David Konstan, Department of Classics
New York University, New York
Despite the intensive study of emotions across a variety of disciplines over the past three decades, it still seems difficult, if not impossible, to define precisely what an emotion is. We may take as illustrative a disagreement over the nature of anger. Whereas some investigators insist that “Anger is associated with justice concerns, or the protection of individual rights, fairness, and autonomy” (Horberg, Oveis, & Ketner, 2011), others maintain with equal conviction that “Neither personal anger nor empathic anger is a truly moral emotion or source of truly moral motivation” (Batson 2011). Such differences in the interpretation of specific emotions are reproduced at the level of the nature of emotion itself. Thus, a prominent student of emotions to lament: “Emotion researchers face a scandal: We have no agreed upon definition for the term — emotion — that defines our field. We therefore do not know what events count as examples of emotion and what events theories of emotion must explain” (Russell, 2012). A study of the history of emotion, which is itself a sub-discipline of the comparative study of emotions across cultures, may help explain why emotions seem to resist definition.
It is, I think, fair to say that the most influential ancient account of the emotions is to be found in the second book of Aristotle’s Rhetoric, where Aristotle surveys a set of emotions, or as he calls them, pathê (plural of pathos), that includes anger and the remission of anger, love, hatred, fear, shame, envy, gratitude, pity, indignation, and a competitive passion that he calls zêlos (the ancestor of the English words “zeal” and “jealousy”), along with mentions of a few others, such as over-confidence (the opposite of fear) and contempt (see Konstan, 2006). What is striking about this list is that it bears a plausible resemblance to what we today (writing in English, at all events) might subsume under the general category “emotion.” There are some outliers, to be sure: gratitude does not always make it into modern inventories of emotion, and pity, oddly enough, is also largely missing; in turn, some modern classifications include rather basic or semi-automatic responses such as disgust and surprise, or what we might think of as moods or states such as happiness and sadness, that are not registered by Aristotle as pathê. But given the range of variation in modern classifications, Aristotle’s itemization seems reasonable enough.
In his treatise on rhetoric, Aristotle provides a precise definition of each pathos, along with advice on how to arouse or assuage it – this is, after all, a handbook for orators, whose job it is to manipulate the passions of their audience. His definition of anger, for example, runs: “Let anger be a desire, accompanied by pain, for a perceived revenge, on account of a perceived slight on the part of people who are not fit to slight one or one’s own” (Rhetoric 2.2, 1378a31-33). Of course, Aristotle was speaking about the Greek term orgê, not “anger,” and we must be on our guard against an uncritical equation of emotion terms across different languages; as Anna Wierzbicka has observed, psychologists of the evolutionary school tend to “absolutize the English folk-taxonomy of emotions” (1999: 171). But allowing that the Greek term, as Aristotle defines it, comes reasonably close to the English “anger,” we may still note some specific features. First, Greek anger is stipulated to be a desire of a certain sort, namely for revenge; it is thus described basically as what we might call an action tendency. Second, Aristotle attends primarily to the stimulus to anger and neglects, in this context at least, either bodily states or facial expressions that may accompany it. This again is natural, perhaps, in a treatise on rhetoric. Third, and most remarkably, Aristotle restricts the causes of anger to a slight or belittlement; anger is not a response to mere aggression or pain, such as an attack by an enemy or stubbing one’s toe. Finally, there is the curious qualification that not all insults arouse anger, but only those coming from people who are somehow unsuited to deliver them.
Taking these several factors together, it is clear that for Aristotle, anger involves high-level cognition. Recognizing a slight or putdown involves an appraisal of another’s motives and intentions, as well as social status; revenge, in turn, is not simply striking back but presupposes a notion of reciprocity. This is why Richard Lazarus could observe that “those who favor a cognitive-mediational approach must also recognize that Aristotle’s Rhetoric more than two thousand years ago applied this kind of approach to a number of emotions in terms that seem remarkably modern” (Lazarus, 2001: 40; cf. Hinton, 1999: 6). Aristotle would seem, then, not only to have been the first thinker to specify the gamut of emotions in a modern way, but also a forerunner of the cognitive interpretation of emotion.
