Crowdsourcing as a means of replenishing “Languages of the World” database

Makarova Elena Andreevna

The present work looks at the fourth version of the “Languages of the World” of IL RAS database. Its main distinctness is the shift from hierarchical data representation, when the features were organized in form of a binary tree, to paradigmatic. We developed a list of 124 features, each having, on average, eight possible values. For a long time the only source of information for the database was the encyclopedia of the same name. Nevertheless, as soon as all languages described in the encyclopedia are added to the database, search for alternative sources of information will become our pressing problem. The solution we suggest is to create questionnaires and conduct a crowd-sourcing project. Linguistic questionnaires are most frequently used by specialists when describing new languages. The main feature of the suggested crowdsourcing project is that not only specialists in certain languages, but also native speakers without any linguistic education will be able to take part in it. This brings us to the necessity of creating two types of questionnaires. The first type is designed for specialists and includes mainly direct questions that presuppose not only knowledge of the language, but also linguistic education. The second type is designed for native speakers who do not have a field-specific education. The second type of the questionnaire includes, as a rule, tasks to make phrases and sentences according to some given criteria and tasks to translate something. The main advantage of such crowd-sourcing project is the opportunity to work distantly with a big number of informants and to considerably replenish the database by adding new languages into it.

Keywords:database, languages of the world, paradigmatic data representation, crowdsourcing, questionnaire, native speaker