PRIMARY 5 MATHEMATICS LEARNING GUIDE · BATCH 13 · GUIDE 52
Special cases test whether a mathematical idea still works when the situation reaches an edge, an extreme or an unusual but valid condition. Boundary cases test the smallest, largest, zero, equal or limiting values allowed by the problem. Counterexamples test general claims by searching for one valid case that makes the claim false.
These habits protect learners from overgeneralising patterns, formulas and heuristics. A rule that works in ordinary examples may fail at zero, at equality, at a remainder of zero, at maximum capacity, or when a denominator/reference quantity changes.
This guide develops the Primary 5 reasoning job of testing general claims, probing edges and using counterexamples. It is distinct from Logical Deduction & Elimination: deduction filters cases under known rules; this page questions whether the proposed rule itself survives all valid cases.
Return to the Primary 5 Mathematics Learning Hub. Related owners: Logical Deduction & Elimination · Find a Pattern · Mathematical Modelling.
1. A general claim is stronger than one example
“This works for 20” is not the same as “this always works”. To support an “always” claim, the relationship must survive every valid case in the stated domain.
Primary 5 learners should therefore learn to ask: what happens at the smallest case, largest case, equal case, zero case or a deliberately awkward case?
2. Special cases are chosen because they reveal structure
If a pattern concerns connected squares, test one square, two squares and perhaps a very small row. If a grouping rule concerns remainders, test a case with no remainder. If a comparison concerns “more than”, test the equal case where the difference is zero.
Good special cases are diagnostic.
3. Boundary case: zero
Suppose someone claims, “Multiplying by a number always makes a quantity larger.” Counterexample: multiply by 0. The result becomes 0, not larger.
Even within positive numbers, multiplying by a proper fraction such as 1/2 makes a positive quantity smaller.
Zero and fractions reveal why the original rule is too broad.
4. Boundary case: one
Multiplying by 1 leaves a number unchanged. Dividing a positive number by 1 also leaves it unchanged. One is an identity boundary that often distinguishes “increase” from “no change”.
5. Boundary case: equality
If A and B become equal, their difference is 0. This is the boundary between “A greater than B” and “B greater than A”. Transfer and comparison methods should behave correctly at this equal stage.
6. Boundary case: no remainder
When dividing 120 by 30, the quotient is 4 remainder 0. A grouping rule that says “always add one more container when there is a remainder” must handle remainder 0 separately: no extra container is needed.
7. Boundary case: smallest complete group
If teams require exactly 6 students, 6 students form one complete team. Five form none. These cases identify the threshold where one group becomes possible.
8. Boundary case: full capacity
If a bus holds at most 40 people, 40 requires one bus. Forty-one requires two. The jump occurs exactly after the capacity boundary.
Testing 39,40,41 is more informative than testing 20,25,30.
9. Boundary case: percentage 0%
0% of a quantity is 0. A 0% discount leaves the original price unchanged. A rule about discounts should still make sense at this boundary.
10. Boundary case: percentage 100%
100% of a quantity is the whole quantity. A 100% discount reduces a non-negative price to 0. A 100% increase doubles the original. These edge values clarify the difference between “100% of” and “100% more than”.
11. Boundary case: fraction equal to one whole
5/5, 8/8 and 20/20 all equal 1. They represent the entire reference whole. Any rule treating every fraction as “less than one” is false because improper and whole-equivalent fractions exist.
12. Counterexample to “larger denominator means larger fraction”
Compare 1/3 and 1/5. The larger denominator gives the smaller fraction when numerators are equal and the whole is positive. The example disproves the naive claim.
13. Counterexample to “larger numerator always means larger fraction”
Compare 3/10 and 2/3. The larger numerator 3 does not produce the larger fraction. Different denominators matter.
A general claim about fractions must control the reference whole and denominator.
14. Counterexample to “perimeter determines area”
A 1×9 rectangle and a 4×6 rectangle both have perimeter 20? Check: 1×9 perimeter=20; 4×6 perimeter=20. Areas are 9 and24. Same perimeter, different area.
