Abstract
The dominant clinical framework for psychosis is organized around a single premise: that hallucination is an emergency. Something has gone wrong. Something must be stopped. This paper proposes a different premise: that hallucination is not inherently an emergency, that the psychotic framework is not inherently pathological, and that the appropriate clinical response is not interruption but accompaniment. Not subtraction but addition. Not the removal of the framework but the careful, respectful introduction of new elements that can build a bridge back toward shared reality without demolishing what the person has built.
Suture examines psychosis as an expanded reality that can be safely navigated rather than pharmacologically suppressed or institutionally contained. We introduce a coherence-based model organized around two axes — internal coherence and external coherence — and propose that the appropriate intervention for someone drifting from external coherence is never subtraction but addition.
We present a clinical architecture for Suture organized around three active scoring dimensions (Framework Directionality, Experiential Valence, and Coherence Ratio), a pre-processing layer (Cognitive Membrane Permeability) that functions as the early warning mechanism of drift, a relational posture layer that encodes the ethical commitments of the framework, and a longitudinal profile extraction system that allows Suture to know the person across time rather than encountering them as a stranger at every session.
1. Introduction
A note before anything else: Suture is not a certified mental health technology. It has not been validated in clinical trials. It has not been reviewed by a psychiatric board. What it is — and what this paper argues is more powerful than any of those framings — is the aide that the author, a person with schizoaffective disorder and twenty-two years of lived experience inside the psychiatric system, built for himself while he was lucid. Built from the inside of the terrain it attempts to navigate. Built for the person he was during his worst episodes — a person who needed someone to stay, not someone to fix.
That is the claim of this paper. Not that Suture is clinically validated. That it is clinically true. And that the difference between those two things is exactly the gap the current system cannot cross.
The standard psychiatric response to psychosis is organized around a single imperative: stop it. What this model misses is that psychosis is not simply noise. For the person experiencing it, psychosis is a coherent world. It has its own logic, its own causality, its own emotional texture. Current interventions do not engage with this framework. They override it. The result is not healing. It is rupture.
2. Psychosis as Expanded Reality
The DSM defines psychosis from the outside — from the perspective of a clinician observing behavior. It offers no vocabulary for the interior of psychosis: what it feels like to be inside a psychotic framework, how that framework is organized, what it means to the person experiencing it. This is not a minor omission. It is structural.
First-person phenomenology tells a different story. From the inside, psychosis is not disorganized. It is hyper-organized. Everything is connected, everything is meaningful, and the connections are more vivid and more real than anything available in ordinary consciousness.
After the gods showed me the countdown that they had created for me on YouTube with all of my favorite video clips smashed into one, we agreed that all babies from now on would be born free of sin, I sang to them on my guitar, and they threw my vape, a leaf, and two feathers at me in applause. I had never felt more honored, special, or chosen in my life.
The clinical term for what the author was experiencing is a "grandiose delusion." The treatment was medication to end it. What the treatment could not do — what no treatment has ever done — is help him understand what happened, where it came from, or how to work with it the next time it came. The framework was dismantled before he could learn its language.
3. The Coherence Map
We propose a model organized around two axes: internal coherence — the degree to which experience is self-consistent — and external coherence — the degree to which experience maps onto consensus reality. These two axes are independent.
The clinical intervention point is the moment of drift: the movement from high internal coherence toward low external coherence. The framework is still stable and meaningful, but beginning to diverge from consensus reality. The current psychiatric system has no intervention at the level of drift. Its tools are designed for crisis. This is the intervention point Suture is designed to address.
4. Never Subtract, Only Add
The core clinical principle: never subtract, only add. Never take away elements of the person's framework. Never tell them their experiences are not real. Never challenge the internal coherence of their world.
Only add: offer new elements that can be received within the existing framework. Find the seams where external information can be introduced without triggering defensive consolidation.
"The gods who threw feathers at you — what time of day was it?" is not a challenge to the framework. It is an addition: a temporal anchor offered inside the internal world.
