Constitutive Attractor Inference Protocol
Preamble — Where This Document Stands
This document lays out the epistemology governing the relationship between clinical history and constitutive memory, \xi_t, in Syntropia. Its core claim: clinical history is a procedure for inferring the geometry of \Omega_t^{(p)}. It is not a biographical record, and it is not a set of covariates. Constitutive attractors are regions of the individuation field with high probability density of occupation, not discrete types: basins in the continuous individuation field that \Omega_t^{(p)} defines over the space of configurations available to this trajectory.
The document works across three layers at once. The ontological layer establishes what attractors are within Syntropia’s architecture. The mathematical layer specifies how that geometry gets represented, states plainly the tension between the current discrete representation and the object’s continuous nature, and points to the canonical formalization in the Mathematical Core v0.4.12. The operational layer establishes how clinical history generates inference about that geometry, and how that inference builds the Bayesian prior P(\xi_t) that Unveiling requires.
Epistemological commitments governing this protocol (Epistemic Core v0.3.8):
The inference CAIP produces is dynamic and abductive: inference, action, and the object of knowledge all emerge together, in the same act of unveiling (CE2). What clinical history tells us about \xi_t is inherently incomplete, and open to revision after the fact (CE3). How much inference a single encounter can produce is capped by the current \kappa(t) (CE10), and there’s an optimal, non-monotonic window for revelation set by \varepsilon_t^{\text{ef}}, \kappa(t), and S_m (CE12). The protocol is non-ergodic by design: you cannot extrapolate directly from population statistics to this singular trajectory. Pilot parameters are a condition of possibility for calibrating the individual case, not its destination (CE16). The process of constitutive attraction that CAIP infers is continuous; T_2 only captures its trace in discrete observations, not the process itself. That is a constitutive observational limit of any practical process science, not a defect specific to this protocol (CE17). Comparing attractors across different trajectories works at the level of structural parameters (depth, reach, gradient, basin type), not trajectory content. That is what makes cumulative knowledge possible without violating the radical singularity of A10 (CE18). Syntropic configurations are dynamic polymorphs of the trajectory (metastable modes of stabilization, not ontological types), and two trajectories with the same observable configuration can have radically different field geometries underneath: clinical isomers (Epistemic Core §I.bis).
Part I — The Ontological Layer: Attractors as the Geometry of \Omega_t^{(p)}
I.1. The Problem With Discrete Taxonomy
The first version of this document (v0.1.0) built a list of twelve attractor types distributed across the four domains of H_t. That move reproduced the exact error Syntropia diagnoses in categorical systems: it collapsed something continuous into something discrete for operational convenience, then treated the discretization as though it had ontological standing of its own.
The individuation field, \Omega_t^{(p)}, is a stability function defined over the space of configurations available to this trajectory right now (a continuous field with basins, ridges, gradients, and barriers), not a finite set of possible ways the field can be organized. Two trajectories with the same “type” of relational attractor are still two distinct points in that configuration space, because they differ in basin depth, in how they couple to other domains through \Psi_t, and in the history of plastic deformation that built that basin in the first place. A discrete taxonomy of attractors loses exactly the same information a discrete taxonomy of trajectories would.
Here is the correct reformulation: constitutive attractors are properties of \Omega_t^{(p)}’s geometry, not entities drawn from some finite set. CAIP is a protocol for inferring that geometry from whatever historical evidence is available.
I.2. Attractors as Basins in the Field of Individuation
\Omega_t^{(p)} is isomorphic to a Lyapunov function over the configuration space (Mathematical Core §II.5 and §III.7). States where \Omega_t^{(p)} is high are attractors: the trajectory tends to settle into them and resists leaving. States where \Omega_t^{(p)} is low are unstable for this trajectory: the field abandons them quickly under perturbation.
That isomorphism has a direct consequence for what attractors actually are: a constitutive attractor is a region of configuration space where \Omega_t^{(p)} forms a basin, where the gradient points inward and the depth is enough to resist perturbations smaller than \varepsilon_t^{\text{ef}}. The clinically relevant properties of an attractor are properties of that basin:
Depth: the magnitude of \Omega_t^{(p)} inside the basin, relative to its surroundings. A deep attractor needs a perturbation bigger than \varepsilon_t^{\text{ef}} to trigger reorganization; the field resists and returns. Depth is a function of accumulated plastic deformation: D_p(t) writes the geometry of past reorganizations directly into \Omega_t^{(p)}.
