A Network Based Account of Word Processing and Semantic Cognition

In this project, we propose to study word meaning using a network based on word associations.  We want to enlarge our existing word association sets in Dutch and in English, which are already the world’s largest sets in both languages, to the size of an adult’s mental lexicon.  Using these data sets, we want to derive three kinds of models for word processing and semantic cognition.  The first model is a simple (first-order) associative network in which concepts are related whenever one concept was given as a direct association to the other concept.  However, such a network is rather sparse.  Therefore, the second model elaborates on the first-order network by adding a random walk process for retrieval of semantic and lexical knowledge.  We can assume that the word association based network, even when derived from huge numbers of empirical data, is characterized by missing links between concepts.  In the third model, we want to solve this problem by deriving a second-order network, using a Bayesian approach that is based on unobserved latent concept features.  The three models will be evaluated, after applying them to the Dutch and the English data, in their predictions of:

• data from large-scale word processing tasks
• judged semantic relatedness
• the evolution of semantic networks at increasing ages