Words and languages are the direct representations of our thoughts and perceptions. Accordingly, the only access to an individual's thoughts and psyche is through the language that he/ she speaks or writes. Psychological disorders and mood states such as depression and PTSD therefore, can be identi ed and predicted by continuously analyzing a patient's discourse. In doing so, we utilize the method of predictive linguistics, which determines cognitive mood states by computationally modeling the notions of 'mind axiology' and 'emotional states'. Our proposed system expresses cognitive states in terms of axiology, i.e., the system of positive and negative values associated with concepts that are universally accepted. Axiological notions provide more insights about the various cognitive states because a patient's words are only intelligible if axiological elements like conception, perception, and intention are taken into consideration. e framework for our analytics engine consists of multiple modules responsible for coherently and systematically retrieving, parsing and processing a patient's discourse. e backbone of this system relies on a learning algorithm that accounts for various valuation criteria such as time-based, intrinsic, consequent and contextual value analysis, while also building a nested network of mood states and their dependencies. By executing this mechanism and correlating our computational data to results calculated by a psychological assessment testing, we are better positioned to classify a patient's mood states-e.g., as being depressive or not. is work will have implications in the area of cognition research and practical applications such as automated neurological and psychiatric assessments, advanced human-computer interaction (HCI), as well as others.