| In all ages,time fails to dim the brilliance of debate radiating from either flea market or large-scale presidential campaign.Given this,some studies have been carried out on debating discourses,which were mainly approached from sophistry and pragmatics.However,studies on the linguistic factors affecting the performance of debates were rarely involved.Therefore,to fill the research gap,this study is dedicated to investigating the causal relationship between certain linguistic factors in argumentative discourse and the performance of debate,thereby exploring the criteria for speech assessment in debating events.The analytical framework derives from Systematic Functional Linguistics(SFL),construction grammar and computational linguistics,based on which four linguistic factors are finally involved in this study,namely,logico-semantic relations,discourse cohesion,construction and discourse richness(a newly proposed factor for synthesizing elements in computational linguistics).Since it is difficult to measure these linguistic factors directly,we resort to their sub-categories(or observed variables)shown as follows:(1)logico-semantic relations: projection,elaboration,extension,enhancement;(2)discourse cohesion: conjunction,reference,ellipsis & substitution;(3)construction: word construction,idiomatic construction,comparative construction,passive construction,ditransitive construction;(4)discourse richness: lexical density,lexical sophistication,lexical variation,syntactic structure,syntactic complexity,readability.The corpus of this study consists of 152 debating speeches from World University Debate Championship(WUDC),which are further divided into four groups based on the scores.Two cases are randomly chosen from the highest-score group and lowest-score group for a case study.To proceed the research in an orderly way,this paper mainly addresses following three core questions:(1)What are the distinctive features of each observed variable in the case randomly chosen from the highest-score group and the lowest-score one?(2)How do four factors,namely,logico-semantic relations,discourse cohesion,construction and discourse richness,relate to each other and the performance of debate?(3)What's the hierarchy of effect from high to low exerted by four factors on the performance of debate? To probe into these issues,this study,at the very beginning,employs a quantitative method to analyze the distinctive features of observed variables in the case study,and then based on above analysis qualitatively constructs the structural equation model(SEM)with the help of Mplus7 for the sake of clarifying the causal relationship.The findings of this study mainly include three aspects.First of all,in the qualitative analysis of cases,high-and low-score cases do share some similarities in the aspects of extension,conjunction,word construction and comparative construction in terms of the number of their markers.The rhetorical function of comparative construction is also revealed in both cases.Compared with low-score case,high-score case shows quite a few differences,for example,less use of markers of projection featured by high frequent company with first-person pronouns,more frequent use of third-person pronouns,more prominent phenomena of nominalization.Secondly,four factors all have a positive impact on the performance of debate but explain the variance of holistic debating score with a different proportion: 20.4%(logico-semantic relations),24%(discourse cohesion),6.5%(construction),49.8%(discourse richness).Besides,logico-semantic relation is highly related to discourse cohesion and moderately correlated to discourse richness and construction.Thirdly,the hierarchy of effect from high to low of four factors on the performance of debate is: discourse richness > discourse cohesion > logico-semantic relations > construction.These findings attempt to suggest an alternative framework of linguistic competence for debating delegates and shed light on the judgement of debating performance. |
PDF Full Text Request