This is a simple tool I built to learn how to type using a Korean keyboard (Korean Dubeolsik 두벌식 layout) and recently added Japanese: keeby.app

you know 500 japanese words. you can read them. you can recognise them on flashcards. but the second someone actually talks to you, it all turns to mud. here's whats going on:

There's a community of language learners who believe reading is the enemy of listening. They watch thousands of hours of video with zero text on screen, convinced subtitles stop your brain from learning to hear. The research says the opposite: 1/

i've been making mnemonics for hiragana with nano-banana this one is forち ("chi"), i.e. chi-teh i mean it's definitely memorable!

You understand everything in your target language. But when you speak, the grammar comes out wrong. You've been told to just listen more. It doesn't work. Here's why:

Some people watch 1,500 hours of content in their target language and still can't hold a basic conversation. Not because they're lazy. Because of a variable nobody in that community talks about: 1/

Tolerate ambiguity. Get your reps in. Trust the process. This is the standard advice for the intermediate plateau. It treats the problem as willpower. It's not. The cause has a name, and once you see the maths, more hours alone can't fix it:

The word "you" is a grammatical error. So is "bird," "goodbye," and "apron." Every English speaker alive is speaking a language built on mistakes:

Children aren't better at learning languages because their brains are more flexible. They're better because they haven't spent 20+ years training their brain to ignore the exact patterns they now need to notice:

The most popular method for adult language learning is modelled on how babies learn. But babies don't do what you think they do: 1/

Studying a word once today is more effective than studying it 8 times right now. This isn't a productivity hack. It's one of the most replicated findings in learning science:

Vocabulary lists are one of the first things people reach for when learning a language. They feel productive. They feel intuitive. And they are actively making your long-term comprehension worse:

The word "light" can describe a morning, a beer, a comedy, a meal, a shade of red, and an infantry battalion. If you memorised the dictionary definition, you'd still have no idea what it means in most sentences. True of almost every word:

Every language learner hits the same wall. You look up a word. You memorise the translation. You see it in a different sentence and it means something completely different. This isn't a gap in your vocabulary. It's a flaw in how you think about words:

"Learn like a baby. Use comprehensible input. Your brain will just learn naturally." I followed this advice for thousands of hours. Here's what the research actually says:

I haven't learned any kanji yet. Still on hiragana, 17 days in. But I've been reading about how to approach it and the debate is wild.

I've been copy-pasting my way through learning Japanese. Need to look up a word? Copy-paste. Search YouTube? Copy-paste. Write a message? Copy-paste. It's faster than typing when you have no idea where the characters are on your keyboard. And I got sick of it.

I started learning Japanese 17 days ago. Flashcards weren't working. I'd see the character, hear the sound, forget it 10 minutes later. So I'm building something I haven't seen anywhere else.

NEW BLOG POST! Declarative vs. procedural memory in language learning: What every learner & teacher should know Continue reading: matbury.com/wordpress/in...

new interview with Adele Goldberg on the constructionist approach to understanding how language works recommended listening for all language learners (skip to 47min mark) www.patreon.com/posts/121-le...

Thanks for the opportunity, Because Language! 👇

here’s what every language learner should know about memory: we have two main memory systems. • declarative memory stores facts—like the capital of France, or where you parked the car. • procedural memory stores patterns—like how to ride a bike or brush your teeth. a thread:

This is fascinating and intuitively resonant. 👇🏽

the problem with grammar study is that the brain does not construct sentences using textbook grammar rules—language exists in our heads as networks of associations between 'chunks', and the patterns documented in textbooks emerge from those associations 🧵 here's what I mean:

the next big problem in language learning, and potentially the last problem is Algorithmically Optimised Input a thread:

the next big problem in language learning, and potentially the last problem is Algorithmically Optimised Input a thread:

@adelegoldberg.bsky.social i tried to summarise the core ideas of Construction Grammar and mentioned your work in this video i'm not a trained linguist so i'm sure there are inaccuracies, feedback is very welcome! youtu.be/u_qGcp3XpAs?...

thank you to whoever was my first subscriber was!

in the 50yrs since Krashen's research on lang. acquisition, a lot has happened in the field of linguistics in my first youtube video, i talk about a new theory of how language is represented in the brain, and why we now know how comprehensible input works. link in replies: