Four chunks per song. That's the constraint, and it's the least intuitive decision in the whole system.
A three-minute song could carry thirty new phrases comfortably. Most people's instinct — ours included, at first — is that more would be better. It isn't, and the reason is a hard limit inside your head.
The Bottleneck Is a Small Number
Working memory, the mental workspace you use to hold and manipulate things in the moment, holds only a handful of items at once. This has been known since Miller's 1956 paper on the magical number seven, and every revision since has moved the number down rather than up.
Present somebody with thirty new phrases and the workspace overflows immediately. What gets encoded isn't thirty items — it's a blur, with the first two and the last one surviving. The other twenty-seven were heard and lost.
So a song stuffed with material doesn't teach more. It teaches roughly the same amount and wastes the rest of the runtime.
What You Do With the Space Instead
Cap it at four and something becomes possible that couldn't happen otherwise. The remaining runtime goes on repetition and recombination.
The four appear, then reappear attached to different objects, then in question form, then combined with each other, then inside longer sentences built entirely from parts already heard.
By the end you've met each core chunk fifteen or twenty times in varying contexts — which is what actually moves something from conscious recall into automatic production. Not exposure to more items. Repeated exposure to fewer.
Four Multiplies Anyway
Here's what makes the constraint cost nothing.
Four core chunks aren't four phrases. "I want" attaches to every object you know. "Do you have" attaches to the same list. Combine them and you get questions and answers; nest them and you get complex sentences.
Four chunks generate thirty or more usable phrases inside a single song, and every one is composed of parts you've now heard twenty times. The output is large. The number of new things you had to hold at once was four.
The Discipline Is in the Leaving Out
Every song could contain more. The work is deciding what not to include — the opposite of how language material is usually built, where volume is the selling point and courses boast about how many words they contain.
Volume is easy. Restraint is what makes it stick.
About Outputly
Four high-frequency chunks per song, multiplying into 30 or more usable phrases across three minutes. Across 100 songs that's over 3,000 phrases — built from a few hundred core chunks you've each heard dozens of times.