MATH 203 Calculus I • Concordia University, Montreal

Revision sheet: inverse functions and logarithms (MATH 203)

This sheet is not a summary of section 1.6 of Thomas' Calculus: you already have the course notes. It answers one question only, what makes students lose marks on inverse functions and logarithms in MATH 203 at Concordia University, and which precise gesture avoids each loss. The inverse trigonometric functions are left to their own chapter.

The idea of an inverse is rarely the problem. The marks go in the algebra inside it: an unknown that appears twice and is never collected, a logarithm law that does not exist, and a domain that nobody checked at the end. Your scientific calculator computes ln⁡\ln and log⁡\log only, so it will not catch any of these for you.

Mark this sheet as read or add it to your favourites: a free account, no password, keeps your read sheets and favourites from one visit to the next and tells you which chapter to tackle next. Create your space, an email is enough.

The thread of the chapter

When the unknown sits in two places, xx in a numerator and a denominator, exe^x and e−xe^{-x}, xln⁡3x\ln 3 and xln⁡7x\ln 7, it must be COLLECTED on one side and FACTORED out before anything else; and since an inverse swaps inputs and outputs, every answer ends with a domain check, the piece where ff is one-to-one, the positive arguments of ln⁡\ln, the range that becomes a domain.

This chapter is part of MATH 203, Calculus I (Concordia)

The essentials

An inverse exists where f is one-to-one, and it is found by collecting the unknown

  • • One-to-one: f(a)=f(b)f(a) = f(b) forces a=ba = b; no horizontal line meets the graph twice. One pair of inputs with the same output settles a no.
  • • f−1(b)=a  ⟺  f(a)=bf^{-1}(b) = a \iff f(a) = b. Domain of f−1f^{-1} = range of ff, range of f−1f^{-1} = domain of ff; the graph is reflected in y=xy = x.
  • • Method: y=f(x)y = f(x), clear the denominators, COLLECT every xx term on one side, FACTOR xx out, divide, rename. y(x+4)=1−3xy(x + 4) = 1 - 3x gives x(y+3)=1−4yx(y + 3) = 1 - 4y.
  • • If a ±\pm or a square root appears, the domain of ff chooses: x−1=2−y\sqrt{x - 1} = 2 - y needs y≤2y \le 2, so g−1(x)=1+(2−x)2g^{-1}(x) = 1 + (2 - x)^2 for x≤2x \le 2 only.
  • • f−1(x)f^{-1}(x) is never 1f(x)\frac{1}{f(x)}: the −1-1 means undoing ff by composition.
-2-112345678-2-112345678y = g(x)y = g⁻¹(x)not g⁻¹y = x
g(x)=2−x−1g(x) = 2 - \sqrt{x - 1} and its inverse are mirror images in y=xy = x; the inverse is the solid half of the parabola, the dashed half is not g−1g^{-1}.

The restriction is part of the formula. Without x≤2x \le 2, the formula 1+(2−x)21 + (2 - x)^2 is a whole parabola, which is not one-to-one and is the inverse of nothing.

Logarithms: every law carries a condition

  • • log⁡ax=y  ⟺  ay=x\log_a x = y \iff a^y = x, defined for x>0x > 0 only; a logarithm can be negative, its argument cannot.
  • • ln⁡(xy)=ln⁡x+ln⁡y\ln(xy) = \ln x + \ln y, ln⁡xy=ln⁡x−ln⁡y\ln\frac{x}{y} = \ln x - \ln y, ln⁡xr=rln⁡x\ln x^r = r\ln x: for x>0x > 0 and y>0y > 0 only.
  • • Change of base: log⁡ax=ln⁡xln⁡a\log_a x = \frac{\ln x}{\ln a}, the only way to get log⁡320\log_3 20 from a calculator that has ln⁡\ln and log⁡\log.
  • • ln⁡(ex)=x\ln(e^x) = x for every xx, but eln⁡x=xe^{\ln x} = x for x>0x > 0 only: e2ln⁡x=x2e^{2\ln x} = x^2 holds for x>0x > 0.
  • • No law for ln⁡(x+y)\ln(x + y), for (ln⁡x)2(\ln x)^2, for (ln⁡x)(ln⁡y)(\ln x)(\ln y), for ln⁡xln⁡y\frac{\ln x}{\ln y}.

