Mathematical-Reasoning ยท Unit 22 ยท Video 4 ยท Interactive Practice

Domain and Codomain Decide: x2+1x^2 + 1 and eโˆ’xe^{-x}

IKey Formulas

StatementNameWhat it depends on
For all s,tโˆˆAs, t \in A: f(s)=f(t)โ‡’s=tf(s) = f(t) \Rightarrow s = tf:Aโ†’Bf : A \to B is an injectionThe domain AA only; changing BB cannot make two inputs share or stop sharing an output
For all yโˆˆBy \in B there is xโˆˆAx \in A with f(x)=yf(x) = yf:Aโ†’Bf : A \to B is a surjectionThe codomain BB, and the domain AA through the range
x2+1x^2 + 1: neither on Rโ†’R\mathbb{R} \to \mathbb{R}; onto only on Rโ†’T\mathbb{R} \to T; one-to-one only on Zโˆ—โ†’N\mathbb{Z}^* \to \mathbb{N}One formula, three verdictsT={yโˆˆRโˆฃyโ‰ฅ1}T = \{y \in \mathbb{R} \mid y \ge 1\}, Zโˆ—={xโˆˆZโˆฃxโ‰ฅ0}\mathbb{Z}^* = \{x \in \mathbb{Z} \mid x \ge 0\}
g(x)=eโˆ’xg(x) = e^{-x}, g:Rโ†’R+g : \mathbb{R} \to \mathbb{R}^+; for b>0b > 0, g(โˆ’lnโกb)=bg(-\ln b) = bA bijectionb>0b > 0 makes lnโกb\ln b defined; for f:Rโ†’Rf : \mathbb{R} \to \mathbb{R} with the same rule, โˆ’1-1 is never an output

Key Insight: A function is its rule together with its domain and its codomain. Change either set and you have a different function, whose verdicts must be checked again.

IIVisualization 1 โ€” One Formula, Twelve Choices of Sets

Ten of the twelve choices of domain AA and codomain BB turn x2+1x^2 + 1 into a different function with its own verdicts; with domain R\mathbb{R} or [0,โˆž)[0, \infty) and codomain N\mathbb{N}, the rule gives outputs outside N\mathbb{N}, so it defines no function at all.

IIIVisualization 2 โ€” Solving eโˆ’a=be^{-a} = b Backward

The input a=โˆ’lnโกba = -\ln b exists exactly when b>0b > 0, so the codomain decides whether eโˆ’xe^{-x} is onto.

๐Ÿ’ก The horizontal line at any height meets the curve at most once, which suggests one-to-one; the proof takes lnโก\ln of both sides of eโˆ’a=eโˆ’be^{-a} = e^{-b} and is the same for ff and gg, because they share the domain R\mathbb{R}.

IVQuiz Questions

Problem 1 ยท Absolute Value on โ„

Given: h:Rโ†’Rh : \mathbb{R} \to \mathbb{R}, h(x)=โˆฃxโˆ’2โˆฃh(x) = |x - 2|. Which statement is true?

โœ… Correct! h(0)=h(4)=2h(0) = h(4) = 2 breaks one-to-one, and โˆ’1-1 is in the codomain but โˆฃxโˆ’2โˆฃโ‰ฅ0|x - 2| \ge 0 never equals it.
โŒ Not quite. Two inputs the same distance from 22 share an output: h(0)=โˆฃโˆ’2โˆฃ=2h(0) = |-2| = 2 and h(4)=โˆฃ2โˆฃ=2h(4) = |2| = 2, with 0โ‰ 40 \ne 4.
โŒ Not quite. The codomain is all of R\mathbb{R}, and an absolute value is never negative, so โˆ’1-1 is never an output.
โŒ Not quite. A bijection needs both properties, and hh has neither: h(0)=h(4)h(0) = h(4), and h(x)=โˆ’1h(x) = -1 has no solution.
Show solution

Not an injection. One pair of inputs is enough: h(0)=โˆฃ0โˆ’2โˆฃ=2h(0) = |0 - 2| = 2 and h(4)=โˆฃ4โˆ’2โˆฃ=2h(4) = |4 - 2| = 2, but 0โ‰ 40 \ne 4.

Not a surjection. For every real xx, โˆฃxโˆ’2โˆฃโ‰ฅ0|x - 2| \ge 0. So โˆ’1โˆˆR-1 \in \mathbb{R}, the codomain, is not h(x)h(x) for any xx, and โˆ’1-1 is not in the range.

