We present a mathematical approach to defeasible reasoning based on arguments. This approach integrates the notion of specificity introduced by Poole and the theory of warrant presented by Pollock. The main contribution of this paper is a precise, well-defined system which exhibits correct behavior when applied to the benchmark examples in the literature. It aims for usability rather than novelty. We prove that an order relation can be introduced among equivalence classes of arguments under the equi-specificity relation. We also prove (...) a theorem that ensures the termination of the process of finding the justified facts. Two more lemmas define a reduced search space for checking specificity. In order to implement the theoretical ideas, the language is restricted to Horn clauses for the evidential context. The language used to represent defeasible rules has been restricted in a similar way. The authors intend this work to unify the various existing approaches to argument-based defeasible reasoning. (shrink)

We discuss five kinds of representations of rationales and provide a formal account of how they can alter disputation. The formal model of disputation is derived from recent work in argument. The five kinds of rationales are compilation rationales, which can be represented without assuming domain-knowledge (such as utilities) beyond that normally required for argument. The principal thesis is that such rationales can be analyzed in a framework of argument not too different from what AI already has. The result is (...) a formal understanding of rationales, a partial taxonomy, and a foundation for computer programs that represent and reason with rationales.The five kinds of rationales are as follows: (c)ompression and (s)pecialization, which yield rules, and (d)isputation, which yields a decision. These are modeled as potentially changing the focus of the dispute. Then there are (f)it, a rationale for rules, and (r)esolution, a rationale for decisions. These cannot be modeled as simply; they force disputation to a meta-level, at least temporarily. (shrink)

Logical models of argument formalize commonsense reasoning while taking process and computation seriously. This survey discusses the main ideas which characterize di erent logical models of argument. It presents the formal features of a few main approaches to the modeling of argumentation. We trace the evolution of argumentationfrom the mid-80's, when argumentsystems emerged as an alternative to nonmonotonic formalisms based on classical logic, to the present, as argument is embedded in di erent complex systems for real-world applications, and allows more (...) formal work to be done in different areas, such as AI & Law, case-based reasoning and negotiation among intelligent agents. (shrink)

Hart's "Ascription of Responsibility and Rights" is where we find perhaps the first clear pronouncement of defeasibility and the technical introduction of the term. The paper has been criticised, disavowed, and never quite fully redeemed. Its lurid history is now being used as an excuse for dismissing the importance of defeasibility.

This paper critically evaluates the celebrated paper of Hanks and McDermott on temporal projection, non-monotonic reasoning, and the frame problem. First I argue against their intuitions, and a fortiori, against their proposed solution. Next, I suggest how the solution they desire could be obtained, were they willing to represent the problem a bit differently.

This game3 was designed to investigate protocols and strategies for resourcebounded disputation. The rules presented here correspond very closely to the problem of controlling search in an actual program. The computer program on which the game is based is LMNOP. It is a LISP system designed to produce arguments and counterarguments from a set of statutory rules and a corpus of precedents, and applied to legal and quasi-legal reasoning. LMNOP was co-designed by a researcher in AI knowledge representation and by (...) a trained computer scientist who was an editor of Washington University Law Review at the time. LMNOP is based on the idea of a non-demonstrative or defeasible rule: i.e., a rule that admits exceptions. It adopts a representational convention that supposes there is an implicit preference of more specific rules over less specific rules. In fact, it automatically adjudicates between competing arguments when one argument meets the broader criterion of being more specific than another. The convention is based on an idea origianlly presented by David Poole [85], and is embedded in a system of determining which arguments are ultimately warranted, which originally appeared in the literatures of epistemology and ethics, by Pollock [87]. This system evolves from work by the first author since 1987; the full statement of the theory is in [92]. Prakken [92] is one example of the idea’s application to the legal domain. LMNOP also draws heavily on the model of legal reasoning and analogical reasoning put forward by Edwina Rissland and Kevin Ashley [89, 90]. Similarities to their legal casebased reasoning program, HYPO, are no accident; LMNOP seeks to improve on HYPO. A description of LMNOP is forthcoming. (shrink)

Hart’s "Ascription of Responsibility and Rights" is where we find perhaps the first clear pronouncement of defeasibility and the technical introduction of the term. The paper has been criticised, disavowed, and never quite fully redeemed. Its lurid history is now being used as an excuse for dismissing the importance of defeasibility.