The rules in table form

Each row reads from left to right: the assumptions, then the result. A red cell is not an answer, it is the finding that the form settles nothing and the instruction to rewrite it. Every case is followed by a worked example.

The laws of logarithms, and the ones students invent

Read a line as: the expression of the first column, under the condition of the second, equals the third. The red lines are rules that do not exist: each is refuted by two numbers, and the method says what to do instead.

ExpressionConditionEquals
log⁡a(xy)\log_a(xy) x>0x > 0, y>0y > 0 log⁡ax+log⁡ay\log_a x + \log_a y

Example: log⁡64+log⁡69=log⁡636=2\log_6 4 + \log_6 9 = \log_6 36 = 2.

ln⁡xy\ln\frac{x}{y} x>0x > 0, y>0y > 0 ln⁡x−ln⁡y\ln x - \ln y

Example: ln⁡8−ln⁡2=ln⁡4\ln 8 - \ln 2 = \ln 4, which checks x=4x = 4 in ln⁡(x+4)−ln⁡(x−2)=ln⁡x\ln(x + 4) - \ln(x - 2) = \ln x.

ln⁡xr\ln x^r x>0x > 0 rln⁡xr\ln x

Example: e−2ln⁡5=eln⁡5−2=125e^{-2\ln 5} = e^{\ln 5^{-2}} = \frac{1}{25}.

log⁡ax\log_a x x>0x > 0, a>0a > 0, a≠1a \ne 1 ln⁡xln⁡a\frac{\ln x}{\ln a}

Example: log⁡320=ln⁡20ln⁡3≈2.7268\log_3 20 = \frac{\ln 20}{\ln 3} \approx 2.7268, between 22 and 33 since 9<20<279 < 20 < 27.

ln⁡(x+y)\ln(x + y) x,y>0x, y > 0 no law rule that does not exist

Example: ln⁡(1+1)=ln⁡2≈0.69\ln(1 + 1) = \ln 2 \approx 0.69, while ln⁡1+ln⁡1=0\ln 1 + \ln 1 = 0.

What to do: Isolate the exponential first, or multiply by exe^x to get a quadratic: ex+6e−x=5e^x + 6e^{-x} = 5 becomes u2−5u+6=0u^2 - 5u + 6 = 0.

(ln⁡x)2(\ln x)^2 x>0x > 0 not 2ln⁡x2\ln x rule that does not exist

Example: At x=ex = e: (ln⁡e)2=1(\ln e)^2 = 1, while 2ln⁡e=22\ln e = 2.

What to do: Set t=ln⁡xt = \ln x and solve the quadratic in tt; only ln⁡(x2)\ln(x^2) equals 2ln⁡x2\ln x.

ln⁡xln⁡y\frac{\ln x}{\ln y} x,y>0x, y > 0, y≠1y \ne 1 not ln⁡xy\ln\frac{x}{y} rule that does not exist

Example: ln⁡8ln⁡2=3\frac{\ln 8}{\ln 2} = 3, while ln⁡82=ln⁡4≈1.39\ln\frac{8}{2} = \ln 4 \approx 1.39.

What to do: A quotient of logarithms is a CHANGE OF BASE: ln⁡xln⁡y=log⁡yx\frac{\ln x}{\ln y} = \log_y x.

The conditions of the first four lines are what produces the rejected roots: combining two logarithms into one enlarges the domain, and the test at the end restores it.

The mistakes that cost marks

These are the errors I correct most often in session. Each one costs marks on a paper, even when the reasoning behind it is right.

1. Leaving x on both sides of an inverse formula

the whole question

What not to write

“y(x+4)=1−3xy(x + 4) = 1 - 3x, so x=1−3x−4yyx = \frac{1 - 3x - 4y}{y}.”

