Comprehensible input is language you can understand, spoken or written, even though it contains a little you haven't learned yet. Stephen Krashen's input hypothesis says this is how languages are acquired: "we acquire language when we understand messages". The idea is influential but debated. Reading stories written at your level is one of the simplest ways to get it.
Where does the idea come from?
Krashen first published the input hypothesis in 1977 and developed it through the 1980s, most fully in Principles and Practice in Second Language Acquisition (1982) and The Input Hypothesis: Issues and Implications (1985). It is one of five linked hypotheses, sometimes called the monitor model.
| Hypothesis | The claim, in brief |
|---|---|
| Acquisition–learning | Picking a language up subconsciously, as children do, is separate from consciously learning its rules. Only acquisition builds the ability to use it. |
| Monitor | Rules you have studied can only check and correct what you say or write, and only when you know the rule, are thinking about correctness and have time. |
| Natural order | Grammar is acquired in a roughly fixed order that teaching doesn't change. English third-person -s is easy to explain but comes late. |
| Input | You progress by understanding language slightly beyond your current stage, which Krashen wrote as i+1. |
| Affective filter | Anxiety, low confidence or boredom act like a filter that stops input getting through. |
What does i+1 mean for a learner?
In Krashen's shorthand, i is the stage you have reached and +1 is the next one. The input has to be understood, not merely heard or seen. A radio station in a language you don't speak gives you plenty of input but none that is comprehensible. Krashen's claim is that if you understand enough, the next structures you are ready for will turn up on their own, so they don't need to be drilled one rule at a time.
Reading fits the idea neatly. In a 2004 conference paper Krashen describes reading for pleasure as comprehensible input, and notes that a text needs to be about 98% comprehensible before a reader can pick up new vocabulary from it, citing Hu and Nation (2000). His later work adds a second condition: the input should be compelling. How interesting learners find it matters, not just whether they can follow it.
What does the research support, and what is still argued over?
The evidence in favour
Krashen's own survey of in-school free reading found that pupils given reading time did at least as well as those given traditional instruction in 51 of 53 comparisons. In studies lasting longer than a school year, the readers were nearly always ahead. Independently, Jeon and Day's 2016 meta-analysis of extensive reading in a second language pooled 49 studies with 5,919 participants and found a small to medium effect on reading proficiency.
The criticism
The hypothesis has been contested from the start. McLaughlin raised methodological objections in 1978, and a 1987 paper by White was titled Against comprehensible input. Later reviewers argue that the theory is hard to test, vague in its key terms and overstates what input can do alone. Its sharp split between acquisition and learning has never been demonstrated, and competing accounts give more weight to skill-building and to producing language yourself.
Where that leaves a learner
Even researchers who are sympathetic tend to treat input as necessary but not sufficient. Butzkamm and Caldwell argue that learners also need to understand how the language is built. Paul Nation recommends reading in large amounts alongside direct vocabulary study, because a word picked up from a single meeting in a text is fragile and soon lost. Krashen himself accepts that pleasure reading alone won't deliver academic language.
How does Contos use comprehensible input?
- One level per text. Every chapter is written at a CEFR level from A1 to B2. At A1 that means short sentences and very common words, so a beginner can follow the plot from the first page.
- Tap when you're stuck. Tap a word and its meaning appears in your own language. One unknown word doesn't have to stop the story.
- A reason to keep going. Every chapter ends on a cliffhanger. That's our answer to Krashen's point about compelling input.
- Some deliberate study, too. The words you tap become flashcards. Strictly, that's "learning" rather than acquisition, but it follows Nation's advice to reinforce words before they fade.
- Levels that follow your vocabulary. Contos moves you up when you know enough words: 500 for A2, 1,500 for B1 and 4,000 for B2.
Open chapter 1 of a series at your level and see how much you can follow without tapping.
Read free in your browserFrequently asked questions
Who came up with comprehensible input?
The linguist Stephen Krashen. He first published the input hypothesis in 1977 and set it out in full in his books of 1982 and 1985, as one of five hypotheses about how second languages are acquired.
Is comprehensible input the same as immersion?
No. Immersion means being surrounded by a language. Input only counts as comprehensible if you understand it. A beginner dropped into fast native conversation hears a great deal but understands little of it, which is why material pitched at your level matters.
Has the input hypothesis been proven?
No. Studies of reading programmes support the idea that understood input helps, but critics argue that the hypothesis is hard to test and that it underplays practice, output and explicit teaching. Treat it as a useful guide, not settled science.
Can you learn a language from input alone?
That is the main point of disagreement. Krashen argues acquisition comes from understood input, but even he accepts that pleasure reading won't cover academic language. Most researchers recommend combining plenty of input with some direct study of vocabulary and grammar.
How do I know if a text is comprehensible enough?
You should follow the story and only stop now and then. Krashen cites research suggesting about 98% of the words should be familiar. Our page on extensive reading looks at those coverage figures in more detail.