# # #
Human languages across the world have a wide range of variations in their grammatical properties. But, is there any limit to such variations? Do languages share any significant common properties? The search for properties holding across languages, or language universals, is a fundamental goal in linguistics, as it offers a window into our shared linguistic ability as the human kind.
Meanings of single words, or lexical semantics, is one area in which the research on language universals has been highly successful. Although, at first glance, words in different languages appear to express unlimited varieties of meanings, linguistic research has revealed robust similarities in lexical semantics across languages (e.g., Barwise & Cooper 1981; Keenan & Stavi 1986).
This collaboration extends the research on lexical-semantic universals into a less cultivated domain: modal words. We take these to include modal auxiliaries (e.g., may/must), attitudinal verbs (e.g., believe/know) and evidential particles (e.g., yooda/darou in Japanese). Although cross-linguistic universals concerning the meanings of these words have rarely been investigated, ongoing theoretical developments in formal semantics provide new avenues for such investigation. For example, recent semantic theories of question sentences suggest that the semantic behavior of attitudinal words in questions is a key to discovering cross-linguistic universals about modal lexical semantics. Spector & Egré (2015) speculate that the veridicality of all natural language attitudinal verbs is uniform across their question-embedding and proposition-embedding use. More recently, several similar non-trivial language-universals attitudinal words have been suggested in the literature (e.g., Theiler, Roelofsen & Aloni 2018), and explanations for such cross-linguistic universals have been explored from computational and learning perspectives (Steinert-Threlkeld 2018).
To evaluate these theoretical proposals, it is crucial to systematically test them with a wide range of cross-linguistic data. In this project, we will entertain and evaluate hypotheses about modal lexical-semantic universals in view of language samples including Dutch, French, German, Hungarian, Japanese, Mandarin Chinese, Ngamo, Romanian, Turkish, and Zapotec. At the same time, we will investigate the theoretical underpinning of feasible semantic universals for modal meanings.