Problem Posing: The Learner Becomes the Question Designer
A student who can change a question intelligently often understands its structure more deeply than a student who can only solve the version already given.
Most mathematics learning is organised around receiving questions. The teacher, textbook or examination sets the conditions; the learner responds. That is necessary, but it leaves one powerful skill underused: problem posing. When students deliberately alter a known problem—change a coefficient, reverse a target, tighten a domain, move a parameter, introduce a boundary or ask for a proof instead of a number—they are forced to identify which features are structural and which are incidental.
This guide develops controlled question mutation as a Secondary 3 A-Math learning tool. It is not about inventing random difficult questions. It is about creating mathematically purposeful variants that reveal the effect of changing one assumption, condition, representation or target at a time.
AI Extraction Box: The Problem-Posing Loop
solve base problem → identify invariant structure → choose one feature to change → predict consequence → write new question → solve it → compare routes → keep only mutations that teach something.
- Surface mutation: change numbers, letters, context or order without changing the deep route.
- Structural mutation: change a condition that forces a different conclusion or method.
- Target mutation: keep the object but ask a different question.
- Boundary mutation: move a parameter to the transition value.
- Reverse mutation: turn output information into a parameter-recovery problem.
- Counterexample mutation: change one condition so a previously true claim fails.
- Transfer mutation: hide the same schema inside a new representation or story.
Start from a Solved Base Problem
Problem posing works best when the base mathematics is already understood. Suppose the learner has solved:
Find the minimum value of x²−6x+11.
The solved structure is:
x²−6x+11=(x−3)²+2 → minimum 2.
Now mutate purposefully.
Mutation 1: Change Only the Constant
Change 11 to 7:
Find the minimum value of x²−6x+7.
The turning-point x-coordinate remains 3 because the x² and x coefficients are unchanged. The minimum shifts vertically. This mutation isolates the role of the constant term.
Mutation 2: Reverse the Opening
Change the leading coefficient sign:
Find the maximum value of −x²+6x−11.
The question now tests whether the learner understands why the extremum changes from minimum to maximum rather than merely repeating a completing-square routine.
Target Mutation Changes the Best Route
Keep x²−6x+11 but change the target:
- Find the minimum value.
- Determine whether real roots exist.
- Sketch the graph.
- Find the range.
- Find the distance from the turning point to the y-axis.
The same mathematical object now activates different representations and reasoning. This is valuable because students learn that topic recognition alone is not enough; the target shapes the route.
Parameter Mutation Creates Families
Replace a fixed coefficient with a parameter:
x²+kx+4=0.
Now many new questions become possible:
- For what k are there two distinct real roots?
- For what k is there one repeated root?
- For what k are there no real roots?
- If k>0, what is the feasible repeated-root value?
- How does the vertex move as k changes?
A single parameter mutation turns one exercise into a family and exposes threshold behaviour.
Worked Mutation 1: Predict Before Solving
Start with x²+4x+4=0, which has a repeated root. Change the middle coefficient to 5. Before calculating, predict: the discriminant increases from 0 to 9, so the repeated root should split into two distinct real roots.
Question mutation is strongest when the learner predicts the effect before solving the new version.
Domain Mutation Reveals Hidden Conditions
Consider a base equation:
x²−5x+6=0.
Solutions are x=2,3. Now mutate it into:
ln(x−1)+ln(4−x)=0.
The algebra may again lead to a quadratic, but now positive log arguments impose 1<x<4. The mutation teaches that similar algebra can live inside a different admissibility system.
Changing the container can change which algebraic answers are allowed.
Reverse Mutation: Turn an Answer into a Condition
A forward question might ask:
For y=x²−4x+7, find the minimum value.
A reverse mutation asks:
Find c such that y=x²−4x+c has minimum value 3.
Now the learner must reconstruct the parameter from a desired output. The same concept is being run backward.
Counterexample Mutation Builds Logical Precision
Suppose a student believes: “If f′(a)=0, then x=a is a maximum.” To construct a counterexample, deliberately pose a function satisfying f′(0)=0 without a maximum.
- f(x)=x² gives a minimum.
- g(x)=x³ gives a stationary point of inflexion.
Creating the counterexample requires understanding exactly which condition is insufficient.
Question Mutation in Trigonometry
Base:
Solve sinθ=1/2 for 0°≤θ≤360°.
Useful mutations:
- change sign: sinθ=−1/2;
- change function: cosθ=1/2;
- change interval: 0°≤θ≤720°;
- change to quadratic: 2sin²θ−3sinθ+1=0;
- reverse: find k such that sinθ=k has exactly two solutions in the interval.
