An empirical observation that states that, in a population consisting of many different types, the proportion belonging to the nth most common type is approximately proportional to . The law was put forward by George Kingsley Zipf (1902–1950), a Harvard linguistics professor, in the context of the frequencies with which words appear in texts. It has been found to be a good approximation in many other contexts (for example, the incidence of family names in South Korea). See also Benford's law.