I think Lenat is headed in the right direction but someone needs to include a knowledge base about learning.
We often attribute 'understanding' and other cognitive predicates by metaphor and analogy to cars adding machines and other artifacts but nothing is proved by such attributions.
