Basic olfactory terms: towards a universal semantic space of odors

In order to determine whether there is a “universal” olfactory-semantic space, this project investigates semantic differences and similarities between vocabularies used to describe odors in culturally diverse languages.

The semantic organization of olfactory vocabularies is derived with a novel, data-driven method that uses natural texts (corpora), on the one hand, and more traditional, lexical-typological methods, on the other.

In part 1 of the project, our corpus-based method is used to derive the olfactory vocabularies of English, Swedish, Italian and Thai.

The method automatically identifies odor expressions based on their olfactory association (quantified from their corpus distributions), and then derive the semantic space of those expressions using distributional-semantic word embedding models.

Part 2 consists of a lexical-typological study of the olfactory vocabularies of a sample of linguistically diverse languages with well-developed, dedicated olfactory vocabularies. Here, elicitation-based methods such as exemplar listings and similarity judgments are used.

Part 3 consolidates the findings from the corpus-based and lexical-typological studies, in order to draw more general conclusions about how languages map out the perceptual space of odors and whether olfactory vocabularies differ in culturally relevant ways.

This work will show whether there is “universal” olfactory-semantic space, and will thus provide a key theoretical contribution to the field of linguistics.