Terms like “easy” and “difficult” imply an act of evaluation which in turn depends on the capacity to compare one task with another, which in turn depends on prior experience of tasks with differing levels of difficulty. Thus, when we acquire a second language, we can say that its derivational morphology, for example, is difficult to learn, while its relativization processes, say, are relatively easy. Such remarks are meaningful only because we already know a language and can measure features of the second language against those of the first. If we had no previously learned a language, we would have no standard of comparison; moreover it is at least in part the nature of what we have already learned that determines whether what we are now about to learn will turn out easy or difficult for us. Now, if we say that something is easy for a two-year-old to learn, we cannot possibly mean any of this; all we can mean is that the child is somehow preadapted to learn that thing, rather than other things, or that in terms of the present theory, he is programmed to learn it. […] things that children learn early, effortlessly, and errorlessly turn out repeatedly to be key features of creole languages, which the children of first creole generations acquire in the absence of direct experience, we can then assume that such early, effortless, and errorless learning results, not from characteristics of the input, or from the efforts of the mother- since the features involved are often too abstract to be known to any but the professional linguist- but rather from the functioning of the innate bioprogram that we have hypothesized.
- Bickerton (1981), Pag. 146