Context Resolution of Verb Particle Constructions for English to Hindi Translation

18836

Authors

nan

Abstract

. Verb Particle Constructions (VPCs) are flexible in nature and hence quite complex and challenging to handle. As a consequence, VPCs generate a lot of interest for NLP community. Despite their prevalence in English they are not handled very well, and hence often result in poor quality of translation. In this paper we investigate VPCs for English to Hindi translation. An English VPC can have different meanings in Hindi based on what its neighboring entities are. The paper focuses on finding the correct Hindi verb for an English VPC. We also discuss some rules for VPC identification, and approaches for resolving the context of a VPC for English to Hindi machine translation. Keywords: VPC, English to Hindi Machine Translation, Semantics, Separability. 1

Keywords

nan

Full Content

Context Resolution of Verb Particle Constructions for English to Hindi Translation Niladri Chatterjee, Renu Balyan Department of Mathematics, Indian Institute of Technology Delhi, India {niladri.iitd, renu17775}@gmail.com Abstract. Verb Particle Constructions (VPCs) are flexible in nature and hence quite complex and challenging to handle. As a consequence, VPCs generate a lot of interest for NLP community. Despite their prevalence in English they are not handled very well, and hence often result in poor quality of translation. In this paper we investigate VPCs for English to Hindi translation. An English VPC can have different meanings in Hindi based on what its neighboring entities are. The paper focuses on finding the correct Hindi verb for an English VPC. We also discuss some rules for VPC identification, and approaches for resolving the context of a VPC for English to Hindi machine translation. Keywords: VPC, English to Hindi Machine Translation, Semantics, Separability. 1 Introduction Quality of a translation often depends upon how correctly the verb of the source language sentence is translated into the target language. In a sentence the verb acts as the binding agent, and is therefore considered the most important component of a sentence. Consequently, verbs need to be handled in a more systematic way for a Machine Translation (MT) system. This is more pertinent with respect to translation from English to Indian languages, Hindi in particular. This is because English verbs are often polysemic, whereas Hindi has different verbs for each of the senses. For illustration, according to WordNet1, the verb “break” has 59 senses, “make” has 49 senses, “give” has 44, “get” has 36 senses and so on. Almost for each of these senses a different and specific verb exists in Hindi. Consequently, selection of an appropriate Hindi verb is very important during translation from English to Hindi (or any other target language). This can be determined by identifying the sense in which the English verb is used. But it has been observed that this is not so straightforward, as often other features like semantics of other components of the sentence have to be looked into for correct translation. English verbs can be classified as: “Single verbs” and “Phrasal verbs” (Singh, 2003). Single verbs are formed using only single words, e.g. “go”, “decide”, “take”, and “give”. Phrasal verbs are made of two or more words, like “take off”, “make up”, “do away with”, “put up with” etc. Our focus in this paper is on translation of this special category of verbs, the ‘phrasal verbs’ from English to Hindi. The motive here is to capture the correct sense and hence provide correct translation of phrasal verbs in Hindi, as their structures. In the rest of the paper we shall be using the term VPC instead of “phrasal verb”. The difficulty in handling English VPCs is often due to their flexible syntactic structure in a sentence. Consider, for example, VPCs comprising a verb followed by a preposition. This can happen in two ways: a) Intransitive verbs are followed by some prepositions e.g. …

Content truncated to 500 words

Labeled Text

. Verb Particle Constructions (VPCs) are flexible in nature and hence quite complex and challenging to handle. As a consequence, VPCs generate a lot of interest for NLP community. Despite their prevalence in English they are not handled very well, and hence often result in poor quality of translation. In this paper we investigate VPCs for English to Hindi translation. An English VPC can have different meanings in Hindi based on what its neighboring entities are. The paper focuses on finding the correct Hindi verb for an English VPC. We also discuss some rules for VPC identification, and approaches for resolving the context of a VPC for English to Hindi machine translation. Keywords: VPC, English to Hindi Machine Translation, Semantics, Separability. 1

Publication Details

Paper ID:
18836
Year:
2011
Publication:
TIPSTER
Publication Type:
conf
DOI:
nan