A Quantitative Study On Part-of-speech Frequency Of Polyfunctional Words Based On British National Corpus

The part of speech of an English word can be ambiguous.For example,we can find some words in dictionary are recorded to have more than one part of speech.Such words are called polyfunctional words 'or multi-category words.Obviously,in practice,the word classes of a polyfunctional word are not equally applied.The distribution can only be observed in running texts.Therefore,the present research investigates the frequency distribution of the word classes of polyfunctional words based on British National Corpus(BNC)by quantitative method.First of all,all polyfunctional words are extracted from BNC.These polyfunctional words are classified into different groups according to polyfunctionality(that is,the number of parts of speech that a word carries).After investigating the part-of-speech frequency of polyfunctional words,we find that the frequency distributes unevenly.Among the parts of speech of a polyfunctional word,the one with highest frequency is called its dominant part of speech.Then,all polyfunctional words are classifies into ten groups(noun group verb group,adjective group etc.),in which the frequency distributions of dominant parts of speech are studied.The result indicates that verb group and adjective group are two exceptions because in them,noun is the most frequently used dominant part of speech.However,in the rest eight groups,the most frequently used dominant parts of speech are what groups themselves.Meanwhile,the Popescu-Altmann function is found to well fit the data in these ten groups.Thirdly,repeat rate and relative repeat rate are introduced to measure frequency concentration(that is,the degree that a polyfunctional word concentrates on one of its parts of speech).By observing the mean value of relative repeat rate of each polyfunctionality,we find that with polyfunctionality increasing,the frequency concentration decreases firstly and then increases.Therefore,Menzerath's law is found to capture the data perfectly.Finally,we study the frequency concentrations of dominant parts of speech in ten groups.The frequency concentrations are listed in descending order.We find that article and interjection themselves respectively are the most concentrated dominant part of speech in article group and interjection group.The result of each group and its most concentrated dominant part of speech is clearly shown in a hierarchy.Theoretically,the present research tests the previous hypotheses about parts of speech in English by quantitative methods and finds mathematical functions to fit the data extracted from British National Corpus.The functions and models become a part of regulations for parts of speech in languages.Practically,the research makes English learners clearer about the part-of-speech frequency distribution of polyfunctional words and the features of dominant parts of speech.The understanding may be helpful for their further study.