How Knowledge Transforming Works | When Writing Changes the Writer's Knowledge

eduKateSG Learning Node Series · 0135

A student can know a great deal about a topic and still write a weak essay because writing is not the same job as knowing.

The learner opens the document, remembers one relevant fact, writes it down, remembers another, writes that down too, then keeps moving until the word count is reached.

The sentences may be accurate. The knowledge may be genuine. Yet the writing can remain shapeless because nothing forced the ideas to change relationship with one another.

Carl Bereiter and Marlene Scardamalia gave this difference a powerful name. Novice writers often rely heavily on knowledge telling: retrieve relevant content and place it into text. More mature writing can involve knowledge transforming: treat writing as a problem space where content, purpose, audience and form constrain one another, forcing the writer to reorganise what they know.

Knowledge transforming begins when writing stops being the output channel for existing ideas and becomes a system that changes the ideas while they are being expressed.

The 50-Second Read

Knowledge telling retrieves content that seems relevant and converts it into text with relatively little transformation.
Knowledge transforming coordinates two problem spaces: the content problem—what do I actually think or know?—and the rhetorical problem—what must this text do for this reader in this situation?
The writer moves repeatedly between these spaces. A rhetorical need can expose a content gap; a new content insight can change the structure of the argument.
Bereiter and Scardamalia argued that moving from knowledge telling to knowledge transforming is not simply adding more writing skill. It requires a reconstruction of the composing process.
Planning, drafting and revision are recursive rather than a rigid three-stage sequence.
Good prompts create genuine rhetorical problems instead of asking only for topic recall.
Revision becomes transformative when it changes meaning, organisation, evidence or argument—not merely grammar and spelling.
Feedback is useful when it pushes the writer back into the problem space rather than replacing the writer’s decisions.
AI can support knowledge transformation by generating counterarguments, identifying gaps and testing structure, but it can also collapse the process if it writes the solution before the student has a problem to solve.
The destination is not “write more academically.” It is to use composition as disciplined thinking.

Canonical Owner Boundary

This Learning Node owns the transformation of a writer’s knowledge through the interaction between content problems and rhetorical problems during composition. How Generative Learning Works owns the broader learning principle that generating explanations, maps or summaries can deepen learning. How Knowledge Building Works owns collective improvement of public ideas in a community. This page owns an individual composing mechanism: how the demand to communicate well can force the writer to reorganise what they know.

1. Knowledge Telling Is Not Stupidity

Knowledge telling is efficient.

If someone asks, “What did you do yesterday?”, you do not need a complex rhetorical model. You retrieve events and tell them.

Novice school writing often uses the same process. The topic activates related knowledge. Genre cues suggest what kind of statement belongs. The writer retrieves one idea after another and converts them into sentences.

This can produce coherent short answers and straightforward narratives. The problem begins when the task requires selection, prioritisation, argument, audience adaptation or conceptual reorganisation.

2. The Hidden Rule of Knowledge Telling: If It Is Relevant, Put It In

A student writing about climate change remembers greenhouse gases, then polar ice, then electric cars, then recycling, then sea-level rise.

Each item is relevant to the topic.

But relevance alone does not create argument.

Knowledge telling tends to use associative retrieval: one idea cues another. The text grows locally. The writer asks, “What else do I know?” more often than “What does this paragraph need to prove?”

3. Knowledge Transforming Adds a Second Problem Space

Bereiter and Scardamalia’s mature model introduces interaction between two spaces.

Content problem space: what is true, important, connected, uncertain or missing in the subject matter?
Rhetorical problem space: what must the text accomplish for this audience, purpose, genre and situation?

The writer moves between them.

“My claim is too broad for the evidence I have” is a rhetorical problem that forces a content decision. “This new source contradicts my second paragraph” is a content problem that forces a rhetorical decision.

The writing process begins changing the writer’s knowledge structure.

4. Writing Creates Problems That Thinking Alone May Not Expose

People often feel that they understand an idea until they try to explain it precisely.

The blank between two sentences reveals a missing causal link. A paragraph requires a distinction the writer had never articulated. A claim suddenly looks too strong once evidence has to be placed beside it.

