1
$\begingroup$

I just got a simple question regarding the use of L'Hopitals method for finding limits. Usually L'Hopitals method can be used to find limits like $$\lim_{x\to0}\frac{\sin x}{x} = \lim_{x\to 0}\dfrac{\dfrac{d}{dx} \sin x}{\dfrac{d}{dx} x} = \lim_{x\to 0}\cos x$$

Here if we plug $0$, we can find the limit of the original function $\dfrac{\sin x}{x}$ at $0$ using the $\cos x$ function. Put $0$ in, and you will get $1$, which is correct. However, if we replace $x$ with $\infty$, we don't get the right limit.

$$\cos x$$ $$\cos(\infty)$$

Which is not right for the limit of the original function, as $$\lim_{x\to\infty}\frac{\sin x}{x} = 0$$ Using the new function which we get via L'Hopital's method does not help get that. Is this like a special case? In what cases could then L'Hopital's way not work?

$\endgroup$
12
  • 3
    $\begingroup$ Why do you think that L'Hospital's rule works for $x \to \infty$? $\endgroup$
    – Somos
    Aug 2 '17 at 23:45
  • 3
    $\begingroup$ $\sin x$ has no limit as $x \to \infty$ (and also does not approach either $\infty$ or $-\infty$) so the limit is not in one of the valid indeterminate forms for applying l'Hopital. $\endgroup$ Aug 2 '17 at 23:50
  • 2
    $\begingroup$ @Somos Because it is proven it works $\endgroup$
    – Peyton
    Aug 2 '17 at 23:51
  • 1
    $\begingroup$ The main issue with L'Hospital's Rule is that most often one does not pay attention to all the hypotheses under which it works. Perhaps the rule is thought as a thumb rule of differentiate and plug instead of a genuine mathematical theorem. One of the hypotheses of the rule is that the limit of the $f'/g'$ (expression obtained after differentiation of numerator and denominator) should have a limit (finitely or infinitely). $\endgroup$ Aug 3 '17 at 2:41
  • 1
    $\begingroup$ L'Hôpital only applies when your limit is either $\dfrac00$ or $\dfrac{∞}{∞}$. $\endgroup$ Sep 19 at 15:15
4
$\begingroup$

L'Hopital has the form if $A,$ then $B.$ It doesn't say $A$ iff $B.$ Thus the limit of $f/g$ can equal $L$ (B) even if the limit of $f'/g'$ doesn't exist.

Incidentally, L'Hopital works in cases of $\text { ? }/\infty.$ In other words, if $\lim g(x) = \infty$ and $\lim f'(x)/g'(x) = L,$ then $\lim f(x)/g(x) = L.$

$\endgroup$
1
$\begingroup$

L'Hospital's Rule has a lot of hypotheses. You seem to be ignoring the hypotheses on the limits of the numerator and denominator of the original ratio. Let's check a statement of the Rule.

From Calculus: Early Transcendentals, 4th ed. by Rogawski, Adams, Franzosa (p. 250):

Assume $f$ and $g$ are differentiable in an interval $(b,\infty)$ and that $g'(x) \neq 0$ for $x > b$. If $\lim_{x \rightarrow \infty} f(x)$ and $\lim_{x \rightarrow \infty} g(x)$ exist and either both are zero or both are infinite, then $$ \lim_{x \rightarrow \infty} \frac{f(x)}{g(x)} = \lim_{x \rightarrow \infty} \frac{f'(x)}{g'(x)}$$ provided that the limit on the right exists. A similar result holds for $x \rightarrow -\infty$.

(This quote contains an astoundingly common error. As Rogawski explains on p. 72, infinite limits do not exist. A better phrasing of the second sentence is "If it is the case that $\lim_{x \rightarrow \infty} f(x)$ and $\lim_{x \rightarrow \infty} g(x)$ both exist and are zero or it is the case that both limits are infinite [which in Rogawski includes either sign], then ...". This error is common among those who have studied real analysis because in that setting one works in the extended reals, so $\infty$ and $-\infty$ are points in the working set of numbers and can be the value of a limit. In (just) the reals, this is not the case.)

Applying this to your example, $f(x) = \sin x$ and $g(x) = x$. While it is the case that $$ \lim_{x \rightarrow \infty} g(x) = \lim_{x \rightarrow \infty} x = 0 \text{,} $$ we find that
$$ \lim_{x \rightarrow \infty} f(x) = \lim_{x \rightarrow \infty} \sin x $$ does not exist. (The sine function takes every interval $(b,\infty)$ to $[-1,1]$, so the limit fails to exist. One can arrive at the same conclusion by studying $\lim_{x \rightarrow 0^+} \sin(1/x)$, but with the added advantage that the graph fits on a finite piece of paper.)

So, no, your example is not a special case. L'Hospital's Rule already constrains to which limit expressions it can be applied, excluding the example you give.

$\endgroup$
0
$\begingroup$

You can directly compute the limit of $\frac{\sin(x)}{x}$ as $x \to \infty$, which is 0 and not an indeterminate form - this, L'Hopital's doesn't apply.

$\endgroup$
5
  • 1
    $\begingroup$ Just because you can calculate the limit in another way doesn't automatically mean l'Hopital doesn't apply. e.g. you can calculate $\lim_{x \to \infty} \frac{x}{x}$ easily enough without l'Hopital, but l'Hopital also works on it. $\endgroup$ Aug 3 '17 at 0:49
  • 2
    $\begingroup$ L'Hopital can be applied whenever the denominator $\to \infty.$ $\endgroup$
    – zhw.
    Aug 3 '17 at 0:51
  • $\begingroup$ @zhw. How do you figure? By OP's example, this should be false. $\endgroup$
    – platty
    Aug 3 '17 at 2:15
  • 4
    $\begingroup$ By "can be applied" I mean that if $g\to \infty,$ and if $f'/g'\to L,$ then $f/g\to L.$ $\endgroup$
    – zhw.
    Aug 3 '17 at 3:17
  • $\begingroup$ @zhw. : Then your comment is not an improvement of this Answer in the context of the Question. $f'/g' = \cos(x)$ does not have a limit as $x \rightarrow \infty$. For careful statements of L'Hopital's Rule, your first comment ignores necessary hypotheses on the existence of various limits, claiming, for instance, $\lim_{x \rightarrow \infty} \frac{x \sin x}{x}$ exists, which is false (although your second comment does correct for this particular example). $\endgroup$ Sep 19 at 14:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.