This page is the exhaustive drill room for higher-order derivatives . Before touching any example, we lay out a scenario matrix : a checklist of every kind of situation that higher-order derivatives can throw at you. Then every example below is stamped with the cell it fills, so by the end you have seen them all .
Intuition Why a matrix first?
A derivative is just "how fast does the previous thing change?" But the situations vary: the function might be a polynomial that eventually dies, a trig function that cycles forever, a physics word problem with units, or a graph where signs flip. If we don't list the cases, we secretly assume the reader has met them. So we make the list visible and then knock every box down.
Definition Reminder — what
f ( n ) ( x ) means
Throughout this page, f ( n ) ( x ) is the ==n -th derivative== of f : the result of differentiating f a total of n times. So f ( 1 ) = f ′ , f ( 2 ) = f ′′ , f ( 3 ) = f ′′′ , and in general f ( n ) ( x ) = d x d [ f ( n − 1 ) ( x ) ] , with f ( 0 ) = f . The parentheses around the n mean "apply the derivative n times" — it is not a power. (Full treatment in the parent note .)
Cell
Case class
What makes it different
Filled by
A
Polynomial — all orders until it dies
Each derivative lowers degree; eventually 0
Ex 1
B
Product / more complex algebra
Power rule alone isn't enough; degree still finite
Ex 2
C
Trig — never dies, cycles
Derivatives repeat with period 4
Ex 3
D
Kinematics word problem (units! )
position → velocity → acceleration with metres/seconds
Ex 4
E
Sign of f ′′ across all regions (concave up / down / zero)
Every sign case + genuine inflection
Ex 5
E′
False candidate : f ′′ = 0 at a point with no sign change
Zero of f ′′ that is not an inflection point
Ex 6
F
Degenerate input : linear & constant functions
f ′′ = 0 everywhere — no curvature
Ex 7
G
Limiting / asymptotic behaviour of high orders
What happens as n → ∞ (exponential vs. polynomial)
Ex 8
H
Exam twist : notation trap d x 2 d 2 y vs ( d x d y ) 2
Reading the symbol correctly
Ex 9
Every cell A–H (plus E′) is covered below. Signs are handled in all directions (positive, negative, zero), including the crucial trap where f ′′ = 0 but the concavity does not flip; degenerate and limiting inputs also get their own examples so you never meet an unshown scenario.
Worked example Example 1 — differentiate
x 5 until nothing is left
Find f ′ ( x ) through f ( 6 ) ( x ) for f ( x ) = x 5 .
Forecast: guess — after how many derivatives does x 5 become the number 0 ? Write your guess before reading on.
f ′ ( x ) = 5 x 4 . Why this step? Power rule : bring the exponent 5 down as a multiplier, drop the exponent by one.
f ′′ ( x ) = 20 x 3 . Why? Differentiate 5 x 4 : 5 ⋅ 4 = 20 , exponent 4 → 3 .
f ′′′ ( x ) = 60 x 2 . Why? 20 ⋅ 3 = 60 , exponent 3 → 2 .
f ( 4 ) ( x ) = 120 x . Why? 60 ⋅ 2 = 120 , exponent 2 → 1 .
f ( 5 ) ( x ) = 120 . Why? 120 ⋅ 1 = 120 , and x 1 → x 0 = 1 , leaving the pure number 120 .
f ( 6 ) ( x ) = 0 . Why? The derivative of any constant is 0 — the slope of a flat line.
Verify: f ( 5 ) ( x ) = 120 = 5 ! (five factorial = 5 ⋅ 4 ⋅ 3 ⋅ 2 ⋅ 1 ). A degree-n polynomial has f ( n ) = n ! and f ( n + 1 ) = 0 — here n = 5 , so it dies at the 6th derivative. ✓
Intuition The "death" of a polynomial
Each derivative shaves one degree off. Start at degree 5, and after 5 shaves you reach degree 0 (a constant), and one more shave gives 0 . This is why polynomials cannot have interesting behaviour "forever" — unlike trig (Cell C).
Worked example Example 2 — a product expanded first
Let g ( x ) = x 2 ( x − 3 ) = x 3 − 3 x 2 . Find g ′′ ( x ) and g ′′′ ( x ) , and check whether x = 1 is a genuine inflection point.
Forecast: it's degree 3 — how many derivatives until it dies?
Expand: g ( x ) = x 3 − 3 x 2 . Why this step? Expanding turns a product into a sum, so we can use the power rule term-by-term instead of the product rule — simpler and less error-prone here.
g ′ ( x ) = 3 x 2 − 6 x . Why? Power rule on each term: x 3 → 3 x 2 , − 3 x 2 → − 6 x .
g ′′ ( x ) = 6 x − 6 . Why? Differentiate again: 3 x 2 → 6 x , − 6 x → − 6 .
g ′′′ ( x ) = 6 . Why? 6 x → 6 , and − 6 (a constant) → 0 .
