General Linguistics Colloquium (WiSe 2025/2026)

This presentation introduces the development of digital linguistic resources for Kartvelian languages, focusing on the integration of existing and emerging corpora for Georgian and Megrelian within a unified annotation and alignment framework. Georgian, the most widely spoken Kartvelian language, already possesses a Universal Dependencies (UD) treebank, while Megrelian remains a severely under-resourced and endangered language with no previous computational resources. To bridge this gap, we present the Megrelian Corpus (XMF), a newly developed resource based on fieldwork data collected in Samegrelo, Georgia (2022–2024). The talk discusses shared challenges of corpus creation, lemmatisation, and UD alignment in agglutinative and morphologically rich Kartvelian languages. The presented resources lay the foundation for future UD-based modeling and cross-linguistic research within the Kartvelian family.

...ABSTRACT...

...ABSTRACT...

...ABSTRACT...

...ABSTRACT...

...ABSTRACT...

This presentation examines the relationship between prosody and grammaticalization through the German modal particle doch. It investigates how prosodic weakening reflects different functional stages, focusing on conjunctional, adverbial, and modal particle uses, as well as the fixed cluster doch mal. Using naturally occurring spoken data, the analysis combines functional annotation and acoustic measurements (F0, duration, accent, boundary) to compare prosodic realizations across functions. The findings are expected to show a clear correlation between grammaticalization and prosodic reduction. This provides empirical evidence for prosody as a reliable diagnostic of grammaticalization.

...ABSTRACT...

...ABSTRACT...

...ABSTRACT...

This presentation provides an updated account of dizque in contemporary Spanish, showing how its current distribution challenges core principles of grammaticalization theory. The term, originating from the univerbation of diz que ("he/she says that..."), has traditionally been described as an adverb with epistemic or reportative meanings. However, our corpus-based analysis reveals that dizque is widely used across Latin American varieties also as an adjective, and even appears to have further developed into a prefix. Remarkably, we identify instances in which dizque functions as a verb -a usage that had disappeared from written records in the 17th century. We argue that this atypical trajectory constitutes a breach of the principle of unidirectionality, long regarded as the most basic tenet of grammaticalization.

...ABSTRACT...