👂 Sound science — why speech goes past you
Ten years of study, and a film in that language still goes past at half comprehension. Most people blame their memory. But the problem is not how many words you have stored — it is that the sound is being filtered out before it ever reaches the part of you that knows them.
A word you recognise instantly on the page and not at all in the ear is not a word you learned badly. Your brain has classified the sound itself as noise it cannot process and discarded it before meaning was ever attempted. Which is why another thousand flashcards does not move it.
In its first months a baby can tell apart sounds from every language on earth. By around twelve months old that has narrowed: the ear keeps the distinctions its own language actually uses and lets the rest collapse together. This is often explained as a matter of frequency — one language low, another high — but the bands do not carry the argument ([[frequency-myth]]). The truer description is not that you cannot hear the sound. It is that you cannot split it. The cue separating r from l, or b from d, goes past in about 0.04 seconds, and to an ear that stopped tracking that cue the two arrive as one thing ([[fast-forword-004sec]]).
Nobody has a poor ear; every ear has simply been sharpened on a different set. Japanese speakers famously take English r and l as one sound. English speakers hear the three Korean k sounds as one, and hear the Spanish pero and perro as the same word. Speakers of languages without tone hear four Mandarin tones as one syllable said with feeling. Run it in any direction and the pattern holds — the language you grew up in decided which differences would stay audible to you, and it decided that before you could read.
One useful way to lay out what happens between hearing and speaking: ① the auditory cortex takes the sound in → ② Wernicke’s area works out the meaning → ③ Broca’s area assembles it back into speech → ④ the motor cortex stores it as habit. For most adult learners the block sits at the very first gate. The auditory cortex does not hand on a sound unlike the ones it grew up with. The other three gates can be in perfect order and nothing starts, because nothing arrives — which is the plainest account of why more vocabulary never worked.
Reading and listening run on different circuits. Classroom language teaching almost everywhere builds the first one — grammar, translation, texts — so the reading side grows out of all proportion while the listening side stays close to empty. A large reading brain and an empty listening one still cannot hold a conversation. Nor does a listening exam settle it: test audio is read out slowly and clearly, and real speech has a rhythm and a speed that no test recording has. On putting the order back, see [[listening-first-order]].
A parrot picks up whatever its household speaks, one language or three, because it has no filter at all — sound goes in exactly as it arrives. Humans have the opposite problem: a highly developed first-language filter that turns unfamiliar sound away at the door. Unable to hear it, we fall back on memorising our way into speech, and memorised speech drains away with time. That loop is what most language education has been repeating.
Dropped into the middle of a place with the ear still closed, you are simply standing in loud noise. Learners write to us from language courses abroad saying they follow less than half of the class. The order is backwards — exposure only turns into learning once the ear can take the sound in at all.
The quickest way in is an unexpected one: use a language you already understand. Played at 2× to 4×, it costs you nothing to follow the meaning, so all of the effort goes into raising the speed at which the brain handles sound ([[listening-effort]]). Carry that faster processing across to the language you are learning and speech at native speed starts to arrive slowly. Thirty minutes a day is the dose, and thirty days is the shape of the routine. (Our free Korean course is built on exactly this order, if you would like to see the whole thing running: brain-hz.com/korean — 60 lessons, shadowing and speak-first modes, uncompressed audio, free and no account.)
Reading about it opens nothing — it takes training.
Source
Adapted from part one of Oh Seung-jong, 100 Questions on Daechung English (2026) — the items on the first-language ear and the auditory filter. The four gates (auditory cortex → Wernicke → Broca → motor cortex) follow the KBS Special “The Real Reason You Cannot Hear English” (broadcast 18 December 2011) as the author summarised it. On why we do not lean on the frequency-band explanation see [[frequency-myth]]; on the science behind speed listening, [[wav-speed-listening]]; on hearing and brain health, [[hearing-dementia]]. For the training itself, [[what-is-shadowing]].
Read the original →※ This article summarizes and restructures the key points of the source above.
🧭 How the method works · 11 more on this