Effects of addressee and communication channel on the realisation of verb-final schwa in German spontaneous speech

In spontaneous speech, German first-person singular verbs can be realised with or without final schwa, e.g. *ich male* ‘I paint’ can be pronounced as [ʔɪç ˈmaː.lə] with schwa or as [ʔɪç ˈmaːl] without schwa. Although the conventional form in written German is the variant with schwa, verb-final schwa is less frequently realised in German spontaneous speech (Kohler & Rodgers 2001). However, certain situational factors such as the presence of a non-native addressee might lead to a pronunciation closer to written language – thus leading to more frequent realisations of word-final schwa. In this corpus-based study, we employ a mixed design to compare whether participants realise schwa more or less frequently in verbal inflectional suffixes depending on two different conditions: speaking to a non-native addressee compared to a German native addressee (condition ADDRESSEE) and conversing via the video-conference tool Zoom (Zoom Video Communications, Inc., San Jose, CA) compared to a face-to-face dialogue (condition CHANNEL). Although it has been shown that speakers tend to shift into the non-native addressee register (NNAR) to accommodate their interlocutors’ (presumed) need for pronunciation closer to the prescribed norm in written language (Bradlow & Bent 2002, Campbell-Kibler 2010, Roche 1998), it remains unclear whether this effect can be found for non-native addressees with mid to high proficiency in German. Furthermore, how NNAR interacts with other registers such as the channel of communication is also unknown. Our understanding of register follows the model of Lüdeling et al. (2024), according to which register refers to a complex of situational parameters and linguistic phenomena. In a previous study (Lange et al. 2024), we examined the influence of the parameter SITUATION on the phenomenon of schwa realisation by comparing the variants task-solving and free dialogue. In the current follow-up study we focus on the parameters ADDRESSEE and CHANNEL. We will take into account the interaction of situational factors as well as the possibility that a linguistic phenomenon can itself be a situational parameter. For instance, if the interlocutor does not produce schwas themselves, does the speaker then align with this? We use version 2 (to be published) of the Corpus of Non-Native Addressee Register (CoNNAR, v1: Lüdeling et al. 2023).