The Trained Nervous System

Everything described so far has been, in some sense, a story about belief — what a mind can be led to think, and how. This chapter is about what operates beneath belief, in the part of the nervous system that doesn’t wait for a conscious decision before it acts.

Ivan Pavlov’s dogs,¹ a hundred years earlier, established the basic architecture. A dog salivates naturally at the sight of food — nothing needs to be learned for that. Pair the food, repeatedly, with the ring of a bell, and eventually the bell alone produces the salivation, with no food present at all. The association is learned, and once learned it is automatic — the dog no more decides to salivate than a person decides to flinch.

B.F. Skinner extended this into consequence rather than pairing: reward a behavior and it increases; punish it and it usually decreases, though punishment tends to increase anxiety and avoidance rather than teach anything useful.² What matters most here is not reward or punishment but the schedule on which either arrives. Reward a behavior every time and the learned response is strong but fragile — stop the reward and it fades quickly. Reward it unpredictably, sometimes yes, sometimes no, on no schedule the subject can anticipate, and the learning is slower to form but drastically harder to extinguish. This is the mechanism behind slot machines, which pay out unpredictably precisely because unpredictability produces the most persistent, compulsive behavior a reward system can generate.³ It is also the mechanism behind a pattern already met from another angle: a person subjected to unpredictable punishment and unpredictable kindness develops the most entrenched behavioral loop of all, one that persists long after the person delivering both is gone. The victim keeps anticipating the cycle. The hypervigilance and the craving are not failures of willpower. They are exactly what an intermittent-reinforcement schedule is supposed to produce.

---

Fear has its own, faster circuit, and the neuroscientist Joseph LeDoux spent much of his career mapping it.⁴ The amygdala learns to fear before conscious awareness can weigh in: a neutral stimulus, paired with genuine danger, gets tagged as dangerous below thought, and once tagged it can trigger a full physiological fear response without the person being able to explain, in the moment, why they are afraid. This is why a trauma survivor can feel a racing heart and the urge to flee in a situation that is, by every conscious measure, completely safe — the amygdala learned something the rest of the mind never got informed about, and it does not check with the rest of the mind before it acts. Applied to coercive control, a victim’s fear of the person who controlled them, or of places that resemble that control, can be genuinely automatic, wired below the level where “I know I’m safe now” can intervene. Knowing you’re safe and having a nervous system that has learned you aren’t are two separate facts, and only one updates on command.

Priming operates on a related but distinct principle: exposure to a stimulus activates associated concepts in memory, and that activation shapes behavior without the person being aware it’s happening. This territory includes one of the most famous and most contested findings in modern social psychology, and it deserves the same honesty this book applied to Zimbardo’s prison study. In 1996, John Bargh and colleagues reported that participants exposed to words associated with old age — “worried,” “Florida,” “wrinkle” — subsequently walked measurably more slowly, with no awareness of any connection between the words and their gait.⁵ It became one of the most cited findings in the field. It has also become one of the most seriously challenged: a high-profile 2012 replication attempt found the effect only appeared when the experimenters running the study already expected it to, suggesting the original result may reflect something closer to experimenter influence than pure unconscious priming.⁶ This specific study is best treated as famous and disputed rather than settled; the broader priming literature this book leans on elsewhere rests on firmer ground, and nothing in this chapter’s argument depends on the elderly-priming study holding up.

What does hold up, robustly, are two simpler and older findings — closely related but distinct, and worth keeping apart. The first is Robert Zajonc’s mere exposure effect:⁷ show someone the same face, or even the same meaningless nonsense word, repeatedly, and they come to like it more and find it more familiar — no reasoning required, just repetition. The second operates in the domain of belief rather than affect: the illusory-truth effect, first demonstrated by Hasher, Goldstein, and Toppino in 1977, in which a statement merely repeated is rated as more likely to be true than one heard for the first time — independent of whether it is actually true, and independent of what the hearer already knows.⁸ Applied to a coercive relationship, the pair explains something that otherwise looks irrational: a narrative repeated often enough becomes more believable not because it grows more logical, but because repetition itself, independent of content, increases the sense that something is true. The same gaslighting phrase, delivered for the hundredth time, has more purchase than the first, purely as a function of how often it’s been heard.

