Where in Semantic Space Do Spanish Neologisms Emerge?
Proceedings of the Workshop Neology and Large Language Models
Abstract
English neologisms, or newly coined words, have previously been shown to emerge in sparser semantic neighborhoods (filling semantic gaps) and near other neologisms (in growing semantic areas). In this work, we investigate where in semantic space Spanish neologisms emerge, and whether this mirrors English neologism development. We find that Spanish neologisms, in comparison to non-neologisms, do indeed appear both nearer to other neologisms and further from non-neologisms. We additionally investigate the prevalence of loanwords from other languages through time in Spanish neologism production and manually assess the topics that appear as loanwords at four years: 1810, 1900, 1950, and 1990. Our findings show that on average, the Spanish neologisms in our dataset have fewer neighboring words in semantic space compared to non-neologisms and tend to cluster more tightly in the semantic space, indicating that patterns of neologism emergence span languages. This suggests that novel methods for neologism detection may be cross-lingually applicable, with these features serving as multilingual predictors of neologism emergence.