Number, Relative Frequency, Entropy, Redundancy, Familiarity, and Concreteness of Word Senses: Ratings for 150 Serbian Polysemous Nouns

Abstract

We collected several measures of ambiguity for 150 Serbian polysemous nouns. Ambiguity measures were derived separately for dictionary senses, and the senses provided by native speakers. In a sense collection task, participants listed all senses of the given word they could think of. Collected senses were categorized in two ways by preserving fine grained semantic intuition of the speakers as much as possible, and by mapping them onto dictionary categories. In addition, we collected familiarity and concreteness ratings of each dictionary sense, and each sense provided by participants. Based on the senses provided, we calculated the number of senses, the proportion of each sense, entropy and redundancy of sense probability distribution. In order to control for the possible influence of idiosyncratic answers, all ambiguity measures were additionally corrected based on sense frequencies and familiarity ratings. Finally, participants rated word familiarity and word concreteness. The provided measures are to be applied in the research of the processing of polysemous words with a specific accent on the processing effects of meaning uncertainty and balance of sense probabilities. Additionally, they are to help understand the relation between concreteness and polysemy, the relation between semantic intuition and dictionary senses and so forth. All of the collected senses, their frequencies, familiarity and concreteness ratings, as well as lexical ambiguity measures, word familiarity and word concreteness ratings are provided in the supplementary material.