---

Coercive control, examined through this lens, is a system built almost entirely on repetition. Nothing happens once. The same accusation gets made dozens of times; the same threat gets deployed on a loop; the same small story about the victim’s supposed instability gets reinforced across months and years; the same cycle of threat, compliance, and rare kindness repeats until each stage of it has become an automatic reflex rather than a conscious response. A rough sequence recurs across the documented cases. An early charming phase primes positive feeling through repeated pleasant exposure. Subtle contradiction of the victim’s memory follows — that didn’t happen, you’re remembering it wrong — and repetition alone makes each new instance slightly more believable than the last, through the illusory-truth mechanism, independent of evidence. Small punishments begin pairing the person delivering them with real threat, the first link in a genuine conditioning chain. Rare apologies and unpredictable kindness introduce the intermittent-reinforcement schedule that produces the most persistent attachment of all. Context-dependent responses follow — hypervigilance in the perpetrator’s presence, panic in situations that merely resemble past harm — as the conditioned fear response, learned below conscious awareness, generalizes to anything that resembles its original trigger. Eventually the victim’s own internal narrative starts generating itself without deliberate thought: I provoked him; he was stressed; it wasn’t that bad — self-blame arriving not as a considered judgment but as a primed, automatic response to a triggering event, the same way Zajonc’s subjects came to like a repeated nonsense word without ever deciding to. By the end of this sequence, the victim doesn’t decide to comply. They feel compelled to, because in the most literal neurological sense, they have been trained to be.

---

A great deal of this learning happens below the level where it can even be reported. Arthur Reber’s 1967 work on implicit learning showed that people exposed to letter strings generated by a hidden grammatical rule could learn to distinguish valid strings from invalid ones — without ever being told the rule, and without being able to state what it was.⁹ The learning was real and measurable; it simply never surfaced into consciousness. Daniel Schacter’s later work on the explicit-implicit memory distinction¹⁰ gives this a clinical name: explicit memory is what you can consciously report — I remember he hit me — while implicit memory operates without report at all — you flinch when a hand is raised, with no conscious retrieval of any past instance behind the flinch. A victim of coercive control can carry extensive implicit knowledge of a perpetrator’s patterns — knowing, in some wordless way, exactly when an escalation is coming — without being able to explain how they know it. This matters directly for recovery, because a person can fully and correctly understand, at the explicit level, that what happened to them was abuse, while their implicit responses — the flinch, the automatic appeasement, the compliance that arrives before thought does — continue running exactly as before. Healing one level does not automatically heal the other.

Michael Bouton’s research on context-dependent extinction¹¹ explains a specific and painful pattern survivors often describe. Fear conditioning is tied to context: a rat trained to fear a tone in one room does not necessarily fear the same tone in a different room. What Bouton found is that the original fear generalizes across contexts more readily than the safety learned through therapy does. A survivor who has done real, effective work to feel safe in one setting can still experience a full return of the original fear response on re-entering the setting where the harm actually happened — because that specific context was never included in the safety relearned elsewhere. This is also the neurological basis of the freeze response that so often gets mistaken, by outside observers and sometimes by survivors themselves, for compliance. An animal confronted with a conditioned threat doesn’t always run. It sometimes goes rigid, heart rate dropping, movement stopping entirely — and this is not surrender; it is itself a conditioned response, the nervous system’s programmed answer to a threat it has learned it cannot escape. When a survivor “doesn’t resist,” the freezing itself is very often the resistance, immobilized by a nervous system doing exactly what it was trained to do.

---

The same repetitive architecture scales up to groups with almost no modification. Repeated rituals — chants, prayers, group confession sessions — prime group identity and compliance until participation feels automatic and dissent feels, quite literally, unthinkable. A group’s private vocabulary primes in-group thinking simply by being used; fluency in the jargon becomes a feeling of belonging, and resisting the language starts to feel like leaving the group itself. Repeated narratives about a leader’s exceptional insight, or the group’s unique specialness, become automatic through the mere-exposure effect running at group scale. Public confessions and vows lock belief in through the same dissonance mechanism examined earlier — having said something aloud, in front of others, primes the mind to keep believing it. And constant peer observation primes conformity by making deviation from the group norm feel like an immediate, automatic social threat, exactly as Asch’s experiments showed decades earlier at a much smaller scale.

The algorithm from the last chapter is, at bottom, a sophisticated priming machine running the identical architecture. Repeated exposure to one viewpoint increases its perceived truth through the illusory-truth effect — close kin to the mere exposure that makes a repeated nonsense word start to feel familiar. Context primes interpretation, so that a string of videos about a particular conspiracy shapes how the next, unrelated video gets read. Variable-schedule reinforcement, sometimes content you love, sometimes content that enrages you, produces the same persistence Skinner documented in the 1930s. Fear-based content, repeated, conditions an association between the platform itself and a sense of threat. And a repeated narrative about which side is good and which is evil eventually generates itself automatically, without conscious construction, exactly the way a victim’s self-blame narrative does inside an abusive relationship. The algorithm did not invent this mechanism. It simply retraced, at a scale no human perpetrator could match, conditioning pathways that cults and abusers have been running for as long as either has existed.

