Can Errror-Statistical Inference Function Securely?

Among the important tasks for a theory of ampliative inference are the classification of the kinds of “material postulates”1 that are needed for specific inferences and the explanation of how those assumptions function in such inferences. The present paper undertakes an analysis of Deborah Mayo’s error-statistical (ES) account of scientific evidence in order to clarify just these points as well as to explain and illustrate the importance of the philosophically neglected notion of the security of an inference [21]. After finding that, on the most straightforward reading of the ES account, it does not The term is John Norton’s [18]. Norton proposes that material postulates and not formal inductive schemas underwrite inductive inference. The perspective of the present paper is that while Norton rightly emphasizes the importance of material postulates in underwriting specific inductive inferences and explaining the growing power of inductive reasoning, the material postulates do not themselves tell us how they should be used. See also [25].