Language acquisition theories have changed a great deal over the last half century. The most popular notions—for example that language learning ability declines with age—are based on ideas half a century old and not supported by the last several decades of research.
The idea that language-learning ability deteriorates with age is largely a myth. What is true is that very small children have one very important advantage in language learning.
Actually they are not that fast. It takes about four years to become fluent, but by no means adept, in one’s native language—and remember that children are “studying” the language almost every waking hour for that time. But their enormous advantage is that they don’t know any other language.
In some ways this is a disadvantage. It cuts out many of the shortcuts second-language learners can (and should) use—definitions of words, grammar explanations, etc. This actually does slow them down in relation to older second-language learners.
But far, far outweighing that disadvantage is the advantage that they learn language organically. They learn to associate words directly with concepts, not through an external medium (another language).
Many current language acquisition theories now accept the concept of Universal Grammar (UG). This means essentially that basic grammar is part of the inborn human mental apparatus. This theory is necessitated by the fact that small children learning language actually know more than they do, or can, explicitly learn.
What small children are in fact doing is mapping the local dialect (the native language) onto universal grammar structure that already exists in the brain. This is why, despite their great differences, the fundamental similarities of human languages are even more striking. They all have something like nouns, something like verbs, and something like adjectives, for example, and they all link them together to form descriptions of objects, events, and concepts in patterns that, while widely various, essentially work in the same way.
In the sentence:
Mary threw the ball at Susan
Mary is the grammatical subject, the ball is the direct object, and Susan the indirect object. English marks these simply by word order. If we say
The ball threw Susan at Mary
the ball has become the grammatical subject, Susan the direct object, and Mary the indirect object, and while the sentence may be absurd, it is perfectly grammatical and understandable.
German and Latin mark these three cases (subject, direct object, and indirect object) by declining the verbs with case-endings, making word-order largely interchangeable. Japanese does the same thing by appending a particle to each noun (ga,* wo, and ni respectively). But all the languages are doing the same thing. They are taking Universal Grammar and expressing it through different local conventions.
So each child begins, not with the tabula rasa (blank slate) postulated by the early-modern language acquisition theories of Rousseau and his followers in the 18th century, but with the fundamental structure of grammar in place and ready to have actual words and local structures mapped onto it.
To this child, the local words and structures she learns become the “incarnation” of Universal Grammar. She does not feel that she is learning “a language” but that she is learning Language. And in a sense, she is. She is learning, or more exactly actualizing, Universal Language in the particular form in which her local culture embodies it.
This is why the “native language” is nearly always more fluent than later-learned languages. To the mind, the native language is Language per se, while any second language is learned through the medium of the native language and therefore is mapped not to Universal Grammar but to the native language. It is a second-level mapping that can never be as organic as the primary mapping.
Many non-specialists are still influenced by 1960s language acquisition theories based on the exploded theory of the brain’s plasticity being greater early in life (it isn’t) and therefore language acquisition being easier at that time. They may point to the fact that a child learning a second or third language very early in life can pick it up to near-native level quite naturally.
The truth here is that the mapping of the local (native) language onto Universal Grammar is still not complete and “hardened”. Small children, even when fluent, still make many mistakes that even the most uneducated adult or older child never makes. Thus with the mapping still in a fluid state, a second or third language is mapped partly onto the first and partly onto Universal Grammar. In the case of a child exposed to two languages from birth, both may be mapped onto Universal Grammar and thus both will be “native”.
Now that we have looked at the basis of language acquisition theories, let us see how we can apply them to learning language in general and Japanese in particular, and especially to the problem of how we can try to map a second language to the Universal Grammar structure in our brains rather than to our first language. Because this is the key to near-native proficiency.
*Those who would argue that if wa is used the grammatical subject has not been marked should read my essay I Am Not an Eel.