Since reading Martha Nussbaum’s Anger and Forgiveness, I have found myself continually more attracted to her concept of transition-anger. That is: the main, and perhaps only, place where anger is a helpful emotion is on its first arising, where it signals to us that something is wrong or unjust; after that, one should transition “off the terrain of anger toward more productive forward-looking thoughts”. (Nussbaum capitalizes “Transition-Anger”, but that seems an awkward usage to me.)
I’ve found the concept of transition-anger very helpful for the argument of my upcoming book (which is more focused than my original concept was, so anger now plays a larger role in it). More even than that, though, I think the basic idea of transition-anger can and should be expanded to other emotions: it is not only anger which is most valuable on first arising. Nussbaum doesn’t consider that approach in Anger and Forgiveness, and there wasn’t a need for her to do so since the book wasn’t about other emotions, but only about anger. But it’s worth talking about here.
Observing my own emotional life, I have noted there is a set of four emotions that I feel very often – most of them daily – and they all cause me trouble and suffering. Yet I see how each can potentially be valuable on first arising. Anger is one of them; the other three are fear, shame, and self-pity. Let’s go through them in turn.
Fear is perhaps the clearest case. We are accustomed to thinking of fear as serving a signal function, as transition-anger does. Like anger, fear gives us the warning that something is wrong; we appreciate the fear response that comes unthinkingly when something is about to fall on our heads or burn us. The problem comes when fear sticks around – sometimes even with no object, nothing to be afraid of. That superfluous fear is what psychologists distinguish as anxiety, and I know it all too well; my therapist tells me I have generalized anxiety disorder. I would love to be able to get myself down to transition-fear: keeping my fear as the first signal that something is dangerous, but then moving to more forward-looking thoughts of avoiding the danger.
Something very similar applies to shame. It too can serve a useful signalling function – this time about something wrong that we ourselves have done. That first feeling of shame can warn us that we have done somebody wrong. Yet just like fear and anger, too often shame sticks with us, returns to us, making us feel worthless in a way that paralyzes us and prevents us from moving to make it right or prevent similar wrongs in the future. Take the signal offered by the initial shame, and transition on to more forward-looking thoughts.
It is not so obvious that self-pity is best suited to a similar kind of transition, but I think that it is. Especially, I think mourning serves as a kind of transition self-pity. Feeling sorry for ourselves can be a dangerous trap that we wallow and get lost in. Yet at the same time, if we do value external goods – as I think we should – then their losses are real losses, and we need to acknowledge them as such, not just with thoughts but with feelings. When the Greek philosopher-general Xenophon learned of his son’s death in battle, it was reported that Xenophon “did not even shed tears, but exclaimed, ‘I knew my son was mortal.'” Unlike the founders of Stoicism, who admired Xenophon, most of us would look at Xenophon and see something wrong with him, with his not feeling sorry for himself. He should have felt pain and suffering at such a great loss. For him to be so unmoved indicates that it was no real loss, which strongly suggests that he did not love his son in the first place.
Yet something would also be off if the loss of a son were to leave a father in a permanent state of sorrow – if he could not eventually move on, or if he took his own life as a result. After all, we cannot imagine a son wanting his father to be destroyed by his death – not if he loved his father himself! To feel no sorrow when we lose a loved one suggests that we have failed them by not truly loving them. But if we spend the rest of our lives mired in misery over their loss, we are failing them in a subtler sense: they loved us and so wished for our well-being, and our sorrow now stands in the way of our fulfilling that wish.
Thus self-pity too serves as an initial signal that something has gone wrong, but in a way that we should move away from. There is a reason why most societies have mourning rituals, ways of expressing grief in a limited period of time after a loss. (And while I’ve focused on the case of a loved one’s death here, mourning can be valuable for many other sorts of losses, including political ones.)
Here Nussbaum’s own life may serve as an example. Just a few short years after the publication of Anger and Forgiveness, Nussbaum’s only daughter Rachel, a campaigner for animal rights, died of a drug-resistant infection. I can only imagine the terrible pain that Nussbaum must have gone through at the time; surely if any situation merits self-pity it is the loss of one’s only child. Yet I notice that now, four years later, she has just come out with a book on animal rights, her daughter’s signature issue, dedicated to her memory. This seems to me a most admirable example of a move off the terrain of self-pity toward more productive forward-looking thoughts.
Very astute analysis, Amod. Thank you for your insights!
Thanks!
Amod, your account here of the transition emotions reminds me of some concepts in emotion-focused therapy (EFT), a conceptual framework for psychotherapy developed by Les Greenberg and colleagues. Here’s a concise summary of the relevant concepts, quoted from the current version of the Wikipedia article about it:
What you and Nussbaum call the emotional transition seems to correspond to principles 4 to 6 in that Wikipedia quotation.
I think your account in the post above could be improved by making a distinction similar to the EFT distinction between primary adaptive and primary maladaptive emotion responses. Something like that distinction is probably used by Nussbaum too, but it doesn’t really come through in this blog post.
