Paraconsistent logics feel like the “outside the box” box to me. They avoid the most obvious difficulties, but there is no good explanation of why any particular paraconsistent logic should be the right solution to this particular problem and, while they are relatively well known, they have not produced many interesting results. The problem of ‘normal’ uncertainty, in particular, was solved with probability theory, not nonclassical logics.
If a system believes A and not A, it has already made a mistake. If a well-designed system represents A and not A, it must be something other than a belief in the usual sense. Representing a specific non-belief thing using paraconsistent logic could work, but just throwing paraconsistent logic at the problem in order to avoid difficulties due to explosion seems like a dead end.
Paraconsistent logics feel like the “outside the box” box to me. They avoid the most obvious difficulties, but there is no good explanation of why any particular paraconsistent logic should be the right solution to this particular problem and, while they are relatively well known, they have not produced many interesting results. The problem of ‘normal’ uncertainty, in particular, was solved with probability theory, not nonclassical logics.
If a system believes A and not A, it has already made a mistake. If a well-designed system represents A and not A, it must be something other than a belief in the usual sense. Representing a specific non-belief thing using paraconsistent logic could work, but just throwing paraconsistent logic at the problem in order to avoid difficulties due to explosion seems like a dead end.
I’m not advocating them, just presenting them informally to see if they are of use for UDT.
Okay, I wasn’t quite sure what you personally thought of them. I obviously don’t object to informing people about various ideas in logic.