A difficult topic is not always difficult because the topic itself is difficult.

Sometimes the learner is standing on an unstable floor.

Secondary algebra becomes hard because fraction operations are slow.

Differentiation becomes fragile because algebraic manipulation is unreliable.

Inference becomes inconsistent because vocabulary access is too weak to preserve the sentence relationship.

Science explanation fails because the concept is known but cause-and-effect relationships are not yet stable enough to support precise language.

Training dependency is the deliberate identification and stabilisation of a prerequisite capability because a later skill currently depends on it strongly enough that weakness below is limiting performance above.

The important word is dependency.

Not every older topic deserves review.

Only the foundation that is currently constraining the next useful performance should become an active repair target.

---

Quick Read: Find the Floor Under the Skill

Target Skill → Identify Required Components → Test the Suspected Dependency → Repair if Needed → Reconnect → Retest the Target Skill

The key questions are:

• What does this skill require the learner to already know or do?
• Which prerequisite is actually failing?
• Is the dependency missing, slow, inaccurate or hard to retrieve?
• Does repairing it improve the target skill?
• Can the learner now perform the higher-level task without extra support?

Dependency Is Different From Chronology

Something taught earlier is not automatically a prerequisite.

And something taught later is not automatically independent.

Training dependencies are structural, not merely chronological.

If Mira struggles with quadratic equations, reviewing every Primary Mathematics topic would be wasteful.

But if factorisation is unstable, that is highly relevant.

If Jonas struggles with Secondary comprehension, revisiting all Primary English may be unnecessary.

But if pronoun reference repeatedly breaks sentence meaning, that dependency matters immediately.

The question is not:

What came before?

It is:

What must currently work for this higher-level task to work?

Prior Knowledge Changes New Learning

A 2025 review in Learning and Individual Differences synthesised sixteen mechanisms through which prior knowledge can influence learning, including encoding, comprehension, chunking, cognitive load, metacognition, transfer and interference.

The important lesson is not simply “more prior knowledge is always better.”

Prior knowledge can help, hinder or interact with instruction differently depending on the learner and the task.

For training, this means a suspected dependency should be tested rather than assumed.

The learner may have enough of the prerequisite already.

Or the prerequisite may be present but too slow to support the higher-level task.

Dependency Can Fail in Different Ways

Missing.
The learner never acquired the prerequisite.

Fragile.
The learner can perform it sometimes but not reliably.

Slow.
The skill is accurate but consumes too much attention.

Context-bound.
The learner can use it in isolation but not inside the target task.

Interfered.
A competing rule activates instead.

Unrecognised.
The learner knows the prerequisite but does not notice when it is required.

Each failure mode needs a different repair.

Mathematics Dependencies

Mathematics is especially dependency-rich.

Later topics repeatedly reuse earlier structures.

• fractions support algebraic fractions;
• equality supports equation solving;
• ratio and proportion support similarity and rates;
• factorisation supports quadratic equations;
• algebraic manipulation supports trigonometry and calculus;
• graphs support functions, rates and modelling.

If a higher topic is failing, trace downward only until the bottleneck becomes visible.

This is closely related to Training Decomposition.

Decompose the task.

Then ask which component depends on an unstable foundation.

Mira: Differentiation Fails Because Algebra Is Expensive

Mira understands differentiation rules.

But long questions collapse during simplification.

The tutor isolates two stages.

• Can Mira choose and perform the differentiation rule?
• Can she simplify the resulting algebra accurately?

The first is stable.

The second is not.

Therefore more differentiation worksheets are inefficient.

The immediate dependency is algebraic manipulation.

Repair it briefly.

Then return to differentiation.

If differentiation performance improves, the dependency hypothesis is supported.

English Dependencies

English dependencies are less visually obvious but equally real.

• vocabulary supports sentence comprehension;
• pronoun reference supports passage coherence;
• sentence relationships support inference;
• idea development supports paragraph quality;
• paragraph architecture supports whole composition;
• reading fluency can support attention available for interpretation.

