Are LeDoux’s survival circuits basic emotions under a different name? (Scarantino 2018)
A commentary in the Current Opinion in Behavioral Sciences themed issue on Survival circuits — edited, notably, by Dean Mobbs and LeDoux himself, so this is an invited disagreement rather than a hostile one. See andrea-scarantino.
The paper is the wiki’s first source to engage LeDoux (2012) directly, and it does something more useful than agree or disagree: it isolates exactly how much of the survival-circuit proposal is a claim about the brain and how much is a proposal about vocabulary. Its answer — almost all vocabulary — is worth taking seriously precisely because it comes from the theorist LeDoux’s framework was supposed to displace.
The title question, answered
Scarantino’s answer is nearly, but not quite, and the residue matters.
What the two agree on (his explicit list):
- Adaptive behaviours and feelings are implemented by different circuits. Both are against Panksepp here — see can-we-know-animal-feelings.
- The subcortical mechanism reflexively generates a global motive state instantiated in preparatory physiology and brain arousal. LeDoux calls it the defensive organismic state (global-organismic-state); Scarantino calls it activated basic fear; Mobbs calls it activated reactive fear. Same box in the diagram, three labels.
- That state does not necessarily involve a subjective feeling.
- A general-purpose cognitive system regulates it into instrumental (not merely reactive) defensive behaviour — the two-level control structure of Gallistel (1980). Both treat cortical regulation as what makes the circuit adaptive rather than reflexive.
- Feelings, when they occur, arise via higher-order representation in cortical networks. See higher-order-theory-of-consciousness.
What they disagree on: what to call the subcortical mechanism, and correspondingly what “fear” denotes.
| subcortical motive mechanism | the felt state | |
|---|---|---|
| LeDoux | defense survival circuit | fear |
| Scarantino | basic fear | feeling of fear |
| Mobbs | reactive fear | cognitive fear |
Scarantino’s Figure (p. 79) is built to make this vivid: an identical box-and-arrow flow (threat → sensory system → subcortical defensive circuit → reflex-like responses / cognitive circuit → instrumental defensive behaviours), with only the labels highlighted in red as the site of dispute.
One recorded asymmetry: both agree the global motive state need not be felt, “although LeDoux is more tentative than I am in granting the defensive organismic state causal powers over behavior.” That is not a terminological difference, and the paper does not develop it. It is the one place a real empirical disagreement peeks through.
New BET: what makes it new
Scarantino’s target is not only LeDoux but traditional BET (Ekman). Two of Ekman’s auxiliary assumptions are discarded while the “hard core” (basic emotions are evolutionary adaptations, pan-cultural, with homologs in other species) is retained — an explicitly Lakatosian manoeuvre, hard core vs protective belt.
1. Terminology → explication. Traditional BET labels its posits with folk categories, which “entitles BET’s opponents to find counterexamples by gathering data about anything to which the folk psychological categories apply.” Scarantino concedes the counterexamples land: folk emotion categories “do not designate natural kinds: their extensions are too heterogenous for any scientifically interesting generalizations to be true of all of their members.”
His diagnosis of why is the sharpest line in the paper: folk categories “were not introduced into the language to carve nature at its joints, but rather to negotiate social transactions among language users. If folk emotion categories happened to carve nature at its joints without having been introduced for that purpose, this would be an unlikely and happy coincidence.” The coincidence has not materialized.
The remedy is Carnapian explication — replace the folk category with a narrower, scientifically tractable one (basic fear ⊂ fear), as memory science replaced folk “memory” with short-term/long-term memory. This reprices the entire evidential landscape: “A critique of a theory of basic fear based on data about fear writ large is like a critique of a theory of short-term memory based on empirical data about memory writ large.” See basic-emotions.
The guard against this being a blank cheque: an explication can still fail on similarity (is it recognizably the folk concept?) or fruitfulness (does it support interesting generalizations?). Scarantino asserts New BET passes both and cites his earlier work rather than showing it here.
2. Mode of operation → input-output open. Ekman’s affect programs are input-open (learning shapes elicitation) but output-closed (a mandatory response cascade). Scarantino rejects the output half — see affect-programs. This is the wiki’s most direct statement of why the bodily-signature literature may have been chasing a badly-motivated target, and it is an evolutionary argument, not an evidential one: to solve abstractly-defined problems like “dealing with dangers or slights,” emotions “cannot afford to produce anything more than impulses to behavior that are flexibly adaptable to the specific circumstances,” with rare exceptions where speed is everything (looming objects). Ekman + Frijda: evolved adaptations that issue action tendencies with control precedence.
Direct relevance to autonomic-specificity-of-emotion: Scarantino accepts the negative empirical record (“high correlations and specificities have not been demonstrated empirically”) that Friedman (2010) contests, then denies it refutes BET, because output-closure was never a good hypothesis. This is a third position on that debate — not “the signatures are there” (Friedman/Levenson) and not “their absence refutes discrete emotions” (Barrett), but “their absence was predictable and BET should never have promised them.”
HPC kinds: variability as prediction, not refutation
The metaphysical engine. Basic emotions are homeostatic property cluster kinds (Boyd 1991) — defined by clusters of relational, changing, historical properties co-occurring often (not always) because of imperfect underlying mechanisms, with no necessary or sufficient properties. Boyd’s argument, which Scarantino leans on: these are the only natural kinds the special sciences are likely to find; biological species work this way.
