Elements of a Social Semantics for Argumentative Dialogue
Rodger Kibble · 2004
This paper represents an initial attempt to pull together some parallel strands in argumentation theory, multi-agent systems, and philosophy of language, centering around the notion of commit- ment. A framework for argumentative dialogue is proposed which is inspired by recent work in multi-agent communication, but is also intended to be broadly applicable to human argumentation. The no- tion of is elaborated in terms of Brandom's notion of entitlement, which is itself fleshed out using Habermasian validity claims. The appeal to intersubjectively observable deontic statuses in place of mentalistic notions of belief and intention provides a social semantics in Singh's sense, and allows for a plausible account of the emergence of Walton's argumentation schemes as well as rhetorical devices such as RST's Evidence and Justify. 1 Argumentation-based communication This paper represents an initial attempt to pull together some parallel strands in argumentation theory, multi-agent systems, and philoso- phy of language. The speech act theories of Austin (1) and Searle (14) have had unanticipated applications in agent technology, in- spired initially by the demonstration in (4) that Searle's systematic analysis of speech acts such as promising, requesting, asserting in terms of preconditions and outcomes, and the beliefs and intentions of participants, could naturally be formalised in terms of planning op- erators. This has led to important developments in the field of multi- agent and human-computer communication. However, agent design in terms of notions such as belief and in- tention faces the software engineering problem that it is not generally possible to identify data structures corresponding to beliefs and in- tentions in heterogenous agents (18), let alone a of mind enabling agents to reason about other agents' beliefs. This problem has been addressed by developing alternative semantics based on in- tersubjectively observable notions of commitments (15). This devel- opment is in some ways anticipated by moves in the philosophy of language to eliminate or at least downgrade mentalistic notions in favour of social constructs, including Brandom's inferential seman- tics (2, 3) and Habermas's theory of communicative action (5, 6, 7). A goal of the work reported in this paper is to utilise some of this analysis in order to develop a more fine-grained conceptualisation of notions like commitment and challenge in the context of computa- tional modelling of argumentative dialogue. As autonomous software agents play an increasing role in elec- tronic commerce, with e-science and even e-government on the hori- zon, it will be essential for citizens, consumers and the intelligent agents themselves to be able to judge the trustworthiness and reliabil- ity of agents they encounter in the virtual world. An important recent