So hh is neither. (With codomain [0,โˆž)[0, \infty) instead, hh would be onto, since h(2+y)=yh(2 + y) = y for yโ‰ฅ0y \ge 0, but still not one-to-one.)

Problem 2 ยท Restricting the Domain

Given: k:[3,โˆž)โ†’Rk : [3, \infty) \to \mathbb{R}, k(x)=x2โˆ’6xk(x) = x^2 - 6x. Which statement is true?

โœ… Correct! On [3,โˆž)[3, \infty), k(s)=k(t)k(s) = k(t) forces s=ts = t; but k(x)=(xโˆ’3)2โˆ’9โ‰ฅโˆ’9k(x) = (x - 3)^2 - 9 \ge -9, so โˆ’10-10 is never an output.
โŒ Not quite. Restricting the domain fixed one-to-one, but the codomain is still R\mathbb{R}, and k(x)=(xโˆ’3)2โˆ’9โ‰ฅโˆ’9k(x) = (x - 3)^2 - 9 \ge -9 misses โˆ’10-10.
โŒ Not quite. On all of R\mathbb{R}, k(2)=k(4)=โˆ’8k(2) = k(4) = -8, but 22 is not in the domain [3,โˆž)[3, \infty). And the range is only [โˆ’9,โˆž)[-9, \infty).
โŒ Not quite. The pair 2,42, 4 that shares an output on R\mathbb{R} is gone: for s,tโ‰ฅ3s, t \ge 3 the inputs never share an output. Only onto fails.
Show solution

Injection. Let s,tโˆˆ[3,โˆž)s, t \in [3, \infty) with k(s)=k(t)k(s) = k(t). Then s2โˆ’6s=t2โˆ’6ts^2 - 6s = t^2 - 6t, so

s2โˆ’t2โˆ’6(sโˆ’t)=(sโˆ’t)(s+tโˆ’6)=0.s^2 - t^2 - 6(s - t) = (s - t)(s + t - 6) = 0.

Since sโ‰ฅ3s \ge 3 and tโ‰ฅ3t \ge 3, s+tโˆ’6โ‰ฅ0s + t - 6 \ge 0, with equality only when s=t=3s = t = 3. Either way s=ts = t: if s+tโˆ’6โ‰ 0s + t - 6 \ne 0 then sโˆ’t=0s - t = 0, and if s+tโˆ’6=0s + t - 6 = 0 then s=t=3s = t = 3.

Not a surjection. Completing the square, k(x)=(xโˆ’3)2โˆ’9โ‰ฅโˆ’9k(x) = (x - 3)^2 - 9 \ge -9 for every xx. So โˆ’10โˆˆR-10 \in \mathbb{R} is not in the range.

The domain decided the injection verdict; the codomain R\mathbb{R} still contains numbers below โˆ’9-9 that are never outputs.

Problem 3 ยท Choose the Codomain

Given: the rule k(x)=x2โˆ’6xk(x) = x^2 - 6x on the domain [3,โˆž)[3, \infty) from Problem 2. For which codomain CC is k:[3,โˆž)โ†’Ck : [3, \infty) \to C a bijection, and which input does an arbitrary yโˆˆCy \in C come from?

Which codomain?

Which input x gives k(x) = y?

โœ… Correct! With C=[โˆ’9,โˆž)C = [-9, \infty), every yโˆˆCy \in C has y+9โ‰ฅ0y + 9 \ge 0, so x=3+y+9x = 3 + \sqrt{y + 9} is a real number at least 33, and k(x)=yk(x) = y.
โŒ Not quite. k(3)=9โˆ’18=โˆ’9โˆ‰[0,โˆž)k(3) = 9 - 18 = -9 \notin [0, \infty), so the rule does not even define a function from [3,โˆž)[3, \infty) to [0,โˆž)[0, \infty).
โŒ Not quite. k(3)=โˆ’9k(3) = -9 must lie in the codomain, and the open interval (โˆ’9,โˆž)(-9, \infty) leaves it out, so this is not a function.
โŒ Not quite. That is Problem 2: with codomain R\mathbb{R}, the value โˆ’10-10 is never reached.
โŒ Not quite. This does solve (xโˆ’3)2=y+9(x - 3)^2 = y + 9, but it is at most 33, so for y>โˆ’9y > -9 it is not in the domain [3,โˆž)[3, \infty).
โŒ Not quite. Check forward with y=7y = 7: x=16โˆ’3=1x = \sqrt{16} - 3 = 1, which is not in the domain, and k(1)=โˆ’5โ‰ 7k(1) = -5 \ne 7.
โŒ Not quite. Completing the square gives (xโˆ’3)2=y+9(x - 3)^2 = y + 9, not yโˆ’9y - 9; also yโˆ’9\sqrt{y - 9} is undefined for y<9y < 9.
Show solution

