Two students disagree.
Student A says:
“This new idea works like something I already know.”
That can be the beginning of understanding.
It can also be the beginning of a misconception.
Because analogy is powerful precisely when two things are not identical.
A useful analogy borrows structure.
It does not erase difference.
Electric current may be explained partly through water flow.
An equation can be explained partly through a balance.
A cell can be compared with a factory.
A memory system can be compared with a library.
A computer network can be compared with a transport network.
Each can help.
Each can also mislead.
The intellectual skill is therefore not:
find something similar.
It is:
find the right relational similarity, preserve it carefully, and refuse to carry across the parts that do not belong.
That is analogical reasoning.
A useful Wintour House definition is:
Analogical reasoning is the controlled mapping of a relational structure from one situation—the base—to another situation—the target, so that shared relationships become easier to see and new inferences can be proposed without mistaking surface similarity or unsupported correspondences for evidence.
The words relational structure matter.
Imagine a heart and a pump.
They look very different.
Yet both can participate in a relation involving movement of fluid through a system.
Or a school timetable and a computer scheduler.
Visually unrelated.
But both allocate constrained resources across competing tasks.
Or an equation and a balance.
Again, very different objects.
But both can preserve equality while corresponding changes are applied to each side.
Analogy is powerful precisely because objects can differ while their roles and relationships align.
Dedre Gentner’s 2025 Open Encyclopedia of Cognitive Science overview defines analogy around shared systems of relations rather than simple object similarity and highlights mapping, candidate inference, schema abstraction and alignable differences as central outcomes of analogical processing.
That is why analogical reasoning belongs permanently in a serious learning architecture.
It is one of the ways knowledge travels.
The Wintour House question is therefore:
If a learner became excellent at ten analogical-reasoning operations, which ten would still matter when textbooks, diagrams, simulations and AI-generated analogies changed?
Before the Top 10: Similarity Is Not Yet Analogy
A tiger and a lion are similar.
Both large cats.
Four legs.
Fur.
Teeth.
That is similarity.
Now compare:
A predator pursuing prey
with
a company pursuing market share.
The objects are radically different.
But a relational structure may be mapped:
actor,
target,
competition,
resources,
strategy,
response.
That is closer to analogy.
The distinction matters because learners are easily attracted by surface resemblance.
Two problems both mention cars.
They feel similar.
Yet one may be a constant-speed problem and the other an acceleration problem.
Another two problems look completely different.
One is about mixing paint.
Another about combining investment portfolios.
Yet both may share the same proportional structure.
Strong analogical reasoners learn to ask:
What is similar about the relationships, not merely the things?
That question changes everything.
1. Learn to Name the Base and the Target Explicitly
An analogy has two sides.
The base is the better-understood situation.
The target is the situation we are trying to understand.
If these roles remain vague, the mapping becomes vague too.
Suppose:
Equation is like a balance.
Base:
physical balance.
Target:
equation.
What do we already know about the balance?
Two sides.
A relationship of equality.
Changing one side alone may disturb the balance.
Corresponding changes can preserve it.
Now ask:
Which of those relations belong in the equation?
The distinction is surprisingly useful because students often mix the two systems together midway through an explanation.
A good learner can say:
“I am borrowing this relationship from the base to understand this feature of the target.”
That sentence keeps ownership clear.
It also prevents the analogy from quietly replacing the actual concept.
Eventually the learner should understand equations as mathematics.
Not as literal weighing machines.
The base is scaffolding.
The target is the destination.
Worth learning because: clearly separating the familiar base from the unfamiliar target makes it easier to see what is being transferred and prevents the analogy from becoming the concept itself.
2. Learn to Map Roles and Relationships, Not Surface Objects
This is the central skill.
Suppose we compare:
solar system
and
atom.
Historically, this analogy was seductive.
Central object.
Smaller objects around it.
Visually appealing.
But surface resemblance does not mean the underlying physics is equivalent.
The learner therefore asks:
What role does each object play?
What relation connects the parts?
