Formalization Of Systemic Functional Grammar

As it is widely known that Systemic Functional Grammar is for application, whichwas named as “appliable linguistics” by Halliday. His Systemic Functional Grammarhas been widely applied in various fields, such as stylistics, translation and foreignlanguage teaching. In those various fields, Halliday has paid great attention to artificialintelligence. The key problem confronted was formalization when we apply SystemicFunctional theory in natural language processing. He thinks applying in artificialintelligence is an important approach to verify Systemic Functional Grammar as well asa precious opportunity to give his grammar better play.In overseas, there are three primary phases for Systemic Functional Grammarapplying in natural language generation: a) in1950s-1960s, Halliday took the lead inapplying Scale and Category Grammar in researching linguistic principle of mechanicaltranslation and proposed a new approach, thesaurus, towards mechanical dictionary; b)in1970s, Winograd, for the first time, constructed a natural language processing system,SHRDULE, based on systemic grammar. Since then, a plenty of natural languagegeneration systems came into being, including PROTEUS, PENMAN, SLANG,COMMUNAL and so on; c) since1980s, the applying points divert to corpus studiesincluding spoken corpus and multimodal corpus; during this period, Halliday exploredtheory and approach of semantic computation combined with intelligent computation.This paper, combined with Penman, a natural language processing systemresearches on Nigel Grammar proposed by Bill Mann and Matthiessen. The maincontents include the procedure of Penman, the mechanism about how Nigel Grammarworks with its environment, and the formalizing method of systemic network and itsimprovements.Two primary approaches are introduced: a) Literature research, i.e. historydocument method. A large number of literatures about SFG and its application innatural language processing should be collected. In addition, literatures oncomputational linguistics and intelligence should be collected, too. b) Formalization.The primary parts applied are probabilistic logic and fuzzy logic. Specifically speaking,through the instance of Nigel Grammar, we’ll strive to figure out how to transform SFGinto a more formal grammar that is able to be represented on computer smoothly. So some appropriate refinements on Systemic Functional Grammar are inevitable, which isa good way to promote its own development.Through the research, we find that a) in traditional Systemic Functional research,researchers are inclined to functional parts. However, when confronted to naturallanguage processing, we must focus more on systemic grammar, because systemicgrammar is essentially a formalizing method of that theory. b) this research contributesa lot to understanding the motivation why we should study Systemic FunctionalGrammar combined with corpus study. It helps develop a system of possibility into asystem of probability.However, some limits and defects still exist. This thesis, for example, spends morespace on theoretical research which concerns more on text while many researches havemoved to multimodal state in linguistic framework, and it is a worthy way to combineSystemic Functional Grammar with Chinese, which are what we are going to do.