One valid pair disproves the claim that equal perimeter forces equal area.
15. Counterexample to “same area means same perimeter”
A 2×6 rectangle has area12 and perimeter16. A 3×4 rectangle also has area12 but perimeter14. Same area does not determine perimeter.
16. Counterexample to “bigger-looking shape has bigger area”
A long thin rectangle may look larger in one direction while having less area than a more compact rectangle. Appearance is not measurement.
Use dimensions, not visual impression.
17. Special cases in pattern claims
Suppose stages 2,3,4 appear to follow a rule. Test stage1. Some patterns have a special starting condition because the first shape has no previous neighbour or shared edge.
Boundary stages often reveal hidden assumptions.
18. Special cases in repeated cycles
If a cycle length is 4, positions with remainder 1,2,3 and0 correspond to the first, second, third and fourth items. The remainder-zero case is a special boundary and should be taught explicitly.
19. Special cases in transfer problems
If transfer amount is 0, the state does not change. If transfer amount equals exactly half the initial difference in an internal two-group system, the groups become equal.
These cases reveal the relationship between transfer and difference change.
20. Special cases in constant-difference problems
If equal amounts are added to A and B, their difference remains constant even when that equal amount is 0. The zero-change case confirms the invariant at the boundary.
21. Special cases in rate
A constant rate model at 0 minutes should produce 0 accumulated quantity if there is no starting quantity. If a claimed rule gives nonzero output at time0, it may contain an unacknowledged initial amount.
22. Special cases in money models
If unit price is fixed, buying 0 items should cost $0 unless there is a fixed charge. Testing 0 items can reveal whether a model should be “cost=price×quantity” or “cost=fixed fee+price×quantity”.
23. Boundary cases reveal hidden fixed charges
A taxi model that predicts $4 at distance0 suggests a starting fee. This is not an error if the context includes one. Boundary testing helps interpret model parameters.
24. Counterexamples test universal language
Words such as always, every, all, never invite counterexample search. To disprove “all multiples of 3 are odd”, one example—6—is enough.
To prove the claim true, however, examples alone are not enough.
25. “Sometimes” claims need at least one true example and one false example
Claim: “Adding two numbers gives an even result.” Sometimes true: 3+5=8. Sometimes false: 3+4=7.
Showing both kinds of cases establishes that the result depends on the inputs.
26. “Never” claims can be disproved by one example
Claim: “An odd number plus an odd number is never even.” Counterexample: 3+5=8. Therefore the claim is false.
27. Extreme-value testing
Within a permitted range, test the smallest and largest inputs. If a rule is supposed to stay within capacity or bounds, extremes often reveal failure first.
For example, if a package may weigh at most 20 kg, testing 20 kg and 20.1 kg clarifies the boundary.
28. Boundary cases in inequalities
“At most 40” includes 40. “Less than 40” does not. “At least 12” includes 12. “More than 12” does not.
Testing the boundary number itself prevents language errors.
29. Counterexamples improve definitions
If a learner says “a parallelogram is a slanted rectangle”, show a rectangle that is also a parallelogram under inclusive classification. The counterexample forces a property-based definition rather than an appearance-based one.
30. Special cases in geometry classification
A square satisfies the properties of a rectangle and a rhombus. Special cases reveal hierarchy. The most specific familiar name does not erase broader properties.
Classification rules should survive these nested cases.
31. Special cases in composite figures
If a removed corner has width0 or height0, no area is actually removed. A subtraction formula should reduce correctly to the original rectangle area.
Testing zero-size missing pieces checks the model.
32. Boundary cases in water-level problems
Water depth 0 means empty tank. Water depth equal to tank height means full tank. Any computed depth above tank height requires overflow or signals an impossible in-tank state.
33. Special cases in grouping
If total equals exactly one group size, quotient=1 remainder0. If total is one less, no complete group is possible under exact grouping. If one more, one group plus remainder1.