This is why writing can be epistemic: the act of composing creates new questions about the writer’s own knowledge.

5. Flower and Hayes: Writing Is Recursive Problem Solving

Linda Flower and John Hayes’ influential cognitive process theory of writing challenged simple stage models where writers plan once, draft once and revise at the end.

Real writing is recursive. Writers plan while drafting, discover problems while revising, generate new ideas while editing, and change goals when the emerging text exposes a better route.

This recursive architecture fits knowledge transforming well. Transformation happens because processes interact rather than because the writer follows a perfect linear sequence.

6. The Rhetorical Problem Gives Knowledge a Job

“Write everything you know about photosynthesis” invites knowledge telling.

“A classmate says plants get their food from soil. Write an explanation that would change their mind using evidence and a model of matter” creates a rhetorical problem.

Now the writer must decide:

What misconception must be displaced?
What evidence will matter to this reader?
Which mechanism needs explanation?
What can be assumed?
What sequence will make the explanation intelligible?

The content becomes organised by communicative purpose.

7. Audience Is Not Decoration

School writing sometimes treats audience as a label at the top of the page: “Write for parents.”

But a real audience changes knowledge selection.

A technical explanation for engineers can rely on specialist concepts. The same explanation for Primary students needs different examples, vocabulary and sequencing. A persuasive letter to a school principal must anticipate institutional constraints differently from a social-media post to peers.

When audience meaningfully changes the decisions, rhetoric transforms content.

8. Purpose Forces Selection

Strong writing is partly the discipline of leaving relevant material out.

A writer may know ten facts but need only three. Choosing which three requires a model of the purpose.

If the purpose is to explain why a policy failed, background facts matter only if they contribute to the causal account. If the purpose is to evaluate whether the policy was justified, different evidence becomes central.

Knowledge transforming begins when “relevant to the topic” is no longer a sufficient criterion.

9. Argument Forces Ideas Into Relationships

A list of facts can coexist without conflict.

An argument must establish relationships: cause, support, qualification, exception, comparison, consequence and priority.

The moment the writer asks “Does this evidence actually support my claim?”, content has entered a new structure.

The writing task is now transforming knowledge rather than merely displaying it.

10. Counterargument Is a Transformation Engine

A writer who only retrieves supporting evidence can maintain a simple model.

A counterargument creates disequilibrium.

If a strong objection exists, the writer must revise the claim, qualify it, distinguish cases or find stronger evidence.

This is one reason argumentative writing can improve thinking when the task is genuine. The writer cannot keep every initial idea unchanged and still solve the rhetorical problem.

11. Planning Is Not Merely Making an Outline

An outline can be a list of what the writer plans to say. That still may be knowledge telling at a larger scale.

Transformative planning asks:

What problem does the reader need solved?
What is my central claim?
What would count as sufficient evidence?
Which idea must come before another?
Where will a sceptical reader object?
What is unnecessary even if interesting?
What do I still need to learn before drafting?

The plan becomes a model of reasoning, not a table of contents.

12. Research Changes When the Writer Has a Problem

Students often research by collecting information first and deciding what to say later.

This creates source piles.

Knowledge transforming gives research a question structure. The writer seeks evidence to resolve uncertainty: Do studies support this claim? Are there exceptions? What mechanism explains the pattern? Which source is strongest?

Search becomes inquiry rather than accumulation.

13. Drafting Can Generate New Knowledge

Writers sometimes discover what they think only after writing a sentence they did not anticipate.

This is not mystical inspiration. Externalising language changes the cognitive environment. The sentence becomes an object that can be reread, compared and criticised.

Ideas that were simultaneous in working memory become sequential on the page. Contradictions become visible. Missing links acquire locations.

Writing extends thinking by making thought inspectable.

14. Revision Is Where the Difference Becomes Obvious

Novice revision often concentrates on surface features: spelling, grammar, punctuation, replacing repeated words.

Those edits matter. But they can leave the knowledge structure untouched.

Transformative revision asks larger questions:

Is the thesis still defensible?
Does paragraph three belong?
Does the evidence justify the strength of the claim?
Is the causal explanation reversed?
Would the argument be stronger if two sections changed order?
What has this draft made me realise that the introduction no longer reflects?

