A curve is steep at one point, flat at another and perhaps turning somewhere between. Differentiation gives us a rigorous way to measure that local change. Instead of guessing from a sketch, we can calculate a gradient, write a tangent, locate a turning point and explain whether a quantity is increasing or decreasing. The most important skill is not remembering a page of rules; it is choosing the correct rule and explaining what its answer means.
This independent eduKateSG Ministry of Education V3.0 learning reference develops 2027 Singapore G3 Additional Mathematics differentiation, from power and chain rules to products, quotients, trigonometric and exponential derivatives, tangent and normal equations, stationary points, second derivative testing, connected rates and straight-line motion. It includes twenty-six fully worked original examples, a twelve-stage learning protocol, common errors and a three-student Depth–Load–Transfer workshop.
This resource is not issued or endorsed by MOE or SEAB, and its numbers and applications are fictional. Mathematical results depend on the stated function, interval and assumptions. Physical rate problems are teaching models, not measurements from real engineering systems.
1. The Meaning of a Derivative
The derivative f′(x) describes the instantaneous rate at which f changes with x. On a smooth graph it equals the gradient of the tangent at that x-value. If f′(2)=5, a small increase of x near 2 produces an approximate rise of five times that small change in y, provided the usual differentiability assumptions hold. It does not mean the original function always increases by exactly five everywhere.
2. The Official 2027 G3 Additional Mathematics Scope
The SEAB 2027 G3 Additional Mathematics K341 syllabus, Calculus C1 includes derivatives of rational powers, sin x, cos x, tan x, eˣ and ln x; constant multiples, sums, differences, product and quotient rules, the chain rule, tangents and normals, increasing and decreasing functions, stationary maxima, minima and stationary inflexions, second derivative classification, connected rates, and motion applications. Integration is also in C1 but receives its own separate explanation elsewhere. G3 Additional Mathematics is distinct from G3 Mathematics, and students should match preparation to their actual subject.
3. Gradient Comes Before Formula
A tangent is the straight line that best matches a smooth curve’s direction at the point of contact. For a straight line, the gradient is constant; for a curve, the tangent gradient can vary from point to point. Differentiation therefore links graphical reasoning with algebra. Mark the precise x-value before substituting into the derivative.
4. The Power Rule Has a Domain
For xⁿ where n is a rational number and the real expression is defined and differentiable, d(xⁿ)/dx = nxⁿ⁻¹. The result works for ordinary positive integer powers and also for many roots and negative powers. At points where the original function is undefined, or where its tangent is not finite, a written formula must not be used as though a real derivative existed.
5. Four Questions Before Differentiating
Ask: What variable is changing? What does the expression contain—sum, product, quotient or nested function? Over which domain is it meaningful? What will the derivative be used to find: a gradient, stationary point or rate? This simple audit prevents many rule-selection mistakes.
6. Eight Rooms of Differentiation Reasoning
Room one reads the function and variable. Room two notes its domain. Room three recognises the structure of the expression. Room four chooses a derivative rule. Room five simplifies correctly. Room six substitutes a requested point or solves f′(x)=0. Room seven interprets sign, second derivative or geometric direction. Room eight checks a new task independently. Repairs should target the first room the student cannot explain.
7. Twelve-Stage Function–Rule–Meaning Protocol
Stage 1 — Read the input and output
Write what x and y represent, especially if a rate problem gives metres and seconds. dy/dx is a rate relative to x, not automatically a speed per second.
Stage 2 — Mark domain and restrictions
For ln x, require x>0; for 1/x, exclude x=0. Differentiation cannot rescue a forbidden input that the original function never possessed.
Stage 3 — Decide whether the task needs a derivative
If asked for y at x=2, substitute into the original function. If asked for tangent gradient at x=2, differentiate first, then substitute.
Stage 4 — Classify the algebraic structure
A sum of powers is simpler than a product of two functions. A bracket power with an inner linear expression signals the chain rule. Look before applying memorised rules.
Stage 5 — Apply the elementary rule
Use power, exponential, natural-log or trigonometric derivatives carefully. For trig calculus the standard formulae assume angle measurements in radians.
