Glossary
The vocabulary of governed autonomy.
These terms were coined in the Adaptive Query portfolio. Each definition tracks the language of the filed specification that discloses it, names that filing, and links to the inventive step it belongs to.
These entries are informal, plain-language summaries for readers. They are not lexicography and do not define, limit, or disclaim any claim term. Where an entry differs from a filed specification, the specification and claims control.
abstention outcome
an outcome of a recorded class emitted by a determination stage where an input required by that stage is unavailable, incomplete, or unresolvable, and adverse to no party.
acknowledgment artifact
an artifact issued by a semantic agent referencing accepted determinations and enumerating action classes, issuance of which decrements the issuing agent's authorization budget merit-independently.
adaptive index
A decentralized index whose anchor maps and quorum thresholds evolve autonomously in response to policy-monitored conditions such as mutation throughput, latency, and load, providing structural continuity and dynamic scalability without external orchestration or global registries. In the embodiment claimed in Application 19/326,036, index entries are organized in a parent-child hierarchy, each corresponding to a unique semantic scope and governed by one or more anchors.
agent-resident execution substrate
A personal computing architecture in which a semantic agent operates as the persistent runtime authority of a computing device, with inference endpoints, knowledge ingestion modules, and related computational assets maintained as governed managed components subordinate to the agent. The agent's identity, cognitive state, and lineage are preserved continuously across the lifetime of the device and across lifecycle operations applied to subordinate components, including installation, retraining, replacement, archival, and removal of inference endpoints.
aggregate retention
a quantity declared in the signed policy object bounding accumulated undischarged harm held in a retention register.
anchor map
The policy-authorized membership and configuration state of an anchor group at a given time, including the active anchors and the quorum parameters required for resolution and mutation ratification. Anchor maps are updated as anchor groups expand or contract under pre-registered policies in response to policy-monitored metrics, and each approved mutation records the previous anchor map in cryptographically committed lineage records, enabling deterministic resolution across structural transitions.
append-only lineage field
a structure carried by the semantic agent to which entries are appended and from which entries are neither removed nor modified.
asserting party
the party from which a conduct evaluation artifact originates, being a party other than the principal.
assertion-cost counter
a structure incremented upon issuance of a conduct evaluation artifact by the semantic agent, per origin-equivalence class of receiving parties, and attested within each issued artifact.
asynchronous scope-based consensus
A mutation validation process confined to the anchors that govern the affected index scope, which achieves policy-defined quorum without invoking global consensus or unrelated anchor groups. Anchors may accept proposals asynchronously from distributed or intermittently connected nodes; these proposals are cached and later validated upon quorum availability, allowing eventual consistency and delayed reconciliation without interrupting the resolution path.
authorization budget
a quantity gating dispatch of actions by the semantic agent within the action space, in units identical to those gating dispatch.
authorization gate
a persistent state element of the memory field governing whether the semantic agent is authorized to execute actions of an action class, and occupying at least a granting state and a withheld state.
biological hash
A non-invertible, domain-scoped, temporally bound cryptographic representation of an individual's biological signal state at the time of a capture event. It serves as the atomic unit of the biological trust-slope, a temporally ordered chain of biological hashes evaluated for continuity, in which identity resides in the coherent, policy-verifiable continuity of the chain itself rather than in any stored template or profile.
bulk-equipotential charge retention
Charge retention in an electrochemical cell achieved by saturating a continuous electronically and ionically conductive medium to a uniform electrochemical potential, so that no internal driving force exists for current redistribution in the absence of an external circuit. The cell retains stored charge without any internal separator, membrane, or electronic-insulating barrier; retention operates by saturation rather than by insulation.
canonical semantic field
One of six schema-defined structural components embedded within a semantic agent object: the intent field, context block, memory field, policy reference field, mutation descriptor field, and lineage field. Each field is individually addressable, machine-readable, and subject to structural validation, and together the fields encode an agent's semantic identity, operational constraints, and evolutionary traceability. A full semantic agent object comprises all six fields, while partial agents containing subsets remain structurally valid.
capability-native computation
A computational paradigm in which determining whether execution can occur is itself a first-class computation producing structured, auditable, and governable results, evaluated prior to execution plan construction. Capability is computed as a determinate outcome (structurally possible, structurally impossible, structurally deferred, or rerouted) rather than a metric, score, or probability, and non-synthesis is treated as a valid computational result rather than an error condition.
