An Open Source Punjabi Resource Grammar

Proceedings of Recent Advances in Natural Language Processing, pages 70–76, Hissar, Bulgaria, 12-14 September 2011. An Open Source Punjabi Resource Grammar Shafqat Mumtaz Virk University of Gothenburg, Sweden virk@chalmers.se Muhammad Humayoun University of Savoie, France humayoun@gmail.com Aarne Ranta University of Gothenburg, Sweden aarne@chalmers.se Abstract We describe an open source computational grammar for Punjabi; a resource-poor language. The grammar is developed in GF (Grammatical framework), which is a tool for multilingual grammar formalism. First, we explore different syntactic features of Punjabi and then we implement them in accordance with GF grammar requirements, to make Punjabi the 17th language in the GF resource grammar library. 1. Introduction Grammatical Framework (Ranta, 2004) is a special-purpose programming language for multilingual grammar applications. It can be used to write multilingual resource or application grammars (two types of grammars in GF). Multilingualism of the GF grammars is based on the principle that same grammatical categories (e.g. noun phrases and verb phrases) and syntax rules (e.g. predication) can appear in different languages (Ranta, 2009a). A collection of all such categories and rules, which are independent of any language, makes the abstract syntax of GF grammars (every GF grammar has two levels: abstract and concrete). More precisely, the abstract syntax defines semantic conditions to form abstract syntax trees. For example the rule that a common noun can be modified by an adjective is independent of any language and hence is defined in the abstract syntax, e.g.: Very big blue house fun1 AdjCN : AP → CN → CN ; However, the way this rule is implemented may vary from one language to another; as each language may have different word order and/or 1In GF code, cat and fun belong to abstract syntax. On the contrary, lincat and lin belong to concrete syntax. agreement rules. For this purpose, we have the concrete syntax, which is a set of linguistic objects (strings, inflection tables, records) providing rendering and parsing. We may have multiple parallel concrete syntaxes for one abstract syntax, which makes the GF grammars multilingual. Also, as each concrete syntax is independent from others, it becomes possible to model the rules accordingly (i.e. word order, word forms and agreement features are chosen according to language requirements). Current state-of-the-art machine translation systems such as Systran, Google Translate, etc. provide huge coverage but sacrifice precision and accuracy of translations. On the contrary, domain-specific or controlled multilingual grammar based translation systems can provide a higher translation quality, on the expense of limited coverage. In GF, such controlled grammars are called application grammars. Writing application grammars from scratch can be very expensive in terms of time, effort, expertise and money. GF provides a library called the GF resource library that can ease this task. It is a collection of linguistic oriented but general-purpose resource grammars, which try to cover the general aspects of different languages (Ranta, 2009a). Instead of writing application grammars from scratch for different domains, one may use resource grammars as libraries (Ranta, 2009b)2. This method enables to create the application …

Content truncated to 500 words