This result is used, with standard frequency interpretations of probabilities, in an analysis of the foundations of statistical inference. This analysis establishes the likelihood function as the appropriate basis from.which statistical inferences can be made directly, for the familiar general purposes of informative inference. For the numerical values of the likelihood function, this analysis provides direct interpretations in terms of probabilities of errors. These probabilities admit frequency interpretations of the usual kind, but they are not in general defined with reference to the specific experiment from.which an outcome is obtained: they express intrinsic objective properties of the likelihood function itself, which this analysis shows to be appropriately relevant and directly useful for purposes of informative inference. The relations of this analysis of problems of informative inference to problems of testing statistical hypotheses, decision-making, conclusions, and Bayesian treatments of inference problems are discussed briefly.