Predicting the rate of novel words from word-level semantic measures in a taboo game setting.

The present study investigates novel word production rates for word meaning expressed by single words in an exploratory-confirmatory design. To capture a wide variety of semantic nuances of word meaning that influence individual processes of word creation, we quantify distributional, categorical, and psychological semantic factors of single words. These features are tested for their correlation with novel word response rates in an online study employing the taboo game paradigm, where speakers are presented with a word for which they have to express one meaning as accurately as possible in a single word. They are not allowed to use the target word itself. We first investigate a range of semantic predictors from a distributional semantic model (word2vec), a hierarchical taxonomic network model (WordNet), and psychological semantic ratings in an exploratory study. In a next step, we conduct a larger, preregistered confirmatory study to examine whether the identified effects can be replicated. Based on the results from these studies, we conclude that participants produce a higher rate of novel words for more concrete word meanings, as well as for word meanings that have lower connectivity within a network of taxonomically related words. (PsycInfo Database Record (c) 2026 APA, all rights reserved).