Sensicon: An Automatically Constructed Sensorial Lexicon

Abstract

Connecting words with senses, namely, sight, hearing, taste, smell and touch, to comprehend the sensorial information in language is a straightforward task for humans by using commonsense knowledge. With this in mind, a lexicon associating words with senses would be crucial for the computational tasks aiming at interpretation of language. However, to the best of our knowledge, there is no systematic attempt in the literature to build such a resource. In this paper, we present a sensorial lexicon that associates English words with senses. To obtain this resource, we apply a computational method based on bootstrapping and corpus statistics. The quality of the resulting lexicon is evaluated with a gold standard created via crowdsourcing. The results show that a simple classifier relying on the lexicon outperforms two baselines on a sensory classification task, both at word and sentence level, and confirm the soundness of the proposed approach for the construction of the lexicon and the usefulness of the resource for computational applications.

Sensory language across lexical categories

“…Finally, we hope to have shown on the methodological side that many interesting questions can be asked, and answered, by using already existing datasets. In our case, we used humanly generated property ratings (Lynott & Connell, 2009), noun ratings (Lynott & Connell, 2013), verb ratings (Winter, 2016a), sound and motion ratings (Medler et al, 2005), a manually annotated lexicon (Strik Lievers, 2015) and the automatically generated Sensicon (Tekiroğlu et al, 2014) to address questions about language and perception, as well as about the semantics of lexical categories more generally. While there was a lot of noise in the used data sources, applying the principle of converging evidence (Lakoff & Johnson, 1999: 79-80;Gries, Hampe, Schönefeld, 2005) through the incorporation of multiple data sources allowed us to draw confident conclusions.…”

“…Dataset 3 is the Sensicon (Tekiroğlu et al, 2014), a resource which includes 22,684 lexemes together with their degree of association with the five senses. A second list has subsequently been published by the same authors (Tekiroğlu et al, 2015), however, this list does not include verbs and thus cannot be used to address the present research hypotheses.…”

“…Finally, verbs were significantly over-represented for touch, under-represented for sight, and crucially, overrepresented for sound. on data from the Sensicon (Tekiroğlu et al, 2014) To look at all datasets together, we average the standardized residuals for each cell across the ratings, the Strik Lievers (2015) word list and the Sensicon. This…”

Understanding the Psycho-Sociological Facets of Homophily in Social Network Communities

“…Sensicon [58] is a sensorial lexicon, which is comprised of words with sense association scores pertaining to the five basic senses: sight, hearing, smell, touch, and taste. For example, when the human mind comes across the word "apple", it will automatically visualise the appearance of an apple, stimulating the eye-sight, feel the smell of the apple in the nose and the taste on the tongue.…”

FrameNet-like Annotation of Olfactory Information in Texts

“…This turn, however, has received little attention from the NLP community, probably due to a number of inherent challenges related to the intangible nature of scents and odours. The few existing works in the field have mainly focused on building resources aimed at capturing and modelling sensory vocabularies (Tekiroglu et al, 2014b) and analysing how the different senses interfere from a lexical point of view (Winter, 2019). This lack of attention may be partly explained by the fact that Western languages, which are prevalent in NLP studies, do not contain rich vocabularies for describing odorants as opposed to other senses (Majid and Burenhult, 2014).…”