Re:The Pain of Real Linear Types in Rust
by Mathnerd314 on Tuesday May 09, 2017 @02:29PM
(#14302647)
Well, if relevance logic is part of linear logic, and it's a true subsumption relationship, then you're saying Belnap doesn't think relevance logic is relevant anymore. So in other words he's pretty much rejected logic altogether.
I guess it makes sense, his most recent papers are about indeterminism (http://www.pitt.edu/~belnap/futurecontingents.pdf), which logic doesn't handle well, but I'm not quite sure how you'd represent the branching-universe model he uses in a type system. I guess you'd need the history operator he uses, m/h = moment m on history h.
Re:The Pain of Real Linear Types in Rust
by hyperpape on Tuesday May 09, 2017 @04:47PM
(#14303674)
I was referring to the subsumption under substructural logic, not a subsumption under linear logic. Belnap seemed to view linear logic as rather different from his own interests in relevance logic.
I guess it makes sense, his most recent papers are about indeterminism (http://www.pitt.edu/~belnap/futurecontingents.pdf), which logic doesn't handle well, but I'm not quite sure how you'd represent the branching-universe model he uses in a type system. I guess you'd need the history operator he uses, m/h = moment m on history h.