What to write

“xy+4y=1−3xxy + 4y = 1 - 3x, so xy+3x=1−4yxy + 3x = 1 - 4y, x(y+3)=1−4yx(y + 3) = 1 - 4y and x=1−4yy+3x = \frac{1 - 4y}{y + 3}. So f−1(x)=1−4xx+3f^{-1}(x) = \frac{1 - 4x}{x + 3}.”

Why: Nothing is solved while xx is on both sides. Collect every term containing xx, factor xx out, and only then divide.

2. Forgetting the restriction or leaving a plus or minus

2 marks

What not to write

“x−1=2−y\sqrt{x - 1} = 2 - y, so x=1+(2−y)2x = 1 + (2 - y)^2 and g−1(x)=1+(2−x)2g^{-1}(x) = 1 + (2 - x)^2.”

What to write

“x−1=2−y\sqrt{x - 1} = 2 - y requires y≤2y \le 2, so g−1(x)=1+(2−x)2g^{-1}(x) = 1 + (2 - x)^2 for x≤2x \le 2. Check: g(5)=0g(5) = 0 and g−1(0)=5g^{-1}(0) = 5.”

Why: Squaring loses the sign condition. The full parabola gives g−1(3)=g−1(1)=2g^{-1}(3) = g^{-1}(1) = 2, two inputs for one output, so it is not an inverse; the figure of the first block shows the half that is.

3. Turning the square of a logarithm into twice the logarithm

the whole question

What not to write

“(ln⁡x)2=ln⁡(x3)+4(\ln x)^2 = \ln(x^3) + 4, so 2ln⁡x=3ln⁡x+42\ln x = 3\ln x + 4 and x=e−4x = e^{-4}.”

What to write

“With t=ln⁡xt = \ln x: t2−3t−4=0t^2 - 3t - 4 = 0, t=4t = 4 or t=−1t = -1, so x=e4x = e^4 or x=e−1x = e^{-1}, both positive, both kept.”

123456789-3-2-1123456y = (ln x)²y = 2 ln x = ln(x²)(e², 4)x
(ln⁡x)2(\ln x)^2 and 2ln⁡x2\ln x are two different curves that cross only at (1,0)(1, 0) and (e2,4)(e^2, 4); everywhere else, replacing one by the other changes the equation.

Why: The power law moves an exponent INSIDE the logarithm. In (ln⁡x)2(\ln x)^2 the square is outside; the two functions only agree at x=1x = 1 and x=e2x = e^2, as the figure shows.

4. Keeping a root that makes a logarithm negative

2 marks, and a solution that does not exist

What not to write

“x+4x−2=x\frac{x + 4}{x - 2} = x gives x2−3x−4=0x^2 - 3x - 4 = 0, so x=4x = 4 or x=−1x = -1.”

What to write

“Domain: x>2x > 2. The candidates are 44 and −1-1; x=−1x = -1 is rejected because ln⁡(−1−2)\ln(-1 - 2) is undefined. Solution: x=4x = 4.”

Why: The quotient law and clearing the denominator both enlarge the domain. The domain is written BEFORE the algebra and applied after it, with the reason.

5. Rejecting a valid negative root by reflex

half the question

What not to write

“log⁡4(x2)=log⁡4(3x+10)\log_4(x^2) = \log_4(3x + 10) gives x=5x = 5 or x=−2x = -2; −2-2 is negative, so it is rejected.”

What to write

“At x=−2x = -2: x2=4>0x^2 = 4 > 0 and 3x+10=4>03x + 10 = 4 > 0, so both logarithms exist. Solutions: x=5x = 5 and x=−2x = -2.”

Why: A candidate is rejected only when it makes an ARGUMENT zero or negative in the original equation. Its own sign does not matter.

6. Taking the logarithm of a sum term by term

the whole question

What not to write

“ex+6e−x=5e^x + 6e^{-x} = 5, so x+ln⁡6−x=ln⁡5x + \ln 6 - x = \ln 5.”

What to write

“Multiply by ex>0e^x > 0: e2x−5ex+6=0e^{2x} - 5e^x + 6 = 0, (ex−2)(ex−3)=0(e^x - 2)(e^x - 3) = 0, so x=ln⁡2x = \ln 2 or x=ln⁡3x = \ln 3.”

