Abstract:

LARA (Learning and Reading Assistant; https://www.unige.ch/callector/text-content/) is a collaborative open project, initiated during Q3 2018, whose goal is to create resources, via crowdsourcing techniques, that help people read L2 texts in foreign/archaic languages. It does this by providing tools that make it easy to transform plain text documents into hypertext versions that give non-native readers various kinds of help. The central idea is that each student receives a personalised version of the text they are reading, marked up so that they can compare all occurrences of any word in their own previous reading progress. Concretely, the screen is divided into two halves, with the text on the left. When the student clicks on a word, the right-hand side shows the personalised concordance for that word. The platform also offers conventional support, including optionally linking words and sentences to translations and human audio, which is recorded using an efficient online tool. Many examples of LARA content are available online at https://www.unige.ch/callector/lara-content/, including material already being used by classroom teachers in Australia, Iceland and Iran. The platform is fully described in the online documentation (Rayner et al 2019) and in papers linked to from the site.

In this paper, we describe recent work in which we have extended LARA in a new direction. Although the platform was originally constructed with pure L2 texts in mind, we have found that it is easy to adapt it so that it can also be used for linguistics papers and language textbooks; these are typically a mixture of L2 and L1 text, with examples in L2 and explanations in L1. For this purpose, we have introduced a simple extension to the LARA markup formalism, which is available in two flavours, depending on whether the L1 or the L2 is the default. Our presentation will be organised around two paradigmatic examples, both available online. The first is a LARA version of "Barngarlidhi Manoo" (Zuckermann 2019), an 80 page children's primer for the South Australian aboriginal language Barngarla. The LARA version's organisation is closely based on that of the original: the text is divided into pages with a table of contents, the same pictures appear, and the general layout is approximately the same. In contrast to the original, it is however possible to listen to any Barngarla word or phrase simply by hovering the mouse over it, or click on any word and get a list of all the places in the book where the word occurs. The second example is a section from a linguistics book (Zuckermann in press): the theme of the passage is variant pronunciations in Israeli Hebrew and related languages, and it is again possible to listen to any word by hovering over it with the mouse.

The LARA tools are open source Python and freely available for download.

References:
[1] M. Rayner, H. Habibi and M. Butterweck. 2019. Constructing LARA Content. https://www.issco.unige.ch/en/research/projects/callector/LARADoc/build/html/index.html
[2] G. Zuckermann. 2019. Barngarlidhi Manoo. LARA version: https://www.issco.unige.ch/en/research/projects/callector/barngarla_alphabetvocabpages/_hyperlinked_text_.html
[3] G. Zuckermann. (in press, due January 2020). Revivalistics: From the Genesis of Israeli to Language Reclamation in Australia and Beyond. New York: Oxford University Press. LARA version: https://www.issco.unige.ch/en/research/projects/callector/revivalisticsvocabpages/_hyperlinked_text_.html

Keywords:

CALL, reading, multimedia, open source, textbooks, indigenous languages.