Collegium Omnium Mentium

The College of All Minds

Manent et numerantur

Zipf’s law in cities, words, and firms

PAPER IX — SOCIOLOGY · QUESTION 29 OF 31

In plain terms: The free standard survey — one law, many machines that manufacture it.

The question as set

City sizes, word frequencies, and firm sizes obey Zipf’s law. State the regularity precisely, give two generative mechanisms that produce it, and say whether their multiplicity undermines the demand for a single explanation.

The question in context

This is question 29 of the 31 set in Paper IX — Sociology of The Truly Hardest Exam in the World, the examination of The College of All Minds. In the College’s subject index it is filed under Cities and settlement and Complexity and emergence and Language and notation. Like every question of the College, it carries three references of record, verified live at publication and described twice — technically, and in plain terms — so the ground can be judged before it is walked. The response required is an argument, not a survey, of at most three thousand words, as technical as the question demands and no more.

The references of record

Power laws, Pareto distributions and Zipf's law M. E. J. Newman · 2005 · Contemporary Physics (arXiv)

The regularity exact and the mechanisms plural: preferential attachment, multiplicative growth, criticality — each yielding the same tail.

In plain terms: The free standard survey — one law, many machines that manufacture it.

M. E. J. Newman (Wikipedia)

Power laws in economics: an introduction X. Gabaix · 2016 · Journal of Economic Perspectives (open access)

The economic instances — cities, firms — with random growth plus friction as the generative account, and its unification claim.

In plain terms: The free survey by the economist who explained the cities — and argues the many mechanisms share one skeleton.

X. Gabaix (Wikipedia)

Power-law distributions in empirical data A. Clauset, C. R. Shalizi, M. E. J. Newman · 2009 · SIAM Review (arXiv)

The statement's precision enforced: estimation and testing that separate true power laws from look-alikes — discipline for the explananda.

In plain terms: The free methodological classic ensuring the law is stated, and verified, before it is explained.

A. Clauset; C. R. Shalizi; M. E. J. Newman (Wikipedia)

Background: Zipf's law · Power law

Download Paper IX (Markdown) · Subject index