The text changes because the writer’s model changed.

15. Revision Can Be Backward Learning

When a writer revises the conclusion, the introduction may need to change. When evidence is weakened, the thesis may need narrowing. When a paragraph becomes stronger, the entire structure may need to be rebalanced.

Later discoveries propagate backward.

This is why writing cannot be perfectly planned in advance. The emerging artifact changes the problem.

16. Writing About Knowledge Is Different From Transforming Knowledge

A student can produce a polished summary of a textbook without changing their own understanding much.

Transformation requires decisions that the source cannot make for the writer.

Compare, reconcile, evaluate, explain, prioritise, apply, synthesise or argue. These operations force knowledge into new relationships.

The more the task can be completed by copying the source’s structure, the less transformation is required.

17. Summary Can Still Transform Knowledge

Summary is not automatically shallow.

A strong summary requires deciding what is central, what is supporting, what can be compressed and how ideas relate.

If students must produce a fifty-word summary of a complex chapter, the compression forces hierarchy.

That can transform the learner’s representation of the material because everything can no longer remain equally important.

18. Explanation Is a Transformation Task

“Explain why” exposes different gaps from “state what.”

To explain, the writer must identify mechanism, sequence, dependency or cause. Missing relationships become visible.

This connects knowledge transforming with self-explanation and generative learning, but writing adds a rhetorical constraint: the explanation must work for a reader who does not share the writer’s internal context.

19. Audience Creates a Test of Completeness

A writer can omit a step because it feels obvious internally.

A real reader exposes that compression.

“How did you get from this claim to that conclusion?” is not merely editorial feedback. It reveals a missing relation in the writer’s model.

Writing for another mind creates pressure toward explicit structure.

20. Knowledge Transforming in Primary Writing

Young writers do not need university essays to practise transformation.

Ask a Primary student to write a short note persuading a class to adopt one playground rule. The learner must choose which reasons matter to classmates, anticipate objections and organise them.

Or ask a child to explain to a younger student why 1/4 is smaller than 1/3 even though 4 is larger than 3.

The content and rhetorical spaces can be simple while the transforming mechanism is already present.

21. Knowledge Transforming in Secondary English

A literature essay becomes transformative when the student must reconcile multiple pieces of textual evidence with a claim.

If one quotation supports the idea that a character is selfish and another reveals sacrifice, the writer cannot simply collect both. The interpretation needs refinement.

Perhaps the character changes. Perhaps self-interest and loyalty coexist. Perhaps the apparent sacrifice is strategic.

The rhetorical demand for a coherent argument changes the literary model.

22. Knowledge Transforming in Science

A laboratory report can be knowledge telling: method, result, conclusion copied into expected sections.

It becomes transformative when the student must explain anomalous data, evaluate uncertainty, compare competing mechanisms and decide what the evidence actually supports.

Writing forces the learner to integrate data and theory rather than keep them as parallel classroom objects.

23. Knowledge Transforming in Mathematics

Mathematical writing is often reduced to showing steps.

But proving, justifying and explaining can transform mathematical knowledge.

“Why must this be true?” requires the learner to distinguish examples from general argument. “Explain why this method fails when x = 0” exposes boundary conditions. “Compare two solution methods and defend which is more efficient here” changes the relationship among procedures.

Writing makes the structure of the mathematics accountable.

24. Knowledge Transforming in Humanities

History and geography become transformative when the task cannot be solved by chronology or fact accumulation alone.

“Which factor mattered most, and under what conditions?” forces prioritisation. “How would the outcome change if this constraint were removed?” forces counterfactual reasoning. “Why do two historians interpret the same event differently?” forces source and framework comparison.

The writer must restructure knowledge to answer the rhetorical question.

25. Prompt Design Decides Which Model the Task Rewards

Prompts can invite knowledge telling even when teachers hope for deep thinking.

“Discuss pollution” is broad enough that students can dump topic knowledge.

“A city has reduced car traffic but air quality has improved only slightly. Explain the most plausible reasons and recommend the next intervention using evidence” creates constraints that require transformation.

