How Lin Che Learns Cognitive Psychology from Scratch
A twenty-act story illustrates how a learner builds, tests, abstracts, transfers, and revises models while studying a complex field.
Chinese original: 林澈如何從零深入學習認知心理學.
Editorial note: Complete English translation of the user-supplied
story.txt, received on September 28, 2026. This illustrative story preserves the source’s claims, examples, and dialogue; it is not independently fact-checked research or evidence of the learner’s mastery. The source title and all twenty acts and concluding passages are retained; headings are formatted for the site. This is a separate story from How A-Heng Truly Learns Cognitive Psychology. Related framework: Cognition as Model Building and Revision.
The following story is not about “a genius learning effortlessly.” It shows the specific cognitive operations that someone with strong learning ability performs when facing an unfamiliar, complex subject. You can apply the protagonist’s approach directly to mathematics, economics, programming, philosophy, law, management, and other fields.
How a Student with Strong Learning Ability Studies Cognitive Psychology in Depth from Scratch
Act 1: He Does Not Immediately Start Reading
Lin Che decides to spend six months truly learning cognitive psychology.
A typical beginning might be:
Find a textbook, start at chapter one, highlight key points, and memorize concepts.
But Lin Che does not do that.
He opens a blank sheet of paper and first writes:
“What exactly do I want to learn to do?”
He thinks for a long time.
Finally, he writes:
I do not want merely to “know many cognitive psychology terms.”
I want to build a model that enables me to explain:
How people perceive information; how they pay attention; how they form memories; how they learn; how they understand; how they reason; why they make certain systematic errors; and why cognition and behavior change when conditions change.
This action looks simple, but it is very important.
Because Lin Che first distinguishes:
Information from knowledge.
He knows that even if he can recite:
“Working memory,” “long-term memory,” “attention,” “cognitive load,” “schema,” “heuristics,” “metacognition,” and so on,
that does not mean he has truly mastered cognitive psychology.
If these concepts are merely scattered terms in his mind, they are only information.
Only when they begin to form relationships:
Attention affects which information enters further processing;
working memory has capacity and processing limitations;
existing knowledge affects how new information is interpreted;
memory retrieval, in turn, affects subsequent judgment and learning;
do they begin to form a knowledge structure.
For the first time, Lin Che genuinely appreciates:
Knowledge is not “knowing a lot”; it is forming structures that can be used.
Act 2: On His First Reading, He Deliberately Does Not Try to “Understand Everything”
In the first week, he opens a cognitive psychology textbook.
The first chapter covers:
Sensation, perception, and attention.
The second covers:
Memory.
The third:
Knowledge representation.
The fourth:
Language.
The fifth:
Problem solving, reasoning, and decision-making.
Following his old habits, he would have stopped whenever he encountered an unfamiliar concept.
But this time he does not.
He first skims the whole book.
When he sees an unfamiliar concept, he asks only:
“Roughly where does this belong in the overall system?”
He draws a very rough diagram:
External information
→ sensation and perception
→ attention
→ working memory
→ interaction with existing knowledge
→ learning and long-term memory
→ retrieval
→ reasoning, judgment, and decision-making
→ behavior
→ new feedback
Many parts of this diagram lack rigor.
But Lin Che is not worried.
Because he knows:
Learning is not putting the correct answer into your head once and for all.
It is more like:
Build a provisional model → encounter more evidence → discover errors in the model → revise the model.
He even writes on the paper:
“This is only Version 0.1.”
This is his understanding of the essence of learning.
Learning is not absorption.
It is:
Updating models.
Act 3: He First Discovers That “I Thought I Understood” Does Not Mean He Understood
One evening, Lin Che finishes reading about working memory.
He thinks it is very simple.
Isn’t the main point just that human working memory has limited capacity?
He closes the book.
Then asks himself:
“Without using the book’s wording, can I explain working memory to a twelve-year-old?”
He begins:
“Working memory is… remembering things briefly…”
Then suddenly stops.
“Are short-term memory and working memory exactly the same?”
He does not know.