Aristotle offers a similarly cognitive account for the other pathê he treats. For example, he defines pity as “a kind of pain in the case of an apparent destructive or painful harm in one not deserving to encounter it,” and which, he adds, “one might expect oneself, or one of one’s own, to suffer, and this when it seems near.” Pity, then, involves a notion of one’s own vulnerability as well as a judgment concerning desert. Indignation is the reverse of pity: we experience it when we see others prospering undeservedly. We feel shame, Aristotle says, when an action reveals our vices: if we flee in battle, for example, it shows that we are cowards, and if we wrong another is a sign that we lack the virtue of justice. Envy is a response to the perception that our social equals are faring better than we are – again a judgment based on comparative worth. Fear too is not just an irrational desire to avoid harm but depends on an assessment of one’s own strength relative to that of one’s antagonists, and so depends on calculation. This is why the Stoics, who were heirs to Aristotle’s classification though they introduced some important changes, denied that animals other than humans can experience emotions in the full sense of the term. They may instinctively flee predators, for example, but since they do not make judgments concerning impending dangers their reaction is not strictly speaking a sign of fear. In one of his consolatory letters the Roman Stoic Seneca affirms that animals do not experience sadness or fear any more than stones do (Consolation to Marcia 5.1). So too, in his essay On Anger, Seneca affirms: “Animals have violence, rabidity, ferocity, aggression, but do not have anger any more than they have licentiousness…. Dumb animals lack human emotions, but they do have certain impulses that are similar to emotions” (On Anger 1.3).
But on Aristotle’s view, the emotions involve more than bare intellect or reason. His accounts of the several pathê indicate that the judgments or evaluations by which they are elicited have a specifically ethical or social quality. Pity and indignation depend on an assessment of whether the other is suffering or prospering deservedly, and envy, of which Aristotle disapproves, nevertheless involves social ranking. Shame presupposes a sense of virtue and vice, and the best kind of love, according to Aristotle, is elicited by an appreciation of a person’s character. Fear may be the least moral of the passions, since it is concerned primarily with one’s own security, but Aristotle notes that we tend to fear, according to Aristotle, people who are unjust or arrogant, who fear us or are our competitors, and those whom we have wronged or who have wronged us (2.5, 1382b8-9). Thus fear too is embedded in the world of social competition and ethical evaluation. And so too is anger: we are angry when we are treated in a way that is incommensurate with our social standing, by those whose status is inferior to ours. For Aristotle, emotions are not just isolable states of excitation but are essentially products of social interactions, and presuppose sensitivity to social status, merit, and morality.
It may seem, then, that Aristotle can be enlisted in support of the thesis of Horberg, Oveis, and Ketner, according to which anger “is associated with justice concerns,” even though we would have to substitute for “individual rights, fairness, and autonomy” alternative values such as honor and status – the morality of another time and place. But this may be to put the question the wrong way round. The word pathos in Aristotle’s time signified broadly any reaction to an external stimulus (it is related to the verb meaning “suffer” or “experience”), including physical conditions such as pain or disease, and by extension it could also denote “misfortune” (compare the English “pathetic”). As the first to narrow down the range of the term so as to refer uniquely to what look like emotions in the modern sense, Aristotle – writing in the context of a rhetorical treatise – selected just those affects that were relevant to the courtroom or the assembly, where one attempted to sway the judgment of the jury or the citizen body by argument and demonstration. This is in part why Aristotle’s analyses have such a cognitive cast: he was thinking of public disputations, not laboratory experiments, as the context for emotion. By the same token, the sentiments that Aristotle picked out for his catalogue of pathê were naturally of an interpersonal and evaluative nature. If Aristotle’s inventory of the pathê became the dominant model for what counts as an emotion today, influencing (at least indirectly) Latin treatments such as those by Cicero and Quintilian and via them, in turn, such eminent theorists of emotion as Thomas Aquinas, Descartes, and Adam Smith, it is not because Aristotle had correctly intuited the natural lineaments of the emotions, as though they were species awaiting their Linnaeus. Rather, we would do better to say that Aristotle produced what looks like the modern class of emotions, but did so for the wrong reasons. His pathê are cognitive and moral because those characteristics were the attributes or differentiae on which he based his selection.
