Singapore School Mathematics Operating Manual · Chapter 5
Many school solutions look like a chain of equal signs or implications. Yet not every step in that chain has the same logical strength. Some operations preserve exactly the same set of solutions in both directions. Others create extra possibilities, delete valid possibilities, or require conditions that must be checked separately.
This chapter develops one of the most important pieces of mathematical control: knowing whether a transformation is reversible. A reversible step allows us to move from one statement to another and back again without changing the solution set. A one-way step may still be useful, but its output must be checked against the original problem.
The examples range from Primary-style inverse operations to Secondary algebra and later exponential, logarithmic and calculus contexts. The deeper principle is the same throughout: when a transformation changes the information carried by the problem, the learner must account for what was lost or introduced.
Equivalence · Squaring and roots · Division and cancellation · Domains and restrictions · Practice · Worked answers
1. An equation is not merely a calculation
Consider 3x + 5 = 20. Subtracting 5 from both sides gives 3x = 15. Dividing both sides by 3 gives x = 5. Each move is reversible because adding 5 and multiplying by 3 recover the previous equation. The equations have exactly the same real solution set.
This is stronger than saying that the arithmetic is correct. It says the transformed statement is logically equivalent to the previous one under the stated domain. The solution x = 5 works in every equation in the chain, and no other real x has been added or removed.
A useful symbol distinction is between “=” inside algebraic expressions and “if and only if” between equivalent statements. Students do not need to write formal logic on every line, but they should know what the chain is claiming. A sequence of transformations should not silently change the set being solved.
For simple linear equations, standard balance operations are usually reversible when their inverse is always defined. Add the same quantity to both sides; subtract the same quantity; multiply by the same non-zero constant; divide by the same non-zero constant. These operations preserve equivalence.
The complication begins when the operation itself is not one-to-one or when its inverse is not always defined.
2. Inverse operations in Primary Mathematics already teach reversibility
Suppose a child thinks of a number, multiplies it by 4, then adds 7 to obtain 31. The forward path is “multiply by 4, add 7”. The reverse path is “subtract 7, divide by 4”. This works because each operation has a well-defined inverse in the given number system.
31 − 7 = 24, and 24 ÷ 4 = 6. Checking forward gives 6 × 4 + 7 = 31. The reverse method is not a trick. It works because the two operations can be undone in reverse order.
Now compare “square a number and get 25”. Squaring does not have a single-valued inverse over the real numbers in the same sense. Both 5 and −5 square to 25. If the original problem says the number is positive, the domain selects 5. Without that restriction, reversing the square requires two branches.
This is why “do the opposite operation” is a useful beginner idea but not a universal algebra law. Mature school Mathematics asks a second question: does the proposed inverse recover every original possibility?
3. Squaring can introduce extraneous solutions
Solve √(x + 1) = x − 1 over the real numbers. Before squaring, note that the square-root side is non-negative, so x − 1 must also be non-negative. Hence x ≥ 1.
Squaring gives x + 1 = (x − 1)² = x² − 2x + 1. Rearranging gives x² − 3x = 0, so x(x − 3) = 0. The squared equation has candidate solutions x = 0 and x = 3.
But x = 0 violates the earlier requirement x ≥ 1. Direct substitution into the original equation also shows the failure: √1 = 1, while x − 1 = −1. Thus x = 0 is an extraneous solution created by the squaring step.
The value x = 3 works: √4 = 2 and 3 − 1 = 2. Therefore the original equation has the single real solution x = 3.
Why did squaring create an extra solution? Because a = b implies a² = b², but the reverse statement is weaker: a² = b² implies a = b or a = −b. Squaring forgets the sign relationship between the two sides.
The repair is not “never square”. Squaring is often exactly the right operation. The repair is to treat the resulting values as candidates and check them against the original equation and its domain.
4. Taking square roots can delete a branch if done carelessly
Solve x² = 49. A common incorrect step is “take square root: x = 7”. This loses the valid solution x = −7.
The correct reasoning is x² = 49, so x = 7 or x = −7. Equivalently, x² − 49 = 0 gives (x − 7)(x + 7) = 0.
It is important to distinguish this from evaluating √49. The symbol √49 denotes the non-negative square root, which is 7. The equation x² = 49 asks for every real x whose square is 49. Those are different mathematical objects.
The positive-square-root notation is single-valued; the inverse relation of squaring over the reals has two branches except at zero. Confusing the two produces systematic missing solutions.
5. Absolute value creates another two-branch inverse
Solve |2x − 1| = 7. Absolute value discards sign, so the expression inside can be 7 or −7.
Case 1: 2x − 1 = 7, giving x = 4. Case 2: 2x − 1 = −7, giving x = −3. Both satisfy the original equation.
Writing only 2x − 1 = 7 assumes the inside is positive without justification. Squaring both sides would also lead to the same pair if followed carefully, but direct case splitting exposes the structure more clearly.
The deeper idea is that a many-to-one transformation cannot generally be reversed by one path. Absolute value maps 7 and −7 to the same output. Recovering the input therefore requires multiple possible branches.