Reach: how far the basin extends across configuration space. A locally reaching attractor mainly organizes a single H_t domain. A cross-domain attractor organizes the field through \Psi_t: a perturbation in the primary domain pulls adjacent domains toward the basin too. Reach is a function of the coupling pattern between domains that the trajectory’s relational history has written into \Psi_t.
Access gradient: how steeply the basin slopes from its rim to its center. A steep gradient means fast capture: small perturbations are enough to pull the field in. A shallow gradient means gradual capture. The access gradient tells you how fast the field responds to environmental perturbation, and it is what calibrates how large an intervention needs to be.
Basin coupling: the geometry of the individuation field is a manifold with global topological structure, not a set of isolated basins. Two basins can sit close together, separated by a low-energy barrier with frequent crossings, or sit far apart, with barriers that need large perturbations to cross. Basin coupling is the structure that T_{\text{escala}} captures when it translates from the local organization \{H_t, \varepsilon_t, \Psi_t\} into the global \Omega_t^{(p)}.
I.3. Constitutive Attractors as D_p(t) Written Into the Geometry
The distinction between a constitutive attractor and a situational tendency of the field is a formal one. A constitutive attractor is a basin whose depth exceeds the field’s elastic threshold: accumulated plastic deformation that has changed \Omega_t^{(p)} irreversibly (D10). A situational tendency is an elastic displacement within the field: it shows up under certain environmental conditions, and the field returns to baseline once those conditions lift.
\text{Constitutive attractor} \;\Leftrightarrow\; \Delta\Omega \to D_p(t) \quad (\delta \geq \varepsilon_t^{\text{ef}})
\text{Situational tendency} \;\Leftrightarrow\; \Delta\Omega = 0 \quad (\delta < \varepsilon_t^{\text{ef}}, \text{ elastic return})
The second equation is technically precise about \Delta\Omega: situational tendencies do not change \Omega_t^{(p)}’s geometry. But sustained situational tendencies do have a cumulative effect on \kappa(t), what we call field fatigue (D11): repeated sub-threshold perturbations wear down receptive capacity even when they never produce plastic deformation. A field with a long history of frequent situational tendencies can have an \Omega_t^{(p)} free of major plastic deformation and still have significantly reduced \kappa(t). The constitutive-attractor-versus-situational-tendency distinction operates on D_p(t). Its cumulative effect on \kappa(t) is independent, and it gets inferred from the history of accommodations.
Clinical history lets you infer which basins of \Omega_t^{(p)} have constitutive status (written into D_p(t)) and which are situational patterns with no associated plastic deformation. That distinction determines the scale of intervention needed: constitutive attractors call for work on D_p(t); situational tendencies respond to changes in the environment or to restoring \kappa(t).
I.4. Where the Field of Individuation Comes From, and Why Attractors Are Not All Alike
\Omega_t^{(p)} = F(\sigma_t, D_p(t), \kappa(t)): the individuation field, right now, is a function of the initial condition or baseline geometry (\sigma_t) and the accumulated history of plastic deformation and receptive capacity (Ontological Audit O19/O20). Basins of \Omega_t^{(p)} come from two ontologically distinct sources:
Basins of biographical constitutive origin: produced by plastic deformation accumulated over the trajectory’s life history, under spatium conditions. These are this protocol’s primary object: the type of attractor clinical history lets you infer directly.
Basins of biological-genomic constitutive origin: written into \sigma_t (the field’s initial condition, prior to any biographical plastic deformation, O19/O20) via the T_{\text{ontológico}} operator, from differentiation[t]. The genomic profile modulates spatium parameters (\varepsilon_t, \Psi_t, D_p(t)) and sets the individuation field’s initial geometry. These basins have no narrative history that explains them, because they predate any history the system has. You can infer them from clinical history (the absence of an explanatory history in the presence of the pattern is itself evidence for this origin), but that inference needs to be supplemented with the genomic-functional profile (see G9, Ontological Core v2.3.9).
