Building a Multilingual Taxonomy of Olfactory Terms with Timestamps

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 4030–4039 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 4030 Building a Multilingual Taxonomy of Olfactory Terms with Timestamps Stefano Menini1, Teresa Paccosi1,2, Serra Sinem Tekiro˘glu1, Sara Tonelli1 1 Fondazione Bruno Kessler – Via Sommarive 18, Trento, Italy 2 Universit`a di Trento – Corso Bettini 84, Rovereto, Italy {menini, tpaccosi, tekiroglu, satonelli}@fbk.eu Abstract Olfactory references play a crucial role in our memory and, more generally, in our experiences, since researchers have shown that smell is the sense that is most directly connected with emotions. Nevertheless, only few works in NLP have tried to capture this sensory dimension from a computational perspective. One of the main challenges is the lack of a systematic and consistent taxonomy of olfactory information, where concepts are organised also in a multi-lingual perspective. WordNet represents a valuable starting point in this direction, which can be semi-automatically extended taking advantage of Google n-grams and of existing language models. In this work we describe the process that has led to the semi-automatic development of a taxonomy for olfactory information in four languages (English, French, German and Italian), detailing the different steps and the intermediate evaluations. Along with being multi-lingual, the taxonomy also encloses temporal marks for olfactory terms thus making it a valuable resource for historical content analysis. The resource has been released and is freely available. Keywords: olfactory information, taxonomies, multilingual resources 1. Introduction In the last two decades, the attention of scholars in the humanities and social sciences has shifted away from the visual and textual dimensions to a multi- sensory perspective, following a so-called ‘sensorial revolution’ (Howes, 2006; Classen, 1999). For exam- ple, recent works by history scholars have dealt with the meaning of odours in particular places and times (Dugan, 2011), or with the role of smells in shaping identity and otherness in the past (Smith, 2006; Tullett, 2016). This turn, however, has had a limited impact on NLP research. Indeed, while the visual dimension is prevalent in texts, linguists have showed that in West- ern languages some senses such as taste and smell are less represented and are expressed with a more limited and ambiguous vocabulary than the visual one (Win- ter, 2019; Majid and Burenhult, 2014). This may ex- plain why multi-sensory studies are still a niche area in the computational linguistics community. However, being able to automatically identify how the different sensory experiences are described would be very rele- vant to accurately carry out different tasks, from emo- tion detection to metaphor identification. Furthermore, accounting for differences in sensory vocabulary across languages and over time would be of great interest for cultural studies, digital humanities and historical con- tent analysis. In this work, we present a taxonomy of olfactory- related terms that we have semi-automatically created. Our goal was to make the process rather fast, taking ad- vantage of existing resources such as WordNet (Miller, 1995) and Google n-grams, while ensuring a high- …

Content truncated to 500 words