Why Compare Derivatives?
When both a numerator and a denominator approach zero, direct substitution gives no value for their quotient. The way each function approaches zero can still be informative: comparing their derivatives compares their instantaneous rates of change near the endpoint.
L'Hôpital's Rule makes that comparison precise. It does not say that every quotient with a zero numerator and denominator can be replaced by a quotient of derivatives. Its conclusion follows under specific hypotheses, and the Cauchy Mean Value Theorem explains why those hypotheses are enough.
The 0/0 Form of L'Hôpital's Rule
We begin with a right-hand version at a finite endpoint \(a\). The functions need only be defined on a short interval to the right of \(a\). Continuity at \(a\) ensures that both functions actually approach zero there.
Here \(\delta>0\), and the derivative ratio is considered only where the derivatives exist and \(g'\) is nonzero. The condition \(g'(x)\ne0\) also ensures that \(g(x)\ne0\) for \(x>a\): if \(g(x)=g(a)=0\) at such a point, Rolle's Theorem applied on \([a,x]\) would give a point between them where \(g'=0\). Thus the original quotient is defined near the endpoint, on the right.
The theorem is a sufficient condition: if its hypotheses hold, the stated conclusion follows. It is not a claim that the hypotheses are necessary in every problem. A two-sided limit can be handled by applying the right-hand result on one side and the corresponding left-hand result on the other; both one-sided conclusions must agree for the two-sided limit to exist.
Why the Cauchy Mean Value Theorem Gives the Result
The Cauchy Mean Value Theorem states that if \(f\) and \(g\) are continuous on \([u,v]\) and differentiable on \((u,v)\), then there is some \(c\in(u,v)\) such that \[ (f(v)-f(u))g'(c)=(g(v)-g(u))f'(c). \] When \(g'(c)\ne0\) and \(g(v)\ne g(u)\), this identity can be rearranged as a ratio of increments equal to a ratio of derivatives. For L'Hôpital's Rule, take \(u=a\) and \(v=x\).
Fix \(x\) with \(a<x<a+\delta\), and apply the Cauchy Mean Value Theorem to \(f\) and \(g\) on \([a,x]\). The continuity and differentiability assumptions in the theorem hold on this interval.
For some \(c_x\in(a,x)\), the theorem gives \((f(x)-f(a))g'(c_x)=(g(x)-g(a))f'(c_x)\). Since \(f(a)=g(a)=0\), this becomes \(f(x)g'(c_x)=g(x)f'(c_x)\).
Because \(g(x)\ne0\) and \(g'(c_x)\ne0\), divide to obtain \(\frac{f(x)}{g(x)}=\frac{f'(c_x)}{g'(c_x)}\).
The intermediate point satisfies \(a<c_x<x\). As \(x\to a^+\), this forces \(c_x\to a^+\), so the assumed limit of the derivative ratio determines the limit of the original ratio.
For completeness, the last step can be written as an epsilon argument when \(L\) is finite. Given any \(\varepsilon>0\), the assumption \(\lim_{t\to a^+}f'(t)/g'(t)=L\) provides \(\eta>0\) such that \(a<t<a+\eta\) implies \[ \left|\frac{f'(t)}{g'(t)}-L\right|<\varepsilon. \] If \(a<x<a+\eta\), then \(a<c_x<x<a+\eta\), so the same inequality holds at \(t=c_x\). Using the quotient identity, \[ \left|\frac{f(x)}{g(x)}-L\right| = \left|\frac{f'(c_x)}{g'(c_x)}-L\right| <\varepsilon. \] This is exactly the definition that \(f(x)/g(x)\to L\) as \(x\to a^+\).
If the derivative ratio tends to \(+\infty\), its definition says that for every \(M>0\), it is greater than \(M\) whenever its argument is sufficiently close to \(a\) on the right. For all sufficiently close \(x>a\), the intermediate point \(c_x\) is in that same neighborhood. The quotient identity then gives \(f(x)/g(x)>M\). This proves that the original ratio tends to \(+\infty\). For a limit of \(-\infty\), use the corresponding condition that the ratio is less than \(-M\).
Worked Examples
Worked Example: An Exponential Difference Divided by \(x\)
Find \(\lim_{x\to0}\frac{e^x-1}{x}\). At \(x=0\), the numerator is \(e^0-1=0\), and the denominator is \(0\), so this is the 0/0 form. Set \(f(x)=e^x-1\) and \(g(x)=x\). Both functions are continuous at zero, differentiable nearby, and vanish at zero. Their derivatives are \(f'(x)=e^x\) and \(g'(x)=1\); the denominator derivative is nonzero.
