In attempting to summarize recent work on functions of granular prefrontal cortex in primates, including man, and possible homologues of these structures in rodents and carnivores, four question need to be asked, the questions of 'where', 'when', 'what' and 'how'. Progress since the Pennsylvania Symposium has been considerable for questions of 'where' and 'when': localization of symptoms ('where') fields a double gradient (up-down and back-to-front) in monkeys, and a right-left difference in man; analysis of time factors ('when') distinguishes early and late lesions, single and serial removals, or succeeds in recording and stimulating at critical moments during performance. However, problems of 'what' and 'how' are still largely unsolved: we do not yet know what various prefrontal symptoms sigify, in terms of normal function, and are only beginning to see how individual prefrontal neurons act and interact. Advances on these questions are likely if one exploits an extended version of those hypotheses about prefrontal physiology that attribute to these structures neither purely sensory nor purely motor functions but consider them instead as sources of 'corollary discharges' whereby the organism presets its sensory systems for the anticipated consequences of its own action.