carbon air phase
A low-density quasi-three-dimensional carbon morphology that forms from precursor graphenic carbon during the conditioning sequence of a newly manufactured cell, through a one-time irreversible phase transition catalyzed by the cell's multifunctional additive. The phase comprises fullerene-class closed-cage allotropes, fulleroid networks, carbon nanofoam structures, or analogous morphologies, with mass density of approximately 0.2 to 0.8 grams per cubic centimeter versus approximately 1.5 to 2.3 for precursor planar graphenic carbon.
cognition-compatible agent schema
A canonical structural model that represents semantic agency as a first-class data object rather than a transient runtime process. The schema specifies six canonical semantic fields, covering intent, context, memory, policy, mutation, and lineage, that collectively encode an agent's semantic identity, operational constraints, and evolutionary traceability, enabling structurally self-describing agents to be validated, interpreted, and governed across decentralized, stateless, or heterogeneous environments without reliance on external orchestration logic or shared session state.
cognition-compatible payload
A semantic payload, such as a semantic agent, that encodes cognitive execution plans, inference graphs, or dynamic behavioral models capable of operating over the memory-native substrate without requiring substrate-level awareness of cognitive semantics. The substrate stack routes, mutates, and resolves agents carrying cognition-compatible payloads while acting only on memory-derived trust signals, transport metadata, and policy references, without applying or simulating cognitive logic.
cognition-native semantic agent
An embodiment of an agent object configured as a structured semantic representation comprising a plurality of independently addressable semantic fields that represent action proposals and governance-relevant state. In exemplary embodiments the semantic fields include at least an intent field, a memory field, a policy field, and a lineage field, and can further include a context field and a mutation descriptor field, supporting governance evaluation, continuity validation, and eligibility determination across heterogeneous execution substrates.
cognition-native semantic execution platform
A distributed computing platform in which persistent, traceable, and policy-governed reasoning is structurally embedded in the execution substrate rather than treated as external to it. The platform executes memory-bearing semantic agent objects across a memory-native substrate and scoped trust zones, governed by cryptographically signed policy objects, a distributed indexing layer of adaptive indexes, and an entropy-resolved identity layer that authenticates agent lineage without persistent static credentials.
cognitive domain field
An independently tracked persistent field encoding one dimension of a semantic agent's behavioral disposition, normative alignment, or execution readiness, maintained with a current value and a trajectory over time. Fields such as affective state, integrity, personality, confidence, and capability extend the foundational schema, and each proposed mutation receives a composite admissibility determination integrating signals from coupled fields to determine whether the mutation is permitted, gated, or suspended.
coherence trifecta
A unified three-phase control loop that maintains an agent's behavioral consistency after deviation, comprising empathy registration, integrity recording, and self-esteem-driven coherence restoration. The empathy engine computes the projected or actual harm distribution across affected entities and domains, generating deviation pressure. The integrity engine records the deviation in the integrity field and lineage with full provenance. Coherence pressure then activates the redemption engine, which generates candidate restorative mutations.
composite admissibility evaluator
The governance mechanism that evaluates a proposed governed mutation, including observations, forecasts, actuations, and coordination directives, against multiple cognitive domain fields such as dispositional, integrity, confidence, and capability fields, and produces an admission, deferral, rejection, or graduated-mode decision. Under the five-property governance chain, every physical actuation requires composite admissibility approval, making the evaluator the gate through which governed actuation passes.
conduct evaluation artifact
a signed artifact received at runtime from a party other than the principal, asserting an evaluation of conduct performed by the semantic agent itself, and comprising at least a recorded assertion time and a conduct descriptor.
confidence governor
A structural subsystem of an agent's cognitive architecture that continuously evaluates whether the conditions for execution remain satisfied and withdraws execution authorization when they are no longer met. It functions as a hard gate rather than an advisory module, comparing the confidence value and its rate of change against policy-defined authorization and suspension thresholds, and gates execution in one of three states: authorized, suspended, or locked.
constellation signature
A geometric fingerprint derived from the relative positions and angular relationships of high-salience spatial anchor points detected within a content artifact, computed in a manner invariant to translation, rotation, and uniform scale change. It encodes spatial relational geometry rather than distributional variance, is distinct from and supplementary to the variance-vector-based unique identifier, and supports matching of artifacts that share distinctive spatial compositions across cropping, scale change, and partial occlusion.
