Abstract: This talk defends a contextualist theory of ‘knowledge’ ascriptions. I argue that in some sentences, the implicit argument of ‘knows’ is bound by a quantifier. The natural readings of these sentences can be generated by contextualist theories, but not by competing interest-relative theories of knowledge. In addition, I argue that the contextualist can explain distinctive patterns in our judgments about sentences in which ‘knows’ is embedded under change-of-state verbs. Along the way, I argue that the most common definitions of ‘encroachment’ and ‘interest relativity’ are seriously flawed.