Depth is partly designed into the rhetorical problem.

26. Graphic Organisers Help Only If They Represent Decisions

A mind map can simply become a colourful list of facts.

An organiser becomes transformative when relationships are required: cause/effect, claim/evidence, problem/solution, similarity/difference, chronology/turning point.

The visual format should make the reasoning architecture visible.

27. Feedback Should Return the Problem to the Writer

“Rewrite this sentence as…” may fix the text while leaving the writer unchanged.

Transformative feedback identifies the rhetorical problem:

Your evidence supports correlation, but your claim says causation.
The reader still does not know why this example matters.
This paragraph answers a different question from the thesis.
You have two explanations here. Which one accounts for the anomaly?
The counterargument is stronger than your current qualification.

The learner must make the repair decision.

28. Feedback Literacy Makes Transformation More Likely

A writer who sees feedback only as correction will patch sentences.

A writer who can interpret feedback as information about the underlying rhetorical problem is more likely to revise structure and meaning.

This is where How Feedback Literacy Works connects directly to knowledge transforming.

29. Peer Review Trains Rhetorical Perspective

Reading another student’s work makes audience problems easier to see because the reviewer does not share the writer’s invisible assumptions.

Peer review becomes useful when students diagnose specific functions rather than exchange praise.

Where did you stop believing the claim? What information did you need earlier? Which paragraph changed your mind? Which term is undefined? What counterexample remains unanswered?

Giving this feedback develops the evaluator the learner later needs for self-revision.

30. Metacognitive Prompts Can Expose the Problem Space

Research on writing instruction has repeatedly explored metacognitive support because complex writing requires monitoring and control.

Useful prompts include:

What is the reader supposed to understand after this section?
What evidence changed your own view?
Where is the weakest link in the reasoning?
What are you assuming the reader already knows?
Which claim is stronger than the evidence permits?
What did you learn while drafting that should change the plan?

These prompts make the normally invisible composing decisions easier to inspect.

31. Knowledge Transforming and Research Notes

Notes organised by source encourage source-by-source writing.

Notes organised by problem encourage synthesis.

Instead of “Source A says…, Source B says…”, organise evidence under questions: What causes the effect? What evidence contradicts the dominant explanation? Which cases do not fit? What mechanism is best supported?

The note system can either preserve fragmented knowledge or prepare transformation.

32. Knowledge Transforming and AI

Generative AI can either deepen or bypass the transformation process.

If the learner enters a prompt and accepts a finished essay, the system has solved most rhetorical problems externally. The student may receive polished text without constructing the knowledge relations that make the text coherent.

Used differently, AI can become a problem generator:

Produce the strongest counterargument to my claim.
Which sentence contains an unsupported leap?
List two different structures for this evidence and explain the trade-off.
What would a sceptical reader ask here?
Which term is doing too much conceptual work?
Generate a case my explanation fails to cover.

The tool should increase the quality of the rhetorical problem, not remove the writer from it.

33. AI Output Still Needs Knowledge Transforming

A fluent draft can create the illusion that the thinking is finished.

The learner should still ask whether the argument fits the task, whether claims are supported, whether the source basis is trustworthy, whether examples are relevant and whether the structure expresses the learner’s actual position.

Editing AI text at the level of commas does not create ownership. Reconstructing the argument can.

34. Cross-Domain Comparison: Engineering Design

An engineer does not list everything known about materials and call it a bridge.

The design problem imposes constraints: load, span, cost, maintenance, weather, regulation and construction method.

Those constraints transform the relevance of knowledge. A material property becomes important because of a design requirement.

Writing works similarly. Rhetorical constraints turn information into decisions.

35. Cross-Domain Comparison: Legal Argument

A lawyer may know many cases, but legal writing requires selecting which authorities apply, distinguishing adverse cases and fitting facts to legal standards.

The argument transforms the raw knowledge base into a position constrained by audience, precedent and burden of proof.

Again, the output problem reorganises knowledge.

36. Cross-Domain Comparison: Teaching

A teacher preparing an explanation often discovers gaps in their own understanding.

Knowing the topic is one thing. Explaining it to a particular learner requires selecting examples, anticipating misconceptions and sequencing ideas.

