1.2 - Reversing a function

1.2 - Reversing a function

An inverse function recovers an input from its output, so the original function must be one–one. In this part of 1.2, you will rearrange rules to find inverses, use domain restrictions to choose the correct branch, and reflect graphs in the line y=xy=x.

Finding an inverse

A machine multiplies a number by 33, then subtracts 44. Its output is 1111. How do you recover the input? Add 44, then divide by 33: (11+4)/3=5(11+4)/3=5. Reversing the calculation reverses both the operations and their order.

Inverse function

For a one–one function ff, the inverse function f−1f^{-1} maps each output of ff back to its unique original input. The domain of f−1f^{-1} is the range of ff, and the range of f−1f^{-1} is the domain of ff.

The notation f−1(x)f^{-1}(x) means the inverse function's output, not the reciprocal 1/f(x)1/f(x).

If a function is many–one, an output does not identify a unique input. It must first be restricted to a one–one domain before an inverse function can be defined.

To find a rule, write y=f(x)y=f(x), solve for the original input xx in terms of the output yy, then rename the input of the inverse as xx. State the new domain using the original range.

Worked example 1: undoing a linear function

Question. Find the inverse of f(x)=3x−4f(x)=3x-4, x∈Rx\in\mathbb R.

Plan. Solve y=3x−4y=3x-4 for xx, then exchange input and output roles.

y=3x−4y+4=3xundo subtractionx=y+43undo multiplication\begin{aligned} y&=3x-4\\ y+4&=3x &&\text{undo subtraction}\\ x&=\frac{y+4}{3} &&\text{undo multiplication} \end{aligned}

The original linear function is one–one and has range R\mathbb R.

Answer. f−1(x)=(x+4)/3f^{-1}(x)=(x+4)/3, with domain and range R\mathbb R.

Check. f−1(f(x))=(3x−4+4)/3=xf^{-1}(f(x))=(3x-4+4)/3=x. In the other order, f(f−1(x))=3(x+4)/3−4=xf(f^{-1}(x))=3(x+4)/3-4=x.

Change only the original domain. The inverse's algebraic expression stays the same, but its legal inputs and possible outputs change.

Worked example 2: swapping domain and range

Question. Find the inverse of g(x)=3x−4g(x)=3x-4, 1<x≤51<x\le5, stating its domain and range.

Plan. Find the original range, then use the rearranged linear rule.

Since the rule increases with xx,

1<x≤5⟹−1<3x−4≤11.1<x\le5\quad\Longrightarrow\quad-1<3x-4\le11.

The same rearrangement as before gives the original input as (y+4)/3(y+4)/3.

Answer. g−1(x)=(x+4)/3g^{-1}(x)=(x+4)/3, with domain −1<x≤11-1<x\le11 and range 1<g−1(x)≤51<g^{-1}(x)\le5.

Check. g−1(11)=5g^{-1}(11)=5, so the included endpoint is preserved. Input −1-1 to the inverse would return 11, excluded from the original domain, so −1-1 is excluded from the inverse domain.

The two composition checks express the meaning of an inverse, but each has its own permitted inputs.

Inverse identities

For a one–one function ff,

f−1(f(x))=xfor x in the domain of f,f^{-1}(f(x))=x\quad\text{for }x\text{ in the domain of }f, f(f−1(x))=xfor x in the range of f.f(f^{-1}(x))=x\quad\text{for }x\text{ in the range of }f.

Not in the formula booklet: learn it.

A correct inverse rule cannot be used outside its domain just because its formula still produces a number.

Watch out

If f(x)=3x−4f(x)=3x-4, the reciprocal is 1/(3x−4)1/(3x-4), while the inverse is (x+4)/3(x+4)/3. Check an inverse by composition: it must undo the original function.

Now reverse a decreasing linear rule with a restricted domain.

Choosing the inverse branch

Rearranging a square introduces two algebraic candidates. The original domain decides which one can return an allowed original input. An inverse function must return one value, so leaving a ±\pm sign is not a finished answer.

Recall that u\sqrt{u} denotes the non-negative square root. From (x+1)2=y+2(x+1)^2=y+2, the two possible signs describe the two sides of the vertex. The following pair changes only the domain to show how that sign is selected.

Worked example 3: the right-hand branch

Question. Find the inverse of f(x)=(x+1)2−2f(x)=(x+1)^2-2, x≥−1x\ge-1, stating its domain.

Plan. Rearrange the square and use x+1≥0x+1\ge0 to choose its sign.

y+2=(x+1)2x+1=y+2original x+1≥0x=−1+y+2\begin{aligned} y+2&=(x+1)^2\\ x+1&=\sqrt{y+2} &&\text{original }x+1\ge0\\ x&=-1+\sqrt{y+2} \end{aligned}

The original minimum −2-2 is attained at x=−1x=-1, so its range is y≥−2y\ge-2.

