본문으로 바로 이동 / Skip to content

PROOF 101

Prove.

The discriminant determines the number of roots

On the real numbers 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

A root of ax² + bx + c = 0 (a ≠ 0) is a number x for which the equation holds upon substitution.

Proof.

(1)a ≠ 0 is the condition for this expression to be quadratic, and is assumed in every block below.convention
(2)The number of roots means the number of all such x.convention

∎

Definition D2

D denotes b² − 4ac.

Proof.

This value is called the discriminant.convention

∎

Lemma L1

When a ≠ 0, for each number u there is exactly one x with 2ax + b = u.

Proof.

(1)Since 2 ≠ 0 and a ≠ 0, we have 2a ≠ 0.A17, A16
(2)Set x₀ = (u + (−b))(2a)⁻¹.(1), A14
(3)2ax₀ = (2a(2a)⁻¹)(u + (−b)) = u + (−b)(2), A11, A12, A14, A13
(4)2ax₀ + b = (u + (−b)) + b = u(3), A7, A8, A10, A9
(5)Suppose 2ax + b = u and 2ay + b = u.assumption
(6)Adding −b to both sides gives 2ax = 2ay.(5), E, A7, A10, A9
(7)Multiplying both sides by (2a)⁻¹ gives x = y.(6), (1), E, A11, A12, A14, A13
(8)Such an x exists and is unique.(4), (7)

∎

Theorem T1

If D > 0, there are exactly two roots.

Proof.

(1)Suppose D > 0.assumption
(2)There exists a unique s > 0 with s² = D.(1), A2
(3)ax² + bx + c = 0 ⟺ (2ax + b)² = D = s²A1, D2, (2)
(4)(2ax + b)² = s² ⟺ 2ax + b = s or 2ax + b = −sA3
(5)Since s > 0, we have s ≠ 0.(2), A5
(6)Since 2 ≠ 0, we have 2s ≠ 0.(5), A17, A16
(7)If s = −s, then s + s = s + (−s) = 0.A10
(8)s + s = s(1 + 1) = 2sA15, A13, A12
(9)Hence s = −s contradicts (6), and s ≠ −s.(6), (7), (8)
(10)There is exactly one x with 2ax + b = s and exactly one x with 2ax + b = −s.L1
(11)If the two coincided, then s = −s, contradicting (9).(9), (10)
(12)There are exactly two roots.(3), (4), (10), (11), D1

∎

Theorem T2

If D = 0, there is exactly one root.

Proof.

(1)Suppose D = 0.assumption
(2)ax² + bx + c = 0 ⟺ (2ax + b)² = 0A1, D2, (1)
(3)If u ≠ 0, then u² ≠ 0.A16
(4)Hence if (2ax + b)² = 0, then 2ax + b = 0.(3)
(5)Conversely, if 2ax + b = 0, then (2ax + b)² = 0 × 0 = 0.A18
(6)There is exactly one x with 2ax + b = 0.L1
(7)There is exactly one root.(2), (4), (5), (6), D1

∎

Theorem T3

If D < 0, no real root exists.

Proof.

(1)Suppose D < 0.assumption
(2)If a real root x existed, then (2ax + b)² = D.A1, D1, D2
(3)0 ≤ (2ax + b)²A4
(4)Hence 0 ≤ D.(2), (3), E
(5)Trichotomy does not allow D < 0 and 0 ≤ D together.(1), (4), A5
(6)No such x exists.(5)

∎

Corollary C

Even when D < 0, on the stage extended by i there are exactly two roots.

Proof.

(1)Suppose D < 0.assumption
(2)A1, A3, L1 use only the field axioms, so they hold over the complex numbers as well.A1, A3, L1
(3)Adding −D to both sides gives 0 < −D.(1), A6, A10, A8, A9
(4)There exists s > 0 with s² = −D.(3), A2
(5)There is i ≠ 0 with i² = −1.A19
(6)(is)² = i²s² = (−1)s²(5), A11, A12, E
(7)(−1)s² = (−1)(−D) = 1 × D = D(4), A20, A12, A13, E
(8)(2ax + b)² = D = (is)², so 2ax + b = ±is.(2), (6), (7), A3
(9)Since s ≠ 0, i ≠ 0, 2 ≠ 0, we have 2is ≠ 0.(4), (5), A5, A17, A16
(10)If is = −is, then is + is = 2(is) = 0, contradicting (9).(9), A10, A15, A13, A12
(11)There is exactly one x with 2ax + b = is and exactly one x with 2ax + b = −is, and they differ.(2), (10), L1
(12)There are exactly two roots.(8), (11), D1

∎

Remark

Can A4 be removed from this proof?

Look at the three cases of the discriminant separately. They do not rest on the same axiom.