Linguistics

“Dad, may you passed me the beans?” a young child might ask.

“No, it’s ‘Dad, may you pass me the beans?'”

“May you please passed me the beans?”

As you can see from this exchange, the child was definitely struggling to get a grasp on the proper tense to explain their situation. Yet, remarkably enough, by the time that child enters school, there is an extremely high probability that the child will have grasped enough rules from the language to be as fluent as an adult speaking the same language. How could this be? 

Since the days of Noam Chomsky’s review of famous psychologist B.F. Skinner’s treatise Verbal Behavior, linguists have generally held the view that children have an innate ability to learn their native language. In his Innateness Hypothesis, Chomsky postulates that the ability to learn the grammar of one’s native tongue is innate and genetic. Without having to hear every single instance of every single rule in English, for example, a native English speaker could form a complex cognitive system in a few years that encompass the grammar that makes up the English language. 

Let us step back for a moment. Many of us remember how hard it is for us to pick up a second language in high school or college, yet sit in awe as we reminisce over the days of how we picked up our native tongue so quickly and effortlessly. We remember sitting through lectures learning where to conjugate a verb, what terms like the future subjective mean, where to place words in a sentence, and a bunch of other rules that felt, for lack of a better word, alien to most of us. (Or maybe interesting? Who knows?) Yet we still struggled to gain supremacy over this ‘foreign’ language. (I always thought that it was weird that in the United States we call second languages ‘foreign’  since English is not ‘native’ to the United States, but I digress.) Why did we have to learn it through rules when we young children could pick it up without being taught? 

Well, there are a few reasons for that, including reasons that to this day are still not well understood. For one, from a linguistic standpoint, children pick up the rules of their native tongue because they have a genetic predisposition. Through another concept pioneered by Chomsky, the idea of a Universal Grammar that encompasses all languages allows babies to be ‘programmed’ by any path that they choose. After all, some universal principles can be present in a language. By the time that we hit puberty, we will have selected a path from the many different paths (languages, in this case) to follow. Through a complex process involving synapse pruning, the brain’s neuroplasticity decreases, strengthening the connections between neurons that are used, while disposing of the rest. Thus, by the time we are adults, our ability to be ‘programmed’ is decreased as we simply have fewer potential pathways for our brain to absorb new information. (We do theoretically have infinite long-term memory, so don’t take this as having our brain ‘stuffed’ with information and we can’t take any more in after a certain age, it’s just some processes will be harder to adapt to since the brain has already strengthened connections in the areas that we use it more in.)

Since our brain was more capable of being molded when we were younger, this partially explains why it can be hard to distinguish between sounds that we never encounter until later in life. Yet, admittedly, scientists do not know how children pick up the rules of their mother tongue. The exact processes are still unknown. We know why it is easier, but we don’t know how to replicate that process completely ourselves. Hence the challenge presented to computers to pick up rules ‘intuitively.’

In the fascinating world of artificial intelligence, computer scientists have been working to understand how to get computers to learn the way that people do. While there have been advances in deep learning and other processes, the human brain to this day remains a far more advanced computing system than any computer for this reason. That explains why a lot of computer science deals with neural networks that are based on the human brain.