He asks again:
“If working memory is limited, why can experts handle very complex information?”
He does not know.
“Why do some people struggle with a string of digits, yet find a familiar, complex sentence less demanding?”
He does not know.
“What role does existing knowledge play?”
Still unclear.
Here, he discovers an important phenomenon:
Familiarity is not understanding.
While reading, a sentence flows smoothly.
The mind produces:
“I know this.”
But “watching someone else complete the reasoning” and “being able to regenerate the reasoning yourself” are completely different things.
He therefore begins following a rule:
After every important concept, answer four questions without looking at the book:
“What is it?”
“Why does this phenomenon exist?”
“What happens if the conditions change?”
“Can I give a new example?”
Only now does training his ability to understand really begin.
Because the essence of understanding is not remembering a definition.
It is:
Building a model in your mind that can run.
Act 4: He Moves from “Terms” to “Variables and Relationships”
When reading about attention, Lin Che initially notes many terms.
Selective attention.
Divided attention.
Top-down processing.
Bottom-up processing.
At first, his notes look like this:
“Selective attention: …”
“Top-down processing: …”
“Bottom-up processing: …”
A few days later, he feels something is wrong with these notes.
He knows what each term means but not how they work together.
So he changes how he takes notes.
He begins asking:
Which variables actually influence attention?
He identifies:
Stimulus salience.
Current goals.
Prior expectations.
Resource limitations.
Task difficulty.
Motivation.
He then asks:
How are they related?
For example:
A very prominent sound may attract attention even when you did not intend to notice it.
But when you are highly focused on a goal, you may overlook information that would otherwise be conspicuous.
He suddenly realizes:
Previously, he saw many “words.”
Now he begins seeing:
Variables → relationships → conditions → outcomes.
This is the first obvious change in his thinking.
He no longer asks:
“What are the knowledge points in this chapter?”
Instead, he asks:
“How does this system actually work?”
Act 5: He Practices Looking for “Essence”
One day, he asks:
“What is the essence of attention?”
At first, he wants to write:
“Attention is selecting information.”
But he stops.
Because he knows “essence” is not a pretty sentence.
He changes the question:
If I remove the specific experiments, terminology, and phenomena, what core problem remains?
He considers:
Human information-processing resources are not unlimited.
So the system must make selections.
In other words:
Limited resources → selection is necessary → selection is jointly influenced by goals and stimuli.
Lin Che does not announce:
“I have found the one ultimate essence of attention.”
He simply writes:
For the question “Why does attention exist, and what problem does it solve?”, one important core structure is:
Information selection under limited processing capacity.
This is how he gradually understands “seeing essence.”
Not asking:
“What is the most mysterious one-sentence definition of this thing?”
But asking:
“If I want to explain this phenomenon, which variables and mechanisms cannot be ignored?”
Act 6: Abstraction Happens for the First Time
Lin Che learns about working memory capacity limitations.
A few days later, he is learning programming.
He finds a very long piece of code with many nested levels.
After a few minutes, he gets lost.
Suddenly, he thinks:
“Could this relate to working memory limitations?”
On the surface:
One is psychology.
The other is programming.
Completely different.
But he removes the surface details.
The psychology case:
Too much information must be held and processed simultaneously
→ cognitive load increases
→ errors increase.
The code case:
Too many variables, conditions, and nested structures must be tracked simultaneously
→ cognitive load increases
→ understanding becomes difficult and errors increase.
He discovers a shared structure underneath:
When the active information a system must maintain simultaneously exceeds what it can process effectively, performance declines.
This is abstraction.
He takes a phenomenon from a psychology textbook,
removes its specific names, examples, and background,
and retains:
A structure usable across situations.
Abstraction is therefore not making an explanation increasingly empty.
It is:
Removing unimportant differences and retaining the common structure that really matters.
Act 7: After Abstraction, Transfer Begins Naturally
Lin Che deliberately sets himself a task.
After every important principle, he must find three completely different situations.
For example:
Working memory limitations.
He thinks of:
Performing too many steps simultaneously while learning mathematics.
Watching navigation, roads, and signs simultaneously while driving in an unfamiliar area.
