Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

An Argumentative Explanation Framework for Generalized Reason Model with Inconsistent Precedents

Wachara Fungwacharakorn, Gauvain Bourgne, Ken Satoh

TL;DR

This work addresses the problem of reasoning under inconsistent precedents in AI and Law by generalizing the reason model to accommodate inconsistent case-bases. It extends the Derivation State Argumentation (DSA) framework to produce Derivation State Arguments (DS-arguments) tied to maximal conclusive sub-bases, with deontic conditions defined via $P_X(s)$ and $O_X(s)$ and governed by the inconsistency set $inc(\Gamma)$. Explanations are provided as admissible dispute trees, and a key result links obligations $O_X(s)$ to the unique, well-founded extension of the DS-argument framework. This yields a principled, explainable mechanism for normative decision-making when precedents conflict, enabling precise justifications of why a court is obligated to decide in a given direction. The approach has potential practical impact for transparent normative reasoning in domains with inconsistent legal precedents and lays groundwork for future refinements such as parallel explanations and dynamic hierarchies.

Abstract

Precedential constraint is one foundation of case-based reasoning in AI and Law. It generally assumes that the underlying set of precedents must be consistent. To relax this assumption, a generalized notion of the reason model has been introduced. While several argumentative explanation approaches exist for reasoning with precedents based on the traditional consistent reason model, there has been no corresponding argumentative explanation method developed for this generalized reasoning framework accommodating inconsistent precedents. To address this question, this paper examines an extension of the derivation state argumentation framework (DSA-framework) to explain the reasoning according to the generalized notion of the reason model.

An Argumentative Explanation Framework for Generalized Reason Model with Inconsistent Precedents

TL;DR

This work addresses the problem of reasoning under inconsistent precedents in AI and Law by generalizing the reason model to accommodate inconsistent case-bases. It extends the Derivation State Argumentation (DSA) framework to produce Derivation State Arguments (DS-arguments) tied to maximal conclusive sub-bases, with deontic conditions defined via and and governed by the inconsistency set . Explanations are provided as admissible dispute trees, and a key result links obligations to the unique, well-founded extension of the DS-argument framework. This yields a principled, explainable mechanism for normative decision-making when precedents conflict, enabling precise justifications of why a court is obligated to decide in a given direction. The approach has potential practical impact for transparent normative reasoning in domains with inconsistent legal precedents and lays groundwork for future refinements such as parallel explanations and dynamic hierarchies.

Abstract

Precedential constraint is one foundation of case-based reasoning in AI and Law. It generally assumes that the underlying set of precedents must be consistent. To relax this assumption, a generalized notion of the reason model has been introduced. While several argumentative explanation approaches exist for reasoning with precedents based on the traditional consistent reason model, there has been no corresponding argumentative explanation method developed for this generalized reasoning framework accommodating inconsistent precedents. To address this question, this paper examines an extension of the derivation state argumentation framework (DSA-framework) to explain the reasoning according to the generalized notion of the reason model.
Paper Structure (4 sections, 3 theorems, 2 figures)

This paper contains 4 sections, 3 theorems, 2 figures.

Table of Contents

  1. Background: The Generalized Reason Model
  2. Background: Abstract Argumentation Framework and Dispute Tree
  3. DSA-Framework for Generalized Reason Model
  4. Concluding Remark

Key Result

Theorem 1

For any fact situation $X$, case base $\Gamma$, and $s \in \{\pi,\delta\}$, $\Gamma {| \space \sim}~ O_{X}(s)$ if and only if

Figures (2)

  • Figure 1: Diagrams depicting priorities induced by $\Gamma_1$ with auxiliary cases from $X_1$
  • Figure 2: DSA-framework from Example \ref{['ex:tax']} (Arguments with solid frames are inside the extension while arguments with dashed frames are outside the extension)

Theorems & Definitions (12)

  • Example 1: fiscal domicile
  • Definition 1: conclusiveness and derivation state
  • Definition 2: maximal conclusive sub-base
  • Definition 3: DS-argument
  • Theorem 1
  • proof
  • Definition 4: DSA-framework for explaining generalized reason model
  • Theorem 2
  • proof
  • Definition 5: explanation
  • ...and 2 more