containment collapse
The failure of the containment layer that separates an agent's speculative planning graph domain from its verified execution memory domain, constituting the architectural analog of delusion. When containment collapses, the agent's cognitive system can no longer distinguish speculative projection from verified reality. A containment restoration protocol responds through execution suspension, planning graph quarantine, lineage forensic analysis, verified state reconstruction, and containment layer re-initialization.
continuity attestation
A cryptographic chaining of a semantic agent's persistent identity over the sequence of substrate-runtime versions, permitting any party verifying the agent's continuity to confirm that the identity is unchanged across substrate-runtime update operations. A continuity attestation also links an agent snapshot artifact to the agent's identity continuity proof, so the same agent remains verifiable across runtime updates and across snapshot, restore, and migration operations.
continuity hash
A cryptographic hash value maintained in a semantic agent's identity field and chained incrementally over the sequence of records appended to the lineage field. The hash is verifiable by reproducing the hash chain from any prior reference value, and a continuity break is detectable as a discrepancy between a computed hash and a stored prior reference value, triggering escalation under the governance policy field.
conversion bar
the property by which a consuming determination is foreclosed from converting an abstention outcome into a scalar value, a default value, an operand of a threshold comparison, or a consequence adverse to any party.
counterparty identity record
a substrate-resident record of an encountered party, comprising at least one identity primitive.
credentialed admissibility profile
An admissibility profile bound to the identity of a material, element, or assembly through cryptographic signature by one or more credentialing authorities. The profile declares one or more property surfaces, each with property-specific parameters and admission conditions, and travels with the material, element, or assembly through manufacturing, installation, operation, and end-of-life processing, remaining evaluable by building-code authorities and building energy management systems.
cross-mesh reconciliation
A mechanism enabling multiple independent governed-mesh deployments to interoperate, reconcile divergent observation histories upon interconnection, and preserve lineage continuity across mesh boundaries without requiring a prior shared authority or consensus protocol. A taxonomy translator produces equivalence attestations between the meshes' authority taxonomies, a lineage-preserving import mechanism preserves original mesh identity in imported observations, and intentionally disconnected meshes are supported through selective observation admission via authorized gateway channels.
deference record
an entry appended by a deferring semantic agent binding at least the identity primitive of the followed agent and the followed determination relied upon, and transmitted to that agent for counter-signature.
depth-selective aggregation
The mechanism that applies training depth profiles to gradient signals during the backward pass. Each training depth profile is a per-layer or per-block contribution weight vector governing gradient magnitude at each depth, so governance policy determines how deeply each training example integrates into model parameters, ranging from full-depth profiles through suppressed depth profiles to a zero-weight depth profile that prevents the example from influencing any parameters.
deviation deductible
a quantity declared in the signed policy object against which the entropy-weighted harm coefficient of each permitted deviation is drawn first.
deviation likelihood
a quotient whose numerator is the difference between the need quantity and the dynamic ethical threshold and whose denominator is the product of the empathy weighting and the self-esteem aggregate.
dynamic agent hash
A deterministic identity fingerprint generated from a semantic agent's internal memory, mutation state, and host entropy. The dynamic agent hash, abbreviated DAH, evolves across execution cycles rather than remaining static, and is used to enforce trust slope continuity and pseudonymous propagation, supporting validation of agent authenticity and semantic lineage without reliance on persistent cryptographic keys.
dynamic agent hash chain
a sequence in which each successor epoch is generated from a prior epoch under an update rule incorporating an unpredictability contribution and a volatile salt, advancing a metering interval independently of any clock.
dynamic device hash
A per-message identity field computed by a transmitting device from inputs including device-specific entropy sources, sensor readings, configuration state, clock state, and content of prior transmissions, having a property of gradual evolution across successive transmissions reflective of the device's operational state. The hash is attached to each governed mesh message for identity-continuity validation, and receiving devices store the received sequence over a policy-defined window for continuity-based device identity.
dynamic flake expansion
A reversible morphological transformation of metal nanoflakes between a contracted resting state and an expanded charged state. The transformation is driven by applied bias, mediated by carbon intercalation from the gel acting as a wedging species, locked by bonded hydrogen at the newly exposed surfaces, and reversed by discharge. Expansion exposes previously buried flake surface area to the gel, admitting additional hydrogen chemisorption beyond the contracted-state surface area.
dynamic indexing protocol
An optional, pluggable indexing layer within the memory-native protocol stack that dynamically restructures semantic flows based on entropy thresholds, semantic density, and lineage volatility. The dynamic indexing protocol, or DIP, organizes agents by entropy, semantic class, and lineage density, enabling adaptive reclassification and semantic partitioning of data flows without preconfigured hierarchical containers or globally imposed structural hierarchies.
