Project A02

Grammatical processing and syntactic change

PI(s): Prof. Dr. Ulrike Demske & Dr. Claudia Felser

In this project we explore the dimensions of syntactic variability from a diachronic perspective. We adopt a multi-methodological approach which combines corpus study and psycholinguistic experimentation to investigate how constraints imposed by the language processing system might affect speakers’ choices of particular syntactic variants. The extraordinary abundance of syntactic variants in Early New High German, which gradually became more constrained during the emergence of present-day Standard German, provides a suitable empirical basis for this.

The project focuses on word order variation in German infinitival complementation, and in particular on the emergence of word order restrictions and the extent to which processing factors might have played a role in this development. Our diachronic corpus data confirm that word order variability declined over time, with discontinuously linearised infinitival complements only attested very rarely in synchronous corpora. Furthermore, our data show that infinitive-embedding verbs which freely allowed for different linearization patterns no longer do so in present-day German. This is likely to be the result of a process of specialisation of word order variants, partly driven by processing (dis-)advantages.

Our experimental findings show that, unlike what might be expected from the perspective of probabilistic approaches to linguistic variation, the acceptability of different word order variants cannot be fully predicted by their corpus frequencies. Despite being rarely attested, discontinuous infinitives are still considered acceptable by present-day German speakers, suggesting that they remain part of the grammar of Standard German as ‘latent’ constructions. We further found that the most frequent word order variant, and the one that speakers are also most likely to choose in production tasks, is the variant that is easiest to process. This indicates that word order variability is at least partly constrained by processing economy principles. The decrease in occurrence of computationally more costly variants in the history of German might reflect the growing influence of spoken German on the emerging standard language.