Jonas may appear weak at inference when one key vocabulary item is actually blocking the passage model.

Or vocabulary may be fine while pronoun reference repeatedly breaks who did what to whom.

Do not label the whole child “weak at comprehension.”

Trace the dependency.

Jonas: Writing Looks Weak Because Ideas Are Thin

Jonas’s grammar is accurate.

His vocabulary is adequate.

But paragraphs remain repetitive.

The tutor asks him to explain the idea orally before writing.

The oral explanation is also thin.

The writing problem is not sentence control.

The dependency is conceptual development.

Train claim → explanation → example → consequence at idea level.

Then return to full paragraphs.

Science Dependencies

Science questions often combine concept knowledge, evidence reading and explanation.

A learner can fail at any layer.

• concept knowledge supports prediction;
• variable reasoning supports experimental interpretation;
• graph reading supports evidence use;
• causal relationships support explanation;
• precision of language supports communication of the reasoning.

If Nadia cannot explain an experiment, first test whether she understands the underlying concept.

If she does, test whether she can read the evidence.

If she can, test whether answer construction is the final bottleneck.

The training target should sit at the lowest unstable dependency that is actually constraining the higher performance.

Nadia: The Concept Is Known but the Evidence Dependency Is Weak

Nadia can explain photosynthesis accurately from memory.

But in an unfamiliar experiment she ignores the measured data and writes the memorised explanation.

The missing dependency is not content knowledge.

It is evidence integration.

Train data → pattern → claim before returning to full explanation.

Now the higher-level answer can use both concept and evidence.

Dependencies Can Be Conceptual, Procedural or Regulatory

Conceptual dependency: later understanding requires an earlier idea.

Procedural dependency: later performance assumes an earlier routine is sufficiently fluent.

Representational dependency: the learner must read or transform a representation before using it.

Language dependency: vocabulary or syntax must be understood before the content relationship can be recovered.

Self-regulatory dependency: the learner must initiate, plan, monitor or check independently for the task to succeed.

Dependencies therefore extend beyond syllabus maps.

Dependency Graphs

A useful mental model is a dependency graph.

The target capability sits at the top.

Below it sit the components required for success.

Below those sit their prerequisites.

Do not repair the entire graph.

Find the first unstable node on the active path.

This idea is increasingly important in educational technology as well. A 2025 ACM Computing Surveys review of prerequisite-relation learning describes prerequisite relations as dependencies among knowledge concepts and learning resources that can support content organisation and personalised learning paths.

The human teaching translation is simple:

Teach the dependency that is blocking the learner—not every dependency that exists in theory.

Dependencies and Training Priorities

A dependency can be weak without being the current priority.

Mira may have slightly slow fraction arithmetic.

If the current quadratic task does not depend heavily on it, the weakness may not deserve immediate repair.

But if algebraic fractions begin next week, the same weakness may become urgent.

Dependency and priority interact.

See Training Priorities.

Dependencies and Training Readiness

Readiness is partly a dependency question.

Can the learner carry the prerequisite load required by the next task?

If not, the learner may need pre-training or a smaller task.

This connects to Training Readiness.

But do not use “not ready” as a permanent label.

Find the dependency that would change readiness fastest.

Dependencies and Training Transitions

Stage transitions reveal hidden dependencies.

Primary Mathematics may allow slow fraction work to survive.

Secondary algebra may make it expensive.

Primary English may tolerate teacher-supported planning.

Secondary work may make self-regulation a dependency.

This is why Training Transitions should include a dependency audit before the new stage accelerates.

Dependencies and Training Interference

A prerequisite can be present but unreliable because a competing rule keeps winning.

Then the repair is not simple relearning.

It may require Training Interference: separate the competing rules, strengthen the discriminating cue, then recombine them.

Do Not Overrepair Foundations

A learner struggles with Secondary Mathematics.

The response is to send the child back through years of Primary worksheets.

This can waste time and damage motivation.

Instead, test the suspected dependency directly.

If it is stable, move upward.

If it is weak, repair the smallest relevant unit.

Then return to the target skill quickly.

