The textbook approach to a new language is grammar → vocabulary → reading → speaking. I tried it, plateaued at month three.

Then I switched: inputs first, grammar later.

The bet#

I bet that the brain figures out grammar by pattern recognition if you feed it enough natural input. Children do this. Adults underestimate how much they can do it too.

What "enough input" means#

For me, daily:

• 30 min Anki (vocabulary spaced repetition)
• 60+ min audio: podcasts, video, YouTube
• ~10 min reading (graded readers, slowly)
• Almost zero grammar study

Six months of this:

• Comprehension of casual speech: ~30% → ~70%
• Vocabulary recognition (passive): ~3000 words
• Grammar understanding: less explicit than textbook learners my level, but comparable functional usage
• Speaking: poor (didn't practice)

Why audio specifically#

Reading lets you cheat. You scan, you skip, you re-read. Audio forces sequential, real-time processing — closer to how the language is actually used.

What didn't work#

• Anki without audio cards: I learned to recognize text, not sound
• Trying to "translate in my head" while listening — too slow, kills comprehension
• Watching anime with subtitles: my eyes did the work, my ears didn't

Now#

I'm at the end of the kind of long project where the temptation to declare "good enough" is strong. I need to keep going for two more years before I can comfortably read novels — which was the goal.

The method has worked. I'm staying with it.

Comments

Comments are moderated. No email, no IP collection.