Navigation – Plan du site

AccueilL’Atelier du CRH16Penser les émotions en Occident :...Their Emotions and Ours: A Single...

Penser les émotions en Occident : de l'Antiquité à nos jours

Their Emotions and Ours: A Single History?

David Konstan


Y a-t-il une histoire de l’émotion dans la longue durée ? Si les émotions sont des invariants à travers l’histoire et les sociétés, du passé au présent, dans ce cas (c’est-à-dire indépendamment des conditions spécifiques qui les suscitent) elles n’ont pas d’histoire en tant que telles. En revanche, si les émotions sont culturellement construites, et donc soumises au changement, est-ce l’émotion particulière qui évolue ou bien une nouvelle qui remplace l’ancienne ? Évidemment on peut affirmer que certaines réponses primitives, quasi-émotionnelles, sont transculturelles, tandis que les émotions complexes sont construites à partir de ces affects élémentaires, et changent d’une culture à l’autre. Cet article offre deux études de cas de l’évolution émotionnelle – la transformation de la pitié en compassion et celle du regret en repentance – et propose, pour conclure, l’idée que nous pouvons tracer des schèmes ou des convergences à travers une série de sentiments, lorsqu’une matrice émotionnelle se dissout et une nouvelle se cristallise.

Haut de page

Texte intégral

  • 1 Barbara H. Rosenwein, “Theories of change in the history of emotions”, in Jonas Liliequist, ed., A (...)

1In a recent paper entitled « Theories of Change in the History of Emotions », Barbara Rosenwein observes that although there have been many fine studies of emotions in those moments when « the emotions of a period are stable », these studies suffer from the drawback that « they are largely unmoored from any larger narrative ».1 But this deficiency is not easily remediable, for, as Rosenwein goes on to note, just those « larger narratives – studies that attempt to account for moments of change and thus span a greater time period – are as yet unsatisfactory or incomplete » (p. 7). Hence, we continue to face the question of how to write a history of emotions for those periods in which « the emotional tenor of a period is transformed », or, more generally, over a time span in which deep changes in the nature of the emotions are understood to have occurred. Rosenwein surveys with a critical eye the deficiencies in the evolutionary view of Norbert Elias, which suffers from grave inadequacies when it is applied outside a limited time period and area, as well as William Reddy’s notion of « emotives » and the idea that the agent of historical change is « emotional suffering » (pp. 10-12), and, finally, with rare objectivity, her own concept of « emotional communities », which, as she points out, covers, like Reddy’s approach, « a very short time period », and « largely fails to discuss the mechanisms of change from one dominant emotional community to another » (p. 14). She also rejects the wider applicability of Jan Plamper’s model, according to which theories of emotion contribute decisively to changes in emotional paradigms, offering as a counterexample the system developed by Thomas Aquinas, who

[…] explained the emotions in a coherent and dazzling manner, making new and original use of Aristotle’s natural physics of motion.... However, he did not catalyse an emotional transformation: his work endorsed the existing order rather than supplying a new paradigm.

It is tempting to look for reasons for change outside the emotional economy of a society in itself, and to see the history of emotions as an epiphenomenon of transformations in the larger social order, whether as a function of economic change, as Marxism might have it, or in other ideological domains, such as religion. But Rosenwein concludes that such change may be a function of multiple factors. And yet, as she says:

Even when we put all these factors together, however, we do not gain a sense of the general shape of emotions’ history over the long haul » (p. 19),

and she concludes:

The shape of a new general narrative of the history of emotions is likely to be less like a round worm – bipartite, with a tail and a head – than like a centipede – with multiple segments (p. 20).

2How, then, shall we proceed to write the history of emotions over the long run, when deep changes in register may have taken place in one or another sentiment or even across the board? One could make a beginning by tracing the evolution of specific emotions, following their turns and changes over the centuries. But the history of concepts or ideas, and this includes the history of emotions, presents a special problem in respect to historical studies. If an idea is to have a history in any strong sense, then it must have undergone some kind of change over the course of time. If it is absolutely stable and immutable, like an animal that has not altered over the past several thousand years, then one can speak of changes in its relation to its environment, perhaps, but not alterations or modifications in the idea itself. In the same way, a given species of insect, for example, is not said to have undergone a modification just because it has changed habitat, but only if it has adapted biologically to the new conditions. In the case of a notion such as fear or anger, we may observe changes in their frequency, intensity, and specific stimuli over time, depending on the social and even physical context: a community that is constantly at war and has cultivated a military ethos may have a different relation to fear than one accustomed to peace, and people who live where diseases are rampant will presumably experience anxiety over illness in ways different from those who occupy more healthful locations, or who have eliminated pestilences through medical advances. But the record of such changes constitutes a history of fear itself only in a derivative sense; for a proper history of fear as such, we should wish to see some sense in which the emotion itself has undergone a transformation. As is well known, an entire school of thought in the modern investigation of the emotions maintains that emotions, or at least what are called basic emotions, are culturally invariant, whether across societies today or in relation to past societies for as long as there are records and indeed throughout the entire existence of homo sapiens. Those committed to this view may wish to know how given emotions were regarded in other times and places, but will not be moved to wonder whether anger or fear as such are different now than they were there and then.