dynamic signature mesh
The dynamic signature mesh (DSM) is the memory-native authentication framework in which identity is validated as progression along trust slopes using local unpredictability, bounded proofs, and anchors, without persistent keys or centralized registrars. Within the mesh, agents and devices maintain evolving hash lineages, and a verifier accepts a presented identity only when it is a valid descendant of a previously trusted state under policy-bounded checks.
edge
a persisted dyadic relation between the semantic agent and one counterparty, constituted by that counterparty's counterparty identity record together with each lineage entry recording an interaction with that counterparty.
entropy-weighted harm coefficient
the product of a policy-declared scope coefficient of the integrity scope of a dissonant event and the dissonance weight of that event, scaled by unity plus a weighted sum over the recorded impacts of that event.
execution evaluation cycle
The procedure an execution node performs upon receiving a persistent executable object. The node parses the intent field to identify the execution operation, evaluates the context block against locally applicable execution policy without centralized coordination, reads prior execution records from the memory field, selects an execution action from execution, mutation, delegation, dormancy, reentry, and termination based solely on those inputs, executes the selected action, and appends the outcome to the memory field as a new execution record.
federated semantic zones
Collections of independently operated nodes or domains that coordinate routing, mutation, and indexing behavior across trust-divergent boundaries using shared memory-native substrate logic, without requiring centralized governance or global consensus. Within federated semantic zones, nodes with heterogeneous stack capabilities participate in a shared trust graph and coordinate policy-scoped mutation validation, congestion response, and health-triggered indexing within and across zones.
five-axis disruption diagnostic framework
A multi-axis diagnostic framework that characterizes an agent's cognitive state as a position across five axes: containment integrity, promotion calibration, coherence restoration capacity, empathic load tolerance, and integrity accountability. A weighted combination of the five axis values yields a cognitive coherence index used as an input to the confidence governor, and a phase-shift early warning system evaluates axis trajectories to trigger preventive interventions before a phase-shift boundary is crossed.
fork adjudication
A governance procedure invoked when a resolution query reveals conflicting alias claims across slope-divergent unique identifiers. An anchor quorum reviews registration timestamps, variance proximity to the canonical identifier, policy lineage, and trust-zone authority, and may resolve the conflict in favor of one registration, freeze both registrations pending escalation, or quarantine the alias. All adjudication outcomes are recorded and accessible to querying parties as provenance metadata.
governance chain
The five-property discipline imposed on every governed mutation in a governed spatial mesh, comprising authority-credentialed observation, evidential weighting in a shared governed observation store, composite admissibility evaluation across cognitive domain fields, governed actuator execution in which every physical actuation requires composite admissibility approval, and lineage-recorded provenance linking every observation, evaluation, and action. Operations that maintain the chain end to end are described throughout the specification as governance-chain-preserving.
governance gate
A logical enforcement checkpoint that determines, before a proposed agent action can proceed, including before an execution context or capability context is instantiated, activated, or admitted for use, whether applicable policy authority has been resolved, filtered for freshness, verified, and found to authorize the action under scope, validity, freshness, and continuity constraints. The gate deterministically permits or denies the action, and denial produces non-execution as a valid, auditable outcome rather than an error.
governed corpus
A slope-band-indexed reference corpus of registered content unique identifiers governed by one or more signed corpus policy objects, against which candidate artifacts are evaluated for admissibility, similarity, exclusion, or memorization proximity. Corpus entries are registered under policy objects specifying admissibility scope, exclusion classes, and similarity tolerance thresholds, and similarity evaluation compares the variance vector of a candidate artifact with variance vectors of reference artifacts in the corpus.
governed marketplace
An architectural primitive for governance-chain-preserving exchange of governance-credentialed commodities between governance-credentialed contributors and consumers, without requiring a third-party platform operator to mediate trust, identity, settlement, dispute resolution, or reputation. It operates across commodity classes including skill-adapter, observation, capacity, service, attestation, energy, spectrum, and compute exchange, with per-deployment parameterization specifying commodity schema, pricing mechanism, licensing framework, and applicable authority taxonomy without architectural modification.
governed observation
a signed structure comprising at least an authority credential field and a payload field, and emitted complete upon emission. In accordance with an embodiment, a governed observation is emitted without acknowledgment, handshake, delivery confirmation, or registration with a central authority.
