KERNEL KNOWLEDGE VERSUS BELT KNOWLEDGE IN DEFAULT REASONING - A LOGICAL APPROACH

In this paper, we formalize in logic the dichotomy between kernel (or hard) knowledge versus belt (or soft) knowledge introduced by Popper and Lakatos in the philosophical logic of scientific discovery. The logic we obtain, default ionic logic, generalizes Reiter's default logic to a continuous and monotonic reasoning system. Default ionic logic gives a solution so some well-known non-monotonic reasoning problems: Lottery paradox, weak implication, disjunctive information, default transformation, normal versus non-normal defaults, the Yale shooting problem and several of its variants, Haugh's Assassin problem, Kautz's Vanishing car problem, and Haugh's Robot problem.