3There is of course another school that regards the emotions as culturally constructed, at least in large part, and hence allows that they may differ from one society to another. But this approach too faces a difficulty when it comes to tracing a history of emotion. For if indeed emotions are subject to change, and precisely for this reason can be said to have a history, how do we decide that we are dealing with one and the same emotion over time, rather than a more fundamental transformation by which something quite new has emerged? In the short run, where the language is continuous and no great ruptures in the culture have occurred, one can track how usage has evolved and even provide some explanation of the causes, as Rosenwein observes. But when shifts in sense and sensibility have been monumental, it may turn out that the new emotional lexicon is incommensurable with the old. Is modern « anger » like ancient Latin ira, or is it more like indignatio, for instance? Is classical verecundia or pudor equivalent to modern « shame », or is it more like « modesty », and if the latter, does it count as an emotion at all? Or again, let’s take classical misericordia: its most natural rendering is « pity », but it is sometimes translated as « mercy » or « compassion » or « sympathy », or even « clemency », though this last might seem to correspond better to Latin clementia. If one wished to write a history of compassion in the Latin West, would it be reasonable in fact to take as its starting point the classical use of misericordia at all? Perhaps some other word is better suited, or perhaps compassion in the modern sense was not in fact part of the classical emotional repertoire at all. Some criterion is needed by which to establish how the ancient and modern vocabularies are related, if indeed they are.

4One approach might be to identify elements in an emotion that are stable over time, even if the complex emotion itself undergoes a radical alteration. Jacques Derrida put his finger on the difficulty in his essay on friendship:

  • 2 Jacques Derrida, Margins of philosophy, trans. A. Bass, Chicago, University of Chicago Press, 1982, (...)

While acknowledging the specific function of a term within its system, we must not, however, take the signifier as perfectly conventional. Doubtless, Hegel’s Idea, for example, is not Plato’s Idea; doubtless the effects of the system are irreducible and must be read as such. But the word Idea is not an arbitrary X, and it bears a traditional burden that continues Plato’s system in Hegel’s system.2

This is all very well, but much too vague. If we take emotion terms like fear or anger, as opposed to a philosophical calque like « idea », what authorizes us to say that the signifier is not « perfectly conventional », especially when the ancient term may have been expressed in a different language, and hence cannot count as the same signifier at all?

  • 3 See David Konstan, « Y-a-t’il une histoire des émotions? », in Philippe Borgeaud, ed., Mythes, rite (...)

5One solution, with which I have myself experimented, involves what we may view as a combination of the two modern approaches mentioned above, namely the innatist school and the cultural constructionists. On this view, the emotions not only change over time but may give way to new emotional regimes and vocabularies that are distinct from the old; nevertheless, certain components of the emotions remain stable, and it is these that permit us to trace an evolution in the emotions themselves over time. The argument presupposes that there exist instinctive or quasi-instinctive responses that are common to all human beings and are shared even by some of the higher species of animals: these include such affects as sympathy with fellow creatures, attachment to the young and the corresponding sense of loss when they are absent, aggressiveness in response to threat, or else a shuddering accompanied by the desire to flee. These elementary reactions are not quite emotions, though they resemble emotions. Thus, brute sympathy is something like pity or compassion, attachment is similar to love, loss is like grief, aggression like anger, and the urge to flee is analogous to fear; but none of these innate reactions is an emotion. The primitive responses are analogous to the innate or basic emotions posited by the universalist school, whereas the complex emotions that are constructed on the basis of these elementary affects may well vary from one society to another, as the social constructionists affirm.3

6Indeed, something like this is the view that Seneca takes of the emotions, as I understand him. In the first book of his On anger, Seneca offers a preliminary definition of the emotion, and explains:

  • 4 Aristotelis finitio non multum a nostra abest; ait enim iram esse cupiditatem doloris reponendi, 1. (...)

Aristotle’s definition is not far different from ours [that is, the Stoics’]; for he says that anger is the desire that pain be paid back.4

  • 5 contra utramque dicitur feras irasci nec iniuria inritatas nec poenae dolorisue alieni causa; nam e (...)

Seneca indicates that there are some differences between the Stoic and the Peripatetic definitions, though it would, he says, be a long story to run through them in detail. « It is objected to both definitions », he continues, « that wild animals grow angry but they are not stirred up by a wrong [iniuria] or for the sake of punishment or pain to another; for even if they accomplish this, it is not what they intend ».5 To this, Seneca replies:

We must affirm that wild animals, and all creatures apart from human beings, are without anger; for since anger is contrary to reason, it does not arise except where reason has a place. 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.

  • 6 For discussion of this passage, see Fabio Tutrone, Filosofi e animali in Roma antica: Modelli di an (...)

Seneca goes on to observe that animals can utter sounds, but they do not have language; their perceptions, moreover, are muddy and confused: « thus, their attacks and outbreaks are violent, but they do not have fears and worries, sadness and anger, but rather things that are similar to these » (1.3.4-8); so too, in the Consolation to Marcia (5.1), Seneca states that animals do not experience sadness and fear any more than stones do. True emotions, on the contrary, require rational assent, and since animals, according to the Stoics, at least, do not possess reason, they cannot experience emotions in the strict sense of the term. In his more precise definition of anger in Book 2 of On anger, Seneca states that it is roused by the presentation or appearance of an offense, but he immediately adds that the emotion does not follow automatically upon the presentation itself, but only when the mind gives its assent. Thus, for anger to arise, one must receive the impression of having endured an offense and must desire to avenge it, and must further join two judgments together, namely that one ought not to have been harmed and that one ought to seek revenge. The immediate impulse, Seneca explains, is a simple thing, whereas a true emotion is compound and contains several elements: one has recognized something, become indignant, judged it to be an offense, and now seeks revenge, and all this cannot occur unless the mind has given its assent to what has struck it. And because animals cannot do this, they do not in fact get angry in the proper sense.6

  • 7 See David Konstan, The emotions of the ancient Greeks: Studies in Aristotle and classical literatur (...)
  • 8 For detailed discussion, see David Konstan, “Reason vs. emotion in Seneca,” in Douglas Cairns and D (...)

