I propose to define one concept of evidence, which, I claim, was lacking. […] Concepts of testimony and authority were not lacking: they were all too omnipresent as the basis for the old medieval kind of probability that was an attribute of opinion. Testimony is support by witnesses, and authority is conferred by ancient learning. People provide the evidence of testimony and of authority. What was lacking, was the evidence provided by things. The evidence of things is not to be confused with the data of sense, which in much modern epistemology, has been regarded as the foundation of all evidence. […] The evidence of things is distinct from testimony, the evidence of witnesses and of authorities. […] The Renaissance had it the other way about. Testimony and authority were primary, and things could count as evidence only insofar as they resembled the witness of observers and the authority of books. Our form of the distinction between these two kinds of evidence testimony and evidence of things, is quite new.