This thesis investigates the structure of knowing, and it argues that existential trust is an epistemological category. The aim of the dissertation is to develop a view according to which all human activity is seen as an activity of a lived body, and in which the understanding of the structure of such activity is regarded as central for the solution even of epistemological problems. This view is not rooted in any one philosophical tradition, but circles around activity of the lived body. It connects thinkers who in other respects belong to different "isms" in philosophy. Central to the dissertation are Aristotle, Dewey, Merleau-Ponty, Wittgenstein, Ryle, Anscombe, Polanyi, and Grene. Michael Polanyi's concept of tacit knowledge, and connected concepts like attend to, attend from, and subsidiary awareness, are presented. Different kinds of subsidiary awareness, not noted by Polanyi, are distinguished. It is also argued that Polanyi has not seen all the implicatons of his view that instruments can be interiorized and be part of the lived body. Conversely, parts of the normally lived body can be exteriorized. Nor has Polanyi seen that one has subsidiary awareness of oneself as a certain kind ofperson. This fact, in turn, is shown to have implications for the way we constitute ourselves as agents. Since we are engendered agents, we always attend from gender. In the last decade, the concept of trust has definitely entered epistemology. Mostly, it has been in terms of trusting testimony and/or testifier. This thesis wants to deepen that account. There is a more fundamental kind of trust, namely trust in oneself and trust in the world. It is called existential trust, and it is of epistemological importance, too. Existential trust, it is claimed, is necessary in all acts of knowing. Probably, this fact is hard to discover without having recourse to a distinction like that between attending from and attending to. Existential trust is shown in the way we attend from something. Observation and thinking are central epistemological categories, of course, but they should be supplemented by trust. Without trust they cannot perform any epistemological work. Linguistically, 'to know' is a state verb. Superficially, it describes only a state, not an activity. This fact, however, does not contradict the view of knowledge put forward in the dissertation. To know is to have a disposition to perform successfully either some kind of action (= knowing how), or to perform some kind of true assertive speech act (= knowing that). Basically, knowing is an activity. / digitalisering@umu