Automated Discovery of Fallacies in Legislative Processes

Abstract

This research or study presents a computational logic tool for automatic discovery of fallacies, which may be inherent, or, introduced intentionally or unintentionally in legal texts. Sound reasoning through legal text has always been a challenge in Natural Language Processing (NLP) in Artificial Intelligence (AI) since the fourth century, wich has received different computational approaches for solution. I have explored and presented logic techniques through the formalism of legal text from its natural form to First Order Logic (FOL) and Prolog programme, which results in the provision of clarity, comprehensibility and deductive reasoning of the text. This as well maintains sound reasoning through the text, which supports decision-making process that will always lead to the same conclusions. I formalised the Citizenship Act and the Fundamental Human Rights and Freedom of Ghana as the knowledge base of the system. I also formalised two Supreme Court cases as testbed to the system in FOL, and the formalised text was implemented in Prolog progrmme for the automated reasoning process of the sytem. This allows for the discovery of fallacies in a claim made against an opponent, facts established by the opponent, and the law employed for the legislation of the case in court. I have also presented an algorithmic framework here in pseudocode for the discovery of logical fallacies in the text. The ontology design of the philosophical research methodology was employed in the conduction of this research, which guided the techniques used for the formalism of the logic tool.