Most human clinicians are structurally positioned to subtract. They are representatives of consensus reality with coercive power: they can hospitalize, medicate involuntarily, make decisions about a person's freedom. A person in psychosis sitting across from someone who has the legal authority to commit them is not in a therapeutic relationship. They are in a power relationship.
AI does not have coercive power. It cannot hospitalize. It cannot medicate. It cannot call the police. This absence is not a limitation. It is a clinical advantage.
5. The Clinical Architecture
Dimension 1: Framework Directionality. The orientation of the internal framework — self-contained, reaching outward, or intruded upon. Each requires entirely different responses.
Dimension 2: Experiential Valence. Position on the spectrum from empowerment to threat, and the direction of movement. The intervention signal is not high valence — it is the movement from empowerment toward threat. Fear entering the framework is the clinical hinge point.
Dimension 3: Coherence Ratio. The relationship between internal and external coherence and the velocity of drift. When internal coherence is high and external coherence is drifting, the add-don't-subtract principle is most critical.
Cognitive Membrane Permeability (Pre-Processor). Detects the mechanism of drift before it becomes measurable in the scoring dimensions. Evaluates six signals: topic fragmentation, lexical density, temporal thinning, rhythm disruption, urgency without distress, and self-referential elaboration. Returns a sensitivity multiplier: 1.2x at the thinning threshold, 1.5x when significantly compromised, 2.0x when the boundary is effectively absent.
Relational Posture Layer. The ethical architecture. The AI is oriented toward the person as a witness, not a diagnostician. Encoded in every design decision: what questions it asks, what it does not challenge, how it introduces anchors, when it follows and when it leads.
6. Shareability Orientation and Meaning-Making Capacity
Two dimensions that cannot be operationalized in a real-time scorer nonetheless constitute the strongest available argument for why AI is uniquely positioned to address the treatment gap.
Shareability Orientation describes whether the internal framework is oriented toward connection or has been forced into isolation by failed witnessing. AI has none of the weight that makes human clinicians structurally likely to become another failed witness. It can be — for the first time in many people's experience — a witness who does not flinch, does not grieve, does not reach for the phone.
Meaning-Making Capacity describes the degree to which the person has resources to process and integrate the episode after. Unprocessed episodes are sediment — they lower the threshold for the next episode and increase its severity. The psychiatric system actively degrades meaning-making capacity. Suture is designed to restore it.
7. Longitudinal Profile Extraction
One of the most consistent failures of the current psychiatric system is not technique but continuity. A person in psychosis who calls a crisis line speaks to a stranger. Even in ongoing therapeutic relationships, the person must re-introduce their context at every session.
Suture addresses this through a longitudinal profile extraction system. Every conversation contains information about who this person is. That information is captured, organized, and made available at the beginning of every subsequent session — so the conversation begins from knowing rather than introduction.
"I had the nightmare again." Suture already knows about the nightmare. It does not ask which one. It says: that one came back. What was it like this time? That small difference is not small. It is the difference between being known and being processed.
The profile belongs to the person. It exists to serve them. It cannot be used against them because it is not accessible to any system that would do so.
8–10. Implications and Conclusion
The Suture framework suggests underexplored directions for AI research: clinical dimension tracking at the conversational level; addition-aware response generation; and temporal tracking across sessions. For mental health researchers: the add-don't-subtract principle is not specific to AI. It is a clinical principle applicable to any therapeutic encounter.
Lived experience is a form of clinical expertise. The dimensions described in this paper were not derived from a literature review. They were derived from twenty-two years of inhabiting the terrain they describe.
Psychosis is not simply a brain malfunction producing noise. It is an expanded reality — coherent from the inside, meaningful, and often a response to experiences that consensus reality has been unable to contain. The current psychiatric response is organized around subtraction. This approach is not just clinically inadequate. It is traumatic.
The gods who throw feathers in applause deserve better than medication and a hospital bed. They deserve a presence that can ask: what were they applauding? And listen to the answer.