Which causal structure is actually shared?
The analogy:
equation ↔ balance
works because the important relation is equality preservation.
Not because numbers resemble metal weights.
Likewise:
circulatory system ↔ transport network
might map:
hub,
route,
flow,
destination,
bottleneck.
But blood vessels are not literally roads.
Vehicles do not reproduce biological transport mechanisms.
The visual resemblance is secondary.
The relational correspondence is primary.
Gentner’s structure-mapping framework emphasises this relational focus: corresponding objects can differ greatly as long as they play parallel roles in the relational structure. The classic 1983 structure-mapping paper made this principle explicit.
A powerful classroom prompt is:
“Do not tell me what the objects have in common. Tell me what the relationships have in common.”
That is analogical reasoning beginning to become disciplined.
Worth learning because: surface similarity can be accidental, while relational similarity is what allows knowledge from one situation to illuminate another.
3. Learn to Build One-to-One Correspondences Before Transferring Anything
A good analogy needs structural discipline.
If one element in the base is mapped to three unrelated elements in the target whenever convenient, the analogy becomes elastic enough to “explain” anything.
So build correspondences.
Balance:
left side ↔ left expression.
Right side ↔ right expression.
Equality ↔ equality.
Add weight to both sides ↔ add same quantity to both sides.
Now the mapping can be inspected.
Another example:
Library.
Book ↔ stored information item.
Shelf location ↔ address or index.
Catalogue ↔ retrieval system.
Borrowing process ↔ access process.
That may help explain some information systems.
But suppose we now say:
librarian ↔ CPU,
bookshelf ↔ RAM,
library visitor ↔ packet,
because each comparison happens to be convenient.
The mapping has lost structural consistency.
Strong analogical reasoning asks:
Are the correspondences coherent across the whole mapped relation?
Gentner’s structure-mapping account emphasises one-to-one correspondence and parallel connectivity as hallmarks of a structurally consistent analogy.
This is a quiet but powerful intellectual habit.
An analogy should survive being written as a mapping table.
If the table becomes chaotic, the analogy may be doing rhetorical work rather than reasoning work.
Worth learning because: one-to-one role mapping keeps an analogy structurally constrained enough to reveal genuine parallels instead of allowing arbitrary similarities to accumulate.
4. Learn to State Where the Analogy Breaks
This may be the single most important classroom habit.
After every important analogy, ask:
Where does this analogy stop working?
Electricity and water.
Useful relationships?
Flow.
Resistance.
Potential difference can sometimes be introduced through pressure differences.
Where does it break?
Electric charge is not literally water.
Electrical energy and material flow do not behave exactly like water in household plumbing.
Equation and balance.
Useful?
Equality preservation.
Break?
An algebraic equation is an abstract relation, not a physical object subject to gravity.
Memory and library.
Useful?
Storage and retrieval.
Break?
Human memory is reconstructive; it is not a fixed shelf of unchanged records.
The learner should become comfortable saying:
“The analogy helps with X, but not Y.”
That is intellectual sophistication.
It is also why the existing PSLE Science owner remains protected. How to Use an Analogy to Learn PSLE Science Without Carrying the Wrong Feature Across owns exactly how Science students avoid importing incorrect features into scientific concepts.
A recent integrative review on analogy competence for science teachers stresses this risk directly: analogies can facilitate learning but can also generate misconceptions when inappropriate features are carried across.
Wintour House therefore makes the boundary explicit.
Every analogy should have an exit door.
Worth learning because: the value of an analogy depends as much on knowing what not to transfer as on noticing the relationship that can be transferred safely.
5. Learn Through Progressive Alignment: Near Analogy Before Far Analogy
Young or novice learners often find it easier to align cases that are obviously similar.
That is not a weakness.
It can be used deliberately.
Suppose a Primary learner is learning the relation:
three items increase in size from left to right.
First compare:
three circles increasing in size
with
three triangles increasing in size.
Surface difference exists.
But the relation remains easy to align.
