Frequency, Collocation, and Statistical Modeling of Lexical Items: A Case Study of Temporal Expressions in an Elderly Speaker Corpus

Frequency, Collocation, and Statistical Modeling of Lexical Items: A Case Study of Temporal Expressions in an Elderly Speaker Corpus1 王聖富 Sheng-Fu Wang 國立臺灣大學語言學研究所 Graduate Institute of Linguistics National Taiwan University sftwang0416@gmail.com 楊靜琛 Jing-Chen Yang 國立臺灣大學語言學研究所 Graduate Institute of Linguistics National Taiwan University flower75828@gmail.com 張瑜芸 Yu-Yun Chang 國立臺灣大學語言學研究所 Graduate Institute of Linguistics National Taiwan University june06029@gmail.com 劉郁文 Yu-Wen Liu 國立臺灣師範大學英語學系 Department of English2 National Taiwan Normal University Yw_L7@hotmail.com 謝舒凱 Shu-Kai Hsieh 國立臺灣大學語言學研究所 Graduate Institute of Linguistics National Taiwan University shukaihsieh@ntu.edu.tw 1 Acknowledgement: Thanks Wang Chun-Chieh, Liu Chun-Jui, Anna Lofstrand, Hsu Chan-Chia, and Liu Yu-Wen for their involvement in the construction of the corpus and the early development of this paper. 2 Graduated. 179 Abstract This study examines how different dimensions of corpus frequency data may affect the outcome of statistical modeling of lexical items. The corpus used in our analysis is an elderly speaker corpus in its early development, and the target words are temporal expressions, which might reveal how the speech produced by the elderly is organized. We conduct divisive hierarchical clustering based on two different dimensions of corpus data, namely raw frequency distribution and collocation-based vectors. Results show when different dimensions of data were used as the input, the target terms were indeed clustered in different ways. Analyses based on frequency distributions and collocational patterns are distinct from each other. Specifically, statistically-based collocational analysis produces more distinct clustering results that differentiate temporal terms more delicately than do the ones based on raw frequency. Keywords: clustering, collocation, corpus linguistics, temporal expression, gerontology 1. Introduction The study of gerontology has gained globe wide attention as the aging population becomes a grave issue in our society nowadays. Much research has noted that aging caused not only physiological changes for elderly people, but also effects on their language production [1], cognitive load [2], context processing speed [3], language performance patterns compared to younger individuals [4], etc. To research gerontology from a linguistic viewpoint, Green [5] proposed that the phenomenon of gerontology could be studied through discourse analysis. Therefore, we collect conversations from the elderly as our speech corpus, and take the corpus as input to exemplify the procedures and usage of lexical modeling. The social roles of elderly people may be embedded in the conversation when they share personal experience or judgment of the past [6] and the present. Thus, we presume that some temporal expressions might pervade as the anchoring points in the conversation-based aging corpus and might help us reveal a certain aspect of the speech behavior pattern the elderly have. Statistical modeling can serve to describe a given set of data, be it diachronic subsets, register, or lexical units. Statistical models often take the so-called “bottom-up” approach which suits most corpus linguists‟ empirical state of mind. Moreover, nice and neat visualization is often a feat in such modeling techniques, to an extent that some of the models are called “graph models” [7]. When the proper behavior of lexical units and the structure of the lexicon are applied, statistical modeling may …

Content truncated to 500 words