7This deeply cognitive view of emotions opens up a space for changes in the nature of the emotions over time, as opposed to innate or instinctive reactions, although Seneca does not pursue this possibility. Insofar as anger is defined as a desire for revenge – the point on which Aristotle and the Stoics agree – one might argue, for example, that this description is not necessarily universal but is rather characteristic of an honor-based society, such as that of classical Greece and Rome, where vengeance plays a central role.7 People do say today that they are angry because they have stubbed their toe on a chair, for example, yet they can hardly desire revenge, unless perhaps they personify the offending piece of furniture. Anger would thus seem to be different for us than it was for Aristotle and Seneca, since it does not involve an intentional offense but can be aroused by simple pain or frustration (frustration itself is an interesting instance of an emotion, if it is one, that is characteristically modern). What both the ancient and the modern versions of anger have in common is the substrate of “violence, rabidity, ferocity, aggression”, as Seneca puts it; in fact, some contemporary views of anger treat it as very like bare animal ferocity, in Seneca’s description, which is why anger commonly appears as one of the basic or universal emotions in modern inventories. What is common to the modern and ancient conceptions of anger – if I may generalize in this way for the sake of the argument – is that they share the underlying affect, which in today’s language has come to represent almost the whole of the emotion. We thus have continuity over time in the affect that subtends anger even as we recognize diversity in the emotion itself. This kind of analysis can of course be extended beyond anger; in fact, I believe that there is evidence to suggest that Seneca himself supposed that for every emotion there was an elementary affect corresponding to it, in the same way that aggression does to anger.8 If this is right, then Seneca may offer a model for how to write the history of emotions generally: it is just the story of how certain basic impulses or affects that are instinctive and non-rational assume the character of emotions in the full sense of the word under determinate, but potentially varying, social conditions. The affect abides through the changes in the conception and experience of emotions across different eras and cultures.

8It is interesting to observe that a similar correlation between certain instinctive responses that human beings share with animals and fully realized human emotions has been proposed in the context of modern experimental psychology. Thus, Gerrod Parrott suggests that we may

  • 9 W. Gerrod Parrott, “Ur-Emotions: The common feature of animal emotions and socially constructed emo (...)

use the term ur-emotion to refer to the commonalities shared by otherwise different emotions of various species.9

Parrott goes on to observe that

there are many differences between the emotions marah (in Indonesian), ikari (in Japanese), song (in Ifaluk), and anger (in English), but in all of them the ur-emotion of antagonism is evident – all four are aimed at an object that is appraised as interfering in some way with one’s concerns, and all four give rise to a motivation to stop that interference in different, culturally specific ways.... The recognition of these components across cultures leads to the intuition that there is something universal about emotions, but it is a mistake to suppose that there exist universal ‘basic emotions’ – marah, ikari, song, and anger are not the same emotion! Rather, it is the presence of the ur-emotion of antagonism that provides the intuition of universality. (p. 248)

9There are, however, some evident difficulties with the above view. First, there is no agreement today on what might constitute a basic set of affects that all human beings and even many animals may be said to share, and hence no apparent way to relate them to emotions in the proper sense, whether on a one-to-one basis or as combinations of more elementary feelings. Thus, Jaak Panksepp has identified what he describes as the “primary-process lexicon and neural data base for at least 7 primal emotional neural networks”, which he lists as

seeking, rage, fear, lust, care, panic/grief, and play, all arising from ancient brain circuits which are probably critically important for the vast varieties of tertiary-process emotions that humans readily discuss, without much neuroscientific understanding.

Panksepp goes on to remark:

  • 10 Cited from correspondence via the list-serve of the International Society for Research on Emotions (...)

Perhaps it is simply the lack of ability or desire to integrate the mass of existing evidence from the other animals into our understanding of the human animal which, because of its dynamic and flexible linguistic abilities, makes so much of human emotion research currently a “language game” where rigorous scientific definitions of higher emotional processes become forever debatable, perhaps next to impossible. Such difficulties do not hamper primary-process animal research where there already exists a conceptually clear primary-process lexicon.10

  • 11 “Afterword”, in Charles Darwin, The expression of the emotions in man and animals, 3rd ed., with In (...)

In turn, the doyen of the evolutionary approach to the emotions, Paul Ekman, who has published among other things a revised edition of Darwin’s path-breaking study, The expression of the emotions in man and animals, lists the fundamental emotions that are universally identified on the basis of facial expressions as anger, disgust, sadness, enjoyment, fear and surprise (the two last sometimes conflated into one), although he suggests as well that contempt and perhaps the complex of shame and guilt may qualify.11 With no consensus among investigators on the elementary affects, or even on how to decide what they are (not to mention that both Panksepp’s and Ekman’s models have been subjected to robust criticism), it is difficult to treat them as the invariant components underlying the shifting history of the emotions.

10But there is a further problem, which in a way is even more intractable, namely that emotions that seem to be based on, or equivalent to, elementary or universal responses sometimes go out of or come into existence entirely, which would seem to vitiate any attempt to treat a given emotion as the variable expression of some underlying affective constant. This circumstance, and the related fact that what is considered an emotion may itself change over time, call for a different approach to the history of emotions, which I sketch very summarily in my conclusion. But first, let me offer two examples of such radical shifts from classical antiquity to the early Christian period: one in which what had earlier been considered an emotion seems to have dropped out, and a second in which a new emotion seems to have appeared where no such sentiment existed previously.

11My first case study centers on the idea of pity. In every classical catalogue of the emotions of which I am aware, pity occupies a central place; like anger and fear today, and indeed in classical antiquity as well, pity was something like a prime or exemplary emotion. When Aristotle defines the emotions in his treatise on rhetoric, he writes:

Let the emotions be all those things on account of which people change and differ in regard to their judgments, and upon which attend pain and pleasure, for example anger, pity, fear, and all other such things and their opposites (2.1, 1378a20-23; for different lists of pathê, cf. Nicomachean Ethics 1105b21ff.; On the Soul 403a16-17).

  • 12 See David Konstan, « The concept of “emotion” from Plato to Cicero », Méthexis t. 19, 2006, p. 139- (...)

