic apparatus and visceral effectors. Dr. Watson also fails to bring his `reaction tendencies' into their broader setting in the emotions.
  7. Being careful, of course, not to confuse this meaning of 'protopathic' with that now current in the English psycho-analytic school.
  8. For example, see Troland, L. T., 'A System for Explaining Affective Phenomena,' J. of Abnorm. Psychol., February, 1920.

Valid HTML 4.01 Strict Valid CSS2