The author sees enormous potential for expert system applications in law. Although many existing projects have contributed considerably to the understanding of this field, Susskind feels that insufficient attention has been given to the presuppositions used in building legal expert systems. In short, efforts in computer-assisted legal reasoning have yielded far less spectacular results than those in other disciplines such as medicine and geology.
The major thrust of this work surveys those aspects of recent jurisprudence (mainly analytical jurisprudence) that are central to expert systems in law. Susskind deals with the nature of legal rules, legal reasoning, legal knowledge, and legal systems. He concludes that there are no theoretical obstacles that would hinder the development and design of rule-based systems of limited scope.
Susskind’s book is intended for readers trained in law. He argues correctly that legal expert systems should only be used by persons having legal expertise. As a result, his book assumes a great deal about legal reasoning that non-lawyers do not know. I am not suggesting that the author attempt to rectify this problem, because no single book could teach legal reasoning even to a professional in another field. The problem is that legal reasoning differs so significantly from logical analysis as used in other fields that legal text usually leaves non-legally trained readers with a false impression of comprehension. This presents a serious dilemma. Expert systems for law must be developed by people who are expert in both law and computer science, and few lawyers are expert in computer science. Susskind, therefore, urges that lawyers become qualified for this task.
This work reexamines traditional jurisprudence in a new context. The author believes that jurisprudence can and should supply the models for building expert systems used in law, since these systems make assumptions about the nature of law and legal reasoning. Consequently, the book focuses on a reexamination of analytical jurisprudence in order to find insights that can be used to develop legal expert systems. Susskind has succeeded in finding common ground in the theories of diverse thinkers on jurisprudence and in focusing these theories on the presuppositions underlying expert systems.
Susskind’s discussion of forward and backward chaining illustrates his successful synthesis of legal analysis and computer logic. He views legal reasoning as recursive, and within legal expert systems recursion is demonstrated in both forward and backward chaining. He then poses the critical question of whether we want expert systems that reason forward or backward through the law. He concludes that judges require no more than forward reasoning while trial lawyers require backward reasoning for most cases. Susskind relates this discussion to the work of Dewey, Harris, Hutcheson, and Frank.
This work is noteworthy in that it successfully integrates legal theory and computer science. In addition, it presents a synthesis of theory and practice. Susskind has written an important book that should remain the standard in the field of legal expert systems for some time.