Aristotle’s word for « emotion » here is pathos, and since pathos had a wide range of meanings, including misfortune, disease, and other such conditions, Aristotle needed not only to provide an abstract definition but also, as I have suggested elsewhere, some examples of what he was including under the term in this particular context; to this end, pity was as self-evident a candidate as anger and fear.12 Pity figures likewise at Aristotle Rhetoric 3.19, 1419b24-26, and in Cicero’s De oratore 2.206, where the list includes « affection, hate, anger, envy, pity, hope, joy, fear, and grief » (cf. Cicero Brutus 188, Dionysius of Halicarnassus Demosthenes 22 [1.322 Usher], where Dionysius says that upon reading Demosthenes’ speeches « I feel one emotion after another – disbelief, anguish, terror, contempt, hatred, pity, goodwill, anger, envy – every emotion in turn that can sway the human mind » [trans. Usher, slightly modified]). Now, I have found that pity is strikingly absent in modern discussions of the emotions, perhaps in part because it is regarded nowadays as expressing a certain contempt, and hence a politically incorrect sentiment, or else something resembling an attitude, such as arrogance, rather than an emotion proper (David Hume noted this negative side of pity). But the excision of pity from the list of emotions goes back much further (it is absent from Thomas Hobbes’ list of “seven simple passions” in Leviathan Part 1, chapter 6, and from Descartes’ “principal passions” in Les Passions de l’âme articles 68-69). It had already been dropped, at least to all appearances, by Galen:

  • 13 ἔστι δὲ πάθη ψυχῆς, ἅπερ ἅπαντες γινώσκουσι, θυμὸς καὶ ὀργὴ καὶ φόβος καὶ λύπη καὶ φθόνος καὶ ἐπιθυ (...)

And there are passions of the soul which everybody knows: anger, wrath, fear, grief, envy, and violent lust. In my opinion, excessive vehemence in loving or hating anything is also a passion; I think the saying ‘moderation is best’ is correct, since no immoderate action is good.13

But Galen’s inventory of emotions is selective, and serves mainly as a preface to his own special definition of emotion as being quick to feel excessive affection or hatred for any object whatsoever. When we come to Gregory of Nyssa, however, we find a more deliberate kind of exclusion:

  • 14 Taking ἀνέλεος ἀπήνεια as adjective and noun, hence deleting the comma in the TLG text, adopted fro (...)
  • 15 Ἕκαστον γὰρ τῶν ἐν ἡμῖν παθῶν, ὅταν ἐπικρατήσῃ, δεσπότης τοῦ δουλωθέντος γίνεται· καὶ καθάπερ τις τ (...)

Each of the passions in us, when it takes control, becomes master of the person enslaved, and like an oppressive tyrant, having seized the citadel of the soul, it uses our own subordinates to maltreat its victim, employing our own mental processes as agents for its purpose. So anger, so fear, cowardice, insolence, the sensations of pain and pleasure, hatred, contentiousness, pitiless unkindness,14 malice, flattery, grudges and insensitivity, and all the passions in us which are reckoned negative, make up a list of tyrants and masters who to win power reduce the soul to slavery like a captive of war.15

Gregory not only fails to include pity in what is rather an extensive list, but explicitly characterizes a pitiless kind of harshness as deplorable. Indeed, shortly afterwards Gregory observes:

  • 16 οὐ γὰρ ἔστιν οὔτε ἀδακρυτὶ ζῇν τὸν ἐπεσκεμμένον δι᾽ ἀκριβείας τὰ πράγματα, οὔτε ἐν λυπηροῖς εἶναι ν (...)

It is impossible, quite impossible, for one who has studied matters carefully, either to live without tears or to avoid reckoning the person plunged in life’s pleasures to be in a sorry plight. One may observe something similar with dumb beasts, whose natural condition is pitiable; for what could be more pitiable than being deprived of reason?16

Far from being a negative pathos, pity is an appropriate response to the unfortunate state of others, including even animals; this is quite different from the view of Aristotle, who held that pity is elicited only by the undeserved misfortune of others. But the more radical departure from Aristotle and the classical tradition generally is that pity, for Gregory, scarcely counts as an emotion at all, but is rather more like mercy, which is to say, a kindly or humane disposition that recognizes sympathetically the deficiencies in others, even including non-human creatures, not with a view necessarily to altering their condition – one cannot make an animal into a rational being, nor ought one to – but simply as a way of acknowledging their distance from the divine. We can perhaps detect beneath the surface differences between Gregory’s and Aristotle’s views of pity a common element of raw empathy, elicited not by rational judgment but by sheer perception of another’s condition. But if so, then this fellow feeling is inflected so differently by the two thinkers that what was for Aristotle one of the most typical pathê or emotions has become for Gregory something akin to a virtue, like philanthrôpia or epieikeia. Indeed, Philo of Alexandria, in his discussion of the prohibition against boiling a lamb in its mother's milk (Exod 23:19, 34:26; Deut 14:20), had already remarked:

[…]since there is an unlimited supply [of animals], anyone who boils the flesh of lambs or kids or some other animal in the milk of its mother betrays a cruel depravity of character, which has amputated that passion [pathos] that is most related and akin to the rational soul, namely pity (De virtutibus 144).

  • 17 Cf. Summa theologiae Supplementum Question 94, on “The relations of the saints towards the damned”, (...)

12On the Latin side, in turn, we may note that Thomas Aquinas’ list of the passiones animae omits misericordia; Aquinas’ inventory is sorted into two types, six labeled “concupiscible” and including love and hate, desire and aversion, and delight and distress, and five “irascible”, which comprise hope and despair, confidence and fear, and finally anger. Elsewhere, Thomas discusses the possibility of rational or volitional pity, which would not count as an emotion proper.17 But Thomas’ most extensive analysis of pity is to be found in the Secunda Secundae, Question 30, where Thomas worries not only whether pity is a virtue but whether it might not be the supreme virtue (he decides that it comes in second to love or caritas). Without going into the details of his argument, I may cite the translation of the Summa by the Fathers of the English Dominican Province, where the question in question is introduced by the editorial comment:

  • 18 The translation was published by Benziger Brothers in 1947.