Solve g ′′ ( x ) = 6 x − 6 = 0 ⇒ x = 1 , then test the sign of g ′′ on each side . Why? A zero of g ′′ is only a candidate — it is a genuine inflection point only if g ′′ actually changes sign there.
g ′′ ( 0 ) = 6 ( 0 ) − 6 = − 6 < 0 → concave down to the left of 1 .
g ′′ ( 2 ) = 6 ( 2 ) − 6 = + 6 > 0 → concave up to the right of 1 .
Sign flips − → + , so x = 1 is a real inflection point.
Verify: g ′′ ( 0 ) = − 6 < 0 and g ′′ ( 2 ) = + 6 > 0 confirm the sign change at x = 1 ✓, so the inflection is genuine (not a false candidate). And g ( 4 ) = 0 , confirming a degree-3 polynomial dies after 4 derivatives. ✓
Worked example Example 3 —
cos x and the period-4 cycle
Let h ( x ) = cos x . Find h ( 1 ) through h ( 4 ) and a formula for h ( n ) .
Forecast: unlike x 5 , this one never becomes 0 . Guess what h ( 4 ) equals compared to h .
h ′ ( x ) = − sin x . Why this step? The derivative of $\cos$ is − sin (as the cosine wave falls, its slope is negative sine).
h ′′ ( x ) = − cos x . Why? Differentiate − sin x : derivative of sin is cos , keep the minus.
h ′′′ ( x ) = sin x . Why? Differentiate − cos x : − ( − sin x ) = sin x .
h ( 4 ) ( x ) = cos x = h ( x ) . Why? Differentiate sin x → cos x . We are back to the start after 4 steps.
General formula: h ( n ) ( x ) = cos ( x + 2 nπ ) . Why? Each derivative shifts a cosine left by a quarter-turn 2 π ; doing it n times shifts by 2 nπ .
Verify: put n = 2 : cos ( x + π ) = − cos x ✓ (matches step 2). Put n = 4 : cos ( x + 2 π ) = cos x ✓. The cycle has period 4 .
Intuition Why this period-4 cycle IS the Taylor-series engine for trig
A Taylor series rebuilds a function near x = 0 from the values h ( 0 ) , h ′ ( 0 ) , h ′′ ( 0 ) , h ′′′ ( 0 ) , … of all its derivatives. For cos x those values, in order, are 1 , 0 , − 1 , 0 and then they repeat (h ( 4 ) ( 0 ) = 1 again, h ( 5 ) ( 0 ) = 0 , h ( 6 ) ( 0 ) = − 1 , … ). Because the pattern is 1 , 0 , − 1 , 0 , 1 , 0 , − 1 , 0 , … , only the even -order terms survive, alternating in sign:
cos x = 1 − 2 ! x 2 + 4 ! x 4 − 6 ! x 6 + ⋯
The coefficient of x n is exactly n ! h ( n ) ( 0 ) — so the period-4 cycle you saw in the figure is literally the reason every other term drops out and the signs flip + , − , + , − . That is the concrete link, not a throwaway phrase.
The figure below plots h = cos x (white) together with its first three derivatives h ′ = − sin x (cyan), h ′′ = − cos x (amber), and h ′′′ = sin x (dashed cyan). The vertical dotted white line at x = 0 marks off the derivative values 1 , 0 , − 1 , 0 used above; each curve is the previous one shifted left by a quarter-turn, and the amber arrow shows that the 4th derivative lands back on the white cos x curve — the period-4 cycle made visible.
Figure 1 — Cell C: cos x (white solid) and its derivatives −sin x (cyan solid), −cos x (amber solid), sin x (cyan dashed) plotted over one full turn from 0 to 2π on a blueprint grid; a dotted vertical line at x=0 reads off the repeating derivative values 1, 0, −1, 0, and an amber arrow points from the far-right of the amber curve back up to the white cos-x curve to show that a fourth differentiation returns to the start — the period-4 cycle that powers the cosine Taylor series.
Worked example Example 4 — position → velocity → acceleration
A drone rises with height s ( t ) = t 3 − 9 t 2 + 15 t (metres, t in seconds). Find velocity v ( t ) , acceleration a ( t ) , and the instant the acceleration is zero.
Forecast: acceleration is the second derivative. Guess whether it's zero at a whole-number time.
v ( t ) = s ′ ( t ) = 3 t 2 − 18 t + 15 (m/s). Why this step? Velocity is the first time-derivative of position — how fast height changes. Units: metres per second.
a ( t ) = v ′ ( t ) = 6 t − 18 (m/s²). Why? Acceleration is the derivative of velocity = the second derivative of s , written s ¨ . Units: (m/s) per second = m/s².
