Agent Concepts

Concepts

trigger-rule-pair

13 min read

Trigger rule pair

Description

The meta-criterion: a heuristic only earns its keep when it is paired with an identifiable invocation condition. A bare rule (“do X”) is not yet a trigger-rule pair — it’s an aspiration or an instruction that may never be retrieved at the moment it’s relevant. A trigger-rule pair has the form: “When [situation-shape S], apply [rule R].” The trigger is what makes the rule addressable at the right moment; without it, the rule sits inert in memory and fails to execute when the situation calls for it.

This concept is directly architecturally relevant to any retrieval-driven reasoning system: the retrieval layer needs trigger patterns to know when to surface a concept or insight, not just what the insight says. A knowledge-base entry without trigger keywords or a “When it applies” section is an incomplete trigger-rule pair — it has the rule but lacks the invocation condition.

Triggers

User-initiated: User is writing or reviewing instructions, rules, heuristics, principles, or guidelines. Any time a rule is stated without a “when”: that’s the moment to surface the trigger-rule-pair concept and prompt for the invocation condition.

Agent-initiated: Agent scans an insight or rule-like statement and detects a bare principle (rule without trigger). Candidate inference: “this heuristic is missing its invocation condition — what situation-shape should fire it?”

Vocabulary cues: “heuristic,” “rule,” “principle,” “always,” “never,” “our convention is,” “best practice,” “standing instruction” — followed by a content statement but not preceded by a conditional (“when X,” “if X,” “in situations where”).

Situation-shape signals: A document containing many bare rules (style guides, team wikis, coding-standard docs) is a trigger-rule-pair audit target — each rule is a candidate for trigger-elicitation.

Exclusions

  • Truly unconditional rules — some invariants apply everywhere with no trigger needed (e.g., “never use plaintext for credentials”). The trigger is effectively “always” and stating it explicitly adds no information.
  • Exploratory context — in a brainstorm or divergent-thinking mode, pairing every heuristic with a trigger prematurely constrains exploration. The concept earns its keep when converging on operationalizable rules.
  • When the trigger is fully obvious from context — “when reviewing code, look for security vulnerabilities” — the trigger adds little if the document is already scoped to code review.

Structure

The rule half is what gets executed; the surface half is what the situation must present for the rule to activate; cadence constrains when within a cycle the trigger fires (e.g., “at session start,” “when a new dependency is introduced,” “after each deployment”). Cadence is optional — some triggers are event-driven rather than cadence-driven.

Relationships

  • doctrine (creation relationship) — trigger-rule-pair is the structural criterion that turns a doctrine fragment into a complete doctrine. Every complete doctrine is a trigger-rule pair; not every trigger-rule pair is elevated to doctrine status.
  • cadence (composition relationship) — cadence-driven triggers fire at temporal intervals (“at session start”) rather than event-based; the combination is trigger-rule-pair with a cadence-shaped trigger.
  • asymmetric-gate (specialization relationship) — an asymmetric gate is a trigger-rule pair where the rule is “apply the expensive check” and the trigger is “when the cheap gate passes.” Trigger-rule-pair is the containing category.
  • active-gate-vs-passive-audit (composition relationship) — the “gate vs. audit” choice is a downstream decision from the trigger: what does the rule do when triggered? Gate (block) or audit (record)?

Examples

CLAUDE.md instructions

Domain: computer-science

“Read X before doing Y” is a trigger-rule pair; “write good code” is not. The quality of CLAUDE.md files correlates with trigger-rule-pair density.

Cognitive science: production rules in ACT-R have the same structure (condition → action). The trigger half corresponds to the condition side of a production.

Domain: psychology

Cognitive science: production rules in ACT-R have the same structure (condition → action). The trigger half corresponds to the condition side of a production.

Architectural decision records (ADRs)

Domain: computer-science

good ADRs state “when this architectural question arises (trigger), prefer approach X (rule).”

Code review norms

Domain: computer-science

“when you see a function over N lines, flag it” (trigger-rule pair) vs. “keep functions short” (bare rule).

Stiell, I. G., Greenberg, G. H., McKnight, R. D., Nair, R. C., McDowell, I., & Worthington, J. R. (1992). A study to develop clinical decision rules for the use of radiography in acute ankle injuries. Annals of Emergency Medicine, 21(4), 384–390.

