Abstract:Consistency of case bases is a way to avoid the problem of retrieving conflicting constraining precedents for new cases to be decided. However, in legal practice the consistency requirements for case bases may not be satisfied. As pointed out in (Broughton 2019), a model of precedential constraint should take into account the hierarchical structure of the specific legal system under consideration and the temporal dimension of cases. This article continues the research initiated in (Liu et al. 2022; Di Florio et al. 2023), which established a connection between Boolean classifiers and legal case-based reasoning. On this basis, we enrich the classifier models with an organisational structure that takes into account both the hierarchy of courts and which courts issue decisions that are binding/constraining on subsequent cases. We focus on common law systems. We also introduce a temporal relation between cases. Within this enriched framework, we can formalise the notions of overruled cases and cases decided per incuriam: such cases are not to be considered binding on later cases. Finally, we show under which condition principles based on the hierarchical structure and on the temporal dimension can provide an unambiguous decision-making process for new cases in the presence of conflicting binding precedents.

What problem does this paper attempt to address?

### Problems the Paper Attempts to Solve The paper attempts to address the issue of precedent conflicts in legal case reasoning. Specifically, it focuses on how to handle inconsistent precedents (i.e., mutually contradictory binding precedents) in the decision-making process of new cases. In legal practice, the consistency requirement of the case database may not be met, thus necessitating a model to handle these inconsistencies. ### Main Contributions 1. **Model Extension**: Based on previous research, the paper combines the Boolean classifier model with legal case reasoning and further enriches these classifier models by introducing the organizational structure of courts and temporal relationships. 2. **Formalization of Precedent Concepts**: Within the extended framework, the paper formalizes the concepts of "overruled cases" and "cases decided per incuriam." These cases no longer have binding force on subsequent cases. 3. **Decision Process Conditions**: The paper explores how principles based on hierarchical structure and temporal dimensions can provide a clear decision-making process in the presence of conflicting binding precedents. ### Background and Motivation - **Application of Machine Learning in Law**: Machine learning is used to predict the outcomes of legal proceedings, but its independence and transparency are questioned. - **Need for Symbolic Methods**: To verify the robustness of machine learning algorithms, symbolic methods are needed to ensure the normative correctness and accuracy of the algorithms. - **Challenges of Precedent Constraints**: Existing models assume that the case database is consistent, i.e., there are no precedent violations. However, this assumption may not hold in actual legal case databases, thus requiring new methods to handle inconsistent precedents. ### Methods 1. **Organizational Structure**: Defines an organizational structure that describes the hierarchical relationships of courts and which courts' decisions are binding on others. 2. **Temporal Relationships**: Introduces temporal preordering to determine the chronological relationships between cases. 3. **Relevance Relationships**: Defines relevance relationships between cases to determine which precedents are relevant to new cases. 4. **Precedents and Binding Precedents**: Based on relevance relationships, defines the concepts of precedents and supporting precedents, and further defines binding precedents. 5. **Overruled Cases and Cases Decided Per Incuriam**: Formalizes the concepts of overruled cases and cases decided per incuriam, which no longer have binding force on subsequent cases. 6. **Decision Process**: Explores how to provide a clear decision-making process through principles based on hierarchical structure and temporal dimensions in the presence of conflicting binding precedents. ### Conclusion By introducing organizational structure, temporal relationships, and relevance relationships, the paper extends the Boolean classifier model to better handle the issue of precedent conflicts in legal practice. This provides a new theoretical foundation and method for future legal case reasoning.

When Precedents Clash

Case Law Grounding: Using Precedents to Align Decision-Making for Humans and AI

What About the Precedent: An Information-Theoretic Analysis of Common Law

Toward representing interpretation in factor-based models of precedent

Precedent and the Rule of Law

Reasoning with previous decisions: beyond the doctrine of precedent

"Precedent" in the Narrative of Chinese Legal History

Incorporating Precedents for Legal Judgement Prediction on European Court of Human Rights Cases

Empirical analysis of Binding Precedent efficiency in the Brazilian Supreme Court via Similar Case Retrieval

United Kingdom, United Courts? Hierarchical Interactions and Attention to Precedent in the British Judiciary

Case Repositories: Towards Case-Based Reasoning for AI Alignment

The Formation of Precedent and the Case-guiding System in China

Precedent-Enhanced Legal Judgment Prediction with LLM and Domain-Model Collaboration

Probe on the New Perspectives and Approaches on Precedents Study

Bisimulation Between Base Argumentation and Premise-Conclusion Argumentation

On the Role of Negative Precedent in Legal Outcome Prediction

Roles and the Moral Practice of Precedent

PILOT: Legal Case Outcome Prediction with Case Law

Algorithmic Learning Foundations for Common Law

Towards Explainability in Legal Outcome Prediction Models

Judicial Pronouncements: Unravelling the Significance of Obiter Dicta and Ratio Decidendi in Legal Precedents