Set a ( t ) = 0 : 6 t − 18 = 0 ⇒ t = 6 18 = 3 s. Why? Zero acceleration is where velocity momentarily stops changing (a velocity extremum).
Verify: at t = 3 : a = 6 ( 3 ) − 18 = 18 − 18 = 0 ✓. Check units flow: metres → m/s (÷ time) → m/s² (÷ time again). Each derivative divides by one more second. ✓
Worked example Example 5 — concavity across all signs
For f ( x ) = x 3 − 3 x , describe the concavity for x < 0 , x = 0 , and x > 0 , and find the inflection point.
Forecast: where does the curve switch from "frowning" ⌢ to "smiling" ⌣ ?
f ′ ( x ) = 3 x 2 − 3 . Why this step? Power rule term-by-term; we need f ′ before f ′′ .
f ′′ ( x ) = 6 x . Why? Differentiate 3 x 2 − 3 : the constant − 3 vanishes.
Case x < 0 : f ′′ = 6 x < 0 → concave down ⌢ . Why? Negative second derivative means the slope is decreasing — the curve bends downward.
Case x > 0 : f ′′ = 6 x > 0 → concave up ⌣ . Why? Positive second derivative means slope is increasing .
Case x = 0 : f ′′ = 0 and it switches sign around it → inflection point at ( 0 , 0 ) . Why? An inflection point is exactly where the bend flips.
Verify: test points: f ′′ ( − 1 ) = − 6 < 0 ✓ (down), f ′′ ( 1 ) = 6 > 0 ✓ (up). Sign genuinely changes at 0 , so ( 0 , 0 ) is a real inflection point. ✓
The figure below draws f = x 3 − 3 x in white. The region x < 0 is shaded cyan (there f ′′ = 6 x < 0 , so the curve is concave down , ⌢ ) and the region x > 0 is shaded amber (there f ′′ = 6 x > 0 , concave up , ⌣ ). The amber dot at the origin marks the inflection point where the shading — and the bend — switches.
Figure 2 — Cell E: the cubic f = x³ − 3x drawn in white on a blueprint grid; the left half-plane (x<0) is shaded cyan and labelled concave-DOWN because f''=6x is negative there, the right half-plane (x>0) is shaded amber and labelled concave-UP because f'' is positive there, and an amber dot with an arrow marks the genuine inflection point at the origin (0,0) where f''=0 and the concavity flips.
Worked example Example 6 — the inflection point that isn't
For f ( x ) = x 4 , we find f ′′ ( 0 ) = 0 . Is x = 0 an inflection point?
Forecast: a common exam trap — students see f ′′ = 0 and declare an inflection. Guess whether x 4 really changes its concavity at the origin.
f ′ ( x ) = 4 x 3 . Why this step? Power rule; we need one derivative before the second.
f ′′ ( x ) = 12 x 2 . Why? Differentiate 4 x 3 : 4 ⋅ 3 = 12 , exponent 3 → 2 .
Solve f ′′ ( x ) = 12 x 2 = 0 ⇒ x = 0 . Why? This is only a candidate — a zero of f ′′ . We must test the sign on each side before claiming anything.
f ′′ ( − 1 ) = 12 ( − 1 ) 2 = + 12 > 0 → concave up to the left.
f ′′ ( 1 ) = 12 ( 1 ) 2 = + 12 > 0 → concave up to the right.
The sign is positive on both sides — it touches 0 at the origin but does not flip. Why does this matter? An inflection point requires the concavity to actually change . Here x 4 is concave up on both sides (a wide flat-bottomed bowl), so x = 0 is a false candidate , not an inflection point.
Verify: f ′′ ( − 1 ) = 12 and f ′′ ( 1 ) = 12 — same sign, no flip. ✓ Contrast with Example 5 where f ′′ = 6 x genuinely changed sign. The rule: f ′′ = 0 is necessary but not sufficient for an inflection point; you must always confirm a sign change.
f ′′ = 0 , therefore inflection point" — the classic trap
Why it feels right: the inflection test starts by solving f ′′ = 0 , so it's tempting to stop there.
The fix: solving f ′′ = 0 only finds candidates . For x 4 at x = 0 , f ′′ = 12 x 2 ≥ 0 everywhere — concave up on both sides, never flipping — so it is not an inflection point despite f ′′ ( 0 ) = 0 . Always finish with the sign-change check.
The figure below overlays the genuine case f = x 3 − 3 x (from Example 5, white) with the false case f = x 4 (amber). At the origin the white cubic crosses from concave-down to concave-up (real inflection), while the amber quartic just kisses the axis from a concave-up bowl on both sides (no inflection).
Figure 3 — Cell E′: two curves near the origin on a blueprint grid — the white cubic x³−3x, which genuinely changes concavity from down to up at x=0 (a real inflection, marked with an amber dot), and the amber quartic x⁴, whose f''=12x² is positive on both sides so it stays a concave-up bowl and only touches its minimum at x=0 (cyan cross), illustrating that f''(0)=0 alone does NOT guarantee an inflection point.