Domain: medicine-and-health

The Ottawa Ankle Rules give a clinician an explicit condition under which an ankle X-ray is warranted. An ankle radiographic series is indicated only if there is pain in the malleolar zone and any one of: bone tenderness at the posterior edge or tip of the lateral malleolus, bone tenderness at the posterior edge or tip of the medial malleolus, or an inability to bear weight both immediately after the injury and in the emergency department. The rule’s value is that it converts a vague clinical instinct (“order imaging when a fracture seems plausible”) into a checkable triggering condition, which is what makes it reliably applicable across clinicians and what lets it safely reduce unnecessary radiographs.

This is a trigger-rule pair in clinical medicine. The_trigger is the articulated decision criterion — the specific, recognizable findings (malleolar tenderness at named sites, or failure to bear weight) whose presence is what fires the rule. The_rule is the action the trigger authorizes: obtain the radiographic series. A bare clinical maxim like “image when in doubt” is a rule without an addressable trigger — it never tells the clinician precisely when doubt should convert to action, so it cannot be applied consistently or audited. The Ottawa Rules earn their keep precisely by supplying the missing invocation condition.

Inference: a clinical decision rule is the medical form of the trigger-rule pair, and its diagnostic and economic power comes from the trigger half being operationalized — stated as observable findings rather than as an aspiration. The discipline of replacing “use judgment” with “when these specific signs are present, take this specific action” is what turns expert intuition into a transmissible, evaluable rule.

Pattern-language tradition (Alexander et al.): each pattern has a “forces” section describing the situation that calls for the pattern — equivalent to the trigger.

Domain: architecture-and-design

Pattern-language tradition (Alexander et al.): each pattern has a “forces” section describing the situation that calls for the pattern — equivalent to the trigger.

Domain: military-sciences

Standing rules of engagement condition the use of force on recognized triggering conditions rather than leaving it to unbounded discretion. Under the U.S. Standing Rules of Engagement, the inherent right of self-defense is activated by a “hostile act” — an attack or other use of force against U.S. forces — or by demonstrated “hostile intent,” defined as the threat of imminent use of force. When a commander or service member recognizes hostile intent, the rule that fires is the authorization to use necessary and proportional force in self-defense, even before an actual attack occurs. The doctrine deliberately turns a fraught judgment into a paired structure: a defined condition to be recognized, and a defined response that the recognition licenses.

This is the trigger-rule pair in the use-of-force domain. The_trigger is the recognized situation-shape — “hostile act” or “hostile intent” — whose identification is what fires the rule; the doctrine even specifies how to read imminence (assessed from all facts and circumstances, not requiring an instantaneous attack), because the trigger must be articulable to be applied under pressure. The_rule is the response the trigger authorizes: the exercise of self-defense, bounded by the further requirements of necessity and proportionality. An order to “defend the unit” with no specification of what conditions justify force would be a rule without an invocation condition — dangerously inert at the moment of decision, and impossible to apply consistently across personnel.

Inference: rules of engagement are trigger-rule pairs precisely because the stakes make an articulable trigger non-negotiable — a use-of-force rule that cannot say when it applies is not a usable rule at all. The structural discipline of naming the recognizable condition that licenses an action, and bounding the action once licensed, is what lets a doctrine execute reliably under stress rather than collapsing into ad-hoc judgment.

Domain: law

Levi describes the doctrine of precedent as a three-step process of “reasoning by example”: a court sees that the material facts of the case before it are similar to those of an earlier case, announces the rule of law the earlier case stood for, and applies that rule to the new facts. The crucial first step is a matching of fact-patterns — the judge decides that the present situation falls within the same category as the precedent. If the match holds, the prior ruling governs; if the present case is materially distinguishable, the precedent does not fire and a different rule (or a new one) applies. Levi stresses that the matching is itself an act of judgment — the later court decides which facts of the earlier case were the material ones — so the trigger is not mechanical, but it is unmistakably a trigger: the rule’s application is conditioned on a recognized similarity of situation-shape.

