Wednesday, 21 March 2007

"Tricky" Argumentation Frameworks

Amongst the set of "tricky" frameworks are those with loops.

Odd-loop frameworks pose problems as follows:

- 1-loop, argument X is attacked by argument A which attacks itself (X <- A -> A). Does A's attack on X have any meaning, given that it attacks itself?

- 3-loop, X <- A -> B -> C -> A. What does A's attack on X mean, given that the argument (C) that defends it from B also attacks it?

- 5-loop, A -> B -> C -> D -> E -> A. Which argument(s) do we believe, if any?

Even-loop frameworks pose problems as follows:

- 2-loop, A -> B -> A. Can we believe one of A or B, or not?

- 4-loop, A -> B -> C -> D -> A. Which argument(s) do we believe, if any?

No comments: