Walter B. Cannon

James’s former undergraduate student, later his most consequential critic. The Cannon–Bard theory is the principal challenge to James–Lange and the source of the “undifferentiated arousal” premise that runs through activation theory and Schachter–Singer.

The critique of James–Lange

From studies of sympathectomized cats (which still hissed and showed piloerection), Cannon concluded autonomic feedback is not necessary for emotional behavior. His arguments: (1) the same visceral changes occur across different emotions and even non-emotional states → undifferentiated arousal; (2) visceral responses are too slow; (3) viscera have too few afferents to carry emotional distinctions. Feelings and bodily responses are separate outputs of central (thalamic) processing. See autonomic-specificity-of-emotion.

Lasting effect

Per Mandler, Cannon’s importance “is not so much that [his criticisms] destroyed the James–Lange theory, but rather that they were influential in producing an extensive research tradition in the psychophysiology of emotion” — the very tradition (Ax, Ekman, Levenson) that later rebutted his non-specificity claim. See friedman-2010-jamesian-perspective.

The second overturning: “autonomic” was the wrong word

The wiki holds Cannon almost entirely as James’s critic. Berntson & Khalsa (2021) record a separate and equally consequential Cannon claim, and the accumulated case against it.

Cannon called the “interofective” system autonomic because it acts “automatically, without direction from the cerebral cortex” (Wisdom of the Body, 1939) — visceral regulation as low-level reflex. That premise organized twentieth-century autonomic physiology, and the last decade has dismantled it from the tracing end:

• Three distinct prefrontal networks — movement, cognition, affect — reach the monkey adrenal medulla in three synapses (Dum et al. 2016).
• Distinct cortical networks drive the stomach’s parasympathetic (rostral insula, mPFC) and sympathetic (motor, S1) limbs (Levinthal & Strick 2020).
• Mouse insular cortex shifts feeding by aversive sensing and anticipates expected satiety.

Berntson & Khalsa’s summary of the change: “there are major rostral neural contributions to visceral regulation, which in turn generally receive a wide array of ascending visceral afferents.” The central-autonomic-network page exists to hold the system Cannon said did not exist.

Note the irony worth keeping. Cannon’s two big claims about the viscera — too undifferentiated to constitute emotions, and too automatic to be centrally directed — have both been substantially reversed, but by different literatures that rarely cite each other. And homeostasis, his other lasting contribution, is now argued to be predictive as well as reactive, which softens his third premise too.

Cannon vs the “adrenalizing” of emotion

Per Dror, Cannon insisted on distinguishing “strong” (sympathetic) from “soft” (parasympathetic) emotions and studied adrenalin only in rage/fear. Schachter & Singer’s universalization of sympathetic arousal to every emotion ran directly against this and against 1940s parasympathetic findings. See origins-of-two-factor-theory.