Stage 6 — Include the inner derivative
For f(g(x)), multiply f′(g(x)) by g′(x). Omitting this inner rate is the classic chain-rule error.
Stage 7 — Keep product or quotient terms together
For uv use u′v+uv′; for u/v use (vu′−uv′)/v². Differentiate the numerator and denominator independently only if a valid algebraic simplification justifies it.
Stage 8 — Simplify without changing restrictions
Factor expressions when useful, but do not cancel a denominator at a forbidden input or create a claim that f′ exists where f did not.
Stage 9 — Use the derivative for the actual question
For tangents, find f′(a) and f(a). For stationarity, solve f′(x)=0. For increasing intervals, inspect the sign of f′ over the specified domain.
Stage 10 — Test the classification
A stationary point may be a maximum, minimum or stationary inflexion. A zero second derivative is not automatically proof of inflexion or either turning type.
Stage 11 — Verify the result
Differentiate an expanded form as a second check when practical. Substitute a point back into the original curve and tangent line; test signs on both sides of a stationary value.
Stage 12 — Transfer to an unfamiliar function
After guided practice, solve a new mixed-rule function with different coefficients and explain the rule choice without a template. This is the independent-learning checkpoint.
8. Twenty-Six Worked Examples: A to Z
9. Worked Example A — A first power derivative
For f(x)=x⁴, f′(x)=4x³. The exponent four comes down as a multiplier, and the power decreases by one. Check by considering how much faster x⁴ changes when x moves away from zero.
10. Worked Example B — Differentiate a sum
For y=3x⁵−2x²+7, dy/dx=15x⁴−4x. The derivative of constant seven is zero because a constant output has no change as x changes.
11. Worked Example C — A negative exponent
For y=x⁻²=1/x² with x≠0, dy/dx=−2x⁻³=−2/x³. The formula does not assign a derivative at x=0, because y itself is undefined there.
12. Worked Example D — A fractional exponent
For y=√x=x^(1/2), dy/dx=(1/2)x^(−1/2)=1/(2√x), for x>0. At x=0 the usual two-sided derivative is not available; domain reasoning matters.
13. Worked Example E — Find a gradient at a point
If f(x)=2x²−3x, then f′(x)=4x−3 and f′(2)=5. Five is the tangent gradient at x=2, not the y-coordinate of the point.
14. Worked Example F — Equation of a tangent
For y=x²−4x+1 at x=3, y=−2 and gradient 2x−4=2. Therefore y+2=2(x−3), or y=2x−8. Verify that (3,−2) satisfies both equations.
15. Worked Example G — Equation of a normal
At the same point, the normal is perpendicular to the tangent of gradient 2, so its gradient is −1/2. Its equation is y+2=−(x−3)/2. A normal is not always obtained by merely negating the tangent gradient.
16. Worked Example H — A product of polynomials
For y=x²(x+1), use u=x² and v=x+1. Then y′=2x(x+1)+x²=3x²+2x. Expanding first gives x³+x² and the same derivative, which provides a useful audit.
17. Worked Example I — A product with a sine function
For y=x² sin x (x in radians), y′=2x sin x+x² cos x. Differentiating only x² and leaving sin x unchanged would omit the second term required by the product rule.
18. Worked Example J — A quotient simplified first
For y=(x²+1)/x with x≠0, rewrite y=x+1/x. Then y′=1−1/x². Algebraic simplification can be quicker than quotient differentiation when the domain remains clear.
19. Worked Example K — Use the quotient rule directly
For y=(2x+1)/(x−1), x≠1, the numerator of y′ is 2(x−1)−(2x+1)=−3. Thus y′=−3/(x−1)². Reversing the subtraction changes the sign.
20. Worked Example L — A bracket to a power
For y=(3x−2)⁵, dy/dx=5(3x−2)⁴×3=15(3x−2)⁴. The factor three is the derivative of the inside expression.