For example, you said: “the main, and perhaps only, place where anger is a helpful emotion is on its first arising, where it signals to us that something is wrong or unjust.” True enough, but it could also be the case that the signal is wrong: our initial judgment “that something is wrong or unjust” can be incorrect. And this can be true of the other emotion categories as well: the cognitive aspect of them can be mistaken, in some cases persistently so. In the case of fear, you said that “the problem comes when fear sticks around”—but more precisely, the problem comes when maladaptive fear sticks around, or else starts as adaptive but becomes maladaptive, with all its sequelae. If adaptive fear sticks around, it can be helpful: living in fear when there is a real persistent threat is adaptive because it can keep us out of danger. So your already strong account of emotion here would be further strengthened by making that distinction clearer, it seems to me.
I suspect, judging from what you’ve written about emotion in the past (as I recall it), that you may tend to downplay such cognitive distinctions in emotions: e.g., your 2021 post “Emotions are not primarily judgements”. In the comments on that post, Benjamin C. Kinney agreed that “Emotions drive ideas more than the other way around”, and that’s true enough in a way, but it’s also a very imprecise way of speaking, because there is already cognition in the emotions. In every credible theory of emotions that I’m familiar with, there is an important cognitive aspect even in the “first arising” of emotion responses (see, e.g., the Wikipedia article “Theory of constructed emotion” and references therein), which is why EFT makes the primary adaptive/maladaptive distinction.
Thanks, Nathan. I agree with the adaptive/maladaptive point. You’re right that Nussbaum doesn’t emphasize the point that anger’s initial judgement can itself be wrong (and neither have I in posts above), but I agree that it can and I think she would too. The point is that, while these emotions are so often harmful, it’s on the first arising where they are most likely to help and be useful – to be adaptive, in Greenberg’s terms. They call our attention to aspects of our situation that we might not notice without them – and while they certainly can do so wrongly, they do so rightly often enough that that first arising is something helpful and necessary (unlike the later times they stick around).
When you say “there is already cognition in the emotions”, what precisely do you mean by that?
Thanks for the response! By “cognition” I mean all the mental (brain) processes, conscious and unconscious, involved in attention, interpretation, and memory. When I said “there is already cognition in the emotions”, I was thinking especially of interpretation and memory: there is some interpretation and remembering happening (which I may not be consciously aware of) before I start feeling the physiological aspects of emotion like “butterflies”, tension, pain, tears, or whatever.
But the brain is complex and layered, and there are various pathways through which all the processes that constitute cognition and emotion can happen. Some pathways produce faster, more “automatic” emotion, and other pathways produce emotion with more layers of cognition. But it’s my understanding that the typical brain is very interconnected, so there’s always already some kind of cognition in the emotions.
I’m always trying to learn more about how recent neuroscience research is changing emotion theory. I’ve found that it’s all so complex that prose alone isn’t sufficient to think about and communicate the subject: other forms like diagrams are necessary to keep all the relevant knowledge well organized and coherent.
By the way, in my comment I may have conflated the truth value of an emotional judgment (or what its truth value would be if we made the judgment explicit in language) with its adaptive value. Truth and adaptiveness are both kinds of success, but they are not necessarily the same: judgments could be adaptive without being true, such as “positive illusions”. But perhaps the more we think about something, the less adaptive falsity becomes, because falsity can infect our reasoning and prevent the systematization of knowledge.
Okay, I think I’m in agreement with you on most of this, at least. For me the “(which I may not be consciously aware of)” caveat is key. The Nussbaumian theory that I object to in the 2021 post tends to imply a rationality and transparency in the nature of emotion, which I don’t think is a view that can be sustained. Our emotions come to us unbidden in ways that we’re often not aware of.
Moreover, we often experience them without cognitive content or with misleading cognitive content. Thus anxiety is often just experienced as fear that isn’t fear of something – and sometimes it’s a displaced fear. So for example I recently took a plane trip after having just heard some very concerning family health news, and I felt terrified “of” missing the flight – but it was pretty clear that the flight wasn’t what was really causing me to be afraid, even though that was the content of the fear/anxiety. I think that fits with the point you’re making here – there was cognitive information that triggered the fear, it just got processed in a complex and distorted way. Such that in the experience of the emotion itself, the somatic or visceral dimension is more prominent than the cognitive cause or content.
I agree with all you said, and I would add that in anxiety disorders it’s common to make interpretations or inferences from somatic or visceral sensation; this is captured well in the title of the article: Arnoud Arntz, Michael Rauner, & Marcel van den Hout (1995), “‘If I feel anxious, there must be danger’: ex-consequentia reasoning in inferring danger in anxiety disorders”, Behaviour Research and Therapy, 33(8), 917–925.
Here’s an article on the “adaptiveness” of emotion dysregulation: Tsachi Ein-Dor & Gilad Hirschberger (2018), “On sentinels and rapid responders: the adaptive functions of emotion dysregulation”, in: Lench, H. C. (ed.), The Function of Emotions: When and Why Emotions Help Us (pp. 25–43), Cham: Springer.
Here’s an example of the kind of complex diagram of cognitive processes that is hard to capture in prose: Figure 1 in: J. Benjamin Hutchinson & Lisa Feldman Barrett (2019), “The power of predictions: an emerging paradigm for psychological research”, Current Directions in Psychological Science, 28(3), 280–291. The authors differentiate between a “representational framework” perhaps somewhat like Nussbaum’s theory (part A of the diagram) and a “predictive-processing framework” (part B of the diagram).
I just found, but haven’t yet read, what looks like a great up-to-date summary and critique of emotion theories: Demystifying Emotions: A Typology of Theories in Psychology and Philosophy by Agnes Moors (Cambridge University Press, 2022).