Why Sign Charts Need Care
A second-derivative sign chart is a compact way to record where a function is concave up or concave down. Its simplicity can be misleading: a chart is reliable only when its intervals respect the function’s domain and its test points actually lie in those intervals. Even a correctly computed sign chart identifies possible inflection points, not inflection points automatically.
Three questions prevent many common errors: Where is the function defined? Which intervals have a constant sign of \(f''\)? And, at a proposed inflection point, does the function really change concavity there?
Concavity, Candidates, and Inflection Points
Suppose \(f\) is twice differentiable on an interval. Where \(f''(x)>0\), the graph of \(f\) is concave up; where \(f''(x)<0\), it is concave down. A sign chart records these signs interval by interval. Points where \(f''\) is zero or undefined can mark boundaries between intervals, but they do not by themselves settle the concavity on either side.
A zero of \(f''\), or a point where \(f''\) does not exist, is therefore a candidate for an inflection point. It must be checked. A point where the function itself is undefined cannot be a point on its graph, even if the signs of \(f''\) differ on the two sides.
For a sign chart, divide the domain at the relevant zeros and undefined points of \(f''\), as well as at breaks in the domain of \(f\). Then test the sign on each resulting interval. A domain break matters even if it is not a zero or undefined point of \(f''\) considered in isolation: no sign test can bridge a point where the function is not defined.
Why One Test Point Works on Each Interval
The sign-chart method rests on a basic continuity fact. If \(f''\) is continuous on an interval and never zero there, it cannot switch from positive to negative (or vice versa) inside that interval.
To justify the theorem, suppose instead that \(f''\) had opposite signs at two points \(a,b\in I\). Continuity and the Intermediate Value Theorem would then imply that \(f''(c)=0\) for some point \(c\) between \(a\) and \(b\). That contradicts the assumption that \(f''\) is nonzero throughout \(I\). Therefore the sign cannot change on \(I\).
This reasoning explains why finding every relevant boundary is essential. Within each interval, the test-point argument applies only when \(f''\) is continuous there and has no zeros there. A missed zero or domain break can leave an interval that should have been split, making a single test point inadequate for the region as labelled.
Worked Example: A Domain Break Is Not an Inflection Point
Worked Example: A Domain Break Is Not an Inflection Point
Consider \(f(x)=\frac{x+1}{x-2}\). Find its concavity intervals and decide whether \(x=2\) is an inflection point.
First record the domain. The denominator cannot be zero, so \(x\ne 2\). Rewrite the function to differentiate:
Differentiate once and then again:
The domain break at \(x=2\) divides the domain into \((-\infty,2)\) and \((2,\infty)\). For a test point \(x=0\) in the first interval,
For a test point \(x=3\) in the second interval,
Thus \(f\) is concave down on \((-\infty,2)\) and concave up on \((2,\infty)\). The signs differ across \(2\), but \(f(2)\) is not defined. There is no point of the graph at \(x=2\), so \(x=2\) is not an inflection point.
Check the boundary directly: the original denominator is \(2-2=0\), confirming that the function has no value there. The sign change describes the concavity on the two separate domain intervals; it does not make the missing point an inflection point.
Worked Example: Choose Test Points in the Intervals You Label
Worked Example: Choose Test Points in the Intervals You Label
Let \(f(x)=\frac{x^4}{12}+\frac{x^3}{6}-x^2\). Find the concavity intervals and inflection points. The derivatives are
The function is defined for all real \(x\). The zeros of \(f''\) are \(-2\) and \(1\), so the intervals to test are \((-\infty,-2)\), \((-2,1)\), and \((1,\infty)\).
Choose one point from each interval. For \(x=-3\), which is in \((-\infty,-2)\),
For \(x=0\), which is in \((-2,1)\),
For \(x=2\), which is in \((1,\infty)\),
The function is concave up on \((-\infty,-2)\) and \((1,\infty)\), and concave down on \((-2,1)\). Concavity changes at both \(-2\) and \(1\). Since \(f\) is a polynomial, it is continuous at both points, so both are inflection points.
Compute their coordinates to identify the points on the graph:
The inflection points are \((-2,-4)\) and \((1,-\frac{3}{4})\). As a check, substituting the \(x\)-coordinates into \(f''(x)=(x+2)(x-1)\) gives \(f''(-2)=0\) and \(f''(1)=0\), and the interval tests verify the required concavity changes.
A common mistake is to label \((-2,1)\) as the first test interval and use \(x=-3\) to test it. That value is not in the interval: \(-3<-2\). Its result, \(f''(-3)=4>0\), tells us about \((-\infty,-2)\), not \((-2,1)\). Correct the mistake by writing the ordered boundaries first, then choosing a point strictly inside each stated interval.
