Conference paper of Jocelyn Aznar introducing Bazabaza:Lex and its functionalities for bridging the lexicography of a language to its resources
Jocelyn Aznar gave a talk titled Bridging the Ethnographic Aspects of Oral Language Lexicography to the Resulting Language Resources at the Linguistic Data and Language Comparison in Light of the ‘Quantitative Turn’ and ‘Big Data’ workshop and symposium, held from May 7th to May 9th.
The lexicography of oral languages is inherently both a qualitative and a quantitative endeavor. It is qualitative because it requires interviewing speakers to collect new material, a “deeply ethnographic process” (Telban 1997), and quantitative because working with the lexicon involves tools and processes to manage the quantity of information. Existing lexicographic tools, from “baseline” spreadsheets to dedicated software like Toolbox (Robinson, Aumann & Bird 2007) or Fieldwork Flex (Rogers 2010), often fail to maintain the connection between the ethnographic origins of linguistic data—specifically, the speakers and the circumstances of its production—when publishing the lexicographic output (whether they are .pdf dictionaries or .csv databases). This entextualisation (Urban 1996) removes the voices of the speakers, alienating them from the output. This presents an ethical issue, particularly when lexicographers are outsiders, as it risks misrepresenting them as the sole authority on the language, even to the speakers themselves.
Addressing this issue is a core motivation behind the online platform Bazabaza:Lex. The platform offers multiple ways to maintain the connection between ethnographic practices and the lexicographic output by:
• Providing flexible metadata to describe lexical resources and items (segments, words, sequences, recordings, texts, languoids (Nordhoff & Hammarstrom 2011; Good & Cysouw 2013), projects), linking the situations of production and the relevant people, regardless of their nature: utterances, oral narratives, interviews, elicitations, or multilingual metalinguistic discourses.
• Assigning a unique, stable, and meaningful URL to every item and resource, ensuring re-users can access their associated information. This grounds the analysis made during the lexicographic process, from data collection to interpretation, contributing to making the process and the data more discoverable, accessible, citable, transparent, and reproducible (Bird & Simons 2003).
In this talk, I will introduce my methodology for documenting the Nisvai lexicon and demonstrate how it highlights the limitations of existing lexicographic tools. I will then showcase Bazabaza:Lex and explain how its flexible framework allows for the management of heterogeneous lexical resources. I will conclude by emphasizing the importance of acknowledging the diversity of actors and sources—many of whom are potentially multilingual—contributing to more ethical and transparent lexicographic methodologies.
Reducing the distance between contributors and language resources questions the rigid boundaries traditionally upheld in dictionaries and challenges the pervasive ideal of linguistic purism, which can obscure the actual multilingual nature of data and prevent researchers from accurately reflecting the realities of linguistic field research.