His argument here (that I found most convincing) was children with hydranencephaly. Many of them have very little cortex but still seem to experience a roughly normal range of emotions in appropriate context.