본문으로 바로 이동 / Skip to content

PROOF 45

Prove.

The intermediate value theorem

On the completeness of the reals ℝ.

A way to read this proof

First separate what is assumed from what is to be proved. At each step, ask which definition or earlier result permits the move. If it is unclear, return to that premise rather than skipping to the conclusion.

Definition D1

c is the supremum of S if c is an upper bound of S and no number smaller than c is an upper bound of S.

Proof.

(1)An upper bound is a number not smaller than any element of S. The supremum is the smallest upper bound.convention

∎

Definition D2

f is continuous at p if for every ε > 0 there is δ > 0 such that x ∈ [a,b] and |x − p| < δ imply |f(x) − f(p)| < ε.

Proof.

(1)δ may depend on p and ε. This definition is a property at the single point p.convention

∎

Lemma L1

Let S = {x ∈ [a,b] : f(x) < k}. Then c = sup S exists and a ≤ c ≤ b.

Proof.

(1)f is continuous on [a,b] and f(a) < k < f(b).assumption
(2)Since f(a) < k, we have a ∈ S, so S ≠ ∅.(1), convention
(3)Since S ⊆ [a,b], b is an upper bound of S.convention, D1
(4)Therefore c = sup S exists.(2), (3), A7
(5)Since a ∈ S and c is an upper bound, a ≤ c.(2), (4), D1
(6)Since b is an upper bound and c is the least upper bound, c ≤ b.(3), (4), D1

∎

Lemma L2

f(c) < k does not hold.

Proof.

(1)Suppose f(c) < k.assumption
(2)Since f(b) > k, c ≠ b, and since c ≤ b, c < b.(1), L1, A3
(3)Set ε = k − f(c). Then ε > 0.(1), A1, A2
(4)Since c ∈ [a,b], there is δ > 0 such that |x − c| < δ implies |f(x) − f(c)| < ε.L1, D2, assumption
(5)b − c > 0.(2), A1, A2
(6)Choose h with 0 < h < δ, h < b − c.(4), (5), A6
(7)Set x₁ = c + h. Then c < x₁ < b and x₁ ∈ [a,b].(6), L1, A1, A2, A4
(8)x₁ − c = h and −δ < h < δ, so |x₁ − c| < δ.(6), (7), A5
(9)|f(x₁) − f(c)| < ε, so f(x₁) < f(c) + ε = k.(4), (8), A5, (3), A1
(10)Therefore x₁ ∈ S.(7), (9), convention
(11)Since x₁ > c, c is not an upper bound of S. This contradicts D1.(7), (10), D1, L1
(12)f(c) < k does not hold.(11)

∎

Lemma L3

f(c) > k does not hold.

Proof.

(1)Suppose f(c) > k.assumption
(2)Since f(a) < k, c ≠ a, and since a ≤ c, a < c.(1), L1, A3
(3)Set ε = f(c) − k. Then ε > 0.(1), A1, A2
(4)Since c ∈ [a,b], there is δ > 0 such that |x − c| < δ implies |f(x) − f(c)| < ε.L1, D2, assumption
(5)c − a > 0.(2), A1, A2
(6)Choose h with 0 < h < δ, h < c − a, and set c′ = c − h. Then a < c′ < c.(4), (5), A6, A1, A2
(7)Suppose x ∈ S and x > c′.assumption
(8)Since c is an upper bound, x ≤ c, and since x > c − h, we have −δ < x − c ≤ 0 < δ.(6), (7), L1, D1, A1, A2
(9)Since |x − c| < δ, we have |f(x) − f(c)| < ε.(8), A5, (4)
(10)f(x) > f(c) − ε = k.(9), A5, (3), A1
(11)Since x ∈ S, f(x) < k. This contradicts (10).(7), convention, (10), A3
(12)Therefore every element of S is at most c′. Then c′ is an upper bound smaller than c, contradicting D1.(11), (6), D1, L1
(13)f(c) > k does not hold.(12)

∎

Theorem T

If f is continuous on [a,b] and f(a) < k < f(b), then there exists c between a and b with f(c) = k.

Proof.

(1)c = sup S exists and a ≤ c ≤ b.L1
(2)Neither f(c) < k nor f(c) > k holds.L2, L3
(3)Therefore f(c) = k.(2), A3
(4)Since f(a) < k = f(c), c ≠ a, and since f(b) > k = f(c), c ≠ b.assumption, (3)
(5)Since a ≤ c ≤ b, we have a < c < b.(1), (4), A3

∎

Remark

Can A7 be removed from this proof?

A7 is used exactly once in this proof, at line (4) of L1. Check how much survives its removal.