Dr Chaithra Puttaswamy (Indian Institute of Technology Kanpur)
Period of award
January 2024 - January 2027
The Leverhulme Trust
Every human language contains a myriad of intricate and interdependent grammatical patterns, but what occurs when languages change? Linguists have uncovered much about change in individual items, for instance, verbs meaning “going to” regularly develop into expressions of the future. Here we seek to push knowledge beyond individual items, to the dynamism of systems. We identify linguistic subsystems with an uncommon degree of elaboration and test whether they arise crucially from systemic, feedback mechanisms. To develop this bold new picture of language change, we sample notable languages worldwide, compare them to less-elaborated siblings, and trace fourteen centuries of Turkic texts.
Co-headed verb (CHV) systems include auxiliary, vector, and light verbs. In the average European language, there are around 8 such verbs (e.g. 5 in Russian, 10 in Italian, and 12 in English). However, some languages have a much more extensive system containing up to 24 in Tamil, 33 in Tuvan, and 38 in Murrinh-Patha. We call these Extensive Co-headed verb Systems (ECSs). We target ECSs because they are a closed grammatical subsystem with evidence of rapid and extensive growth in languages, such as the Turkic languages whose ancestor had around 5 CHVs in the 7th Century AD. Still, some of its descendants have 8 times that much. Our key objective is to research such growths and to understand the systematic change within verbal systems. To achieve our objectives we use both synchronic and diachronic approaches.
The project evaluates three hypotheses:
H1 — ECSs constitute a coherent class of synchronic system, exhibiting cross-linguistic similarities.
H2 — The diachronic rise of ECSs exhibits specific, recurrent similarities in multiple language families.
H3 — The rise of ECSs has a temporal profile characterised by rapid growth fuelled by analogical extensions.
For H1 we build an openly accessible dataset of languages with at least 20 CHVs from across the world, which will be the basis of the synchronic analysis. In this work phase, we produce a multivariate typology of ECSs that will serve as a baseline for the other work phases.
For H2 (and to support H1) we have identified three language families with a high ratio of ECSs: Turkic, Dravidian, and Non-Pama-Nyungan. Based on the variation with these families (e.g. in Turkic, Turkish has 6 CHVs, while Tuvan has 33) we will be able to analyse the internal structural differences to infer the diachrony of CHV systems while keeping other parameters of variation relatively constant due to the closeness of the languages.
Evaluating H3 requires a diachronic analysis, which will be conducted on the Turkic family. Given its fourteen centuries of well-documented written records, we will map the rise and use of CHVs both in ECSs and less complex CHV systems. We sample historical variations systematically from four lineages: Oghuz (e.g. Turkish), Karluk (e.g. Uzbek), Kipchak (e.g. Kazakh), and Bolgar (Chuvash). With a detailed map of the evolution of these related and yet different CHV systems, we can confidently verify H3 – that the rise of ECSs has a temporal profile characterised by rapid growth fuelled by analogical extensions, followed by deceleration.