Towards describing Tibetan syntax: From word segmentation to rewrite rules through a semi-automated workflow

Abstract

The first task in Tibetan Natural Language Processing is word segmentation. We present our lightweight segmentation tool that is based on lexical resources. It can be executed within InDesign and the user can update it with the manual corrections of its output. We then propose a semi-automated workflow aiming at syntactic analysis that uses utterance simplification and intonation cues to get precise information about the syntactic structure of the Tibetan language. Native speakers, even if they are non-specialists, are thus able to provide us with precise information about the structure of utterances. This will allow the scientific community to obtain resources enabling the study of Tibetan syntax. Moreover, the extra task we have included allows for the easy generation of educational materials that the informants can benefit from.

Algorithmic description of the decomposition and checking of a Classical Tibetan syllable

“…We implemented this decomposition in a very simple software 21 and tested it against all the data produced for (Hildt, 2016), based on (ཚ ་ཏན་ཞབས་ ང་།, 2003). The only two differences were two different choices for ambiguous syllables, so the main rules succeed on 100% of the ~18000 syllables.…”

“…We have tested our spell checker against the data of (Hildt, 2016) and have only found expected discrepancies, due to the treatment of syllables built on , , , and as exceptions.…”