Later compare:
three sounds increasing in volume,
or
three numbers increasing in magnitude.
Now the learner is moving farther from the original surface.
This is progressive alignment.
Start with cases easy enough to compare.
Use the comparison to make the relation visible.
Then move toward cases whose surfaces differ more strongly while the relational structure remains.
Gentner’s work on analogy and abstraction describes progressive alignment as a pathway by which learners first form relatively concrete relational abstractions and then become capable of more distant mappings.
Recent child research gives a concrete illustration. In a 2025 Cognitive Science study on structural alignment and spatial construction, highly alignable comparisons supported stronger transfer of a bracing principle than low-alignability comparisons, while far transfer remained harder.
The educational lesson is elegant.
Do not always begin with the cleverest, most distant analogy.
Begin with the one the learner can actually align.
Then stretch.
Worth learning because: learners often reach abstract relational understanding more reliably when they first master a close analogy and then progressively transfer the same structure into increasingly different contexts.
6. Learn to Generate Candidate Inferences—Then Mark Them as Provisional
Once two structures align, analogy does something interesting.
It suggests new information.
Suppose:
Base system:
A bottleneck in one part of a transport network constrains overall flow.
Target:
A production system has a structurally similar bottleneck.
Candidate inference:
Improving the bottleneck may increase whole-system throughput more than improving already-fast stages.
That inference may be useful.
But analogy has not proved it.
It has suggested something worth testing.
Gentner’s 2025 synthesis calls these candidate inferences: when an aligned relational system contains additional connected information in the base, that information can be projected to the target as a possible inference.
The word candidate should be taught explicitly.
Not:
“The analogy shows that…”
Try:
“The analogy suggests that we should test whether…”
That sentence preserves epistemic hygiene.
In Science:
candidate inference → experiment.
In Mathematics:
candidate conjecture → proof or counterexample.
In History:
candidate parallel → source analysis.
In research:
candidate hypothesis → evidence.
Analogy is a powerful generator.
It is not a final verifier.
Worth learning because: analogies become tools for discovery when learners can project new possibilities without confusing those possibilities with evidence that the target actually behaves the same way.
7. Learn to Compare More Than One Analogy for the Same Target
One analogy highlights one structure.
A second analogy may reveal another.
Suppose we explain the brain as:
computer.
Useful for:
information processing,
input/output,
some aspects of storage.
Danger:
the learner may over-mechanise human cognition.
Now another analogy:
ecosystem.
Useful for:
interacting systems,
adaptation,
distributed relationships.
Different strengths.
Different failures.
Together they reveal that no single analogy owns the target.
Or consider:
an atom as a solar system,
an atom as a probability distribution,
an atom as a quantum state.
Different historical and conceptual roles.
The learner should eventually ask:
What does Analogy A illuminate?
What does Analogy B illuminate?
Where do they disagree?
Which one fits this specific question?
Multiple analogies can prevent over-commitment to one source structure.
They also force the learner to return to the target itself.
The question becomes:
Which properties of the target are stable across several useful mappings?
That is much stronger than memorising one favourite analogy.
This is a clean boundary with Top 10 Comparison Skills Worth Learning.
Comparison owns systematic similarity and difference checking.
Analogical Reasoning uses comparison to map relational systems and transfer inference.
Worth learning because: using several analogies prevents a learner from mistaking one helpful perspective for a complete model of the target.
8. Learn to Use the Difference Between Two Analogies as Information
Analogies reveal similarities.
They also reveal important differences.
Suppose two companies are compared to biological organisms.
Both:
consume resources,
respond to environments,
change over time.
But one key difference appears:
organisms reproduce biologically;
companies reproduce strategies, franchises or organisational forms in very different ways.
That difference is not a failure of the comparison.
It teaches us something about the target.
Gentner’s framework describes alignable differences: once two systems are structurally aligned, differences occupying corresponding roles become especially noticeable.
This can be educationally powerful.
Compare:
fraction multiplication
with
whole-number multiplication.