21. Worked Example M — A square root of an expression
For y=√(2x+1), dy/dx=(1/2)(2x+1)⁻¹/²×2=1/√(2x+1), where x>−1/2. Writing 1/(2√(2x+1)) forgets the inner factor two.
22. Worked Example N — A sine chain
For y=sin(2x), dy/dx=2cos(2x). The 2 multiplies the result because the input angle changes twice as fast as x.
23. Worked Example O — A cosine chain
For y=cos(3x), y′=−3sin(3x). Both the negative sign from differentiating cosine and the inner factor three are essential.
24. Worked Example P — A tangent derivative
For y=tan x, y′=sec²x where the function is defined, using radians. At x=π/4, the derivative is sec²(π/4)=2, so the curve’s instantaneous gradient is two there.
25. Worked Example Q — A growing exponential
For y=e^(4x), y′=4e^(4x). The exponential retains its form while the chain rule contributes the inner derivative four.
26. Worked Example R — A natural logarithm
For y=ln(5x), real x>0, the derivative is 5/(5x)=1/x. This does not mean ln(5x)=ln x; the functions differ by ln5, whose derivative is zero.
27. Worked Example S — A logarithm of a quadratic
For y=ln(x²+1), all real x are allowed because x²+1>0. The derivative is 2x/(x²+1). The denominator comes from differentiating ln(u), and 2x from the inside.
28. Worked Example T — A stationary minimum
For f(x)=x²−6x+5, f′=2x−6=0 at x=3. The point is (3,−4). Since f′′=2>0, it is a local minimum; alternatively the derivative changes from negative to positive.
29. Worked Example U — Stationary is not always a turning point
For f(x)=x³, f′=3x², so x=0 is stationary. However f′ stays nonnegative on both sides, and the curve keeps increasing through zero; it is a stationary point of inflexion, not a local maximum or minimum.
30. Worked Example V — Two stationary points
For f(x)=x³−3x, f′=3x²−3=0 at x=−1 and x=1. f′′=6x, so (−1,2) is a local maximum and (1,−2) a local minimum. Check the y-values with the original function.
31. Worked Example W — The maximum of a quadratic
For f(x)=−x²+4x+1, f′=−2x+4 gives stationary x=2; f(2)=5 and f′′=−2<0. The stationary point (2,5) is a maximum.
32. Worked Example X — Increasing and decreasing intervals
For f(x)=x²−4x, f′=2x−4. The function decreases for x<2, has zero gradient at x=2 and increases for x>2. The derivative sign, not just its zero, establishes the intervals.
33. Worked Example Y — A connected-rate model
For a circular region of area A=πr², differentiation with respect to time gives dA/dt=2πr(dr/dt). If r=3 cm and increases at 0.2 cm/s, then dA/dt=1.2π cm²/s at that instant. Units verify the interpretation.
34. Worked Example Z — Straight-line motion
A particle has displacement s(t)=t³−6t²+9t metres. Velocity v=ds/dt=3t²−12t+9 m/s, so v(1)=0. Acceleration a=dv/dt=6t−12 m/s², hence a(1)=−6 m/s². Zero velocity at an instant does not imply zero acceleration.
35. Tangent and Normal Are Two Different Lines
The tangent follows the local gradient of a curve. A normal, at a smooth point with nonzero finite tangent gradient m, has gradient −1/m. At a horizontal tangent, the normal is vertical and cannot be written y=mx+c; at a vertical tangent, the ordinary derivative dy/dx may not exist. Always start from the actual point on the curve.
36. What the First Derivative Sign Tells Us
Positive derivative means a differentiable function is locally increasing; negative derivative means it is locally decreasing. Solving f′=0 identifies candidates for stationary points, but separate intervals and derivative signs determine what happens before and after each candidate.
37. What the Second Derivative Can and Cannot Prove
At a stationary point, f′′>0 confirms a local minimum and f′′<0 a local maximum under the standard smoothness assumptions. If f′′=0, the test is inconclusive: inspect derivative signs or higher-order behaviour. The classic x³ example prevents a misleading rule that every zero gradient is an extremum.
