Abstract
Arguments are advanced that a theory of possible worlds cannot be a theory of meaning for modal statements, And lewis's version of the theory in his "counterfactuals" is used as a particular stalking-Horse. (a) 'possible world', Though used referentially, Is defined in a way that makes it non-Referential, And moreover, The theory does not supply or validate proposals for criteria that individuate worlds; hence the theory seems incomprehensible. (b) the theory yields no useable account of truth-Conditions for modal statements. (c) a conceptualist version of the theory cannot overcome these problems. (d) lewis's similarity theory of possible worlds as a semantics for counterfactuals sustains in many cases (including one he cites) intuitively untrue counterfactuals and rejects many intuitively correct ones