Complex Predicates in Languages: Emergence, Typology, Evolution – COMPLETE

1. Creation of a typological questionnaire
A typological questionnaire will be drafted by the four project leaders, with feedback from the post-docs, docs, and participants, and filled in by all members of the consortium. Its development will need extensive preliminary checking on data through exchange between the PIs and with participants and partners.
The primary unit of observation in the questionnaire will be individual constructions in individual languages concerning the semantic, morphosyntactic, and lexical aspects be developed into a questionnaire, and discussed with other participants of the project.
1. Online database and statistical analysis
The project’s recruited positions (postdocs, PhD students) will contribute to creating an online database on Verbal Complex Predicates, in close cooperation with our engineers; they will fill it with their own data, whether first-hand or second-hand. Once that database is setup and tested through, language experts will fill it with their own data.
That database will in turn be mined for typological generalizations and statistical analyses:
› In order to check how much the four types of complex predicates form “natural classes”, we will use NeighborNet to calculate a distance matrix showing the clustering of construc¬tions, and of their morphosyntactic properties.
› The typologically informed taxonomy of lexical verbs and on the independent confirmation of typological generalizations can start out from clusterings in NeighborNet, followed by other methods currently used by Bisang & Malchukov in their grammaticalization project (e.g. correlation analysis, hierarchical cluster analysis, random forest analysis). The final decision on these methods can only be taken in the 3rd year on the basis of our data, in cooperation with the French postdoc.
› The studies on grammaticalization and lexicalization will be based on a list of grammatical erbs (grammaticalization) and of lexical domains, extracted from the questionnaire.
› Questions on areal vs. phylogenetic biases will be visualized by plotting the results from the preceding section onto WALS-style maps (Haspelmath et al. 2005; Dryer & Haspelmath 2013)..

During the first eighteen months of the project, the team achieved all the planned objectives on schedule.
The project leaders (PI), with the participation of the doctoral and post-doctoral students recruited, developed a typological questionnaire on verbal complex predicates during distance meetings, taking up the theoretical points of interest for each construction provided. This questionnaire was submitted to and discussed with all the project participants at the start-up conference held at the University of Mainz from 14 to 16 November 2022. It was then amended on the basis of these discussions during distance meetings between the PIs, doctoral students and post-doctoral students, and distributed to all participants in April 2023. This questionnaire served as a basis for two PIs to draft two model chapters, which were distributed to participants in August 2023. Together, they are intended to serve as a guide for participants in writing their articles for the collective volume planned for the end of the project.
The November 2022 conference also enabled 25 participants to present and discuss an initial analysis of verbal complex predicates in their specialist language(s), to discuss two theoretical problems on the subject (the hierarchy of inter-clausal relations and comparative concepts in typology) with the project's experts, and to present and discuss the database model.
The questionnaire was converted into a web interface (HTML/Javascript) using a MySQL database. The programming language is Python combined with Flask Framework, WTForms, Blueprints and Jinja2 Templating. The database is in the final stages of testing and will be open to all participants as planned at the beginning of next year. The web interface allows users to enter a new language, constructions, examples, grammaticalization and lexicalization paths and to suggest new ones, as well as to check and modify all the data. When adding a new construction, conditional selections are possible. The ComPLETE database is accompanied by a detailed manual (discussed together and written by a post-doc) containing examples and instructions on how to answer each question.

The participants' articles will be submitted to the IPs between 1 December 2023 and 1 February 2024 and will first be proofread internally by the IPs during the first half of 2024.
In 2024, all participants will be able to complete the BDD.
In the second half of 2024, the core team works on quantitative assessment of the database; the French postdoc selects and starts running relevant methods of statistical analysis.
The second conference with approx. 30 participants & partners is due early 2025. They will be asked to submit a chapter before attending the conference. The core team writes the position paper.
The Handbook of Complex predicates is submitted. Work on the database is completed; the online database is made public.
In addition to the planned deliverables and milestones mentioned in B, the doctoral and post-doctoral students of the project have submitted a proposal for a workshop on complex predicates for the SLE conference to be held in Helsinki from 21 to 24 August 2024 (https://societaslinguistica.eu/sle2024/wp-content/uploads/sites/6/2023/10/Exploring-the-Limits-of-Complex-Predicates-1.pdf). If accepted, 21 abstracts have already been received.