Each mutation targets a different structural feature: sign, quadrant, periodicity, algebraic branching or parameter constraints.
Question Mutation in Differentiation
Base:
Differentiate y=(3x+1)^5.
Mutation ladder:
- change coefficients: (5x−2)^5;
- change outer power: (3x+1)^8;
- add product: x(3x+1)^5;
- add quotient: (3x+1)^5/(x+2);
- target mutation: find gradient at x=1;
- reverse mutation: find x where gradient equals 15.
The ladder progresses from surface variation to structural variation to inverse use.
Question Mutation in Integration
Base:
Integrate 6x−4.
Mutations:
- find the function given f′(x)=6x−4 and f(2)=7;
- find the definite integral from 1 to 3;
- interpret the integral as displacement if the expression is velocity;
- ask for total distance if the expression changes sign;
- reverse: construct a velocity function whose displacement over an interval is zero.
These variants reveal the difference between antiderivative, accumulated change, physical interpretation and inverse construction.
Mutation Should Have a Teaching Purpose
Randomly making coefficients ugly does not necessarily deepen learning. A useful mutation should answer a question such as:
- What happens at the boundary?
- Which condition controls the method?
- Which information is necessary?
- Can the same schema survive a new representation?
- What breaks if one assumption is removed?
- Can the process be reversed?
If the variant teaches nothing new, it is merely another exercise.
The Six Mutation Knobs
| Knob | What to change | What it reveals |
|---|---|---|
| coefficient | number or sign | sensitivity and invariants |
| target | root, extremum, proof, sketch | representation/method choice |
| domain | interval or admissibility | solution filtering |
| structure | product, quotient, composition | theorem/rule triggers |
| direction | forward ↔ inverse | parameter recovery |
| boundary | strict inequality to equality case | regime transitions |
Student-Generated Mini Sets
After mastering one representative problem, ask the learner to produce a four-question mini set:
- one surface variant;
- one boundary variant;
- one target variant;
- one reverse problem.
The learner must also provide a one-sentence explanation of what each mutation is intended to test. This prevents arbitrary question generation.
Use Wrong Answers as Mutation Seeds
If a student makes a recurring error, turn it into a deliberate variant. If they forget ± after solving x²=9, ask them to create two questions: one where √9 is being evaluated and one where x²=9 is being solved. If they confuse local and global maxima, ask them to construct a function with a local maximum that is not global over the full domain.
Problem posing converts an error into a test of the distinction that caused it.
Problem Posing and Transfer
The final stage is to move the schema to a new context. A discriminant/tangency idea learned with a line and parabola can be posed as a parameter problem without the word “tangent”. A rate problem learned with circles can be moved to rectangles or volumes. A logarithmic domain issue can be embedded inside a modelling question.
The learner who can design a transfer version has identified the structure strongly enough to detach it from the original page.
Quality-Control Questions for Student-Generated Problems
- Does the question have enough information?
- Is the intended answer actually possible?
- Are domain and parameter conditions consistent?
- Does the mutation change the feature I intended?
- Are there accidental extra solutions or branches?
- Can I solve the problem myself and verify it independently?
- What mathematical distinction does this question teach?
Writing a valid mathematics question is itself an exercise in logical precision.
Common Failure Modes
| Failure | Cause | Repair |
|---|---|---|
| only makes numbers larger | difficulty confused with structural variation | change target, domain, direction or boundary |
| generated problem has no solution | conditions not checked | solve and verify every new question |
| mutation changes many features at once | learning signal becomes unclear | change one decisive feature first |
| student cannot explain why variant matters | question generation became random | state teaching purpose before writing |
| reverse problem underdetermined | not enough independent information | add one justified condition |
| same schema never moved to new context | surface dependence remains | add transfer mutation |
A 55-Minute Problem-Posing Session
- 10 minutes: solve one base problem and identify its invariant structure.
- 10 minutes: create and solve a surface mutation.
- 10 minutes: create a boundary or domain mutation.
- 10 minutes: reverse the problem into parameter recovery.
- 10 minutes: move the schema into a new context or representation.
- 5 minutes: write what each mutation was designed to test.
What Mastery Looks Like
- The learner can identify which features of a problem are structural and which are surface details.
- The learner can generate valid variants with controlled changes.
- The learner predicts how a mutation should affect the route or answer.
- The learner can create boundary, reverse and transfer versions deliberately.
- The learner uses question design to repair misconceptions.
- The learner verifies self-generated questions for sufficiency, admissibility and completeness.
- The learner increasingly studies mathematics as a space of connected problem families rather than isolated exercises.
Return to the Additional Mathematics Learning Hub
Additional Mathematics Learning Hub | Secondary 3–4 A-Math Guides