The one Latin word "misericordia" signifies either pity or mercy. The distinction between these two is that pity may stand either for the act or for the virtue, whereas mercy stands only for the virtue”.18

The resulting version, in which the single term misericordia is alternately rendered as “pity” or “mercy”, makes a mishmash of Thomas’ analysis.

  • 19 See Lactantius Epitome 33.6-8; Inst. Div. 3.23.8-10; cf. 6.10.2-4, 11-12.

13No doubt there were several reasons for the shift in the understanding of eleos or misericordia between the classical and the Christian periods. Eleos is a principal attribute of God both in the Septuagint, where it renders the Hebrew hesed (e.g., Gen 19:19, 24:12, etc.) or rahmim (e.g., Deut 13:18, Isaiah 47:6, etc.), among other terms, and in the New Testament. Although the Stoics had attacked all pathê, including pity, as inappropriate to the sage and hence to deities, God’s capacity for pity, like his capacity for anger, had to be defended by Christian writers (whether Christ had human passions was much debated in antiquity; cf., e.g., Gregory of Nyssa Against Apollinarius 3.1.136). Pity was also becoming more acceptable as a sentiment in the later Roman Empire, and one might argue that part of the success of Christianity lay in its ability to express emerging values of mercy and clemency under the radically new social conditions of late antiquity. Some, like Lactantius, explained that, contrary to the view of the Stoics, emotions were not bad per se, since God bestowed them on human beings and hence they must have a positive purpose; rather, emotions are bad when employed to a vicious end, and good when in the service of virtue. Thus, anger at adultery, for example, is a good response, and pity too is appropriate when the sufferer is innocent.19 But the move to disqualify pity as an emotion was under way, and it was in the process of being reconfigured as what we know as pity or compassion.

14Turning now to my second case, the Greek term metanoia may literally be rendered as “after-thought”, and in the classical lexicon it signifies a change of mind, though it sometimes bears the sense of the modern regret, that is, sorrow that things have not turned out the way one had wished or expected. This word, and the related term metameleia, do not appear in Aristotle’s inventory of the emotions in the Rhetoric, any more than the Latin equivalent paenitentia is to be found in Cicero’s discussions. Philo of Alexandria, indeed, seems to have included regret (metanoia) among the virtues, although he assigns it a second-class status: just as health is the greatest good for the body and recovery next best, he reasons, so too changing one’s mind “is not ranked in the first and highest class of goods, but as winning second prize in the next class” (De virtutibus 176). Classical philosophers generally have little to say about regret or remorse, and where it is mentioned, it often carries a negative connotation. The object of the philosopher is not to err; the best that Aristotle has to say for regret is that it is a sign that the original action was not performed intentionally (Nicomachean Ethics 1110b18-23; cf. Plato Laws 866D-E). Otherwise, “wicked people [phauloi] are full of regret” (1166b24-25), whereas virtuous people are properly pleased with themselves.

  • 20 Πάντα τε εὖ ποιεῖ ὁ νοῦν ἔχων, καὶ γὰρ φρονίμως καὶ ἐγκρατῶς καὶ κοσμίως καὶ εὐτάκτως ταῖς περὶ τὸν (...)

15The Stoics agreed that regret or a change of mind pertained to ordinary people as opposed to the sage, and was a sign of a flawed character; as Arius Didymus puts it, whereas the person who has good sense does everything well, the base person or phaulos, who has no experience of correct behavior, does everything badly, and for this reason is changeable and afflicted by regret (metameleia) over each of his actions.20 Arius goes on to explain that

  • 21 Εἶναι δὲ τὴν μεταμέλειαν λύπην ἐπὶ πεπραγμένοις ὡς παρ᾽ αὑτοῦ ἡμαρτημένοις, κακοδαιμονικόν τι πάθος (...)

[…] regret (metameleia) is pain for things that have been done in error by oneself, an unfortunate and conflictual passion [pathos] of the soul; for to the extent that one enmeshed in regrets suffers for what has happened, he is irritated with himself for having been responsible for it.21

  • 22 Μεταμέλεια δὲ λύπη ἐπὶ ἁμαρτήμασι πεπραγμένοις ὡς δι᾽ αὑτοῦ γεγονόσιν, p. 12 Kreuttner = SVF 414.1- (...)

The use of the term pathos here suggests that the Stoics, at least, did regard metameleia as an emotion, and this would seem to be confirmed by the inclusion of metameleia in the long list of pathê compiled by Andronicus in his περὶ παθῶν under the class of “pain” (one of the four broad Stoic categories of emotion, the others being pleasure, fear, and desire), along with pity, envy, emulousness, jealousy, and fully eighteen other painful emotions, minutely differentiated.22

16But if metameleia is an emotion, it raises a potential paradox for the Stoics that Cicero, for example, seized upon. As a passion, as opposed to a mere judgment, the Stoics could only view regret as something to be eliminated, like anger, pity, and all other pathê. Yet there is something different about regret and the other emotions, for one might come to regret or have a change of heart concerning precisely one’s own lack of virtue, and one would have thought that such a sentiment was worthy of approval. Thus Cicero writes in the Tusculan Disputations (3.77):

  • 23 et tamen non satis mihi videtur vidisse hoc Cleanthes, suscipi aliquando aegritudinem posse ex eo i (...)

