Elements of Argumentation

Abstract

Background and techniques for formalizing deductive argumentation in a logic-based framework for artificial intelligence. Logic-based formalizations of argumentation, which assume a set of formulae and then lay out arguments and counterarguments that can be obtained from these formulae, have been refined in recent years in an attempt to capture more closely real-world practical argumentation. In Elements of Argumentation, Philippe Besnard and Anthony Hunter introduce techniques for formalizing deductive argumentation in artificial intelligence, emphasizing emerging formalizations for practical argumentation. Besnard and Hunter discuss how arguments can be constructed, how key intrinsic and extrinsic factors can be identified, and how these analyses can be harnessed for formalizing argumentation for use in real-world problem analysis and decision making. The book focuses on a monological approach to argumentation, in which there is a set of possibly conflicting pieces of information (each represented by a formula) that has been collated by an agent or a pool of agents. The role of argumentation is to construct a collection of arguments and counterarguments pertaining to some particular claim of interest to be used for analysis or presentation. Elements of Argumentation is the first book to elucidate and formalize key elements of deductive argumentation. It will be a valuable reference for researchers in computer science and artificial intelligence and of interest to scholars in such fields as logic, philosophy, linguistics, and cognitive science.

Graph-Based Dispute Derivations in Assumption-Based Argumentation

“…[11], we are concerned with 'efficient' arguments, but, rather than imposing minimality of support of arguments, which needs to be ascertained 'globally', we propose rule-minimality for arguments, which can be ascertained 'locally'. Our notion of rule-minimality is related to the definition of argument structure (Def.…”

“…[9][10][11][12]). This introduces both conceptual redundancy and inefficiency in standard methods of implementation, in that within an argument different rules for deriving the same conclusion may be used, potentially introducing unnecessarily points of attack and requiring additional defence efforts.…”

Bridging the Gap between Abstract Argumentation Systems and Logic

“…, E n be the extensions of (Arg(Σ), R) under a given semantics. 2 For x ∈ L , x is a conclusion of Σ iff ∃a ∈ Arg(Σ) such that Conc(a) = x and a ∈ E 1 ∩· · ·∩E n . We write Output(Σ) to denote the set of all conclusions of Σ.…”

“…It is used, for instance, for handling inconsistency in knowledge bases (e.g. [2,9]) and for decision making (e.g. [1,3]).…”

Bottom-Up Argumentation

“…Several notions of dialectical validity have been defined (e.g. see [5,8,1]) and several systems, for some or several of these notions, are available (e.g. see [10,9]).…”