In short
- A story for learners is a real story with its language controlled, not a language exercise with a plot attached.
- Research on graded readers points to about 98% known words for easy reading, so most of your words must come from the reader's level.
- Simplification fails when it produces choppy, unnatural sentences; the fix is to rewrite whole passages, not to cut words one by one.
- Plan, write freely, check the words, rewrite the hard parts, read aloud, then test the story on a real learner.
Most stories written for English learners fail in one of two ways. Either the language is right and the story is dull, or the story is good and half the words are unknown to the reader. This article covers what makes a story work for a learner, what the research on graded readers says about controlling language, and why simplification done badly is worse than none. It ends with a step-by-step method, a table of faults and fixes, and one passage of mine before and after control.
What makes a story work for a learner?
A story works for a learner when it is a real story first. It has one clear problem, a small cast, events told in the order they happen, dialogue that sounds like people talking, and an ending that pays off the problem. Then, and only then, its language is held to a level the reader can follow. Day and Bamford (1998) call this kind of writing language learner literature and treat it as a genre in its own right, like children's literature.
Publishers of graded readers already work this way. Claridge (2012) interviewed commissioning editors at four major publishers. One told her its stories must be linear, with no flashbacks, no confusion and no multiple narrators. The editors also agreed that a level list is a guide for writers, not a cage. A thriller at a 700-headword level may still need 'kill', 'die' and 'shoot'.
The CEFR describes the reader, not the lesson. The Council of Europe (2020) says an A2 reader can understand short, simple stories about familiar, concrete situations in high-frequency everyday language. At B1, the reader can follow stories with a clear linear storyline. Neither descriptor mentions a grammar point. Both mention a story. So before you think about words, check that your draft has five things:
- One clear problem. Someone wants something and cannot have it yet. The reader knows what it is by the end of page one.
- Few characters. One main character and one or two others, with names that start with different letters.
- Events in order. One scene after another, in time order. No flashbacks, no second narrator.
- Dialogue that sounds spoken. Short turns, interruptions, questions answered with questions.
- An ending that pays off. The problem is solved, or clearly not solved, by something the main character does.
What the research says about language control
Language control means keeping most of the words and structures inside what the reader already knows. The research gives you a target. Hu and Nation (2000) gave learners a short story with different densities of unknown words and measured comprehension. They concluded that fiction read without help needs about 98% of the running words to be known: one unknown word in fifty.
Nation and Deweerdt (2001) show what that looks like in a real book. They compared a level-two simplified version of a classic novel with the first 7,965 words of the original. In the simplified text, 98.6% of the words were in the most frequent 2,000 word families or were names. In the original, the figure was 91.3%. For a reader with 2,000 words, that is one unknown word every seven lines against one in almost every line.
Headwords are how publishers count the words a writer may use. A headword stands for a word family, so 'accident' also covers 'accidents', 'accidental' and 'accidentally'. Nation and Wang (1999) analysed 42 readers from one six-level series and found that a learner needed the vocabulary of a book's own level, not only the levels below, to know 95% of its words. What Are Graded Readers? explains headword counts in more detail.
Sentence length matters too. Crossley, Yang and McNamara (2014) had 48 learners read three authentic passages and simplified versions of each at beginner and intermediate level. The beginner versions were read faster and understood better, and the comprehension advantage held after controlling for proficiency and background knowledge. Short sentences help. The next section shows why chopping sentences is not the same as simplifying.
Why simplification done badly is worse than none
Nation and Deweerdt (2001) summarise the standard criticism of simplified texts. When writers are told to use short, simple sentences, the result can be choppy and unnatural discourse, with poor links between sentences. The reader has to guess how one idea connects to the next, because the writer cut the word that said so. Nation and Deweerdt accept that this is true of poorly written simplifications, and add that many others are a joy to read.
Here is the fault in miniature. 'He opened the door. He saw the dog. He was afraid. He ran.' Every sentence is easy, and the passage is hard, because the reader must supply 'because' and 'so'. 'He opened the door and saw the dog. He was afraid, so he ran.' The same words plus two connectors, and it reads like English.