Remembering ingredients, heat levels, and sequence simultaneously when making a complex dish for the first time.
Their appearances are completely different.
But the structures are similar.
He begins to understand:
Transfer is not “suddenly becoming flexible in application.”
It requires an intermediate step:
First abstract the original knowledge into a structure,
then recognize whether a similar structure appears in the new situation.
So:
Case A
→ abstract model M
→ case B
This is transfer.
From then on, for every important principle he learns, he asks:
“Besides this example, where else does it hold?”
This question gradually changes his learning.
He is no longer only studying cognitive psychology.
He is accumulating a library of models for understanding other phenomena.
Act 8: While Studying Memory, He Investigates Regularities Rather Than Conclusions
Lin Che reads:
Active retrieval usually benefits later memory performance more than repeated passive reading.
Previously, he might have written:
“Do more active recall.”
Now he is not satisfied.
He asks:
“What exactly is this regularity?”
He begins breaking it down.
First:
What are the variables?
Learning method.
Number of retrievals.
Time intervals.
Material difficulty.
Existing knowledge.
Final test format.
Delay before testing.
Then:
“What is the outcome variable?”
It could be:
Recognition.
Free recall.
Application.
Transfer.
Long-term retention.
Next:
“Is this regularity the same under every condition?”
No.
He starts searching for boundaries:
Does it work the same with different materials?
What happens when testing is used before the material is understood at all?
Do immediate tests and tests one week later differ?
What happens when recall is incorrect and there is no feedback?
For the first time, Lin Che understands:
Genuinely understanding a regularity is not knowing “A causes B.”
It is knowing:
Under condition C,
A acts through certain mechanisms
to increase or decrease B,
and the effect may change when conditions change.
This is the ability to understand regularities.
Act 9: He Actively Looks for Counterexamples
Initially, after learning a viewpoint, Lin Che always likes to find examples supporting it.
Later, he discovers:
This makes him prone to illusions.
He establishes a new rule:
Whenever I think I have discovered a regularity, I must actively ask: Under what circumstances would it fail?
For example:
“The more repetitions, the better the memory.”
He asks:
What if it is merely mechanical rereading?
What if each repetition happens in the same environment?
What if I can recognize it but cannot actively retrieve it?
What if I memorize the surface answer without knowing the principle?
He gradually discovers:
Counterexamples do not destroy knowledge; they sculpt it.
Because “regularities” without boundaries are often just slogans.
Mature knowledge usually includes:
Conditions of applicability.
Conditions of failure.
Exceptions.
Mechanisms.
The strength of the evidence.
Act 10: He Discovers That Thinking Itself Can Be Decomposed
Previously, Lin Che would say:
“Let me think.”
Now, he starts asking:
“What operation am I actually performing when I say I am thinking?”
Sometimes he is:
Comparing two theories.
Sometimes:
Classifying.
Sometimes:
Looking for causality.
Sometimes:
Working backward from outcomes to causes.
Sometimes:
Imagining “What would happen if this condition were absent?”
Sometimes:
Breaking a complex system into parts.
Sometimes:
Combining two existing models.
He gradually realizes:
Thinking is not a mysterious mass of activity; it is a series of operations on internal representations.
When he “cannot think of anything,” he no longer simply sits there trying harder to think.
He asks:
“Which operation am I missing now?”
Have I not compared?
Not decomposed?
Not drawn a causal chain?
Not looked for counterexamples?
Not changed the representation?
Not changed the scale?
Not considered time?
For the first time, his thinking becomes manageable.
Act 11: He Encounters a Genuine “Problem” for the First Time
The teacher assigns an open-ended question:
Why can students spend a great deal of time studying and still perform poorly on exams?
Many classmates immediately begin answering.
“Maybe they do not work hard.”
“Their methods are poor.”
“Their memory is poor.”
Lin Che does not answer immediately.
He first rewrites the problem:
The goal is to explain why increased study time does not translate into improved exam performance.
Then he asks:
Which steps lie between “study time” and “exam results”?
