Thanks for the Einstein reference. I wasn't familiar with it. However, my comment expressed little more than well established principles in epistemology and the philosophy of science. Einstein in his essay discusses the same general issue. My favorite demonstration of the difficulty of defining (non-mathematical) knowledge is the Gettier problem. The coolest thing about it is that if we try to come up with a definition of knowledge that is always agreed upon by everybody -- one that fully reconciles what we can intuitively and "logically" consider to be knowledge with all forms of the Gettier problem[1] -- we find ourselves with such a narrow definition that hardly covers anything we consider to be "known". This means that in science, as with everything else, we must adopt more lenient -- or pragmatic -- definitions of knowledge, but those definitions are not rigorous in any mathematical sense, and are always lacking.

[1]: http://en.wikipedia.org/wiki/Gettier_problem

Also, see:

http://en.wikipedia.org/wiki/Philosophy_of_science

http://en.wikipedia.org/wiki/Epistemology

Yes, that is his position, but it's a false dichotomy.