The communicative problem transforms the teacher’s representation of the subject.

This is one reason preparing to teach can itself strengthen learning.

37. A Practical Knowledge-Transforming Writing Protocol

Define the rhetorical problem: audience, purpose, genre, constraints and desired effect.
State the content problem: what question must the knowledge answer?
Build a provisional claim: give the draft something to test.
Map evidence by function: support, qualify, contradict, explain or exemplify.
Identify uncertainty: what do you still not know well enough to write?
Plan relationships: sequence by reasoning, not by the order facts came to mind.
Draft to discover: allow the emerging text to expose new problems.
Interrogate the draft: where does the reader need more, less or different evidence?
Revise meaning first: repair claim, evidence, reasoning and structure before polishing sentences.
Seek a counterview: use peer, teacher or tool to stress-test the argument.
Update the knowledge model: record what the writing process changed in your understanding.
Polish last: grammar and style should serve a structure that already earns them.

38. Failure Mode: The Outline Is a Fact Inventory

The student plans five paragraphs by writing five facts.

Repair: label each paragraph by function—claim, mechanism, comparison, counterargument, consequence—not merely topic.

39. Failure Mode: Research Comes Before the Question

The learner collects twenty sources and later tries to invent an essay that can use them.

Repair: begin with a provisional problem and let research revise it.

40. Failure Mode: Revision Means Copyediting

The student corrects spelling while the thesis remains unsupported.

Repair: revise in layers—argument, structure, paragraph function, sentence logic, then surface accuracy.

41. Failure Mode: Teacher Feedback Rewrites the Text

The student submits weak writing and receives a better version from the teacher.

Repair: return the problem. Ask the learner to choose among alternatives, explain the gap and perform the revision.

42. Failure Mode: Templates Become the Thinking

A structure such as PEEL or claim-evidence-reasoning helps novices organise work, then becomes a rigid container applied whether or not the rhetorical problem fits.

Repair: teach the function beneath the template. Once the function is understood, vary the form.

43. Failure Mode: Sophisticated Vocabulary Hides Knowledge Telling

The essay sounds advanced but remains a sequence of relevant claims with little causal or evidential connection.

Repair: strip the argument into plain language. If the relationships disappear when vocabulary is simplified, the rhetoric was carrying more weight than the reasoning.

44. Failure Mode: AI Removes the Productive Problem

The learner outsources claim formation, structure, evidence selection and revision, then edits wording.

Repair: require the learner to own the problem before using AI—state the claim, map evidence, identify uncertainty, then use the tool to challenge rather than replace those decisions.

45. The Missing-Node Scan

If student essays contain many correct facts but weak arguments, if research becomes copying from sources, if revision focuses almost entirely on grammar, if outlines are inventories rather than reasoning structures, or if learners can explain a topic orally but cannot shape it for a reader, the missing node may be knowledge transforming.

Look for these clues: paragraphs that could be rearranged without changing meaning; evidence placed beside claims without explanation; introductions written before the writer knows the argument and never revisited; feedback producing cleaner sentences but the same conceptual gaps; AI-generated text that students cannot defend; and writers who ask “what else can I add?” instead of “what problem remains unsolved?”

The learner may possess plenty of knowledge. The writing process may simply not be forcing that knowledge to reorganise.

46. The Return Path

Return to the student with a page full of accurate facts.

There are two ways to improve the piece.

One is to add better vocabulary, fix grammar and place the facts in a cleaner order.

The other is to ask what the text is trying to accomplish, what the reader needs, which claim the evidence can actually support, where the argument breaks, what opposing evidence changes and what the writer now thinks after trying to solve those problems.

The second process can change the essay because it changes the knowledge that generates the essay.

Knowledge transforming works when composition becomes a thinking environment: rhetorical demands expose content problems, content discoveries reshape the rhetoric, and the writer finishes with more than a better text—they finish with a better organised mind.

Research and Further Reading

eduKateSG Learning Node Series · 0135 · Previous: 0134 — How Epistemic Curiosity Works.

How Knowledge Transforming Works | Writing Becomes Thinking When Ideas Must Solve the Rhetorical Problem