Cleanthes [the successor to Zeno as head of the Stoic school] seems to me, nevertheless, not to have taken sufficient notice of this point, that distress [aegritudo] can sometimes be felt as a result of the very thing that Cleanthes himself concedes is the greatest evil. For what shall we say about the time when Socrates persuaded Alcibiades, as we have heard, that he was no account of a person and that there was no difference between Alcibiades, who was born in the highest station, and any load carrier, and Alcibiades berated himself and in tears implored Socrates to grant him virtue and drive out his baseness – what shall we say, Cleanthes?23

Cicero makes it clear that the cause of Alcibiades’ distress must have been a genuine evil, even on Stoic terms, and that therefore his reaction was a good one (cf. Graver, p. 196). We cannot be sure that the term metameleia (as opposed, for example, to the simple lupê or “distress”) was employed by Cleanthes, but there is no reason to doubt that sometimes a change of mind might be accompanied by grief or anguish, according to the classical Greeks, and so was not just an intellectual matter, a mere alteration of opinion, but had more the quality of a genuine emotion. Nevertheless, there was never any great interest among classical thinkers in regret or remorse, in comparison with their deep concern with the virtues, for example, and in fact it is not a major element in classical literature generally, whether epic or tragedy or any of the other major genres.

17With the advent of Christianity, and even before that with the translation of the Hebrew Bible into Greek, metanoia and metameleia assumed an entirely new role. Margaret Graver puts the matter succinctly:

If one were to consult Philo, or Origen, or any other Jewish or Christian writer of Greco-Roman antiquity, about the proper function of mental distress, an answer would be ready to hand: remorse and repentance bring about a change in one’s relationship to god, marked by a fuller awareness of one’s responsibilities as a moral agent. This explicitly religious conception of remorse, developed under the influence of the Hebrew shuv or “return” (i.e., a return to god), is not to be found in the secular philosophical tradition (p. 206),

  • 24 For a thorough and critical study of the nature of metanoia in Greek sources, both pre-Christian an (...)
  • 25 ηὐχόμην γὰρ πάντας ὑμᾶς ἐν δάκρυσι ζῆν καὶ μετανοίᾳ διηνεκεῖ, Epistulae 207.4; cf. Enarratio in pro (...)

despite the example of the weeping Alcibiades (for shuv rendered as metanoia, cf. Jer 4.28 and 8.6). Now, in fact there are many variations on the way shuv is rendered in the Septuagint (one important term, in this context, is epistrephô), but after the intertestamental period metanoia assumes the sense of a radical conversion, and it occurs in this sense with some frequency in the New Testament.24 A change of heart, however, even one as deep as that implied by a religious conversion, does not necessarily entail the painful emotional experience that we typically associate with the idea of remorse or repentance. The latter meaning became dominant with the emergence of ascetic practices, when penitence came to signify less a fully realized inner transformation, and the joy that might naturally accompany such a renewal of the self, and more a lifelong discipline of self-punishment motivated by a perpetual consciousness of guilt and sin. Examples of such deliberate mortification of the flesh and spirit are familiar and need not be rehearsed here. What the new uses of the terms metanoia, metameleia, and paenitentia, along with coinages such as katanuxis, contritio, and compunctio, indicate is a new emotional register in the idea of a change of mind or afterthought. Whereas with pity or rather eleos and misericordia we saw how a standard emotion in the classical repertoire was sublated into something more like a virtue, in the case of metanoia and metameleia we see the opposite process, by which a term that had by and large the intellectual sense of a change of opinion acquired a powerful emotional resonance. As Basil, for example, writes in his letter to the Neocaesarians, “I pray that you all live amid tears and perpetual repentance”.25 If metanoia and metameleia do not enter regularly into the formal lists of pathê or adfectus even in Christian texts, in part no doubt because of an abiding Stoicizing perception of the emotions as negative (as in the citation from Gregory of Nyssa above), they are nevertheless so closely associated with grief and tears as to assume the character of a full-fledged passion.

18A long-term history of the emotions must take account of such transformations in the very nature of the sentiments under investigation. Changes in the history of any given emotion undoubtedly respond to various factors, including philosophical or religious positions, developments in social life, the influence of foreign contacts, and much else: Barbara Rosenwein’s centipede is a good heuristic image. But there may also be certain patterns or convergences that occur across a range of sentiments, as a given matrix starts to dissipate and a new one congeals, and these invite us, in Derrida’s words, to attend to “the specific function of a term within its system” – that is, the overall configuration of sentiments in an age, both those embraced within and those excluded from the prevailing conception of emotions. If, as I have suggested, the classical idea of pity and the rather bloodless notion of remorse make sense within the context of a culture based on honor, or at least one where honor was a central value, the new Christian virtue of mercy and the passionate conception of remorse or repentance take their place in a world where sin, guilt, and the need for redemption are uppermost. The pressures of a world view will help to organize the emotional lexicon into a system that is at least to some degree coherent. But any such arrangement will at the same time bear traces of its own complex evolution. When we write the history of emotions, we have to take into account such accumulated alterations that eventually change the landscape, to the point of levelling old mountains, as it were, and raising new ones.

Haut de page


1 Barbara H. Rosenwein, “Theories of change in the history of emotions”, in Jonas Liliequist, ed., A history of emotions, 1200-1800, London, Pickering and Chatto, 2012, p. 7-20. Rosenwein refers in the article to Norbert Elias, The civilizing process: Sociogenetic and psychogenetic investigations, trans. Edmund Jephcott, rev. ed., Oxford, Blackwell, 2000; William M. Reddy, The navigation of feeling: A framework for the history of emotions, Cambridge, Cambridge University Press, 2001; and Jan Plamper, “Fear: Soldiers and emotion in early twentieth-century Russian military psychology”, in Jan Plamper, ed., Emotional turn? Feelings in Russian history and culture = Slavic Review, t. 68, 2009, p. 259-83.

