Mechanism
A relational trust trajectory extends the biological identity architecture from validating an operator's own signals to modeling the trustworthiness of the other parties an agent interacts with. Where the surrounding sections of the chapter establish a human operator's identity by maintaining trust-slope continuity over that operator's own biological signals, the relational trust trajectory mechanism applies the same continuity-based paradigm outward: it models the consistency, reliability, and behavioral continuity of other parties, including other agents, human collaborators, and external systems, across the agent's interaction history.
The trajectory is maintained per entity. For each entity in the agent's relational graph, the system keeps one relational trust trajectory, so the object is the agent's own record of how a specific counterparty has behaved over time rather than a shared or symmetric structure jointly owned by both parties. This per-entity framing is what makes the trajectory other-referential: it records what the evaluating agent has observed of the entity, not what the entity asserts about itself.
What the Trajectory Comprises
A relational trust trajectory comprises four components. The first is a behavioral consistency score, derived from the entity's observed pattern of commitments honored versus commitments violated across successive interactions. The second is a communication reliability score, derived from the entity's observed pattern of stated intentions versus actual actions, including detection of discrepancies between declared state and observed behavioral or biological signals. The third is an event continuity record, a sequence of interaction events in which each event is evaluated for plausibility as a continuation of the prior interaction pattern, analogous to the trust-slope continuity validation applied to biological identity observations. The fourth is a trajectory direction, indicating whether the entity's relational trust is increasing, stable, or declining over the evaluation window.
These four components are not independent scores bolted together. The event continuity record carries the same continuity logic used elsewhere in the chapter for an operator's own biological hashes: each interaction is judged as a plausible successor to the established pattern, not matched against a fixed reference. The trajectory direction is the summary the rest of the platform consults, distinguishing an entity that is becoming less reliable from one that is holding steady or improving.
Computed from External Behavior Only
The relational trust trajectory is computed without requiring access to the other entity's internal state. The agent observes the entity's externally visible behavior, including actions taken, commitments made, outcomes produced, delegation contracts honored or violated, and communication consistency, and evaluates each observation as a plausible continuation of the prior behavioral trajectory. Nothing in the mechanism depends on inspecting the counterparty's own fields, memory, or declared norms.
Where the agent has access to biological signals from the other entity through the biological signal acquisition modalities described earlier in the chapter, the system additionally evaluates communication-biology discrepancies. These are conditions in which the entity's verbal or textual communication diverges from the entity's biological state indicators, such as elevated stress during assurances of calm, or physiological markers of deception during assertions of truthfulness. The discrepancy feeds the communication reliability score, so that a divergence between what an entity says and what its biology indicates is treated as evidence bearing on trust.
Recording in Lineage
The relational trust trajectory for each external entity is recorded in the agent's lineage as a series of relational trust observations. Each observation comprises the interaction context, the observed behavioral consistency, the trajectory update, and the resulting relational trust score. Because the trajectory is written into lineage rather than held as an opaque running figure, the basis for a given trust assessment is reconstructable: an inquiry into why the agent treated a particular entity with caution can be traced back to the specific observations and trajectory updates that produced the entity's current standing.
Coupling to the Empathy Weighting Engine
The relational trust trajectory feeds directly into the empathy weighting engine described in the integrity chapter. The agent's empathy computation for projected harm to or from an external entity is modulated by the relational trust trajectory for that entity. An entity with a declining trust trajectory, indicating increasing behavioral inconsistency, violated commitments, or detected communication-biology discrepancies, receives amplified empathy weighting in the deviation function, causing the agent to exercise greater caution in evaluating actions that involve that entity. Conversely, an entity with a stable or increasing trust trajectory receives standard or reduced empathy weighting, reflecting the lower relational risk associated with behaviorally consistent partners.
This creates a feedback pathway from biological observation through relational trust modeling into normative evaluation. The agent's sense of impact, the empathy computation that governs how strongly it resists a deviation, is informed by the observed reliability of the entities the action would affect. In human-relatable terms, a person exercises greater care in dealings with someone who has previously broken commitments than with someone who has been consistently reliable. The platform implements the same relational trust modulation structurally.
Coupling to Multi-Agent Trust Weighting
The relational trust trajectory also feeds into the multi-agent trust weighting mechanism described in the integrity chapter. When multiple agents participate in group decisions, delegation chains, or quorum-governed operations, each participant's relational trust trajectory modulates the weight given to that participant's contributions. A participant whose observed behavior has been inconsistent contributes less to a group outcome than one whose trajectory is stable or increasing.
The trajectory is distinguished from the integrity trust score defined elsewhere in the platform. The integrity trust score measures an agent's consistency with its own declared norms and is therefore self-referential. The relational trust trajectory measures an entity's consistency as observed by the evaluating agent from external behavioral evidence and is therefore other-referential. The two answer different questions: whether an agent is faithful to its own stated commitments, versus whether a counterparty has been reliable as seen from the outside.
Trajectory Versus Predictive Social Modeling
The relational trust trajectory is also structurally distinct from the predictive social modeling mechanism disclosed in the integrity chapter. Predictive social modeling projects an entity's current cognitive disposition from recent observable behavior, answering what the entity's present state is likely to be. The relational trust trajectory instead accumulates historical consistency across successive interactions, answering whether the entity has been reliable over time. One is a forward projection of present disposition; the other is an accumulated record of past behavior. They draw on overlapping observations but serve different roles in the agent's reasoning.
Disclosure Scope
The relational trust trajectory mechanism, comprising the per-entity trajectory maintained for each entity in the agent's relational graph, the four components of behavioral consistency score, communication reliability score, event continuity record, and trajectory direction, the computation from externally visible behavior including the evaluation of communication-biology discrepancies, the recording of relational trust observations in the agent's lineage, and the coupling of the trajectory into the empathy weighting engine and the multi-agent trust weighting mechanism, is disclosed in the cognition filing (U.S. Application No. 19/647,395 and its international counterpart). This article describes that disclosed mechanism and uses the spec's own terminology. The scope extends to the embodiments in which biological signals from the other entity are available and to those in which only behavioral evidence is observed, provided the trajectory remains an other-referential, continuity-based record of an external entity's observed reliability.