38. Chain Rule Is a Multiplication of Rates
If u=g(x) and y=f(u), then dy/dx=(dy/du)(du/dx). This explains why the derivative of (3x−2)⁵ includes a factor three. Conceptual rate multiplication survives much harder examples, so teach the inside variable explicitly rather than chanting ‘bring down the power’.
39. Units Expose Misinterpretations
If displacement is metres and time is seconds, ds/dt has units metres per second and d²s/dt² has metres per second squared. If the function is a cost in dollars versus quantity in items, its derivative is dollars per item. A correct numerical gradient with incorrect units answers the wrong physical question.
40. Boundaries Matter in Optimisation
An algebraically stationary maximum may not be the greatest value on a restricted interval. For a closed interval, compare allowable stationary points with the endpoints. Likewise, a physically meaningful length cannot be negative merely because the unrestricted quadratic has that formal solution.
41. How to Diagnose an Incorrect Answer
A derivative wrong by a constant multiplier often signals a missing chain factor. A quotient sign error often comes from reversing vu′−uv′. A stationary point that does not lie on the curve comes from substituting into the derivative instead of the original function. Separate method faults from arithmetic slips.
42. Three-Pupil Calculus Workshop
Pupil A identifies domain and expression structure. Pupil B chooses and applies the derivative rule. Pupil C checks the result against a second representation and interprets its sign or units. The tutor rotates roles and inspects the first incorrect transformation. Each pupil then independently solves a different tangent, stationary-point or connected-rate question.
43. eduKate Depth–Load–Transfer Diagnostic
Depth: Can the learner explain why the chain rule needs the inside derivative? Load: Can they combine quotient structure, sign and domain in one written solution? Transfer: Can they classify a new stationary point without a model answer? The immutable Clementi three-pupil teaching reference guides the diagnose–repair–practise–prove approach, while this content is explicitly G3 Additional Mathematics.
44. Parent and Tutor Questions
Ask ‘What is changing?’, ‘Why is this the product rule rather than just the power rule?’, ‘Which expression gives the point on the curve?’, ‘How do you know it is a maximum?’, and ‘Can you test one new function alone?’ Praise a correct explanation and clear independent transfer, not just fast symbolic working.
45. Frequently Asked Questions
Is differentiation only a set of formulas?
No. The rules allow you to calculate how a function changes, but gradients, turning points and rates still need interpretation and valid domain conditions.
When should I use the chain rule?
Use it for compositions such as sin(3x), e^(2x) or (x²+1)⁴. Multiply the outer derivative by the derivative of the inner function.
Is a derivative of zero always a maximum?
No. A stationary point can be a minimum, maximum or stationary inflexion. Inspect signs or a valid second-derivative test.
What is the difference between tangent and normal?
The tangent follows the curve’s instantaneous direction. A normal is perpendicular to that tangent at the same point.
Can I differentiate x to any fractional power?
Use the real power rule where the expression is defined and differentiable. Check restrictions and endpoints, especially roots and negative exponents.
Why do sine and cosine derivatives assume radians?
The standard calculus formulas rely on angles measured in radians. Differentiating degree-input functions needs conversion factors.
Can a student get the right derivative but a wrong application answer?
Yes. Confusing f(a) with f′(a), forgetting units or failing to check domain can spoil a correct symbolic derivative.
What should I practise after learning the power rule?
Mix one product, one nested function and one graph-classification task. Independent rule selection matters more than doing twenty near-identical power examples.
46. Independent Depth and Transfer Exercises
Exercise 1 — Tangent check
For y=x²+2x at x=1, the point is (1,3), derivative is 2x+2 and gradient four. Write y−3=4(x−1) and verify the point.
Exercise 2 — Mixed chain and quotient
Differentiate (x²+1)/(x+2) for x≠−2, using quotient rule. Show numerator 2x(x+2)−(x²+1)=x²+4x−1 over (x+2)².
Exercise 3 — Exponential rate
For y=e^(−3x), find dy/dx=−3e^(−3x). Explain why it is negative for all real x and what that means graphically.
