Why Rolle’s Theorem Is Useful in Proofs
Rolle’s Theorem gives a precise connection between the values of a function at two endpoints and the derivative somewhere between them. If the endpoint values agree, and the function behaves regularly throughout the interval, then its graph must have a horizontal tangent at an interior point.
That conclusion is useful well beyond locating a single derivative value. By choosing intervals between roots, we can force derivative zeros and obtain limits on how many roots a function can have. By applying the theorem to a suitably adjusted function, we can also prove that a function whose derivative vanishes everywhere must be constant.
Statement of Rolle’s Theorem
The hypotheses have separate roles. Continuity on the closed interval ensures that the function attains a maximum and a minimum. Differentiability is required at interior points, where the derivative conclusion is made. Equality of the endpoint values ensures that an interior extremum can be found unless the function is constant.
The theorem guarantees at least one such \(c\), not necessarily exactly one. It also does not say that every point in the interval has derivative zero.
Why the Theorem Holds
Here is a proof using two standard facts from calculus: a continuous function on a closed bounded interval attains a maximum and a minimum, and a differentiable function has derivative zero at an interior local maximum or minimum.
If \(f\) is constant on \([a,b]\), then \(f'(x)=0\) for every \(x\in(a,b)\). Since \(a<b\), the interval has interior points, so the conclusion holds.
If \(f\) is not constant, the extreme value theorem gives points in \([a,b]\) where \(f\) attains its maximum and minimum. At least one of these values differs from the common endpoint value \(f(a)=f(b)\).
If the maximum differs from the endpoint value, it cannot occur at either endpoint, since both endpoints have the same value. If instead the minimum differs from the endpoint value, it cannot occur at either endpoint. Thus, in either case, an extremum is attained at some \(c\in(a,b)\).
The function is differentiable at this interior extremum. Fermat’s theorem therefore gives \(f'(c)=0\), as required.
This proof also explains the theorem’s structure: equality at the endpoints alone is not enough. Continuity and differentiability are what allow the endpoint information to lead to an interior derivative conclusion.
Worked Example: Finding a Derivative Zero Between Equal Values
Worked Example: Finding a Derivative Zero Between Equal Values
Let \(f(x)=x^2-4\) on \([-2,2]\). Use Rolle’s Theorem to find a point where \(f'\) is zero, and verify the result.
First, \(f\) is a polynomial, so it is continuous on \([-2,2]\) and differentiable on \((-2,2)\). Its endpoint values are
The hypotheses of Rolle’s Theorem hold, so some \(c\in(-2,2)\) must satisfy \(f'(c)=0\). Differentiate:
Solving \(2c=0\) gives \(c=0\). This point lies in the required open interval, since \(-2<0<2\). To check the conclusion, substitute it into the derivative:
Thus \(c=0\) is the guaranteed interior derivative zero.
Using Roots to Force Derivative Zeros
A root of \(f\) is a point \(r\) where \(f(r)=0\). If \(f\) has two distinct roots \(r_1<r_2\), then its values at the endpoints of \([r_1,r_2]\) agree: both are zero. Provided \(f\) is continuous on this closed interval and differentiable on its interior, Rolle’s Theorem gives a derivative zero strictly between those roots.
For each adjacent pair of roots, Rolle’s Theorem applies because the function values at both endpoints are zero. The resulting derivative zero lies inside that particular interval. These intervals are disjoint, so the zeros obtained from different intervals must be distinct. This is why choosing intervals carefully matters.
Worked Example: Three Roots Force Two Derivative Zeros
Consider \(f(x)=x(x-1)(x-2)\). Find its roots and use Rolle’s Theorem to locate the zeros of \(f'\) between them.
The factors show that the roots are \(0\), \(1\), and \(2\). Since \(f\) is a polynomial, it is continuous and differentiable everywhere. Expand before differentiating:
Rolle’s Theorem guarantees one zero of \(f'\) in \((0,1)\) and another in \((1,2)\). To find them explicitly, solve the quadratic equation:
Because \(0<\frac{\sqrt{3}}{3}<1\), the value \(1-\frac{\sqrt{3}}{3}\) lies in \((0,1)\), while \(1+\frac{\sqrt{3}}{3}\) lies in \((1,2)\). Check each by substitution. Writing \(s=\frac{\sqrt{3}}{3}\), we have \(s^2=\frac{1}{3}\), so
The two computed derivative zeros are distinct and lie in the two separate intervals required by Rolle’s Theorem.
Root Counts and Uniqueness
The root-count result can be used in reverse as a contradiction argument. If a function has no derivative zeros in an interval, it cannot have two distinct roots there: two roots would force a derivative zero between them. More generally, if \(f'\) has at most \(m\) distinct zeros, then \(f\) has at most \(m+1\) distinct roots, provided the relevant continuity and differentiability conditions hold.
Worked Example: Proving a Root Is Unique
Show that \(f(x)=x^3+x-2\) has at most one real root, and check that \(x=1\) is its root.