Worked example Example 7 — the flat cases
Find f ′′ for (a) the constant f ( x ) = 7 and (b) the line f ( x ) = 4 x − 1 .
Forecast: what is the "bend" of a straight line? Guess before computing.
(a) Constant: f ′ ( x ) = 0 , so f ′′ ( x ) = 0 . Why this step? A constant never changes → slope 0 everywhere → and the slope itself never changes → second derivative 0 .
(b) Line: f ′ ( x ) = 4 , so f ′′ ( x ) = 0 . Why? A line has a constant slope 4 ; a constant slope's rate of change is 0 . No bending.
Verify: both give f ′′ ≡ 0 . That is correct and expected — a straight line (and a flat line) has zero curvature : it is neither concave up nor down. These degenerate cases confirm the rule "f ′′ = 0 everywhere ⇔ graph is a straight line." ✓
f ′′ = 0 at a single point is NOT the same as f ′′ = 0 everywhere
In Example 5, f ′′ = 0 only at x = 0 (with a sign flip) → inflection point. In Example 6, f ′′ = 0 only at x = 0 (no flip) → not an inflection point. Here f ′′ = 0 for all x → a straight line, no inflection anywhere. The word "everywhere" is the whole difference.
Worked example Example 8 — exponential vs polynomial as
n → ∞
Compare the n -th derivatives of p ( x ) = x 3 and E ( x ) = e x as n grows large.
Forecast: which one "survives" being differentiated infinitely many times?
Polynomial p ( x ) = x 3 : p ′′′ ( x ) = 6 , and p ( 4 ) ( x ) = 0 , so p ( n ) ( x ) = 0 for all n ≥ 4 . Why this step? Each derivative lowers the degree; a degree-3 polynomial hits 0 after 4 derivatives and stays there.
Exponential E ( x ) = e x : E ′ ( x ) = e x , so E ( n ) ( x ) = e x for every n . Why? The defining property of e x is that it is its own derivative — differentiating never changes it.
Limit view: as n → ∞ , p ( n ) → 0 but E ( n ) = e x stays fixed. Why this matters? In a Taylor series , every term needs the value f ( n ) ( 0 ) ; polynomials contribute finitely many, but e x contributes forever — which is why e x = ∑ n ! x n is an infinite sum.
Verify: p ( 4 ) ( x ) = 0 ✓ (constant 6 differentiates to 0 ). E ( n ) ( 0 ) = e 0 = 1 for every n , so its Taylor coefficients are all n ! 1 — an infinite, never-dying tail. ✓
Worked example Example 9 —
d x 2 d 2 y versus ( d x d y ) 2
For y = x 4 , compute both d x 2 d 2 y and ( d x d y ) 2 , and confirm they are different objects.
Forecast: do you expect the same answer? Commit to yes/no first.
First derivative: d x d y = 4 x 3 . Why this step? Power rule — the shared starting point for both quantities.
Second derivative: d x 2 d 2 y = d x d ( 4 x 3 ) = 12 x 2 . Why? The superscripts 2 mean "apply d x d twice " — it is bookkeeping of the operator, not a power.
Squared first derivative: ( d x d y ) 2 = ( 4 x 3 ) 2 = 16 x 6 . Why? Here the 2 really is a power — we multiply the slope by itself.
Verify: at x = 1 : d x 2 d 2 y = 12 but ( d x d y ) 2 = 16 . Different! ✓ The notation trap is confirmed: d x 2 d 2 y = ( d x d y ) 2 .
Recall Did every cell land?
Below, each line is written as Question ::: Answer — a study-card format. Cover everything after the ::: and try to answer, then reveal.
What does d x 2 d 2 y actually mean?
What is cos ( 4 ) ( x ) , and why?
What are the units of the second time-derivative of position?
If f ′′ ( x 0 ) = 0 , is x 0 automatically an inflection point?
Which "survives" infinite differentiation: x 3 or e x ?
After how many derivatives does x 5 become 0 ? The 6th derivative; f ( 5 ) = 120 = 5 ! , then f ( 6 ) = 0 .
The 4th derivative of cos x equals cos x — trig derivatives cycle with period 4.
Units of acceleration (2nd derivative of position) are metres per second squared, m/s².
For any line f ( x ) = m x + c , the second derivative is 0 everywhere (constant slope → zero curvature).
If f ′′ ( x 0 ) = 0 , is x 0 an inflection point? Not necessarily — you must check that f ′′ changes sign ; e.g. x 4 has f ′′ ( 0 ) = 0 but no sign change, so no inflection.
Under infinite differentiation, e x stays e x forever, while any polynomial eventually becomes 0 .