CEFR levels are useful when reviewing vocabulary, grammar and sentence complexity. They do not tell a teacher whether students understand a cultural reference, recognise an example or share the background knowledge assumed by the writer.

This distinction matters in CLIL because subject texts often rely on more than language. A computing article may mention a novel, a science text may assume familiarity with a local environmental issue, and a history source may use political references that are obvious to some students and opaque to others.

Three types of difficulty

When students struggle with a source, it helps to separate language difficulty, subject difficulty and contextual difficulty. Language difficulty concerns vocabulary and structure. Subject difficulty concerns the concept itself. Contextual difficulty concerns examples, references and assumptions that the writer does not explain.

The three often overlap. A student may understand every word in a sentence and still miss its point because the example is unfamiliar. Another may know the example but lack the technical concept needed to interpret it. Planning improves when the teacher identifies the actual barrier instead of lowering the whole text.

Classroom example: AI bias and The Great Gatsby

Fourteen students in class 3 AINF studied how bias can enter AI systems through training data. They compared a 416-word B1 version with a 603-word B2 version and completed an 11-question response task about meaning and difficulty.

The source referred to The Great Gatsby. Nine students made a broadly appropriate connection between the cultural source and the risk of bias in training data. Five could not explain it. The adapted language had not solved the problem because the missing information was cultural, not grammatical.

The episode was useful for the subject. Training data reflects what is included, excluded and overrepresented. The difficulty with the reference gave the class a concrete experience of unequal access to background knowledge, which connects directly to the topic of bias.

Review cultural assumptions as well as language

A teacher can ask AI to list every name, cultural reference, metaphor and example in a source, then explain what prior knowledge each one assumes. It can suggest a short gloss, a more internationally accessible example or a question that helps students investigate the reference.

The decision to keep, explain or replace a reference remains pedagogical. If the reference is incidental, a brief explanation may be enough. If it contributes to the subject goal, it may deserve explicit teaching. If it distracts from the lesson and adds no value, replacing it can be sensible.

Questions to add to text review

  • Which people, places, events, texts or institutions are mentioned?
  • What does the writer expect the reader to know already?
  • Will the example be equally accessible across the class?
  • Is the reference important to the content objective?
  • Should I explain it, replace it or turn it into part of the lesson?

In the AI-bias lesson, nine of fourteen students connected the cultural reference to training-data bias; five could not explain the reference. The CEFR level alone did not account for that difference in access.