Temporal Logic

Overview

This area tracks the temporal-specification slice of Formal Methods: LTL-style languages, branching-time logics, automata-shaped tasks, monitors, and temporal objectives that constrain learning, verification, and synthesis over time. Broader verification infrastructure belongs in Formal Methods unless the temporal structure is doing the main explanatory work.

Active Questions

  • Which temporal fragments are expressive enough for safety without becoming operationally brittle?
  • Which safety/co-safety fragments are useful because of finite-prefix semantics, and which remain hard because of realizability complexity?
  • When is a problem primarily about temporal specification, and when is it better routed to broader Formal Methods?
  • How should temporal specifications be grounded in observations, abstractions, or learned state representations?
  • What is the practical boundary between monitoring, masking, and offline synthesis?
  • When should temporal synthesis use pure Safety Games, GR(1)-style objectives, or richer Emerson-Lei Objectives?
  • When does an LTL safety formula need a centralized joint-action shield versus a factored or distributed construction?
  • When can temporal task structure be learned as an automaton rather than specified upfront?
  • How can temporal task structure compose through reusable subtask automata?
  • How do temporal logics change when agents can revise rules, issue private queries, or carry causal state across time?
  • Which temporal safety specifications can be decomposed into local assume-guarantee obligations for multi-agent shields?
  • When do finite-trace temporal values serve as strategy-comparison criteria rather than only monitors or automata tasks?
  • When do real-time requirements need duration-sensitive interval logics rather than event-order logics alone?
  • Which temporal-logic workflows depend on automata-theoretic translations rather than only semantic trace definitions?
  • Which temporal-logic workflows are best understood through monadic second-order definability over words or trees?
  • When does a specification need linear-time trace semantics, and when does it need branching-time quantification over possible futures?

Key Concepts

Key Sources

Adjacent Foundations