Quote: a word's rank in Ulysses times its frequency of occurrence is a constant; i.e. a 45 degree line on log-log paper with steps for low frequencies

Quotation Skeleton

[From M.L. Hanley's index of J. Joyce's novel Ulysses] we have found a clearcut correlation between … sense that they approximate … r x f = C in which r … [and C is a constant] … [p. 26] Clearly the curves [for r and f on log-log paper] conform with considerable … "steps" of progressively increasing size as the line … [The steps are due to integral frequencies, 1,2,3,4, ...]. … [p. 38] [For this relationship to appear, the sample size needs to approximate C * (1+1/2+1/3...+1/n) where n is the number of words. In this case, n=29,899 and 10*C is approximately the length of Ulysses, 260,430].   Google-1   Google-2

Copyright clearance needed for quotation.

Additional Titles

Quote: need the right sample size to get a hyperbolic relationship for rank vs. frequency in Zipf's law

Related Topics

Topic: words in natural languages (40 items)