Varol Akman

Sir John Lyons’s Linguistic Semantics: An Introduction (Cambridge, UK: Cambridge University Press, 1995) is a tolerable addition to the list of half a dozen or so impressive titles he has produced on linguistic subjects over the years. This book was initially planned to be a second edition of his Language, Meaning and Context (Lyons, 1981). However, in the end it turned out to be a successor and replacement. For it is, in the author’s words, a very different book compared to the 1981 volume: it is much longer, treats topics missing in the earlier volume, and is written in a different style.

In traditional linguistic accounts of context, one thinks of the immediate features of a speech situation, that is, a situation in which an expression is uttered. Thus, features such as time, location, speaker, hearer and preceding discourse are all parts of context. But context is a wider and more transcendental notion than what these accounts imply. For one thing, context is a relational concept relating social actions and their surroundings, relating social actions, relating individual actors and their surroundings, and relating the set of individual actors and their social actions to their surroundings.

Founded in 1993, ELEKTRIK: Turkish Journal of Electrical Engineering and Computer Sciences, has gradually become better known and is fast establishing itself as a research oriented publication outlet with high academic standards. In a modest attempt to advance this trend, this special issue of ELEKTRIK brings together five papers exemplifying the state of the art in artificial intelligence (AI). Written by experts, the papers are especially aimed at readers interested in gaining a better appraisal of the applications side of the AI enterprise. In all papers there is a strong emphasis on measuring the benefits of the proposed approaches in experimental contexts. The papers broadly fall into five actively researched, contemporary domains of AI: pattern recognition, genetic algorithms, fuzzy sets, intelligent design, and agents.

The notion of context arises in assorted areas of artificial intelligence (AI), including knowledge representation, natural language processing, intelligent information retrieval, etc. Although the term ‘context’ is frequently employed in descriptions, explanations, and analyses of computer programs in these areas, its meaning is frequently left to the reader’s understanding. My aim in this paper is to offer a swift review of context in AI. I will first identify the role of context in various fields of AI. I will then consider the enterprise of formalising context. The focus of my discussion will be McCarthy's (1993) proposal – a landmark study that inspired almost every logical theory of context since then – and its close variants, studied by his students. (The author is grateful to Prof. Carlo Penco for his invaluable assistance with this work.)

The papers that make up this special issue do not take idealized abstractions of context as their point of departure but rather start with the actual phenomena under study and later generalize. We agree that, more often than not, giving a formal model and providing a theory of a loaded notion – such as context – can lead to important insights. Thus, precise models of context and accompanying theories are useful. However, given the widely different fields, methodologies and worldviews within which people study, an approach that starts with the root phenomena/problems and only attempts normalization and generalization post hoc might be a more productive way to proceed.

Commonsense reasoning about the physical world, as exemplified by "Iron sinks in water" or "If a ball is dropped it gains speed," will be indispensable in future programs. We argue that to make such predictions (namely, envisioning), programs should use abstract entities (such as the gravitational field), principles (such as the principle of superposition), and laws (such as the conservation of energy) of physics for representation and reasoning. These arguments are in accord with a recent study in physics instruction where expert problem solving is related to the construction of physical representations that contain fictitious, imagined entities such as forces and momenta (Larkin 1983). We give several examples showing the power of physical representations.

In this paper we attempt to develop a problem representation technique which enables the decomposition of a problem into subproblems such that their solution in sequence constitutes a strategy for solving the problem. An important issue here is that the subproblems generated should be easier than the main problem. We propose to represent a set of problem states by a statement which is true for all the members of the set. A statement itself is just a set of atomic statements which are binary predicates on state variables. Then, the statement representing the set of goal states can be partitioned into its subsets each of which becomes a subgoal of the resulting strategy. The techniques involved in partitioning a goal into its subgoals are presented with examples.

In Strawson’s Entity and Identity, there are two essays (Chapters 11 and 12), which study the notion of context. In these essays, Strawson advances a threefold distinction regarding how context bears on the meaning of 'what is said' when a sentence is uttered. In this paper, we'll (i) review the original scheme of Strawson and summarize his improvements to his own scheme, and (ii) add our own improvements to make it even more thoroughgoing. We'll also show that unless it is elaborated with several considerations (mostly based on our work regarding context as a social construct and contextualizing as a form of social action) it cannot function as a realistic initiative towards building common sense models of how intended meaning is achieved.

This paper argues that in addition to the familiar approach using formal contexts, there is now a need in artificial intelligence to study contexts as social constructs. As a successful example of the latter approach, I draw attention to 'interpretation' (in the sense of literary theory), viz. the reconstruction of the intended meaning of a literary text that takes into account the context in which the author assumed the reader would place the text. An important contribution here comes from Wendell Harris, enumerating the seven crucial dimensions of context: knowledge of reality, knowledge of language, and the authorial, generic, collective, specific, and textual dimensions. Finally, two recent approaches to interpretation, due to Jon Barwise and Jerry Hobbs, are analyzed as useful attempts which also come to grips with the notion of context. It must be noted that there has been a considerable body of contributions connecting linguistic structure with social context. For example, anthropological linguistics, from Bronislaw Malinowski onwards, has underlined the cultural context of discourse as essential to meaning. This viewpoint became prominent with the emergence of the ethnography of speaking in anthropology. Thus, conversation analysis represents a consistent formal effort to contribute to an analysis of the nature of context. While this paper emphasizes and reviews the literary theory approach, it makes various contacts with works of the latter kind (e.g., the landmark contributions of Erving Goffman, John Gumperz, William Hanks, John Heritage, Dell Hymes, et al.) in order to deliver a more balanced and complete study of the dimensions of context.

Future robots should have common sense about the world in order to handle the problems they will encounter. A large part of this commonsense knowledge must be naive physics knowledge, since carrying out even the simplest everyday chores requires familiarity with physics laws. But how should one start codifying this knowledge? What kind of skills should be elicited from the experts (each and every one of us)? This paper will attempt to provide some hints by studying the mental models of force and motion.