In this paper I want to describe and motivate an approach to knowledge that I call the Causal Indicator Analysis. My strategy will be to sketch (in Part I) the main features of an adequate account of knowledge, then use my sketch (in Part II) to reveal some of the faults of some of the main analyses defended today. I will be particularly interested in discussing the work of Fred Dretske, whose views have significantly influenced my own. With these tasks behind me, I will offer my own account in Part III, and argue that it has the features an account of knowledge should have.