The foundation exists to support the building.

Do not spend the entire renovation underground.

Use the Dependency Test

A simple experimental method is:

Remove or Support the Suspected Dependency → Retest the Higher Skill

If Mira fails a geometry problem, temporarily provide the algebra.

Does the geometry reasoning now succeed?

If Jonas fails comprehension, provide the meaning of the suspected vocabulary item.

Does the inference now succeed?

If Nadia fails an experimental conclusion, identify the variables for her.

Can she now reason from evidence?

This does not prove causation perfectly, but it gives useful diagnostic evidence about where the chain is breaking.

Dependencies and Training Recombination

Once the prerequisite is repaired, put it back inside the target skill.

This is Training Recombination.

Fraction fluency should return to algebraic fractions.

Vocabulary should return to passage interpretation.

Variable reasoning should return to full investigations.

The dependency has done its job only when the higher capability improves.

Dependency Maps Should Shrink Over Time

When a foundation becomes stable, it should stop consuming active training attention.

Let normal use maintain it.

A learner who has repaired fraction fluency does not need permanent fraction drills if later Mathematics keeps the skill alive.

This connects to Training Maintenance and Exit Criteria.

Failure Mode: Foundation Repair Becomes the Whole Programme

The learner spends months “strengthening basics” but rarely returns to the actual target skill.

Repair:

define a dependency exit criterion and recombine quickly.

Failure Mode: The Wrong Dependency Is Chosen

Jonas struggles with composition.

He receives grammar worksheets.

Grammar was already stable.

The actual dependency was idea development.

Repair:

use evidence from the whole performance, not assumptions about what weak students usually need.

Failure Mode: Dependency Is Present but Too Slow

Mira can manipulate fractions correctly but needs so much attention that the higher algebra collapses.

This is not a conceptual gap.

It is a fluency dependency.

Use short correct repetition until the operation becomes cheaper, then retest the higher task.

Failure Mode: Dependency Is Strong in Isolation but Missing in Context

Nadia can identify variables on a dedicated worksheet.

Inside an unfamiliar investigation, she does not recognise that variable analysis is needed.

The dependency is not missing.

Recognition is missing.

Use mixed and varied cases so the prerequisite becomes available in context.

A Parent Dependency Audit

• What higher skill is currently failing?
• What does it genuinely depend on?
• Has the suspected prerequisite been tested directly?
• Is the dependency missing, fragile, slow or context-bound?
• Does repairing it improve the target skill?
• Are we reviewing more old material than necessary?
• When can this foundation leave active training?

A Tutor Dependency Audit

• What is the target capability?
• Which component first fails?
• What prerequisite does that component require?
• Can I support or remove the suspected dependency temporarily to test the hypothesis?
• What is the smallest repair?
• How soon will I recombine?
• What evidence will show the dependency is no longer the bottleneck?

The Deeper Idea: Advanced Performance Is Built on Invisible Cheap Foundations

Experts appear to do complex things directly.

But many lower-level operations have become stable enough to disappear from conscious attention.

The algebra is available.

The vocabulary is available.

The sentence relationship is available.

The variable logic is available.

When those dependencies fail, higher reasoning becomes expensive.

Good training does not worship foundations for their own sake.

It keeps the necessary ones stable enough that the learner can spend attention higher up the chain.

A prerequisite deserves active training only while its instability is still limiting something more valuable above it.

Research Foundations

Useful recent anchors include the 2025 review of sixteen mechanisms through which prior knowledge affects learning and the 2025 ACM Computing Surveys review Prerequisite Relation Learning: A Survey and Outlook, which examines how prerequisite relationships among concepts and learning objects can support educational sequencing and personalisation. These literatures should not be reduced to rigid prerequisite trees. The practical principle here is diagnostic: identify the specific dependency currently constraining performance, repair it at the smallest useful scale, and verify the repair by returning to the higher-level task.

Continue Through How Training Works

Read this alongside Training Transitions, Training Readiness, Training Priorities, Training Decomposition, Training Recombination, Training Interference and Training Maintenance.