Shared operation.
Important difference:
multiplying by a proper fraction can reduce magnitude.
That difference helps dismantle the overgeneralised rule:
“multiplication makes bigger.”
Or compare:
democracy in two countries.
Shared institutional roles.
Different electoral rules.
Now the difference becomes analytically meaningful precisely because the systems were aligned.
A sophisticated analogy therefore asks two questions.
What is shared?
What corresponding difference matters?
Worth learning because: once structures are aligned, differences become easier to interpret because the learner can see exactly which role changed and what consequence followed.
9. Learn to Search for a Counter-Analogy or Non-Example
A good analogy can become seductive.
We like elegant mappings.
So test them.
Suppose:
“Learning is like building a muscle.”
Useful:
practice,
adaptation,
recovery.
Now counter-analogy:
“Learning is also like building a network.”
Different implications:
connection,
organisation,
retrieval paths.
Which features of the muscle analogy survive?
Which were accidental?
Or produce a non-example.
If all systems with “flow” behave like water, does information flow through a conversation obey the same conservation rules?
No.
Good.
The analogy now has boundaries.
This is a handoff to How to Improve Critical Thinking and Top 10 Verification Skills Worth Learning.
Critical Thinking asks whether the claim survives alternatives.
Verification asks whether it deserves acceptance.
Analogical Reasoning supplies a useful local test:
Can I construct a case where the surface looks similar but the relational structure differs—or where the relation looks similar but a crucial condition breaks?
That forces the analogy to earn its scope.
Worth learning because: searching for counter-analogies stops an elegant comparison from expanding beyond the relational conditions that actually support it.
10. Learn to Transfer the Relation Without Needing the Original Analogy Present
This is the educational finish line.
At first:
equation ↔ balance.
Later:
student solves equations correctly.
Eventually:
the learner no longer needs to imagine a literal balance every time.
The relational principle has been abstracted:
operations preserving equality must apply correspondingly to both sides.
The analogy has done its job.
Now remove it.
Likewise:
worked-example analogy.
Then new isomorphic problem.
Can the learner recognise the same underlying structure despite new surface features?
This transfer is difficult.
Classic analogical-reasoning research repeatedly finds that people often fail to retrieve a prior analogous case when surface features change, even when the relevant relational structure is available in memory.
So the final test is:
Remove the analogy.
Change the objects.
Change the story.
Keep the relation.
Can the learner recognise it?
If yes:
the analogy has become knowledge.
Worth learning because: the ultimate purpose of analogy is not permanent dependence on the familiar example but independent recognition and use of the underlying relation in new contexts.
The Top 10 Analogical Reasoning Skills as One System
The Wintour House route is:
BASE/TARGET → RELATIONAL MAP → STRUCTURAL CONSISTENCY → BOUNDARY → PROGRESSIVE ALIGNMENT → CANDIDATE INFERENCE → MULTIPLE ANALOGIES → ALIGNABLE DIFFERENCE → COUNTER-ANALOGY → INDEPENDENT TRANSFER
The quieter version is:
Know which case you understand and which case you are trying to understand. Map the relationships, not the decoration. Check that the roles line up. State where the analogy breaks. Use close comparisons before distant ones. Let the analogy suggest possibilities, not prove them. Then remove the analogy and see whether the relational principle survives.
That is analogical reasoning.
Not metaphor for decoration.
Not “this reminds me of…”
Not surface similarity.
Not proof.
Not evidence by resemblance.
Not carrying every feature from one domain into another.
Analogy is disciplined relational transfer.
Analogical Reasoning Is Not the Same as Comparison
Top 10 Comparison Skills Worth Learning asks:
How are A and B similar?
How do they differ?
Analogical Reasoning asks a stricter question:
Do elements in A and B play corresponding roles inside the same relational structure?
Comparison may stop at similarity and difference.
Analogy can continue into inference and transfer.
Analogical Reasoning Is Not the Same as Abstraction
Top 10 Abstraction Skills Worth Learning owns the removal of irrelevant detail while preserving structure.
