Key points are not available for this paper at this time.
We provide the first proof of confluence for cut elimination in multiplicative-exponential linear logic proof-nets that is not based on Newman’s lemma or strong normalization, not even indirectly. To achieve our goal, we take inspiration from Tait and Martin-Lof’s method based on parallel reduction for the lambda-calculus, even though the wilder realm of untyped proof-nets makes the proof subtler and more challenging.
Guerrieri et al. (Mon,) studied this question.