YAWN Research · Working Paper 001Version 1.0
Interception
Awareness between emotion and action
Interception is the moment an activated pattern becomes observable before it becomes behavior.
- Author
- David Forman
- Role
- principal of @dave
- Published
- Status
- Public working paper · not peer reviewed
Abstract
Narrowed thesis
A reaction is a trajectory, not a verdict.
This paper proposes Interception as a measurable event: an agent detects an emerging, emotion-linked action tendency, represents it as an object of attention, and changes the probability of what happens next.
The evidence supports a narrower thesis than the metaphor first suggests. Affect labeling, decentering, self-distancing, reappraisal, and well-timed pauses can alter components of an emotional or behavioral response. They do not reliably erase emotion, guarantee wise action, or prove that emotion uniquely frees humans from learned programming.
Definition
Operational definition
What must happen for an Interception to count?
Feeling calmer is not sufficient proof. The operational question is whether awareness changed the next selectable move.
Momentary Interception
- Detect an activated emotional state and its action tendency.
- Represent the state as a revisable object of attention.
- Reopen at least one behavior option before commitment.
Completed YAWN cycle
- Choose one bounded move.
- Observe a consequence that can count as proof.
- Update confidence and the future response policy without rewriting history.
Mechanism
The possibility window
A response begins inside a relationship between an agent and a situation. Observation is not outside that system; it is the system gaining an inspectable model of its present trajectory.
Relationship is structurally first; observation is operationally first.
- 01Relationship event
Something changes in the agent–arena relationship.
- 02Affective update
Body state, appraisal, prediction, attention, and action readiness reorganize.
- 03Meta-awareness
The active state becomes distinguishable from the world it is modeling.
- 04Bounded move
A label, delay, question, reappraisal, exit, request, or deliberate action can compete.
- 05Proof and update
Consequences update the relationship, confidence, and future response policy.
Calibration
Claim register · predict before you read
How much of the idea survives contact with evidence?
Before you read the evidence: how well does each claim hold up? The register grades you.
People can sometimes label, distance from, decenter from, or reappraise an emotional response in ways that change physiology, experience, or behavior.
Meta-awareness plus a response gate can sometimes loosen the coupling between an activated state and its next action.
The complete YAWN Interception protocol improves goal-consistent behavior beyond a time-matched neutral pause.
Intense emotion generally makes people more flexible; it often narrows attention and strengthens a prepotent response.
Emotion uniquely wakes humans from programming; nonhuman animals also show flexible and metacognitive-like behavior.
These are qualitative synthesis judgments. They are not frequencies, effect sizes, or calibrated empirical probabilities.
Evidence
Open evidence notebook
Eight inspectable studies, two levels of relevance.
No included study tests Interception as operationally defined. These studies test component mechanisms and design analogues.
Component-mechanism evidence
Pauses, labeling, distancing, decentering, and a useful null.
104 participants + 58-participant replication · 6,262 trials
A five-second break in conflict
Forced breaks of 5, 10, or 15 seconds reduced retaliatory sound-blast intensity on high-negative-affect trials. With immediate responding, negative expression predicted aggression (r = .33); during a break, the association was approximately zero (r = .01).
A brief gate can weaken the immediate coupling between negative affect and retaliation.
Limit: This is a laboratory aggression task, not durable habit change, and it must not be generalized to abuse or coercive control.
88 spider-fearful adults · randomized experiment
Affect labeling during exposure
One week after exposure, the affect-labeling group showed a larger decrease in skin-conductance response than reappraisal, distraction, or exposure alone (pairwise d = 0.64–0.85). Self-reported fear did not differ between groups.
Naming an emotion can alter physiological responding and may influence approach in some contexts.
Limit: The approach advantage was only marginal versus distraction, and the result does not show that labeling makes fear disappear.
63 + 79 adults · two experiments
When labeling reverses direction
Labeling reduced distress for high-intensity negative images, but in the second experiment increased distress for low-intensity images (strategy × intensity partial η² = .23). Immediate versus delayed labeling did not reliably change the effect.
Affect labeling is context-sensitive rather than a universal off switch; intensity belongs inside the model.
Limit: Image-viewing distress is not equivalent to consequential real-world behavior.
177 online + 144 laboratory participants · preregistered replication
Self-distancing during anger
Under immersed reflection, recalled anger impaired learning to avoid disadvantageous choices relative to fear. The anger-specific disadvantage was attenuated under self-distancing in both experiments.
Perspective can change how an emotional state couples to later decisions.
Limit: The induced emotions were mild and the gambling task is only a proxy for consequential real-world choice.
345 adults · up to 42 phone surveys across one week
Decentering in daily life
Momentary decentering and affect were mutually reinforcing over time. Greater decentering was associated with less persistence of negative affect and dysphoria, while several prospective results were mixed.