Analogical comparison can produce abstraction.
But analogy requires two structured cases and a mapping between them.
Abstraction may happen without any analogy at all.
One compresses.
The other aligns and transfers.
Analogical Reasoning Is Not the Same as Pattern Recognition
Top 10 Pattern Recognition Skills Worth Learning detects repeated structure across observations.
Analogy maps relational structure between cases.
A learner may notice:
“All these questions contain percentages.”
Pattern.
Then miss that one percentage problem has the same deeper structure as an apparently unrelated ratio problem.
Analogy can cross surfaces that pattern recognition anchored too tightly to appearance.
Analogical Reasoning Is Not the Same as Explanation
Top 10 Explanation Skills Worth Learning owns constructing the causal, logical or structural bridge that makes an outcome intelligible.
An analogy can help explain.
But analogy is not explanation by itself.
“Electricity is like water” is not yet an explanation.
Which relation?
Why?
Under what condition?
Where does it break?
Explanation can use analogy.
Analogy needs explanation to keep its mapping honest.
Analogical Reasoning Is Not the Same as Synthesis
Synthesis combines several information structures into a larger model.
Analogy maps a relational structure from one case to another.
A learner can make a brilliant analogy using two cases and perform no multi-source synthesis.
A learner can synthesise twenty studies without using any analogy.
Different operations.
Analogical Reasoning Is Not Evidence
This boundary deserves its own heading.
“Country A is like Country B.”
Therefore?
Nothing yet.
“AI is like electricity.”
Therefore?
Nothing yet.
“The brain is like a computer.”
Therefore?
Nothing yet.
An analogy can suggest what to inspect.
It cannot establish that the target inherits every property of the base.
The correct intellectual sequence is:
ANALOGY → CANDIDATE INFERENCE → TEST
Not:
ANALOGY → CONCLUSION
That single distinction prevents enormous amounts of bad reasoning.
For Primary Students
Primary analogical reasoning should begin with visible relations.
“This key opens this lock. What does that relationship remind you of?”
“These two stories have different characters. What happens in the same order?”
“These two number patterns use different numbers. What rule stays the same?”
At younger ages, close alignment matters.
Do not demand far analogy immediately.
Compare two obviously parallel cases.
Name the relationship.
Then slowly change the surface.
A 2026 preschool study involving 220 children modelled analogical reasoning through increasingly complex attributes including rule understanding, relational inference and application, and multidimensional problem solving. It is a diagnostic model rather than an intervention trial, so it should not be treated as proof of one universal developmental sequence, but it reinforces the view that analogical reasoning has decomposable cognitive demands.
A good Primary routine is:
What matches? What relationship matches? What does not match?
Simple.
Powerful.
For Secondary Students
Secondary learners should begin using analogies deliberately across subjects.
Mathematics:
Which old problem has the same structure?
Science:
Which familiar system helps illuminate this mechanism?
English:
Which character conflict resembles another text structurally, despite different settings?
History:
Which cases appear similar—and which crucial conditions make the analogy dangerous?
At this level, analogy should become less teacher-supplied.
Ask students to generate their own.
Then audit them.
Base? Target? Mapping? Boundary?
That turns analogy from illustration into reasoning.
For JC Students
JC students need analogies that survive higher resolution.
Economics:
Can a biological feedback system illuminate market adjustment?
Perhaps.
Which relations map?
Where does agency make the comparison break?
Physics:
Which mathematical structure from one domain is isomorphic to another?
Chemistry:
Which model helps make an invisible process intelligible?
GP:
Historical analogies appear constantly in political and social argument.
Are the supposedly analogous cases actually structurally comparable?
Which background conditions differ?
At JC level, the learner should become suspicious of analogy precisely when it feels rhetorically irresistible.
Elegant analogy deserves more checking.
Not less.
Analogical Reasoning in Mathematics
Mathematics is full of structural analogies.
Addition and multiplication.
Arithmetic sequences and linear functions.
