We build a generic methodology based on learning and reasoning to detect specific attitudes of human agents and patterns of their interactions. Human attitudes are determined in terms of communicative actions of agents; models of machine learning are used when it is rather hard to identify attitudes in a rule-based form directly. We employ scenario knowledge representation and learning techniques in such problems as predicting an outcome of international conflicts, assessment of an attitude of a security clearance candidate, mining emails for suspicious emotional profiles, mining wireless location data for suspicious behavior, and classification of textual customer complaints. A preliminary performance estimate evaluation is conducted in the above domains. Successful use of the proposed methodology in rather distinct domains shows its adequacy for mining human attitude-related data in a wide range of applications.