| A1 | String = sense | Surface identity decides the cluster. | Different propositions collapse. | Compare argument patterns and entailments. |
| A2 | Cross-source creep | A BPMN use is folded with a PROV use inside F.3. | The interpretation basis changes unnoticed. | Finish each local claim first; test any relation in F.9. |
| A3 | Over-granulation | SLO and its full form become separate claims without a difference in use. | Friction rises without gain. | Consolidate source-blessed aliases. |
| A4 | Under-granulation | Diagram node and real occurrence share one claim. | Later inferences contradict each other. | Split on argument or entailment conflict and add a counterexample. |
| A5 | Imported definition | A dictionary replaces the selected source passages. | The result is no longer source-local. | Ground the claim in the actual passages. |
| A6 | Label drift | Plain adds scope not present in Tech or the claim. | The cold reader learns the wrong concept. | Keep both labels within the same claim. |
| A7 | Substantive leakage | A sense line contains policies, equations, or kind criteria. | The lexical note is asked to establish a substantive fact about another value. | Route the claim to its direct pattern. |
| A8 | Edition blend | Two editions with changed usage share one unqualified claim. | Replay and comparison become unreliable. | State each basis; merge only if evidence supports the same claim. |
| A9 | Cue worship | Similar collocations are treated as proof. | Correlation replaces source meaning. | Use cues to locate passages, then test propositions. |
| A10 | Time blur | A design-time use and a run-time occurrence are clustered together. | MethodDescription and Work collapse. | Split and use F.11, A.3, and A.15 for the substantive distinction. |