When a learner meets something unfamiliar, one of the most powerful questions is:

What is this like that I already understand?

That question is the beginning of analogy.

We use analogies constantly in teaching because new ideas rarely enter an empty mind. They land inside a network of older models, experiences and structures.

A function can be compared with a machine: an input enters, a rule acts, an output emerges.

An equation can be compared with a balance: valid transformations preserve equality.

A paragraph can be compared with a room: every sentence should contribute to the same space rather than opening unrelated doors.

A scientific model can be compared with a map: useful because it preserves selected relationships, incomplete because it is not the territory itself.

But analogies can also mislead.

A balance metaphor can help explain equivalent operations, but algebra is not literally a physical scale. A machine analogy can illuminate functions, but some mathematical structures do not behave like ordinary machines. A map metaphor can clarify model limits, but not every model has a simple spatial relationship to reality.

Training analogy is the deliberate mapping of a familiar relational structure onto a new problem so the learner can use old understanding to organise new understanding—while also learning where the mapping stops.

The final phrase matters.

A strong learner does not merely know the analogy.

The learner knows what the analogy preserves and what it does not.

---

Quick Read: Map Relationships, Not Decorations

A useful analogy-training sequence is:

Familiar Case → New Case → Align the Parts → Map the Relationships → Predict → Test → Identify the Breaking Point → Transfer Again

The key questions are:

• What corresponds to what?
• Which relationship is being carried across?
• Which similarities are merely surface-level?
• What does the analogy predict?
• Where does it stop being reliable?
• Can the learner use the relationship without the original analogy later?

Why Analogies Help

New problems become easier when the learner can organise them using an existing structure.

Mira understands physical balance before she fully understands algebraic equivalence.

So the tutor uses the balance as an entry point.

If the same amount is removed from both sides of a balanced scale, balance remains.

If the same valid operation is applied to both sides of an equation, equality remains.

The surface systems differ.

The relational idea transfers.

This is the central insight of structure-mapping approaches to analogy: useful analogies depend more on aligned relationships than on superficial resemblance.

A July 2026 open-access commentary in the Canadian Journal of Science, Mathematics and Technology Education applies structure-mapping theory to mathematics teaching and argues that an analogy should be judged by how well its relational structure supports the mathematical idea, not simply by how familiar or memorable the story is.

The Familiar Thing Is Not the Goal

Analogies are bridges.

A bridge matters because it helps the learner cross.

The learner should not have to remain on it forever.

If Mira can only understand equations when she imagines a physical scale, the analogy has not yet fully transferred into mathematical structure.

Eventually she should be able to say:

Equivalent transformations preserve the equality relationship.

The analogy helped construct that relationship.

Then the relationship becomes independent of the analogy.

Surface Analogy vs Structural Analogy

Two things can look similar and behave differently.

Two things can look completely different and share the same structure.

This distinction is essential.

A recipe problem and a map-scale problem do not look alike.

Yet both may preserve a proportional relationship.

A story about friendship and a newspaper-style article do not look alike.

Yet both may require the same evidence-bounded inference.

A plant experiment and an electrical experiment look completely different.

Yet both can preserve the same experimental logic:

Change → Measure → Control → Compare → Conclude

Training analogy teaches learners to see the second kind of similarity.

Analogy and Training Invariants

Training Invariants asks what stays the same as the surface changes.

An analogy gives us two different surfaces and asks whether the same relationship exists in both.

That means a good analogy is really a test of invariance.

The learner should be able to identify:

• the familiar objects;
• the new objects;
• the familiar relationships;
• the corresponding new relationships;
• the invariant structure that justifies the mapping.

If the learner only remembers the story, the analogy has remained decorative.

Analogy and Training Contrast

Analogy becomes stronger when a learner compares two candidate analogies.

Which one maps the important relationship more accurately?

Which one shares more surface features but fewer structural features?

This connects to Training Contrast.

For example, consider two analogies for a mathematical function.

Analogy A: a vending machine where one button produces one selected item.

Analogy B: a person choosing randomly from several outputs after receiving the same input.

The first can support the single-output relationship more cleanly.

The second breaks the function condition if one identical input can produce several outputs unpredictably.

Contrast teaches why one analogy fits better.

The Mapping Table

For difficult analogies, make the correspondence explicit.

Balance scale → Equation

• left pan → left side of equation;
• right pan → right side of equation;
• equal weight → equal value;
• same action on both sides → equivalent transformation;
• balance preserved → equality preserved.

This prevents the learner from keeping only the image.

The mapping becomes relational.

Mathematics Analogy: Equation as Balance

Mira has learned the school shorthand “move it across and change the sign.”

This often works procedurally.

But it can hide the mathematics.

The balance analogy repairs the relationship.

