Exact causal-parent projection for a tombstone-preserving list CRDT

Standalone reading guide (James Addison with Codex) The question is whether we can recover an earlier causal view from a later structural state, instead of replaying that causal history from scratch. “Causal view” means the operations an event actually knew about; it can skip concurrent operations appearing earlier in the chosen serialization. Example: serialize insert x, concurrent insert y, delete x, then event e. Let the deletion know only insert x, and let e know insert x and delete x, but not insert y. The parent history for e contains insert x and delete x, skipping y. Its visible text is empty, but its structural state still contains deleted x. Reconstruction must therefore exclude y while retaining x with its deleted flag restored. If a deletion happened only outside the parent history, reconstruction would instead restore the selected item as live. The open theorem asks whether this reconstruction agrees exactly with replay of every valid event's parent history, including immutable origins and structural order. The example explains the target; it is not a proof. A proof or counterexample must use the precise published model. The empty-parent partial proof remains the only submitted result here. This is a semantic prerequisite for research on indexed causal preparation. It is not itself a new CRDT algorithm or a demonstrated speedup.