Arguments and Artifacts for Dispute Resolution

In a social context cultural differences, individual interests, and partial awareness are often the causes of disputes. Alternative Dispute Resolution (ADR) is usually considered to be alternative to litigation, and can also be used to allow disputing parts to find an agreement. A dispute resolution is not an easy task and usually involves more entities including mediator or arbitrator with multiple dialogue sessions. In the paper we focus the attention on dispute resolution system in artificial society proposing a model and a technology to support the persuasive processes. The persuasion is the principal form dialogue used in an ADR system where agents exchange arguments to support their positions. The general architecture proposed to build an ADR system exploits two artifacts abstractions – Co-Argumentation Artifact and Dialogue Artifact – that provide the right abstractions to coordinate the agents during the argumentative process. The technological support for the artifacts is provided by the TuCSoN infrastructure, also exploiting a meta-programming technique in Prolog. Finally, in the paper we present a simplified example of the execution of a persuasion dialogue ground on the commitments. I. ALTERNATIVE DISPUTE RESOLUTION People develop systems and methods in order to settle conflicts in a fair way. Human societies define norm systems, infrastructure (such as court) and methods (such as trial) to achieve the dispute resolution. In a global business process scenario there is a increasing need of speed-up the processes, and to make faster the conflict resolution. The new systems have to support legal process when for instance a negotiation is broken, they have to combine mediation and legal service to avoid litigation. Alternative Dispute Resolution (ADR) is usually considered to be alternative to litigation. It also can be used as a colloquialism for allowing a dispute to drop or as an alternative to violence. ADR is generally classified into at least four subtypes: negotiation, mediation, collaborative law, and arbitration. Walker and Daniels [1] underline that legal negotiation is a part of traditional dispute resolution system rather than a component of the ADR movement. The legal negotiation directly occurs among agents that represent the disputants in a context similar to a courtroom. Arguments have a central role in the process of formalising legal system, and in the trial, too. The paper [2] contains a survey of logic in computational model on legal argument. The authors present the main architecture of legal arguments with a four layer architecture: 1) logical layer, 2) dialectical layer, 3) procedural layer, and 4) strategic layer. Disputants use arguments in order to persuade the other parts of the dispute and also the decision makers—juries, judges, clients and attorneys. In [3] the use of arguments in an ADR systems is considered, and an analysis of arguments in different contexts such as arbitration, mediation and multi-party facilitation is presented. Argumentation plays an important role in conflict resolution systems, where it drives the ADR to obtain a successful solution of the dispute. The argumentation process promotes the values of justice, equality and community that are desirable in a dispute resolution system. In an open agent society, the same issue as in human society holds: it is undesirable to resolve dispute by litigation. The development of a system for internal resolution of disputes in virtual organisations is presented by Jeremy Pitt et al in [3], which proposes a norm-government MAS and an ADR protocol specification for virtual organization exploited by intelligent agents. ADR supplies a theoretical bases for Online Dispute Resolution (ODR) as defined in [4]. ODR has the purpose to extend the ADR process, moving it towards virtual environments while providing computation and communication support. In ODR, the role of technology used to facilitate the resolution of disputes between parties is crucial. It provides a structured communication, as well as an informed environment that helps to the successful conclusion of the conflict. ODR could be seen as an instance of an ADR system, with a communication infrastructure and Artificial Intelligence (AI) techniques aiming at supporting the parties toward agreements. The reasoning and argumentation capabilities of the parties are achieved by exploiting AI methods. Walton and Godden [5] show that argument-based dialogue, in particular persuasion dialogue, contributes to the construction of effective dispute resolution system. The main type of dialogue usually considered by ADR is negotiation, which could be interpreted as a particular sort of communication for the purpose of persuasion. In argumentation theory both types of dialogues are present: persuasion dialogue and negotiation dialogue. These two types of dialogue have a different structure and different goals, and in the context of ODR systems should be managed by different procedural rules. A fundamental problem in ODR and ADR systems is that it is difficult to structure and process the information exchanged between negotiating parties. In order to resolve this problem in this work we propose to build a ADR system based on the A&A meta-model [6] with Co-Argumentation Artifact (CAA) [7] and Dialogue Artifact (DA) [8] abstractions. We aim at providing a framework for conflict resolution in an agentbased society supplying a supporting infrastructure in order to manage arguments, to retrieve information and to bargain. Our framework provides structured information based on logic tuple along with the control of dialogue processes through a mediated form of communication over a programmable infrastructure. These two features are useful in order to build MAS in a scalable and flexible architecture, and also to build ADR that supports multi-party dialogue sessions. The aim of the work is to provide a more formal (functional) connection among the two types of argumentation artifacts CAA and DA in order to support a dialogue for dispute resolution. In particular we make explicit a set of functionalities useful during the dialogue to control the relation with the argumentative commitment store. For the CAA we collect a list of operations to manage a commitment store based on the argumentation system. On the other hand for the DA we describe by operational semantics the use of the CAA operations during the dialogue. The result is a powerful architecture where it is possible to specify a dialogue grounded on the state of the commitment store enabling a partially automate dialogue execution through DA and CAA infrastructure. Using that dialogue specification the DA can automatically drive the sequence of actions based on the state of the CAA. In Section II we introduce the architecture of the framework with the definition of the specific CAA and DA; in Section III we explain the argumentation and dialogue system by introducing the new operators to describe the interaction with the commitment store; finally in Section IV we present the case study, implementing a persuasion dialogue protocol.