ABSTRACT
The primary goal of artificial intelligence (AI) is the simulation of human intelligence by computers. Certainly, there have been considerable successes in automating aspects of human intelligence – computers can now be programmed to perform expert tasks such as disease diagnosis and financial planning. However, it remains as true of modern AI as it was true of AI in its formative years that a complete computational characterisation of human intelligence is beyond the grasp of this cognitive scientific discipline (later, we will examine the views of Hilary Putnam on why he believes this to be the case). Indeed, it is mundane tasks such as natural language processing that are proving most difficult to automate. Within the automation of this processing, pragmatics occupies a unique and influential position in relation to AI. Of all the branches of linguistics, pragmatics is alone concerned with the study of language use. This emphasis on use is at the heart of any concept of intelligence, AI 1 and human – intelligence consists not in competences, but in the use of competences in the performance of tasks. AI is thus concerned to model not speakers’ linguistic competence per se, but how speakers use this competence in communication. 2 Moreover, to the extent that AI is concerned to model the pragmatic notion of language use, pragmatics is in a position to influence the development of AI models of language processing. In the next section, we consider the features that any AI model of language processing must satisfy in order to have pragmatic validity. In the same way that pragmatics can influence AI, we will argue subsequently that AI has the potential to shape and contribute insights to pragmatics. AI’s capacity to so influence pragmatics is revealing of its multidisciplinary orientation (indeed, the same multidisciplinary orientation that we have taken to characterise pragmatics throughout this book). Cawsey comments as follows on AI’s multidisciplinary orientation and the conceptual interchange that such an orientation makes possible:
AI is a fascinating subject to study as it overlaps with so many different subject areas, and not just computer science. These include psychology, philosophy and linguistics. These different subjects contribute in different ways to our understanding of how we can act and communicate intelligently and effectively. Insights from these (and other) areas help us to get computers to do tasks requiring intelligence, which in turn sheds further light on human intelligence, feeding back into these related disciplines.
(1998: 1–2)