Fractions and ratios.
Geometric transformations.
Vectors.
Matrices.
A powerful Mathematics learner asks:
Which problem I already know has the same skeleton?
Not:
Which problem has the same nouns?
This is how transfer grows.
A worked example about trains may transfer to water tanks if the underlying rate structure is recognised.
A proof strategy in one geometric configuration may suggest a route in another.
But the mapping must be checked.
The same formula shape does not guarantee the same domain conditions.
Analogy proposes.
Mathematics verifies through derivation, proof or counterexample.
Analogical Reasoning in Science
Science is one of analogy’s great homes.
We use models for things too small, too large, too fast, too slow or too abstract to observe directly.
But Science also shows analogy’s danger most clearly.
The existing PSLE Science page therefore remains protected:
How to Use an Analogy to Learn PSLE Science Without Carrying the Wrong Feature Across.
The cross-domain Wintour rule is:
map the relationship, then write the boundary.
Science analogy should never end at:
“It is like…”
Continue:
“It is like X with respect to Y, but unlike X with respect to Z.”
That sentence is much safer.
Analogical Reasoning in English and GP
Literature loves analogy.
Character A mirrors Character B.
One social system parallels another.
One image functions like another.
But strong interpretation requires evidence.
A structural analogy between two characters may reveal a useful relationship.
It does not mean the characters are equivalent.
GP requires even greater care.
Public argument frequently relies on analogies:
“This technology is the new printing press.”
“This conflict is another Vietnam.”
“Data is the new oil.”
“AI is electricity.”
These phrases can be intellectually productive.
They can also smuggle conclusions into the argument.
Ask:
What relation is being mapped?
Which relevant conditions differ?
What policy conclusion is being projected?
Does that conclusion depend on a feature the two cases do not share?
Now rhetorical analogy becomes analysable.
Analogical Reasoning in Studying
Students can use analogy actively.
After learning a concept, ask:
What does this remind me of structurally?
Then:
Where does that comparison break?
This is much better than a decorative mnemonic.
Suppose a learner understands feedback loops in ecosystems.
Can that help understand feedback in economics?
Maybe.
Map it.
The analogy may reveal a higher-order schema.
That is one route toward transfer.
But the learner should eventually retrieve the schema without needing the original case.
Otherwise the analogy remains a crutch.
Analogical Reasoning in Research
Researchers use analogy to generate models and hypotheses.
A process in one domain resembles a process in another.
Interesting.
Now what?
Map.
State the common relation.
Generate a candidate inference.
Then test.
Historical examples from science show analogy’s creative role, and Gentner’s 2025 review notes its importance in scientific discovery as well as learning, problem solving and communication.
Research should therefore respect analogy without worshipping it.
Analogy can open a door.
Evidence decides whether there is actually a room behind it.
Analogical Reasoning in the Age of AI
AI can generate analogies instantly.
“Explain quantum computing with a cooking analogy.”
Done.
“Explain inflation to a ten-year-old using balloons.”
Done.
“Give me five analogies for working memory.”
Done.
This is useful.
It is also dangerous.
AI optimises for an understandable mapping.
That mapping may contain a subtle structural error.
So a strong AI workflow asks for more than an analogy.
Ask:
Map each element explicitly.
Which relation is the analogy intended to preserve?
List three ways the analogy breaks.
What misconception could a student develop from this analogy?
Give me a second analogy that highlights a different property.
Which inference from the first analogy should not be transferred?
Now AI becomes an analogy generator plus an analogy critic.
Better still:
have the student generate the analogy first.
Then let AI challenge the mapping.
This preserves cognitive ownership.
The learner does the transfer.
The machine provides pressure.
The Analogy Paradox: A More Familiar Analogy Can Be More Misleading
Familiarity feels safe.
But familiar systems contain many strongly represented properties.
Those properties can leak into the target.
Water is familiar.
That makes water-flow analogies attractive.
It also makes it easy to import assumptions about literal fluid, storage, leakage and consumption where they do not belong.