Differentiate:
For every real \(x\), \(x^2\ge 0\), so \(3x^2+1\ge 1>0\). Thus \(f'\) has no zeros. Suppose, for contradiction, that \(f\) had two distinct real roots \(r_1<r_2\). The polynomial is continuous on \([r_1,r_2]\), differentiable on \((r_1,r_2)\), and satisfies \(f(r_1)=f(r_2)=0\). Rolle’s Theorem would then give a point \(c\in(r_1,r_2)\) with \(f'(c)=0\), contradicting \(f'(x)>0\) for every real \(x\). Therefore \(f\) has at most one real root.
To verify the stated root, substitute \(x=1\):
So \(1\) is a root, and the at-most-one result shows that it is the unique real root.
Proving a Function Is Constant
A particularly useful consequence of Rolle’s Theorem is that a function with zero derivative throughout an interval must be constant there. To prove it, compare the function values at any two points \(a<b\) in the interval. If those values differed, subtract the line joining the two endpoint values. The adjusted function would have equal endpoint values, so Rolle’s Theorem would force its derivative to vanish somewhere. That would force the original derivative to equal the nonzero slope between the two values, contradicting the assumption that the derivative is always zero.
Worked Example: Zero Derivative Forces Equal Values
Suppose \(f\) is continuous on \([0,3]\), differentiable on \((0,3)\), and \(f'(x)=0\) for every \(x\in(0,3)\). Prove that \(f(0)=f(3)\).
Assume instead that \(f(0)\ne f(3)\). Define the slope between the endpoint values by
The assumption \(f(0)\ne f(3)\) implies \(m\ne 0\). Now define \(g(x)=f(x)-mx\) on \([0,3]\). This function is continuous on \([0,3]\) and differentiable on \((0,3)\), and its endpoint values agree:
Rolle’s Theorem gives some \(c\in(0,3)\) such that \(g'(c)=0\). But differentiating \(g(x)=f(x)-mx\) gives \(g'(x)=f'(x)-m\), and hence
This implies \(m=0\), contradicting \(m\ne 0\). Therefore the assumption \(f(0)\ne f(3)\) is false, and \(f(0)=f(3)\).
A Common Pitfall: Equal Endpoints Are Not Enough
A frequent error is to check \(f(a)=f(b)\) and immediately invoke Rolle’s Theorem. The theorem also requires continuity on the entire closed interval and differentiability on its interior. Here is an example that shows why endpoint equality alone does not justify the conclusion.
Worked Example: Checking the Hypotheses Before Applying Rolle
Define \(f\) on \([-1,1]\) by \(f(x)=1\) for \(x\ne 0\), and \(f(0)=0\). The endpoint values agree:
A tempting but invalid argument is: “The endpoint values are equal, so Rolle’s Theorem guarantees a point \(c\in(-1,1)\) with \(f'(c)=0\).” The hypothesis of continuity on \([-1,1]\) fails, because \(f(0)=0\) while \(f(x)=1\) for every \(x\ne0\). Thus Rolle’s Theorem cannot be applied. In this particular example \(f'(x)=0\) for \(x\ne0\), so its conclusion happens to hold; that does not make the attempted application valid.
For a sharper illustration, define \(h(x)=x\) when \(x<0\), and \(h(x)=x-2\) when \(x\ge0\), again on \([-1,1]\). Then
The function is discontinuous at \(0\). At every interior point other than \(0\), its derivative is \(1\), and at \(0\) its derivative does not exist. There is therefore no interior point where its derivative is zero. Equal endpoint values, without the required regularity, do not guarantee Rolle’s conclusion.
A Reliable Proof Strategy
When a problem asks for a derivative zero, a root-count bound, or a constant-function conclusion, first identify the points that should be the endpoints of the interval. Then check the theorem’s hypotheses on that interval, rather than assuming they hold.
For a pair of roots or equal function values, use those points as endpoints. For a root-count argument, use adjacent roots to obtain disjoint intervals.
Check continuity on the closed interval, differentiability on its interior, and equality of the endpoint values.
Write that Rolle’s Theorem gives a point strictly inside the chosen interval where the derivative is zero.
Compare the forced derivative zero with the derivative information in the problem. A contradiction can establish uniqueness or rule out unequal function values.
The key is to make the interval do the work. Correctly chosen endpoints turn equality of function values into a guaranteed interior derivative zero; repeated across disjoint intervals, that simple step yields useful conclusions about roots and constant functions.
Check Your Understanding
For each question, identify the interval and the hypotheses that make the conclusion valid.
- If \(f\) is continuous on \([a,b]\), differentiable on \((a,b)\), and \(f(a)=f(b)\), what does Rolle’s Theorem guarantee?
- If a function has four distinct roots and satisfies the relevant regularity conditions between them, how many distinct derivative zeros are guaranteed?
- Why does the absence of derivative zeros imply that a function can have at most one root on an interval where Rolle’s Theorem applies?
- In the constant-function argument, why is it useful to subtract the line with slope \(\frac{f(b)-f(a)}{b-a}\)?
- For the function that equals \(1\) except at \(0\), where it equals \(0\), which hypothesis of Rolle’s Theorem fails on \([-1,1]\)?
