In the paper we offer a logical explication of the frequently used, but rather vague, notion of point of view. We show that the concept of point of view prevents certain paradoxes from arising. A point of view is a means of partial characterisation of something. Thus nothing is a P and at the same time a non-P (simpliciter), because it is a P only relative to some point of view and a non-P from another point of view. But there is a major, complicating factor involved in applying a logical method that is supposed to provide a formal and rigorous counterpart of the intuitively understood notion: ‘point of view’ is a homonymous expression, and so there is not just one meaning that would explain points of view. Yet we propose a common scheme of the logical type of the entities denoted by the term ‘point of view’. It is an empirical function: when applied to the viewed object in question, it results in a (set of) evaluating proposition(s) about the object. If there is an agent applying the criterion, the result is the agent’s attitude to the respective object. The paper is organised into two parts. In Part I we first adduce and analyse various examples of typical cases of applying a point of view to prevent paradox. These cases are examined according to the type of the viewed object: a) the viewed object is an individual and b) the viewed object is a property or an office. In Part II we then show that the method described in Part I can be applied also to the analyses of agents’ attitudes. We explain how an agent can believe of something that it is a P and at the same time a non-P: the agent applies different viewpoint criteria to the viewed object. The inversion of perspective consisting in the perspective shifting from the believer on to the reporter in the case of attitudes de re, and from the reporter to the believer in the case of attitudes de dicto, is also analyzed. We show that there is no smooth logical traffic back and forth between such attitudes unless some additional assumptions are added, and prove that they are not equivalent. By way of conclusion, we explicate the notion of conceptual point of view and analyze cases of viewpoints given by conceptual distinction. We show, finally, that the proposed scheme of the type of point of view can be preserved, this time, however, in its extensional version.