Worked Example: A Zero of the Second Derivative May Not Be an Inflection Point
Worked Example: A Zero of the Second Derivative May Not Be an Inflection Point
Consider \(f(x)=x^4\). Its derivatives are
The second derivative is zero at \(x=0\). To determine what happens on either side, test \(x=-1\) and \(x=1\):
The function is concave up on both sides of \(0\). Its concavity does not change there, so \((0,0)\) is not an inflection point, even though \(f''(0)=0\). The point is on the graph because \(f(0)=0\), but being on the graph and having \(f''=0\) are not enough.
For contrast, \(g(x)=x^3\) has \(g''(x)=6x\), which is negative for \(x<0\) and positive for \(x>0\). The function is continuous at \(0\), so its concavity changes there and \((0,0)\) is an inflection point. Substitution confirms \(g(0)=0\) and \(g''(0)=0\); the signs on both sides establish the change.
Worked Example: An Undefined Second Derivative Can Still Mark an Inflection Point
Worked Example: An Undefined Second Derivative Can Still Mark an Inflection Point
Consider \(h(x)=x^{1/3}\), defined for every real \(x\). For \(x\ne 0\),
The second derivative is undefined at \(0\), so \(0\) is a candidate boundary. If \(x<0\), then \(x^{5/3}<0\), and therefore \(-\frac{2}{9}x^{-5/3}>0\). If \(x>0\), then \(x^{5/3}>0\), and therefore \(-\frac{2}{9}x^{-5/3}<0\). Thus \(h\) is concave up to the left of \(0\) and concave down to the right.
The function is continuous at \(0\), with \(h(0)=0\). Its concavity changes there, so \((0,0)\) is an inflection point, even though \(h''(0)\) does not exist. The check is the value \(h(0)=0\), together with the opposite signs of \(h''\) on the two sides.
A Reliable Sign-Chart Routine
Use this routine when a concavity problem involves zeros, undefined values, or domain breaks. Each step guards against a different error.
Mark every point where \(f\) is undefined. Such a point splits the domain, regardless of whether a formula for \(f''\) suggests a sign change there.
Record where \(f''=0\) and where \(f''\) is undefined, while retaining the domain restrictions inherited from \(f\).
Use the domain breaks and relevant zeros or undefined points to partition the domain into open intervals. Do not test across a listed boundary.
Choose a point inside that interval, evaluate \(f''\), and record only its sign. If the assumptions for the sign-chart theorem do not hold throughout an interval, split it further.
Confirm that \(f\) is defined and continuous at the point and that concavity changes on its two sides. A zero or undefined value of \(f''\) alone is insufficient.
Common Errors and Their Corrections
| Error | Why it fails | Correction |
|---|---|---|
| Omitting a domain break | A sign test may then span two separate parts of the domain. | Write the domain before constructing the intervals. |
| Using a test point outside its labelled interval | The computed sign applies to the test point’s actual interval, not the one intended. | Check the inequalities defining the interval before evaluating \(f''\). |
| Calling every zero of \(f''\) an inflection point | The second derivative may have the same sign on both sides. | Compare concavity immediately to the left and right. |
| Calling every undefined value of \(f''\) an inflection point | The function may be discontinuous or undefined there; alternatively, its concavity may not change. | Check that the function is defined and continuous, then verify a change in concavity. |
Why These Checks Matter
Concavity describes the shape of a graph over intervals, while an inflection point is a specific point on the graph where that shape changes. Confusing these claims can lead to incorrect descriptions even when the derivatives are calculated correctly. In particular, a sign change across a missing point describes two separate intervals; it does not join them into one curve or supply the missing point.
The key habit is to treat a sign chart as evidence that must be interpreted, not as an automatic list of answers. Establish the intervals, test each interval correctly, and then check each proposed inflection point against the function and its concavity on both sides.
Check Your Understanding
For each question, explain the reasoning as well as the conclusion.
- If \(f''(c)=0\), what additional evidence is needed before \(c\) can be identified as an inflection point?
- Why can a sign change in \(f''\) across a point where \(f\) is undefined not establish an inflection point there?
- When may one test point determine the sign of \(f''\) throughout an interval?
- For \(f(x)=x^4\), why is \(x=0\) not an inflection point even though \(f''(0)=0\)?
- For \(h(x)=x^{1/3}\), why can \(x=0\) be an inflection point even though \(h''(0)\) is undefined?
