What the Intermediate Value Theorem Says
A continuous function can change its output as its input moves across an interval. The Intermediate Value Theorem (IVT) gives a precise guarantee about which outputs must occur during that change.
The key information comes from two places: the function's values at the endpoints, and whether the function is continuous throughout the interval between them. If a target output lies between the endpoint outputs, continuity guarantees that the function attains it somewhere on the interval.
Here, “between, inclusive” means
The theorem guarantees at least one input \(c\) with the required output. It does not generally identify that input, and it does not say there is only one such input.
Read the Hypotheses Carefully
The IVT is a conditional statement: the guarantee follows when all its hypotheses hold. Each condition has a specific role.
- A closed interval: The function must be defined on \([a,b]\), including both endpoints, with \(a<b\).
- Continuity throughout the interval: The function must be continuous at every point between the endpoints and at the endpoints, using one-sided continuity there.
- A target between the endpoint values: The proposed output \(y\) must satisfy the inclusive inequality in the theorem.
Continuity outside \([a,b]\) is not required. For example, a function may be continuous on \([a,b]\) even if it is not defined for inputs outside that interval. What matters is that it is defined and continuous everywhere on the interval used in the claim.
If the target is strictly between the endpoint values, then it cannot equal either endpoint output. Therefore the input supplied by the theorem must lie in the interior:
If the target equals an endpoint value, the theorem still applies, but its guaranteed input may simply be that endpoint. The inclusive version is useful because it covers both endpoint targets and strictly intermediate targets.
Why Continuity Leads to the Guarantee
Imagine moving the input from \(a\) to \(b\), while following the output of \(f\). If the starting output is below a target \(y\) and the ending output is above it, a continuous function cannot make that change by jumping from one side of \(y\) to the other. It must take the value \(y\) somewhere along the way. The same reasoning applies when the endpoint outputs occur in the reverse order.
This picture explains the role of continuity, but the theorem is stronger than a drawing: it applies to every function satisfying the hypotheses, whether or not its graph is easy to sketch. The endpoint-value condition can also be written as a sign test. To see this, if \(f(a)\leq f(b)\), a value \(y\) between them makes \(y-f(a)\geq0\) and \(y-f(b)\leq0\); if \(f(b)\leq f(a)\), the signs reverse. Thus, in either order,
This equivalence is a way to check the endpoint requirement; it is not a replacement for continuity. The theorem's central guarantee comes from having both the endpoint condition and continuity on the entire interval.
Applying the Theorem
Worked Example: A Target Between Increasing Endpoint Values
Let \(f(x)=3x-2\) on \([0,2]\). Determine whether the IVT guarantees an input \(c\in[0,2]\) such that \(f(c)=1\).
First, \(f\) is a linear function, so it is continuous at every real number and in particular throughout \([0,2]\). Its endpoint values are
The target \(1\) satisfies \(-2\leq1\leq4\). All the IVT hypotheses hold, so there exists \(c\in[0,2]\) such that \(f(c)=1\). In this example we can find and check the input directly:
Substitution verifies the result: \(f(1)=3(1)-2=1\). The theorem guarantees existence; the direct calculation supplies the particular input here.
Worked Example: Endpoint Values in the Reverse Order
Let \(f(x)=x^2\) on \([-2,1]\). Does the IVT guarantee that \(f\) takes the value \(2\)?
The polynomial \(x^2\) is continuous on the whole interval. The endpoint values are
The endpoint values are in decreasing order, but that causes no problem. The target satisfies \(1\leq2\leq4\), so the theorem guarantees some \(c\in[-2,1]\) with \(f(c)=2\). To check a specific input, solve \(c^2=2\). The two possible real solutions are \(c=\sqrt{2}\) and \(c=-\sqrt{2}\). Since \(1<\sqrt{2}\), the positive solution is outside the interval; since \(1<\sqrt{2}<2\), the negative solution satisfies \(-2<-\sqrt{2}<-1\), so it lies in \([-2,1]\).
This confirms the guaranteed output at an input in the interval. The theorem itself did not require the endpoint values to be listed in increasing order.
Worked Example: A Continuous Function That Is Not Monotone
Consider \(f(x)=x^2-2x\) on \([0,3]\), with target \(y=2\). The function is a polynomial, so it is continuous throughout the interval. Calculate its endpoint values:
Because \(0\leq2\leq3\), the IVT guarantees at least one \(c\in[0,3]\) such that \(f(c)=2\). To check an input explicitly, solve:
Completing the square gives
Since \(1-\sqrt{3}<0\), that solution is outside \([0,3]\). The other solution satisfies \(0<1+\sqrt{3}<3\), because \(0<\sqrt{3}<2\). Substitution checks it:
The function is not monotone on this interval, but monotonicity is not an IVT hypothesis. Continuity and the endpoint-value condition are enough for the guarantee.
A Common Mistake: Ignoring Continuity
A common error is to check that a target is between the endpoint values and then apply the IVT without checking continuity. The endpoint test alone does not justify the conclusion.
Worked Example: Why the Continuity Check Matters
Define \(g:[-1,1]\to\mathbb{R}\) by
A mistaken argument might note that \(g(-1)=1\), \(g(1)=2\), and \(3/2\) is between \(1\) and \(2\), then claim that the IVT guarantees \(g(c)=3/2\). That conclusion is not justified: \(g\) is not continuous at \(0\). In fact, for every \(x\) in the domain, \(g(x)\) is either \(1\) or \(2\), so there is no input with \(g(x)=3/2\).
The correction is to verify all the hypotheses before applying the theorem. Here the target condition holds, but continuity fails, so the IVT gives no guarantee. This example does not contradict the theorem; it fails one of its assumptions.
What the Theorem Does and Does Not Tell You
The IVT is an existence result. Once its hypotheses hold, it says that at least one input produces the chosen intermediate output. It does not usually tell you where that input is, how to calculate it exactly, or whether other inputs produce the same output.
In particular, do not add a monotonicity condition: the theorem does not require the function to keep increasing or keep decreasing. Nor does it require a formula that can be solved algebraically. Its strength is that a continuity argument can guarantee an output even when finding the corresponding input would be difficult.
To use the theorem reliably, identify the interval first, verify continuity on the whole interval, calculate both endpoint outputs, and check that the target lies between them. Only then state the existence conclusion with the input \(c\) in the correct interval.
Choose endpoints \(a<b\) and make sure the function is defined on \([a,b]\).
Verify that \(f\) is continuous throughout \([a,b]\), including at its endpoints.
Calculate \(f(a)\) and \(f(b)\), then check the inclusive between condition for \(y\).
Conclude that some \(c\in[a,b]\) satisfies \(f(c)=y\); use \(c\in(a,b)\) when \(y\) is strictly between the endpoint outputs.
Check Your Understanding
For each question, pay attention to both the endpoint values and continuity.
- State the hypotheses and conclusion of the Intermediate Value Theorem for \(f\) on \([a,b]\).
- If \(f(a)=5\), \(f(b)=-1\), and \(f\) is continuous on \([a,b]\), which values \(y\) are covered by the theorem?
- If \(y\) is strictly between \(f(a)\) and \(f(b)\), where must the guaranteed input \(c\) lie?
- Does the IVT require \(f\) to be increasing or decreasing? Explain briefly.
- A function has endpoint values \(0\) and \(4\), and a target value \(2\). What additional hypothesis must be checked before using the IVT?