2 Jacques Derrida, Margins of philosophy, trans. A. Bass, Chicago, University of Chicago Press, 1982, p. 254.

3 See David Konstan, « Y-a-t’il une histoire des émotions? », in Philippe Borgeaud, ed., Mythes, rites et émotions, Geneva, Librairie Droz, 2009, p. 15-28; idem, “Not Quite Emotions: Sentiments that Did not Make the Grade”, in Natascha Adamowsky, Robert Felfe, Marco Formisano, Georg Toepfer und Kirsten Wagner, eds., Affektive Dinge: Objektberührungen in Wissenschaft und Kunst, Göttingen, Wallstein Verlag, 2011, p. 113-26.

4 Aristotelis finitio non multum a nostra abest; ait enim iram esse cupiditatem doloris reponendi, 1.3.3.

5 contra utramque dicitur feras irasci nec iniuria inritatas nec poenae dolorisue alieni causa; nam etiam si haec efficiunt, non haec petunt, 1.3.4.

6 For discussion of this passage, see Fabio Tutrone, Filosofi e animali in Roma antica: Modelli di animalità e umanità in Lucrezio e Seneca, Pavia, Edizioni ETS, 2012, p. 228-34; Tutrone discusses in detail and with rich bibliography every passage in both philosophers in which animals figure importantly.

7 See David Konstan, The emotions of the ancient Greeks: Studies in Aristotle and classical literature, Toronto, University of Toronto Press, 2006, esp. p. 259-62.

8 For detailed discussion, see David Konstan, “Reason vs. emotion in Seneca,” in Douglas Cairns and Damien Nelis, eds., Emotions in the classical world: Methods, approaches, and directions, Heidelberg, Steiner Verlag, 2016.

9 W. Gerrod Parrott, “Ur-Emotions: The common feature of animal emotions and socially constructed emotions”, Emotion Review, t. 4, 2012, p. 247-48; quotation on p. 247-48. See also W. Gerrod Parrott, “Ur-emotions and your emotions: Reconceptualizing basic emotion”, Emotion Review, t. 2, 2010, p. 14-21; and Nico H. Frijda and W. Gerrod Parrott, “Basic emotions or ur-emotions?”, Emotion Review, t. 3, 2011, p. 406-15 (the issue contains a special section on basic emotion theory), where ur-emotions are defined as « intentional states » accompanied by a « mode of action readiness » (the list of such modes – eighteen in all – includes acceptance, attending, avoid, reject, desire, exuberance, domination, submission, tenseness, and inhibition). However, when Frijda and Parrott affirm that « Ur-emotions are elicited by events as appraised » (p. 410), they would seem to part company with the classical analyses, or else to be using the term “appraise” in a very latitudinarian sense in which non-human animals too can be said to appraise situations.

10 Cited from correspondence via the list-serve of the International Society for Research on Emotions (posted 2 January 2013); see also Jaak Panksepp, “Cross-species affective neuroscience decoding of the primal affective experiences of humans and related animals”, PLOS ONE t. 6(8): e21236 doi:10.1371/journal.pone.0021236, for detailed discussion of so-called basic emotions and a defense of Panksepp’s own set.

11 “Afterword”, in Charles Darwin, The expression of the emotions in man and animals, 3rd ed., with Introduction, Afterword and Commentary by Paul Ekman, London, Harper Collins 1998 (orig. 1872), p. 390-91; for a survey of various lists of basic emotions, see Robert C. Roberts, Emotions: An essay in aid of moral psychology, Cambridge, Cambridge University Press, 2003, p. 185-92.

12 See David Konstan, « The concept of “emotion” from Plato to Cicero », Méthexis t. 19, 2006, p. 139-51.

13 ἔστι δὲ πάθη ψυχῆς, ἅπερ ἅπαντες γινώσκουσι, θυμὸς καὶ ὀργὴ καὶ φόβος καὶ λύπη καὶ φθόνος καὶ ἐπιθυμία σφοδρά. κατὰ δὲ τὴν ἐμὴν γνώμην καὶ τὸ φθάσαι πάνυ σφόδρα φιλεῖν ἢ μισεῖν ὁτιοῦν πρᾶγμα πάθος ἐστίν; On diagnosing and curing ailments of the soul K; trans. Paul W. Harkins, Galen on the passions and errors of the soul, Columbus, Ohio State University Press, 1963, p. 32.

14 Taking ἀνέλεος ἀπήνεια as adjective and noun, hence deleting the comma in the TLG text, adopted from Migne, Patrologia graeca 44.1193-1302.

15 Ἕκαστον γὰρ τῶν ἐν ἡμῖν παθῶν, ὅταν ἐπικρατήσῃ, δεσπότης τοῦ δουλωθέντος γίνεται· καὶ καθάπερ τις τύραννος καταλαβὼν τῆς ψυχῆς τὴν ἀκρόπολιν, δι᾽ αὐτῶν τῶν ὑποχειρίων κακοῖ τὸ ὑπήκοον, ὑπηρέταις τοῖς ἡμετέροις λογισμοῖς πρὸς τὸ δοκοῦν ἑαυτῷ καταχρώμενος· οὕτως ὁ θυμὸς, οὕτως ὁ φόβος, ἡ δειλία, τὸ θράσος, τὸ κατὰ λύπην τε καὶ τὸ καθ᾽ ἡδονὴν πάθος, μῖσος, ἔρις, ἀνέλεος ἀπήνεια, φθόνος, κολακεία, μνησικακία τε καὶ ἀναλγησία, καὶ πάντα τὰ κατὰ τὸ ἐναντίον ἐν ἡμῖν νοούμενα πάθη, τυράννων ἐστί τινων καὶ δεσποτῶν ἀπαρίθμησις, τῶν πρὸς τὸ ἴδιον κράτος τὴν ψυχὴν οἷόν τινα δορυάλωτον καταδουλούντων.

