Predicting word concreteness and imagery

Abstract

Concreteness of words has been studied extensively in psycholinguistic literature. A number of datasets has been created with average values for perceived concreteness of words. We show that we can train a regression model on these data, using word embeddings and morphological features. We evaluate the model on 7 publicly available datasets and show that concreteness and imagery values can be predicted with high accuracy. Furthermore, we analyse typical contexts of abstract and concrete words and review the potentials of concreteness prediction for image annotation.