Abstract

Semantic contextualism claims that sentences ascribing knowledge or lack thereof (sentences like "S knows that p" and "S doesn't know that p") are context dependent: they express different propositions in different contexts of utterance. "Knows that" is either indexical or elliptical and refers to different relations in different circumstances. Invariantism argues in turn that the knowing relation is just one and the proposition expressed by a given knowledge ascription does not depend on context. A special case of invariantism is interest-relative invariantism proposed recently by Jason Stanley. According to |R| knowledge is conceptually linked to practical interests. Whether or not true beliefs count as knowledge depends on the costs of being wrong; on the stakes in a given situation (I may know that the bank will be open on Saturday if I have no important business to be done in the bank; if however I have an impending bill coming due I will not count as knowing that the bank will be open on Saturday even though my evidence as regard bank opening hours has not changed). It is argued in the paper that the difference between various contexts in which knowledge ascriptions are made is not a difference in stakes. Moreover knowledge has to be distinguished from willingness to be sure. One may know something but not be sure about it and may be sure about something but not to know it. The higher the stakes the more sure one usually wants to be, but the height of the stakes does not have such an impact on knowledge