Exercise 4 — Two tests for stationarity
For y=x³−12x, solve 3x²−12=0 giving x=±2. Use second derivative 6x to classify and substitute into the curve for coordinates.
Exercise 5 — Connected rate
For V=s³, let s=2 cm and ds/dt=0.1 cm/s. Derive dV/dt=3s²(ds/dt)=1.2 cm³/s. State that the result is instantaneous.
Exercise 6 — Unfamiliar transfer
A new graph has derivative f′(x)=(x−1)². Explain why x=1 is stationary but f′ does not change sign there. Do not classify it as a maximum without further evidence.
Extra Transfer Check — A Tangent Really Is Local
Take f(x)=x². At x=2, the curve passes through (2,4) and f′(2)=4, so its tangent is y=4x−4. At x=2.01, the true curve gives 4.0401 while the tangent gives 4.04. The small difference 0.0001 illustrates local approximation: tangent values become close near the point of contact, but the tangent is not the entire curve.
Extra Transfer Check — A Zero Second Derivative Is Inconclusive
Compare f(x)=x³ and g(x)=x⁴ at x=0. Both have first and second derivatives equal to zero there. Yet x³ passes through a stationary inflexion while x⁴ has a local minimum. When the second-derivative test produces zero, inspect the first derivative’s signs or the local function behaviour before naming the point.
Extra Transfer Check — Average Versus Instantaneous Rate
For f(x)=x², the average rate between x=1 and x=3 is (9−1)/(3−1)=4. The instantaneous rate at the midpoint x=2 is also f′(2)=4, but this equality is a consequence of that particular quadratic and interval. In general, calculate the required secant gradient or tangent gradient rather than treating them as interchangeable.
Extra Transfer Check — An Endpoint May Be an Extremum
The real function f(x)=√x has domain x≥0. Its least value is zero at the endpoint x=0, although the standard finite derivative is not available at that boundary. A problem asking for a greatest or least value on a restricted domain must consider endpoints as well as interior stationary points.
Extra Transfer Check — Roots and Stationary Points Are Not the Same
For y=x²−4, roots occur when y=0, giving x=−2 and x=2. Its stationary point occurs when y′=2x=0, giving x=0 and y=−4. The derivative solves a gradient question; the original equation solves an intercept question. Writing the wrong equation first can produce perfectly calculated but irrelevant numbers.
Extra Transfer Check — Acceleration While Stopped
Suppose a fictional particle has displacement s(t)=t²−4t metres for t≥0. Velocity is v(t)=2t−4 m/s, which equals zero at t=2 s. Acceleration is a(t)=2 m/s², so the object can be momentarily at rest while its velocity continues to change. Explain the distinction before deciding whether a negative velocity means negative speed.
Extra Transfer Check — Rate Units Stay Attached
If A is a measured area in cm² and t is time in seconds, dA/dt is in cm²/s. If the same numeric function instead represents distance in metres versus time, its derivative has units m/s. The symbol dy/dx gives a numerical rule, but the physical interpretation comes from defining what x and y measure.
Extra Transfer Check — An Independent Diagnostic
Give a student f(x)=x²eˣ and ask for the derivative without a prompt naming the rule. A correct solution is f′(x)=2xeˣ+x²eˣ=eˣ(x²+2x). Next ask for a tangent gradient at x=0, giving zero. This two-step task distinguishes remembering a product formula from applying it, simplifying and interpreting a requested point.
47. Closing Principle
Understand the derivative as a local rate; identify the structure; apply a valid rule; protect the domain; use the result to answer the exact tangent, turning-point or rate question; and verify through a different representation. That reasoning is transferable mathematics.
Related eduKateSG References and Verified Sources
Connect to Reference 213 — Quadratic Graphs, Reference 246 — A-Math Polynomials, Reference 273 — Binomial Expansions, the MOE V3.0 library, the Clementi instructional standard, and SEAB 2027 K341 Additional Mathematics C1.
MOE V3.0 Reference Sequence Navigation
Read the preceding Reference 273 and browse the MOE V3.0 library. Keep each result tied to the original function and question. · Next: Reference 275