An observer stance tracks greater emotional mobility in ordinary life.
Limit: The design is observational and cannot establish that awareness caused the change.
50 participants for behavior/craving · 49 for EEG · randomized experiment
A useful urge-surfing null
After cigarette-cue exposure, a 15-minute mindfulness exercise did not improve Go/NoGo accuracy or reaction time versus usual coping. The craving difference was trend-level, not significant; a neural difference appeared without demonstrated behavioral benefit.
A brain measure is not enough; action and consequence must remain the proof target.
Limit: The brief intervention, student-smoker sample, and single session constrain generalization.
Mechanism and engineering analogues
Useful for measurement and design; not direct emotion-action proof.
Published analysis N = 30 · OpenNeuro contains 34 participant records
Reappraisal as a distributed mechanism
During reappraisal, right ventrolateral prefrontal activity related to reported emotion change through separable amygdala and nucleus-accumbens pathways; the modeled pathways jointly explained roughly half the variance in self-reported affect change.
The result is consistent with distributed appraisal and valuation pathways rather than a single rational switch.
Limit: This is a neural mediation analogue, not behavioral proof of Interception.
37 analyzed adults · 6,061 available randomized points from 8,274 recorded decision-point rows
Intervening at a decision point
HeartSteps walking suggestions increased near-term activity by about 24% (about 59 steps; p = .02) on average, while the effect declined across the six-week study.
Micro-randomization is an engineering precedent for testing whether a prompt changes the next behavior at the moment it is delivered.
Limit: HeartSteps is not emotion evidence, and its public repository contains data rather than the article's analysis code.
Catalog
An open emotion-language corpus
Labels are handles, not hidden essences.
GoEmotions contains 58,009 English Reddit comments annotated with one or more of 27 emotion labels or neutral. On its 54,263-comment agreement-filtered split, the published BERT baseline reached macro-F1 .46 across all 28 labels.
This is an inspectable vocabulary for expressed language, not an ontology of private feelings. Its English-Reddit domain and annotation process introduce cultural, platform, and agreement bias.
Worked emotional-catalog record
Before sending the message
Face heat, tight chest, rapid speech; unpleasant, high activation, intensity 8 of 10.
Ranked possibilities: anger, then hurt, then fear. The labels are provisional and can overlap.
Observed: “message unanswered for two days.” Interpretation: “I am being dismissed.”
Urge: accuse and send now. Relationship goal: obtain clarity while preserving respect.
Save the draft, pause, and ask one factual question before sending.
Did the message, tone, or outcome change? What should update the next response?
The goal is not to suppress anger. It is to preserve the information in anger without granting its first action tendency unilateral authority.
Protocol
Falsifiable next step
Turn the metaphor into a micro-randomized study.
A four-arm micro-randomized study can test the full protocol against recording and time alone.
Record only
Emotion, urge, and intended action.
Neutral pause
A time-matched delay without emotion framing.
Label + gate
Body signal, emotion hypothesis, urge, and pause.
Full Interception
Label, relationship goal, two alternatives, bounded move, and proof.
- Primary: whether the initially intended action occurred within 15 minutes.
- Secondary: later-rated alignment with the person's stated relationship goal.
- Learning: which states, intensities, contexts, and prompts predict a useful opening.
Boundaries
What this paper does not claim
Agency needs more than awareness.
01Emotion is not noise to conquer. It changes attention, prediction, learning, and action readiness—sometimes usefully, sometimes dangerously.
02Awareness is not automatically agency. Viable alternatives, capacity, permission, and an environment that permits action also matter.
03One intercepted moment does not rewrite a habit. Durable change requires repetition, consequence, and consolidation.
04Failure to intercept under threat, trauma, exhaustion, intoxication, or coercion is not a moral failure. Control capacity is state- and context-dependent.
05Emotion labels are provisional models shaped by culture, context, language, and learning.
06This conceptual paper is not medical advice, therapy, diagnosis, or crisis care. Acute risk calls for safety, support, and qualified help—not an interface prompt.
Sources
Background references
Primary sources, open where possible.
Study-specific paper, data, code, and project links appear on each evidence card. These sources frame the broader mechanism and its boundaries.
- 01Bernstein, A. et al. (2015). Decentering and related constructs: A critical review and metacognitive processes model.
- 02Lieberman, M. D. et al. (2007). Putting feelings into words: Affect labeling disrupts amygdala activity in response to affective stimuli.
- 03Khalsa, S. S. et al. (2018). Interoception and mental health: A roadmap.
- 04Barrett, L. F. (2017). The theory of constructed emotion: An active inference account of interoception and categorization.
- 05Wessel, J. R. et al. (2016). Surprise disrupts cognition via a fronto-basal ganglia suppressive mechanism.
- 06Rosati, A. G., & Santos, L. R. (2016). Spontaneous metacognition in rhesus monkeys.
