If all that is meant is that a given object cannot be known except through the knowledge that relates to it, this is indeed true. Yet knowledge itself is not grasped by something other than knowledge; rather, universal or absolute knowledge is apprehended through a particular, determinate act of knowing. Hence the term “knowledge” in one of the two premises does not carry the same meaning it bears in the other, so circularity does not arise. When someone says, “Nothing can be known except by knowledge,” what he intends is that every object of knowledge is only grasped by means of knowledge.
As for the objection: “If knowledge were known by something other than itself, circularity would follow,” we reply that the definiendum here is not a specific cognition of a particular object; it is knowledge in its universal sense. Universal knowledge is known by a restricted, particular cognition, and there is no contradiction in this—just as one piece of information may be conveyed by another, so information in its absolute or general sense can be conveyed by a specific, particular report.
Some answer by saying: the dependence of conceiving something other than knowledge upon the occurrence of knowledge of it—that is, by means other than knowledge itself, not upon conceiving knowledge—cuts off the alleged circle. This is one of the two replies, for the occurrence of knowledge of any object depends on the occurrence of a particular cognition, not on conceiving that cognition; what the definition seeks is to delimit the conception of knowledge, not to bring about the acquisition of knowledge of those objects.
The second reply is as follows: conceiving anything other than knowledge depends on the occurrence of knowledge of that object, not on the occurrence of the knowledge that is being defined—namely, the comprehensive, exclusive universal. This latter is what the definition aims to fix. Even if conceiving something other than knowledge did depend both on having knowledge of the object and on conceiving that knowledge, such dependence would still not remove the need for a definition. Indeed, it may be said that whoever knows something might also know that he knows it, and the earlier discussion would return.
To clarify: in the external world there are determinate things, such as a particular human being or a particular act of knowing