Many people think that in order to learn a foreign language you have to start speaking it as soon as possible. Ideally, they say, you should go live in the country where they speak it. Realistically, they say, if you don’t take those immersive steps you will have no chance of learning it.

This advice is very common, but it doesn’t always come from those who have successfully achieved fluency through sustained effort.

The main determining factor of success in language learning is whether the learner sticks it through to the end (although really there is never an “end”). What better way to demoralize yourself than to try to stumble through greetings with a waiter at a French restaurant or an employee at an Italian coffee bar? Having had those experiences myself I can assure you that they aren’t conducive to sticking it through in your language learning. They make you want to quit, they make you think that their rapid speech is something unattainable.

That’s why I stick to the approach of getting really good at reading and listening comprehension before I start working on my speaking ability. It took me about 6 months of daily Greek study before I felt confident enough in my abilities to start to tell people I was learning Greek. In the beginning when I would mention that I was learning Greek to a native speaker, they would start speaking to me in the language and I would freeze up.

What I tend to do is massive amounts of comprehensible input with content that interests me and is targeted to my level, until I reach a point where I need noticeably less translation, pausing, and rewinding. Comprehensible input is a core part of the input hypothesis, developed by the linguist Stephen Krashen, where most emphasis is placed on the dual competencies of spoken and written language comprehension. Understanding the spoken and written word of native speakers is the main competency that fuels language learning. It is the seed that allows you to start speaking later. This could be achieved with anything from graded readers and podcasts for learners to TV shows with subtitles, depending on your level and interests.

Usually, once I realize I am pausing and rewinding podcasts a bit less, and can talk to myself about some basic things in an accent that isn’t totally embarrassing, I will find an online tutor that I can video chat with. The key point here is that this usually happens many months after I start learning. Not only do I think that this isn’t a problem, I know that it is ideal for me. It is not procrastination, it is not a waste of time.

Waiting this long before starting to speak actually improves the conversations you have with a conversational buddy. Imagine starting to speak with someone from day one. What are you going to talk about? Mostly boring stuff. If you start speaking once you have a sufficient base, your first conversation can be about movies, TV shows, travel, or any number of more engaging topics. You might seriously struggle in your output, but you won’t struggle to understand what your partner is saying.

One final point about this. I often find that progress in my speaking ability goes from really bad to amazing in a very short amount of time when I follow this method. I don’t know the exact neuroscience behind it, but there is some sort of activation that has to happen for you to begin to speak really well, and this activation is only going to be as good as your underlying comprehension in the language. The bigger your “base” of comprehension, the quicker you will improve your speaking ability once you start.