Three Ways That Curriculum Fails

Why students fail to make progress when curriculum lacks specificity, systematicity and subject sensitivity

As a general principle, children fail to make progress through the curriculum for three reasons. The curriculum is insufficiently specific, insufficuently systematic, or insufficiently subject sensitive.

Obviously,, these are not the only things that matter, but they are - I think - the most important becaue they are all things we can work to ameliorate. If we can’t specify what is to be learned, if we don’t think carefully about sequence and return, or if we ignore the distinctive structure of subjects, then even the best intentions tend to descend into vagueness, incoherence and unhelpful impositions. We may cover content, but coverage is not progress. In fact, coverage is often inimical to effective instruction.

1. Insufficient specificity

If you’re unable to specify something, you’re unlikely to be able to teach it. And if you don’t teach something, it shouldn’t be a surprise that many students fail to learn it.

This ought to be obvious, but it’s astonishing how often curriculum planning proceeds as if broad aims were enough. We wave at lofty ambitions and say we want children to become better readers, more confident writers, more critical thinkers. But what do these things actually consist of? Wanting students to be more analytical, critical or creative is fine as far as it goes, but it doesn’t go very far. Such ambitions are too vague to guide instruction. How can teachers teach such ephemera as “confidence” or “understanding texts more deeply” unless these are broken down into concrete knowledge, concepts and procedures?

If we can’t describe what children need to know, what they need to be able to do, and what they need to understand about the relationship between the two, then teaching descends to guesswork. Assessment is even worse with students being judged on their ability to do things they’ve not been taught to do, rather than attending to the precise components from which skilled performance emerges.

In practice, insufficient specificity reveals itself in curriculum language that sounds impressive but says nothing. Statements such as “explore a range of texts” or “develop inference skills” may gesture towards something worthwhile, but they don’t tell us enough. Which texts? What kind of inference? Based on what knowledge? Using what vocabulary? Through what sequence of examples and practice?

What’s needed is the ability to decompose large, abstract goals into teachable constructs. By this I mean components of learning that can be named clearly, modelled explicitly, practised deliberately and checked with precision. They need to be small enough to teach but substantial enough to matter. If a curricular goal can’t be broken down in this way, it can’t be taught.

This is harder than it sounds because most of the things we value in education are complex performances rather than single skills. Reading with insight, writing with control, reasoning with subtlety are not indivisible capacities that can somehow be poured directly into children. They’re composite achievements, built from layers of knowledge, habit and procedural competence. The work of curriculum must be to identify the parts from which that performance is made.

One useful way to do this is to think in terms of precursors and successors. A precursor is the knowledge or capability that needs to be in place for new learning to make sense. A successor is what that learning later makes possible. Curriculum content never exists in isolation. Every element sits in relation to what has gone before and what might come next. Ignore those relations and coherent teaching becomes much harder.

This diagram above illustrates what it means to decompose a curricular goal into teachable components. At the centre sits the construct itself, judiciously selecting vocabulary. On the left are some of its precursors: spelling strategies, morphological knowledge, etymological knowledge, knowledge of synonyms and knowledge of word classes. This is the prior knowledge that make it possible to select words judiciously. If students don’t know how words are built, what they mean, how they function grammatically or how they can be spelt and distinguished from near alternatives, then asking them to choose vocabulary - judiciously or otherwise - is little more than wishful thinking. On the right are some likely successors: writing a range of texts, reading a wide range of texts, and developing knowledge of pragmatics and semantics. In other words, once students can select vocabulary with care and precision, this supports more mature reading, more controlled writing and a deeper understanding of how meaning shifts according to context and use. Curricular content should be understood relationally. A construct is shaped by what needs to come before it and by what it later makes possible.

Take reading. Suppose we want students to infer a character’s motives from a story. That may sound like one thing, but it rests on a whole set of precursors. Students need to decode fluently enough that the mechanics of reading don’t swamp meaning. They need sufficient vocabulary to grasp the literal sense of the passage. They need enough syntactical knowledge to follow the relationships within and across sentences. They need background knowledge of the world and often of genre. They need to notice significant details and hold them in mind. And they need to understand that character can be revealed indirectly through speech, action, description and the responses of others. Only then does inference become teachable in any meaningful sense.

And even then, what’s being taught isn’t some generic, context-free skill of inference. It’s something much more precise: how to move from textual evidence to a warranted claim. Once that becomes secure, it opens up successors. Students may compare the way different writers construct character. They may track changes in motive across a text. They may detect irony, ambiguity or tension. But these later achievements depend on the earlier components having been made visible, taught explicitly and practised to the point of mastery.

The same is true of writing. Consider the common ambition that students should write analytically about a text. Again, this sounds reasonable enough, but it’s far too broad to guide instruction unless it is decomposed. Analytical writing depends on many precursors. Students need relevant knowledge of the text. They need the vocabulary to name what they notice. They need sentence-level control sufficient to express a clear claim. They need to understand the difference between retelling and analysing. They need to know that quotations are evidence, not decoration. They need some grasp of the grammatical and rhetorical structures that make analytical prose possible: subordination, apposition, causal explanation, concessive phrasing, the ability to integrate reference to methods without losing clarity.

