Languages change as they gain more speakers. When a language grows, its expressiveness expands and contracts at once: Its vocabulary expands, but its grammar becomes simpler. This paradox in language can help us understand why simple ideas take hold rapidly in large cultures, while complicated ones tend not to become mainstream.
Small, isolated languages tend to have (pdf) relatively complex grammar. Linguists think that’s because small groups form linguistic niches, with language structures specific to their community’s uses. As more people learn to the language, and its uses broaden, esoteric and difficult-to-learn grammar becomes less useful to the larger group. That causes the grammar to become simpler.
But if that’s the case, shouldn’t vocabulary get simpler, too?
Researchers developed a computer model to try to untangle this paradox. The results of their simulation, published in a recent paper, suggest that the main difference between vocabulary and grammar is how easy or difficult they are to learn. In large groups, easily learned concepts—like new vocabulary words—flourish, according to the simulation, and challenging concepts like grammatical rules do not. In smaller communities, difficult concepts spread more readily.
The simulation works by rating different concepts’ easiness-to-learn (measured by the number of times a simulated user would need to be exposed to it to adopt it as a “convention.”) This is trying to emulate the real-world fact that, while you can start using a new word correctly after hearing it just a couple times, learning how to properly conjugate a verb might require weeks of practice. The latter requires more exposure to the concept.
As the researchers increased the size of the network, these “hard” concepts made up a decreasing proportion of the conventions learned by at least 10% of the population, even though they were introduced at the same rate.
Smaller groups were better at learning tricky things like grammar, but worse at spreading the simpler stuff. (For the statistically interested, the researchers’ code is all freely available.)
This explains the vocabulary-grammar discrepancy. But it also has implications beyond language.
Large cultural groups, the research suggest, provide a boost for easy ideas while creating friction for difficult ones. A meme can take over the internet instantly, becoming a familiar part of culture for millions of people overnight. The complicated structures of bebop music, on the other hand, had to be forged in small communities. Trying to bring bebop to a large group quickly would have resulted in its simplification, akin to what happens to grammar, thus stripping it of the complexity that made it novel.
The internet has had the dual effects of connecting people to a single, massive collective culture and allowing tiny niche communities to develop. It turns out language has been doing the same thing for thousands of years.