Intuition What this page is
The parent note gave you the two master formulas. This page throws every kind of problem at those formulas — different bases, nested functions, negative inputs, products, exponents with variables, a real-world rate, and an exam trap. If a cell in the matrix below exists, you will see it worked out fully here.
The only tools we use are already proven:
d x d ln x = x 1 , d x d ln u = u u ′ , d x d log a x = x l n a 1 .
Here u means "some function of x hiding inside the log," and u ′ is its derivative. Everything else is careful bookkeeping.
Before working anything, let us list every case-class this topic can generate. Each row is a "cell." The worked examples below are tagged with the cell they cover, so together they fill the whole grid.
#
Cell (case class)
What makes it different
Example
C1
Plain ln of a polynomial
pure chain rule u ′ / u
(a)
C2
Non-e base log a
extra ln a 1 factor
(b)
C3
Simplify-first (log laws)
avoid chain rule entirely
(c)
C4
Negative / absolute-value input
domain, ln ∣ x ∣
(d)
C5
Product/quotient inside
log laws turn it into a sum
(e)
C6
Variable in the exponent
logarithmic differentiation
(f)
C7
Real-world rate (word problem)
units, interpretation
(g)
C8
Exam twist / degenerate
limiting value, hidden constant
(h)
We will hit all eight cells with eight examples.
The single fact behind every example is the shape of ln x and its slope 1/ x . Look at the figure: the height climbs but ever more slowly, and the little slope triangles get flatter as you move right.
Intuition Why this shape matters for examples
Every worked answer below will end in something over "the inside." When the inside is big, the slope is small — exactly the flattening you see in the figure. Keep this picture in mind as a sanity check: a log derivative is never negative just because x is large; it shrinks toward zero.
Worked example Differentiate
f ( x ) = ln ( x 2 + 4 )
Forecast: guess the answer before reading. It should be "derivative of the inside over the inside." What is the inside's derivative?
Identify the inside: u = x 2 + 4 .
Why this step? d x d ln u = u u ′ only works once we name u clearly.
Differentiate the inside: u ′ = 2 x .
Why this step? Power rule on x 2 gives 2 x ; the constant 4 contributes 0 .
Assemble: f ′ ( x ) = u u ′ = x 2 + 4 2 x .
Why this step? This is the Chain Rule specialised to ln .
Verify: the inside x 2 + 4 is always ≥ 4 > 0 , so ln is defined for all real x — no domain worry. At x = 0 the slope is 4 0 = 0 : correct, because x 2 + 4 has a minimum there (flat bottom), so its log is flat too. ✓
Worked example Differentiate
g ( x ) = log 10 ( 7 x )
Forecast: the 7 is a constant multiplying x . Will it survive into the answer, or cancel?
Change of base: g ( x ) = ln 10 ln ( 7 x ) .
Why this step? We only proved the clean rule for ln ; convert first (see Logarithm Laws ).
Split the top with a log law: ln ( 7 x ) = ln 7 + ln x , so g ( x ) = ln 10 ln 7 + ln x .
Why this step? ln 7 is a pure constant — its derivative is 0 , so the 7 will vanish.
Differentiate: g ′ ( x ) = ln 10 1 ⋅ x 1 = x ln 10 1 .
Why this step? l n 10 1 is a constant multiplier; only ln x has a slope.
Verify: using the master formula directly, d x d log 10 u = u l n 10 u ′ = 7 x l n 10 7 = x l n 10 1 . Same answer — the 7 cancels, exactly the constant-inside phenomenon. ✓
Worked example Differentiate
h ( x ) = ln x + 1 x 3
Forecast: this looks like a quotient-plus-chain nightmare. Log laws make it a two-term sum. Guess how many terms the answer has.
Expand with log laws: h ( x ) = ln x 3 − ln ( x + 1 ) 1/2 = 3 ln x − 2 1 ln ( x + 1 ) .
Why this step? ln B A = ln A − ln B and ln A p = p ln A turn one hard log into two easy ones (Logarithm Laws ).
Differentiate term by term:
h ′ ( x ) = 3 ⋅ x 1 − 2 1 ⋅ x + 1 1 .
