Herdan-Heaps Law Corresponds to Lotka's Law: A Dynamic Perspective from Simon's Model. Issue 2 (2nd April 2016)

• Record Type: Journal Article
• Authors: Zhang, Yifei; Shan, Shi; Yan, Chunning; Qiu, Junping; Xiong, Li
• Is Part Of: Journal of quantitative linguistics. Volume 23:Issue 2(2016)
• Page Start: 211
• Page End: 227
• Publication Date: 2016-04-02
• Subjects: Linguistics -- Statistical methods -- Periodicals; 410.151
• DOI: 10.1080/09296174.2016.1142329
• Languages: English
• ISSNs: 0929-6174

Abstract

Abstract: Herdan-Heaps law and Lotka's law are two important laws in linguistics and many other fields, which are often found to coexist in many languages. Herdan-Heaps law describes the type-token relation between number of distinct words and text length. Lotka's law concerns the fraction of words with a given number of word occurrences. Utilising a variant of the Simon model, this work demonstrates that if the growth rate of different words follow Herdan-Heaps law, with an exponent in the interval (0, 1), then the exponents of Lotka's law and Herdan-Heaps law are identical. A biparameter power law distribution, i.e. the Waring distribution, is derived within the framework of Simon's model. Estimators of the Waring distribution parameters are determined and numerical illustrations are provided.