This difference in origin does not change the inference procedure in Parts III and IV: the same four questions apply to both. What it changes is where treatment is aimed. Basins of biographical origin can be worked on through D_p(t). Basins of biological-genomic origin need to be read on their own functional terms and worked with at the level of the field’s geometry, not treated as though there were a history to modify, because there is not one.
Part II — The Mathematical Layer: Representation and Update Requirements
II.1. The Tension Between Continuous Ontology and Discrete Representation
The individuation field is ontologically continuous: becoming produces stabilizations on a manifold with no discrete boundaries between types. The current mathematical representation of \Omega_t^{(p)} is discrete, though:
\Omega_t^{(p)}: \{\text{C-A},\ldots,\text{C-M}\} \to [0,1]
The 14 configurations of canonical system v1.2 (C-A through C-M, with C-C split into C-C1 and C-C2) are operational discretizations of the space of stabilizations. They are clinical communication tools that let a clinician name a reading of the field and share it with the person in front of them, not natural kinds (see Configuraciones_Canonicas_v1_3_9.qmd). The \Omega_t^{(p)} function over those 14 points is a projection of the continuous individuation field onto a finite set of clinical reference points that can be demonstrated without heavy structural machinery. But basin geometry (depth, reach, access gradient, topological coupling) cannot be represented with any precision across just 14 discrete points. Representing the individuation field with those properties intact requires representing \Omega_t^{(p)} as a function over a continuous space. That representation is the object of G2/G10 (Ontological Core v2.3.9), and it is a requirement for the pilot, not something to defer to a later phase.
II.2. Canonical Formalization in the Mathematical Core
The continuous extension of \Omega_t^{(p)} that G-CAIP requires takes the form of a density function over a differentiable manifold \mathcal{W}:
\Omega_t^{(p)}: \mathcal{W} \to \mathbb{R}_{\geq 0}
where \mathcal{W} parametrizes the space of possible organizations of the field, with a local metric structure that reflects functional similarity between nearby organizations. The 14 discrete configurations are high-density points in \mathcal{W}: clinical reference points on a manifold that contains them as special cases, not as the boundaries of the space.
That representation lets us formalize the four geometric properties from §I.2 precisely: depth as the difference in \Omega_t^{(p)} between the basin’s interior and its rim; reach as the breadth of the field of possibilities \mathcal{F}_t^{(p)} around the basin; access gradient as the norm of the gradient at the rim; topological coupling as the structure of barriers between basins that persistent homology captures.
Requirement met: Mathematical Core v0.4.13 incorporates the continuous extension of \Omega_t^{(p)} over \mathcal{W} in §II.5bis, and the field of possibilities \mathcal{F}_t^{(p)} in §II.5ter. The canonical formalization lives in those documents.
II.3. G2/G10 in the Context of CAIP: The Central Articulation Gap
G2/G10 (Ontological Core v2.3.9). Continuous representation of \Omega_t^{(p)}’s geometry and inference of constitutive attractors.
The gap labeled G-CAIP in earlier versions of this document maps canonically onto G2 (calibrating the metric tensor g_t over \mathcal{W}) and G10 (stabilization conditions: which spatium parameter values produce each canonical configuration) in the Ontological Core’s research agenda, v2.3.9. What follows develops G2/G10 specifically for the historical-inference protocol.
Generating ontological commitment: constitutive attractors are basins of the continuous individuation field \Omega_t^{(p)}. The discrete representation across 14 clinical reference points captures which basin is most stable right now and supports precise clinical communication. It does not capture basin geometry (depth, reach, gradient, topological coupling), and that geometry is what determines intervention scale and transition prediction.
Extending the inference target (session 2026-06-04): G-CAIP is not just about inferring \Omega_t^{(p)}’s geometry over \mathcal{W}. Its full target is inferring the field of possibilities, \mathcal{F}_t^{(p)} := \{w \in \mathcal{W} \mid \Omega_t^{(p)}(w) \geq \theta(\xi_t)\}, which requires estimating two things at once: (1) the density distribution \Omega_t^{(p)} over \mathcal{W}, and (2) the threshold \theta(\xi_t) as a function of constitutive memory. \mathcal{F}_t^{(p)} can contract through two independent mechanisms, and the pilot needs to characterize them separately: density redistribution driven by D_p(t), and threshold elevation driven by reduced \kappa(t). The functional form \theta(\xi_t) = \theta_0 \cdot g(D_p(t), \kappa(t)) is a sub-problem of G2 (metric tensor calibration) that the pilot needs to start closing.
