This so compelling a thread I just had to get into it. The input has to be compelling not b/c input that is input i.e. uptake, won’t be processed as input but simply b/c it has to be attended to to be uptake (I believe VanPatten makes the distinction between input and uptake i.e. input that is received and processed). The processing may not be conscious but the attention is. The infant becoming fant (play on words: in-fant is Latin for not-speaking) attends to everything; 15 year olds attend to their bodily functions. At least that’s the way I see it.
Re natural order: the way I have always understood it (I first heard of it reading Krashen and then discussion of it in other books on SLA; I recently put in my blog a piece on John McWhorter’s comment that in a language pared down to minimal levels of grammar in the creolization process, there is a natural order in which “parts” i.e. grammatical features, are lost). No one knows what order what features are acquired in (awkward I know), so we present input to learners in a way that repeats chosen structures. When we begin to choose the structures focused on on the basis of a curriculum or need or some other basis, we have no idea whether we are following the natural order or not b/c we don’t know the natural order. Perhaps over time and with as much money poured into SLA as into developing new tanks, we could find out. Until then, we select 3 or so structures to “present” as circled input and the learners seem to acquire. At least that’s what my friends in tprs tell me. Does it matter which structures are chosen? I don’t know. Who does? But when you target certain structures just bear in mind, you might be going against the natural order.
So how about Krashen’s oft-repeated dictum: speak it and they will find it, i+1 that is?That makes sense to me. So how many tprs teachers select randomly structures to circle?