The project Verbs' constructional patterns across languages: a multi-dimensional investigation is funded through the European Union funding – NextGenerationEU – Missione 4 Istruzione e ricerca - componente 2, investimento 1.1 “Fondo per PNR e PRIN” grant n. 20223XH5XM, CUP F53D23004570006. The project aims at investigating constructional patterns of verbs across languages and is embedded in current research on verbal constructions. With this project, we aim to further develop both typological and diachronic comparisons of verb classes and their similar patterns in order to extract general coding tendencies in terms of argument structure constructions and valency patterns. We build on results reached by members of the team in the following earlier projects: the ValPaL project to which some members of our research team contributed, the PRIN 2015 project Transitivity and argument structure in flux (20159M7X5P) and the PaVeDa project.
PaVeDa – Pavia Verbs Database is an open-source relational database for investigating verb argument structure across languages (Zanchi et al. 2022), which intends to expand and enhance the Valency Patterns Leipzig (ValPaL) database (Hartmann et al. 2013) developed within the Leipzig Valency Classes Project. Funded by the DFG between 2009 and 2013, the project carried out a large-scale cross-linguistic comparison of valency classes. It was inspired by Levin (1993), a classical study of syntactic classes of verbs in English, which argued that a semantic classification of verbs can be achieved through applying syntactic diagnostics. Yet, this study was not followed up cross-linguistically, which leaves open the question of which aspects of these classifications are universal and which are language-particular. Similarly, valency dictionaries are few in number and mostly deal with European languages, thus they cannot fill the gap. To make progress in the cross-linguistic study of valency classes, the members of the Valency Classes Project (Andrej Malchukov, Bernard Comrie, Iren Hartmann, Martin Haspelmath, Bradley Taylor & Søren Wichmann) assembled a group of contributors, who collaborated on providing a consistent set of cross-linguistic data.
The online database ValPaL, edited by Iren Hartmann, Martin Haspelmath, & Bradley Taylor, contains data for 80 verb meanings from 36 languages: Ainu (Southern Hokkaido), Balinese, Bezhta, Bora, Chintang, Eastern Armenian (standard Eastern Armenian), Emai, English, Even, Evenki, German (Standard), Hokkaido Japanese, Hoocąk (Wisconsin Hoocąk), Icelandic, Italian (Standard Italian), Jakarta Indonesian, Jaminjung (Both Jaminjung and Ngaliwurru varieties), Japanese (Standard), Ket, Korean (Spoken Korean as used in and around Seoul), Mandarin Chinese (PTH), Mandinka, Mapudungun, Mitsukaido Japanese, Modern Standard Arabic, Nen, Nǀǀng, Ojibwe (Odawa), Russian, Sliammon, Sri Lanka Malay. The data stored in the ValPaL database are based on a database questionnaire for a selected sample of 80 verb meanings, which are conceived of as representative of the verbal lexicon and have been reported in the literature to show distinctive syntactic behavior both within and across languages. Apart from valency frames (or more precisely, coding frames, as manifested by flagging/case-marking and/or indexing/ agreement), the contributors provide information about major argument alternations, both uncoded alternations (such as the Dative alternation in English) and verb-coded alternations (e.g., passive).
In spite of the research carried out within the ValPaL project, no systematic comparative study on diachronic developments across languages is available. The PaVeDa project intends to expand and enhance the ValPaL database with more languages and further features and is configured to contrastively display valency patterns simultaneously in different languages. Within this project, the Pavia team cooperates with a number of international partners who provide sets of data for the new languages uploaded in the database. For the time being, the datasets from several ancient languages (Old Latin, Ancient Greek, Gothic, Old English, Classical Armenian, Old High German) and modern languages (Modern Greek) have been uploaded in the database, along with the modern languages stored in the ValPaL database. As for the additional features, an intermediate level of annotation to the original ValPal have been added, the alternation class, which categorizes language-specific alternations into four cross-linguistic types (valency re-arranging, valency augmenting, valency decreasing, argument identifying). Since alternations are primarily language-specific, this level should serve as the initial comparative tool.
While the ValPaL database does not allow for contrastive visualization of constructions across the languages it contains, developers of the PaVeDa database designed a special layer of annotation that allows generalizing over language-specific patterns, and makes them visually comparable. Work on ancient languages also brought to methodology redesign, as ancient languages can only be studied based on corpus data rather than relying on the native speakers’ knowledge. This practice brings about a usage-based methodology that we count on implementing for modern languages too, linking the data on constructional patterns to existing digitalized corpora. In the near future, we aim to further develop both typological and diachronic comparison by adding more languages, both ancient and modern, from language families already represented in the ValPal database (Indo-European and Afro-Asiatic) as well as from families that are not represented (Uralic and Turkic).
