Contracts in code?

Abstract

A contract written in natural language describes the parties’ rights and obligations. It cannot, however, ‘do’ anything. It produces legal not physical effects. Traditional contracts do not instruct computers to perform any operations. Contracts expressed in code could. The question arises: can we convert contractual documents from being passive embodiments evidencing agreements into active files containing computer instructions? The ability to guarantee that contracts are performed as agreed, an idea underlying ‘smart contracts,’ depends on the ability to express obligations in code or, more specifically, on the ability of code to express obligations exactly as agreed. Not every obligation can, however, be encoded - be it due to its very nature or to the limited expressiveness of programming languages. Highlighting the interplay between form and substance, this paper presents the complexities involved in expressing contracts in code. Before contemplating the technical aspects of ‘smart contracts’ we must confront some preliminary questions: what are we trying to encode – and for what purpose? Can the entire contract be expressed in code? Should we focus on translating legal prose into code or on encoding specific tasks that serve to discharge obligations? This paper provides the groundwork for more ambitious projects in the area of ‘smart contracts.’