More on the input hypothesis: is an output-feedback loop necessarily less efficient that pure input for improving speaking ability?

Assume a learner wants speak Japanese fluently in 2 - 4 years and is not super concerned about developing an accent. The learner has currently been doing input only for one year. The learner decides to replace some of the input time with an output-feedback loop (e.g. speaking with a professional teacher with feedback). Will this necessarily be less efficient? What percentage of input time would need to be replaced by an output feedback-loop before it would be counter-productive? This seems like an issue of opportunity costs and marginal utility.

For example, I’ve heard many people recommend to only meet with a tutor to practice speaking 1-2 times a week. Is this because efficiency decreases after that? Assume I have infinite money and have no problem meeting with a tutor to practice speaking 5 times a week instead of practicing input with that same time. Would it be wasteful in terms of opportunity costs?

Haven’t gotten a heck of a lot of practice in speaking Japanese yet, but speaking from my 5 years of French experience, the more practice you get, the better. Full immersion is the absolute best-case scenario (that is, for someone who understands enough to make it not completely useless jibberish) but until travelling to Japan is possible, your best option is to get as much speaking practice in as possible to improve your accent and speed.

Having a native or fluent speaker as a tutor is great and should be done as much as your schedule and budget allow since that helps prevent you from developing bad habits, but failing that, try just picking out random sentences from your day-to-day life and translating them into Japanese. If you can get a full sentence, say it aloud. Do as much listening as possible (the obvious one is anime but dramas, music, and radio are all just as good, radio probably being the closest to the speed, tone, and accent of native speakers). The goal is to maximize your exposure to the spoken language.

So no, I don’t see any reason why seeing a tutor 5 times a week would harm anything. It’s not like vocab memorization and SRS, where you need breaks in order to fully commit it to memory. The more exposure, the better. My guess is the reason they said 1-2 times per week is because that’s what’s manageable in any normal person’s schedule and budget. See what works, and the more immersive, the better. Listen to Japanese podcasts, with or without subtiles in your free time. Even if you don’t understand it, if you can expose yourself to the accent over and over again, you’ll better be able to mimic it.


As long as you’re trying to think about how to use the language sometimes (I want to say X./How do I say X?) and not just constantly consuming then I think any combination of approaches is fine. Keeping your brain engaged is the key and the truth is you can only reliably output after being fed some examples to go off of. If you never output then it’s like you never have to study for a test because the test never comes up.

Speaking is a physical skill. You get better at physical skills by practicing them.

The ideal situation would be “controlled” immersion, in which you get all the input you can stand and are only allowed to communicate in Japanese and are surrounded by people who are both fluent speakers and willing to correct you.


You can form the equation, but you’re not going to get a satisfactory answer to this. Comparatively, some output practice improves efficiency compared to no output, but I don’t think anyone can give you a universally applicable ratio since a) no one’s studied what it might be and b) it probably changes with competence.

I was being partially rhetorical. I maInly just wanted to emphasize the marginal utility and opportunity cost aspects of the post. (In truth, I’m fishing for perspectives on the general ideas of the post.) For example, your response indirectly suggests that ratio is likely not zero and might increase (more output-feedback) with competence.

In that case, more is probably better…after all, you’re still practicing input as well during speaking practice.