ProofComplete

Applications of Derivatives - Key Proof

We present proofs of Fermat's Theorem and the Second Derivative Test, two fundamental results that connect local extrema to derivatives.

ProofFermat's Theorem

Theorem: If ff has a local maximum or minimum at cc and f(c)f'(c) exists, then f(c)=0f'(c) = 0.

Proof: Suppose ff has a local maximum at cc (the minimum case is similar). Then there exists δ>0\delta > 0 such that f(c)f(x)f(c) \geq f(x) for all x(cδ,c+δ)x \in (c - \delta, c + \delta).

Consider the difference quotient from the right: f(c+h)f(c)hfor 0<h<δ\frac{f(c+h) - f(c)}{h} \quad \text{for } 0 < h < \delta

Since f(c)f(c+h)f(c) \geq f(c+h), we have f(c+h)f(c)0f(c+h) - f(c) \leq 0. With h>0h > 0, this gives: f(c+h)f(c)h0\frac{f(c+h) - f(c)}{h} \leq 0

Taking the limit as h0+h \to 0^+: f+(c)=limh0+f(c+h)f(c)h0f'_+(c) = \lim_{h \to 0^+} \frac{f(c+h) - f(c)}{h} \leq 0

Similarly, for the left-hand limit with δ<h<0-\delta < h < 0: f(c+h)f(c)h0(negative numerator, negative denominator)\frac{f(c+h) - f(c)}{h} \geq 0 \quad \text{(negative numerator, negative denominator)}

Taking the limit as h0h \to 0^-: f(c)=limh0f(c+h)f(c)h0f'_-(c) = \lim_{h \to 0^-} \frac{f(c+h) - f(c)}{h} \geq 0

Since f(c)f'(c) exists, we have f+(c)=f(c)=f(c)f'_+(c) = f'_-(c) = f'(c). Therefore: f(c)0andf(c)0f'(c) \leq 0 \quad \text{and} \quad f'(c) \geq 0

This forces f(c)=0f'(c) = 0. \square

Remark

The proof uses a subtle but important technique: the one-sided limits must both exist and equal f(c)f'(c), and they are constrained by opposite inequalities. The only value satisfying both constraints is zero.

ProofSecond Derivative Test

Theorem: Suppose ff'' is continuous near cc and f(c)=0f'(c) = 0.

  • If f(c)>0f''(c) > 0, then ff has a local minimum at cc
  • If f(c)<0f''(c) < 0, then ff has a local maximum at cc

Proof: We prove the minimum case; the maximum case is analogous.

Assume f(c)=0f'(c) = 0 and f(c)>0f''(c) > 0. Since ff'' is continuous and f(c)>0f''(c) > 0, there exists δ>0\delta > 0 such that f(x)>0f''(x) > 0 for all x(cδ,c+δ)x \in (c - \delta, c + \delta).

Because f(x)>0f''(x) > 0 on this interval, ff' is strictly increasing on (cδ,c+δ)(c - \delta, c + \delta).

For x(cδ,c)x \in (c - \delta, c): Since ff' is increasing and f(c)=0f'(c) = 0, we have f(x)<f(c)=0f'(x) < f'(c) = 0. Thus ff is decreasing on (cδ,c)(c - \delta, c).

For x(c,c+δ)x \in (c, c + \delta): Since ff' is increasing and f(c)=0f'(c) = 0, we have f(x)>f(c)=0f'(x) > f'(c) = 0. Thus ff is increasing on (c,c+δ)(c, c + \delta).

Therefore, ff decreases to the left of cc and increases to the right of cc, which means ff has a local minimum at cc. \square

Remark

The second derivative test is actually a consequence of the first derivative test combined with the monotonicity theorem. The sign of ff'' determines whether ff' is increasing or decreasing, which in turn determines the behavior of ff near the critical point.

ExampleWhen the Second Derivative Test Fails

Consider f(x)=x4f(x) = x^4 at x=0x = 0.

We have f(0)=0f'(0) = 0 and f(0)=0f''(0) = 0, so the second derivative test is inconclusive.

However, f(x)=x4>0f(x) = x^4 > 0 for all x0x \neq 0, so ff clearly has a local (and absolute) minimum at x=0x = 0.

This shows that f(c)=0f''(c) = 0 doesn't rule out an extremum—it just means we need to use other methods (like the first derivative test or higher derivatives) to determine the nature of the critical point.

These proofs illuminate why the derivative tests work, grounding computational techniques in rigorous mathematical reasoning.