Why comprehensible input still matters.

Stephen Krashen published Principles and Practice in Second Language Acquisition in 1982. Forty-four years later, his central claim is still the most contested idea in language teaching, and it is still mostly right.

What Krashen actually said

The Input Hypothesis is narrower than its critics suggest and broader than its fans usually claim. Krashen argued that humans acquire language in one way: by understanding messages, or what he called comprehensible input. Production, conscious study, error correction, and grammar drill might shape what students notice and what they remember about a rule. None of these things produce acquisition.

That is the claim. The rest is commentary.

For forty years critics have lined up to point out that this is too strong. Surely some output matters. Surely some attention to form matters. Surely some learning happens through deliberate practice. These criticisms have a point. Krashen overcorrected against the audiolingual orthodoxy of his era, and the field has spent decades adjusting the dials. But the dial has never moved back to where the textbook industry would prefer.

The thing he actually got right

Watch a Spanish 1 student trying to read an authentic news article. The affective filter is real. Their face freezes. They scan for cognates, give up, switch to English. Now watch the same student read a graded version of the same story, with vocabulary they mostly know, illustrated, with difficult words glossed in the margin. Something different happens. They lean in. They turn the page. They figure out the third paragraph from context. Two weeks later they remember the verb for “to invent” because Carla in the story invented something funny.

This is what Krashen pointed to. He gave it a name. The field has argued about the boundary conditions for four decades, but the central observation has held up across thousands of studies. Reading and listening, when mostly understandable and somewhat interesting, produce acquisition. Worksheets do not.

The acquisition system is the one a child uses. The learning system is the one your AP textbook uses. You cannot test your way into acquisition.

What we got wrong in the nineties

The first wave of CI teaching produced some embarrassing classrooms. Teachers, hearing that input was sufficient, stopped teaching grammar entirely. They stopped giving feedback. They stopped asking students to write or speak. The pendulum had swung.

Two responses came back. Roy Lyster's work on counterbalanced instruction (Lyster, 2007) showed that pure CI without form-focused feedback was insufficient when the target structure violated learners' L1 expectations. Norris and Ortega's meta-analysis (2000) found that explicit, form-focused instruction had a large effect size, d = 1.13, when it was embedded in meaningful communication.

These findings did not refute Krashen. They refined him. Input is the engine. Output, feedback, and focus-on-form are the steering. You cannot steer a car that is not moving, and you cannot move a car with only a steering wheel.

What modern comprehensible input looks like

A current CI classroom is busier than the 1990s version. The teacher still leads with story-based input, embedded readings, and authentic resources at appropriate levels. There is more pushed output (Swain, 1985). There is more deliberate focus-on-form during meaningful tasks (Lightbown and Spada, 2013). There is more spaced retrieval of vocabulary (Cepeda et al., 2006). There is more explicit teaching of reading strategies.

The pedagogy has not abandoned Krashen. It has built infrastructure around his central insight.

What this means for your Monday

If you teach world languages in 2026, here is the simplest test of whether your daily plan honors what the research actually says. Take one lesson. Count the minutes students spend reading or hearing connected target-language input that they can mostly understand. If the number is under fifteen, you are running an old textbook. If it is over twenty-five, you are running something closer to what acquisition requires.

Everything else (the grammar pop-up, the vocabulary chart, the writing prompt) is supporting infrastructure. Worthwhile, but supporting. The input is the lesson.