< Back to previous page

Project

The role of similarity measures in predicting the strength of arguments involving negative evidence.

The primary aim of the project is to look at the relation between people's judgments in category-based property induction tasks and different ways of deriving similarity measures in conceptual space. Category-based property induction involves making a judgement about whether a member of a category has a particular property given that another member of the category has that property. For instance, how likely is it that sparrows have a syrinx given that robins do? This paradigm ensures that in order to make their judgments participants primarily rely on the way individual pieces of evidence are represented in their semantic memory. Prior research has clearly shown a relation between inductive strength and the similarity between evidence and conclusion (e.g., Rips, 1975; Osherson, et al., 1990). Here we want to compare similarity measures derived from human judgments with those from semantic networks based on word associations and text corpora in predicting argument strength.
Date:1 Aug 2010 →  31 May 2011
Keywords:Similarity, Induction, Concepts, Categories, Semantic networks
Disciplines:Animal experimental and comparative psychology, Applied psychology, Human experimental psychology