Classical phonological constraints apply to individual candidates. Yet, some authors have proposed constraints that instead apply to sets of candidates, such as distinctiveness constraints (Flemming 2002) and Optimal Paradigm faithfulness constraints (McCarthy 2005). As a consequence, the classical constraints need to be ''lifted'' to sets by summing across the set. Is this assumption of constraint summation typologically innocuous? Or do the classical constraints make different typological predictions when they are summed? Extending a result by Prince (2015), we characterize those models of constraint interaction for which constraint summation is typologically innocuous. As a corollary, typological innocuousness is established for OT and HG.