The Role of Decisions and Evaluations in Judicial Reasoning

A Challenge for the Computational Approach to Legal Theory

Abstract

One of the main objectives of the computational models to legal theory is to design computer programs capable of performing (certain kinds of) legal reasoning. A common objection to this thesis is that legal reasoning involves evaluations and decisions – tasks that, arguably, can only be carried out by humans. A possible way to answer this objection consists in following an argument put forward by Alchourrón and Bulygin in a very insightful paper on the limits of logic in legal reasoning and the application of artificial intelligence to law and state that, at least in some cases (the “easy” ones), the role of evaluations and decisions – and, therefore, the discretionary power held by the courts – is rather limited. In fact, the argument goes, in easy cases judges do not perform genuine evaluations, but they restrict themselves to record the evaluations of the social group to which they belong. From this consideration it should follow that, at least in easy cases, legal and judicial reasoning can be automated and performed by (deemed) intelligent machines without too much difficulty. Contrary to this view, in this paper, I will argue that value judgements and decisions are all-pervasive in legal (viz., judicial) reasoning and, consequently, the idea that some legal decisions could be seamlessly delegated to computers is more problematic than proponents of the computational approach might assume.