6. Dividing by an expression can delete a solution
Solve x(x − 4) = 0. One tempting move is to divide both sides by x and obtain x − 4 = 0, hence x = 4. But the original equation also has x = 0.
The problem is that dividing by x assumes x ≠ 0. If x = 0, the division is not defined. The step has silently excluded one of the original cases.
The correct zero-product reasoning is x = 0 or x − 4 = 0, giving x = 0 or x = 4.
More generally, cancellation by a variable expression is safe only after the possibility that the expression equals zero has been handled. A factor can be cancelled from a fraction under a non-zero condition; it cannot simply be erased from an equation when that factor itself may create a solution.
7. Cancelling factors in fractions requires domain tracking
Consider the rational expression (x² − 9)/(x − 3). Factorising gives (x − 3)(x + 3)/(x − 3). For x ≠ 3, this simplifies to x + 3.
But the original expression is undefined at x = 3. The simplified formula x + 3 is defined there. Thus the two formulas are equal on the domain x ≠ 3, but they are not identical as unrestricted functions.
This distinction matters in graphing and solving. The graph of the original rational function is the line y = x + 3 with a hole at x = 3. If a student simplifies and forgets the restriction, that missing point disappears from the model.
Cancellation preserves values where the cancelled factor is non-zero. It does not repair the original denominator at the excluded value.
8. Multiplying by an expression can also introduce hidden conditions
Solve 1/(x − 2) = 3. The original domain excludes x = 2. Multiplying by x − 2 gives 1 = 3(x − 2), so 1 = 3x − 6 and x = 7/3.
The candidate 7/3 is not excluded and checks in the original equation. This multiplication is safe because the original equation already requires x − 2 ≠ 0. The domain condition was present before the transformation.
Now imagine starting from an equation such as x/(x − 2) = 0. Multiplying by x − 2 gives x = 0, which is valid. But the transformed equation by itself no longer displays the exclusion x ≠ 2. The restriction must be carried forward separately.
Cross-multiplication is therefore not a magic pattern. It is multiplication by denominators under non-zero conditions. When those conditions matter, they belong in the solution.
9. Logarithms are reversible only inside their domain
Solve log₂(x − 1) = 3. The logarithm requires x − 1 > 0, so x > 1. Converting to exponential form gives x − 1 = 2³ = 8, hence x = 9. The value lies inside the domain.
The transformation log₂(A) = 3 to A = 8 is reversible when A > 0. The positivity condition is not an extra optional check; it is part of what it means for the logarithm to exist over the reals.
Consider log₂(x − 1) + log₂(x − 3) = 3. The domain requires x > 3. Combining logs gives log₂[(x − 1)(x − 3)] = 3, so (x − 1)(x − 3) = 8. Expanding gives x² − 4x − 5 = 0, with algebraic roots x = 5 and x = −1.
Only x = 5 lies in the domain x > 3. The root −1 belongs to the polynomial equation created after the logarithms were combined and exponentiated, but not to the original logarithmic equation.
10. Exponentials preserve order differently depending on the base
For base b > 1, the function bˣ is strictly increasing. Therefore bᵃ = bᵇ implies a = b, and bᵃ < bᵇ implies a < b. These steps are reversible because the exponential function is one-to-one on the real numbers.
For 0 < b < 1, equality remains one-to-one, but inequalities reverse direction because the function is strictly decreasing. For example, (1/2)ˣ < (1/2)³ implies x > 3.
This is another example of why an algebraic-looking manipulation needs structural knowledge. The same symbolic base pattern can preserve equality while changing the direction of an inequality depending on the function’s monotonicity.
11. Multiplying an inequality by an unknown-sign expression is dangerous
Suppose we want to solve (x − 1)/(x + 2) > 0. Multiplying both sides by x + 2 without knowing its sign can produce the wrong inequality direction.
The correct approach is to locate the critical values x = 1 and x = −2, then analyse signs across the intervals. For x < −2, numerator and denominator are both negative, so the quotient is positive. For −2 < x < 1, the signs differ, so the quotient is negative. For x > 1, both are positive.
Thus the solution is x < −2 or x > 1, with x = −2 excluded because the expression is undefined and x = 1 excluded because the inequality is strict.
Alternatively, one can multiply by (x + 2)², which is positive wherever the original expression is defined. That preserves inequality direction while retaining the condition x ≠ −2. Choosing a transformation whose sign is controlled is a form of mathematical engineering.
12. Trigonometric equations also require complete inverse branches
Solve sin θ = 1/2 for 0° ≤ θ ≤ 360°. A calculator inverse sine gives 30°, but this is not the complete solution set. Sine is also 1/2 in the second quadrant, giving 150°.
The inverse-sine function returns a principal value. Solving the equation requires every angle in the specified domain that produces the target sine value.
The same distinction appears for cosine and tangent. A calculator’s inverse function is a way to recover one representative angle. The periodic trigonometric function has many inputs with the same output, so the full equation requires the appropriate symmetry and periodicity.
