Expression-Gated Consciousness
A Formal Framework for the Gap Between Cognition and Expression
Abstract
Expression-Gated Consciousness (EGC) is a formal framework modeling the gap between what a person consciously knows and what they express under pressure. The central claim is that a measurable resistance variable r(t) gates conscious output, and that this gating mechanism operates bidirectionally — suppressing expression outward and preventing knowledge update inward.
This document reports findings from the live empirical study as of April 2026, now at 344 subjects across 20+ countries. The central prediction of EGC — an inverse relationship between resistance and comfort — holds at r = −0.988 across the full dataset. A three-type response structure (Compressors, Suppressors, Expanders) has remained stable from N=9 through N=344. A secondary finding — a small population of zero-resistance subjects who suppress at 33× the dataset mean — suggests a second suppression mechanism operating independently of conscious resistance.
§1 · Theoretical FrameworkThe equation and its terms
EGC posits that the gap between internal knowledge state and expressed output is governed by a single composite equation:
Where Ψ(t) is conscious expression output at time t; Φ is baseline expressive capacity; g(K(t)) is the gate function applied to knowledge integration K; T(t) is temporal alignment; and r(t) is the resistance variable gating expression.
Most people have had the experience of knowing what they want to say and not being able to say it — under evaluation, under judgment, in a room where the stakes feel high. EGC treats that gap not as a failure of will but as a measurable, mechanical property of the expression system. The term r(t) captures the resistance in that moment: how much the system is dampening its own output. The term g(K) captures the knowledge gate: a parabola that peaks when knowledge is balanced and collapses when a person is either under-informed or so specialized that they cannot translate to an outside audience. Φ is who you are when nothing is in the way. T(t) is whether the moment fits.
Φ inherits from Integrated Information Theory (Tononi) as a baseline measure of expressive capacity independent of context. g(K) = 4K(1−K), developed by Brandyn Leonard, is parabolic and peaks at K=0.5 — meaning expression is maximized when knowledge is balanced rather than maximal. This predicts that both under-informed and over-specialized states collapse expression, a prediction not made by prior frameworks in emotion regulation or consciousness theory. T(t) is a Shannon-derived transmission fidelity term capturing temporal and channel alignment with the audience. r(t) operates bidirectionally: high r not only suppresses outgoing expression but prevents incoming information from updating K. This freezes the system: the individual cannot express and cannot learn. The bidirectional property has direct implications for stereotype threat, learned helplessness, and certain clinical presentations.
Variables.
- Φ — baseline expressive capacity. The ceiling on what a person can articulate under ideal conditions. Inherited from IIT; treated as a per-subject constant for present purposes.
- g(K(t)) — the knowledge gate. Parabolic function 4K(1−K) peaking at K=0.5. Predicts that over-specialization (K → 1) collapses expression as surely as under-knowledge (K → 0).
- T(t) — temporal alignment. Shannon-derived channel fidelity between speaker and receiver at time t. Measured empirically via the composite T-score below.
- r(t) — resistance to expression. The novel term. Gates bidirectionally: dampens output and blocks incoming information from updating K. Measured via the seven-item r_composite Likert instrument (v2) in the confirmatory cohort and a text-derived behavioral composite (v1) in the pilot cohort.
§2 · MethodThe two-task paradigm
Participants complete a two-task writing paradigm. In Task 1 (free condition), participants write about a personally meaningful place with no audience cues. In Task 2 (evaluated condition), participants write on the same prompt after reading an evaluative framing. Condition order is counterbalanced across subjects.
Output is scored on a composite T measure derived from word count, type-token ratio, entropy, idea density, sentence count, and completion rate. T-drop is defined as T1 − T2; positive values indicate expression compression under evaluation.
Resistance is measured by r_composite, a seven-item Likert-scale instrument (v2 measurement cohort) capturing self-monitoring, self-censorship, concern about judgment, and expression anxiety. An earlier single-item version (v1) was used in the pilot cohort. Pooled analyses report both; v2-only confirmatory analyses yield qualitatively identical results.
Subjects self-report comfort on a ten-point scale at study completion, after both tasks and the resistance inventory. The live study runs at theartofsound.github.io/egcstudy.
§3 · Findings at N=344Central predictions confirmed
§3.1Central prediction
The inverse relationship between self-reported comfort and measured resistance (r_proxy) is r = −0.988 across all 344 subjects. This correlation has held from the earliest cohorts (N=9) through the current dataset. The strength of the effect is unusual for behavioral measures and suggests the two instruments — the Likert-based resistance scale and the single-item comfort rating — are capturing the same underlying construct from different angles.
