Every living language is always changing, one speaker-to-speaker interaction at a time. A new pronunciation, word, or grammatical form ("innovation") starts with a handful of speakers and spreads — or fades away — purely through social contact. Sociolinguists model this as a diffusion process on a social network: each node is a speaker, each edge a social tie, and each tick some speakers hear the new variant from a neighbor and decide whether to start using it themselves.
Real sound changes such as the "cot–caught" merger in North American English or "th-fronting" in British English spread through exactly this kind of social-network contagion, typically following an S-shaped adoption curve first described by linguists William Labov and Gillian Sankoff.
A 3D social network of simulated speakers shows a new word variant spreading, or dying out, purely through contact between neighbors — the same social-diffusion process that drives real sound change and vocabulary shift.
Each speaker adopts the new variant based on the fraction of neighbors already using it, amplified or dampened by a prestige parameter — producing the classic S-shaped adoption curve linguists observe in real sound changes.
Set the innovation's social prestige, contact rate, network reach, and how many speakers start out as innovators, then watch amber "new variant" speakers spread — or shrink back to blue-grey — across the network in real time.
Sociolinguists William Labov and Gillian Sankoff first documented the S-curve pattern of language change by tracking real sound shifts like the "cot–caught" merger across generations of speakers.