A useful analogy therefore balances:
enough familiarity to support mapping,
not so much unexamined detail that the base overwhelms the target.
The Analogy Paradox: A More Distant Analogy Can Produce Better Transfer—But Only After the Learner Can Align It
Far analogies are powerful.
They reveal structure beyond surface.
But they are hard.
Gentner’s work on progressive alignment makes this boundary particularly clear: highly surface-dissimilar cases can support broad relational abstraction once the learner can align them, but presenting distant cases before that ability is established can simply produce confusion.
So:
close first.
Far later.
Not because easy is always better.
Because alignment is the cognitive gate.
The Analogy Paradox: The Best Analogy Eventually Becomes Unnecessary
At first:
equation = balance.
Useful.
Later:
the learner understands equality preservation mathematically.
The balance can disappear.
Excellent.
A teaching analogy has succeeded when the learner no longer depends on it.
That is quiet luxury.
Scaffolding that knows when to leave.
The Wintour House Test: Does Analogical Reasoning Survive When AI Can Generate Perfect Analogies?
Suppose AI becomes extraordinary at analogy generation.
It knows the learner’s age.
Prior knowledge.
Vocabulary.
Subject.
Misconceptions.
It can produce the perfect comparison instantly.
Does human analogical reasoning disappear?
No.
Because somebody still has to decide:
which relation matters,
whether the mapping is structurally consistent,
which surface features are irrelevant,
where the analogy breaks,
which candidate inference is legitimate,
whether another analogy reveals a different structure,
and whether the relation transfers once the analogy is removed.
The scarce skill becomes analogy governance.
That is why Analogical Reasoning belongs permanently in the Skills Worth Learning series.
The mature learner can eventually say:
I know which case I am borrowing from and which case I am trying to understand. I can map the corresponding roles and relations. I know where the mapping stops. I can use close examples to discover a relation and later recognise that relation across very different surfaces. I can generate a new inference without pretending the analogy proved it. I can compare multiple analogies and test them with counterexamples. And when the scaffold disappears, I still own the relation.
That is analogy becoming transfer.
Research Anchors
The ten skills above are a Wintour House editorial synthesis, not a claim that cognitive science has validated one universal ten-part taxonomy of analogical reasoning.
Gentner’s 2025 Open Encyclopedia of Cognitive Science overview provides a clear current general framework. It defines analogy in terms of shared relational systems rather than object similarity and describes analogical mapping through structural consistency, relational focus, systematicity, candidate inference, schema abstraction and alignable differences. It also highlights a longstanding difficulty: people often fail to retrieve structurally analogous prior cases when their surface features differ.
A 2023 systematic review of analogy teaching retained 19 empirical articles containing 24 sample elements. Most reported positive outcomes from analogy-based teaching, especially for learning novel concepts, but the review found disproportionate representation of higher education and exact-science contexts and identified important gaps across other educational levels and disciplines.
Recent child research illustrates progressive alignment in practice. The 2025 Cognitive Science study on structural alignment and spatial construction found stronger transfer when children compared highly alignable structures; far transfer, especially generative transfer, remained more difficult.
The 2026 preschool cognitive-diagnostic study involved 220 children and identified four hierarchical cognitive attributes: rule understanding, relational inference and application, two-dimensional problem solving and multi-dimensional problem solving. It is an assessment model rather than evidence for one particular instructional programme.
Finally, the recent Analogy Competence for Science Teachers review stresses the dual nature of analogy: it can support conceptual understanding and language, but can also generate misconceptions when irrelevant properties of the base are transferred to the target.
The strongest defensible Wintour House conclusion is therefore:
Analogical reasoning is not noticing that two things seem alike. It is controlled relational transfer: define the base and target, align corresponding roles and relations, preserve structural consistency, make the boundary explicit, move progressively from near to far cases, treat projected conclusions as candidate inferences, use multiple and counter-analogies to expose limits, and finally demonstrate transfer when the original analogy is no longer present.