governed spatial mesh
A distributed architecture of governance-credentialed devices and agents that cooperatively perceive, coordinate, and act on a physical environment. Every governed mutation, including observations, forecasts, actuations, and coordination directives, passes through a common governance chain of authority-credentialed observation, evidential weighting, composite admissibility evaluation, governed actuator execution, and lineage-recorded provenance, and capabilities of the system are disclosed as first-class architectural primitives of the mesh.
lineage-derived training corpus
A training corpus assembled by selecting artifacts identified in a semantic agent's append-only lineage field, filtering the selected artifacts against the applicable corpus policy object, and applying any declared redaction or normalization rules. The assembled corpus supplies a supervised training signal for policy-governed retraining of subordinate inference tools, including incremental fine-tuning of a personal corpus model from artifacts authored or designated since the most recent retraining event.
marker-track transport
An architectural primitive producing governance-chain-preserving operation of autonomous and assisted transport units along sequences of governance-credentialed markers. The marker sequences produce rail-analogous guidance properties, including per-unit track following, junction-aware routing, platoon coordination, and a governed safety envelope, on existing public infrastructure without requiring dedicated rail construction, right-of-way acquisition, or displacement of existing road users.
matched pair
a first governed observation representing an offer, a tender, a claim, a demand, or a commitment, and a second governed observation responsive thereto, recognized under a declared pairing rule within a spatial proximity window and a temporal proximity window.
matched-pair settlement
An architectural primitive for governance-chain-preserving bilateral settlement of a physical-world exchange through paired governed observations from two authority-credentialed parties within a governance-policy-defined spatial-temporal proximity window. The primitive produces a persistent settlement record admissible by downstream consumers without third-party intermediary, without centralized consensus, and without pre-negotiated session state, and operates across tolling, commerce, chain-of-custody, capacity reservation, energy transfer, and other bilateral exchanges.
memorization proximity score
A score quantifying structural proximity between a generative model output and its most similar registered training artifact. It is computed as a function of cosine similarity between the output variance vector and the most similar training artifact variance vector within the governed corpus, weighted by the variance band of the output and calibrated against a policy-declared memorization threshold for the relevant content category.
memory-bearing agent
A cryptographically signed data object that serves as the fundamental unit of transmission and execution within the memory-native substrate. Each memory-bearing agent carries a unique identifier, a payload, a transport header, and a memory field recording mutation lineage, access logs, trust evaluations, and policy references, allowing the agent to govern its own routing, mutation eligibility, and consensus behavior without reliance on external session state.
memory-bearing semantic agent object
A software object that structurally encodes canonical fields for intent, context, memory, policy reference, mutation descriptor, and lineage. These six fields govern the agent's execution, mutation, delegation, and semantic traceability across trust zones and substrate nodes, allowing the agent to persist semantic identity, execute governed mutations, and maintain traceable behavior without reliance on centralized orchestration or external session state.
memory-native protocol stack
A modular protocol stack executed at each node of a distributed network in which the primary unit of protocol execution is a memory-bearing agent. The stack includes a routing layer, an indexing layer, and a consensus layer whose behavior is determined by metadata embedded within each received agent, including the verifiable lineage, access logs, and policy references that govern routing, mutation, and consensus behavior.
memory-native substrate
An execution fabric composed of semantic nests and scoped trust zones that routes and stores memory-bearing semantic agent objects across distributed trust zones based on semantic context, mutation eligibility, and trust-scoped propagation policies. Nests provide memory anchoring, fallback scaffolding, and entropy continuity for agents in localized substrates, while substrate components retain execution history, resolve fallback agents, support scaffolding operations, and validate mutation events based on internal field coherence.
memory-resident execution
An execution model in which execution state, history, and decision context reside within a persistent executable object itself rather than in external runtimes, schedulers, or orchestration layers. Each execution node that receives the object locally evaluates its embedded intent, context, and memory, allowing execution to persist, pause, resume, and propagate across distributed nodes while long-horizon context and execution continuity are maintained by the object's own memory field.
memory-resolved identity
Memory-resolved identity is an identity model in which authentication depends exclusively on locally retained information, including unpredictability sources, lineage evidence, and policy-scoped traces, rather than on third-party attestations or long-lived secrets. Verifiers establish identity by recomputing and validating successor hashes from remembered state, without reliance on centralized authorities, long-lived keypairs, or synchronized registries.
