Skip to content
Topos Uranos

Summary

This lesson studies functions whose values grow or decrease without bound. It completes the common scheme of the definitions of limit with the box that was missing, the one for the value near ∞\infty, and it distinguishes infinite limits, which are a precise way of having no limit, from the other forms of divergence. A single lemma, the reciprocal lemma, reduces infinite limits to limits equal to zero, so that the rules for the sum, the product and the quotient are deduced from those of the algebra of limits; only comparison, the theorem on unbounded monotonic functions and composition require an argument of their own. With them, two results that the earlier lessons left half done are completed: the quotient whose denominator tends to zero and the rational functions whose numerator has the higher degree. The lesson also shows why indeterminate forms are not rules, and it ends with composition with infinite limits.

Learning objectives

  1. Define infinite limits, at a point, on one side and at infinity, with the scheme of the challenge and the response, and distinguish them from the other forms of divergence.
  2. Prove the reciprocal lemma and deduce from it and from comparison the rules for the sum, the product and the quotient with infinite limits.
  3. Recognize indeterminate forms and show with examples that they can give any result.
  4. Compute infinite limits of polynomials, rational functions, trigonometric functions and compositions, and prove the theorem on unbounded monotonic functions.

The missing box

In the lesson on limits as the variable tends to infinity a common scheme was described: every definition of limit is a challenge about the values with a response about the variable, and the only thing that changes is what it means to be near each of them. For the variable, all the forms have already been used: near a number aa, on one side of it, or near ±∞\pm\infty. For the values, only one: near a number LL. This lesson completes the scheme with the remaining form, in which the value is near ∞\infty, that is, f(x)>Nf(x) > N, or near −∞-\infty, that is, f(x)<Nf(x) < N.

DefinitionInfinite limits

Let aa be an accumulation point of the domain AA of ff. We say that ff tends to infinity as xx tends to aa, and we write lim⁡x→af(x)=∞\lim_{x \to a} f(x) = \infty, if for every real NN there exists a δ>0\delta > 0 such that f(x)>Nf(x) > N for every xx in AA with 0<∣x−a∣<δ0 < \abs{x - a} < \delta. Likewise, lim⁡x→af(x)=−∞\lim_{x \to a} f(x) = -\infty if for every NN there exists a δ>0\delta > 0 such that f(x)<Nf(x) < N at those xx.

Written with quantifiers, the definition has the usual form; the challenge is now an upper half-plane, that of the points with ordinate greater than NN:

lim⁡x→af(x)=∞↔(∀N)(∃δ>0)(∀x∈A)(0<∣x−a∣<δ→f(x)>N)\lim_{x \to a} f(x) = \infty \leftrightarrow (\forall N)(\exists \delta > 0)(\forall x \in A)\left(0 < \abs{x - a} < \delta \rightarrow f(x) > N\right)

The variants are obtained by changing only the response, as in the earlier lessons: lim⁡x→a+f(x)=∞\lim_{x \to a^+} f(x) = \infty if the condition holds for a<x<a+δa < x < a + \delta; lim⁡x→∞f(x)=∞\lim_{x \to \infty} f(x) = \infty if there exists an MM such that it holds for x>Mx > M; and so on in the other cases. In what follows, PP denotes any of the targets of the variable (aa, a+a^+, a−a^-, ∞\infty or −∞-\infty), and “near PP” means in a punctured interval, on one side or on a ray, as the case may be. All the proofs in the lesson use only one common property: two conditions that hold near PP hold simultaneously near PP, by taking the smaller of two radii or the larger of two bounds. The theorems on finite limits that are cited at an arbitrary target PP (uniqueness, local boundedness, preservation of sign, the algebra of limits and the lemmas of the lesson on the algebra of limits) hold at all of them by the lesson on one-sided limits and by the inversion lemma of the preceding lesson, whose correspondence δ=1M\delta = \frac{1}{M} between the bound and the radius carries infinite limits over in the same way, because it does not depend on the form of the challenge.

It is worth insisting on what the notation says. If lim⁡x→af(x)=∞\lim_{x \to a} f(x) = \infty, the function has no limit at aa in the sense of the lesson on the limit of functions of a real variable, because it is not bounded near aa, and local boundedness says that a function with a limit is. The equality with ∞\infty describes a precise way of having no limit. We say that ff diverges at PP if it has no finite limit there; tending to ∞\infty or to −∞-\infty are two ways of diverging, but not the only ones. The function sin⁡1x\sin \frac{1}{x} diverges at 00 and is bounded; the function 1xsin⁡1x\frac{1}{x}\sin\frac{1}{x} diverges at 00, is not bounded near 00 and, nevertheless, tends neither to ∞\infty nor to −∞-\infty, because it vanishes at points as close to 00 as we like.

PropositionThe reciprocals of x2x^2 and of xx at zero

lim⁡x→01x2=∞\lim_{x \to 0} \frac{1}{x^2} = \infty. On the other hand, lim⁡x→0+1x=∞\lim_{x \to 0^+} \frac{1}{x} = \infty and lim⁡x→0−1x=−∞\lim_{x \to 0^-} \frac{1}{x} = -\infty, so that 1x\frac{1}{x} tends neither to ∞\infty nor to −∞-\infty as xx tends to 00. The response to the challenge NN is the square root of the reciprocal of NN. In the figure, the upper half-plane is the challenge; the strip adjacent to 00, the response.

Proof

  1. δ=1N\delta = \dato{d}{\frac{1}{\sqrt{N}}}

    Let NN be a real number; it suffices to consider N>0N > 0, because a response to a larger challenge responds to all the smaller ones. We take as the response the square root of its reciprocal.

  2. 0<x2<1N→1x2>N0 < x^2 < \frac{1}{N} \rightarrow \resaltar{\frac{1}{x^2} > N}

    If 0<∣x∣<δ0 < \abs{x} < \delta, the square of xx is positive and less than 1N\frac{1}{N}; taking reciprocals of positive numbers reverses the inequality.

  3. 0<x<1N→1x>N,−1N<x<0→1x<−N0 < x < \frac{1}{N} \rightarrow \frac{1}{x} > N, \qquad -\frac{1}{N} < x < 0 \rightarrow \frac{1}{x} < -N

    For 1x\frac{1}{x} from the right, the proposition on the reciprocal from the preceding lesson gives the response δ=1N\delta = \frac{1}{N}. From the left, 1x\frac{1}{x} is the opposite of 1−x\frac{1}{-x}, and −x-x lies to the right of 00: the values are less than −N-N.

  4. lim⁡x→0+1x=∞≠−∞=lim⁡x→0−1x\lim_{x \to 0^+} \frac{1}{x} = \infty \neq -\infty = \lim_{x \to 0^-} \frac{1}{x}

    In every punctured interval with center 00 there are values of 1x\frac{1}{x} greater than 11 and less than −1-1: no δ\delta responds to the challenge N=1N = 1 for ∞\infty, nor to the challenge N=−1N = -1 for −∞-\infty.