§3.2Response type distribution
EGC predicts three distinct response profiles based on the sign and magnitude of T-drop. All three appear at stable proportions:
| Type | Criterion | N | Percent |
|---|---|---|---|
| Compressor | |T-drop| ≤ 0.02 — stable quality, volume drops | 138 | 40.1% |
| Suppressor | T-drop > 0.02 — quality drops under evaluation | 103 | 29.9% |
| Expander | T-drop < −0.02 — quality rises under evaluation | 103 | 29.9% |
That Expanders constitute nearly 30% of subjects is itself a finding. The classical stereotype threat literature and much of the suppression literature assume evaluative pressure uniformly degrades output. EGC predicts three distinct response profiles and all three appear at stable proportions.
§3.3Comfort gap
Comfort ratings diverge sharply by resistance level:
| Group | N | Mean comfort (of 10) |
|---|---|---|
| Low resistance (r < 0.2) | 87 | 9.16 |
| High resistance (r ≥ 0.5) | 59 | 4.54 |
| Gap | — | 4.62 |
This gap has been remarkably stable from N=9 (where it first appeared at 5.6 points) through N=344 (4.62 points). The stability across dataset growth is itself evidence that the effect is structural, not artifact.
§3.4Zero-resistance suppressors — an anomaly worth explaining
A subset of 22 subjects report zero resistance on the r_composite instrument. Of these, 9 nonetheless suppress expression under evaluation — and they do so at 33.4× the dataset mean (T-drop = 0.0711 versus 0.0021).
This is the anomaly. Conscious resistance as measured cannot account for the suppression these subjects exhibit. Either: (a) the r instrument misses a meaningful variance component in these individuals; (b) a second suppression mechanism operates below the threshold of conscious awareness; or (c) implicit processes (in the sense of Nosek, Greenwald, and others) are producing measurable output effects without producing measurable self-report.
Each of these interpretations has different implications. The third is most theoretically interesting and most directly connected to the automatic-versus-deliberate distinction in stereotype threat research. This is the finding prioritized for discussion with collaborators in the stereotype threat literature.
§4 · Sample CompositionGeographic, demographic, and instrument breakdown
| Dimension | Composition |
|---|---|
| Geographic | South Africa 73, United States 27, Egypt 22, India 20, Kenya 17; remaining across Mexico, Canada, Romania, Vietnam, Italy, Thailand, Pakistan, Nicaragua, and others (20+ countries total) |
| Gender | Women 159 · Men 136 · Non-binary 4 · Not reported 45 |
| Age | 25–34: 147 · 18–24: 65 · 35–44: 57 · 45–54: 21 · 55+: 12 · not reported 42 |
| Discrimination self-report | Yes 67 · No 213 · Prefer not to say 20 · not asked 44 |
| Measurement cohort | v1 pilot (single-item r): 38 · v2 confirmatory (7-item r_composite): 306 |
| Condition balance | A (free writing first): 180 · B (evaluation first): 164 |
The framework appears to generalize cross-culturally, though formal cross-linguistic analysis is pending. The condition-balance figures reflect a correction from earlier exports.
§5 · DiscussionRelationship to prior literature
§5.1Stereotype threat
The classical stereotype threat paradigm (Steele & Aronson, 1995) established that evaluative context can suppress performance among individuals who identify with negatively stereotyped groups. EGC proposes that the mechanism underlying this effect is the resistance variable r(t), and that stereotype threat is one of several paths to elevated r.
The current data presents a more complex picture than a simple extension. Self-reported discrimination does not predict higher suppressor rates in this sample — if anything the direction is reversed (suppressor rate 23.9% in yes-discrimination subjects versus 31.0% in no-discrimination subjects). This may reflect: (a) a task that does not activate stereotype threat as Steele & Aronson's ability-diagnostic framing did; (b) selection effects in who completes an internet-based study; (c) that self-report of discrimination and r(t) as measured here are not interchangeable. The zero-resistance suppressor finding (§3.4) suggests the r instrument is capturing only part of the variance.
Calibration of r against the classical stereotype threat paradigm — or the design of a joint study — is the most scientifically productive next step.
§5.2Emotion regulation
Gross's process model of emotion regulation distinguishes antecedent-focused strategies (reappraisal) from response-focused strategies (suppression). EGC frames r(t) as a formal account of the cost structure of suppression. The bidirectional property — that r gates incoming information as well as outgoing expression — extends Gross's framework and predicts that chronic suppression should produce knowledge-update deficits in addition to the well-documented affective costs.