Why this step? Each term is now a plain ln ; the second uses d x d ln ( x + 1 ) = x + 1 1 .
Tidy: h ′ ( x ) = x 3 − 2 ( x + 1 ) 1 .
Why this step? Just cosmetic — leave it as a sum, it is already clean.
Verify: the domain of the original is x > 0 (need x 3 > 0 and x + 1 > 0 ). At x = 1 : h ′ ( 1 ) = 3 − 4 1 = 4 11 = 2.75 . Reasonable and positive — the function is increasing there. ✓
Worked example Differentiate
p ( x ) = ln ∣ x ∣ , and evaluate p ′ ( − 2 )
Forecast: ln of a negative number is undefined — but ln ∣ x ∣ is fine for x < 0 . Do you think the slope formula flips sign on the left?
For x > 0 : ∣ x ∣ = x , so p ( x ) = ln x and p ′ ( x ) = x 1 .
Why this step? No absolute value to worry about on the positive side.
For x < 0 : ∣ x ∣ = − x , so p ( x ) = ln ( − x ) . Chain rule with u = − x , u ′ = − 1 :
p ′ ( x ) = u u ′ = − x − 1 = x 1 .
Why this step? The two minus signs cancel — the slope is 1/ x on both sides.
Combine: p ′ ( x ) = x 1 for all x = 0 .
Why this step? One tidy rule covers the whole domain minus the forbidden point x = 0 .
Verify: p ′ ( − 2 ) = − 2 1 = − 0.5 . The picture in the figure below shows the mirror-image branch; on the left the curve falls as x increases toward 0 , so a negative slope is right. ✓
ln ∣ x ∣ has no derivative at x = − 2 because logs need positives."
Why it feels right: you were told ln needs x > 0 . The truth: the argument ∣ − 2 ∣ = 2 is positive; the absolute value fixes the domain. Only x = 0 is excluded. Fix: derivative x 1 holds for every x = 0 .
Worked example Differentiate
q ( x ) = ln ( x 2 e 3 x )
Forecast: you could chain-rule the product, or split with a log law. Which is fewer steps?
Split: q ( x ) = ln x 2 + ln e 3 x = 2 ln x + 3 x .
Why this step? ln ( A B ) = ln A + ln B , and ln e 3 x = 3 x because ln undoes e (inverse of exp ).
Differentiate: q ′ ( x ) = 2 ⋅ x 1 + 3 = x 2 + 3.
Why this step? Both terms are now elementary; no chain rule needed at all.
Verify (via the messy route): with u = x 2 e 3 x , u ′ = 2 x e 3 x + x 2 ⋅ 3 e 3 x = e 3 x ( 2 x + 3 x 2 ) . Then u u ′ = x 2 e 3 x e 3 x ( 2 x + 3 x 2 ) = x 2 2 x + 3 x 2 = x 2 + 3 . Identical — log laws saved us the algebra. ✓
Worked example Differentiate
y = ( 2 x ) x for x > 0
Forecast: the exponent is a variable, so neither the power rule nor a x rule applies alone. What tool pulls an exponent down? Logs. See Logarithmic Differentiation .
Take ln of both sides: ln y = x ln ( 2 x ) .
Why this step? ln A p = p ln A drags the variable exponent x into a product we can differentiate.
Simplify the right: ln ( 2 x ) = ln 2 + ln x , so ln y = x ln 2 + x ln x .
Why this step? Splitting first makes the product rule cleaner.
Differentiate both sides (Implicit Differentiation on the left):
y y ′ = ln 2 + ( ln x + x ⋅ x 1 ) = ln 2 + ln x + 1.
Why this step? Left side: d x d ln y = y y ′ (chain rule). Right: product rule on x ln x .
Solve for y ′ : y ′ = ( 2 x ) x ( ln 2 + ln x + 1 ) .
Why this step? Multiply back by y = ( 2 x ) x .