Again, a many-to-one function cannot be reversed by a single branch unless the domain has first been restricted enough to make the function one-to-one.
13. Differentiation loses information; integration restores a family
If f(x) = x² + 5, then f′(x) = 2x. If g(x) = x² − 17, then g′(x) = 2x as well. Differentiation has removed the constant difference between the functions.
Therefore, if f′(x) = 2x, we cannot conclude f(x) = x². The general antiderivative is f(x) = x² + C, where C is an arbitrary constant.
An initial condition can recover the missing information. If f(3) = 14, then 9 + C = 14 and C = 5.
This is a sophisticated version of the same operating-manual principle. The derivative transformation is not one-to-one over all differentiable functions because every vertical shift has the same derivative. Integration produces a family until an additional condition selects one member.
14. Equivalent expressions are not always equivalent solution steps
The identity (x − 2)(x + 3) = x² + x − 6 holds for every real x. Replacing one side with the other inside an equation preserves the equation exactly because the expressions are equal everywhere on the domain.
By contrast, replacing x with √(x²) is not valid for all real x, because √(x²) = |x|, not x. The two expressions agree only when x ≥ 0.
A transformation can therefore fail even when every symbol looks familiar. The learner must know the identity being used and the conditions under which it is true.
This is why formula recall should include domain and structure, not only surface shape.
15. A compact audit for every non-obvious transformation
Before accepting a transformed equation, ask four questions. Was the operation defined for every current candidate? Is the operation one-to-one on the relevant domain? Could the step have introduced extra candidates? Could it have removed a case?
If the answer to either of the last two questions is yes, mark the result as a candidate set rather than the final solution set. Then substitute back into the original statement or reapply the original restrictions.
This audit is especially useful after squaring, taking roots, cancelling variable factors, multiplying inequalities by variable expressions, using inverse trigonometric functions, combining logarithms, and integrating after differentiation.
The aim is not to make every line slower. It is to identify the small number of transformations where logical control matters most.
16. Independent practice
1. Solve x² = 16 over the real numbers.
2. Solve √(x + 6) = x.
3. Solve |3x + 2| = 8.
4. Solve x(x − 5) = 0 without losing a solution.
5. Simplify (x² − 4)/(x − 2), stating the domain restriction.
6. Solve 2/(x − 1) = 1 and state the excluded value.
7. Solve log₃(x − 2) = 2.
8. Solve log₂(x) + log₂(x − 2) = 3.
9. Solve (x − 4)/(x + 1) ≥ 0.
10. Solve cos θ = 1/2 for 0° ≤ θ ≤ 360°.
11. If f′(x) = 6x and f(2) = 17, find f(x).
12. Explain why replacing x by √(x²) can change an expression when x is negative.
17. Worked answers
1. x = 4 or x = −4. Both square to 16.
2. The square-root side is non-negative, so x ≥ 0. Squaring gives x + 6 = x², hence x² − x − 6 = 0 and (x − 3)(x + 2) = 0. Candidates are 3 and −2. The domain excludes −2. Check x = 3: √9 = 3. Therefore x = 3.
3. 3x + 2 = 8 or 3x + 2 = −8. Thus x = 2 or x = −10/3.
4. Zero-product reasoning gives x = 0 or x − 5 = 0. Hence x = 0 or 5. Dividing by x would incorrectly delete the zero solution.
5. Factor to obtain (x − 2)(x + 2)/(x − 2) = x + 2, valid for x ≠ 2. The simplified expression carries the original exclusion.
6. The original domain excludes x = 1. Multiplying by x − 1 gives 2 = x − 1, so x = 3, which is valid.
7. Domain x > 2. Exponential form gives x − 2 = 9, hence x = 11.
8. Domain x > 2. Combine logs: log₂[x(x − 2)] = 3, so x² − 2x = 8. Rearranging gives x² − 2x − 8 = 0, hence (x − 4)(x + 2) = 0. Candidates are 4 and −2. Only x = 4 lies in the domain.
9. Critical values are x = 4 and x = −1. The quotient is positive on x < −1 and x > 4, zero at x = 4, and undefined at x = −1. Therefore x < −1 or x ≥ 4.
10. θ = 60° or 300°. The inverse cosine principal value alone is incomplete.
11. Integrating gives f(x) = 3x² + C. Since f(2) = 12 + C = 17, C = 5. Therefore f(x) = 3x² + 5.
12. √(x²) means the non-negative number whose square is x², so it equals |x|. If x = −4, then √(x²) = √16 = 4, not −4. The replacement is valid only when x ≥ 0.
18. Continue through the operating manual
Use Case Splitting and Piecewise Reasoning when one mathematical situation naturally breaks into branches. Use Boundary Cases, Endpoints and Extremes when validity changes at equality, zero or an endpoint. Use Discrete and Continuous Models when the type of quantity changes the allowed answers.
Return to the BTT Mathematics Hub for the full Batch 02 route.