He draws:
Time invested
→ actual time attending
→ processing method
→ initial understanding
→ memory encoding
→ retention
→ retrieval
→ recognition of problem types
→ knowledge transfer
→ execution during the exam
→ results
Looking at this chain, he suddenly realizes:
“Study time” is only a variable at the beginning.
Many intermediate steps actually affect the result.
Someone sitting at a desk for five hours
does not mean five hours of effective processing.
Even understanding something
does not mean being able to retrieve it one week later.
Even retrieving a definition
does not mean transferring it to an unfamiliar problem.
The difficulty in this question is therefore not:
“The answer is too hard.”
It is:
The initial problem representation is too crude.
Now he genuinely understands:
The essence of a problem is a gap between the current state and the goal state, with the effective path unknown.
The first step in solving it is usually not finding an answer.
It is:
Representing the problem correctly.
Act 12: Once the Problem Is Re-represented, Solutions Suddenly Multiply
Lin Che now breaks “working hard but doing poorly on exams” into several possible bottlenecks.
If the bottleneck is attention:
Change the environment.
If it is understanding:
Increase self-explanation, examples, and causal derivation.
If it is retention:
Add spaced learning.
If it is retrieval:
Increase closed-book recall.
If it is transfer:
Add varied problems and practice across situations.
If it is uncorrected errors:
Increase high-quality feedback.
He suddenly understands:
Problem solving is not selecting the “best method” from a pile of methods.
More important is:
First identifying where the system is stuck.
This is “bottleneck thinking.”
The same surface problem:
“Not learning well.”
May have completely different underlying causes.
If the diagnosis is wrong,
more effort sometimes just means doing the wrong thing more forcefully.
Act 13: He Trains His “Learning Ability”
The six-month plan reaches its second month.
Lin Che begins reviewing his progress once a week.
He does not ask:
“How many pages did I read this week?”
He asks:
“What changed in my models?”
For example, he originally thought:
Learning meant repeated exposure to information.
Now:
Learning includes encoding, integration with existing knowledge, retrieval, feedback, correction, and other steps.
Originally:
Forgetting meant a memory disappeared.
Now he knows:
“Unable to retrieve” and “no trace remains at all” are not the same thing.
Originally:
Understanding happened at the moment of reading.
Now he begins thinking:
Genuine understanding requires regeneration, prediction, application, and revision.
These changes let him see for the first time:
Learning ability can be understood as the ability to update models.
People with strong learning ability are not necessarily right the first time.
The opposite may even be true.
They may be very willing to admit:
“My previous model was wrong.”
What really creates a difference is:
The speed of discovering errors, the speed of revising models, and the ability to verify them again after revision.
Act 14: He Discovers That More Knowledge Can Sometimes Make Learning Faster
In the third month, he begins reading harder research papers.
Something strange happens.
In the first month, reading ten textbook pages was very slow.
Now, although some papers are harder, he can quickly grasp their core.
Why?
Previously:
Attention, retrieval, encoding, schema, and working memory
were isolated terms.
Now, each has a place.
When new information arrives, he does not have to understand it from scratch.
He can “attach” it to an existing model.
For the first time, he truly understands:
Knowledge is not a burden. Well-organized knowledge is the infrastructure for understanding new knowledge.
This is also why experts and beginners may see completely different things in the same information.
Beginners see:
Many details.
Experts more readily see:
Familiar structures.
Act 15: He Investigates “Understanding Other People”
Cognitive psychology gradually involves judgment, decision-making, social cognition, and related questions.
Lin Che starts using models in everyday life.
During a group assignment, one classmate says:
“I think our current plan is pretty good. We do not need to change it.”
Another immediately says:
“He is just lazy and does not want to do more work.”
Previously, Lin Che might also have accepted that explanation.
Now, he pauses.
He thinks:
A statement is merely an observed output.
Inferring a psychological state directly from one statement involves considerable uncertainty.
He lists several possibilities in his mind:
The classmate genuinely thinks the plan is good.
He thinks revision would cost too much.
He worries about missing the deadline.
He has not understood the other options.
