ON THE PRODUCTION OF KNOWLEDGE