Why Check for Every Intersection?
Solving an equation such as \(f(x)=g(x)\) produces possible intersection coordinates, but a list of plausible solutions is not automatically a complete list. A missed intersection can change the description of the region bounded by the curves and affect later calculations. The important question here is narrower: how can you verify that the proposed list includes every intersection on the interval?
The key is to study the difference between the two functions. Every intersection corresponds to a zero of that difference, so finding all intersections becomes a root-checking problem on a specified interval. Algebra can sometimes settle the question exactly. When it cannot, graphs and interval tests can guide a proof, but a visual estimate by itself is not a guarantee of completeness.
Why the Difference Function Captures Intersections
Suppose the curves are \(y=f(x)\) and \(y=g(x)\). At a common point with \(x\)-coordinate \(c\), their \(y\)-coordinates must agree: \(f(c)=g(c)\). Subtracting \(g(c)\) from both sides gives \(f(c)-g(c)=0\), which is \(D(c)=0\). Conversely, if \(D(c)=0\), then \(f(c)-g(c)=0\), so \(f(c)=g(c)\); the point \((c,f(c))\) lies on both curves.
This equivalence is the reason the method works. It also gives a useful checklist: identify the difference, solve or analyze \(D(x)=0\), keep only coordinates in the proposed interval, and check that the original functions are defined at each coordinate.
An interval matters. A function may have zeros outside the interval that are not relevant to the requested intersection list. Also, if the interval is closed, its endpoints are included and must be checked; if an endpoint is not included, a root there is not an intersection on the interval. Be explicit about which interval is being considered.
Factor the Difference When Possible
For polynomial differences, factoring is often the most direct way to certify a complete list. A factorization reveals the roots, and solving each factor equal to zero accounts for every root of the product. Check that each resulting coordinate lies in the interval. A repeated factor still gives only one distinct coordinate, even though it appears more than once in the factorization.
Worked Example: Factoring to Find Every Intersection
Find and verify every intersection of \(f(x)=x^3-3x\) and \(g(x)=0\) on \([-2,2]\).
First form the difference:
Factor out \(x\), then factor the difference of squares:
A product is zero exactly when at least one factor is zero. Thus the only real roots are \(x=0\), \(x=\sqrt{3}\), and \(x=-\sqrt{3}\). All three lie in \([-2,2]\), because \(0\) lies between the endpoints and \(\sqrt{3}<2\), so also \(-2<-\sqrt{3}\).
Check by substitution into the difference:
The factorization also certifies that no other real roots exist: every zero must make one of the three displayed factors zero. The intersection coordinates are therefore exactly \(-\sqrt{3}\), \(0\), and \(\sqrt{3}\). Since \(g(x)=0\), the corresponding points are \((-\sqrt{3},0)\), \((0,0)\), and \((\sqrt{3},0)\).
Factoring provides more than candidate values: when the factors are fully accounted for, it proves that the list is exhaustive. By contrast, solving only one factor and stopping does not establish that the other factors have no roots in the interval.
Count Distinct Coordinates, Not Factor Repetitions
A root can occur more than once in a factorization. For an intersection list, the relevant count is the number of distinct \(x\)-coordinates, not the total number of repeated factors. A repeated root must not be counted as two different intersections just because its factor appears squared.
Worked Example: A Repeated Root Gives One Intersection Coordinate
Find every intersection of \(f(x)=x^2\) and \(g(x)=2x-1\) on \([-1,2]\).
The difference is
A square is zero only when its base is zero, so \(D(x)=0\) exactly when \(x=1\). The coordinate \(1\) belongs to \([-1,2]\). Substitution confirms the equality:
Thus the curves intersect at \((1,1)\), and there is only one distinct intersection coordinate. The squared factor does not create a second value of \(x\). Since \((x-1)^2=0\) has no other solutions, the list is complete.
When You Cannot Factor: Use Interval Evidence Carefully
Some differences do not factor conveniently. A graph can help locate possible roots, but its pixels and scale limit what it can show. Two roots close together may look like one, and a root where the graph just meets the horizontal axis may be difficult to distinguish from a near miss. A graph is useful for proposing where to investigate; it does not, by itself, prove that every root has been found.
