Dates but not causes

Partly right, and too simple to act on. Test it against the three-way CLIL frame - content, language, cognition.

• Content: they can list dates and estates, so the content was accessed and stored. The failure is not missing information.
• Cognition: the test asked for causal explanation; the lesson asked 'what happened?'. The activity was recall and the assessment was reasoning. That mismatch alone would produce this result in a learner's first language.
• Language: the textbook's language (nominalisation, long noun phrases, embedded clauses) is well above B1, so comprehension was partial. More importantly, the language needed to express causation - 'this led to', 'as a result', 'because they were excluded' - was never taught. Learners have the causes and no way to phrase them.
• The four copied answers are not dishonesty. They are what a learner does when they can recognise the right paragraph but cannot generate the sentence.

So: language is a real part of the problem, but 'their English' locates the fault in the learners and skips the two things the teacher controls - the difficulty of the text and the fact that the language of causal reasoning was never taught.

Separate the three language strands rather than treating 'English' as one problem.

• Language of learning (content-obligatory vocabulary): estate, privilege, tax, harvest, bread price, assembly. Pre-taught with visuals, not glossed in passing.
• Language for learning (the classroom and thinking language): because, therefore, this led to, as a result, one consequence was. Taught explicitly with a substitution table on the board.
• Language through learning: the phrasing they need and don't yet have, captured as it comes up in the lesson and kept on a visible 'useful language' board.

Scaffolds:

• Split the text. Jigsaw: four groups each get one cause (financial crisis, food scarcity, inequality of the estates, Enlightenment ideas) in a simplified version, then re-form groups of four and teach each other. This reduces the reading load to B1 and makes the language necessary.
• Cause–consequence organiser: boxes with arrows, filled before writing.
• Model answer, one paragraph, annotated for the causal connectives.
• Sentence frames for the weakest: 'The Third Estate were dissatisfied because ___. This led to ___.'

Product: first an oral explanation to a partner using the organiser, then a five-sentence causal paragraph. Speaking before writing, because the paragraph is where the B1 ceiling bites hardest.

Two rubrics, marked separately, then read together.

• Content rubric: are the causes correct, and are the causal links between them accurate? A learner who writes 'the Third Estate were angry because bread cost too much' with grammar errors scores full content marks.
• Language rubric: use of causal connectives, range of topic vocabulary, comprehensibility - but not general accuracy, which is not what this unit is for.

Reasoning: if language is folded into the content mark, you cannot tell a learner who does not understand from a learner who understands and cannot say it. That distinction is the whole point of CLIL assessment, and the four copied answers are exactly the case where it matters - the current test cannot tell you which of the two they are.

Caveat to state openly: separating the rubrics does not remove the confound. A learner who genuinely cannot express causation may still be unable to demonstrate understanding in this format. So add one low-language check of the same content - an annotated timeline or a card-sort of causes and consequences - so that content understanding has a route to the surface that does not depend on B1 writing.