Answer. f−1(x)=−1+x+2f^{-1}(x)=-1+\sqrt{x+2}, for x≥−2x\ge-2.

Check. f(1)=2f(1)=2 and f−1(2)=−1+2=1f^{-1}(2)=-1+2=1. The inverse always returns a number at least −1-1, as required.

On the other branch, the square root is still non-negative. It is the expression before the root that must take the negative sign.

Worked example 4: the left-hand branch

Question. Find the inverse of g(x)=(x+1)2−2g(x)=(x+1)^2-2, x≤−1x\le-1, stating its domain.

Plan. Use the original condition x+1≤0x+1\le0 after rearranging.

y+2=(x+1)2⟹x+1=−y+2⟹x=−1−y+2.y+2=(x+1)^2 \quad\Longrightarrow\quad x+1=-\sqrt{y+2} \quad\Longrightarrow\quad x=-1-\sqrt{y+2}.

The original range is again y≥−2y\ge-2, since the vertex is included.

Answer. g−1(x)=−1−x+2g^{-1}(x)=-1-\sqrt{x+2}, for x≥−2x\ge-2.

Check. g(−3)=2g(-3)=2, and g−1(2)=−1−2=−3g^{-1}(2)=-1-2=-3. Choosing the plus sign would return 11, outside the original domain.

An odd power behaves differently. The real cube root u3\sqrt[3]{u} is the unique real number whose cube is uu; for example, −83=−2\sqrt[3]{-8}=-2. Cubing is one–one on all real numbers, so undoing a cube does not introduce a ±\pm choice. Restrictions still determine the inverse domain.

Worked example 5: undoing a restricted cube

Question. Find the inverse of h(x)=(x+1)3−2h(x)=(x+1)^3-2, x<−1x<-1, stating its domain.

Plan. Take the real cube root, and find the range from the negative value of x+1x+1.

y+2=(x+1)3⟹x=y+23−1.y+2=(x+1)^3 \quad\Longrightarrow\quad x=\sqrt[3]{y+2}-1.

For x<−1x<-1, x+1<0x+1<0, so (x+1)3<0(x+1)^3<0 and y<−2y<-2. Every output below −2-2 occurs.

Answer. h−1(x)=x+23−1h^{-1}(x)=\sqrt[3]{x+2}-1, for x<−2x<-2.

Check. h(−3)=−10h(-3)=-10, and h−1(−10)=−83−1=−3h^{-1}(-10)=\sqrt[3]{-8}-1=-3. Input −2-2 would return the excluded endpoint −1-1.

For a reciprocal square, first isolate the square. A strict restriction on one side of a vertical asymptote selects the sign, while the reciprocal also excludes a limiting output.

Worked example 6: a reciprocal square

Question. Find the inverse of k(x)=1/(x+1)2−2k(x)=1/(x+1)^2-2, x<−1x<-1, stating its domain.

Plan. Isolate (x+1)2(x+1)^2, then choose the negative root because x+1<0x+1<0.

y+2=1(x+1)2(x+1)2=1y+2rearrange the reciprocalx+1=−1y+2choose the negative branchx=−1−1y+2\begin{aligned} y+2&=\frac{1}{(x+1)^2}\\ (x+1)^2&=\frac{1}{y+2} &&\text{rearrange the reciprocal}\\ x+1&=-\frac{1}{\sqrt{y+2}} &&\text{choose the negative branch}\\ x&=-1-\frac{1}{\sqrt{y+2}} \end{aligned}

The original range is y>−2y>-2: the reciprocal square is positive, can be arbitrarily small or large, and is never zero.

Answer. k−1(x)=−1−1/x+2k^{-1}(x)=-1-1/\sqrt{x+2}, for x>−2x>-2.

Check. k(−2)=−1k(-2)=-1, and k−1(−1)=−2k^{-1}(-1)=-2. More generally, if x>−2x>-2, the inverse returns a number below −1-1, and substituting it into kk gives x+2−2=xx+2-2=x.

The input restrictions and the algebraic sign must agree throughout the inverse's domain.

Watch out

The domain x<−1x<-1 in the reciprocal example describes original inputs. It selects the minus sign in the inverse rule. It is not the inverse domain: that comes from the original range y>−2y>-2.

Use the same reasoning with a shifted asymptote and a larger numerator.

Reflection in y equals x

If ff maps aa to bb, its inverse maps bb to aa. Therefore every point (a,b)(a,b) on the original graph becomes (b,a)(b,a) on the inverse graph. Swapping coordinates is a reflection in the line y=xy=x, drawn with equal scales on the axes.

Graph of an inverse

For a one–one function, reflect its graph in the mirror line y=xy=x to obtain the graph of its inverse. Included and excluded endpoints retain their inclusion status as their coordinates swap.

An xx-intercept (a,0)(a,0) becomes the inverse's yy-intercept (0,a)(0,a). A yy-intercept (0,b)(0,b) becomes the inverse's xx-intercept (b,0)(b,0). Only intercepts lying on the permitted original graph are available.