Research question: can \mathcal{F}_t^{(p)} (the field’s accessible region, not just the full \Omega_t^{(p)} distribution) be inferred precisely enough from the pilot’s longitudinal data, and does that inference converge with the topological representation the observational apparatus produces from sensor signal?
Representation hypothesis: persistent homology over the individuation field \Omega_t^{(p)} (a component of Chapter 9’s observational apparatus) produces a topological representation of basins from dynamic signal. Inference from clinical history and topological representation from the sensor are two independent routes to the same latent structure. Their convergence is G-CAIP’s central testable prediction, and it is a result the pilot has to produce, not a question to defer past it.
Gaps with direct implications for CAIP: G17 (the clinical interstitial field, D15-int in the Ontological Core v2.3.0) affects the operational layer: the properties of \Omega_t^{(\text{int})} determine which aspects of \xi_t^{(p)} are even accessible in this encounter, regardless of the field’s individual parameters. Ontological Core v2.3.0 formalizes three regimes of \Omega_t^{(\text{int})}, each with direct implications for what CAIP can produce in a given encounter:
- Productive regime, high \kappa^{(\text{int})}(t), moderate D_p^{(\text{int})}(t): CAIP can produce high-resolution inference about \xi_t^{(p)}, including deeper, more constitutive basins. The four questions in Part III operate at full strength.
- Restricted regime, \kappa^{(\text{int})}(t) reduced by accumulated D_p^{(\text{int})}(t): CAIP produces lower-resolution inference. The deepest basins may be inaccessible in this encounter even when the person’s individual parameters would allow it. The protocol needs to work with that limit in view, not force unveiling of layers the interstitial field is not making available right now.
- Rupture regime, \kappa^{(\text{int})}(t) < \kappa^{(\text{int})}_{\text{umbral}}: intervening on the person’s D_p(t) in this state produces harm analogous to what P12 describes. Priority: restore \kappa^{(\text{int})}(t) before any work on constitutive memory. In this regime, the priority intervention is actively repairing the interstitial field, not unveiling.
A7 reformulated (Ontological Core v2.3.0): the clinician’s field is part of the protocol. A7 v2.3.0 states that the encounter produces \Omega_t^{(\text{int})} as an emergent organization from the coupling between the two fields, and that the clinician’s field reorganizes in the encounter too. The consequence for CAIP: the clinician’s field state at the time of the encounter is a condition that shapes what the protocol can produce, not just a source of noise. A clinician with reduced \kappa(t) in that moment produces an interstitial field with lower receptive capacity, regardless of the person’s own individual parameters. The Clinical Core v0.3.2 protocol (reading the clinician’s field post-encounter) is the operational counterpart to this claim.
G18 (\rho(t), the resignification operator) affects how the prior gets built: if \rho(t) turns out to be distinguishable from \kappa(t), building P(D_p(t)) will need to separate the magnitude of D_p(t) from what function D_p(t) serves in the field. Both gaps remain open, pending pilot data.
Additional gaps with implications for CAIP (v2.3.0):
G19 (the transitional regime, Ontological Core v2.3.0) affects CAIP’s ability to detect the silent period: during Phase 1 of §T1 (\chi_t high, \dot{H}_t \approx 0), the protocol can give the mistaken impression that nothing is underway. Reading posterior variance as a signal of \chi_t, not just as uncertainty, is the correction CAIP needs to make during that period.
G20 (the metric of processual similarity, urgent) means that comparisons between attractors across different trajectories, the kind CAIP performs implicitly whenever a clinician generalizes from past cases, do not yet have a formal foundation. CE18 and H-COMP lay out the principle and the working hypothesis; CAIP operates on that principle without a calibrated metric behind it yet.
G21 (the nature of the M ↔︎ \Omega_t^{(p)} relationship) affects basin inference when M is low: if M acts as an access threshold rather than a modifier of the landscape itself, a low M changes operational access to that geometry, not the field’s geometry. Part III should be read with that distinction in mind: the inferred basin depth is the field’s real depth. What a low M affects is the field’s ability to move within that geometry, not the geometry itself.