- A typological questionnaire on verbal complex predicates

- A typological database for verbal complex predicates

- Book chapters
Malchukov, Andrej and Patryk Czerwinski (in press). Complex predicates in North Asian languages. In: Edward Vajda (ed.) The Languages and Linguistics of Northern Asia, WOL Vol. 10, Berlin: Mouton de Gruyter, 1065-1119. doi.org/10.1515/9783110556216-019

- Presentations at conferences
25 presentations at the Conference of the ComPLETE project, University of Mainz. November 14-16, 2022 by Aplonova, Ekaterina ; Bisang, Walter, Alexandre François, Andrej Malchukov, Martine Vanhove, Uyên-To Doan-Rabier & Frédérique Mélanie. 2022; Boyeldieu, Pascal; Butt, Miriam; Creissels, Dennis; Czerwinski, Patryk; Fedden, Sebastian; Forker, Diana; François, Alexandre; Güldemann, Tom & Lee Pratchett; Haig, Geoffrey; Haspelmath, Martin; Helmbrecht, Johannes ; Inman, David; Jacques, Guillaume; Malchukov Andrej & Patryk Czerwinski; Meyer, Ronny; Moyse-Faurie, Claire; Robert, Stéphane; Skribnik, Elena; Souag, Lameen; Treis, Yvonne; Van Valin, Robert; Vuillermet, Marine; Mithun, Marianne.
Aplonova, Ekaterina. 2023. Serial verb constructions in Ginyanga (Kwa): towards a corpus study. CALL-52, University of Leiden. August 28-30, 2023.
Krauße, Daniel. 2023. Coverbs, Converbs and Serial Verbs: Differences and Similarities. Invited talk at the Linguistic Colloquium at Mainz University, Germany.
Bisang, Walter. Oct 2023.The two components of discourse grammar and cross-linguistic variation, a com¬pari-son of converb con¬structions, switch-reference construc¬tions and serial verb constructions in discourse. Invited speaker at Workshop on “Discourse grammar and complementation” at Sun Yat-sen University, Guangzhou, PR China.
Krauße, Daniel. 2023. Non-contiguous complex predicates: A view from Wagiman. 2n? FEMIDAL workshop (Formal & Experimental Methods and In-depth Description of Australian Indigenous Languages) in Fontainebleau, France.

- Presentations in seminars
François, Alexandre. 2023. Predicados complexos e estrutura argumental em Vanuatu. Séance du séminaire “Descrição, Tipologia e Linguística Histórica: entre a Melanésia e a Amazônia” (niveau master, doctorat, post-doctorat), à l’invitation du Programa de Pós-Graduação em Linguística, Universidade de Brasília, Brésil.
Vanhove, Martine. 2023. Auxiliaries in Beja. Séminaire du LLACAN, 14/02/2023
Krauße, Daniel. 2023. The ComPLETE database of complex predicates: A view from Melanesia. Invited talk at the Séminaire du Lattice in Montrouge, France.

- Dissemination
Aplonova, Ekaterina & Pozdniakova, Viktoria. 2023. A narrative corpus of Ginyanga. In Nikitina, Tatiana, Ekaterina Aplonova, Izabela Jordanoska, Ekaterina Biteeva, Abbie Hantgan-Sonko, Guillaume Guitang, Olga Kuznetsova, Rebecca Paterson, Elena Perekhvalskaya & Lacina Silué (eds.) The SpeechReporting Corpus: Discourse Reporting in Storytelling. CNRS-LLACAN & LACITO, discoursereporting.huma-num.fr

Although the term “complex predicates” is well established in linguistics, it is still a theoretical challenge. Complex predicates are generally defined as sequences of phonologically independent words, which together behave like a single predicate; but this definition covers a broad range of constructions whose boundaries are not clear. For that reason, we focus on verb-based complex predicates (VCP), defined as grammatical constructions serving as a predicate to a single subject, and involving at least two lexical items belonging, synchronically or diachronically, to the class of verbs. This definition includes serial verbs, converbs, light verbs and auxiliaries: these share certain functional properties and sometimes follow parallel historical dynamics. The “ComPLETE” project aims to carry out a systematic analysis of complex predicates on formal and semantic grounds, from the perspectives of synchrony, diachrony and areality.
As the available data is limited to certain language families and geographic areas, ComPLETE will base its analyses on a wide range of language families, in line with the expertise of the participants. First, the project will develop a typological questionnaire for guaranteeing cross-linguistic comparability. Based on the answers provided by linguistic experts, three research questions will be addressed: (1) Can we identify cross-linguistic dependencies between the semantic domains associated with complex predicates (e.g. tense-aspect-mood, spatial orientation) and specific formal properties? (2) How can we model the diachronic evolution of these structures, in terms of grammaticalization and lexicalization? (3) How are complex predicates distributed across the world’s linguistic areas or language phyla?
The core teams from France and Germany are composed of four linguists with a long experience in linguistic fieldwork, cross-linguistic comparison and typology. Each country will recruit one PhD student and one postdoctoral researcher: in Germany, the postdoc (50%) will be responsible for coordination; in France, s/he will conduct a synthesis on a linguistic area, and bring in expertise in statistics. Together, they will collaborate with computational engineers to build a database.
The ComPLETE team will also receive contributions from 14 full-fledged participants and 14 external partners – linguists with expertise on a large number of language families, and on a wide range of theoretical issues.At the end of three years, the ComPLETE project will present a joint volume with a position paper and chapters by all contributors. Our database and analyses will be publicly accessible online through the open-source platform of Cross-Linguistic Linked Data (CLLD).