Trust Diagnostics

Normative

Expansion · Bond · Calibrating Trust to Behavior

Mechanism

Trust Diagnostics is the practice of asking whether trust is warranted before you increase exposure. It separates trust from the feelings that often travel with it: warmth, familiarity, liking, gratitude, shared identity, admiration, low friction. Those can be evidence, but they are not trust.

Trust is domain-specific vulnerability under risk. You trust a colleague to critique your argument, a surgeon to operate, a friend to keep a confidence, an institution to enforce its rules, an AI system to report uncertainty, a partner to tell you when something is wrong. Each is a different trust object. A person may be trustworthy in one domain and not another. Collapsing those domains is one of the ordinary ways trust goes wrong.

The question is not "Do I trust them?" The question is "For what, at what exposure, on what evidence?"

The tool exists because people overgeneralize trust constantly. Someone is kind, so they are trusted with competence. Someone is competent, so they are trusted with integrity. Someone shares your values, so they are trusted with power. Someone has kept small commitments, so they are trusted with catastrophic downside. The mistake is not that any one signal is irrelevant. The mistake is treating a signal from one dimension as if it authorizes trust across the whole person or institution.

Trust Diagnostics holds against two failures. Control withholds trust until certainty arrives, which means cooperation never scales beyond low-risk exchanges. Decay extends trust because the relationship feels right, because distrust feels unkind, or because the person once earned trust in a different domain. The Range form is trust proportioned to domain, evidence, risk, and consequence.

Practice

The diagnostic question is: "What trust am I extending, and what behavior warrants that level of exposure?"

Use it before giving someone authority, sharing sensitive information, relying on a report, accepting advice, escalating access, or letting a relationship move from low-stakes cooperation into high-stakes dependence.

The scan is not a moral verdict. It is a calibration tool. A person can be low-trust in one domain and high-trust in another. A person can be benevolent but not competent, competent but not honest under pressure, honest but unreliable, reliable in calm settings and unreliable when status is threatened. The point is to get the trust specific enough that behavior can actually discipline it.

In the Wild

A team promotes its most technically capable engineer into a managerial role. The evidence for ability is strong, but the trust object has changed. The question is no longer "Can this person solve hard technical problems?" It is "Can this person exercise authority over people in a way that preserves honest signal, distributes credit fairly, and makes hard tradeoffs without using status defensively?" The old evidence matters, but it does not answer the new trust question.

A friend is warm, loyal, and emotionally present. You trust them with personal distress. That trust may be warranted. But it does not automatically mean you should trust their investment advice, medical interpretation, or strategic reading of a professional conflict. Trust Diagnostics preserves the friendship by refusing to make one kind of trust do every job.

An institution publishes a code of conduct and has handled small conflicts well. That warrants some trust. Then a serious allegation implicates a senior member whose status benefits the institution. The trust object changes under cost. The diagnostic now asks whether the institution's integrity holds when enforcing its rule threatens its own standing. The earlier behavior is evidence, but the higher-stakes domain requires stronger evidence.

Before you say "I trust them," complete the sentence. Trust them with what? Under what pressure? At what downside? On what evidence?

That small act changes the relationship. It protects you from suspicion as default because you can extend trust where evidence warrants it. It protects you from naive trust because you no longer treat warmth, identity, or past competence as a universal license. Trust becomes something you can calibrate instead of something you either feel or refuse.

Lineage

Roger Mayer, James Davis, and David Schoorman's 1995 organizational trust model is the primary lineage for the ability, benevolence, and integrity scan. Their work separated trust from trustworthiness and treated trust as willingness to be vulnerable to another party. The Codex adds two operational emphases: domain specificity and Range calibration. This page also adds observed reliability as a practice-level check, not as a fourth MDS antecedent, because trust calibration needs to know whether conduct repeats under time and pressure.

Niklas Luhmann's work on trust as a way of reducing social complexity belongs in the background. Trust lets people act without knowing everything. That is why it is necessary and dangerous at the same time. You need enough trust to move, but the simplification can hide the very facts that should have changed the trust.

Julian Rotter's work on interpersonal trust and later organizational research on cognition-based and affect-based trust help explain why people trust from different channels. Some trust is built from observed competence and reliability. Some is built from emotional bonds. Both can be legitimate. Both can mislead when transferred into the wrong domain.

Risk and vulnerability are the common thread across the trust literature. If nothing can be lost, the word "trust" is doing little work. Trust becomes meaningful when exposure is present. The exposure question keeps the tool anchored: what can be harmed if the trust is misplaced, and does the evidence warrant that risk?

Cross-references

Within the category. Good Faith as Default governs the opening posture before evidence has accumulated. Trust Diagnostics governs the assessment once exposure is increasing. Graduated Reciprocity is the practice for building evidence through small reciprocal steps. Skin in the Game tests whether consequence alignment supports the trust being requested.

Across the Workshop. Calibrating Confidence to Evidence is the Foundation neighbor. Confidence answers to evidence; trust answers to behavior. Checking Your Map Against Reality is the Knowledge neighbor. A trust diagnostic is only as good as the map of behavior it uses, so report fidelity, signal selection, and name-reality fit all matter.

Limitations. Trust Diagnostics can drift into a control checklist that makes cooperation feel like a compliance audit. Use it when exposure is increasing or behavior is unclear, not as a way to monitor every ordinary interaction. The tool can also create false precision. Ability, benevolence, integrity, and reliability are judgment categories, not measurements with clean units. Treat the scan as a disciplined read, not a score.