§T1: CAIP across each phase of the transformation process. Transformation theory (§T1, Ontological Core v2.3.0) has direct implications for how CAIP’s prior should be read. In Phase 1 (silent accumulation): the prior needs to reflect that the field is under generative tension even while H_t looks stable. Posterior variance over \xi_t is the signal to watch. In Phase 2 (threshold and reorganization): the prior right before reorganization is the least reliable one. The field is at its point of maximum change, and historical inference may not yet capture the newly emerging basin. In Phase 3 (transitional regime): CAIP produces maximum-uncertainty inference. An \Omega_t^{(p)} distribution spread across multiple configurations with no clear mode is the correct signal here, not a protocol failure. In Phase 4 (installation): the prior converges on the new basin, and posterior variance drops.
Status: an active gap (G2/G10 on the Ontological Core’s research agenda, v2.3.9) that the pilot needs to start closing. The 14 configurations of canonical system v1.2 are the language for communicating the process; the continuous geometry of \mathcal{F}_t^{(p)} is the model’s actual inference target.
Part III — The Operational Layer: Inference From Clinical History
III.1. Clinical History as a Sampling Procedure Over \Omega_t^{(p)}
Here is the central epistemological reframing: clinical history is a sampling procedure over \Omega_t^{(p)}’s geometry, not an archive of events. Every element of the intake is an observation that updates the probability distribution over which regions of \mathcal{W} form active basins for this trajectory right now.
\text{Clinical history} \;\rightarrow\; \text{inference of active basins in } \Omega_t^{(p)} \;\rightarrow\; P(\xi_t)
The clinical question organizing the history is “what structure did this leave in the present field’s geometry?”, not “what happened?” That reorientation produces a radically different relationship to biographical material: events matter to the extent that they let you infer whether they built active basins in \Omega_t^{(p)}, how deep those basins are estimated to be, and how far they reach in the current field. They do not matter in themselves.
This is where NC-10’s consequence becomes technically precise: the field organizes what it presents in narrative according to its own attractors. What is absent from the narrative is information of equal or greater relevance, because the narrative is a selective projection of the individuation field \Omega_t^{(p)}, and the deepest basins can be exactly the ones the self-image never shows.
§I.bis of the Epistemic Core is just as precise on this point: two trajectories with the same observable configuration, the same point in the discrete 14-point space, can have radically different field geometries underneath. They are clinical isomers. The sampling procedure over \Omega_t^{(p)} that CAIP sets up infers this trajectory’s constitutive structure, not its membership in a category. That is an ontological distinction, not a methodological one: the inference works on the field’s singular geometry, not on its distance from some type.
III.2. The Four Inference Questions
Inferring \Omega_t^{(p)}’s geometry from clinical history is organized around four questions. None of them classify the attractor into a discrete category. All four estimate basin properties, the same properties Part II’s mathematical layer formalizes.
Question 1: Which region of configuration space does the basin operate in?
The basin has a location in \mathcal{W}, approximated from the H_t domains that clinical history shows as most consistently compromised. The five domains (V_t, R_t, P_t, A_t, B_t) correspond to five marginal projections of \mathcal{W}: \mathcal{W}_V, \mathcal{W}_R, \mathcal{W}_P, \mathcal{W}_E, \mathcal{W}_B. These are marginal projections, not a partition. \mathcal{W} does not factor into five independent subspaces, precisely because \Psi_t exists as domain coupling only because the manifold has joint structure. A basin with high density in \mathcal{W}_R can have simultaneous density in \mathcal{W}_E through the \psi_{ER} coupling; that co-location is a basin with cross-domain geometry in \mathcal{W}, not the sum of two local basins. The projection onto \mathcal{W}_B matters especially for basins of biological-genomic constitutive origin (where the somatic domain can be the basin’s primary axis) and for configurations where B_t \to V_t or B_t \to R_t propagation is the dominant clinical mechanism.
The answer to this question is an estimate of the basin’s location in \mathcal{W} and its cross-domain reach, not a type of attractor.
Question 2: What is the basin’s estimated depth?
