The Limit Laws


LEMMA


a lemma is an auxiliary theorem, a result that justifies its existence only by virtue of its prominent role in the proof of another theorem.

(1) If

xx0<ϵ2andyy0<ϵ2|x-x_0| < \frac{\epsilon}{2} \quad and \quad |y-y_0| < \frac{\epsilon}{2}

then

(x+y)(x0+y0)<ϵ|(x+y) - (x_0+y_0)| < \epsilon

Proof:

(x+y)(x0+y0)=(xx0)+(yy0)xx0+yy0<ϵ2+ϵ2=ϵ\begin{aligned} \\ |(x+y) - (x_0+y_0)| = |(x-x_0)+ (y-y_0)| \\ \leq |x-x_0| + |y-y_0| \\ < \frac{\epsilon}{2} + \frac{\epsilon}{2} \\ = \epsilon \\ \end{aligned}

(2) If

xx0<min(1,ϵ2(y0+1))andyy0<ϵ2(x0+1)|x-x_0| < min(1, \frac{\epsilon}{2(|y_0| + 1)}) \quad and \quad |y-y_0| < \frac{\epsilon}{2(|x_0| + 1)}

then

xyx0y0<ϵ|xy-x_0y_0| < \epsilon

Proof:

Since xx0<1|x-x_0| < 1 we have

xx0xx0<1|x| - |x_0| \leq |x-x_0| < 1

so that

x<1+x0|x| < 1 + |x_0|

Thus

xyx0y0=x(yy0)+y0(xx0)xyy0+y0xx0<(1+x0)ϵ2(x0+1)+y0ϵ2(y0+1)<ϵ2+ϵ2<ϵ\begin{aligned} |xy-x_0y_0| = |x(y-y_0) + y_0(x-x_0)| \\ \leq |x| \cdot |y-y_0| + |y_0| \cdot |x-x_0| \\ < (1 + |x_0|) \cdot \frac{\epsilon}{2(|x_0| + 1)} + |y_0| \cdot \frac{\epsilon}{2(|y_0| + 1)} \\ < \frac{\epsilon}{2} + \frac{\epsilon}{2} \\ < \epsilon \\ \end{aligned}

(3) If y00y_0 \neq 0 and

yy0<min(y02,ϵy022)|y-y_0| < min(\frac{|y_0|}{2}, \frac{\epsilon |y_0|^2}{2})

then y0y \neq 0 and

1y1y0<ϵ|\frac{1}{y} - \frac{1}{y_0}| < \epsilon

Proof:

We have

y0yyy0<y02|y_0| - |y| \leq |y - y_0| < \frac{|y_0|}{2}

so y>y0/2|y| > |y_0|/2. In particular, y0y \neq 0, and

1y<2y0\frac{1}{|y|} < \frac{2}{|y_0|}

Thus

1y1y0=y0yyy0<2y01y0ϵy022<ϵ|\frac{1}{y} - \frac{1}{y_0}| = \frac{|y_0 - y|}{|y| \cdot |y_0|} < \frac{2}{|y_0} \cdot \frac{1}{|y_0|} \cdot \frac{\epsilon |y_0|^2}{2} < \epsilon

Theorem


If

limxaf(x)=landlimxag(x)=m\lim_{x \to a} f(x) = l \quad and \quad \lim_{x \to a} g(x) = m

then

(1)limxa(f+g)(x)=l+m(2)limxa(fg)(x)=lm(3)limxa(1g)(x)=1mifm0\begin{aligned} & (1) \lim_{x \to a} (f+g)(x) = l + m \\ & (2) \lim_{x \to a} (f \cdot g)(x) = l \cdot m \\ & (3) \lim_{x \to a} (\frac{1}{g})(x) = \frac{1}{m} \quad if \quad m \neq 0 \\ \end{aligned}

Proof


The hypothesis means that for every ϵ>0\epsilon > 0 there are δ1,δ2>0\delta_1, \delta_2 > 0 such that, for all xx

if0<xa<δ1,thenf(x)l<ϵ,if \quad 0 < |x-a| < \delta_1, \quad then \quad |f(x) - l| < \epsilon, andif0<xa<δ2,theng(x)m<ϵ,and \quad if \quad 0 < |x-a| < \delta_2, \quad then \quad |g(x) - m| < \epsilon,

This means (since, after all, ϵ/2\epsilon/2 is also a positive number) that there are δ1,δ2>0\delta_1, \delta_2 > 0 such that, for all xx

if0<xa<δ1,thenf(x)l<ϵ2,if \quad 0 < |x-a| < \delta_1, \quad then \quad |f(x) - l| < \frac{\epsilon}{2}, andif0<xa<δ2,theng(x)m<ϵ2,and \quad if \quad 0 < |x-a| < \delta_2, \quad then \quad |g(x) - m| < \frac{\epsilon}{2},

Now let δ=min(δ1,δ2)\delta = min(\delta_1, \delta_2). if 0<xa<δ0 < |x-a| < \delta, then 0<xa<δ10 < |x-a| < \delta_1 and 0<xa<δ20 < |x-a| < \delta_2 are both true, so both

f(x)l<ϵ2andg(x)m<ϵ2|f(x) - l| < \frac{\epsilon}{2} \quad and \quad |g(x) - m| < \frac{\epsilon}{2}

are true. But by part (1) of the lemma this implies that

(f+g)(x)(l+m)<ϵ|(f+g)(x) - (l+m)| < \epsilon

This proves (1).

To prove (2) we proceed similarly, after consulting part (2) of the lemma. If ϵ>0\epsilon > 0 there are δ1\delta_1, δ2>0 \delta_2> 0 such that, for all xx,

if0<xa<δ1,andf(x)l<min(1,ϵ2(m+1))if \quad 0 < |x-a| < \delta_1, \quad and \quad |f(x) - l| < min(1, \frac{\epsilon}{2(|m| + 1)}) andif0<xa<δ2,andg(x)m<ϵ2(l+1)and \quad if \quad 0 < |x-a| < \delta_2, \quad and \quad |g(x) - m| < \frac{\epsilon}{2(|l| + 1)}

Again let δ=min(δ1,δ2)\delta = min(\delta_1, \delta_2), if 0<xa<δ0 < |x-a| < \delta, then

f(x)l<min(1,ϵ2(m+1))|f(x) - l| < min(1, \frac{\epsilon}{2(|m| + 1)}) g(x)m<ϵ2(l+1)|g(x) - m| < \frac{\epsilon}{2(|l| + 1)}

So, by the lemma, (fg)(x)lm<ϵ|(f \cdot g)(x) - l \cdot m| < \epsilon, and this proves (2)

Finally, if ϵ>0\epsilon > 0 there is a δ>0\delta > 0 such that, for all xx,

if0<xa<δ,theng(x)m<min(m2,ϵm22)if \quad 0 < |x-a| < \delta, \quad then \quad |g(x) - m| < min(\frac{|m|}{2}, \frac{\epsilon |m|^2}{2})

But according to part (3) of the lemma this means, first, that g(x)0g(x) \neq 0, so (1/g)(x)(1/g)(x) makes sense, and second that

(1g)(x)1m<ϵ|(\frac{1}{g})(x) - \frac{1}{m}| < \epsilon

This proves (3).


Reference: Calculus Micheal Spivak. 5. Limits

分类

Published at:
June 30, 2026
Keywords:
Math
Calculus
Limits