On the History and Future of Sociolinguistic Data

Abstract

Recordings of natural speech play a central role in the diverse subdisciplines of linguistics. The reliance on speech recordings is especially profound in sociolinguistics, where scholars have developed a range of techniques for eliciting and analyzing natural speech. However, sociolinguists have rarely focused explicitly on the storage, management, and preservation of their data – the interfaces to their data – and this lack of focus has had consequences for the advancement of the field. In this essay, I briefly review the history of data‐management practices within sociolinguistics, insofar as these practices have been discussed in the literature. I then propose new ways to consider and approach natural speech recordings as data for sociolinguistic analysis and provide examples from the North Carolina Sociolinguistic Archive and Analysis Project, a Web‐based digitization and preservation project, to highlight the analytical as well as theoretical benefits of more rigorous considerations of ‘data’ within sociolinguistics.

Corpora from a sociolinguistic perspective

“…This trend of closed data appears to be changing and it is now the case that more groups of sociolinguistic researchers are making their data available to colleagues and to the public (cf. KENDALL, 2008;CHILDS et al 2011), but it remains the case that sociolinguistics has so far not been able to benefit from the kind of peer review only possible when datasets are widely available for review and re-analysis. 9 This has also, of course, limited the ability of other (i.e.…”

“…KENDALL, 2007KENDALL, , 2008 and the Online Speech/ Corpora Archive and Analysis Resource (OSCAAR; <http:// oscaar.ling.northwestern.edu>; cf. KENDALL, 2010) are two examples of ways that one might approach implementing this sort of model.…”

“…These variable codes are stored along with their time-stamps and users can later return directly to the moment in the audio associated to each extracted variable at the click of the mouse. See Kendall (2007Kendall ( , 2008 for more on SLAAP's variable analysis features.…”

“…Users can extract phonetic information directly from this view (only pitch data is illustrated in Figure 3). Transcripts in SLAAP are dynamic entities and can be reformatted in numerous ways, from textual representations, like the columnar format suggested by Ochs (1979), to various graphical formats (screenshots of these other transcript views and a fuller discussion of SLAAP's transcript model are available in KENDALL, 2007KENDALL, , 2008. Traditional corpus analysis features are available, such as in Figure 4, which displays the highest frequency bigrams (on the left) and a sample concordance (on the right; for the phrase "high school") from the same transcript shown in Figures 2 and 3 The connection between the audio recording and the transcript (and other annotation layers) is not the only step available towards spoken language corpora that fit the "re-conceived" model of Figure 1 LIBERMAN, 2001) provides a formal framework for the development of these kinds of interfaces to data.…”

“…In Kendall (2008), I proposed a model for considering data within sociolinguistics that attempts to maintain close connections between layers of annotation or metadata. Crucially, this involves being explicit about layers of abstraction (steps away from the original source data) in our annotation and metadata creation processes.…”