If:

3x + 5 = 17

subtracting 5 from both sides preserves equality.

Nothing literally “moves.”

The equation is transformed equivalently.

Now ask Mira where the analogy breaks.

Physical balances involve mass and gravity.

Algebraic equality is a mathematical relation.

The analogy preserves the symmetry of valid operations, not every physical property of a scale.

Mathematics Analogy: Function as Machine

The machine analogy can help students see input, rule and output.

Input x.

Rule: multiply by 2 and add 3.

Output: 2x + 3.

The 2026 Canadian mathematics-education article mentioned earlier examines teacher-made analogies for the formal definition of a function and argues that analogy quality should be judged by structural alignment and the mathematical relationships it makes available.

So use the machine analogy carefully.

Ask which parts map well.

Ask which do not.

Then leave the machine behind and return to functions themselves.

Mathematics Analogy: Graph as Story of Change

Some students see a graph as a picture to memorise.

Use the analogy of a story.

The horizontal axis tells where or when we are in the story.

The vertical axis tells what quantity is changing.

The gradient tells how quickly the relationship changes.

The turning point marks a change in direction.

Then test the analogy.

Can Mira tell the story of an unfamiliar graph?

If yes, the analogy has become a reading strategy.

English Analogy: Paragraph as a Room

Jonas writes paragraphs where sentences are individually good but do not belong together.

The tutor says:

Imagine the paragraph is one room. Does every sentence belong in this room?

This creates a structural check.

A sentence can be beautiful and still belong in another room.

The analogy helps Jonas separate sentence quality from paragraph coherence.

Later, the word “room” can disappear.

Jonas can ask directly:

Does this sentence serve the paragraph’s function?

English Analogy: Evidence as Payment

Jonas sometimes makes strong claims without enough passage evidence.

Use a payment analogy.

The stronger the claim, the more evidence it costs.

“Concerned” may require modest evidence.

“Furious” requires stronger evidence.

“Terrified” requires different evidence again.

The analogy is not literal.

Its relational point is useful:

Claim strength should remain proportionate to evidential support.

Science Analogy: Model as Map

Nadia sees a simplified diagram and assumes every detail is literally true of the real system.

Use a map analogy.

A transport map does not preserve every distance and angle exactly.

It preserves selected relationships that help the user navigate.

Scientific models work similarly.

They simplify reality to make selected relationships easier to reason about.

Ask Nadia:

• What relationship does this model preserve?
• What detail does it ignore?
• What prediction can it support?
• Where might it become inadequate?

The analogy teaches both usefulness and limits.

Science Analogy: Experimental Control as Fair Comparison

Young learners often understand “fairness” socially before they understand experimental control scientifically.

The tutor can begin there.

If two runners race different distances, comparing finishing times does not isolate running speed fairly.

Likewise, if two experimental groups differ in several important conditions, the result cannot be attributed cleanly to one factor.

Again, the analogy eventually needs refinement.

Scientific control is not moral fairness.

The analogy only preserves the logic of making comparisons interpretable.

Analogies Can Reveal Misconceptions

Ask the learner to extend the analogy.

If the extension becomes incorrect, the error can reveal the learner’s model.

Mira says that because a balance returns to level after equal action, every equation transformation should preserve visual symmetry.

Now the analogy has been overextended.

That is useful evidence.

Jonas treats evidence as payment and assumes one quotation always “pays for” one claim regardless of relevance.

Again, the analogy has become too literal.

Good analogy teaching includes a breaking-point question:

Where does this comparison stop helping?

Analogy and Training Perturbation

Training Perturbation can stress-test an analogy.

Change one feature in the familiar case.

Predict what the corresponding feature should do in the new case.

If the mapping holds, the learner gains structural confidence.

If the mapping breaks, the boundary becomes visible.

For example, if the “function machine” receives one input and randomly produces different outputs, the analogy no longer preserves the function condition.

Perturbation shows which parts are essential.

Analogy and Training Nonexamples

One useful exercise is to show an analogy that almost works but fails structurally.

Ask the learner to reject it.

This creates a nonexample of analogy itself.

The learner has to distinguish:

• familiarity from explanatory value;
• surface resemblance from relational alignment;
• memorable stories from useful models.

This connects to Training Nonexamples.

Learner-Generated Analogies

Eventually, ask students to generate their own analogies.

This is difficult in a useful way.

The learner must decide which structure is essential enough to carry across.

Mira is asked:

Can you invent a real-world system that behaves like a function?

Jonas is asked:

What is paragraph coherence like?

Nadia is asked:

What familiar system helps explain why a scientific model is useful but incomplete?

Then test the analogy.

Which correspondences are valid?

Where does it break?

Generating an analogy becomes a diagnostic task for understanding.

