A family of discrete heavy‑tailed distributions produced by a stochastic growth process in which new categories (e.g., words, items) appear with a fixed probability and existing categories accrue occurrences in proportion to their current frequency (a preferential‑attachment mechanism). The model yields skewed rank‑frequency patterns often used to describe word frequencies and some bibliometric phenomena.

Principle

Sequential entry plus preferential attachment produces a right‑skewed frequency distribution whose tail exponent is determined by the rate of introduction of new categories; the process explains how early or frequent items become disproportionately common.

Demonstration

Illustrative scenario: simulate a sequence where at each step with probability p a new token enters and with probability 1−p an existing token is sampled proportional to its count; after many steps the empirical frequency histogram shows a long tail consistent with the Yule–Simon form parameterized by p.

Misapplication

Fitting a Yule–Simon form to observed heavy tails and inferring that preferential‑attachment is the causal mechanism without testing alternatives (e.g., multiplicative growth, fitness models) or without checking sequential growth assumptions.

Consequence

Provides a mechanistic candidate for observed power‑law‑like distributions in linguistics and bibliometrics and suggests modeling choices, such as including entry rates or attachment kernels, when simulating or interpreting frequency data.

Reversal

Finite‑size effects, truncation, sampling bias, or alternative generative mechanisms can produce similar heavy tails; in static or non‑sequential systems the Yule–Simon process is not a valid causal model.

Boundary

Clearly within: systems built by sequential additions with preferential reuse (e.g., word use in growing corpora). Boundary case: corpora assembled from pooled independent sources without sequential growth. Clearly outside: datasets produced by independent draws from a fixed distribution with no growth dynamics.

Semantic Tension

Mechanistic explanation (preferential attachment) ↔ Descriptive fit (power‑law as statistical model): a good fit does not prove the mechanism.

Synthesis

The Yule–Simon distribution supplies a simple, testable generative hypothesis for skewed frequency data in growing systems; it should be treated as one mechanistic model among alternatives and evaluated against process‑level evidence.