The Constructional Changes Of English Marginal Auxiliaries

This dissertation mainly discusses the constructional changes of three English marginal auxiliary construction subschemas on the �central modal � main verb' cline from a DCxG perspective.In contrast with traditional research on modality which is primarily based on the researchers' own introspection with decontextualized examples,this study adopts a corpus-based quantitative approach,with data extracted and semantically annotated mainly from COHA.Drawing on several multivariate statistical techniques in R software,the diachronic changes in each subschema are displayed,and the results are visualized in a linguistically interpretable way.In terms of Marginal Modal Constructions,it is found that most of its members display a preference for to-infinitive complements.In a case study of need vs.need to,after conditional inference trees and random forests analyses,the results show that after 1960 s,the frequency of to-infinitive complement is gradually increasing in both assertive and non-assertive contexts.The most important language-internal factors include clause orientation,transitivity,animacy and inversion in a descending order.As to the emergence of those �merger� cases,it is arguably due to the reanalysis of the highly entrenched adverbial use of needs,which serves as a transitional stage for the shift from modal need to lexical need to.Regarding want to,its deontic meaning of necessity with second person singular subject is hypothesized to be under the influence of ought to and need to successively as is observed in the linguistic motion chart.In addition,the author assumes that those �want + bare-inf.� cases might be attributed to the constructional contamination effect from another formally similar construction help(to)motivated by paradigmatic analogy.However,so far this usage is only limited to English-based Pidgins/Creoles in fiction,and whether it may encroach into the territory of mainstream American English still remains to be further observed.Concerning Semi-auxiliary Constructions,it is shown that both �BE adj.TO' and �BE pp.TO' are prototypically used to express deontic modality,and the global productivity of this construction schema almost remains the same since the 19 th century.Indeed,there also exist a small proportion of epistemic �BE adj./pp.TO' constructions,and they are hypothesized to be derived from the corresponding �BE adj./pp.THAT' constructions via semantic and paradigmatic analogy.In other words,the to-infinitive is prototypically used to encode deontic modality,to be distinguished from the thatclause complement functioning as a marker of epistemic modality.This claim is further supported by a case study of be supposed to from EEBO corpus data.Judging from all this,the author proposes a �prototype-based fuzzy set model of constructional contamination',and several micro-constructions which can express both epistemic and deontic modality function as �the bridges� to trigger this constructional contamination.With regard to Catenative Constructions,a case study of seem to vs.appear to is conducted.By means of semantic vector space modeling,cluster analysis and correspondence analysis,the overall trend of their semantic development is displayed respectively.In general,both of them undergo a gradual shift in collocational preferences since 1900s: seem to tends to combine with those conceptual/perceptual verbs to express the speaker/writer's subjective attitude towards real-world events,whereas appear to mainly collocates with stative verbs as an objective description and representation of the physical world.Based on all the discussions above,a tentative network of Modality Construct-icon is established,and various node-external inheritance links at different schematic levels both inside this constructional network and from other external macroconstructions are visualized.In addition,the multiple sources that may simultaneously influence the semantic and/or morphosyntactic changes of a certain micro-construction are displayed in this network-based model.Finally,the author contends that those marginal auxiliaries in English should not be marginalized;rather,they are the most splendid part in the study of English modality,and are well worth a further exploration.This is in line with the basic tenet of construction grammar,namely to pay due attention to those �idiosyncratic' and �peripheral' linguistic phenomena that are often downplayed in generative approaches.