Date: 2022
Type: Thesis
The uncanny valley of computable contracts : analysis of computable contract formalisms with a focus towards controlled natural languages
Florence : European University Institute, 2022, EUI, LAW, PhD Thesis
IDELBERGER, Florian, The uncanny valley of computable contracts : analysis of computable contract formalisms with a focus towards controlled natural languages, Florence : European University Institute, 2022, EUI, LAW, PhD Thesis - https://hdl.handle.net/1814/75015
Retrieved from Cadmus, EUI Research Repository
The possibility of using computable contracts — similar to the smart contracts currently used for blockchain applications — has launched a new era of digital collaboration and experimentation, heralding new practices alongside legal and interdisciplinary challenges. Computable contracts — that is, legal contracts that can be automated but are, according to my definition, ideally also readable by humans — are of primary importance for transposing legal contracts to the digital world in a functional, accessible, and inclusive way. This framing of the issue touches upon law, socio-legal considerations, linguistics, computer science, and userinterface questions. This work chronologically describes the struggles and development process that language inventors, legislators, legal professionals, and computer scientists have encountered in addressing the challenge of computable contracts. For the contractual and law-related struggles, I put current technological research efforts into the context of past and present legislators. Moreover, I present computable contracts as a form of computable law in the context of the contract as a business tool as well as in the context of transnational legal theories that describe the law as an amalgam of international business, public and private interests, and technology, forming a normative but fluid whole. Based on these explorations, I define criteria that useful, accessible, and inclusive computable contracts should respect to be functional, meet the standards of a legal contract and represent its contents, and be accessible and inclusive as a text. Based on these criteria, I then compare several formalisms, specifically a state machine formalism, a blockchain programming language (Solidity), a logic programming language (Prolog), a general controlled natural language (CNL) and my own explorative implementation of one such language. I then analyse how far these formalisms meet the criteria, highlighting common issues and benefits and comparing them to each other and with natural language contracts. Finally, I introduce the term ‘uncanny valley’, originally used in research into the perception of humanoid robots, into work with CNLs. I discovered the possibility for this transposition during my research and hope this term makes the phenomenon of CNLs and their possible uses and limitations more prominent. Along the way, I undertake exercises of categorising computable contract terminology and contracting technology that may be useful for the present work but also the field of computable law in general. As a result, the field of law and technology research may benefit from improved terminology and delineation of categories, and the research into human-readable computable contracts from improved insights and examples of how the analysed languages meet the criteria to create a useful, human-readable (accessible) and inclusive computable contract. These categorisations and comparisons show that our use of and interaction with technology is still severely limited, especially in law and contracts, to the detriment of users, efficiency, quality and comfort. For example, contracts are a tool to create business relationships, and while businesses have adapted to electronic transmission, the way of writing and implementing contracts has not fundamentally changed. Moreover, a core aspect of law is the language of law, and for decades people have researched and laboured to understand both language and the law and the relationship of both to technology. However, while language is of primary importance in law, technology still struggles with understanding language, especially legal language. In this period of ever more complex technologies governed by computer code and law, it is essential to bring code and law together. Specifically, contracts, in this case, should be properly included in digitisation and not only as artefacts transposed from the physical world, such as PDF documents. This work shows which textual artefacts of contracts can be transposed into computable instructions with CNLs and where the limitations are. In conclusion, CNLs are currently a good bet for creating humanreadable but also computable contracts that still use natural language text, excluding alternative custom interfaces such as graphical interfaces. At the same time, CNLs are hard to develop and difficult to use correctly.
Cadmus permanent link: https://hdl.handle.net/1814/75015
Full-text via DOI: 10.2870/638745
Series/Number: EUI; LAW; PhD Thesis
Publisher: European University Institute
LC Subject Heading: Contracts