Testing endpoint values can provide rigorous evidence when paired with a theorem and a root-count argument. If \(D\) is continuous on \([a,b]\) and \(D(a)\) and \(D(b)\) have opposite signs, the Intermediate Value Theorem guarantees at least one zero in \((a,b)\). That statement gives at least one, not exactly one. To certify a unique zero, one can also show that \(D\) is strictly monotonic on the interval. A strictly increasing or strictly decreasing function can have at most one zero there.
Worked Example: Certifying a Root with a Sign Test and Monotonicity
Verify how many intersections the curves \(f(x)=x^3+x-1\) and \(g(x)=0\) have on \([0,2]\), and locate the intersection coordinate within a short interval.
The difference is \(D(x)=x^3+x-1\), a polynomial, so it is continuous on \([0,2]\). Its endpoint values are
Because the endpoint values have opposite signs, the Intermediate Value Theorem guarantees at least one root between 0 and 2. To rule out any additional roots, differentiate:
For every real \(x\), \(3x^2\geq 0\), so \(D'(x)=3x^2+1>0\). Therefore \(D\) is strictly increasing on \([0,2]\). A strictly increasing function cannot take the value zero at two different inputs: if \(x_1<x_2\), then \(D(x_1)<D(x_2)\). Consequently there is at most one root. Together with the existence result, this proves there is exactly one intersection coordinate on \([0,2]\).
To narrow down its location, test \(x=\frac35\) and \(x=\frac7{10}\):
Continuity guarantees a root between these two test values, and strict increase guarantees that it is the only root on \([0,2]\). The intersection coordinate lies in \(\left(\frac35,\frac7{10}\right)\). The exact coordinate need not be expressible in a simple form for the completeness check to be conclusive.
What Interval Tests Do—and Do Not—Establish
Sign tests are powerful, but their conclusion must be stated precisely. Opposite signs at two endpoints guarantee at least one root between them. They do not alone prove that there is exactly one. For that, add a valid uniqueness argument, such as strict monotonicity, or use exact algebra to list all roots. Conversely, equal signs at two endpoints do not prove that there are no roots between them: a function may cross the axis twice, or touch it and return to the same side.
A practical check is to divide the proposed interval into smaller subintervals and test the difference at their endpoints. Each sign change guarantees at least one root in that subinterval, assuming continuity. But endpoint tests alone can still miss roots that do not produce a sign change. Treat this procedure as a guide unless you can also justify how many roots are possible in each subinterval.
Set \(D(x)=f(x)-g(x)\), and state the interval being checked.
Confirm that both functions are defined at any candidate coordinate and that the coordinate belongs to the stated interval.
Factor when possible. Otherwise, use sign tests with a justification that establishes existence and, where needed, uniqueness.
Substitute each candidate into \(D\), and explain why the method rules out additional zeros in the interval.
Common Pitfall: Losing a Domain Restriction
Algebraic manipulation can introduce candidates that were not in the domain of the original functions. For example, consider \(f(x)=\frac{x-1}{x-1}\) and \(g(x)=0\) on \([0,2]\), with \(f\) undefined at \(x=1\). A careless calculation might set the numerator equal to zero and report \(x=1\) as an intersection:
That conclusion is invalid because the original left side is undefined at \(x=1\). For every \(x\neq 1\) in the interval, \(f(x)=1\), which is not equal to \(g(x)=0\). Thus there are no intersections on \([0,2]\). The correction is to retain the original domain restriction throughout the algebra and check every candidate in the original equation.
Check Your Understanding
For each question, explain how you would verify whether the intersection list is complete, not merely how you would guess candidate coordinates.
- For \(f(x)=x^2-4\) and \(g(x)=0\) on \([-3,1]\), write the difference, factor it, and state every intersection coordinate in the interval.
- If \(D(a)<0\) and \(D(b)>0\), with \(D\) continuous on \([a,b]\), what does the Intermediate Value Theorem guarantee? What additional kind of argument could establish uniqueness?
- Why does the factorization \(D(x)=(x-2)^2(x+1)\) give two distinct root coordinates rather than three?
- Why can equal signs at the endpoints of an interval not, by themselves, establish that the difference function has no zero inside?
- What must you check if algebra produces a candidate intersection at a value where one of the original functions is undefined?