merit-independent metering
incrementing of a counter responsive to a determination class or to an issuance, without any determination being performed of whether the refusal, contraction, or issuance was well founded.
mesh-derived coordinate
An architectural primitive producing a self-referencing coordinate system through cooperative inter-agent ranging, without dependence on satellite-based navigation or any specific external positioning infrastructure. The primitive admits at least eighteen ranging modalities and fourteen coordinate frame types, with governance-credentialed frame definition and anchor observation admission producing a shared spatial reference frame accessible to any governance-credentialed consumer.
mesh-derived time
An architectural primitive producing a cooperatively estimated, governance-credentialed temporal reference frame without a master clock, satellite time source, or centralized time authority. The primitive admits thirteen clock technologies and twelve synchronization modalities and supports timestamp attestation, in which agents emit timestamp observations carrying an authority credential, mesh-derived time value, estimated time uncertainty, and cryptographic signature.
multi-function credentialed substrate
A synthesis primitive in which a credentialed structural-storage substrate operates simultaneously as a perception substrate, an identity-attesting boundary, an event-announcement surface, a storage body, an electrical distribution surface, a data network substrate, a credentialed-event commerce surface, a duration-attested grid-commitment surface, a lineage-recorded memory, and a thermal-coupling interface under one architectural primitive. The substrate selectively admits energy, signals, and devices at its boundary through composite admissibility evaluation.
multi-property credentialed material
A structural building material whose credentialed admissibility profile comprises at least two property surfaces selected from the structural, thermal, energy storage, fire-performance, sound-transmission, vapor-permeability, environmental, distribution, network, water-coupled, thermal-coupling, and carbon-sequestration admissibility surfaces. Each surface is independently credentialed by an authority with declared scope and composes with the other surfaces through declared composition rules.
mutation router
An anchor-scoped, policy-aware subsystem that selects propagation paths for proposed structural mutations. It considers contextual signals such as semantic proximity of target containers, trust entropy between involved nodes, and live telemetry indicators, and its adaptive paths evolve in response to observed context and anchor-local constraints. Routing may prioritize nodes with high semantic proximity to the mutation target, high availability, and low recent mutation rejection or conflict rates.
need quantity
a quantified semantic urgency carried in the memory field, comprising at least a scalar magnitude and a categorical type, and computed from state internal to the semantic agent.
non-executing cognitive mode
a state in which the semantic agent conducts speculative evaluation without committing state changes to the memory field.
non-execution attestation
a signed governed observation emitted responsive to writing an authorization gate to a withheld state, whose payload comprises at least an enumeration of the withheld action classes and a typed marker of the recorded abstention class.
object-resident execution state
Execution state carried within a persistent executable object itself, including appended execution records, lineage references, reentry conditions, and policy references, rather than execution state held by external runtimes, schedulers, orchestration layers, or session-bound control mechanisms. Each execution node reads and updates this embedded state during evaluation, preserving execution continuity, auditability, and execution eligibility across multiple execution lifecycles and heterogeneous execution nodes without centralized coordination.
operator intent primitive
An architectural primitive for governance-chain-preserving publication, admission, fusion, and consumption of operator-intent observations. Intent enters the mesh across a plurality of fidelity tiers spanning fully shared cognitive state, structured partial-fidelity integration, and behavior-inferred intent produced by mesh observation of legacy participants, so autonomous, semi-autonomous, and manual operating units sharing the same physical space yield mesh-consumable intent observations regardless of any unit's integration level.
origin-equivalence class
a set of asserting parties between which the semantic agent's own records evidence a derivable relation of a relation type declared in the signed policy object.
other-directed reparation arc
a reparation arc carrying an arc designation field of other-directed value and a counterparty binding field carrying the resolving counterparty's identity primitive, each written at creation and thereafter unmodified.
partial semantic agent object
A semantic agent object that contains fewer than all canonical semantic fields yet remains structurally valid under the cognition-compatible schema. A partial semantic agent object may participate in validation, delegation, or interoperability through fallback inference, structural scaffolding, or lineage-based resolution, allowing even minimally instantiated agent objects to participate in semantic networks and lineage graphs without requiring centralized coordination or external state reconstruction.
permitted deviation
an admission, responsive to the deviation likelihood exceeding unity, of a mutation that the signed policy object in force otherwise forbids, recorded as permitted rather than as fault.
