If the refinement procedure lacks a unique outcome, it cannot coherently 'favor' a solution in contr...