He actually disagrees but does not want conflict.
He is simply tired.
He has invested heavily in the current plan and therefore prefers it.
Lin Che does not say:
“The real reason must be one particular explanation.”
He begins learning to understand people probabilistically.
Observe behavior, build several possible models, then look for evidence that can distinguish them.
He asks the classmate:
“What problem worries you most if we revise it?”
The classmate replies:
“Time. We have to submit it on Friday. I do not think the other plan is bad; I think we do not have enough time.”
Lin Che suddenly realizes:
If he only looks at “he refuses to revise,”
it is easy to infer:
An attitude problem.
But adding one variable:
Time constraints
makes the entire behavior much easier to understand.
Act 16: He Begins Understanding Groups
Later, something else happens in the group.
Five people attend a meeting.
The teacher proposes a plan.
Everyone says:
“Fine.”
Afterward, three people privately say:
“I actually do not think it works very well.”
Lin Che is very interested in this.
He discovers:
Explaining group behavior only through “each person’s individual thoughts” is difficult.
You must also include:
Roles.
Power.
Norms.
The cost of making a public statement.
The distribution of information.
How others will see me.
What others believe everyone else thinks.
He suddenly discovers:
An individual-level model cannot fully explain group phenomena.
He upgrades his model:
Individual behavior ≈
goals + beliefs + incentives + ability + constraints.
Group behavior also requires:
Power structures + social norms + coordination problems + identity + common knowledge.
This is abstraction working again.
He is not memorizing what happened in one particular group.
He is trying to extract:
What structures can recur across different groups?
Act 17: He Questions His Favorite Explanations
By the fourth month, Lin Che has learned quite a lot.
A danger appears.
He starts easily explaining everything with theories he has just learned.
Someone forgets something:
“Memory retrieval failure.”
Someone makes a poor decision:
“Cognitive bias.”
Someone ignores advice:
“Confirmation bias.”
Suddenly, everything has a theoretical explanation.
This gives him a sense of control.
But one day, he asks himself:
“If a theory can explain everything, how can I know whether it is actually correct?”
He becomes alert to the danger.
From then on, he adds a rule:
An explanation must produce predictions that distinguish between alternatives.
For example:
If someone errs mainly because they “did not notice the information,”
making the information more salient may improve the result.
If the main reason is “not understanding at all,”
increasing salience alone may have little effect.
If the main reason is “knowing but being unable to retrieve,”
providing a cue may help considerably.
For the same surface error,
different models should generate different predictions.
Now he begins genuinely understanding:
Good models do not merely tell stories after the event. They should also predict unfamiliar situations.
Act 18: He Begins Seeing “The Essence of Essence”
In the fifth month, he looks back at his first month’s notes.
The pages are filled with:
Definitions.
Terms.
People’s names.
Experiment names.
Now he reorganizes the whole of cognitive psychology.
This time, he does not follow textbook chapters.
Instead, he asks:
“If I compressed the whole course into a few core questions, what would they be?”
He arrives at a set of questions:
How is external information selected?
How are limited processing resources allocated?
How is new information integrated with existing models?
How is information stored and retrieved?
How do people infer the world from incomplete information?
How do goals control cognition and behavior?
How does the system update in response to errors?
How does existing knowledge both help cognition and create biases?
How do different situations change the same person’s judgment?
Looking at these questions, for the first time he feels he is beginning to “see the layer underneath.”
Textbooks contain more than a thousand pages.
There are thousands upon thousands of experiments.
But much of the content revolves around a few core tensions:
Limited resources.
Incomplete information.
Existing knowledge.
Predictions and errors.
Goals and selection.
Environment and constraints.
Learning and adaptation.
This is what he understands as:
The essence of essence is not finding one magical sentence, but identifying a few variables, relationships, mechanisms, and constraints with strong explanatory power.
Act 19: Has His Ability Really Improved?
Six months later, a friend asks:
“So are you really good at cognitive psychology now?”
Lin Che does not answer by counting the theories he has memorized.
He takes out a research article he has never read before.
First, he looks at the research question.
