Constructed emotion
Emotions are not fixed, pre-wired reactions waiting in the brain to be triggered. On this view the brain builds an emotion in the moment, taking ambiguous signals from the body and the situation and labelling them with concepts learned through experience and culture; the feeling is the result of that act of categorising, not a readout from a dedicated emotion circuit. It is set against the older picture of emotions as universal, biologically discrete states with their own facial and bodily signatures.
Lisa Feldman Barrett's theoretical framework holding that emotions are not biologically pre-wired, evolutionarily fixed categories (the classical view she argues against) but are constructed in the moment by the brain from interoceptive prediction, conceptual knowledge, and situational context. The brain categorises ambiguous bodily and situational signals using emotion concepts learned through experience and culture; the resulting emotion is a categorisation of the body's predicted state, not a readout of a hardwired emotion module.
Etymology§
The phrase theory of constructed emotion is Barrett's coinage, refined across the late 2000s and 2010s. Earlier framings she used include conceptual act theory (Barrett 2006, Perspectives on Psychological Science) and the current name with the 2017 trade book How Emotions Are Made (Houghton Mifflin Harcourt). The theoretical commitment traces to her mentor James Russell's core affect dimensional model and to wider constructionist traditions in psychology, but Barrett gave the construct its modern neuroscientific grounding via the embodied-predictive- interoception-coding (EPIC) framework of Barrett and Simmons 2015.
Constructed emotion is the theoretical position that emotion is not a readout but a categorisation. The brain receives ambiguous body- state and situational signal and applies an emotion concept — fear, anger, sadness — to label what is happening; the felt experience of emotion is the consequence of the categorisation, not its cause. The argument runs against the classical or basic emotion view (Paul Ekman's tradition, Jaak Panksepp's affective- neuroscience programme) that emotions are universal, biologically discrete categories with specific facial expressions and physiological signatures.
For interoception specifically, constructed emotion is the theoretical sibling of interoceptive inference: both place interoception inside a predictive-processing framework, both treat feeling as constructed rather than read out, both make the agranular visceromotor cortices the prediction- generating layer rather than Craig's right-anterior-insula-as-feeling-substrate. Where Seth 2013 makes the theoretical case, Barrett and Simmons 2015 makes it with a specific architectural model (EPIC) and corticocortical-connectivity grounding.
The theory has generated substantial opposition from the basic- emotion tradition — Ekman, Joseph LeDoux's amygdala-and-fear programme, the affective-neuroscience tradition of Panksepp. The disagreement is not merely empirical; it is about what emotion is as a category, and what evidence would settle the question. Barrett's 2017 trade book is the popular-facing argument; the journal literature has continued to accumulate on both sides.