This is the trigger-rule pair operating across cases rather than within a single statute. The_trigger is the recognized correspondence between the new case’s material facts and the precedent’s fact-pattern — the “this situation is one of those” recognition that has to occur before any prior holding becomes relevant. The_rule is the holding the precedent supplies: the legal consequence that the earlier court attached to that fact-pattern, now carried forward. Stare decisis is, in this light, a retrieval mechanism: the body of precedent is a store of rules indexed by the situation-shapes that should surface them, and the lawyer’s craft is largely the craft of arguing that a given trigger does, or does not, match.

Inference: a precedent with no articulable triggering fact-pattern is inert — it can state a rule, but nothing tells a future court when to reach for it. The same structural requirement that makes a heuristic operational makes case law operational: the rule must be paired with a recognizable invocation condition, and “is this case like that one?” is exactly the trigger-matching question that pairing demands.

Domain: law

A statute such as “if a driver’s blood-alcohol concentration is at or above 0.08, that driver is guilty of driving under the influence” has the form of a conditional. Kelsen makes this structure the centerpiece of his theory of law: a legal norm, he argues, is a hypothetical judgment of the form “if A is, then B ought to be” — if the defined condition (the delict) obtains, then the consequence (the sanction) ought to follow. He contrasts this deliberately with a law of nature, which asserts “if A is, then B is”: the legal norm substitutes imputation for causation, an “ought” linking condition to consequence by an act of legislative will rather than by physical necessity. The legal condition and the legal consequence are the two halves Kelsen insists every legal rule joins.

This is the trigger-rule pair stated as jurisprudence. The_trigger is Kelsen’s condition — the recognizable factual situation (the Tatbestand) whose occurrence is what fires the norm: here, the measurable fact “BAC ≥ 0.08.” The_rule is the legal consequence the trigger activates — the imputed status, duty, or sanction (“guilty of DUI,” and the punishment that attaches). A statute that named a consequence with no condition, or stated a prohibition with no specification of when it applies, would be defective precisely the way a bare heuristic is: it would lack the invocation condition that lets it be applied to the right case at the right moment.

Inference: well-drafted law is a corpus of explicitly conditioned rules, which is exactly why statutes are addressable — an official can match a concrete situation against the stated condition and know which consequence to apply. The structural lesson generalizes: a rule becomes operational only once its triggering condition is articulated as a recognizable pattern, and a system of rules without triggers cannot execute, no matter how sound each rule is in isolation.

Wells, P. S., Anderson, D. R., Rodger, M., Forgie, M., Kearon, C., Dreyer, J., Kovacs, G., Mitchell, M., Lewandowski, B., & Kovacs, M. J. (2003). Evaluation of D-dimer in the diagnosis of suspected deep-vein thrombosis. New England Journal of Medicine, 349(13), 1227–1235.

Domain: medicine-and-health

The Wells clinical model for deep-vein thrombosis assigns points for findings such as active cancer, recent immobilization, localized tenderness along the deep venous system, and asymmetric calf swelling, subtracting points when an alternative diagnosis is at least as likely. The cumulative score then triggers a branch in the diagnostic pathway: a score of one or less classifies DVT as “unlikely,” sending the patient to a D-dimer test that, if negative, rules out DVT without imaging; a score of two or more classifies DVT as “likely,” sending the patient directly to compression ultrasonography. The score is not the diagnosis — it is the condition that determines which downstream rule applies.

This is the trigger-rule pair embedded in a diagnostic algorithm. The_trigger is the stratified pretest probability — the score crossing its threshold (”≥ 2 → likely,” ”≤ 1 → unlikely”) is the recognizable condition that fires a particular next action. The_rule is the action that the threshold authorizes: in the “unlikely” branch, “obtain a D-dimer and rule out on a negative result”; in the “likely” branch, “proceed to ultrasonography.” The score’s job is to make the rule addressable — without a quantified trigger, a clinician facing a swollen leg has no principled way to decide which test sequence the patient should enter.

Inference: a scored decision rule is a trigger-rule pair whose trigger is a threshold on an aggregated signal rather than a single observation. The structural point is identical to the simpler clinical rules — the rule executes only when its invocation condition is articulated and met — but the trigger here is graded, which lets a single algorithm route different patients to different rules by where their score lands. Defining the threshold is defining the trigger.

Graph View

Backlinks