So basic emotions “lack essences and fixed boundaries, show massive internal variability due to the need of adapting responses to changing circumstances, lack one-to-one mappings with any folk psychological categories, but nevertheless carve nature at its joints.”
This is the move that converts the constructionist’s best evidence into the nativist’s expectation, and it is why the wiki should not file Scarantino as simply “pro basic emotions.” He concedes more to Barrett than Friedman does — folk categories are not natural kinds, signatures were never coming — while keeping evolved, causally potent, hard-wired affect programs. Compare LeDoux, who keeps the circuits by renaming them; Scarantino keeps them by reconceiving what a kind is. Two different escapes from the same meta-analytic pressure (lindquist-2012-brain-basis-of-emotion).
The argument against letting common sense legislate
LeDoux’s stated reason for surrendering “fear” to the feelings camp is pragmatic: the gravitational pull of common sense is so strong that claims about implicit emotion get read as claims about explicit emotion. His example — “although I consistently emphasized that the amygdala circuits operate nonconsciously, I was often described in both lay and scientific contexts as having shown how feelings of fear emerge from the amygdala.”
Scarantino’s three replies:
- Common sense is not univocal. Folk psychology routinely uses emotions as motive latent variables that explain behaviour — “Maria slapped Lucas because she was angry, Tom shot the bear because he was afraid.” The identification of emotions with feelings is one folk usage, not the folk usage.
- Common sense shouldn’t legislate anyway. Quine: no conceptual truths in science; even Euclidean parallels went. “Why should the affective sciences be the exception to this rule?” The analogy offered: lay confusion at Pluto’s 2006 demotion is no reason for astronomers to revise the definition of planet.
- The pragmatic argument may run backwards. If most contemporary affective science already separates emotion from feeling, then reserving “fear” for the feeling is the more confusing convention — inviting readers to translate LeDoux’s cortical-circuit results about fear into their own scheme and “mistakenly conclude that LeDoux has shown that the motivational system which evolved to deal with dangers is cortically implemented.”
Reply (3) is the one with teeth for this wiki, and Scarantino backs it with a survey showing the emotion/feeling split is the cross-traditional majority: Ekman (feelings not a “sine qua non” of basic emotions), Damasio (emotion = body-state changes with the function of initiating responses; feeling = the experience of those changes, which “open[s] the door for some measure of willful control” — see somatic-marker-hypothesis), Scherer (subjective experience is one of five components; any three co-occurring suffice, so some emotions have no feeling at all), Russell (core affect need only be accessible, not accessed, for an emotion episode — see core-affect).
That list is the paper’s strongest empirical-ish contribution: LeDoux’s terminological reform is presented as tracking common sense, and Scarantino shows it cuts against the field’s actual working consensus.
Mobbs’ middle position
Recorded because the themed issue’s other editor occupies the cell between the two principals. Mobbs (2018) distinguishes reactive fear (automatic, reflex-like — freezing, fight-or-flight; Fanselow & Lester’s circa-strike defense) from cognitive fear (involves a feeling, and arises when there is time to “organize and strategize escape”; if the threat is potential or non-imminent, the result is anxiety instead).
Unlike LeDoux, Mobbs will use a qualified “fear” for subcortical defensive circuits understood behaviourally. Scarantino accepts the reflex-like/strategizing distinction but rejects the definitional bundling of a feeling into cognitive fear: “I continue to think that we should not mix phenomenological and motivational dimensions.” He splits it into the interplay between basic fear and cognitive regulation (motivational) and the feeling of fear (experiential), noting the feeling can in principle attach to reactive fear too, whenever it is targeted by a higher-order representation.
Where this leaves the wiki
Filed as the new debate what-should-emotion-terms-refer-to, which is genuinely distinct from the wiki’s existing three:
- locationist-vs-constructionist-brain-emotion asks whether emotion categories map to circuits.
- where-are-feelings-constituted asks where feelings happen.
- can-we-know-animal-feelings asks whether we can attribute them to animals.
- This one asks what the word “emotion” should pick out — and Scarantino’s contribution is showing that a live dispute between two frameworks can be almost entirely this, while the neuroscience underneath is shared.
Not a hard contradiction with existing content. Scarantino contradicts nothing the wiki asserts about LeDoux; he agrees with the wiki’s reading of LeDoux’s architecture and disputes its labels. Where he does bear on existing pages (basic-emotions, autonomic-specificity-of-emotion), it is by adding a position, not by overturning a recorded claim.
One correction the paper forces on the wiki’s framing, though. can-we-know-animal-feelings records LeDoux’s DLPFC-based skepticism as hostage to where-are-feelings-constituted — if feelings are insular, other mammals have insula and the premise fails. Scarantino tightens this: he cites LeDoux & Brown (2017) approvingly for locating higher-order representation in “general networks of cognition, which involve regions in the prefrontal cortex, posterior parietal cortex, and insular cortex.” So LeDoux’s own later, co-authored position lists the insula among the substrates of emotional consciousness — which the 2012 paper does not. Whether that is a shift, or the insula entering as a mere workspace component rather than as Craig’s medium of feeling, cannot be settled from this commentary. Flagged on where-are-feelings-constituted and higher-order-theory-of-consciousness; resolving it needs LeDoux & Brown (2017) itself, which is not in raw/.