Comprehensible input is the linguist Stephen Krashen's idea that people acquire a language when they understand messages that contain a little language just beyond their current level, written as i+1. According to Krashen, speaking and writing are the result of acquisition, not its cause; learners need to listen to and read a great deal of material they can understand, in a relaxed state of mind.
The idea has had a strong influence on language teaching since the 1980s, and it is also the idea quoted most often in descriptions of immersion courses. But it is also one of the most criticised theories in the field. This article sets out Krashen's five hypotheses, their strengths and weaknesses, and how to use the part that holds up when you assess a language class.
The five hypotheses of the Monitor Model
Krashen developed these ideas from the late 1970s and set them out in full in Principles and Practice in Second Language Acquisition (1982) and The Input Hypothesis: Issues and Implications (1985). Together they are usually called the Monitor Model, and they consist of five hypotheses:
- Acquisition and learning are two different processes. Acquisition is an implicit process, like a child picking up their mother tongue, and it produces the ability to use a language fluently. Learning is conscious knowledge of rules. Krashen argued that what is learned does not turn into what is acquired.
- The Monitor. Learned knowledge only works as an "editor" that corrects sentences, when the speaker has time, is paying attention to form and knows the rule. In fast conversation the monitor hardly has time to run.
- The natural order. Grammatical structures are acquired in a fairly stable order that does not depend on the order in which they are taught. This idea rests on studies of English morphemes by Heidi Dulay and Marina Burt in the early 1970s.
- Input. Learners move from level i to i+1 by understanding input that contains i+1, helped by context and knowledge of the world. There is no need to teach i+1 deliberately; if there is enough input and it is understood, i+1 will be present by itself.
- The affective filter. Anxiety, low self-confidence and lack of motivation put up a "filter" that stops input from reaching the part of the mind that does the acquiring. The filter concept was put forward first by Dulay and Burt, and Krashen built it into his model.
The five hypotheses lock together into a tidy system: to acquire a language you need comprehensible input with a low filter; studying grammar only helps you correct sentences when you have time. That tidiness is exactly why the model spread so easily, and also why researchers examined it so closely.
What i+1 means in an actual lesson
The notation i+1 is easily mistaken for a formula: if you know 1,000 words, give the learner a text with 1,001. That is not what Krashen meant. In his view i+1 cannot be measured precisely and teachers do not need to target each structure; as long as the input is rich and the learner understands most of it, the next-level material will turn up often enough on its own. He called this "casting a wide net" rather than aiming at one grammar point at a time.
Take an example. Your child, at beginner level, listens to a teacher telling a story about a trip to the market, with pictures, gestures and real objects. Your child understands the story thanks to the context, even though it contains words and structures they have not learned yet, such as the past tense of irregular verbs. After meeting those structures many times in stories they understand, your child starts to use them without anyone explaining the rule. This is the picture Krashen had in mind.
On the other hand, a listening exercise in which your child understands only a few per cent is not comprehensible input, even if it is "standard native-speaker English". A text your child understands completely, with nothing new in it, does not move them forward either. Later reading researchers such as Batia Laufer and Paul Nation produced estimates that are often used: to read a text comfortably, a learner needs to know roughly 95 to 98 per cent of the words in it. These estimates are independent of Krashen, but they help to picture how much "comprehensible" actually is.
The main critiques, in the order they appeared
It cannot be tested. In 1987 Barry McLaughlin, in Theories of Second-Language Learning, and before him Kevin Gregg (1984), argued that concepts such as "acquisition", "learning", "i+1" and "filter" were not defined clearly enough to be measured, so the model could hardly ever be shown to be wrong. A theory that cannot be wrong is hard to call a scientific theory in the strict sense.
Input alone is not enough. Merrill Swain observed pupils in French immersion classes in Canada: after years of hearing and reading French every day, they understood almost like native speakers, yet their speaking and writing still contained many grammatical errors. In 1985 she proposed the output hypothesis: only by speaking and writing, and being pushed to express themselves accurately, do learners notice the gaps in their own language.
Interaction matters more than one-way input. Michael Long argued that input becomes most comprehensible when learners negotiate meaning: asking again, asking for repetition, having things rephrased for them. His interaction hypothesis shifted the emphasis from "hearing a lot" to "back-and-forth exchange".
Conscious attention is necessary. In 1990 Richard Schmidt put forward the noticing hypothesis: learners have to notice a form in the input before it has a chance of being acquired. This contradicts Krashen's view that acquisition is entirely unconscious.
Explicit teaching works. Syntheses of many studies on grammar teaching, most notably the 2000 paper by John Norris and Lourdes Ortega, show that deliberate teaching produces measurable and fairly lasting progress, contrary to the hypothesis that learning does not turn into acquisition. What is still debated is whether that progress can be used in natural conversation.