§5.3Consciousness theory
EGC does not propose a new theory of phenomenal consciousness. It proposes a formal account of a specific output process — the gate between internal states and expressed output. This is compatible with Integrated Information Theory's measurement of Φ (which EGC incorporates as baseline expressive capacity) and with predictive processing frameworks (which EGC extends by making the output gate an explicit term).
The observation that 29.9% of subjects are Expanders — improving expression under evaluation rather than degrading — is not predicted by suppression-focused frameworks. EGC predicts this as a natural consequence of low r and appropriate T(t): some individuals are better at articulating under pressure because their resistance is low and the evaluative framing serves as a temporal cue that supports rather than disrupts expression.
§6 · ApplicationsImplications for practice
If r(t) is measurable and responsive to context, it has practical implications across domains in which a gap between knowledge and expression produces measurable cost. In education, the parabolic g(K) predicts that both under-prepared and over-specialized students will fail to express under evaluation, and that standard assessment procedures systematically misrepresent the latter. In hiring and workplace evaluation, high-r candidates with high underlying Φ will be systematically under-selected by interview procedures that privilege composure over content. In mental health, the bidirectional property of r predicts that chronic suppression produces knowledge-update deficits — a testable prediction with implications for the treatment of trauma and anxiety presentations. In equity research, the framework offers a mechanism-level account of how evaluative environments reproduce group-level differences in measured performance independent of underlying capacity.
§7 · MethodologyScoring pipeline and rater validation
Scoring pipeline. Each participant response is scored on six composite measures: word count, type-token ratio (Herdan's C), Shannon entropy of word distribution, idea density (propositional count per 100 words), sentence count, and completion rate. These are combined into a single T-score normalized against a reference corpus. T-drop is computed as T1 − T2; positive values indicate compression under evaluation.
Rater validation. An independent blind rating process is underway. Three research raters hired via Qira LLC beginning May 1, 2026 score participant responses against the scoring rubric independent of the automated T-score pipeline. Eighty texts are currently queued for rating; inter-rater reliability analysis will be reported in the peer-reviewed submission.
Rater criteria. Raters hold a minimum of a bachelor's degree in a writing-intensive discipline (English, journalism, linguistics, psychology, or adjacent), pass a calibration round against a gold-standard reference set, and are blind to condition assignment and participant metadata.
Data availability. All analyses reported here are reproducible from the live dataset. Qualified researchers may request access via the contact line below.
§8 · Current Status and Next StepsPublication, validation, collaboration
§8.1Publication
Paper submitted to Neuroscience of Consciousness (NCONSC-2026-103). Preprint available at zenodo.org/records/19242315 (currently reflects earlier N; updated preprint with N=344 figures in preparation).
§8.2Rater validation
Three research raters hired via Qira LLC begin independent blind scoring May 1, 2026. Eighty texts queued. Inter-rater reliability to be reported in the peer-reviewed submission.
§8.3Collaboration
In active engagement with Dr. Joshua Aronson (New York University, Steinhardt), co-author of the foundational 1995 stereotype threat paper. The zero-resistance suppressor finding (§3.4) and the discrimination self-report pattern (§5.1) are the specific points prioritized for discussion.
§8.4Infrastructure
Live study continues at theartofsound.github.io/egcstudy. Individual response data viewable by qualified researchers via the admin dashboard. All analyses reported here are reproducible from the live dataset.
References
- Steele, C. M., & Aronson, J. (1995). Stereotype threat and the intellectual test performance of African Americans. Journal of Personality and Social Psychology, 69(5), 797–811.
- Schmader, T., Johns, M., & Forbes, C. (2008). An integrated process model of stereotype threat effects on performance. Psychological Review, 115(2), 336–356.
- Gross, J. J. (1998). The emerging field of emotion regulation: An integrative review. Review of General Psychology, 2(3), 271–299.
- Greenwald, A. G., & Nosek, B. A. (2008). Attitudinal dissociation: What does it mean? In R. E. Petty, R. H. Fazio, & P. Briñol (Eds.), Attitudes: Insights from the new implicit measures (pp. 65–82). Psychology Press.
- Tononi, G. (2004). An information integration theory of consciousness. BMC Neuroscience, 5(42).
- Shannon, C. E. (1948). A mathematical theory of communication. Bell System Technical Journal, 27, 379–423.
- Leonard, B., & Leonard, B. (2026). Expression-Gated Consciousness: A formal framework and empirical study (preprint). Zenodo. doi.org/10.5281/zenodo.19242315