The More Words, the Better

Clark’s paper talks about how children learn ‘-er’ endings and instrument nouns. The study found that children often construct new works to fill gaps in their vocabulary; for example, a child will infer that “typewrite” is someone who typewrites based on the noun “typewriter.”

How does one know when a word can be transformed to a noun by adding ‘er’ (i.e. write -> writer) and when it cannot (i.e. typewrite -> typewriter)? And broadly, how does one know when it’s appropriate to apply a general rule to a specific instance?

In Symsys1, we talked about how learning often follows an inductive fashion. For example, learning the concept of “dog” involves seeing several examples of dogs, which inform the concept of a dog in our minds. Similarly, by learning inductively, we do not have to list all negative examples or exceptions to the definition—without being told, we know that a table is not a dog. On the other hand, it’s difficult to learn the concept of a “dog” through deductive learning because we are not able to provide a comprehensive yet specific enough definition that compensates for all dogs in the universe.

Thus, when children learn language, they learn not from generalizations, but rather from specific examples. There is no general rule that tells us how to distinguish between words like ‘writer’ from words like ‘typewriter.’ Instead, we start to realize which words can be transformed by adding ‘-er’ through examples and experiential knowledge, eventually forming a kind of “instinct” for language.