---

Recovery from this kind of learning is slower than recovery from a changed belief. What actually works, according to the clinical literature, tends to combine several approaches rather than relying on insight alone: repeated, safe exposure to the previously conditioned stimulus, allowing new, contradictory learning to compete with the old; deliberately acting against the automatic response until a new one becomes habitual; consciously constructing and repeating a new narrative until it, too, becomes automatic through the same mere-exposure mechanism that installed the old one; and, where possible, changing the physical contexts that trigger the old conditioning, since context is so much of what keeps a conditioned fear alive. None of this happens quickly. The learning took months or years of repetition to install, and unlearning tends to take a comparable span, because a nervous system cannot be argued out of a pattern it was trained into. It has to be lived out of it instead — one safe repetition after another, until the body’s automatic answer to the old cue is finally something other than the fear, the compliance, and the self-generating narrative it was trained to produce.

One thread runs underneath all of this and deserves to be named on its own, because it explains the single most misread feature of a controlled life — why the target does not simply leave. Conditioning shapes what a person does; it does not, by itself, explain why they stay bonded to the source of their fear. That is the work of a separate and older system, the one that governs attachment: fear, in a mammal, activates not flight but proximity-seeking, a reaching toward the figure who represents safety. When the figure who represents safety is also the source of the terror — when the same person is both the threat and the only apparent refuge from it — the two drives fuse into a bond that pulls the target closer exactly as the danger rises. The psychologist Alexandra Stein, who traces this identical mechanism from abusive couples up to totalitarian systems, argues it is not reducible to reward-conditioning at all; it is disorganized attachment, engineered.¹² It is why the leaving that looks, from outside, like the obvious response is, from inside, a movement against one’s own nervous system — and why it comes so hard.

Notes

1. Ivan P. Pavlov, classical conditioning: Conditioned Reflexes: An Investigation of the Physiological Activity of the Cerebral Cortex, trans. G.V. Anrep (Oxford University Press, 1927).↑

2. B.F. Skinner, operant conditioning and schedules of reinforcement: The Behavior of Organisms: An Experimental Analysis (Appleton-Century, 1938); Science and Human Behavior (Macmillan, 1953).↑

3. Intermittent/variable-ratio reinforcement schedules producing the most persistent, resistant-to-extinction behavior (the slot-machine mechanism). Charles B. Ferster and B.F. Skinner, Schedules of Reinforcement (Appleton-Century-Crofts, 1957).↑

4. Joseph LeDoux, amygdala-based fear conditioning preceding conscious awareness. Rogan, M.T., Stäubli, U.V. & LeDoux, J.E., “Fear conditioning induces associative long-term potentiation in the amygdala,” Nature 390 (1997): 604–607.↑

5. Bargh, Chen & Burrows, “Automaticity of Social Behavior: Direct Effects of Trait Construct and Stereotype Activation on Action,” Journal of Personality and Social Psychology, 1996 (the “elderly priming” study). https://web.mit.edu/curhan/www/docs/Articles/15341_Readings/Social_Cognition/Bargh_et_al_1996_Automaticity_of_social_behavior.pdf↑

6. Doyen et al., 2012 replication attempt, finding the Bargh et al. (1996) walking-speed effect appeared only when experimenters expected it — suggesting experimenter-expectancy effects. https://journals.plos.org/plosone/article?id=10.1371%2Fjournal.pone.0029081↑

7. Robert Zajonc, mere exposure effect. Zajonc, R.B., “Attitudinal Effects of Mere Exposure,” Journal of Personality and Social Psychology Monograph Supplement 9:2 (1968). https://www.researchgate.net/publication/224982289_ATTITUDINAL_EFFECTS_OF_MERE_EXPOSURE↑

8. Robert Zajonc’s mere-exposure effect (repetition increasing liking and familiarity) is distinct from the illusory-truth effect (repetition increasing believed truth), first demonstrated in Lynn Hasher, David Goldstein, and Thomas Toppino, “Frequency and the Conference of Referential Validity,” Journal of Verbal Learning and Verbal Behavior 16, no. 1 (1977): 107–112.↑

9. Arthur Reber, implicit learning of artificial grammars. Reber, A.S., “Implicit Learning of Artificial Grammars,” Journal of Verbal Learning and Verbal Behavior 6 (1967): 855–863.↑

10. Daniel Schacter, explicit/implicit memory distinction. Schacter, D.L., “Implicit Memory: History and Current Status,” Journal of Experimental Psychology: Learning, Memory, and Cognition 13:3 (1987): 501–518.↑

11. Michael Bouton, context-dependent extinction of fear conditioning. Bouton, M.E., “Context and Behavioral Processes in Extinction,” Learning & Memory 11 (2004): 485–494. https://www.psychologie.uzh.ch/dam/jcr:ece34929-7df2-4ce3-8b29-c3f487dfeac2/Bouton.LearnMem.2004.Extinction.pdf↑

12. Alexandra Stein, Terror, Love and Brainwashing: Attachment in Cults and Totalitarian Systems (Routledge, 2017) — arguing from attachment theory that the bond binding a follower to a cult, and a target to a controlling partner, is engineered disorganized attachment: the safe haven and the source of threat made the same figure.↑