The codomain. k(x)=(xโˆ’3)2โˆ’9k(x) = (x - 3)^2 - 9. For xโ‰ฅ3x \ge 3, (xโˆ’3)2(x - 3)^2 takes every value in [0,โˆž)[0, \infty), so the range is [โˆ’9,โˆž)[-9, \infty). A bijection needs codomain = range, and every output (including k(3)=โˆ’9k(3) = -9) must lie in the codomain. So C=[โˆ’9,โˆž)C = [-9, \infty).

Work backward. Let yโˆˆCy \in C. From (xโˆ’3)2โˆ’9=y(x - 3)^2 - 9 = y we get (xโˆ’3)2=y+9(x - 3)^2 = y + 9, and y+9โ‰ฅ0y + 9 \ge 0, so xโˆ’3=ยฑy+9x - 3 = \pm\sqrt{y + 9}. The domain needs xโ‰ฅ3x \ge 3, so take x=3+y+9x = 3 + \sqrt{y + 9}.

Verify forward.

k(3+y+9)=(y+9)2โˆ’9=y+9โˆ’9=y.k\big(3 + \sqrt{y + 9}\big) = \big(\sqrt{y + 9}\big)^2 - 9 = y + 9 - 9 = y.

Together with the injection from Problem 2 (which depends only on the domain), k:[3,โˆž)โ†’[โˆ’9,โˆž)k : [3, \infty) \to [-9, \infty) is a bijection.

Problem 4 ยท Reciprocals on Two Domains

Given: p:(0,โˆž)โ†’(0,โˆž)p : (0, \infty) \to (0, \infty) and q:Rโˆ’{0}โ†’Rq : \mathbb{R} - \{0\} \to \mathbb{R}, both defined by the rule 1x\dfrac{1}{x}. Which statement is true?

โœ… Correct! Both are one-to-one; pp reaches every y>0y > 0 via x=1yx = \tfrac{1}{y}, while qq misses 00.
โŒ Not quite. The codomain of qq contains 00, and 1x=0\tfrac{1}{x} = 0 has no solution, so qq is not onto.
โŒ Not quite. Unlike x2x^2, the reciprocal never merges a negative input with a positive one: 1s=1t\tfrac{1}{s} = \tfrac{1}{t} gives s=ts = t on all of Rโˆ’{0}\mathbb{R} - \{0\}.
โŒ Not quite. pp is onto: for yโˆˆ(0,โˆž)y \in (0, \infty), the input x=1yx = \tfrac{1}{y} is in (0,โˆž)(0, \infty) and p(1y)=yp(\tfrac{1}{y}) = y.
Show solution

Both are injections. If 1s=1t\tfrac{1}{s} = \tfrac{1}{t} with s,tโ‰ 0s, t \ne 0, multiply both sides by stst to get t=st = s. This works on either domain.

pp is a surjection. Let yโˆˆ(0,โˆž)y \in (0, \infty). Work backward: 1x=y\tfrac{1}{x} = y gives x=1yx = \tfrac{1}{y}, which is defined and positive because y>0y > 0. Verify: p(1y)=11/y=yp\big(\tfrac{1}{y}\big) = \dfrac{1}{1/y} = y. So pp is a bijection.

qq is not a surjection. 0โˆˆR0 \in \mathbb{R}, but 1x=0\tfrac{1}{x} = 0 would need 1=0โ‹…x=01 = 0 \cdot x = 0. So 00 is in the codomain but not in the range.

Every other real yโ‰ 0y \ne 0 is reached by x=1yx = \tfrac{1}{y}, so changing the codomain of qq to Rโˆ’{0}\mathbb{R} - \{0\} would make it a bijection too.

Solved: 0 / 4