<p>The purpose of this dissertation is to argue that epistemic contextualism, which proposes that the word ‘know’ is a context-sensitive term, is seriously deficient and therefore indefensible. Since epistemic contextualists claim that their semantic theory of ‘know’ contributes not only to a linguistic model of knowledge ascription but also to a unified solution to some important puzzles in epistemology, I divide my thesis into two basic parts. In the first part (i.e., Chapters 2 and 3), I argue that the proponents of both binary and ternary accounts of the supposed context-sensitivity of ‘know’ fail to provide a reasonable linguistic model of knowledge ascription. My argument in Chapter 1 indicates that ‘know’ cannot be treated as a binary context-sensitive term that is similar to paradigmatic indexical terms or gradable adjectives. Chapter 2 takes contrastivism as a representation of the ternary account of the supposed context-sensitivity of ‘know’ and argues that this theory is in an even worse position because it even fails to capture the supposed phenomena of the context-sensitivity of knowledge ascription. The second part (i.e., Chapters 4, 5, and 6) argues that epistemic contextualism does not provide us with a really satisfactory solution to the puzzles of skepticism, the epistemic closure principle and fallibilism. On the contrary, its rival, invariantism, with some support from pragmatics, psychology of belief and experimental philosophy, is able to solve the above puzzles in a quite nice way. At the end of my thesis (i.e., Chapter 7), I systematize the observations, the evaluations and the critiques of epistemic contextualism from the previous chapters and indicate that epistemic contextualists even fail to establish their supposed phenomena of the context-sensitivity of ‘know.’</p> / Doctor of Philosophy (PhD)