There are four [questions] concerning every fact: (1) whether it is, (2) what it is, (3) of what sort it is, (4) why it is. We apprehend from evidence that these accidental things are, and we judge the 'what it is' from the definition of them. From a definition we do not know that the accidental attributes are, but what they are, of what sort and why [sc. they are what they are]. Thus through the fact that the accidental attributes are, we know the 'what it is' of the definition. But through the 'what it is' of the definition [we know] the remaining questions about the accidental attributes.