Language Comparison with CLDF

Robert Forkel (University of Passau)

15.04.2024, 14:15–15:45
Geschwister-Scholl-Platz 1 (Hauptgebäude), F 007

Linguists have been interested in language comparison for a long time. And while comparing languages is often done qualitatively rather than quantitatively, even then, language data – ideally lots of it – plays an important role, e.g. as glossed examples. In this talk I will revisit the history of language comparison in particular in the digital age, i.e. once we deal with databases rather than printed books. The focus of the talk will also be on large-scale language comparison, i.e. comparison of tens, hundreds or thousands of languages – the subfield of linguistics that has been called "Diversity Linguistics" by Haspelmath and others. This kind of comparative work has mostly been done in Typology or Historical Linguistics and the relevant types of language data are often confined – due to the sheer number of languages to cover - to typological surveys or comparative wordlists. But even such rather simple types of data pose significant problems when it comes to comparison. Ensuring comparability – i.e. the prerequisites for actual comparison – turns out to be a significant task, in particular when computers are to be enlisted for analysis – and thus the full context of a "linguist's brain" is not available. The talk will investigate how CLDF, a more recent development in the field, addresses the – also more recent – needs of fully automated or computer assisted language comparison. A sidetrack running through this history of language comparison in the digital age – the role of web-based databases – will also be highlighted, in particular looking at its viability as a longterm publication strategy – and how CLDF might address shortcomings of this model.