Key examples: opening the database door to language teachers

We are developing a comparative grammar database for three Alaskan Athabascan languages (Koyukon, Lower Tanana, and Upper Tanana). Each item (utterance) in this database is presented with an interlinear morpheme translation as well as formal and functional tags. Our goal is to make this tool useful to different audiences, including fellow linguists, but particularly to teachers and learners of Athabascan languages, where there is a dearth of mid- and higher- level pedagogical materials. Meeting this need proves to be challenging. All three languages are severely endangered, yet very little pedagogical material is available and the few extant descriptions are written in such a way as to be largely inaccessible to individuals without considerable training in linguistics. Since many teachers of Alaskan Athabascan languages not only do not have that training, but also may not themselves be fluent speakers, this poses a serious problem. Our database has the potential to become a rich source of materials for teachers and learners, but as consultation with teachers showed, it needs to include a device that could guide them to basic forms, in the way that a pedagogical grammar might do. At present, users may search on form (e.g. ‘Why’?) or function (e.g., Question:Open-ended) to find data relating to the lessons they need to plan. However, the response to the search is a set of sentential examples displaying a number of structures of vastly different complexity, without any commentary indicating their pedagogical level. This is of course the nature of a database rather than a grammar, which typically gives the searcher one or two ‘perfect’ key examples. Using wh-question formation as an example, we will illustrate a metric based on a combination of frequency, structural considerations, and social complexity establishing a level for selected database entries comparable to the key examples mentioned above. This information will be accessible to users through a ‘construction summary,’ which also provides accessibly written commentary on each key example. It will be possible to generate a list of ‘similar examples’ illustrating the same structure, and a list of ‘related constructions’ which have a similar function or form. In this way, the interface unites the reliance on a few select examples of traditional grammaticography with the data-intensive approach of a database while paying attention to the needs of language teachers, providing both information and autonomy for the creation of lesson plans