Analogies and Training Readiness

An analogy only helps if the familiar side is actually familiar enough to support the new side.

If the learner does not understand the source system, the analogy creates two problems instead of one.

So check Training Readiness.

• Does the learner know the familiar case?
• Can the relevant relationship be stated?
• Is the mapping simple enough to process?
• Are irrelevant similarities likely to distract?
• Can the learner eventually leave the analogy behind?

Analogies and Training Granularity

Use an analogy at the scale of the relationship that needs explaining.

Do not force one giant metaphor to explain an entire subject.

Equation balance may explain equivalent transformation.

It does not need to explain every feature of algebra.

A map analogy can explain model simplification.

It does not need to explain all of scientific epistemology.

This is the role of Training Granularity: keep the teaching unit large enough to preserve the relationship and small enough to avoid unnecessary baggage.

Analogies and Transfer

The final test of analogy is not whether the student remembers it.

It is whether the student recognises the relationship in a new case.

After the balance analogy, give Mira a new equation without mentioning balances.

After the map-model analogy, give Nadia an unfamiliar scientific model and ask what it preserves and omits.

After the evidence-payment analogy, give Jonas a new inference and ask whether the claim is adequately supported.

The analogy has succeeded when the underlying rule survives after the story disappears.

Failure Mode: Cute but Structurally Weak

An analogy can be memorable and still be poor teaching.

The story is funny.

The relationship is wrong.

Ask whether the correspondence supports the actual concept.

Memorability is not enough.

Failure Mode: Surface Similarity Wins

The learner notices that both systems are round, moving, colourful or familiar but misses the actual structural relationship.

Repair by explicitly mapping roles and relations.

Ask:

What is playing the same role in both systems?

Failure Mode: Analogy Overextension

The learner treats every property of the familiar case as though it must apply to the new one.

This is why every important analogy should include a limit statement.

This analogy helps with X. It stops helping when we assume Y.

Failure Mode: Analogy Becomes Permanent Scaffolding

The learner can only solve when the metaphor is restated.

Fade it.

Move from familiar story to structural language.

Then test the structure without the story.

Mira’s Analogy Training

Mira understands the equation-as-balance analogy.

The tutor gives three equations and asks her to explain each transformation using the balance relationship.

Then the balance language disappears.

She now explains:

I subtract the same quantity from both sides because I need an equivalent equation.

The analogy has done its job.

Jonas’s Analogy Training

Jonas uses the “claim costs evidence” analogy.

At first the tutor asks, “Can you afford this word?”

Later Jonas no longer needs the metaphor.

He asks whether the passage contains enough evidence for the strength of the inference.

The underlying calibration has become explicit.

Nadia’s Analogy Training

Nadia uses the map analogy for scientific models.

The tutor then gives three unfamiliar diagrams.

For each, Nadia identifies:

• what relationship the model highlights;
• what details it suppresses;
• what question the model helps answer;
• what conclusion would exceed the model.

The analogy has grown into model literacy.

The Parent Analogy Audit

• Does the analogy explain a relationship or merely make the lesson entertaining?
• Does my child understand the familiar side?
• Can the child identify what corresponds to what?
• Can the child state what the analogy predicts?
• Can the child state where the analogy breaks?
• Can the child solve later without the analogy being repeated?

The Tutor Analogy Audit

• What structure am I trying to teach?
• What familiar domain preserves that structure?
• What irrelevant similarities might mislead?
• Which correspondences should be made explicit?
• Where does the mapping stop?
• How will I fade the analogy?
• What new task will test structural transfer?

The Deeper Idea: Analogy Is Controlled Reuse of Understanding

Learning would be impossibly slow if every new problem had to be built from zero.

We reuse structure.

Old relationships become templates for seeing new ones.

But intelligent reuse requires discipline.

Find the mapping.

Test it.

Find the invariant.

Find the boundary.

Then carry the structure forward without carrying the metaphor forever.

A good analogy does not replace the new idea. It helps the learner build a structure strong enough that the analogy can eventually disappear.

Research Foundations

A useful current starting point is the 2026 Canadian Journal of Science, Mathematics and Technology Education article on effective analogies in mathematics, which applies structure-mapping theory to analogy quality. This connects naturally with recent work on variability and transfer learning, because analogical transfer requires a learner to recognise common structure across changing surfaces. The broader evidence base on analogy is mature, but classroom effectiveness still depends on careful mapping, prior knowledge, explicit comparison and attention to where the analogy fails.

Continue Through How Training Works

Read this alongside Training Invariants, Training Contrast, Training Nonexamples, Training Perturbation and Training Recombination.

Next: How Training Works | Training Self-Explanation — Explain Why the Step Works, Not Just What to Do.