Why: There is no law for ln⁡(a+b)\ln(a + b). When the unknown sits in exe^x and e−xe^{-x}, multiplying by exe^x gathers it into one quadratic in u=exu = e^x.

7. Writing the change of base upside down

1 to 2 marks

What not to write

“log⁡320=ln⁡3ln⁡20≈0.3667\log_3 20 = \frac{\ln 3}{\ln 20} \approx 0.3667.”

What to write

“log⁡320=ln⁡20ln⁡3≈2.7268\log_3 20 = \frac{\ln 20}{\ln 3} \approx 2.7268, consistent with 32<20<333^2 < 20 < 3^3.”

Why: The argument goes on top: log⁡ax=ln⁡xln⁡a\log_a x = \frac{\ln x}{\ln a}. Bracketing by powers of the base catches the inversion in five seconds.

8. Reflecting the graph in the wrong line

the sketch, 2 to 3 marks

What not to write

“The graph of ln⁡x\ln x is the graph of exe^x reflected in the yy-axis.”

What to write

“The graph of ln⁡x\ln x is the reflection of the graph of exe^x in the line y=xy = x: (0,1)(0, 1) becomes (1,0)(1, 0).”

-4-3-2-112345-3-2-112345y = eˣy = e⁻ˣy = ln xy = x
Reflected in the yy-axis, exe^x becomes the dashed e−xe^{-x}; reflected in y=xy = x, it becomes ln⁡x\ln x, the inverse.

Why: The yy-axis reflection replaces xx by −x-x and gives e−xe^{-x}; the inverse swaps the COORDINATES, which is the mirror y=xy = x.

Which method to choose

Which method, by the FORM of the equation

Look at where the unknown sits before writing anything: the form picks the method

  • If powers with unrelated bases, the unknown in both exponents → take ln⁡\ln of both sides, distribute, collect the xx terms, factor

    Example: 32x+1=7x3^{2x + 1} = 7^x: x(2ln⁡3−ln⁡7)=−ln⁡3x(2\ln 3 - \ln 7) = -\ln 3, x≈−4.3715x \approx -4.3715

  • If exe^x and e−xe^{-x} together → multiply by exe^x, set u=exu = e^x, solve the quadratic, reject u≤0u \le 0

    Example: ex+6e−x=5e^x + 6e^{-x} = 5: u=2u = 2 or u=3u = 3, x=ln⁡2x = \ln 2 or ln⁡3\ln 3

  • If a sum or difference of logarithms → write the domain, combine into one logarithm, remove it, test each candidate

    Example: ln⁡(x+4)−ln⁡(x−2)=ln⁡x\ln(x + 4) - \ln(x - 2) = \ln x: x=4x = 4, x=−1x = -1 rejected

  • If logarithms in different bases → change of base to a single base first

    Example: log⁡2x+log⁡4x=6\log_2 x + \log_4 x = 6: 32log⁡2x=6\frac{3}{2}\log_2 x = 6, x=16x = 16

  • If ln⁡x\ln x squared, or ln⁡x\ln x in a quadratic → set t=ln⁡xt = \ln x, solve for tt, then x=etx = e^t

    Example: (ln⁡x)2=ln⁡(x3)+4(\ln x)^2 = \ln(x^3) + 4: t=4t = 4 or −1-1, x=e4x = e^4 or e−1e^{-1}

  • If the unknown in the denominator of an exponent → isolate the power, take ln⁡\ln, solve for the reciprocal, then invert

    Example: 80⋅2−9/h=5080 \cdot 2^{-9/h} = 50: h=9ln⁡2ln⁡(8/5)≈13.27h = \frac{9\ln 2}{\ln(8/5)} \approx 13.27

The last step of every branch is the same: test each candidate in the ORIGINAL equation, and write the reason for each rejection.

How the answer is expected to be written

A marker ticks steps. Here they are in order, with the concluding sentence expected word for word.

