Research

The Arapesh “suitcase miracle”: The interpretive value of reproducible research

Authors:

Abstract

Recent discussions about reproducibility in linguistics emphasize that access to the original documentation allows analyses to be checked against the data that underlie them. But every annotated recording is itself always an interpretation, with some information and perspectives highlighted and some obscured. For this reason, access to the original documentation offers something better than scientific accountability: it offers a view into the underlying acts of interpretation made by other researchers and the consultants and assistants they worked with. These acts of interpretation – even problematic or “incorrect” ones – can generate insight into both linguistic structure and the social relations developed and mobilized in the context of research. This point is illustrated through analysis of legacy documentation of the Bukiyip variety of Arapesh (Papua New Guinea), which was collected fifty years ago by SIL linguist Robert Conrad and transcribed by a native speaker. Analyzing the transcribed texts against the original recordings shows the transcriptions to be problematic, with significant stretches of speech elided and “corrections” introduced by the transcriber. Yet these discrepancies do something more valuable than disconfirm the texts’ validity: they reveal linguistic patterns that would otherwise go undetected, and in some cases are only detectable now because of the discrepancies.

Keywords:

Arapeshreproducible research
  • Year: 2021
  • Volume: 21
  • Page/Article: 37-69
  • DOI: 10.25894/ldd16
  • Published on 31 Dec 2021