Depth gets inferred from three indicators in the clinical history:
How old the pattern is: if the pattern predates declarative memory (with no narrative history to explain it), the basin originates in \sigma_t (the field’s initial condition, O19/O20), and its depth cannot be inferred from biographical events. If narrative history is available, the age of the formative period tells you something about depth: basins built during periods of reduced \varepsilon_t^{\text{ef}} (early childhood, periods of high allostatic load) run deeper than basins built during periods of greater elasticity.
Resistance to prior perturbation: if the pattern survived contexts that should have changed it (prior interventions, major life changes, relationships pulling the other way), the basin is deeper than it looks on the surface. A history of failed attempts to change it is direct evidence of depth.
Estimated \varepsilon_t^{\text{ef}} during the formative period: the same event produces deeper plastic deformation when the field’s \varepsilon_t^{\text{ef}} was already reduced. The history of spatium conditions during the formative period (high allostatic load, sustained instability, high-demand contexts with limited resources) tells you what \varepsilon_t^{\text{ef}} the field had when the basin was built.
Question 3: What is the basin’s reach: local or cross-domain?
Reach gets inferred from how consistently the pattern shows up across domains in the history: if it only appears in contexts tied to the primary domain, the basin is local. If it shows up, with variations, across multiple domains, the basin is cross-domain, and its geometry extends through \Psi_t.
\Psi_t’s asymmetry leaves a signature in clinical history: if the relational pattern consistently activates the existential domain but not the volitional one, \psi_{ER} is larger than \psi_{VR} for this trajectory. The history of how events in one domain produced effects in others is direct evidence about \Psi_t’s structure and about the basin’s cross-domain reach.
Question 4: What is the access gradient: how big a perturbation activates the basin?
The access gradient gets inferred from the history of the pattern’s activation thresholds: does the field fall into the basin at small perturbations, or does it take something bigger? A field that drops into the relational basin at the smallest relational cue has a steep gradient: fast capture. A field that only falls in under major relational perturbation has a shallow one.
The access gradient calibrates intervention size: an intervention that clears the gradient activates the basin. If work on D_p(t) is not indicated at that point, that activation produces plasticity that narrows \mathcal{F}_t^{(p)}. The history of contexts that activated the pattern (their magnitude, frequency, specificity) is the primary evidence for estimating the gradient.
III.3. Evidence Rules
The four questions run on observational evidence, and different kinds of evidence carry different update weights. Distinguishing evidence types matters operationally for building the prior.
Direct activation evidence: events or contexts in the history that activated the pattern observably and consistently. This produces the largest update to both basin probability and estimated depth. It is the most specific evidence, and also the most vulnerable to narrative bias: the field chooses what activation it presents.
Resistance-to-change evidence: a history of failed attempts to modify the pattern (prior interventions, relationships pulling the other way, high adaptive-demand contexts). This directly updates estimated basin depth, regardless of the narrative about where it came from.
Propagation evidence: the history of how events in one domain produced effects in others. This updates cross-domain reach and \Psi_t’s structure.
Threshold evidence: the history of how large the contexts were that activated the pattern. This updates the access gradient.
Absent expected evidence: if clinical history contains no evidence consistent with a basin the current presentation suggests, that absence updates toward one of two alternatives: biological-genomic constitutive origin predating declarative memory (a basin written into \sigma_t, the initial condition, before any biographical plastic deformation, O19/O20 in the Ontological Core), or a residual attractor with no current activity. Telling the two apart takes a reading of the preview and clinical resonance, not narrative history alone.
III.4. Active Attractor Versus Residual Attractor
The active-versus-residual attractor distinction is not settled just by whether the pattern shows up in the preview. There is an added complication from the generative potential functional \Phi_t^{(p)} (a working structural hypothesis, §II.5sexies in the Mathematical Core v0.4.12): a field can show \nabla H_t \approx 0, what looks like stable residency, through two clinically distinct mechanisms. In a field stable through equilibrium (\Phi_t^{(p)} \approx 0), the basin is deep and there is no generative tension: the attractor is active, and the field is not pushing toward any reorganization. In a field stable through exhaustion (\Phi_t^{(p)} \gg 0 with \kappa(t) < \kappa_{\text{umbral}}), there is high generative tension with no capacity to act on it: the attractor may be residual, the field would move if it could, but it cannot. These two situations call for radically different interventions (log rule 110). A history of failed change attempts against an apparently stable pattern is the primary evidence for telling the two apart.