This note corrects a lemma in the recent paper 1] of one of the authors by rst correcting problems with Poole's rule for speci city of arguments. It also responds to the criticism of Touretzky, et al. 9].

Formal accounts of negotiation tend to invoke the strategic models of conflict which have been impressively developed by game theorists in this half-century. For two decades, however, research on artificial intelligence (AI) has produced a different formal picture of the agent and of the rational deliberations of agents. AI's models are not based simply on intensities of preference and quantities of probability. AI's models consider that agents use language in various ways, that agents use and convey knowledge, that agents plan, (...) search, focus, and argue. Agents can choose their language, apply their knowledge, change their plans, continue their search, shift their focus, and rebut another's arguments. (shrink)

Carlos Alchourron was a scholar in the old tradition, with a vast culture and a passion for knowledge. His initial research, with Eugenio Bulygin on Normative Systems ( Alchourron-Bulygin 71]), led him to the realization that legal reasoning is actually representative of a more general kind of reasoning. He subsequently concluded that classical mathematical logic was not appropiate for formalizing this ampliative and non-deterministic kind of reasoning. His line of attack shows clearly in the characteristics of the AGM system of (...) belief revision AGM 85]. The language of mathematical logic was preserved and the only big departure from that tradition is the addition of a formalism to represent changes of a theory. The key Error: Illegal entry in bfrange block in ToUnicode CMapError: Illegal entry in bfrange block in ToUnicode CMapError: Illegal entry in bfrange block in ToUnicode CMapError: Illegal entry in bfrange block in ToUnicode CMapError: Illegal entry in bfrange block in ToUnicode CMapError: Illegal entry in bfrange block in ToUnicode CMapError: Illegal entry in bfrange block in ToUnicode CMapError: Illegal entry in bfrange block in ToUnicode CMapError: Illegal entry in bfrange block in ToUnicode CMapError: Illegal entry in bfrange block in ToUnicode CMapelement is a non-constructive choice function that provides for a selection of \worlds" (maximally consistent extensions of a theory), which allows a consistent view of the revision to be applied.. (shrink)

Dialectic is the fancy word for debate. AI contributes to the logic and processing of argument and uses ideas of argument in its models of communication; as it continues to do this, the computational study of dialectic, like the computational study of argument, is inevitable.

We provide a retrospective of 25 years of the International Conference on AI and Law, which was first held in 1987. Fifty papers have been selected from the thirteen conferences and each of them is described in a short subsection individually written by one of the 24 authors. These subsections attempt to place the paper discussed in the context of the development of AI and Law, while often offering some personal reactions and reflections. As a whole, the subsections build into (...) a history of the last quarter century of the field, and provide some insights into where it has come from, where it is now, and where it might go. (shrink)

There is the idea that rational belief for a single individual can be constructed via a process of unilateral argument. To preempt antipathy between the AI communities that can claim the idea that rational belief can be so constructed, we trace the idea to the beginning of this century, to Keynes' dispute with Russell over logic and probability. We review how Keynesian ideas were revived in AI's work on non-monotonic reasoning and parallel developments in philosophical logic.

Most of the papers in this collection are from the First International Workshop on Deontic Logic in Computer Science, DEON91, held in Amsterdam in December 1991. AI (especially AI and law, and knowledge representation) and formal system specification are the computer science communities that would seem to be most interested. In fact, this reviewer, a researcher in AI, was surprised to find common ground with a visiting researcher in distributed systems by discussing the contents of this book: he being in (...) the same field as Wieringa, and I being in the same field as Meyer. (shrink)