Finding an inverse function

When to use it: Any question that asks for f−1f^{-1}, its formula, its domain or one of its values

  1. 1 Prove that ff is one-to-one on its domain, or restrict the domain until it is, and say which piece you keep.
  2. 2 Find the range of ff: it will be the domain of f−1f^{-1}.
  3. 3 Write y=f(x)y = f(x), clear the denominators, collect every term containing xx on one side and factor xx out.
  4. 4 Divide, stating what must not be zero; if a ±\pm appears, choose the sign with the restricted domain.
  5. 5 Rename xx and yy, write the domain next to the formula, and check one value through ff and back.

Concluding sentence

“ff is one-to-one. From y(x+4)=1−3xy(x + 4) = 1 - 3x we get x(y+3)=1−4yx(y + 3) = 1 - 4y, so f−1(x)=1−4xx+3f^{-1}(x) = \frac{1 - 4x}{x + 3} for x≠−3x \ne -3. Check: f(0)=14f(0) = \frac{1}{4} and f−1(14)=0f^{-1}\left(\frac{1}{4}\right) = 0.”

The trap: Stopping at a line with xx on both sides, or writing the formula without its domain.

Marking: Typically 2 marks for one-to-one and the range, 4 for the algebra, 2 for the domain, 2 for the check.

Check before you hand in

Five minutes of checking recover more marks than one more problem started in a hurry.

The typical problem, taken apart

An inverse where the unknown is e to the x, twice

Let f(x)=3ex−1ex+2f(x) = \frac{3e^x - 1}{e^x + 2}. Show that ff is one-to-one, find its range, and find f−1f^{-1} with its domain.

Every step must be justified as on a MATH 203 final.

-5-4-3-2-112345-11234y = 3y = -1/2(0, 2/3)y = f(x)
The graph rises between the lines y=−12y = -\frac{1}{2} and y=3y = 3 without touching them; the algebra below proves what the picture suggests.

Step 1

Rewrite f(x)=3(ex+2)−7ex+2=3−7ex+2f(x) = \frac{3(e^x + 2) - 7}{e^x + 2} = 3 - \frac{7}{e^x + 2}. As xx increases, ex+2e^x + 2 increases, 7ex+2\frac{7}{e^x + 2} decreases, so ff increases: ff is one-to-one.

Why

Splitting off the constant puts the unknown in ONE place, which makes both the monotonicity and the range readable without any derivative.

Step 2

ex+2e^x + 2 takes every value in (2,∞)(2, \infty), so 7ex+2\frac{7}{e^x + 2} takes every value in (0,72)\left(0, \frac{7}{2}\right) and ff every value in (−12,3)\left(-\frac{1}{2}, 3\right).

Why

The range is announced before the formula: it will be the domain of f−1f^{-1}, and the next steps must find it again.

Step 3

y(ex+2)=3ex−1y(e^x + 2) = 3e^x - 1, so yex+2y=3ex−1ye^x + 2y = 3e^x - 1. Collect: yex−3ex=−1−2yye^x - 3e^x = -1 - 2y, and factor: ex(y−3)=−(1+2y)e^x(y - 3) = -(1 + 2y), so ex=1+2y3−ye^x = \frac{1 + 2y}{3 - y}.

Why

This is the gesture of the chapter: exe^x appears in two terms, and only collecting and factoring isolates it. Dividing needs y≠3y \ne 3, which the range guarantees.

Step 4

ex>0e^x > 0 requires 1+2y3−y>0\frac{1 + 2y}{3 - y} > 0, that is −12<y<3-\frac{1}{2} < y < 3. Then x=ln⁡1+2y3−yx = \ln\frac{1 + 2y}{3 - y}, so f−1(x)=ln⁡1+2x3−xf^{-1}(x) = \ln\frac{1 + 2x}{3 - x} on (−12,3)\left(-\frac{1}{2}, 3\right).

Why

The condition of ln⁡\ln gives back exactly the range of step 2: two independent roads to the same set is the best check of the domain.