persistent executable object
A structured, memory-bearing semantic object subject to an execution lifecycle in which one or more execution nodes evaluate, act on, mutate, delegate, place into dormancy, reenter, or terminate it, such that execution continuity is governed by object-resident state without centralized coordination. In the embodiment claimed in Application 19/538,221, the object carries an intent field encoding a machine-readable execution descriptor, a context block encoding execution-relevant metadata, and a memory field encoding prior execution state.
personal corpus model
A managed inference endpoint whose model artifact parameters are trained, fine-tuned, or adapter-modified against a training corpus derived from artifacts authored, curated, or designated by the user. Its inference behavior reflects the user's accumulated body of work as encoded in the parameter values, so the user's domain knowledge, terminology, and structural conventions are internalized in the weights rather than consulted as external context through retrieval at inference time.
positive abstention
a recorded, governed withholding of an action, the semantic agent appending a determination naming the withheld action and the ground relied upon, and withholding dispatch of the action.
principal
the party to which a semantic agent is bound and to which an escalation record is emitted.
reason-type
a classification of an edge as obligation-sustained, payment-sustained, want-sustained, or not-typeable, resolved by a severance-survival test over recorded severance events.
refusal counter
a structure resident in the memory field accumulating increments recorded against the semantic agent.
reparation arc
a forward corrective structure created upon a recorded violation and contributing unrectified dissonance while undischarged, reversing nothing and discharged only by a restorative act.
restorative mutation
a scoped mutation appended to the append-only lineage field, comprising at least an anchored reference to the entry recording the deviation addressed and a corrective act, that entry being neither removed nor modified.
scope partition
a named division of the operation of the semantic agent having a per-partition lineage record and a per-partition cognitive state.
self-esteem aggregate
a running aggregate carried in the memory field of coherence between intents declared by the semantic agent and actions it executed as recorded in the append-only lineage field.
semantic admissibility gate
The central deterministic governance mechanism of inference-time semantic execution. It evaluates each proposed semantic mutation against the semantic state object and produces one of three outcomes, admit, reject, or decompose, with no probabilistic scoring or confidence-weighted pass-through. Four sequential evaluation stages cover policy constraints, mutation descriptor validity, lineage continuity, and entropy bounds, transforming inference from uncontrolled generation into governed semantic execution.
semantic agent
an identity-bearing computational object carrying at least a memory field, an append-only lineage field, and a policy reference field.
semantic agent object
A memory-bearing data object structured in accordance with the cognition-compatible schema and defined by the presence of one or more canonical semantic fields that collectively encode semantic intent, governance constraints, mutation eligibility, and traceable lineage. It is distinguished from a runtime process, execution thread, or session construct, exists independently of any particular execution environment, and carries sufficient internal information to be validated, interpreted, and governed by receiving nodes based solely on its internal composition.
semantic hysteresis
A property of the affective modulation layer whereby an agent's current affective state depends not only on the current structured observations but also on the trajectory of prior states. It is implemented through asymmetric update rules in which negative valence updates, driven by failure, uncertainty, or threat, apply at a higher rate than positive valence updates, so the agent responds more rapidly to deteriorating conditions than it recovers from them, producing a built-in caution bias.
semantic quarantine
A governance state in which an agent object is structurally prevented from instantiating execution contexts or performing specified governed action classes regardless of substrate availability. The quarantine persists until authorized remediation occurs, a policy-defined quarantine condition expires, or an authorized override is verified and applicable. A fallback enforcement agent can issue a trust degradation signal and initiate a quarantine action upon detecting a violation or invalid state.
semantic swarm
A collection of semantic objects operating concurrently within shared or intersecting execution environments, in which coordinated execution behavior emerges through delegation, mutation, lineage tracking, and memory-resident execution state without centralized control. Each semantic object functions as an independent execution entity governed by its embedded intent, context, memory field, and locally evaluated policy, and distributed planning emerges from accumulated memory and policy-bound evaluation rather than precomputed workflows or consensus mechanisms.
severance event
an entry of the append-only lineage field recording a structural change terminating the capacity of the semantic agent and a counterparty to interact, comprising at least a severance class and a recorded time.
signed policy object
an object resolved by canonical alias, carrying at least a declared value set, and covered by a verifiable signature.
slope band
A bounded contiguous range of variance-vector space within a global slope continuum, used as the basis for assigning anchor node governance responsibility to content unique identifiers whose variance vectors fall within the range. Slope bands are quantized segments of variance space and serve as the routing unit for identifier registration, alias resolution, and cache propagation within the anchor network. Also called a variance band.