The second fault is thinness. Claridge (2012) reports the view of David Hill, a long-time reviewer of graded readers, that many original stories for learners are thin, with simpler plots and weaker characters than adapted classics. Thinness comes from writing the language first. Start from a word list and you get sentences that fit the list. Start from a character with a problem and you get a story, then fix the words.
The third fault is the disguised lesson. If every sentence exists to show the past simple, the reader can tell. A story about last summer will be full of past tenses without you forcing them. How Is a Story Levelled to the CEFR? shows how editors judge a finished story, and none of their checks reward a grammar drill.
A step-by-step method
The method separates the two jobs, story and language, so that neither spoils the other. Use a level list to find the words you did not notice were hard, then decide, word by word, whether the story needs each one.
- Plan the story in five lines. Who is the main character, what do they want, what stops them, what do they do about it, how does it end. No five lines, no story yet.
- Write it freely. Write the whole story in the English that comes naturally, without stopping to check words. You need a complete, living draft before you control anything.
- Check the words against a level list. Paste the draft into a vocabulary profiler, or check it by hand against a published list for the level. Mark every word outside the level and ignore names. Writing at A2: Vocabulary and Grammar Control explains which public lists exist.
- Rewrite the hard parts, keeping the text whole. Work passage by passage, not word by word. Replace a rare word with a known one, show it in a context that carries it, or say the same thing another way. Keep the connectors and the rhythm.
- Read it aloud. Every sentence you stumble on, a learner will stumble on. Dialogue you cannot say in one breath is too long.
- Test it on a learner. Give it to one reader at the target level. Ask them to count the unknown words on one page, then to tell you what happened.
Common faults and how to fix them
These are the faults that come up most often when a draft is checked. Each has a fix that keeps the story intact.
| Fault | What it looks like | Fix |
|---|---|---|
| Too many characters | Five names in two pages; the reader cannot remember who is who. | One main character and one or two others. Cut or merge the rest. |
| Events out of order | A flashback on page three explains what happened before page one. | Tell it in time order. If the past matters, start the story there. |
| Choppy sentences | 'He ran. He was scared. He fell.' | Join with 'and', 'but', 'because', 'so'. Vary sentence length so the passage has a rhythm. |
| Textbook dialogue | 'Hello. How are you? I am fine, thank you.' | Give each line a purpose. Let people interrupt, disagree and answer a question with a question. |
| A rare word carrying the plot | The key moment turns on 'verdict' or 'ransom', which the reader does not know. | Use a known word, or show the new word three times in clear contexts before it matters. |
| Pronoun tangles | 'He told him that he had seen his brother.' | Repeat the names. Learners do not mind repetition; they mind confusion. |
| The ending explains the moral | 'And so Ali learned that honesty is always best.' | Cut the sentence and let the last action show it. See How to Write a Moral Without Preaching. |
One passage before and after language control
Here is a passage of my own, written freely and then controlled for a low level, around A2. The story is the same in both versions.
Before language control (78 words)
Mara had been dreading the recital for weeks, and now, as the audience settled into an expectant hush, her fingers trembled above the keys. The hall smelled of polish and old velvet. She glanced at her grandmother in the third row, who had taught her the piece on a battered upright piano, and remembered the promise she had made: whatever happened, she would finish. She took a breath, ignored the stifled cough from somewhere behind her, and began.
After language control (84 words)
Mara was afraid of the concert. She thought about it every day for three weeks. Now the hall was quiet, and everyone was waiting. Her hands were above the piano, and they were shaking. She looked at her grandmother in the third row. Her grandmother taught her this music at home, on an old piano. Mara made her a promise: 'I will finish. I will not stop.' Somebody behind her coughed. Mara did not look. She took a breath and started to play.
