The design and evolved function of anger
Is anger a feeling or a computational system?
Everyone has experienced anger, but it is much more than a feeling. Anger can the thought of as a neurocognitive system–one that regulates a variety of cognitive, physiological, and behavioral processes. At the CEP, we have been developing a theory about the adaptive functions of anger, and using it to derive and test hypotheses about its computational design.
Wait…What? Aren’t emotions subjectively experienced feelings? Why are we talking about anger as a computational system?
Conscious experiences are produced by complex computational machinery. Consider vision. When we open our eyes, we see a world of objects, but studying the phenomenology of perceptual experiences (e.g., how similar are your experiences of yellow and blue?) was the wrong empirical strategy. The science of perception leapt forward when cognitive scientists identified the basic problem that vision evolved to solve (What objects exist and where are they?) and began asking a computational question: How can information collected by your retina—a 2-dimensional sheet of cells—be used to infer the identity and location of 3-dimensional objects in the world?
We are making the same bet about emotions and motivation: that more progress can be made by asking what adaptive problems an emotion like anger evolved to solve, and then asking what properties a computational system would need to solve those problems well. Questions about the subjective experiences these computations create can always be addressed later, when we have a better handle on how the anger system works.
The Recalibrational Theory of Anger
The model we have been testing is the recalibrational theory of anger. It proposes that anger is produced by a neurocognitive system that evolved to deploy bargaining tactics to resolve conflicts of interest in favor of the angry individual. It is triggered by the inference that another person is not putting enough weight on your welfare. When the anger system is activated, it orchestrates responses designed to bargain effectively for better treatment. The functional product of anger (if successful) is to recalibrate the other person’s decision systems, in ways that increase the weight they put on your welfare in the future.
Bargaining power. The two bargaining tools humans have is the ability to confer or withhold benefits, and the ability to inflict costs–through aggression, for example. The greater the benefits an individual controls, or the greater the costs the individual can inflict, the greater the bargaining position he or she has.
If the anger system is designed to bargain for better treatment, then it should be sensitive to your relative bargaining power. If so, then people with more bargaining power will feel entitled to better treatment–their anger system will activate more easily or more frequently than the same system in a person who has less bargaining power. A simple prediction is that perceptions of your own bargaining power will be influenced by conditions or situations that would have given you more bargaining power under ancestral conditions. All else equal, stronger men had more bargaining power ancestrally, through their enhanced ability to inflict costs through aggression. Similarly, more attractive women had had more bargaining power, through their ability to confer benefits. Other conditions should also matter too, but we started with these two.
Formidability and the logic of human anger (PNAS 2009) introduces the recalibrational theory of anger and tests these straightforward predictions. The evidence shows that stronger men and more attractive women are more anger-prone, feel more entitled to better treatment, and prevail more in conflicts of interest. They also more strongly endorse the use of force to resolve conflicts–even in modern situations where their personal strength (or attractiveness) plays no role (such as the use of military force in international conflicts).*
These results undermine theories that attribute anger and aggression primarily to frustration, a history of negative treatment, or a desire for equity. According to the recalibrational theory, strength and beauty are not unique: Anything that increases the social bargaining power of an individual should increase that person’s anger-proneness and feeling of entitlement.
For a brief account, see Andrea Estrada’s UCSB press release.
The grammar of anger
The recalibrational theory of anger predicts that anger will be triggered by the inference that the target is putting too little weight on your welfare. This means that imposing a cost on someone is neither necessary** nor sufficient to trigger anger. Consider this example:
Imagine I ruined your beautiful silk scarf (the cost I imposed). If the scarf got stained because I used it to make a tourniquet to save my child from bleeding to death–a large benefit to me–my imposing that cost should trigger little or no anger in you. The benefit to me is so large that I would impose this cost even if I valued your welfare a great deal. But let’s say the scarf got stained because I used it to wipe ketchup off my face—I was too lazy to reach further for a napkin. The cost I imposed is the same as before (your scarf is ruined), but my doing this for such a small benefit for myself implies that I put very little weight on your welfare. That is likely to trigger your anger system.
We tested this and many other fine-grained predictions that follow from the recalibrational theory in The grammar of anger: Mapping the computational architecture of a recalibrational emotion (Cognition, 2017). This includes predictions about the arguments angry people deploy when they are bargaining for better treatment and the arguments people use to make an angry person less angry. (Hint: These arguments have a conceptual grammar that follows directly from the hypothesis that anger is triggered by the inference that someone puts too little weight on your welfare.
The results of these experiments replicated across all six tested cultures: the US, Australia, Turkey, Romania, India, and Shuar hunter-horticulturalists in the Ecuadorian Amazon. The results contradict key predictions about anger based on equity theory and social constructivism.
**Note that inflicting a cost is not a necessary condition either. The inference that I put little weight on your welfare can also be triggered when I give you a benefit–especially if you notice I only do this to avoid incurring a cost myself! Julian Lim tested this prediction in his dissertation–we are just behind on our paper writing. Stay tuned.
Emotions as computational systems
Our papers on anger, shame, and pride are part of a more encompassing research initiative intended to place the study of emotion and motivation on a new, more rigorous foundation. The goal of this research is to produce detailed maps of the circuit logic of emotions and motivational systems, showing how the details of their problem-solving characteristics match the recurrent adaptive problems our ancestors faced. This program formed the core of the joint proposal Leda Cosmides and John Tooby made to the NIH Director’s Pioneer Award program, granted in 2005. Discussions of this research program can be found in:
Tooby, J., Cosmides, L., Sell, A., Lieberman, D. & Sznycer, D. (2008). Internal regulatory variables and the design of human motivation: A computational and evolutionary approach. In Andrew J. Elliot (Ed.) Handbook of approach and avoidance motivation. pp. 251-271. Mahwah, NJ: Lawrence Erlbaum Associates.
Tooby, J. & Cosmides, L. (2008). The evolutionary psychology of the emotions and their relationship to internal regulatory variables. In M. Lewis, J. M. Haviland-Jones & L. F. Barrett (Eds.), Handbook of Emotions, 3rd Ed. (pp. 114-137.) NY: Guilford. This paper also summarizes the evolutionary psychological framework for analyzing emotions as circuits or programs.
*Journalism gone wild. In 2010, The Sunday Times of London published a piece claiming that researchers at the CEP found a link between blonde hair in women and anger, entitlement, and “warlike” behavior. No such research was done, and we believe the claims of the article are false. As can be seen by a search in our original publication (here) the words “blonde” or even “hair” never appear. Nevertheless, the story spread rapidly throughout the blogosphere and the mainstream news. We took steps to correct the story. Click here to read our letter to the Times.
What we did do was research on the evolved function of anger, and its relationship to variables such as strength and attractiveness (much more interesting, and with facts behind it!).
Afterwards, Accurate accounts of the story were reported here and here.