An enormous number of academic articles have been written on knowledge, knowing, and understanding in many different fields, from linguistics to psychology. For centuries philosophers have studied the nature of the knowledge we have about the world. After all that has already been said, is it possible to add anything truly new to the conversation? Timo Honkela answers in the affirmative.

This view is based on the possibility that computers can help us model various processes and phenomena related to knowledge. When it comes to formulating theories and models of knowledge, computers partly replace pen and paper. Honkela says that important divisions of knowledge include those between verbal and nonverbal, empirical and indirect, and individual and communal. Computers allow us to construct socio-cognitive models of, for example, how languages form, how symbolic representations are learned and used, how social forms of communication and conceptual systems are formed, and the subjective and contextual nature of the use of languages and concepts. Theories and models are constructed on different levels of abstraction, and they are compared with traditional understanding in the social sciences and the humanities. Honkela’s long-term goal has been to construct a theory of meaning in which the aforementioned themes of subjectivity and contextuality are taken into account seriously as empirical phenomena. Along the way, Honkela and his research partners have used various machine learning models to analyse the likes of Grimms’ fairy tales, online forums, interviews, speeches by American presidents, Shakespeare’s sonnets, abstract art, people’s physical activity and its verbal description, and historical newspaper articles.