T
VIEN

The four pieces

The one frame I use to check whether my study week is missing something.

⏱ About 5 min read

Of everything I have read about language learning, this is the frame I found most useful. It comes from Paul Nation, a researcher in New Zealand, who calls it the four strands. The original paper is free to read here, 12 pages long.

The idea: a decent study programme needs all four of the things below, and each should take roughly a quarter of the time.

I like it because it does not tell you what to study. It just gives you a way to check: which piece did I leave empty last week?

What the four pieces are

PieceWhat it meansWhat I actually do
1. Meaning-focused inputListening and reading material you mostly understandWatch shows, listen to podcasts, read fiction
2. Meaning-focused outputSpeaking and writing to convey a real idea, not to do an exerciseJournal, summarise an episode, have a conversation
3. Language-focused learningStudying the language directly: words, grammar, pronunciationAnki, a grammar book, sound drills
4. Fluency developmentRedoing what is already easy, but fasterSkim an easy book, replay an old episode, retell one story several times

The condition on piece 1 is specific: the material has to be at a level where you already know about 98% of the words — a number from Hu and Nation (2000). I go into it in How many words is enough.

Piece 4 is one I had never heard anyone mention before, and the one I neglected longest.

The piece I forgot: fluency development

Piece 4 teaches you nothing new. That is exactly the point.

It takes what you already know and makes using it fast and automatic. Without this piece you become someone who knows a lot of words and still speaks haltingly — because every word takes a second to retrieve.

Three things I do for it:

  • Easy reading, fast. Pick a book where I barely need a lookup, then read for 15 straight minutes, stopping for nothing.
  • Replay old episodes. An episode I have already seen twice, listened to a third time at 1.25x.
  • Tell it three times. Pick a story, tell it in English in 4 minutes. Then tell the same story in 3 minutes. Then 2. The content doesn't change, only the time you have.

The 4–3–2 exercise

That third one is a named exercise in the research, called 4/3/2. Nation describes it in Improving speaking fluency (1989).

The first time you think while you talk. By the third, the sentences come out ready-made. I find it far more effective than telling three different stories in the same nine minutes.

How I actually split my time

Nation says the four should be equal. I never managed that, and I don't think I should pretend otherwise.

My first year leaned hard into piece 1:

StageInputOutputLanguage-focusedFluency
Year 1~70%~5%~20%~5%
Year 2~55%~15%~15%~15%
Year 3 onwards~40%~30%~10%~20%

I still think leaning into input early is right — with nothing in your head, there is nothing to produce. But "leaning" is not the same as "zero".

A piece at zero is the real problem

The real mistake is not an uneven split. It is leaving one piece at zero for months on end.

I left output at zero for nearly two years. By the time I needed to speak, I understood everything being said to me and could not assemble a decent sentence myself.

The weekend self-check

Every Sunday I spend about two minutes on this:

1. How many minutes did I listen and read this week?

Under 5 hours and nothing else matters as much as pulling that number up.

2. Did I speak or write a single sentence of my own?

Three journal sentences count. Zero is the problem.

3. Did I have any session studying the language directly?

Anki, looking one unfamiliar structure up properly, drilling a sound.

4. Did I have any session doing something easy, at speed?

This is the piece I forget most. If every session in a week felt hard, this is usually what is missing.

Sources for this page

✏️ Edit this page on GitHub