16 οὐ γὰρ ἔστιν οὔτε ἀδακρυτὶ ζῇν τὸν ἐπεσκεμμένον δι᾽ ἀκριβείας τὰ πράγματα, οὔτε ἐν λυπηροῖς εἶναι νομίζειν τὸν ταῖς βιωτικαῖς ἡδοναῖς ἐμβαθύνοντα·καθάπερ ἐπὶ τῶν ἀλόγων τὸ τοιοῦτον ἔστιν ἰδεῖν·οἷς ἐλεεινὴ μὲν ἡ τῆς φύσεώς ἐστι κατασκευή·τί γὰρ ἐλεεινότερον τῆς τοῦ λόγου στερήσεως; the translation is that of Stuart George Hall, in Hubertus R. Drobner and Albert Viciano, eds., Gregory of Nyssa: Homilies on the beatitudes, Leiden, E.J. Brill, 2000, p. 44-45. Nyssa offers as one possible definition of eleos: “Pity is a loving self-identification with those vexed by grievous events” (trans. ibid., p. 59).

17 Cf. Summa theologiae Supplementum Question 94, on “The relations of the saints towards the damned”, Article 2: “Mercy or compassion may be in a person in two ways: first by way of passion, secondly by way of choice. In the blessed there will be no passion in the lower powers except as a result of the reason's choice. Hence compassion or mercy will not be in them, except by the choice of reason”, trans. Fathers of the English Dominican Province, Online Edition Copyright © 2008 by Kevin Knight; n.b.: the Supplement was compiled, it seems, by Fra Rainaldo da Piperno, based on Thomas’ commentary on the Fourth Book of the Sentences of Peter Lombard.

18 The translation was published by Benziger Brothers in 1947.

19 See Lactantius Epitome 33.6-8; Inst. Div. 3.23.8-10; cf. 6.10.2-4, 11-12.

20 Πάντα τε εὖ ποιεῖ ὁ νοῦν ἔχων, καὶ γὰρ φρονίμως καὶ ἐγκρατῶς καὶ κοσμίως καὶ εὐτάκτως ταῖς περὶ τὸν βίον ἐμπειρίαις χρώμενος συνεχῶς. ὁ δὲ φαῦλος, ἄπειρος ὢν τῆς ὀρθῆς χρήσεως, πάντα κακῶς ποιεῖ καθ᾽ ἣν ἔχει διάθεσιν ἐνεργῶν, εὐμετάπτωτος ὢν καὶ παρ᾽ ἕκαστα μεταμελείᾳ συνεχόμενος, quoted in Stobaeus Anthology 2.7.11.

21 Εἶναι δὲ τὴν μεταμέλειαν λύπην ἐπὶ πεπραγμένοις ὡς παρ᾽ αὑτοῦ ἡμαρτημένοις, κακοδαιμονικόν τι πάθος ψυχῆς καὶ στασιῶδες· ἐφ᾽ ὅσον γὰρ ἄχθεται τοῖς συμβεβηκόσιν ὁ ἐν ταῖς μεταμελείαις ὤν, ἐπὶ τοσοῦτον ἀγανακτεῖ πρὸς ἑαυτὸν ὡς αἴτιον γεγονότα τούτων, my translation; cf. Margaret Graver Stoicism and emotion, Chicago, University of Chicago Press, 2007, p. 193; Seneca On benefits 4.34.4.

22 Μεταμέλεια δὲ λύπη ἐπὶ ἁμαρτήμασι πεπραγμένοις ὡς δι᾽ αὑτοῦ γεγονόσιν, p. 12 Kreuttner = SVF 414.1-32.

23 et tamen non satis mihi videtur vidisse hoc Cleanthes, suscipi aliquando aegritudinem posse ex eo ipso, quod esse summum malum Cleanthes ipse fateatur. quid enim dicemus, cum Socrates Alcibiadi persuasisset, ut accepimus, eum nihil hominis esse nec quicquam inter Alcibiadem summo loco natum et quemvis baiolum interesse, cum se Alcibiades adflictaret lacrimansque Socrati supplex esset, ut sibi virtutem traderet turpitudinemque depelleret – quid dicemus, Cleanthe?

24 For a thorough and critical study of the nature of metanoia in Greek sources, both pre-Christian and Biblical, see Guy D. Nave, Jr., The role and function of repentance in Luke-Acts, Atlanta, Society of Biblical Literature, 2002.

25 ηὐχόμην γὰρ πάντας ὑμᾶς ἐν δάκρυσι ζῆν καὶ μετανοίᾳ διηνεκεῖ, Epistulae 207.4; cf. Enarratio in prophetam Isaiam p.6 (if not by Basil, then very likely by a contemporary Cappadocian): Ἐὰν οὖν ὀργιζόμενον, ἢ λυπούμενον, ἢ μεταμελόμενον..., etc.

Haut de page

Pour citer cet article

Référence électronique

David Konstan, « Their Emotions and Ours: A Single History? »L’Atelier du Centre de recherches historiques [En ligne], 16 | 2016, mis en ligne le 26 mai 2016, consulté le 30 mai 2024. URL : ; DOI :

Haut de page


David Konstan

Professeur d’études classiques à New York University. Parmi ses livres, on peut citer : Roman Comedy (1983) ; Sexual Symmetry: Love in the Ancient Novel and Related Genres (1994) ; Greek Comedy and Ideology (1995) ; Friendship in the Classical World (1997) ; Pity Transformed (2001) ; The Emotions of the Ancient Greeks: Studies in Aristotle and Classical Literature (2006) ; “A Life Worthy of the Gods”: The Materialist Psychology of Epicurus (2008) ; Before Forgiveness: The Origins of a Moral Idea (2010), et, récemment, Beauty: The Fortunes of an Ancient Greek Idea (2014). Ancien président de l’American Philological Association, il est membre de l’American Academy of Arts and Sciences et membre honoraire de l’Australian Academy of the Humanities.

Haut de page

Droits d’auteur


Le texte seul est utilisable sous licence CC BY-SA 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search