Language Documentation and Description, and What Comes After
In this talk, I reflect on the 20 years since Himmelmann’s (1998) discussion of the separation of language documentation and language description with a concentration on its relevance for comparative work on Australian languages. I briefly review the key concepts underlying Himmelmann’s dichotomy, including ‘data’ versus ‘analysis’ and ‘language’ versus ‘linguistic behavior’. I then discuss the consequences of these dichotomies for comparative work, especially the work on the Chirila lexical database of Australian languages (Bowern 2016) and the associated “grammar bootcamps”. I show that many of the precepts of the data/analysis divide apply even when the source of the linguistic data is not the linguist working with native speakers, but that a concept of “raw” versus “analyzed” data is too simple to capture either linguistic documentation or the “second generation” analysis and comparison that result. I conclude with some illustrations of how we can support language reclamation and language maintenance projects by drawing on existing language documentation.
20 Years of Language Documentation -- and before?
Language documentation has grown immensely as a branch of linguistics during the past 20 years, N. Himmelmann's seminal Linguistics paper of 1998 being regarded as the starting point of the discipline. It is especially the steady development of digital means and methods concerning the storage of audiovisual data and connecting them with textual materials that have been crucial for the success of hundreds of documentation projects world-wide. However, the basic idea of documenting languages on the basis of audiovisual recordings was not totally new in the late 1990s. In the present paper, I will give a an outline of former attempts to document spoken languages in Eurasia, starting with recordings from the Caucasus undertaken by Adolf Dirr before the 1st World War and continuing over several decades of the 20th century, with a view to the question of how to preserve, process, and use such materials in the context of linguistic documentation of today.
Mapping multilingual repertoires: A case study of a rural African region
This talk reports on the results of interdisciplinary collaborative research to document patterns of multilingualism in the Lower Fungom region of Northwest Cameroon, with a focus on efforts to visually represent and analyze the spatial distribution of multilingual knowledge. Lower Fungom is geographically compact, but linguistically quite diverse. It is associated seven to nine local languages, and its inhabitants collectively have knowledge of dozens of other languages spoken outside of the region. The average adult speaks around six distinct languages and, in some cases, also has command over multiple dialects of a single language. This research makes use of a mix of standard documentary linguistic methodologies along with data collection and analysis techniques developed within sociolinguistics, anthropology, geography, and computer science.
A key goal of the project is to develop visual representations of individual-level linguistic knowledge that intuitively present its complicated spatial distribution and move past reductive representations of language distributions that represent a language's spatial distribution via a dot on a map or by subdividing a region into discrete areas to which just one language is assigned. This talk will present results of this work via maps that integrate information on where individuals live, their social relationships, and their linguistic knowledge into a network-based representation. It will discuss the challenges involved in collecting the data needed to generate these maps as well as the design decisions that need to be considered when creating them.
This talk will also discuss how the data collected to generate these maps can be used in quantitative investigations of the relationship between spatial networks to social networks to explore questions such as how individuals' multilingual repertoires are shaped by proximity to other speakers of a given language as opposed to their kinship relations. Taken together, this research points to new ways to explore the relationship between language and space using individual-level data and modern Geographic Information System (GIS) technologies.
Figure: Representation of locations of homes of individuals who are competent in three Lower Fungom varieties: Abar, Kung, and Missong. Small circles represent sampled individuals. Grey circles represent individuals not competent in all three varieties, and orange circles represent indviduals competent in them. Colored straight lines indicate connections between individuals and the villages associated which each of these three varieties. These are overlaid onto a spatial representation of a portion of Lower Fungom including the road and footpath network. Thick lines represent motorable roads and thin lines represent footpaths. Color of lines represents difficulty of movement, with green indicating the least difficulty, red the most difficuty, and orange moderate difficulty.
Mapping Language Practices (and Language Prospects) in Nepal
I will present Geovisualization methods and tools employed by our collaborative, multi-institutional team in a larger program of language documentation in the Manang District of Nepal between 2012 to 2015. We used geo-tagging equipment to collect audio and visual recordings of three types of data: language attitudes and practices interviews, free-form narratives, and elicited vocabulary and grammatical paradigm sets, from representative speakers of four languages spoken in twenty-six Manang villages. Our team designed and developed a Web-based, interactive multimedia atlas to display data points corresponding to the speakers, links to the three types of data gathered in multimedia format, providing friendly user interface for the manipulation and spatial analysis of all the data. This atlas may be viewed here: https://mananglanguages.isg.siue.edu/.
The Manang District is a good candidate for a case study of spatiality because it has undergone rapid environmental, economic and infrastructure changes over the years, including the construction of its first motorable road and population shifts associated with this. Some Manang communities have also witnessed population movements associated with both the rise of boarding schools in remotely located Kathmandu, and also a migrant worker phenomenon that takes young adults to to other countries for long-term employment. Therefore, this study includes modified notions of spatiality (e.g., a modified Euclidean distance appropriate for the geography of Manang; access to the newly emerging motor road; proximity to the Manang District headquarters; and also, a locally popular social-psychological divide between “upper” vs. “lower” regions of Manang residency), alongside traditional social variables (age, gender, formal education, occupation).
Two of the four languages (Nar-Phu and Gyalsumdo) are critically endangered; one is viable but vulnerable (Manange), and one is viable, but with considerable contact-induced language change (Gurung). I consider the systematic investigation of attitudes and practices to be an important first step into better assessing the types and possible causes for structural variation and contact-induced language change, as well as the mechanisms behind endangerment and possible paths to preservation. In a multilingual region like Manang, where languages demonstrate different degrees of viability, there will be variation in terms of how residents view the usage and function/value of their mother tongues. This study hypothesizes that this variation in reported use and function/value is not random, and rather correlates with both social and adjusted spatial factors, which I term “social spaces.”