Culture trumps biology in language development, study argues

Share via

Are the rules of language encoded in our genes, or are they primarily shaped by the speaker’s cultural context?

Leading linguistic thinkers have argued that our brains are hard-wired for languages to follow certain sets of rules. But a team of scientists is challenging that premise in a study published online Wednesday in the journal Nature.

The team used biological tools to construct evolutionary trees for four language families and found that each of the families followed its own idiosyncratic structural rules, a sign that humans’ language choices are driven by culture rather than innate preferences.


The authors say their findings run contrary to the idea of Noam Chomsky’s generative grammar, which says the brain has hard and fast ordering rules for language. They also contradict the “universal rules” of Joseph H. Greenberg, who said languages tended to choose certain patterns over others.

“Culture trumps the innate structure of the human mind,” said study coauthor Russell Gray, an evolutionary psychologist at the University of Auckland in New Zealand. “We need to take much more seriously the role of cultural factors in changing language diversity.”

But many linguists challenged the study’s conclusions and said that, in any event, they did not contradict Greenberg’s ideas.

About 7,000 languages are spoken today, each with a unique blend of sounds, words and structure. Some languages place the verb near the beginning of a sentence, while others stick it at the end. Some have genitives (like the possessive in “Mary’s dog”); others do not.

For decades, linguists have studied the diverse structures of languages with the idea that there are underlying principles all languages follow — principles that many presumed were programmed into our brains.

Chomsky argued that a small set of hard-wired rules determined the underlying structure of all languages. Greenberg statistically analyzed a multitude of languages and identified a universal list of ordering relationships within them. For example, if a language puts verbs before objects (as in the English “eat bagels”), it also will put prepositions before nouns (“in school”). But if it puts verbs after objects, it will have postpositions instead (as Hindi does).


To test whether languages followed such rules as they evolved, Gray’s team built four family trees. One was based on 82 Indo-European languages (including Spanish, English and Hindi), another was based on 400 Austronesian (including Indonesian and Hawaiian), a third analyzed 34 Uto-Aztecan (including Hopi and Yaqui) and a fourth included 73 Bantu (as well as two Bantoid) languages.

Theoretically, if a language changed its verb ordering, for instance, its preposition ordering would switch too. But the data showed that this didn’t happen consistently — some languages would change one rule without changing the other.

Gray’s team also looked at eight of Greenberg’s ordering categories (verb-object, adjective-noun and others) and mapped how they were related within each language family. If Greenberg had been right, they said, the four language families would share similar relationships between these categories.

Instead, the researchers found that while there were strong correlations within families, there were few similarities across them.

“Rather than being constrained by our psychology, they’re constrained by the local features of the language,” Gray said.

However, many linguists criticized the study for misunderstanding the foundational linguistics theories it claimed to contradict.


According to Greenberg’s model, languages change gradually, so it’s not necessarily surprising that the object-verb order could change without an accompanying change in preposition-noun order, said William Croft, a linguist at the University of New Mexico in Albuquerque. “You’d expect one order to change first, then the other,” he said.

“There are many logically possible word orders that languages could choose, and don’t,” UC Davis psychologist and typologist John Hawkins said. And since the study doesn’t consider other factors that contribute to language development, he said, it “leaves you with a somewhat unsatisfactory taste.”