Ontological Text Parser

From dandi08

This is a simple idea, but very hard to explain, as it will need to operate on multiple levels. However, the bulk of the program should be workable by the end of the quarter.

The basic idea is to simulate human cognition through a non-heirarchical ontology, based on the principles and ideas found in our book "Women, Fire, and Dangerous Things", and allow modification and extension of it through natural language typed by the user.

At the same time, it should ALSO allow the visualization, creation and manipulation of specific examples of that ontology, in an imaginary world, and be able to answer questions about that world, based on the ontology, also through the use of natural language.

It's kind of hard to explain in simpler terms, but here is an imaginary example run of the sort of program I am talking about:

Hello.
->Chickens are usually brown. Are they? I didn't know that.

->Imagine a chicken. Ok.

->Is the chicken a bird? Yes. All chickens are birds.

->What color is it? Which "it" do you mean? The chicken?

->Yes Probably brown. Chickens are usually brown.

->The chicken is red. Ok.

->What color is the chicken? Red.

->Eggs are always white. Are they? I didn't know that.

->Imagine an egg. Ok.

->What color is the egg? White. Eggs are always white.

->The egg is brown. But I thought that eggs are always white? What do you mean?
a) The egg is not brown
b) Eggs are not always white
c) THIS egg is an exception
d) Never mind.
-> c Ok.

->What color is the egg? The egg is brown.

I have a lot more specific examples of the sort of behavior I want to emulate, involving many notes on irregularities and ambiguities and how I want to handle them. However, the basic idea is that the program takes natural language and applies it to an imaginary world, which is modeled around an ontology. The project will combine linguistics, python, case studies, and, of course, ontologies.