Relative Frequency as a Determinant of Phonetic Change

Abstract

George Kingsley Zipf philology, can test for himself the validity of the theory in those corollary problems, the full treatment of which in this discussion I have for the sake of brevity been forced to forego. I should mention at the outset that the cumulative evidence in confirmation of my theory was at times of such embarrassing exactitude that I felt compelled to substitute, wherever relevant and attainable, the results of other investigators, made for quite different purposes. Since the results thus incorporated may not be known in every case to the reader, I consistently give necessary references, and, in cases of delayed publication, add the address of library or author. Needless to say, in approaching any such problem as that of mutation in language, one is at once beset by Streitfragen in the very field of Comparative Philology itself. Since it is my desire to establish the theory as far as possible on incontestable grounds, that it may in turn throw light on unsettled issues, I have studiously avoided treating any extraneous problem the solution of which is itself still quite obscure. Moreover, the new Gestalt psychology, which promises to be a boon to philology, I have been obliged, both from its present stage of infancy and from my own ignorance, to disregard almost entirely. But I add here that criticism from no point of view will be more welcome to me than from that of the Koehler-Wertheimer Gestalt psychologists. The presentation of my thesis is intended to begin with the simpler and less complicated sound-changes, proceeding by degrees to the subtler and more delicate.' 1 I need not say that the entire investigation, submitted in partial fulfilment of the requirements for the degree of Ph.D. in Comparative Philology at Harvard University, would have been impossible without the friendly aid of others. First of all, I express my greatest thanks to Professor George H. Chase, Dean of the Graduate School of Harvard University, for his continued and strong encouragement, which alone sustained me through many hours of baffling theoretical work preliminary to the investigation of specific, linguistic changes. In gathering the great mass of statistical evidence I was helped considerably by the kindness of Professor Fuchs of the Sichsisches Stenographisches Landesamt, Dresden; Dr. Christian Johnen of Dtisseldorf; Mr.

Proximity Full-Text Search by Means of Additional Indexes with Multi-component Keys: In Pursuit of Optimal Performance

“…It will also be important to investigate how the proposed indexing structure can be used by modern ranking algorithms. The author assumes that based on Zipf's law [6], our test text collection is sufficient and acceptable for evaluating search performance. Nevertheless, to investigate ranking algorithms' behavior we plan to use collections, such as TREC GOV and GOV2, which are intended to analyze search quality.…”

“…We assume that in typical texts, words are distributed similarly, in accordance with Zipf's law [6]. Therefore, the results obtained with our text collection will be relevant to other collections.…”

“…(to, be, or): (0, 1, 2), (0, 5, 6), (4, 1, 2), and (4,5,6). Only for the first component of the key is the intermediate posting list ordered in increasing order.…”

“…Words appear in texts at different frequencies. The typical word frequency distribution is described by Zipf's law [6]. An example of words' occurrence distribution is shown in Fig.…”

An efficient algorithm for three-component key index construction

“…We showed that three component indexes can be created for relatively large values of MaxDistance (i.e., 5,7,9).…”