T
VIEN

Comprehensible input

How I pick material at the right difficulty, and how I make it easier when it isn't.

⏱ About 3 min read

The original idea

There is one concept I think explains almost everything: comprehensible input.

It is material where you get the main idea, even without understanding every word.

Both extremes are useless:

  • Material you understand nothing of is just noise. It does not matter how much of it you consume.
  • Material you understand 100% of teaches you nothing new.

The useful part sits in between.

There is a number behind this idea: Hu and Nation (2000) found that to read comfortably without a dictionary, you need to know roughly 98% of the words on the page — about 6 unknown words per page of a novel. I go into it in How many words is enough.

The difficulty scale I use

ComprehensionWhat it meansWhat I do
Below 50%Too hardSwitch material. Don't push through.
50–70%Hard, tiringOnly for intensive reading of a short passage
70–90%The golden zoneMost of my time goes here
Above 95%Too easyGood for keeping the habit, but no progress

Making material easier yourself

You don't always have to change material. You can lower the difficulty of the thing you actually like.

1. Know the content in advance

Read a plot summary in your own language first. Knowing what is about to happen frees your brain up to process the language.

2. Watch it a second time

The second pass through the same episode is always far easier than the first pass through a new one. I think this is the most underrated trick there is.

3. Read first, then listen

Read the subtitles or transcript first, then listen again without looking at the text.

4. Slow it down

0.85x genuinely helps. But treat it like training wheels — temporary, not permanent.

Signs you are in the right zone

You can follow the thread of the story and you are curious about what happens next, but the odd sentence still slips past you.

If you have to rewind constantly to follow along, the material is too hard — however much you like it.

✏️ Edit this page on GitHub