L'Hôpital's Rule therefore gives \(\lim_{x\to0}\frac{e^x-1}{x}=1\). Check the derivative calculation: differentiating \(e^x-1\) gives \(e^x\), and differentiating \(x\) gives \(1\). Their ratio tends to \(e^0/1=1\), as required by the theorem's conclusion.
Worked Example: Comparing Two Exponential Differences
Find \(\lim_{x\to0}\frac{e^x-1}{e^{2x}-1}\). Substitution gives \(0/0\). Let \(f(x)=e^x-1\) and \(g(x)=e^{2x}-1\); each is continuous at zero and has value zero there. By the chain rule, \(f'(x)=e^x\) and \(g'(x)=2e^{2x}\). Since \(2e^{2x}>0\), the denominator derivative is nonzero near zero.
Thus \(\lim_{x\to0}\frac{e^x-1}{e^{2x}-1}=\frac12\). The calculation is checked by differentiating both functions: the chain rule supplies the factor \(2\) in \(g'(x)\), and the resulting derivative ratio has limit \(1/(2e^0)=1/2\). Omitting that chain-rule factor would change the answer and would not be a correct application of the theorem.
Worked Example: A Logarithmic Difference Divided by \(x\)
Find \(\lim_{x\to0}\frac{\ln(1+x)}{x}\). The expression is defined for \(x>-1\), and substitution at zero gives \(\ln(1)-0=0\). Take \(f(x)=\ln(1+x)\) and \(g(x)=x\). These functions are continuous at zero, differentiable near zero, and both vanish at zero. Their derivatives are \(f'(x)=1/(1+x)\) and \(g'(x)=1\); in a sufficiently small neighborhood of zero, \(g'\) is nonzero.
It follows that \(\lim_{x\to0}\frac{\ln(1+x)}{x}=1\). To check the derivative ratio, \(1/(1+0)=1\); the denominator derivative is indeed \(1\), so no additional factor is introduced. The theorem gives the same value for the original quotient's limit.
A Common Misuse
A common error is to differentiate the numerator and denominator whenever a quotient appears, without first checking that the original limit has the required form and that the other hypotheses hold. For example, consider \(\lim_{x\to0^+}\frac{1+x}{x}\). A mistaken calculation might differentiate the numerator and denominator and report \(1/1=1\).
The rule does not apply: the numerator tends to \(1\), not \(0\), so the original expression is not of the 0/0 form. Directly, \[ \frac{1+x}{x}=\frac{1}{x}+1\longrightarrow+\infty \qquad (x\to0^+), \] not \(1\). Checking the original form is therefore essential. Also, the quotient of derivatives in the theorem is \(f'/g'\); it is not obtained by differentiating the quotient \(f/g\).
What the Derivative-Ratio Limit Tells You
The Cauchy Mean Value Theorem supplies the central link: for each nearby \(x\), the original quotient equals a derivative ratio evaluated at some point between \(a\) and \(x\). Those intermediate points must approach \(a\) as \(x\) approaches \(a\). Consequently, a finite derivative-ratio limit gives the same finite limit for the original quotient, and an infinite derivative-ratio limit gives the matching signed infinite limit.
This is why the hypotheses matter. Continuity at the endpoint supplies the zero values; differentiability permits the Cauchy Mean Value Theorem; and a nonzero denominator derivative ensures the quotient of derivatives is available. Without the required hypotheses, this theorem provides no conclusion. In a specific problem, first verify the 0/0 form and the local conditions, then compute the derivative ratio and its limit, and only then state the limit of the original quotient.
Check Your Understanding
Use the theorem statement and its derivation to answer the following questions.
- Why must \(g(x)\ne0\) for \(x>a\) under the stated hypotheses?
- When applying the Cauchy Mean Value Theorem on \([a,x]\), where does the point \(c_x\) lie?
- Explain why \(c_x\to a^+\) as \(x\to a^+\).
- If \(f'(x)/g'(x)\to L\) for finite \(L\), which quotient has limit \(L\), and what hypotheses are needed?
- Why does the calculation using \(f(x)=1+x\) and \(g(x)=x\) fail to be an application of the 0/0 form?