What changed. 'Dreading' became 'was afraid of' and 'recital' became 'concert'. The past perfect became the past simple. One 41-word sentence became four short ones, joined with 'and' where the ideas belong together. The promise moved into direct speech, because speech is easier to follow than a reported thought. What stayed: the grandmother, the old piano, the cough, the breath. 'Coughed' may sit above the level, but the context carries it: a sound, behind her, that she ignores.
Testing the story on a learner
No list can tell you what one reader knows. Give a learner at the target level one page from the middle of the story. Ask them to count the words they do not know, then to tell you what happened. Two unknown words in a hundred is the 98% that Hu and Nation (2000) found necessary. Five or six means the level is wrong. A reader who counts two but cannot retell the page has hit a plot turn that depends on something they missed, so make it plainer.
If the story will be recorded, test the recording too, because a pace that suits a fluent reader may be too fast for a learner. How to Narrate a Story for Language Learners covers speed, pauses and clarity.
How iRead handles this
iRead takes original stories from writers, and its editorial team levels each accepted story by CEFR band. Narrators record themselves reading with a phone, and the recording is transcribed and synced to the text automatically. Writers and narrators are paid for accepted work and credited on the story.
The core principle: A story for learners is a real story with its language controlled. Write the story first, then hold the words to the level.
Write for iRead: Writers submit original stories, which iRead's editorial team levels by CEFR band. Writers are paid for accepted work and credited on the story.
Apply to write for iReadKey takeaway
Plan the story in five lines, write it freely, check the words against a level list, rewrite whole passages, read it aloud, and test it on one learner.
Frequently asked questions
How do I write a story for English learners?
Write a real story first: one problem, few characters, events in order, spoken dialogue, a clear ending. Then check the words against a list for the level, rewrite the hard passages so the text stays natural, read it aloud, and test it on one learner at that level.
How many unknown words can a learner story contain?
Hu and Nation (2000) concluded that reading fiction without help needs about 98% known words, or one unknown word in fifty. On a page of a hundred words, that is two unknown words. Names do not count, and a word repeated in clear contexts stops being unknown.
Should I use short sentences?
Shorter than in adult fiction, yes, but not chopped. Nation and Deweerdt (2001) note that short, simple sentences can produce choppy, unnatural text with missing links. Keep 'and', 'but', 'because' and 'so', and vary the length so the passage has rhythm.
Is it better to write an original story or adapt a classic?
Both can work. Claridge (2012) found that publishers disagree: one favoured originals because content can be controlled with the language, while others adapt classics. Claridge also reports a reviewer's view that many originals are thin. Write the story you can make real, then control it.
Sources
- Day, R. R., & Bamford, J. (1998). Extensive Reading in the Second Language Classroom. Cambridge University Press. www.cambridge.org/9780521568296
- Claridge, G. (2012). Graded readers: How the publishers make the grade. Reading in a Foreign Language, 24(1), 106-119. doi.org/10.64152/10125/66668
- Nation, I. S. P., & Deweerdt, J.-P. (2001). A defence of simplification. Prospect, 16(3), 55-67. eric.ed.gov/?id=EJ637434
- Crossley, S. A., Yang, H. S., & McNamara, D. S. (2014). What's so simple about simplified texts? A computational and psycholinguistic investigation of text comprehension and text processing. Reading in a Foreign Language, 26(1), 92-113. doi.org/10.64152/10125/66686
- Hu, M. H.-C., & Nation, P. (2000). Unknown vocabulary density and reading comprehension. Reading in a Foreign Language, 13(1), 403-430. doi.org/10.64152/10125/66973
- Nation, P., & Wang, K. M.-T. (1999). Graded readers and vocabulary. Reading in a Foreign Language, 12(2), 355-380. doi.org/10.64152/10125/66966
- Council of Europe (2020). Common European Framework of Reference for Languages: Learning, teaching, assessment. Companion volume. Council of Europe Publishing. rm.coe.int/common-european-framework-of-reference-for-languages-learning-teaching/16809ea0d4