Natural Kinds

• 
• So why did we say this was a dog instead of a mammal, animal, or thing.
• Rosch and Mervis (e.g. Basic Object in Natural Categories) did some solid work in how people categorise objects in the 70s.
• It turns out that people (cross-culturally) categorise things in a very similar way.
• There are basic level categories (like dogs), and people generally say things are this level.
• Super-categories like mammal, and sub-categories like collie, are much less likely. Moreover, if you force people to answer this way they are slower.
• Another of their papers is Family resemblances... This points out the elements of categories are not necessary and sufficient, but they tend to share a lot of features. This fits in with Wittgenstein, but less well with C4.5.
• (For that matter, why didn't we say it was a picture?)