The University of Canterbury is known internationally for the Origins of New Zealand English (ONZE) corpus (see Gordon et al 2004). ONZE is a large collection of recordings from people born between 1851 and 1984, and it has been widely utilised for linguistic and sociolinguistic research on New Zealand English. The ONZE data is varied. The recordings from the Mobile Unit (MU) are interviews and were collected by members of the NZ Broadcasting service shortly after the Second World War, with the aim of recording stories from New Zealanders outside the main city centres. These were supplemented by interview recordings carried out mainly in the 1990s and now contained in the Intermediate Archive (IA). The final ONZE collection, the Canterbury Corpus, is a set of interviews and word-list recordings carried out by students at the University of Canterbury. Across the ONZE corpora, there are different interviewers, different interview styles and a myriad of different topics discussed. In this paper, we introduce a new corpus – the QuakeBox – where these contexts are much more consistent and comparable across speakers. The QuakeBox is a corpus which consists largely of audio and video recordings of monologues about the 2010-2011 Canterbury earthquakes. As such, it represents Canterbury speakers’ very recent ‘danger of death’ experiences (see Labov 2013). In this paper, we outline the creation and structure of the corpus, including the practical issues involved in storing the data and gaining speakers’ informed consent for their audio and video data to be included.
In this paper Paul Millar outlines the development of the University of Canterbury Quakebox project, a collaborative venture between the UC CEISMIC Canterbury Earthquakes Digital Archive and the New Zealand Institute of Language Brain and Behaviour to preserve people’s earthquake stories for the purposes of research, teaching and commemoration. The project collected over 700 stories on high definition video, and Millar is now looking at using the corpus to underpin a longitudinal study of post-quake experience.
Natural disasters are highly traumatic for those who experience them, and they can have an immense and often lasting emotional impact (Cox et al., 2008). Emotion has been studied in linguistics through its enactment in language, and this field of research has increased over the past decades. Despite this, the expression of emotion in post-disaster narratives is a largely unexplored field of research. This thesis investigates how emotion is expressed in narratives taken from the QuakeBox corpus (Walsh et al., 2013), recorded, following the Christchurch earthquakes, in 2012 and rerecorded in 2019. I take a mixed methods approach, combining computer-based emotion recognition software and discourse analytic techniques, to explore the expression of emotion at both a broad and narrow level. Two emotion recognition programs, Empath (Fast et al., 2016) and Speechbrain (Ravanelli et al., 2021), are employed to measure the levels of positive and negative emotion detected in a wide dataset of participants, which are investigated in relation to the gender and age of participants, and the temporal difference between the first and second QuakeBox recordings. In a second phase, a subset of these participants’ narratives was analysed qualitatively, exploring the co-construction of emotion and identity through a social constructionist lens and examining the societal Discourses present in the earthquake narratives. The findings highlight the relevance of gender in the expression of emotion. Female speakers have higher levels of positive emotion than non-female speakers in the findings of both emotion recognition programs, and there is a clear gendered difference in the construction of identity in the narratives, influencing the expression of emotion. The expression of emotion also appears to be mediated by New Zealand culture. Within this, a Discourse of the Christchurch earthquakes emerges, with motifs of luck, gratitude, and community, which reflects the values of the people of Christchurch at the time. Findings reinforced in both phases of the analysis also indicate differences between the lexical content and acoustic features in the emotion expressions, supporting previous research that argues that the expression of emotion, as a performative act, does not reflect the speaker’s inner state directly. This research adds a new dimension to (socio)linguistic research on emotion, as well as providing insight into how crisis survivors display emotion in their post-disaster narratives.
