In recent papers Fred Sommers and Michael Lockwood have independently argued that the distinction between the ‘is’ of predication and the ‘is’ of identity (henceforth: the
IP-distinction) is not well-founded. This claim is somewhat obscure since, on the theories they advocate, it is not only still possible to distinguish between the ‘is’ of predication and the ‘is’ of identity, but important to do so on pain of turning good arguments bad. Sommers' way of putting it, namely that we don't need identity, is no better. Of course we don't needidentity since, in second-order quantification logic, identity is definable in terms of predication. Perhaps the best way of putting their point is by means of Sommers' claim ( op. cit.p. 500) that identity statements can be formalized by monadic relations. We can therefore mark the IP-distinction by the essentially polyadic nature of identity.