Selectional preferences have a long history in both generative and computational linguistics. However, since the publication of Resnik's dissertation in 1993, a new approach has surfaced in the computational linguistics community. This new line of research combines knowledge represented in a pre‐defined semantic class hierarchy with statistical tools including information theory, statistical modeling, and Bayesian inference. These tools are used to learn selectional preferences from examples in a corpus. Instead of simple sets of semantic classes, selectional preferences are viewed as (...) probability distributions over various entities. We survey research that extends Resnik's initial work, discuss the strengths and weaknesses of each approach, and show how they together form a cohesive line of research. (shrink)

This paper presents the notion that verbal discourse is structured, in form and contents, by metaphorical reasoning. It discusses the concept of “metaphorical network” as a framework for relating the parts of a speech act to each other, since such an act seems to cohere into a meaningful text on the basis of “domains” that deliver common concepts. The basic finding of several research projects on this concept suggest that source domains allow speakers to derive sense from a verbal interaction (...) because they interconnect the topic of discussion to culturally-meaningful images and ideas. This suggests, in turn, that language is intertwined with nonverbal systems of meaning, reflecting them in the contents of verbal messages. Overall, the concept of metaphorical networks implies that human cognition is highly associative in structure. (shrink)