From Guessing to Computing
Tables and graphs are fine for building intuition, but they are impractical as a general method — you cannot tabulate your way through an exam. This concept collects the theorems that let you compute a limit algebraically, in a "mechanical" way, without ever touching a table.
Direct substitution for polynomials
Theorem 9.1. If P(x)=a0+a1x+a2x2+⋯+anxn is a polynomial, then for any real x0,
limx→x0P(x)=P(x0).
In other words: for a polynomial, the limit is just direct substitution. This works because a polynomial has no "trouble points" — it's defined and smooth everywhere, so there's nothing for a limit to disagree with. (A constant function is the degree-0 special case: limx→x0c=c.)
The algebra of limits
Theorem 9.2. Suppose limx→x0f(x) and limx→x0g(x) both exist, and c is a constant. Then all of the following exist, and:
- Constant multiple: x→x0limcf(x)=cx→x0limf(x)
- Sum/Difference: x→x0lim[f(x)±g(x)]=x→x0limf(x)±x→x0limg(x)
- Product: x→x0lim[f(x)g(x)]=x→x0limf(x)⋅x→x0limg(x)
- Quotient: x→x0limg(x)f(x)=limx→x0g(x)limx→x0f(x), provided limx→x0g(x)=0.
These rules extend to any finite number of functions being added, multiplied, etc.
The quotient rule has a hard precondition: it only applies when the denominator's limit is non-zero. If limx→x0g(x)=0, this rule simply does not fire — you cannot conclude the quotient's limit is "undefined" or "infinite" from the rule itself; you must investigate further (rationalising, factoring, or the infinite-limit machinery of Concept 3).
The power rule
Theorem 9.3. If limx→x0f(x) exists, then for any positive integer (in fact any rational) n,
limx→x0[f(x)]n=[limx→x0f(x)]n.
This is really just the product rule applied n times to f⋅f⋯f.
The difference-of-powers limit
Theorem 9.4. For a positive integer n (and in fact for any rational exponent),
limx→ax−axn−an=nan−1.
Why it's true, in brief: factor xn−an=(x−a)(xn−1+xn−2a+xn−3a2+⋯+an−1). Cancel the common factor (x−a) (legitimate since we only ever consider x=a when taking the limit) to leave n terms, each of which tends to an−1 as x→a — giving a total of nan−1. This single result is a shortcut generator: it turns an entire family of "0/0"-looking quotients into instant answers, and — though its proof waits for a later chapter — it is secretly the seed of the derivative of xn.
The three evaluation techniques for a 0/0 form at a finite point
When direct substitution into a quotient produces 00, the quotient theorem is blocked (the denominator's limit is 0), so you cannot stop — you must first rewrite the expression so the offending factor cancels.
1. Factor and cancel. If both numerator and denominator vanish at x0, then (x−x0) is a common factor of both — pull it out and cancel. …