Testing k−1, k, k+1 around the group-size boundary clarifies interpretation.
34. Use counterexamples to repair overgeneralised heuristics
Heuristic: “When there is a remainder, round up.” Counterexample: question asks “How many complete teams?” For 38 students in teams of6, answer is6 complete teams, not7. The heuristic needs a context condition: round up only when every item/person must be accommodated in whole groups.
35. Error map
| Error | Cause | Repair question |
|---|---|---|
| Accepts a rule after three ordinary examples | No edge testing | What happens at zero, one, equality or the maximum? |
| Thinks one confirming example proves “always” | Evidence overgeneralised | Can you find a valid case that breaks it? |
| Uses counterexample against a claim outside its stated domain | Domain ignored | Is this case actually allowed? |
| Misses remainder-zero case | Cycle/group boundary not tested | What happens exactly at a complete cycle/group? |
| Confuses “at most” with “less than” | Boundary inclusion missed | Is the boundary value allowed? |
36. Testing routine for a general claim
- State the claim precisely.
- Identify its allowed domain.
- Test ordinary cases.
- Test zero/one/equality if allowed.
- Test minimum and maximum boundaries.
- Search deliberately for a counterexample.
- If none appears, explain structurally why the claim should hold rather than relying only on examples.
37. Practice laboratory A
- Disprove: “Multiplying a positive number always makes it larger.”
- Disprove: “Equal perimeter means equal area.”
- Test boundary values 39,40,41 for a bus capacity of at most40.
- For a repeating cycle of length5, what does remainder0 mean?
- Disprove: “Every multiple of4 is also a multiple of8.”
38. Practice laboratory B
- Claim: “Adding two odd numbers is always even.” Test several cases and explain structurally.
- Claim: “A fraction is always less than1.” Give counterexamples.
- A tank height is30 cm. Test depths 0,30,31 cm against empty/full/in-tank conditions.
- Teams require6 students. Interpret totals5,6,7 under “complete teams only”.
- Claim: “Every rectangle is a square.” Give a counterexample.
39. Answers
1. Multiply by 1/2: 10×1/2=5, smaller.
2. 1×9 and4×6 both have perimeter20 but areas9 and24.
3. 39 fits one bus; 40 exactly fills one; 41 requires two.
4. It lands on the last item in the cycle.
5. 4 itself is not a multiple of8.
6. Odd+odd=even because each odd is 2k+1; together 2(k+m+1), an even quantity. At Primary 5 this can also be reasoned with paired counters plus one extra from each number making another pair.
7. 5/5=1 and6/5>1.
8. 0 empty; 30 full; 31 impossible without overflow.
9. 5→0 complete teams; 6→1; 7→1 complete team with1 left.
10. A 2×3 rectangle is not a square.
40. Full claim-testing problem
A student claims: “If two rectangles have the same perimeter, the one with the longer length always has the larger area.” Test the claim for rectangles with perimeter20.
Possible whole-number side pairs include:
- 1×9 → area9;
- 2×8 → area16;
- 3×7 → area21;
- 4×6 → area24;
- 5×5 → area25.
As the rectangle becomes less stretched and approaches a square, the area increases. Therefore “longer length always means larger area” is false. The 1×9 rectangle has the longest length but the smallest area in this list.
The counterexample does more than reject the claim: it reveals a deeper pattern worth investigating.
41. Final checkpoint
A strong Primary 5 learner tests mathematical rules at informative special cases, treats zero, one, equality, remainder zero and capacity thresholds carefully, distinguishes inclusive from exclusive boundaries, searches for counterexamples to universal claims, recognises that one counterexample disproves “always”, and uses edge cases to reveal hidden assumptions in models and heuristics.
Return to the Primary 5 Mathematics Learning Hub.
Wintour House V1.0 · CivDJ · eduKate Publishing: ordinary examples show how a rule behaves; boundary cases show what the rule really means; one good counterexample can save a learner from memorising a false generalisation.