I wish to suggest that the controversy over whether anger is a moral or non-moral emotion is an effect, however distant the cause, of the way emotions as a class first came upon the scene in classical antiquity, when morality was built into the very concept of a pathos thanks to the context in which it emerged. What is more, already in antiquity the complex cognitive and ethical conception of emotion was challenged in a variety of ways. For one thing, Aristotle himself had recognized the physical correlates of emotion, and it was no great leap to define emotions in biological terms, as functions of the blood or, in the prevailing medical view, of the dominance of one or another of the four basic humors. More interestingly, the Stoics, who went so far as to define pathê as conditional upon voluntary assent, also introduced the notion of pre- or proto-emotions (the Greek term is propatheia) that were instinctive reactions, independent of any judgment. These reflexes, which Seneca, in his treatise On Anger, calls “the initial preliminaries to emotions” (2.2.6), are defined as “motions that do not arise through our will,” and are therefore irresistible and do not yield to reason. Seneca provides a lengthy and, at first sight, rather puzzling list of these proto-emotions, which include such responses as shivering or goose-pimples when one is sprinkled with cold water, aversion to certain kinds of touch (presumably slimy things and the like), hair rising upon hearing bad news, blushing at obscene language, the vertigo produced by heights, responses to theatrical spectacles and narratives of historical events, songs and martial trumpeting, horrible paintings, and the sight of punishments even when they are deserved – note the specifically non-moral nature of this last reaction, which is presumably something like raw empathy. Although irrational animals are not capable of all these responses – as Mark Twain observed, Man is the only animal that blushes. Or needs to” (in Following the Equator, Pudd’nhead Wilson’s New Calendar) – giddiness and shivering are within their repertoire and presumably the ferocity and aggression they manifest when provoked, and which human beings too can experience, are pre-emotional reactions. The reaction of a human being to a sudden shove, for example, is not quite anger, at least until we judge that we have been belittled by whoever administered it, and in this respect is no different from that of a dog or cat to a similar incitement. We may see here, I believe, an intimation of the distinction that some theorists today draw between emotion and affect, with the latter understood as more elementary and universal reflexes as opposed to ostensibly higher-level cognitive responses. Here too, however, the Stoics’ motivation for positing the category of proto-emotions is very different from today’s arguments. The Stoics maintained that a sage was immune to passions such as anger and fear, since she or he knew that the usual reasons for taking offense or worrying about harm were inconsequential: only one’s own virtue really matters, and the wise were proof against the loss of that. But even a sage might turn pale when caught in a storm at sea, or tense up when jostled; ergo, these reflexes were not true pathê but something more primitive.
Aristotle’s works soon became the object of study in schools and academies, and extensive commentaries were written to explain and interpret them. Of these, the earliest to survive in something like a complete condition was composed by a certain Aspasius, who composed a commentary on the Nicomachean Ethics sometime around the mid-second century AD. When he comes to discussing Aristotle’s conception of the pathê, which he says neither Aristotle nor other of the Peripatetic school adequately defined, Aspasius both documents a controversy among earlier Aristotelians and offers an original interpretation of the pathê in his own right. Aspasius divides the pathê into two broad genera or classes, pleasure and pain, under one or the other of which he subsumes all the specific pathê, such as anger and fear but also the sub-classes of pleasure and pain themselves. According to Aspasius, the pathê may arise prior to any supposition (his word is hupolêpsis), directly as a result of perception, though in other cases a supposition may come first, in line. Either way, the pleasure and pain that are the generic features of any pathos, whatever the differentiae that distinguish them, take the form, Aspasius says, of motions in the non-rational part of the soul, and such a change is accompanied in turn by a corresponding motion in the body. Aspasius has a problem with the notion of appetite or desire (the Greek term is epithumia), since, he says, it may seem to partake of both pleasure and pain (we feel pain when we are hungry but pleasure at the anticipation of eating), and thus require a higher genus under which all pathê fall. But he finally concludes that desires too are either pleasant or painful, and not both. Indeed, he offers anger as an example of the painful kind, since it is defined by Aristotle as a desire for revenge, accompanied by distress. There is more to say about Aspasius’ theory (for those who are interested, I have translated the entire commentary into English), but we may note several features. First, it would appear that even though he was a loyal Aristotelian, Aspasius seems to have departed from the master’s account of the pathê by emphasizing what today is called emotional valence, that is, the positive or negative quality or feel. Aristotle had affirmed that the pathê are accompanied by pleasure and pain, and indeed he offered anger as an example, inasmuch as the anticipation of revenge is pleasant. In Aspasius’ account, however, pleasure and pain are the primary categories (we may note that the Stoics had posited four classes of pathê: pain, pleasure, fear, and desire). Coordinate with this new emphasis, the role of judgment, with its attendant moral and social presuppositions, was at least partly diminished. This, at least, is how I understand Aspasius’ insistence that the pathê may arise independently of supposition or belief. Although I am not unreservedly given to social determinism in respect to intellectual developments, it may not be irrelevant to observe that Aspasius proposed his theory when Greece was under the sway of the Roman Empire, as opposed to the freewheeling Athenian democracy where Aristotle wrote his treatises. Perhaps, as the law courts and political assemblies declined in importance, thinking about emotions moved away from an emphasis on merit and social position and began to look more to the way emotions feel – more in line, it may be, with modern approaches.
