Guide · 10-minute read
You've probably heard that you learn a language by understanding it, not by memorizing rules. That idea is mostly right — with one important catch. Here's the honest version, plus a quick test to see if a Spanish text is the right level for you.
1 · The idea
Input is the Spanish you read or hear. Comprehensible means you understand it — mostly. Not every word, but enough to follow what's going on: who did what, and why it matters.
The term comes from the linguist Stephen Krashen, who argued about forty years ago that we acquire a language when we understand messages in it, not when we study its rules. Ideally the input is a little beyond what you already know: familiar enough to follow, new enough to stretch you.
In practice that means meaning first. You read a story or listen to a conversation because you want to know what happens. New words and grammar ride along with the meaning, and the context helps you work them out.
Meaning first: two sentences from a story
Y cada noche lo estaciona chueco. Medio carro queda en mi lugar.
And every night she parks it crooked. Half her car ends up in my spot.
2 · The evidence
Comprehensible input is popular online, and claims about it range from careful to wildly overconfident. Here is what holds up, graded by how strong the evidence is. Full details and sources are on our science page.
Broad agreement
Researchers from very different camps agree on this. A 2021 review of Krashen's ideas found “a central role for communicatively embedded input in all theories” of second language learning (Lichtman & VanPatten 2021).
Not demonstrated
Krashen's strong claim is that understanding input is both necessary and sufficient: nothing else is needed. That was widely criticized and has never been shown. It's a theory, not a finding (Lichtman & VanPatten 2021).
Evidence: good · effects modest
Reviews of dozens of classroom studies find small-to-medium gains from extensive reading (Nakanishi 2015; Jeon & Day 2016). The most careful estimate, which corrects for bias in earlier studies, is more modest — and comes from months of regular reading (Hamada 2020).
Evidence: good
A review of 24 studies found learners picked up only a share of the new words in a single text, roughly one in six to eight (Webb, Uchihara & Yanagisawa 2023), and fewer were still there months later (Brown, Waring & Donkaewbua 2008).
Evidence: strong
Across 42 studies, readers who could see what words meant learned clearly more of them than readers without that help (Yanagisawa, Webb & Uchihara 2020). Comprehensible input doesn't mean you must guess everything.
Evidence: good
Word-focused tasks taught more words than reading alone (Laufer 2003), and in one study, studying word meanings directly was more time-efficient than picking them up from stories (Hulme & Rodd 2021, with adults in their own language). It isn't guaranteed, though: in a review of 22 studies, many practiced words were forgotten by the later test (Webb, Yanagisawa & Uchihara 2020).
Evidence: strong
Pulling a word out of memory beats reading it again (Karpicke & Roediger 2008), and spacing practice over days beats cramming — for language learners too (Kim & Webb 2022). These are some of the most reliable findings in learning research.
The honest takeaway
Lots of input you understand, plus a little deliberate practice of the words you want to keep.
Input gives you meaning, context and a feel for how Spanish sounds. But from input alone, most new words slip away. Recalling a few of them on later days is what makes them stay. Neither half replaces the other.
You may have heard that you need a set number of hours of input — 1,500 is a popular figure — to understand and speak comfortably. That's a rule of thumb from online learner communities, based on people's own experience. It isn't a research finding, and we don't know of a study that measured it. Hours are a fine way to track your effort; just don't treat any number as a promise.
3 · The numbers
Researchers measure this as text coverage: the share of the running words in a text that you already know. If a story has 100 words and you know 95 of them, your coverage is 95%.
That sounds high, but small gaps add up fast. Here's what different levels look like — each block is a word, and the yellow ones are words you don't know:
Two figures come up again and again. In a classic study, learners of English read a story in which some words had been swapped for made-up ones. At 80% coverage nobody understood it well enough, and at 90% and 95% only a minority did; the authors estimated that about 98% — one unknown word in 50 — is what most learners need to read a text without help (Hu & Nation 2000, as described by Nation 2006). A larger study of 745 college students in Israel suggested two levels: about 98% for reading on your own, and about 95% for reading with some guidance and help (Laufer & Ravenhorst-Kalovski 2010).
But it's not a cliff. Newer studies find that understanding rises gradually as you know more of the words, with no magic point where it suddenly clicks (Schmitt, Jiang & Grabe 2011). A replication of the classic study couldn't fully confirm its results (Kremmel et al. 2023), and a 2025 study found no clear difference between 90%, 95% and 98% for advanced learners (Webb, Pellicer-Sánchez & Wang 2025). Learners and texts vary a lot. All of these studies were with learners of English.
What this means for you: treat 95% and 98% as useful rules of thumb for choosing texts, not as pass/fail lines. The 98% figure assumes you have no dictionary or translation; with help at hand, a bit lower is fine (Nation 2006). It also explains why novels and native TV feel so hard at first: for English, one estimate is that you'd need around 8,000–9,000 word families to know 98% of the words in written texts, and 6,000–7,000 for spoken ones (Nation 2006).
4 · Try it
Below is a real paragraph from one of our stories, El carro chueco (The Crooked Car). Read it, and tap every word you don't know. We'll count your coverage as you go. Be honest — a word you'd only half-guess counts as unknown.
El carro chueco · 9 sentences · 66 words
. , . . .
, . . . .
¿ ?
On a keyboard: Tab into the text, move with the arrow keys, press Enter to mark a word.
I live in a building with a small parking lot. Each neighbor has a spot, with their number on the wall. My spot is next to doña Celia's. She always gets home first.
Doña Celia has a small, old red car. And every night she parks it crooked. Half her car ends up in my spot. So I park in the visitor spot.
Does she do it just to annoy me?
How we count: every running word counts, so a word you don't know is counted each time it appears — tapping it marks all of them. The thresholds are rules of thumb from research on reading (see above), not exact cut-offs.
5 · In practice
At A2 you know the basics, but native TV, podcasts and novels are still packed with words you haven't met. Texts written for learners help: in one review, learners picked up more words from texts written for them than from texts written for native speakers (Webb, Uchihara & Yanagisawa 2023). Here's a simple routine that combines input with a little practice.
Four complete stories with audio for every sentence and English on tap:
Or practice one skill at a time: Spanish reading practice with short texts and translations by sentence, and Spanish listening practice sentence by sentence.
Probably not, if you want to keep the words you meet. Understanding input is central, but the claim that it's all you need hasn't been shown (Lichtman & VanPatten 2021), and from reading alone you keep only some of the new words. Adding a little recall practice is the best-supported way to make them last.
Both count as input. Reading lets you go at your own pace and see every word; listening trains your ear for real speed. Reading while listening combines them — research finds it at least as good as reading alone, and most learners prefer it (Brown, Waring & Donkaewbua 2008).
If you lose the thread of what's happening, it's too hard for now. As a rough guide, more than one unknown word in every 20 (below about 95% coverage) means you'll probably want a translation at hand. Use the tool above to check.
Sources checked November 2026. More on each one on our science page. Found a mistake? Write to support@habloro.app.
In the Habloro app, you read and listen to a short story, keep the words you didn't know, and recall them in their own sentence over the next days.
Habloro
