John, this approach seems very promising. I'm wondering how it accommodates hierarchies of classes or types. For example, how would it represent our knowledge that living things have animals, plants, and other subtypes, and that animals have mammals, birds, fish, reptiles, etc. as subtypes, and that mammals have many subtypes including humans, dogs, cows, tigers, etc. ? It wouldn't necessarily be the type hierarchy that biologists use, but rather the one that we all learn as children and modify as we learn more. And, of course, there are many similar type hierarchies that we learn and use, such as types of tools, types of foods, types of chothing, and on and on.