Worked example 7: reflecting a straight line

Question. Describe a sketch of f(x)=2x+1f(x)=2x+1, x∈Rx\in\mathbb R, and its inverse, identifying both graphs' intercepts and their mirror line.

Plan. Find the original intercepts, swap their coordinates, and use the inverse rule to check them.

The original graph passes through (−1/2,0)(-1/2,0) and (0,1)(0,1). Reflection gives (0,−1/2)(0,-1/2) and (1,0)(1,0).

Rearranging y=2x+1y=2x+1 gives f−1(x)=(x−1)/2f^{-1}(x)=(x-1)/2.

Answer. Draw the original line with gradient 22, its inverse with gradient 1/21/2, and the mirror line y=xy=x. Original intercepts are (−1/2,0),(0,1)(-1/2,0),(0,1); inverse intercepts are (1,0),(0,−1/2)(1,0),(0,-1/2).

Check. The inverse rule gives f−1(0)=−1/2f^{-1}(0)=-1/2 and f−1(1)=0f^{-1}(1)=0, confirming the reflected intercepts.

The mirror line belongs in an inverse sketch even when the two function graphs are easy to recognise.

[DIAGRAM: asset_name: Swapping coordinates; asset_slug: inverse-line-reflection; caption: Reflecting in y equals x swaps each point's input and output; description: Equal-scale coordinate axes show y equals 2x plus 1 in blue, y equals x minus 1 divided by 2 in green, and dashed mirror line y equals x. The point 0 comma 1 reflects to 1 comma 0 and the point 1 comma 3 reflects to 3 comma 1. Corresponding points are joined by faint perpendicular segments.]
Reflecting in y equals x swaps each point's input and output

An asymptote describes a curve's limiting behaviour. At a vertical asymptote, the output grows in magnitude without bound as the input approaches a fixed value. A horizontal asymptote gives a height approached as the input increases or decreases without bound. Reflection swaps horizontal and vertical asymptotes: x=ax=a becomes y=ay=a, and y=by=b becomes x=bx=b.

Replace the straight line by a reciprocal-square curve. The coordinate swap is unchanged, but now it also explains the positions of the asymptotes.

Worked example 8: reflecting asymptotes

Question. For k(x)=1/(x+1)2−2k(x)=1/(x+1)^2-2, x<−1x<-1, describe the graph of its inverse, including its domain, range, asymptotes and the reflection of (−2,−1)(-2,-1).

Plan. Swap the original domain and range, then reflect each feature in y=xy=x.

The original curve has vertical asymptote x=−1x=-1, horizontal asymptote y=−2y=-2, domain x<−1x<-1 and range y>−2y>-2.

Answer. The inverse has domain x>−2x>-2, range y<−1y<-1, vertical asymptote x=−2x=-2 and horizontal asymptote y=−1y=-1. The point (−2,−1)(-2,-1) becomes (−1,−2)(-1,-2). Draw the reflected branch and the mirror line y=xy=x.

Check. Its rule k−1(x)=−1−1/x+2k^{-1}(x)=-1-1/\sqrt{x+2} gives k−1(−1)=−2k^{-1}(-1)=-2. As xx approaches −2-2 from above, the denominator approaches zero and the output decreases without bound; as xx increases, the output approaches −1-1 from below.

The reflected graph lies to the right of its vertical asymptote and below its horizontal one, exactly as its domain and range require.

[DIAGRAM: asset_name: Asymptotes exchange direction; asset_slug: inverse-asymptote-reflection; caption: A vertical asymptote becomes horizontal when a function is reversed; description: Equal-scale axes show the branch y equals one divided by the square of x plus 1 minus 2 for x less than minus 1, and its reflected inverse y equals minus 1 minus one divided by the square root of x plus 2 for x greater than minus 2. Original asymptotes x equals minus 1 and y equals minus 2 are blue dashed lines; inverse asymptotes x equals minus 2 and y equals minus 1 are green dashed lines. The mirror line y equals x and the reflected pair minus 2 comma minus 1 and minus 1 comma minus 2 are labelled.]
A vertical asymptote becomes horizontal when a function is reversed

Return to the machine in the opening: its input 55 and output 1111 exchange places. The algebraic reversal and the graph's reflection express the same idea.

Exam tip

An inverse sketch should show the correct branch, its key features and the mirror line y=xy=x. Reflect the stated function's domain as well as its visible shape.

Reflect a set of features without needing a formula.

Quick Recap

  • An inverse exists for a one–one function, and exchanges its domain and range.
  • Rearrange to recover the original input, then rename variables and state the inverse domain.
  • A square-root sign follows the original branch; a real cube root has no ±\pm choice.
  • Reflect points, endpoints and asymptotes in the line y=xy=x.

The opening machine is reversed by adding 44, then dividing by 33: 1111 returns to 55.