Then he guesses:
“The authors are probably distinguishing between these two possible mechanisms.”
He sees the experimental design.
He guesses again:
“This control condition is probably intended to rule out another explanation.”
He sees the results.
He pauses:
“If this result holds, the original model needs a small revision.”
Then he reads the study limitations.
He has already anticipated two of them.
His friend asks:
“Could you do this before?”
Lin Che says:
“No.”
Only now does he truly understand ability.
Not:
“I have read this article before.”
But:
Even with new input, I can still reliably use an effective set of operations.
He can now:
Identify variables.
Distinguish correlation from causation.
Look for mechanisms.
Check boundaries.
Propose alternative explanations.
Imagine counterexamples.
Compare models.
Generate predictions.
Revise in response to evidence.
Together, these constitute genuine ability.
Act 20: After Six Months, He Uses the Same Method to Study Another Subject
After the six months end, Lin Che begins economics.
His friend asks:
“So you have to start over again?”
Lin Che says:
“Of course I have to learn the content anew, but my way of learning does not have to start entirely from scratch.”
He encounters supply and demand.
His first question is not:
“What is the definition?”
It is:
“What are the core variables?”
He encounters opportunity cost.
He asks:
“What problem does this concept solve?”
He encounters market failure.
He asks:
“Which mechanisms prevent individual choices from naturally producing certain collective outcomes?”
He encounters a model.
He asks:
“Which details have been deliberately ignored? Why?”
He encounters a regularity.
He asks:
“What conditions make it hold?”
He encounters a policy case.
He asks:
“The surface story is different, but is there a familiar structure underneath?”
Now we see genuine transfer.
He is not forcing cognitive psychology’s conclusions onto economics.
He transfers something at a higher level:
How to define problems.
How to identify variables.
How to build models.
How to investigate causality.
How to find boundaries.
How to test with counterexamples.
How to update through feedback.
These are the abilities with genuinely high transfer potential.
Finally: What Did Lin Che Really Learn?
On the surface, he studied:
Cognitive psychology.
At a deeper level, he learned:
How to learn a complex field.
At a still deeper level, he actually trained four things.
The first:
Representation.
Transforming messy reality into:
Variables, relationships, structures, and problems.
The second:
Compression.
Extracting from many cases:
Regularities, invariants, and common structures.
The third:
Inference.
Asking:
If the model is true,
what happens when conditions change?
The fourth:
Updating.
When reality does not match the prediction,
rather than protecting his old views,
he revises the model.
The concepts that initially looked separate all connect.
Knowledge means forming structured, usable models.
Learning means updating models through information and feedback.
Understanding means building models that can explain, derive, and predict.
Abstraction means removing surface differences and retaining key structures.
Transfer means discovering similar structures in another situation and reusing models.
Thinking means comparing, decomposing, combining, reasoning with, and revising models.
A problem is a situation in which a path between the current state and the goal has not yet been found.
Problem solving means using models to find and verify paths under constraints.
Ability means reliably performing these operations even when the situation changes.
Seeing through to essence means identifying, among complex phenomena, the few variables, causal mechanisms, invariants, and constraints that truly govern outcomes.
High-level learning is therefore ultimately not:
Putting more answers into your head.
It is gradually developing the ability to:
Face something you have never seen before and still know how to observe, how to ask questions, how to model, how to verify, and how to revise.
At this point,
what you have gained is no longer merely “knowledge.”
It is a cognitive system that can continuously produce new knowledge.
The most practical way to turn this story into your own training is to make yourself complete the same loop whenever you study a topic in depth: first draw a rough model → identify core variables → reconstruct it in your own words → find three new examples → abstract the common structure → predict what happens when conditions change → actively seek counterexamples → test through problems or reality → revise the model based on errors.
What really matters is that this method does not require you to “see through to essence” on your first attempt. Quite the opposite: seeing essence is often the result of reality continually correcting a model through Versions 0.1 → 0.2 → 0.5 → 1.0. One of the most important differences between experts and beginners may therefore not be how correct experts are initially, but how effectively they expose their errors quickly and then make high-quality updates.