Why widely spoken languages have simpler grammar

STALIN SPOKE Russian as a second language. The Georgian dictator of the Soviet Union had a noticeable accent and is said to have mumbled his case-endings. The tale indicates two things. One is that learning new languages is hard, even with a great deal of exposure. (Stalin started learning Russian at around ten and spoke it all his adult life.) The other is that languages are more complex than they need to be. Not having mastered all Russian’s finer points didn’t keep Stalin from ruling the Soviet Union with a murderously effective iron hand.

Russian really is hard for learners, and a casual comparison might serve the conclusion that big, prestigious languages like Russian are complex. Just look, after all, at their rich, technical vocabularies, and the complex industrial societies that they serve.

But linguists who have compared languages systematically are struck by the opposite conclusion. They tend to find that “big” languages—spoken by large numbers over a big land area—are actually simpler than small, isolated ones. This is largely because linguists, unlike laypeople, focus on grammar, not vocabulary. Consider Berik, spoken in a few villages in eastern Papua. It may not have a word for “supernova”, but it drips with complex rules: a mandatory verb ending tells what time of day the action occurred, and another indicates the size of the direct object. Of course these things can be said in English, but Berik requires them. Remote societies may be materially simple; “primitive”, their languages are not.

Systematically so: a study in 2010 of thousands of tongues found that smaller languages have more Berik-style grammatical bits and pieces attached to words. By contrast, bigger ones tend to be like English or Mandarin, in which words change their form little if at all. No one knows why, but a likely culprit is the very scale and ubiquity of such widely travelled languages.

As a language spreads, more foreigners come to learn it as adults (thanks to conquest and trade, for example). Since languages are more complex than they need to be, many of those adult learners will—Stalin-style—ignore some of the niceties where they can. If those newcomers have children, the children will often learn a slightly simpler version of the language from their parents.

But a new study, conducted at the Max Planck Institute for Psycholinguistics at Nijmegen in the Netherlands, has found that it is not entirely foreigners and their sloppy ways that are to blame for languages becoming simpler. Merely being bigger was enough. The researchers, Limor Raviv, Antje Meyer and Shiri Lev-Ari, asked 12 groups of four strangers and 12 groups of eight to invent languages to describe a group of moving shapes on the screen. They were told that the goal was to rack up points for communicating successfully over 16 rounds. (They “talked” by keyboard and were forbidden to use their native language, Dutch.)

Over time both big and small groups got better at making themselves understood, but the bigger ones did so by creating more systematic languages as they interacted, with fewer idiosyncrasies. The researchers suppose that this is because the members of the larger groups had fewer interactions with each other member; this put pressure on them to come up with clear patterns. Smaller groups could afford quirkier languages, because their members got to “know” each other better.

Neither the more systematic nor the more idiosyncratic languages were “better”, given group size: the small and large groups communicated equally well. But the work provides evidence that an idiosyncratic language is best suited to a small group with rich shared history. As the language spreads, it needs to become more predictable.

Taken with previous studies, the new research offers a two-part answer to why grammar rules are built—and lost. As groups grow, the need for systematic rules becomes greater; unlearnable in-group-speak with random variation won’t do. But languages develop more rules than they need; as they are learned by foreign speakers joining the group, some of these get stripped away. This can explain why pairs of closely related languages—Tajik and Persian, Icelandic and Swedish, Frisian and English—differ in grammatical complexity. In each couple, the former language is both smaller and more isolated. Systematicity is required for growth. Lost complexity is the cost of foreigners learning your language. It is the price of success.