Grant for Semantic Primes

My Kakenhi grant application has been accepted. The title is "Creation and Evaluation of Tacit Knowledge Based on Semantic Primes". Since I met Anna Wierzbicka during my sabbatical at ANU I have been thinking about the smallest units of meaning and how AI could use them to perceive the world and learn. The problem with current models is that their perception doesn't resemble the way we experience the world. Tensor manipulation has dominated the field because it works better than previous approaches, but its success led to the situation that almost no-one tries to go beyond transformers and explore different paths. I believe that innate capabilities are needed but Chomsky or Pinker don't give engineers hints useful enough. For that reason I decided to explore ideas from cognitive linguistics.