Which parts of Krashen still hold up
Despite all the criticism, there is one point almost every school of thought agrees on: there is no acquisition without a great deal of meaningful input. Swain, Long and Schmidt did not reject the role of input; they only said it is necessary but not sufficient. Studies of extensive reading, that is, having pupils read lots of easy books they choose for themselves, show good results for vocabulary and reading comprehension, in keeping with Krashen's thinking.
The affective filter idea also has indirect support: research on foreign language anxiety, beginning with the work of Elaine Horwitz and colleagues in 1986, shows that high anxiety goes together with lower results. Even though a "filter" mechanism in Krashen's exact sense has not been demonstrated, the observation that stressed learners learn less well is backed by plenty of data.
The natural order idea is also partly right: learners of English do tend to master morphemes in a fairly similar sequence, although that sequence is influenced by the mother tongue more than Krashen thought. Vietnamese speakers, for example, tend to drop plural and past-tense endings for longer, because Vietnamese words do not change their form.
Bringing it into the classroom: signs that a course offers comprehensible input
You do not have to accept the whole Monitor Model to make use of the parts that hold up. The table below sets out what parents can observe or ask about when looking at a language course, especially a short course abroad.
| What to look at | Good signs | Signs worth asking about | Link to theory |
|---|---|---|---|
| Amount of meaningful listening and reading each day | Time for reading books at the right level, listening to stories, activities that use the language as a tool | Most of the time spent copying rules and doing gap-fill exercises | Input hypothesis |
| Difficulty of the material | A placement test, classes grouped by level, material your child mostly understands | A mixed-level group all working from the same textbook | i+1, the 95–98% word threshold for reading |
| Chances to speak and be corrected | Pair work, small groups, teachers who correct errors in a way that keeps the conversation going | Only listening to explanations, rarely having to speak | Swain, Long: output and interaction |
| Classroom atmosphere | Mistakes are not criticised in front of the class, there is time to prepare before speaking | Being called to the board without warning, every spoken sentence marked | Affective filter, foreign language anxiety |
The last column shows a sensible way to use the theory: not taking sides, but taking from each school the part that has evidence behind it. A good course usually combines all four rows: plenty of input at the right level, chances to speak, attention to form when needed, in a classroom that does not make your child afraid.
Common misunderstandings when people quote Krashen
The first misunderstanding: "Just listen passively, keep English playing all day and your child will become good at it." Krashen was talking about comprehensible input. Background sound that your child does not understand or pay attention to is not input in his sense. Patricia Kuhl's research with young children on learning from video compared with learning from a real person also shows that social interaction matters more than the sheer amount of sound.
The second misunderstanding: "Krashen proved that studying grammar is useless." This was his hypothesis, and it is precisely the part most strongly challenged by the evidence. A view more widely accepted in the field is that explicit grammar helps learners notice forms in the input and correct themselves, especially teenagers and adults, who are able to analyse.
The third misunderstanding: "Being surrounded by native speakers means getting comprehensible input." A low-level pupil sitting among native speakers who talk fast may understand very little and retreat into a group of friends from home. Input only becomes "comprehensible" when it is adjusted: a class at the right level, leaders who know how to speak slowly and clearly, activities with context. That is why a good immersion programme has to be designed, not just consist of sending the learner somewhere far away.
Frequently asked questions
Do researchers still take Krashen seriously?
His works are still widely cited, and the input hypothesis is the starting point of most textbooks on second language acquisition. Even so, few researchers today accept the whole model; most see input as a necessary condition, combined with interaction, output and attention.
How is the Natural Approach related to Krashen?
It is the teaching method Krashen developed with Tracy Terrell, set out in their joint book in 1983. The method gives priority to meaningful communication, allows learners a silent period of listening before they speak, reduces direct error correction and uses many activities with context to keep the input comprehensible.
How can I tell whether an English book suits my child's level?
A simple way is to have your child read one page and count the words they do not know. If there is more than one unknown word in roughly every twenty, the book is probably still too hard for reading a lot and reading comfortably. Graded readers are made to solve exactly this problem.
If my child is afraid to speak, should I push them to speak straight away?
Many learners go through a period of mostly listening before they feel confident enough to speak, especially in a new environment. Pushing a child to speak in front of a group easily raises anxiety. A better approach is to create chances to speak in small groups, give time to prepare, and then raise the level gradually once your child is used to it.
Does the idea of comprehensible input work for Chinese?
Yes, and for Chinese it matters even more in reading, because Chinese characters make written input harder to understand. Many Chinese courses have graded readers based on the number of characters already learned; learners use pinyin or zhuyin at first to keep the text comprehensible.
Is the affective filter a neurological concept?
No. It is a metaphor for the way negative emotions get in the way of taking in a language; there is no evidence yet of a corresponding mechanism in the brain. But research on anxiety, motivation and self-confidence in language learning shows that emotions do have a real effect on results.