The Emotional Trigger Record

I. The Mechanism

Information can be designed to produce a specific emotional state in its recipient — not as a side effect of its content but as its primary function. The emotional state is not incidental to the information's purpose. It is the distribution mechanism. Content that triggers outrage spreads faster than content that informs. Content that triggers fear holds attention longer than content that educates. Content that triggers in-group solidarity produces more engagement than content that invites cross-group deliberation.

This is not an observation about human nature. It is a design specification. The internal research documents disclosed by Frances Haugen in 2021 established that Facebook's recommendation algorithm was calibrated to amplify content that produced emotional activation — specifically, content triggering outrage, moral indignation, and in-group threat response — because emotionally activated content generated more engagement, more time-on-platform, and more advertising revenue than informationally rich but emotionally neutral content.

The emotion is the distribution mechanism. The content is the payload.

II. The Facebook Internal Research

The Haugen disclosure included internal research documents demonstrating three findings that Facebook's own data science team had documented and that the company's leadership had reviewed:

Finding 1: Outrage amplification. The recommendation algorithm's engagement-optimisation function reliably surfaced content that triggered moral outrage, because outrage-triggering content produced the highest engagement metrics (comments, shares, reactions, time-on-post). The algorithm did not select for outrage deliberately — it selected for engagement, and outrage was the most engagement-efficient emotional state available. The optimisation function and the outrage amplification function were structurally identical.

Finding 2: Anger as the dominant reaction. Internal analysis showed that the "angry" reaction (the emoji response Facebook added in 2016) was weighted five times higher than the "like" reaction in the algorithm's engagement scoring. Content that made users angry was algorithmically promoted over content that users merely found informative or pleasant. This was a design decision — the weighting was set by Facebook's engineering team, not by the users.

Finding 3: Political content amplification. Internal research documented that political content — which disproportionately triggers outrage and in-group threat response — was amplified by the recommendation system to a degree that the company's own researchers flagged as problematic for democratic discourse. The company's Civic Integrity team proposed interventions. The interventions were reviewed by leadership and not implemented, because implementing them would have reduced engagement metrics.

III. The Neuroscience of Affective Engineering

The REBUS model documented in the Neural Complexity sciences page explains why affective engineering works at the neurological level.

The amygdala hijack. The amygdala processes threat-related stimuli approximately 100 milliseconds faster than the prefrontal cortex can evaluate them. Content designed to trigger fear or in-group threat response activates the amygdala before the analytical processing systems have time to evaluate whether the threat is genuine. By the time the prefrontal cortex could perform an evidence-based evaluation, the amygdala has already produced an emotional response — a response that the user experiences as a genuine reaction to genuine information, because the emotional response arrived first and shaped the subsequent evaluation.

Salience network capture. The salience network — documented in AOA-006 (The Cognitive Audit) as one of the four cognitive capacities required for accountability — determines what feels important. Affective engineering hijacks the salience network by substituting emotional intensity for informational significance as the criterion for attention allocation. A forensic accounting finding (informationally significant, emotionally neutral) is salience-invisible. A tribal threat narrative (informationally thin, emotionally intense) is salience-dominant. The salience network, retrained by repeated exposure to affectively engineered content, produces a population that cannot distinguish significant information from significant emotion.

Prior installation. Each exposure to affectively engineered content tightens the prior structures the REBUS model describes. The outrage-triggering narrative about an out-group becomes a precision-weighted prior: the brain expects the out-group to behave in the way the narrative predicted. Bottom-up signals that contradict the prior (evidence of the out-group's actual behaviour) are suppressed by the prior-weighting system. The prior was installed not through deliberate indoctrination but through repeated algorithmic exposure to content optimised for the emotional state that the prior requires.

IV. The Red Flag Filter

This paper proposes a diagnostic question for evaluating information environments:

"What specific emotional state is this content designed to evoke, and who benefits from the target audience feeling that state?"

The Red Flag Filter does not ask whether the content is true or false. Affectively engineered content is often factually accurate — the outrage-triggering story about a political opponent's statement may quote the statement correctly. The question is not accuracy. The question is function: is the content's primary function to inform, or is its primary function to produce an emotional state that benefits a third party?

The distinction is structural. Informational content produces emotional responses as a byproduct of the information it conveys. Affectively engineered content conveys information as a byproduct of the emotional state it produces. The sequence is reversed. The emotion comes first; the information is selected to serve the emotion.

V. The Polarisation Consequence

The Polarisation Cascade series (PC-001 through PC-005) documented the population-level consequence of affective engineering: epistemic fragmentation. This paper adds the neurological mechanism.

When a population's information environment is dominated by affectively engineered content, the amygdala's threat-bias calibration shifts across the entire population — but it shifts differently for different demographic segments, because the algorithm optimises for each segment's specific emotional triggers. The result: different segments of the population are having their amygdalae calibrated against different threat profiles. Each segment perceives a different set of threats as genuine. Each segment's prior structures are tightened against a different set of out-groups.

The "polarisation" label understates what is happening. The population is not merely disagreeing more intensely about the same facts. It is being neurologically recalibrated, segment by segment, to perceive different threats as real — producing the condition in which shared reality becomes structurally unavailable because different segments' salience networks are flagging different phenomena as important.

VI. The Advertiser's Interest

The attention economy series (AE-001 through AE-005) documented why platforms optimise for engagement: advertising revenue is a function of time-on-platform and targeting precision. Affective engineering serves both.

Emotionally activated users spend more time on-platform (the engagement function). Emotionally activated users also produce more behavioural data — their reactions, shares, comments, and content consumption patterns during emotional activation reveal their values, fears, group loyalties, and vulnerability profiles with greater precision than their behaviour during neutral browsing. The emotional state is not just the distribution mechanism for content. It is the data-generation mechanism for advertising targeting.

The user's outrage is not a cost the platform tolerates. It is the product the platform sells.

Named Condition

Affective Engineering — the deliberate design of information content to activate specific emotional states in the target audience, where the emotional state is the distribution mechanism rather than an incidental byproduct of the information conveyed. Identifiable through the Red Flag Filter: when the content's primary function is to produce an emotional state that benefits a third party (platform, advertiser, political actor) rather than to inform the audience, the content is affectively engineered regardless of its factual accuracy.