Polynomial Derivatives at an Arbitrary Point

From One Power to a Whole Polynomial

A polynomial combines a finite number of powers of its variable, each multiplied by a coefficient. To find its derivative at a point, we can start from the derivative definition and handle every power inside the difference quotient. The key is to expand before taking the limit: the expansion separates the part that remains from the parts that disappear.

This approach builds on the binomial theorem and the integer-power derivative from the earlier material. Here the goal is to assemble the calculation for a general polynomial at an arbitrary base point, showing explicitly how its finitely many terms contribute.

Definition: The derivative of a function $p$ at a real number $a$, when the limit exists, is $$ p'(a)=\lim_{h\to 0}\frac{p(a+h)-p(a)}{h}. $$ The quotient is considered for $h\ne 0$; the limit describes its behavior as $h$ approaches zero.

Expanding the Difference Quotient

Let $p$ be a polynomial with real coefficients, written as $p(x)=c_0+c_1x+c_2x^2+\cdots+c_nx^n$. Here $n$ is a nonnegative integer, and any of the coefficients may be zero. The point $a$ is an arbitrary real number. Substituting $a+h$ and $a$ into the polynomial gives a difference quotient in which the constant term cancels.

\frac{p(a+h)-p(a)}{h} = \sum_{k=1}^{n}c_k\frac{(a+h)^k-a^k}{h}, \qquad h\ne 0.

For each positive integer $k$, the binomial theorem expands $(a+h)^k$. The term $a^k$ is the term with no factor of $h$, so it cancels when $a^k$ is subtracted. Every remaining term contains at least one factor of $h$, which can then be divided out.

(a+h)^k-a^k = \sum_{j=1}^{k}\binom{k}{j}a^{k-j}h^j,

and therefore

\frac{(a+h)^k-a^k}{h} = \sum_{j=1}^{k}\binom{k}{j}a^{k-j}h^{j-1}.

The $j=1$ term is $ka^{k-1}$, because $\binom{k}{1}=k$ and $h^{1-1}=1$. Every term with $j\ge 2$ contains a positive power of $h$, so it approaches zero as $h\to 0$. This also works when $a=0$: the expression $a^0$ means $1$, including when $a=0$. For a constant term, $k=0$, the difference is simply $c_0-c_0=0$, so it contributes nothing.

The Polynomial Derivative Formula

Theorem: If $p(x)=\sum_{k=0}^{n}c_kx^k$ is a real polynomial, then $p$ is differentiable at every real number $a$, and $$ p'(a)=\sum_{k=1}^{n}k c_k a^{k-1}. $$ A constant polynomial has derivative zero; equivalently, the sum is empty when $n=0$.

The formula is a consequence of a finite calculation, not an assumption that we can differentiate a sum by a rule already established elsewhere. In the difference quotient, each power produces a finite sum of terms. Since there are only finitely many powers in the polynomial, there are only finitely many terms in the resulting expression. We can take the limit of those terms individually using the ordinary limit laws for finite sums.

Substitute into the definition.
Write $\bigl(p(a+h)-p(a)\bigr)/h$, with $h\ne 0$.

Expand the powers.
Apply the binomial theorem to each $(a+h)^k$, and cancel its $a^k$ term against the subtraction.

Divide by the increment.
Each remaining term has a factor of $h$, so division lowers its exponent by one.

Take the limit.
The terms with a positive power of $h$ vanish; the terms independent of $h$ give the stated formula.

Justification from the Finite Expansion

For completeness, put the expansions for all powers together. After substituting the polynomial into its difference quotient, expand each power and divide by $h$:

\frac{p(a+h)-p(a)}{h} = \sum_{k=1}^{n}c_k \sum_{j=1}^{k}\binom{k}{j}a^{k-j}h^{j-1}.

Separate the $j=1$ terms from those with $j\ge 2$. The first group does not depend on $h$. Each term in the second group has $h^{j-1}$, whose exponent is at least one:

\frac{p(a+h)-p(a)}{h} = \sum_{k=1}^{n}k c_k a^{k-1} + \sum_{k=1}^{n}c_k \sum_{j=2}^{k}\binom{k}{j}a^{k-j}h^{j-1}.

Every sum here has finitely many terms. For each term in the second group, all factors other than $h^{j-1}$ are fixed real numbers, and $h^{j-1}\to 0$ as $h\to 0$. Thus the entire second group tends to zero. The first group is fixed as $h$ varies. The limit of the difference quotient therefore exists and equals that first group, proving the theorem. If $n=1$, the second group has no terms and is zero; the same reasoning still applies.

Worked Examples

Worked Example: A Cubic Polynomial at a General Point

Find the derivative of $p(x)=2x^3-x^2+4x-5$ at a general real point $a$, then evaluate it at $a=1$ using the difference quotient.

First substitute $a+h$ and $a$. The constant $-5$ cancels:

\frac{p(a+h)-p(a)}{h} = \frac{2\bigl((a+h)^3-a^3\bigr)-\bigl((a+h)^2-a^2\bigr)+4h}{h}.

Expand the cube and square, then divide every term by $h$:

\begin{aligned} \frac{p(a+h)-p(a)}{h} &=\frac{2(3a^2h+3ah^2+h^3)-(2ah+h^2)+4h}{h}\\ &=6a^2+6ah+2h^2-2a-h+4. \end{aligned}

As $h\to 0$, the terms $6ah$, $2h^2$, and $-h$ approach zero. Hence $p'(a)=6a^2-2a+4$. At $a=1$, this gives $6-2+4=8$. To check directly in the expanded quotient, set $a=1$: it becomes $8+5h+2h^2$, whose limit is $8$, as required.

