Pitching an Arabic reading text at the right level

What actually makes an Arabic text hard, the unknown-word percentages worth aiming at, and how to bring a real text down a level without turning it into Arabic nobody writes.

5 minute read Updated

The short version

  • For reading on their own, learners need to know about 98% of the words. For a text you will work through together, about 90–95%.
  • Sentence length and subordination make Arabic harder faster than vocabulary does.
  • Tashkīl is a difficulty dial you control independently of everything else. Use it.
  • Simplifying by shortening is safe. Simplifying by substituting rare words for common ones is where texts stop sounding like Arabic.

The two different jobs a text can do

Before levelling anything, decide which of two things the text is for, because they want opposite settings.

Intensive reading is the text you work through as a class: short, dense, deliberately containing things the learners cannot yet do, with you there to explain them. Around 90–95% known vocabulary is right — roughly one unfamiliar word in every ten to twenty. Below 90% the lesson becomes a translation exercise and comprehension stops.

Extensive reading is the text a learner takes away and reads for meaning, at length, without a dictionary. Here the number is much higher — about 98%, or one unknown word in fifty. That feels absurdly easy to a teacher and it is exactly right: the point is volume and fluency, and anything that forces a stop kills both.

Most of the disappointment with reading in Arabic classrooms comes from using an intensive text for an extensive purpose. A page with fifteen new words is not "challenging homework"; it is homework nobody will finish.

What makes an Arabic text hard

Roughly in order of how much difficulty they add:

Unknown vocabulary. The dominant factor, as in any language. Worth measuring rather than estimating — teachers consistently underestimate how much of their own text is unfamiliar.

Sentence length and subordination. Formal Arabic prose tolerates very long sentences, chained with و and ف and layered with relative clauses on الَّذِي. A learner who can read every word in a sixty-word sentence may still not be able to say what it claims. This is the difficulty that surprises people.

Root familiarity. A word can be unknown and still guessable if its root is known and its pattern is recognised. Two texts with the same unknown-word count can differ enormously here, which is why raw counts mislead.

Absent tashkīl. Discussed at length in its own guide. It roughly doubles the work for anyone not yet reading fluently.

Register and variety. A news editorial, a hadith, a children's story and a WhatsApp message are four different languages as far as a Year 9 is concerned. A text can be short, vowelled and entirely beyond a class because of what it is.

Cultural assumption. A text that assumes you know what happens at عِيد, or who a given figure is, is harder than its vocabulary suggests — and this cuts both ways in a mixed class, where some pupils have the background and some do not.

How to actually measure it

Take the text and your class vocabulary list, and count. Not impressionistically — actually count, once, for a text you are unsure about; it recalibrates the guesswork for everything after.

  1. Count the running words in the passage, not the distinct ones. A word repeated eight times is eight words for the reader's purposes and one item for yours.
  2. Mark every running word that is not on the class list and not transparently derivable from a known root.
  3. Divide. That is your unknown percentage.

Then look at the longest sentence and count its words. Anything past about twenty-five in a text for beginners or intermediates deserves a second look regardless of vocabulary, and if a single sentence carries three subordinate clauses it will be the sentence the whole class stalls on.

Automatic readability tools help with the first measurement and are weak at the second and third. Treat the number as a screen, not a verdict.

Bringing a text down a level

In rough order of how safe they are:

Split the sentences. The safest edit available. A sixty-word sentence becomes three twenty-word sentences with almost no loss and a large gain. Arabic takes to this well because the coordinating particles that chain clauses can usually just become full stops.

Add tashkīl. Costs nothing in authenticity, because vowelled text is a real register rather than a learner dialect. Vowel the words you expect to be new and leave the rest.

Cut, do not paraphrase. Removing a whole sentence or paragraph leaves what remains genuine. Rewriting a sentence in easier words leaves something that may not be Arabic anybody would write.

Gloss in the margin. Six words defined beside the text, in Arabic if possible, will lower difficulty more reliably than any amount of rewriting, and it leaves the text intact.

Substitute vocabulary — carefully. Replacing a rare word with a common synonym is fine when a real synonym exists. It often does not; Arabic's near-synonyms carry register differences that a substitution flattens, and the result reads as translationese. If you are substituting more than a handful of words, you are writing a new text, which is a legitimate thing to do — just do it deliberately rather than by accident.

What to avoid: stripping the text of every structure the class has not met. A text with no unknown grammar in it gives a learner nothing to grow into, and reading is one of the main ways grammar is acquired rather than taught.

Differentiating without three worksheets

The same passage can carry three levels at once, which is a great deal less work than three passages.

  • Vary the tashkīl, not the text. Fully vowelled, partially vowelled and bare versions of one passage are one file and three levels, and nobody can tell from across the room which they have.
  • Vary the glossary. Same text, twelve glossed words for one group and four for another.
  • Vary the questions. Retrieval questions, inference questions and evaluative questions on one passage cover a wide range without anyone reading a different thing.
  • Vary how much they read. "Everyone reads the first two paragraphs; go on if you finish" is unglamorous and works.

Where to find texts

Authentic material is easier to source than it was and still needs filtering. News sites give you current, well-written prose at a level that is genuinely hard for school learners — usable in extracts, rarely usable whole. Children's literature is often closer to the right level but assumes cultural knowledge and a vocabulary of the domestic that a foreign-language learner has not been taught. Graded readers written for Arabic learners are the most directly usable and the thinnest on the ground.

Two practical notes. Check the variety before you use anything found online: a lot of what reads as Arabic prose is a dialect in written form, which is fine if you meant it and confusing if you did not. And check what happens to the text when you paste it — copying Arabic out of a PDF frequently reverses or mangles it, and it is worth reading the result before it reaches thirty pupils.

Scroll to Top