There were other developments as well in the way emotions were conceived in classical antiquity. The rise of mystical thinking, associated with revivals of Platonism and Pythagoreanism under the Roman Empire, brought about a new concern with such sentiments as wonder, awe, and ineffable experiences of joy and amazement or even shock (the Greek word is ekplêxis, from which “apoplexy” is derived), along with the contrary feelings of gloom and despair. Both these extremes were further accentuated in Christian texts, with their radical division between the saved and the doomed. The Byzantine statesman and scholar Theodore Metochites, who lived into the fourteenth century AD, enthused about the pleasures of gazing at the heavens, “and how, when the sky is clear, each sight everywhere brings not only wonder but also joy to the roaming eyes, not only inspiration but also a mood that gladdens and sweetens the heart” (Sententious Remarks 43). Once again, it seems that not only did the theory of the emotions change and evolve, but the very conception of emotion and the nature of the items included under this description altered and, in the process, left their mark upon future thinking about this elusive and historically malleable category.
We cannot look to history to resolve today’s dilemmas concerning the emotions, and the extent to which they are moral or cognitive in nature, although earlier theories may provide illuminating insights. What the history of emotion reminds us of is that the very term is variable, and subject to some degree to the prevailing social conditions of any era, including our own. There are doubtless certain constants over time in the sentimental repertoire, such as (perhaps) some of the primitive reflexes that the Stoics called proto-emotions, although one must be careful not to assume that even such elementary responses, not to say more complex pathê, can be mapped precisely onto their ostensible modern equivalents (to the extent that these are uniform). Perhaps the problem with emotion theories today, to which James Russell called attention, is not that we have “no agreed upon definition” of the term, but that we have many, each of which is suited to a particular context and purpose. This is not a bad thing, so long as psychologists and others are aware of the way in which the theory constitutes the object, at least in large measure. This is one of the lessons that the history of emotions can teach us, as it invites us to observe changes in approach and indeed the very object of investigation, and to test our own presuppositions about what constitutes an emotion.
References
Batson, C. D. (2011). What’s Wrong with Morality? Emotion Review 3: 230-236.
Hinton, A. L. (1999). Introduction: Developing a Biocultural Approach to the Emotions. In Alexander Laban Hinton (Ed.) Biocultural Approaches to the Emotions. Cambridge: Cambridge University Press: 1-37.
Horberg, E. J., Oveis, C., & Ketner, D. (2011). Emotions as Moral Amplifiers: An Appraisal Tendency Approach to the Influences of Distinct Emotions upon Moral Judgment. Emotion Review, 3, 237-244.
Konstan, D. (2006). The Emotions of the Ancient Greeks: Studies in Aristotle and Classical Literature. Toronto: University of Toronto Press.
Lazarus, R. S. (2001). Relational Meaning and Discrete Emotions. In Klaus R. Scherer, Angela Schorr and Tom Johnstone (Eds.) Appraisal Processes in Emotion: Theory, Methods, Research. Oxford: Oxford University Press: 37-67.
Russell, J. A. (2012). Introduction to Special Section: On Defining Emotion. Emotion Review, 4: 337.
Wierzbicka, A., (1999). Emotions across Languages and Cultures: Diversity and Universals. Cambridge: Cambridge University Press.