Once the construct is carefully specified, it becomes possible to teach. But this only happens if we are precise about what the component parts actually are. Take something as apparently straightforward as using evidence in literary analysis. It’s easy enough to tell students to support their ideas with quotations, but that instruction conceals at least two distinct demands: selecting meaningful evidence from the text and embedding that evidence fluently within a sentence. Of these, selection is the more demanding task. Students tend to reach for lines they happen to remember, or ones that feel obviously important, rather than choosing the textual detail that best supports the point they are trying to make. Embedding, by contrast, is more technical. It can usually be improved through clear modelling and repeated practice.

So instead of setting the vague task of “write an analytical paragraph”, we can specify a much smaller construct: select and embed relevant textual detail. That construct can then be taught directly. Students can be shown how to choose short, relevant quotations, how to avoid lifting large chunks of text, how to embed quotations naturally within a grammatical sentence, how to use quotation marks correctly, and how to comment on the effect of the quoted language. They can also be warned away from empty formulations such as “This quote shows…” and taught alternatives that do more analytical work.

At this point, modelling becomes crucial. Suppose the point for which we require evidence is, Mark Antony’s eloquence as a speaker makes him an attractive and compelling character. Students now need to find evidence from Julius Caesar that genuinely supports that claim. A novice might reach for, “I am no orator, as Brutus is, / But, as you know me all, a plain blunt man…” because it sounds memorable and clearly belongs to Antony’s speech. But when we examine it more closely, it does less than we might hope. It is a strategic disclaimer, certainly, but not the clearest example of eloquence. That matters because one of the hidden parts of this construct is learning to test a quotation against the claim being made. Does it really support the argument, or is it merely adjacent to it?

A better choice is Antony’s repeated phrase, “honourable men”. Here students can be shown how to notice that repetition, test its usefulness, and decide that it more convincingly supports the point. The phrase allows Antony to appear respectful while steadily hollowing out Brutus’s credibility. It reveals rhetorical control, irony and manipulation all at once. Once that evidence has been selected, the next step is to embed it fluently: When Antony says, ‘honourable men,’ it reveals his growing irony and manipulation, as the phrase becomes increasingly hollow with each repetition. That sentence does not merely drop in a quotation. It integrates it into a grammatical structure and immediately comments on its effect.

Or take something apparently simple like sentence structure. Specifications often say students should vary their sentences or write more sophistication. But that kind of language conceals more than it reveals. What exactly needs to be taught? One answer might be control over subordination. That’s a teachable construct. Its precursors include understanding clause boundaries, recognising the difference between complete and incomplete meaning, and knowing how subordinating conjunctions shape relationships between ideas. Its successors include being able to express cause, contrast, condition, time and concession with greater precision. In turn, that supports more nuanced argument, clearer explanation and more controlled narration.

All that said, perfect specificity is neither possible nor desirable. If we tried to specify everything, curriculum would dissolve into an infinite regress. Every construct could be broken into smaller parts, and those parts into smaller parts again. At some point the attempt to achieve total precision becomes self-defeating. Instead of clarifying what needs to be taught, it overwhelms teachers with detail and obscures the larger performance those details are meant to serve.

So the goal is not perfect specificity but sufficient specificity. The question is not, Have we specified everything? but, Have we specified enough to make effective teaching possible?

A useful test is whether what’s to be taught can be clearly modelled, practised and checked. If a curricular statement is still so broad that different teachers would interpret it in wildly different ways, then it probably isn’t specific enough. If it can’t be turned into examples, explanations, practice tasks and feedback, then it probably isn’t specific enough. If students regularly fail and we can’t tell whether the problem lies in missing knowledge, a misunderstood process or the demands of the task itself, then again, it probably isn’t specific enough.

Sufficient specificity means identifying the level of detail at which teaching decisions become clearer. It means specifying enough that we can see the likely points of difficulty, enough that we can distinguish prerequisite knowledge from the thing currently being taught, and enough that we can explain why one example has been chosen rather than another. Beyond that point, further decomposition may add complexity whilst decreasing instructional value.

Too often curriculum planning jumps straight from vague aims to broad activities. We want students to infer, so we give them inference questions. We want them to write analytically, so we ask them to produce an essay. But the end tasks are not the same as the practice required to master them. If the underlying constructs haven’t been identified, assessment merely tells who was successful despite our lack of specificity.

A specific curriculum asks

• What is the thing we’re actually trying to teach?

• What are its constituent parts?

• Which parts are prerequisite?

• Which parts can be taught now?

• What later learning will this unlock?

• Where are students likely to falter?

• What knowledge will they need in order not to?

Without sufficient specificity, curriculum remains vague, superficial and fragile. With it, teaching has a chance of becoming cumulative, intentional and effective..

2. Insufficient systematicity

Even when curriculum content is well specified, children will not necessarily make progress unless what they need to learn is arranged with sufficient care.