ABSTRACT 

Legal knowledge consists of statutes and old cases. As a statute is a set of 
legal rules, inference by a statute is realized as rule-based reasoning. However, 
legal rules often contain legal predicates (legal concepts). Legal concepts are 
ambiguous and their strict meanings are not fixed until the rules are applied 
to actual facts. To apply legal rules to actual facts, we need rule interpretation 
and matching between legal concepts and concrete facts. To realize this, old 
cases are often referenced and their explanations are reused. Consequently,
legal reasoning can be modeled as a mixed paradigm of rule-based reasoning 
and case-based reasoning. However, it takes this model a long time to search 
for similar cases and to draw conclusions, and a complex mechanism is needed 
to manage several inference engines. To solve these problems by parallel 
inference, we developed a legal reasoning system, HELIC-II, on the parallel 
inference machine. 

Overview of HELIC-II 

HELIC-II draws legal conclusions for a given case by referencing the rel-
evant statute and old cases and outputting them in the form of inference 
trees. HELIC-II consists of a rule-based engine and a case-based engine. The 
rule-based engine refers to legal rules and draws legal consequences logically. 
Following is a legal rule for manslaughter caused by negligence. Every legal 
rule can be represented with this kind of inference rule. 


     manslaughterCausedByNegligence("comment",[article = 210],
           [person(A,[]),person(B,[]),{{A = B }},
              action(_action, [agent = A]),
              negligence(_negligence, [agent = A, action = _action]),
              causation(_causation, [cause = _action, result = _death2]),
              death(_death2, [agent = B])]
              
           [[manslaughterCausedByNegligence
                   (_, [agent = A, action = _action])]]).

This rule contains the legal concept of "negligence". Whether this "negli-
gence" has occurred or not is the key problem that must be judged for each 
case that is like the following case of Mary. 


					- 88 -