A judgmental analysis of linear logic

We reexamine the foundations of linear logic, developing a system of natural deduction following Martin-Löf’s separation of judgments from propositions. Our construction yields a clean and elegant formulation that accounts for a rich set of multiplicative, additive, and exponential connectives, extending dual intuitionistic linear logic but differing from both classical linear logic and Hyland and de Paiva’s full intuitionistic linear logic. We also provide a corresponding sequent calculus that admits a simple proof of the admissibility of cut by a single structural induction. Finally, we show how to interpret classical linear logic (with or without the MIX rule) in our system, employing a form of double-negation translation. This work has been supported by NSF grants CCR-9988281 and CCR-0306313. Bor-Yuh Evan Chang is currently at the University of California, Berkeley, CA and holds a California Microelectronics Fellowship.