We introduce two-sided type systems, which are sequent calculi for typing formulas. Two-sided type systems allow for hypothetical reasoning over the typing of compound program expressions, and the refutation of typing formulas. By incorporating a type of all values, these type systems support more refined notions of well-typing and ill-typing, guaranteeing both that well-typed programs don’t go wrong and that ill-typed programs don’t evaluate - that is, reach a value. This makes two-sided type systems suitable for incorrectness reasoning in higher-order program verification, which we illustrate through an application to precise data-flow typing in a language with constructors and pattern matching. Finally, we investigate the internalisation of the meta-level negation in the system as a complement operator on types. This motivates an alternative semantics for the typing judgement, which guarantees that ill-typed programs don’t evaluate, but in which well-typed programs may yet go wrong.

@article{paperbot2578,
  title = {Ill-Typed Programs Don’t Evaluate},
  author = {Steven Ramsay and Charlie Walpole},
  journal = {Proceedings of the ACM on Programming Languages},
  volume = {8},
  number = {POPL},
  year = {2024},
  doi = {10.1145/3632909}
}