O NE of the objects of the last chapter was to show that, if we wish to ensure that empirical statements can be tested, we need to treat their correctness as dependent not only upon the traditional

objective criterion but also upon the contextual criterion-upon the purposes for which the statements are made. In order to justify this methodological judgment, it is enough to demonstrate that both criteria are necessary and sufficient for testing some empirical statements, and that there is no syntactical or grammatical feature which enables us to identify those statements-if any-for which the objective criterion alone would be sufficient. For if we can demonstrate this (and I think we have done so), it follows that it would be stupid not to make sure, when­ ever a statement is disputed, that its context is adequately determined.