Abstract

IN his recent paper ‘On Concepts of Truth in Natural Languages’ [7] Professor Sommers proposes a new solution to the Liar paradox. He claims that ‘an acceptable natural solution must possess the following two features: Barriers to the Liar paradox are discovered in natural syntax, The natural barriers which keep out the Liar do not also exclude meaningful and harmless linguistic reference’. If we reformulate Sommers’ claim we get the puzzling contention that the natural barriers of a semantic paradox should be looked for in natural syntax. It is puzzling for there is no reason to seek solution of some of the most difficult problems in the semantics of natural languages in the syntax of natural languages. Now, this puzzling air is not due to our introduction of the epithet ‘semantic’ to the Liar paradox. Sommers recognizes the semantic character of the paradox and its solution when he states that the principle which is claimed to be the barrier of the paradox ‘though.… a syntactic principle, the restraints it imposes on natural language are semantic. It sets bounds on the expressive powers of the language by prohibiting certain kinds of state reference or state characterization’.