The psychological construction approach posits that emotions are constructed out of more basic psychological ingredients that are not themselves specific to emotion. Two such basic components were proposed by Russell (2003): core affect and affective quality. “Core affect is that neurophysiological state consciously accessible as the simplest raw (nonreflective) feelings evident in moods and emotions” (p. 148). It is a single integral blend of two dimensions: pleasure-displeasure (which can range from elation to agony) and activation-deactivation (which can range from frenetic excitement to sleep). The feeling is an assessment of one’s current condition. Affective quality is a property of the stimulus: its capacity to change core affect. Perception of affective quality together with core affect allows a person to construct the emotion.
Barrett (2013) believes that psychological construction constitutes a paradigm for the scientific study of emotion that is different from the “faculty” psychology paradigm of the basic emotion and appraisal approaches. She points to three principles of psychological construction that define this difference: the principles of variation, of core systems, and of emergentism and holism. I will not discuss these principles here, but I will say that in many respects the details of these principles bear a striking resemblance to many of the ideas originally expressed by William James (and acknowledged by Barrett and Russell). Whether these ideas will actually change the way researchers on emotional issues behave, remains to be seen. Barrett also calls psychological construction the Darwinian approach to the science of emotion, primarily because pre-Darwin, species were considered fixed whereas post-Darwin, the variety within species could be exploited by natural selection and lead to new kinds. Although there are some similarities between the two approaches, I doubt that most biologists would be impressed with the analogy.
In another recent development, LeDoux (2012) has proposed rethinking the emotional brain in terms of survival circuits. The survival circuits proposed by LeDoux correspond almost exactly to behavior systems as I have defined them, although he places more constraints on which brain circuits would be considered survival circuits. His survival circuits are considered to be “innate” and to have functional [survival] significance for the organism. There are many problems with the concept of “innate” (see Chapter 7) and functional explanations cannot solve causal problems (Hogan 2015). LeDoux also notes that his list of survival circuits does not align well with human basic emotions.
In considering these various approaches to the study of emotion, I would propose that it is the activated behavior system that determines the quality of the emotion. The study of emotion then becomes the study of what behavior systems exist in any organism, what motivational factors activate them, and how they are expressed. Emotions are the subjective aspect of strongly activated behavior systems. A corollary of this conceptualization is that the felt emotion becomes an epiphenomenon: like the whistle of the steam engine, it has no causal significance—which is, of course, consonant with James’ viewpoint. Much of the research on emotion in the past 50 years can be understood in these terms.
Nonhuman Emotion
I have defined emotion as the subjective aspect (feeling) of strongly activated behavior systems. Since we have no access to the subjective experience of any animal (except ourselves), any discussion of nonhuman emotion must rely on investigation of the expression of such strongly activated behavior systems. One of the first systematic studies of the expression of the emotions in man and (other) animals was that of Darwin (1872). Darwin was primarily interested in similarities between animal expression of presumed emotional states such as anger, terror, and joy and human expression of these and other emotions. Darwin assumed that animals such as dogs, cats, horses, and monkeys had such emotional states and tried to show that the expression of these emotions in humans could be traced to their expression in various animals as support for his theory of evolution. Since we know that the nervous systems of all animals have similar components, it should be possible to infer the emotional state of an animal from observations of its behavior. We would be inferring the state of activation of an animal’s various behavior systems, irrespective of whatever subjective experience the animal might be having. In effect, we would be performing a motivation analysis (p. 62). We have already seen examples of this with respect to the upright posture of the herring gull, the zig-zag dance of the stickleback, and waltzing in junglefowl. A similar example is Lorenz’s (1966) analysis of the facial expressions of fear and aggression in dogs (Figure 3.8). In this figure, increasing aggression goes from left to right and increasing fear goes from top to bottom. In (a) the dog is calm and unemotional; in (b) and (c) it is becoming more afraid; in (d) and (g) it is becoming more aggressive. The other figures depict ambivalent expressions. It can be seen that as fear is increasing, the ears and the corners of the mouth are drawn backward and downward; as aggression is increasing, the upper lip is raised and the mouth opened. These examples show that it is possible to ascertain which behavior systems in an animal are activated. But how strongly does the system have to be activated in order to be considered an emotion? And how do we measure strength?
Figure 3.8 Facial expressions of fear and aggression in dogs. Explanation in text. (From Lorenz 1966).
Strength has been measured both behaviorally and physiologically, and recently, “cognitively” as well. An early investigator of “emotionality” in animals (the rat in this case) was Hall (1934). He showed that defecation and urination in a standard situation were valid measures of individual differences in emotionality. Hall considered emotionality a trait, characteristic of an individual. He felt that attempts to differentiate specific emotions were extremely speculative, a view still held by many investigators of human emotion, as we have seen above. Hall, as also most prior and subsequent investigators of animal emotion, was really interested in using animal studies as a model for understanding human emotion. And soon thereafter, many other measures of bodily changes in animals, both behavioral and physiological, began to be used in investigations of various aspects of emotion. Paul et al. (2005) review the various approaches to measuring emotional processes in animals, past and present, including new non-linguistic cognitive measures.
Most studies of animal emotion are directed to understanding human emotion, but the rise of interest in animal welfare has led many investigators to study animal emotion per se. In the context of welfare, it is crucial to discover what makes an animal ‘feel good’ (or, at least, not suffer). However, feelings are subjective and we can never know what an animal feels (see Panksepp 2010 and Dawkins 2015 for recent discussions of animal consciousness). M. Dawkins (2008) suggests that a scientific study of animal suffering and welfare can be based on answers to two questions: Will the situation improve animal health? And, will it give animals something they want? The answer to the second question can be determined by discovering what the animal finds positively and negatively reinforcing (what they want and do not want) in a learning situation. Even here, however, a difference between “wanting” and “liking” (Berridge 2004) makes interpretation of the results not straightforward (an animal may like something, but not want it at this moment). Nonetheless, Dawkins’ approach seems the most reasonable proposal to date. Theoretically, Mendl et al. (2010) have proposed a framework that integrates the discrete emotion approach (i.e., the basic emotion approach above) with the dimensional approach (i.e., the psychological construction approach above) for the study of animal emotion and mood. The cognitive aspects of the dimensional approach allow one to experimentally dissociate “liking” something from currently