I am interested in implementing abduction over a full first order theory ie it may be non-Horn. (Aside: Almost all the references I've seen for abduction operate over Horn theories eg "Modeling Abduction over Acyclic First-Order Logic Horn Theories in Answer Set Programming: Preliminary Experiments". Is there a reason for this other than lack of minimal hypotheses?) Are there frameworks that might be suitably modified for my purposes? Alternatively, does anyone know of a way of coaxing a SMT solver into perhaps giving me something close to what I need?
thanks!