4.1.2Calculus I — Limits & Derivatives

Limit laws — sum, product, quotient, constant multiple

1,652 words8 min readdifficulty · medium

WHAT are the laws?

The two atoms everything is built from: limxac=candlimxax=a\lim_{x\to a} c = c \qquad\text{and}\qquad \lim_{x\to a} x = a

Figure — Limit laws — sum, product, quotient, constant multiple

WHY are they true? (Derivation from scratch)

A limit limxaf(x)=L\lim_{x\to a}f(x)=L means: we can force f(x)f(x) as close to LL as we like by keeping xx near aa. Formally, for every ε>0\varepsilon>0 there is δ>0\delta>0 so that 0<xa<δ    f(x)L<ε.0<|x-a|<\delta \;\Rightarrow\; |f(x)-L|<\varepsilon.

Sum law (full proof)

We want (f+g)(L+M)|(f+g)-(L+M)| small. Regroup and use the triangle inequality p+qp+q|p+q|\le|p|+|q|: (f(x)+g(x))(L+M)=(f(x)L)+(g(x)M)f(x)L+g(x)M.|(f(x)+g(x))-(L+M)| = |(f(x)-L)+(g(x)-M)| \le |f(x)-L| + |g(x)-M|. Why this step? We split the total error into "ff's error" + "gg's error", each of which we can control.

Given ε>0\varepsilon>0, choose δ1\delta_1 making fL<ε2|f-L|<\tfrac{\varepsilon}{2} and δ2\delta_2 making gM<ε2|g-M|<\tfrac{\varepsilon}{2}. Take δ=min(δ1,δ2)\delta=\min(\delta_1,\delta_2). Then both hold at once: (f+g)(L+M)<ε2+ε2=ε.|(f+g)-(L+M)| < \tfrac{\varepsilon}{2}+\tfrac{\varepsilon}{2}=\varepsilon.\qquad\blacksquare Why split ε\varepsilon as ε/2+ε/2\varepsilon/2 + \varepsilon/2? So the two halves add back exactly to ε\varepsilon.

Constant multiple (proof)

cf(x)cL=cf(x)L.|cf(x)-cL| = |c|\,|f(x)-L|. If c0c\neq0, make fL<εc|f-L|<\tfrac{\varepsilon}{|c|}, so the product is <ε<\varepsilon. (If c=0c=0 both sides are 00.) Why divide by c|c|? So that multiplying back by c|c| lands on ε\varepsilon.

Product (sketch of the key trick)

The clever move is add and subtract Lg(x)L\,g(x): fgLM=fgLg+LgLM=g(fL)+L(gM).fg - LM = f g - Lg + Lg - LM = g(f-L) + L(g-M). Why this step? It separates "ff wobbling" from "gg wobbling". Near aa, gg is bounded (close to MM), so g(fL)0g(f-L)\to 0, and L(gM)0L(g-M)\to0. Both go to 00, so the product limit is LMLM.

Quotient

First prove lim1g=1M\lim \tfrac{1}{g}=\tfrac1M (needs M0M\neq0 so gg stays away from 00 nearby), then apply the product law to f1gf\cdot\tfrac1g. Why need M0M\neq0? Dividing by something heading to 00 blows up — the result need not exist.


HOW to use them — worked examples



Recall Feynman: explain to a 12-year-old

Imagine two friends walking toward a meeting spot. One ends up at chair LL, the other at chair MM. If you measure the distance between them, it heads to L+ML+M apart; if you stack them, LML\cdot M. Knowing where each friend is going tells you where any combination is going — unless you try to divide by a friend who walks all the way to zero, because then "share per person" explodes. That's the whole game: limits slide through +,,×+,-,\times, and through ÷\div only if the bottom isn't heading to zero.


Flashcards

Sum law statement
lim(f+g)=limf+limg=L+M\lim(f+g)=\lim f+\lim g=L+M, when both limits exist.
Constant multiple law
lim(cf)=climf=cL\lim(c\,f)=c\lim f=cL.
Product law
lim(fg)=(limf)(limg)=LM\lim(fg)=(\lim f)(\lim g)=LM.
Quotient law + condition
lim(f/g)=L/M\lim(f/g)=L/M, valid only if M0M\neq0.
Two atomic limits everything builds on
limxac=c\lim_{x\to a}c=c and limxax=a\lim_{x\to a}x=a.
Key trick in the product-law proof
Add & subtract Lg(x)Lg(x): fgLM=g(fL)+L(gM)fg-LM=g(f-L)+L(g-M).
Why split ε\varepsilon into ε/2\varepsilon/2 in the sum proof
So the two bounded errors add back to exactly ε\varepsilon.
What to do when the quotient gives 0/00/0
Law fails; factor/cancel/rationalize, then take the limit.
Does lim(f+g)\lim(f+g) existing imply each limit exists
No — counterexample x+(x)\lfloor x\rfloor + (-\lfloor x\rfloor).
Limit of a polynomial at aa
p(a)p(a) — direct substitution (from sum + const-mult + power laws).

Connections

  • Epsilon-Delta definition of a limit — the foundation every law is proved from.
  • One-sided limits — laws apply identically to left/right limits.
  • Indeterminate forms 0 over 0 — what to do when the quotient law fails.
  • Continuity — a function is continuous at aa iff limxaf=f(a)\lim_{x\to a}f=f(a); limit laws ⇒ sums/products of continuous functions are continuous.
  • Squeeze theorem — the tool for limits the algebra laws can't reach.
  • Derivative as a limit — every derivative rule (sum/product/quotient rule) descends from these laws.

Concept Map

foundation of

foundation of

proves

proves via

proves

proves

applied to f times 1 over g

requires

repeated gives

building block for

building block for

Epsilon-delta definition

lim c = c

lim x = a

Triangle inequality

Sum law: L plus M

Constant multiple: cL

Add and subtract Lg trick

Product law: L times M

Quotient law: L over M

M not equal 0

Power law

Compute polynomial limits

Hinglish (regional understanding)

Intuition Hinglish mein samjho

Dekho, limit ka matlab simple hai: jaise-jaise xx kisi point aa ke paas jaata hai, f(x)f(x) kis value ki taraf bhaag raha hai. Ab limit laws kehte hain ki agar tumhe ff aur gg dono ki limit pata hai (LL aur MM), to inke sum, difference, product, aur quotient ki limit bhi turant nikal jaati hai — bas LL aur MM ko usi operation se jod do. Yani limit andar ghus jaati hai operations ke through. Isliye humein har baar epsilon-delta wala lamba kaam nahi karna padta; bade function ko chhote-chhote known tukdo me toR kar reassemble kar lete hain.

Sabse important baat: quotient law me ek shart hai — neeche wali limit MM zero nahi honi chahiye. Agar M=0M=0 aa gaya, to law fail, aur tumhe factor karna padega (jaise x29x3=x+3\frac{x^2-9}{x-3}=x+3 wala trick). Yahan x3x\to3 ka matlab hai xx exactly 33 nahi, sirf paas, isliye cancel karna bilkul legal hai.

Ek aur galti se bachna: agar lim(f+g)\lim(f+g) exist karti hai, iska matlab ye nahi ki ff aur gg ki alag-alag limit bhi exist karti hai. Laws tabhi lagao jab dono pieces ki limit pehle se exist karti ho. Mnemonic yaad rakho: "Same Operation, Same Result — par Zero se divide mat karo." Bas itna pakad lo, to polynomial aur rational functions ki saari limits seconds me ho jayengi.

Go deeper — visual, from zero

Test yourself — Calculus I — Limits & Derivatives

Connections