Introduction
A single sign change can guarantee that a continuous function has a root. The same idea can be applied more than once: if the signs at the endpoints change on several separate intervals, continuity guarantees a root in every one of them.
The separation of the intervals matters. A root found in one interval must not be the very same point as a root found in another. With properly separated intervals, the Intermediate Value Theorem gives not just several root-containing intervals, but several distinct roots.
The multiple-interval result
Recall the sign-change consequence of the Intermediate Value Theorem: if \(f\) is continuous on a closed interval \([a,b]\) and \(f(a)\) and \(f(b)\) have opposite signs, then \(f(c)=0\) for some \(c\) strictly between \(a\) and \(b\). The condition that the signs are opposite can be written as \(f(a)f(b)<0\).
The intervals in the theorem have a gap between consecutive endpoints. In particular, they do not overlap or share an endpoint. Continuity is required on each interval where the theorem is applied; continuity elsewhere is not needed for this conclusion.
Why the result guarantees distinct roots
Apply the sign-change consequence of the Intermediate Value Theorem on the first interval. It gives a root \(c_1\in(a_1,b_1)\). Apply it on the second interval to get \(c_2\in(a_2,b_2)\), and continue through the list. Each application is valid because the function is continuous on that interval and its endpoint values have opposite signs.
It remains to justify that these roots are distinct. If \(i<j\), the intervals are ordered and separated, so every point in \((a_i,b_i)\) is less than every point in \((a_j,b_j)\). In particular, \(c_i<c_j\). Thus \(c_i\ne c_j\), and all the roots obtained are distinct.
This argument is a repeated application of the sign-change consequence of the Intermediate Value Theorem, followed by a separation argument. The theorem does not require solving \(f(x)=0\), nor does it identify the roots exactly.
Worked examples
Worked Example: Three roots of a cubic
Show that \(p(x)=x^3-3x\) has at least three distinct real roots by using the intervals \([-2,-1]\), \([-1/2,1/2]\), and \([1,2]\).
First, \(p\) is a polynomial, so it is continuous on each of these closed intervals. Now evaluate the endpoints of the first interval:
The endpoint values have opposite signs: \((-2)(2)=-4<0\).
For the second interval, calculate:
These values have opposite signs: \(\frac{11}{8}\left(-\frac{11}{8}\right)<0\).
For the third interval:
Again, the endpoint values have opposite signs: \((-2)(2)=-4<0\). The intervals are ordered and separated because \(-1<-\frac12\) and \(\frac12<1\). The theorem therefore gives at least one root in the interior of each interval, and the three roots are distinct. Checking the endpoint calculations confirms that none of the signs used in the argument is zero.
Worked Example: Two roots of a trigonometric function
Show that \(f(x)=\cos x\) has at least two distinct roots in the intervals \([\pi/3,2\pi/3]\) and \([4\pi/3,5\pi/3]\).
The cosine function is continuous, so it is continuous on both intervals. At the endpoints of the first interval, the exact values are:
The product is \(\frac12(-\frac12)=-\frac14<0\), so there is a root in \((\pi/3,2\pi/3)\).
At the endpoints of the second interval:
Here the product is \((-\frac12)(\frac12)=-\frac14<0\), so there is a root in \((4\pi/3,5\pi/3)\). The intervals are separated because \(2\pi/3<4\pi/3\). Thus the two roots guaranteed by the theorem are distinct. The endpoint values are exact, and their opposite signs verify the required condition on both intervals.
Worked Example: At least four roots of a quartic
Show that \(q(x)=x^4-5x^2+4\) has at least four distinct real roots by checking four separated intervals.
The function \(q\) is a polynomial, hence continuous. Its values depend only on \(x^2\), so \(q(-x)=q(x)\). We can calculate the needed values using positive magnitudes. At \(x=9/4\) and \(x=7/4\), the calculations are:
Therefore \(q\) has opposite signs at the endpoints of \([-9/4,-7/4]\) as well as \([7/4,9/4]\), using \(q(-x)=q(x)\).
For the two inner intervals, calculate:
These values show opposite signs at the endpoints of \([-5/4,-3/4]\) and \([3/4,5/4]\), again by \(q(-x)=q(x)\). All four sign changes are strict: the numerators \(1105\), \(-495\), \(-351\), and \(385\) are nonzero. The intervals are ordered and separated: \(-7/4<-5/4\), \(-3/4<3/4\), and \(5/4<7/4\). The theorem guarantees at least four distinct roots, one in the interior of each interval. This conclusion follows from continuity and the endpoint signs; no root has been calculated explicitly.
A common mistake: counting the same root twice
A sign change on each of two intervals does not by itself guarantee two distinct roots if the intervals overlap. For example, let \(f(x)=x\), and consider \([-1,1]\) and \([-1/2,1/2]\). The function is continuous on both intervals, and its endpoint values change sign on each:
A mistaken conclusion would be: “There is a root in each interval, so there are two distinct roots.” The theorem does give a root in each interval, but both conclusions can refer to the same point, \(x=0\). Since the intervals overlap, the distinctness argument does not apply.
The correction is to check the intervals themselves, not only the endpoint signs. To conclude that roots are distinct by this method, use separated intervals, such as intervals ordered with \(b_i<a_{i+1}\). Then any root in one interval must differ from any root in another.
A reliable way to organize the argument
When a problem asks you to prove that a function has several distinct roots, keep the hypotheses visible as you work:
List the closed intervals and verify that they are ordered and separated.
State why the function is continuous on each interval.
Show the endpoint values, including enough arithmetic to verify their signs.
For each interval, opposite endpoint signs guarantee a root in its interior.
Use the separation of the intervals to explain why the roots obtained are different.
Why this method matters
Often an equation is difficult or impossible to solve explicitly, but evaluating a function at selected points is straightforward. A sign change then gives rigorous evidence that a solution exists between those points. Repeating this process on separate intervals turns several simple endpoint checks into a proof of several distinct solutions.
The method is a guarantee, not a formula for locating each root. It also does not say that an interval contains only one root. Its strength is that it establishes a lower bound on how many distinct roots exist, using continuity, endpoint signs, and separation.
Check Your Understanding
Answer these questions using the conditions and reasoning from this tutorial.
- What continuity and endpoint-sign conditions guarantee a root in one closed interval?
- Why must consecutive intervals be separated to conclude that the guaranteed roots are distinct?
- A continuous function has opposite signs at the endpoints of each of four separated intervals. What can you conclude about its roots?
- Why do sign changes on two overlapping intervals fail to guarantee two distinct roots?
- Does a sign change on an interval prove that there is exactly one root in that interval?