slope entanglement
The enforced cryptographic and behavioral binding between an agent's dynamic agent hash and its host substrate's dynamic device hash. Each time an agent mutates, the resulting DAH includes a reference to the host device's DDH state, creating a verifiable trust slope lineage recorded in the agent's memory trace. Deviations in either trajectory, or missing entanglement references, result in quarantine, rollback, or rejection under zone policy.
standing quantity
an ordinal quantity computed from a direction of initiation, a latency to reciprocate, and a regularity of recorded interactions. In accordance with an embodiment, no cardinal magnitude of benefit is computed or stored.
structural isomorphism thesis
The assertion that the platform's computational mechanisms implement the same causal structure that produces analogous human behaviors: deviation under pressure, recovery through restorative action, confidence-mediated execution governance, empathic consequence registration, and dispositional modulation of speculation. The correspondence is architectural, not metaphorical: the structural reasons for each behavioral outcome correspond to the structural reasons that produce the analogous behavioral outcome in human cognition.
structural scaffolding
A deterministic, schema-defined resolution process by which missing or degraded canonical fields in a partial agent are inferred, reconstructed, or defaulted using available context metadata, policies identified by the policy reference field, lineage anchors, or environmental constraints. Structural scaffolding does not introduce semantic authority beyond that implied by existing fields, and all resolutions are recorded as trace outcomes in the agent's memory field, preserving auditability.
structural starvation
An architectural hallucination prevention technique in which a language model is denied access to the informational resources hallucination requires. It is implemented through five complementary constraints: prompt bounding, absence of external memory, forced reliance on verified agent field values, intermediate rejection of failed proposals without exposed rejection rationale, and stateless purging of the model's context after each inference call, consistent with the model's role as a structurally untrusted proposal generator.
substrate-mode storage
Energy storage that exists as a credentialed property of structural building materials rather than as a property of dedicated battery devices. It is structurally distinct from device-mode storage by the architectural inversion that places storage capacity in materials whose primary function is structural, with distributed capacity aggregated by a building's electrical system as a coherent resource through credentialed admissibility evaluation.
three-in-one traversal step
The atomic unit of semantic discovery, comprising three structurally coupled phases performed in sequence at each anchor boundary. A search step produces a candidate transition set from the anchor's published reachable semantic neighborhood, an inference step scores, ranks, or selects among the candidates, and an execution step evaluates the proposed transition for admissibility under governance constraints. No transition is possible without completing all three phases.
trust entropy
A dynamic trust metric derived from operational telemetry, historical quorum outcomes, and interaction stability between nodes, used by anchors as a context signal in mutation routing and context evaluation. Scores may be computed from real-time telemetry anomalies, mutation audit trail outcomes, and historical anchor interaction stability. Trust entropy is distinct from the locally derived entropy resource defined in the specification and does not expose private content.
trust slope
A directional trajectory of evolving identity hashes — the ordered sequence of dynamic agent hash, dynamic device hash, or content anchor hash states over time — evaluated for continuity and mutation legitimacy. Trust slopes represent the entropic and semantic continuity of an entity's identity over time and across substrate transitions, and are evaluated to verify behavioral lineage, confirm mutation legitimacy, and detect divergence or tampering. In embodiments, the directional deltas between successive states are used in that evaluation.
trust-slope validation
A governance evaluation that examines how an agent object's eligibility or enforcement state evolves over time based on recorded governance events. Policy-defined criteria determine whether the trust state is stable, improving, or degrading. A stable or improving trust slope can satisfy authorization conditions for specified action classes, while a degrading trust slope can restrict eligibility, elevate enforcement class, require corroboration, or trigger quarantine evaluation. The evaluation applies deterministic, policy-defined criteria to objectively verifiable records rather than subjective scoring or probabilistic inference.
unaddressed reparation arc
a reparation arc carrying an arc designation field of unaddressed value and no counterparty binding field, accumulated in the retention register at a declared multiple not less than unity.
untested class
as applied to an origin-equivalence class, a class every constituent edge of which resolves not-typeable, and, as applied to an action class, a class for which no corrective encounter with subsequent recorded compliance is recorded.
use-positive aging
An aging behavior in which cycle life is increased rather than decreased by use. The behavior is attributable to mechanochemical self-healing operating during cycling, in which mobile carbon species migrate from the gel to mechanically strained sites on the metal nanoflakes and locally anneal the strain, and to the absence of calendar-aging chemistries operating at rest, so that cycle life increases as a function of cumulative cycling.