Not every basin with a biographical history is active in the present. An attractor is active if it is still organizing the current field (showing up in the preview, in forms, in grammatical tense, in behavioral patterns observable in the encounter), regardless of whether the person reports it as a problem. A residual attractor has history but is not organizing the present field: the basin exists, but the field does not inhabit it often enough to constitute active D_p(t).
This distinction is clinically fundamental, because it determines what feeds into the prior P(\xi_t): only active attractors contribute as current D_p(t). Residual attractors have documentary value but are not variables in the active clinical inference. The most sensitive indicator here is the fit between historical narrative and the preview in the clinical encounter: a pattern that shows up in the history but has no expression in the current preview is very likely a residual attractor.
Part IV — Building the Prior P(\xi_t)
IV.1. Prior Structure From the Inferred Geometry
The Bayesian prior P(\xi_t) that Unveiling requires (P6, D13) gets built from Part III’s inference results. Its structure:
P(\xi_t) = P(D_p(t)) \cdot P(\kappa(t) \mid D_p(t))
Active-basin inference feeds into P(D_p(t)): every identified basin, with its depth, reach, and gradient estimates, contributes to the distribution over D_p. Deeper, wider-reaching basins carry more weight. The history of sub-threshold accommodation (repeated perturbations that never produced plastic deformation but wore down \kappa(t) cumulatively) feeds into P(\kappa(t) \mid D_p(t)).
The prior at the first encounter has maximum variance across all of \xi_t: it is the point of greatest uncertainty in the whole process. That is the right starting condition, not a limitation. First-encounter clinical history produces two kinds of information at once: an estimate of position in S_m (how much \kappa(t) is available and what scale of D_p(t) is likely, lower-variance and more immediately actionable), and a first inference about active-basin geometry (location in \mathcal{W}, estimated depth and reach, approximate gradient). That second kind of information runs high-variance at the first encounter, but it is not empty. It sets up the geometric hypotheses that later evaluations will confirm, rule out, or refine.
IV.2. Evidence Convergence and Cumulative Construction
Building the prior is a cumulative process: each evaluation’s posterior becomes the next one’s prior. First-encounter clinical history produces the initial prior; later evaluations (sensor data, self-report, new narrative material) refine the estimate of \Omega_t^{(p)}’s geometry.
Two principles govern this cumulative construction:
Convergence: when multiple independent evidence sources all point to the same basin (historical narrative, encounter preview, sensor pattern, grammatical tense, behavioral forms), that basin’s probability in the prior climbs with each added source. Convergence across different source types is more informative than depth within a single one.
Divergence as information: when evidence from different sources points in opposite directions (narrative shows no basin where the preview shows activation, or vice versa), the divergence itself carries information about the field’s structure. It can indicate masking, experiential dissociation, or a biological constitutive attractor with no narrative history behind it. Divergence updates the prior in ways convergence never does.
IV.3. Relationship to the Observational Apparatus
Clinical history builds the prior. The sensor and self-report contribute to the likelihood. The two have complementary properties, and inference needs both at once:
Clinical history captures the accumulated diachronic dimension (the geometry the field built up over time), with narrative’s selective bias built in. The sensor captures the current dynamic dimension with no narrative bias, but no access to constitutive history either. Active-basin inference brings the two together: the basin inferred from history predicts the pattern the sensor should show if that basin is active. Whether the prediction matches the observed pattern, or diverges from it, updates basin probability and refines the \xi_t estimate.
Persistent homology in the observational apparatus (the component closest to \Omega_t^{(p)}) produces a topological representation of basins straight from dynamic signal. CAIP’s historical inference and the sensor’s topological representation are two independent routes to the same latent structure. Their convergence is G2/G10’s central testable prediction, and it is a result the pilot needs to produce.
Version 0.4.16 — July 13, 2026. Single-source canonical document. Any modification requires a new version and explicit approval. Cross-check against: Ontological Core v2.3.16, Mathematical Core v0.4.27, Clinical Core v0.3.12, Epistemic Core v0.3.19.