C.11:8 - Common Anti-Patterns and How to Avoid Them
One quick usability test helps here: if the closing line does not state one lawful choice result for the working chooser or team, the current result is still unfinished even if the doctrine survey looks polished.
| Anti-pattern | Symptom | Why it fails | How to avoid / repair |
|---|---|---|---|
| Candidate-formation or search takeover | The text starts constructing complete ways or generating options as if that work were already part of decision doctrine. | The choice rule has no stable option set to compare. | State the existing option set. Use C.38 for same-result way formation, C.39 for a missing explanation of how to obtain the result, or C.40 for feasible variation and examination. Add C.18 when a generation account or archive/front claim is needed. |
| Policy collapse | Exploration or exploitation governance over a candidate pool is written as if it were identical with choosing among current options. | Choice doctrine and candidate-pool policy become indistinguishable. | C.19 remains explicit as the neighboring pattern for selection policy and exploration governance. |
| Planning collapse | Sequencing, replanning, and enactment budgeting are written as if they were already part of the choice calculus. | Planning-side question moves out of C.24 by accident. | Execution order and operational budgeting remain in C.24, even when C.11 says more probing is rational. |
| Inventory without decision rule | The current comparison names many objects and schools but never shows how to move from a live option set through one ChoiceRule to one ChoiceResult. | The pattern becomes one cleaned-up survey rather than one decision discipline. | State one explicit decision-record shape: chooser, option set, comparison basis, dependence layer, applicable probe-worthiness test, one explicit doctrine, and one emitted result. |
| Hidden basis shift | Different options are compared under different belief states, outcome models, or dependence layers without one explicit statement that the basis changed. | The comparison only looks precise; in fact the choice rule cannot be audited. | Keep one shared comparison basis until one named probe or model change updates it, and state explicitly when the dependence layer changes. |
| No closure rule | The text sounds careful but never says what makes choose now, reject current set, probe again, or reroute lawful. | The record never closes into one explicit decision result. | State the closure conditions explicitly and show why the current case satisfies exactly one of them. |
| Undefined load-bearing terms | Terms such as PreferenceOrder, BeliefState, or OutcomeModel appear without local operational clarification. | Core comparison objects stay implicit and the decision question depends on outside theory or undocumented assumptions. | Give one local plain gloss or equivalent operational clarification for each load-bearing term used in the pattern text. |
| Bounded-resource bridge loss | ProbeBudget, ValueOfInformation, or ValueOfComputation are mentioned, but the text silently lets C.19 or C.24 own them. | The theory-side doctrine disappears into neighboring policy or planning prose. | Keep those objects theory-side in C.11; let neighboring patterns consume their outputs without minting the concepts. |
| Result-boundary collapse | The text treats declaring selector-facing set content, or later making it available, as if either were identical with deciding. | Choice doctrine silently absorbs the G.5 result question or the publication question. | Keep both outside C.11: use G.5 to declare the selector-facing result, E.17 for its publication face and return to source, and E.24.PUB for the publication occurrence and availability. |
| Agent-default narrowing | Every chooser is described as one Agent even when the subject is really one team, organization, or other collectivity-bearing system. | The governed chooser is narrowed before the doctrine even starts. | DecisionSubject remains the default, and DecisionSubjectGranularity types the chooser-bearing level. |
| Prestige-branch citation | Active inference or quantum-like work is cited only as one fashionable name. | The text sounds current without stating what limitation is being repaired. | The repaired limitation is stated directly: embodied online updating for active inference, and context or order effects for quantum-like lines. |
| Cost-free deliberation | The text speaks as if probing and computation are free. | Bounded-resource doctrine disappears behind one idealized choice moment. | ProbeBudget, CostToProbe, ValueOfInformation, and ValueOfComputation stay visible in the calculus. |