Verify: note ln 2 + ln x = ln ( 2 x ) , so y ′ = ( 2 x ) x ( ln ( 2 x ) + 1 ) — the same shape as the parent's x x ( ln x + 1 ) , with x → 2 x inside the log. At x = 1 : y = 2 1 = 2 and y ′ = 2 ( ln 2 + 0 + 1 ) = 2 ( ln 2 + 1 ) ≈ 3.386 . ✓
Worked example A quiet zone's loudness in decibels is
L ( I ) = 10 log 10 ( I 0 I ) , where I is sound intensity and I 0 is a fixed reference. How fast does L change with I ? Evaluate at I = 100 I 0 .
Forecast: more intensity means more loudness — but the log means each extra unit of I adds less. Guess: will d I d L grow or shrink as I grows?
Rewrite: L = 10 ( log 10 I − log 10 I 0 ) ; the second term is constant.
Why this step? log 10 I 0 has zero derivative — only log 10 I varies.
Differentiate using the master rule d I d log 10 I = I l n 10 1 :
d I d L = 10 ⋅ I l n 10 1 = I l n 10 10 .
Why this step? Base-10 log carries the l n 10 1 factor (Cell C2 lesson).
Evaluate at I = 100 I 0 : d I d L = 100 I 0 ln 10 10 = 10 I 0 ln 10 1 .
Why this step? Plug the value in and keep I 0 symbolic (it is a fixed constant with units of intensity).
Verify: units are decibels per unit intensity , and the answer shrinks like 1/ I — exactly the flattening from figure s01. Louder sounds need a bigger jump in intensity to gain the same decibel, which is why decibels are logarithmic. Numerically 100 l n 10 10 ≈ 0.04343 dB per unit (taking I 0 = 1 ). ✓
d x d log a x and state what happens as the base a → 1 + .
Forecast: the formula is x l n a 1 . What does ln a do as a slides down toward 1 ? That is the trap.
Master formula: d x d log a x = x ln a 1 .
Why this step? Direct from change of base; ln a is a constant for fixed a .
Degenerate limit: as a → 1 + , ln a → 0 + , so x ln a 1 → + ∞ .
Why this step? Dividing by something shrinking to 0 blows the slope up.
Interpret: log 1 x is undefined — that is why the parent note demands a = 1 . The derivative diverging signals the forbidden base.
Why this step? Calculus warns you at the boundary: an infinite slope means the function itself has collapsed.
Verify: sanity check the other special base a = e : ln e = 1 gives x ⋅ 1 1 = x 1 , recovering plain ln (parent §3). And a = 2 at x = 1 gives 1 ⋅ l n 2 1 ≈ 1.4427 . Both finite and correct; only a = 1 misbehaves. ✓
Recall Which cell does each problem hit? (cover the right side)
d x d ln ( x 2 + 4 ) ::: C1 — plain chain rule, answer x 2 + 4 2 x .
d x d log 10 ( 7 x ) ::: C2 — extra l n 10 1 , answer x l n 10 1 .
d x d ln x + 1 x 3 ::: C3 — split first, answer x 3 − 2 ( x + 1 ) 1 .
d x d ln ∣ x ∣ ::: C4 — x 1 for all x = 0 .
d x d ln ( x 2 e 3 x ) ::: C5 — x 2 + 3 .
d x d ( 2 x ) x ::: C6 — ( 2 x ) x ( ln ( 2 x ) + 1 ) .
Why does log a x break as a → 1 ? ::: C8 — ln a → 0 , slope → ∞ , base 1 forbidden.
Mnemonic One-line strategy picker
"Split, name, blow-up." Split with log laws when there's a product/quotient/power (C3, C5); name the inside u for a raw chain rule (C1); watch the blow-up / hidden constants at the boundary (C2, C8).
Chain Rule — the engine behind C1, C4.
Logarithm Laws — turns C3, C5 into sums.
Logarithmic Differentiation — the only route for C6.
Implicit Differentiation — used inside C6 on d x d ln y .
Derivative of e^x and a^x — the inverse fact simplifying ln e 3 x in C5.
Standard Limit (1+t)^{1/t} → e — the foundation the parent's 1/ x rests on.
Parent topic — the master formulas this page exercises.