Disambiguation of Particles

The importance of particles elements can be seen from the fact that every analysis of particle has talked about the complexities that one has to face while dealing with particles interpretation and classification. Particles are interesting elements from different points of view and demand serious attention to determine their forms and functions as well as their significant role in the grammar of a natural language. Linguistic investigations on the particle elements have been done with different objectives as well as within different theoretical frameworks. Every grammar of a natural language has some discussion on particle elements of that language particularly about their forms, distribution and functional roles. In this work, I have made an attempt to investigate the distribution and classification of particle elements in Hindi and also present an outline of the methods/strategies for their disambiguation. This has been done keeping in mind the requirements of various applications, particularly for Hindi - to - English machine translation. Therefore it becomes imperative to look at, to a great extent, also the mapping patterns of each of the particle elements in Hindi-to-English (machine) translation environment. It is well-known that different interpretations of a particle element may have different mapping patterns across different translation pairs. Therefore, for the purpose of application of this knowledgebase for a machine translation system, it is important to determine the mapping patterns of each interpretation of a Hindi particle element into the target language. For the last several decades, particularly with the emergence of natural language processing and its ever expanding applications, the role of linguistic knowledgebase in the form of linguistic analysis has been felt with greater urgency. Machine translation is one of the most difficult and important among the applications of natural language processing where linguistic resources in the form of computational grammar and corpus are crucially and urgently needed. The knowledgebase for developing a machine translation system requires detailed study of languages pair for various reasons and objectives. This work can be seen as a contribution towards the creation of linguistic knowledgebase for development of machine translation system for Indian languages, particularly Hindi-to-English translation pair. The use of the analysis present here can also be extended to another important area in applied linguistics and particularly for a comparative grammar study that is the development of knowledgebase for language learning-teaching in the form of learning-teaching materials. The present work makes an extensive use of Hindi corpus obtained from different sources including online texts, for collection and classification of different particle elements according to their functional roles. The term particle is one of the terms for categorising the various parts of speech into word classes within the structural approach of linguistic analysis. Within this parameter, the categorization is based on the inflexional properties of words and consequently particles are mainly non-declinable or indeclinable elements. In this sense, particles may be adverbs, conjunctions, prepositions, interjections, sentence adverbs, negation, focus and emphatic expressions, etc. and not the word classes such as noun, verb, adjective, article, or pronoun. Some common particle expressions are modal particles, focus particles, comparative particles, answering particles, negation particles, etc. (Möllering 2001, pp 131). Some of the important research issues to particle expressions across theoretical orientations and across related languages can be presented under three broad issues: a. the roles particles play in the semantic structure of a sentence and the kind of connection they have in the syntactic structure of a language, b. disambiguation strategies and rules for particle disambiguation in different contexts, and c. significance of the study for various application areas such as machine translation and second language learning-teaching. In this work, an attempt has been made to address some of the major issues that pertain to the particle elements, particularly those particle elements that play an important role in determining the meanings / interpretations of a sentence. I focus on the research questions as listed in the above section as the general guiding parameters for the study. In Hindi, in particular and in most of the natural languages, particle elements are known for their complexity with respect to their distribution across various grammatical categories. They occur in various grammatical forms with different grammatical and pragmatic functions. They play an important role in determining the meaning of the structure where they occur. For instance, a single particle can have different meanings or interpretations in different contexts. This is one of the reasons behind a lot of difficulties and confusion in translating a text from one language to another or in deriving the exact meaning of the sentence, especially in the case of machine translation. From the semantic-pragmatic point of view, disambiguation of particle elements is significant as they often are the connecting elements in the semantic structure of a sentence. They play a major role not only in determining the meaning of a sentence but also its various interpretations. That makes it imperative to devise some ways to disambiguate their different meanings in various contexts. This complex issue has been addressed in the book by studying the distribution and classification of the particle elements and that has been used to further disambiguate their meanings and present comprehensive mapping rules for each meaning/interpretation of a particle element. Disambiguation of particle elements plays an important role in conveying the intended information with the exact interpretation. The book raises these issues and attempts to formulate strategies to disambiguate the different roles that the particle elements play. This has been done with particular emphasis on the machine translation systems with a view to improving their performance in identifying and capturing the meaning of particular particles in particular situations. As has been pointed out by Halliday (1960), this kind of study has been found useful and relevant in developing second language learning-teaching materials, too. The primary significance of the study lies in the fact that most of the particles facilitate different interpretations in different contexts and to capture the exact interpretation in a particular context is crucial for a number of applications.