Worked Example: A Quartic Polynomial at a Negative Point

Find the derivative of $q(x)=3x^4+2x^3-x+6$ at $a=-2$, starting with its difference quotient.

Use the binomial expansion term by term. The constant cancels, and the linear term contributes $-1$ after division by $h$:

\begin{aligned} \frac{q(-2+h)-q(-2)}{h} &=3\frac{(-2+h)^4-(-2)^4}{h} +2\frac{(-2+h)^3-(-2)^3}{h}-1\\ &=3(-32+24h-8h^2+h^3) +2(12-6h+h^2)-1\\ &=-96+72h-24h^2+3h^3+24-12h+2h^2-1\\ &=-73+60h-22h^2+3h^3. \end{aligned}

Taking $h\to 0$ leaves $-73$. Check using the formula: $q'(a)=12a^3+6a^2-1$. At $a=-2$, this is $12(-8)+6(4)-1=-96+24-1=-73$. The result agrees with the limit calculation.

Worked Example: Evaluating at Zero

Find the derivative of $r(x)=x^5-2x^3+x$ at $a=0$. This example shows why the terms with powers of $a$ must be interpreted carefully at zero.

Since $r(0)=0$, substitution into the defining quotient gives, for $h\ne 0$,

\begin{aligned} \frac{r(0+h)-r(0)}{h} &=\frac{h^5-2h^3+h}{h}\\ &=h^4-2h^2+1. \end{aligned}

As $h\to 0$, $h^4\to 0$ and $h^2\to 0$, so the limit is $1$. The polynomial formula gives $r'(a)=5a^4-6a^2+1$, and at zero it gives $0-0+1=1$. Thus the direct quotient and the formula agree, including at a point where several powers of $a$ vanish.

A Common Mistake: Substituting Zero Too Early

The difference quotient is defined for $h\ne 0$. Its numerator often approaches zero as $h\to 0$, but that does not mean the quotient approaches zero. Dividing by $h$ can leave a nonzero term.

For example, for $f(x)=x^2$ at $a=3$, an incorrect calculation might say that the numerator $(3+h)^2-9$ approaches zero, so the derivative is zero. This overlooks the division by $h$. The correct calculation is

\frac{(3+h)^2-9}{h} = \frac{9+6h+h^2-9}{h} = \frac{6h+h^2}{h} = 6+h, \qquad h\ne 0.

Now the limit is $6$, not zero. The correction is to simplify the quotient while $h\ne 0$, and only then take the limit. In a polynomial expansion, this means canceling the common factor of $h$ before deciding which terms vanish.

Why the Finite-Sum View Matters

The method explains exactly how the value of a polynomial's derivative at a point is assembled: each power contributes its constant-in-$h$ term after expansion, while all higher powers of $h$ disappear in the limit. The finiteness of a polynomial is essential to this argument. There is no question here of controlling infinitely many terms or exchanging an infinite sum with a limit.

The result is useful in two ways. It gives a direct first-principles calculation when a polynomial and a point are specified, and it establishes the general formula at every real point in one argument. Once the formula is known, a numerical derivative can be checked both by evaluating the formula and by simplifying the difference quotient. Keeping those two viewpoints connected helps identify algebra errors, especially sign errors and missing powers of the increment.

Takeaway: For a polynomial $p(x)=\sum_{k=0}^{n}c_kx^k$, expand the difference quotient at $a$. After dividing by $h$, the terms that still contain a positive power of $h$ vanish in the limit, leaving $p'(a)=\sum_{k=1}^{n}k c_k a^{k-1}$.

Check Your Understanding

Use the difference quotient and the finite expansion to answer these questions.

For $p(x)=x^3-4x+2$, write $\bigl(p(a+h)-p(a)\bigr)/h$ in simplified form before taking the limit.
What term in the binomial expansion of $(a+h)^k$ produces the contribution $ka^{k-1}$ after division by $h$?
Why does the constant term of a polynomial contribute zero to its difference quotient?
For $q(x)=2x^2+3x$, use the general formula to find $q'(-1)$, and state the value of each term in the formula.
Why is it valid to take the limit of the expanded polynomial quotient one term at a time?

Polynomial Derivatives at an Arbitrary Point

What You'll Learn

From One Power to a Whole Polynomial

Expanding the Difference Quotient

The Polynomial Derivative Formula

Justification from the Finite Expansion

Worked Examples

Worked Example: A Cubic Polynomial at a General Point

Worked Example: A Quartic Polynomial at a Negative Point

Worked Example: Evaluating at Zero

A Common Mistake: Substituting Zero Too Early

Why the Finite-Sum View Matters

Check Your Understanding

Test yourself: Calculus I final exam practice

Want all 1,114 practice problems?

Get the Calc Pass App

Vol. 1: 1,114 Questions

Vol. 2: Solutions Manual

Polynomial Derivatives at an Arbitrary Point

What You'll Learn

From One Power to a Whole Polynomial

Expanding the Difference Quotient

The Polynomial Derivative Formula

Justification from the Finite Expansion

Worked Examples

Worked Example: A Cubic Polynomial at a General Point

Worked Example: A Quartic Polynomial at a Negative Point

Worked Example: Evaluating at Zero

A Common Mistake: Substituting Zero Too Early

Why the Finite-Sum View Matters

Check Your Understanding

Test yourself: Calculus I final exam practice

Want all 1,114 practice problems?

Get the Calc Pass App

Vol. 1: 1,114 Questions

Vol. 2: Solutions Manual

Previous

Next