Step 5

Check: f(0)=23f(0) = \frac{2}{3}, and f−1(23)=ln⁡7/37/3=ln⁡1=0f^{-1}\left(\frac{2}{3}\right) = \ln\frac{7/3}{7/3} = \ln 1 = 0.

Why

One value through ff and back takes twenty seconds and catches a swapped fraction or a sign lost while collecting.

The conclusion, written out

“ff is increasing, hence one-to-one, with range (−12,3)\left(-\frac{1}{2}, 3\right). Its inverse is f−1(x)=ln⁡1+2x3−xf^{-1}(x) = \ln\frac{1 + 2x}{3 - x} for −12<x<3-\frac{1}{2} < x < 3.”

The classic mistake on this problem: Taking ln⁡\ln of the numerator and the denominator separately, ln⁡(3ex−1)−ln⁡(ex+2)\ln(3e^x - 1) - \ln(e^x + 2), which leaves xx trapped in two logarithms; or giving the formula with no domain.

Learn by heart

  • • Only a one-to-one function has an inverse; domain of f−1f^{-1} = range of ff, and the graph reflects in y=xy = x.
  • • Unknown in two places: collect it on one side, factor it out, THEN divide or take ln⁡\ln.
  • • ln⁡(xy)=ln⁡x+ln⁡y\ln(xy) = \ln x + \ln y, ln⁡xr=rln⁡x\ln x^r = r\ln x: for POSITIVE x,yx, y only. No law for ln⁡(x+y)\ln(x + y) or (ln⁡x)2(\ln x)^2.
  • • log⁡ax=ln⁡xln⁡a\log_a x = \frac{\ln x}{\ln a}: the argument on top.
  • • ln⁡(ex)=x\ln(e^x) = x for every xx; eln⁡x=xe^{\ln x} = x for x>0x > 0 only.
  • • Equation: domain first, algebra, then test every candidate in the ORIGINAL equation; reject only for a reason.

Frequently asked questions

How do I find the inverse of a function in MATH 203?

Check that the function is one-to-one, restricting its domain if needed. Write y equals f of x, clear the denominators, move every term that contains x to one side, factor x out and divide. Rename x and y, and write the domain of the inverse, which is the range of the original function. Finish by checking one value through f and back.

Why do some solutions of a logarithmic equation have to be rejected?

Combining logarithms and clearing denominators enlarge the domain. The difference ln of x plus 4 minus ln of x minus 2 only exists when x is greater than 2, but the combined equation also accepts x equals minus 1. Reject a candidate only when it makes the argument of a logarithm zero or negative in the original equation; a negative candidate can be a perfectly valid solution.

How do I compute log base 3 of 20 on a calculator that only has ln and log?

Use the change of base formula: log base 3 of 20 equals ln 20 divided by ln 3, about 2.7268. The argument goes on top. Check it by bracketing: 20 is between 3 squared and 3 cubed, so the answer must be between 2 and 3. Getting 0.3667 means the fraction was typed upside down.

Is the square of ln x the same as 2 ln x?

No. Two ln x equals ln of x squared, because the power law moves an exponent that is inside the logarithm. The square of ln x squares the value of the logarithm, and it has no simpler form. At x equals e the first gives 1 and the second 2. In an equation, substitute t for ln x and solve the quadratic in t.

Practise it

Corrected exercises: Inverse functions and logarithms, MATH 203 at Concordia

A method is proved on a paper, not on a sheet. The set for the same chapter takes each of these traps into a problem, with the solution written out step by step.

  • 10 corrected exercises
  • 100 points
  • 150 minutes
Do the exercises
Previous sheet Functions, trigonometric and exponential functions Next sheet Rates of change and limit laws

© Ahmed Squalli Houssaini. Revision sheet published at www.letuteurscientifique.ca/en/fiches/math203-inverse-functions-logarithms. Free for personal and classroom use; republishing it elsewhere requires written permission (legal notice).

See also

Looking for a MATH 203 tutor in Montreal?

Get in touch for a first session. Logarithms and inverses come back in the derivative of ln x, in logarithmic differentiation and in every growth problem of the course, and the algebra has to be automatic by then.

Site by Studio Squalli