1.5 megabytes of information needed before talking

New research from the University of California – Berkeley – has suggested that learning to speak isn’t something which humans are ‘hardwired’ to do, with scientists saying language acquisition from birth – 18 is ‘a remarkable feat of cognition’

In the study, published recently in the Royal Society Open Science journal, researchers calculated that, from infancy to young adulthood, learners absorb approximately 12.5 million bits of information about language – around two bits per minute – to fully acquire linguistic knowledge. If converted into binary code, the amount of information needed to master ones native language is 1.5 MB.

The findings challenge previously held assumptions that human language acquisition happens effortlessly, with senior author Steven Piantadosi saying the findings highlight ‘that children and teens are remarkable learners.’

To illustrate the nuances required to master language, the researchers broke down the process by which a language learner comes to understand what a turkey is. On learning the word “turkey” a young learner typically gathers bits of information by asking, “Is a turkey a bird? Yes, or no? Does a turkey fly? Yes, or no?” and so on, until grasping the full meaning of the word “turkey.”

Comparing the process of acquiring language to that of the basic unit of data in computing (a bit, or binary digit) researchers said that by using the standard definition of eight bits to a byte, it reveals the complexity of language learning.

“When you think about a child having to remember millions of zeroes and ones (in language), that says they must have really pretty impressive learning mechanisms.” Mr Piantadosi said.

Lead author of the study, Frank Mollica, worked with a team of researchers to gauge the amounts and different kinds of information that English speakers need to learn their native language.

Researchers arrived at their results by running various calculations about language semantics and syntax through computational models. Notably, the study found that linguistic knowledge focuses mostly on the meaning of words, as opposed to the grammar of language.

Mr Piantadosi said that a lot of the available research on language learning focuses on syntax and word order, but that this new study shows that syntax represents “a tiny piece” of language learning, and that the main barrier to language acquisition is in learning what “so many words” mean.

The focus on semantics (understanding what the words mean) over syntax (the order in which the words are presented) distinguishes humans from robots, Mr Piantadosi said.

“Our study really highlights a difference between machine learners and human learners,” he added. “Machines know what words go together and where they go in sentences, but know very little about the meaning of words.”

Pre-empting the question as to whether plurilingual people store twice as many bits of information, Mr Piantadosi said it was unlikely in the case of word meanings, many of which are shared across languages. 

“The meanings of many common nouns like ‘mother’ will be similar across languages, and so you won’t need to learn all of the bits of information about their meanings twice,” he said.

The study is available to view here.