Computing Vector Derivatives Componentwise

Why Compute a Vector Derivative Componentwise?

A vector-valued function can describe a moving point in space: at each parameter value, it assigns a position vector. Its derivative describes the instantaneous change in that position. Although the output is a vector, the calculation can be reduced to familiar scalar derivatives, one component at a time.

This method also gives a precise test for whether the vector derivative exists. If each component has a derivative at a parameter value, those derivatives form the vector derivative. If even one component derivative fails to exist, the vector derivative fails to exist there as well.

Definition of a Vector Derivative

Let $I$ be an interval of real numbers and let $\mathbf r:I\to\mathbb R^n$ be a vector-valued function. Write its components as scalar functions $f_1,\ldots,f_n$:

\mathbf r(t)=\langle f_1(t),f_2(t),\ldots,f_n(t)\rangle.

At an interior parameter value $t_0$, the derivative of $\mathbf r$, if it exists, is the limit of the vector difference quotient as $h$ approaches zero:

\mathbf r'(t_0) = \lim_{h\to 0} \frac{\mathbf r(t_0+h)-\mathbf r(t_0)}{h}.

The quotient is defined for nonzero $h$ small enough that $t_0+h$ lies in the domain. The limit must exist as a vector in $\mathbb R^n$. At a domain endpoint, the corresponding one-sided difference quotient is used when a one-sided derivative is under consideration.

Definition: If every component $f_i$ has a scalar derivative at $t_0$, the componentwise derivative vector is $$ \langle f_1'(t_0),f_2'(t_0),\ldots,f_n'(t_0)\rangle. $$ The componentwise result below shows that this vector is exactly $\mathbf r'(t_0)$.

The Componentwise Derivative Theorem

The key fact about vector limits is that a vector limit exists exactly when all of its component limits exist. When they do, the vector limit consists of those component limits. Apply that fact to the difference quotient defining $\mathbf r'(t_0)$.

Theorem: Let $\mathbf r(t)=\langle f_1(t),\ldots,f_n(t)\rangle$, and let $t_0$ be an interior point of its domain. The derivative $\mathbf r'(t_0)$ exists if and only if each scalar derivative $f_i'(t_0)$ exists. In that case, $$ \mathbf r'(t_0)=\langle f_1'(t_0),f_2'(t_0),\ldots,f_n'(t_0)\rangle. $$

Justification. For every nonzero $h$ for which the quotient is defined, subtract the vector values component by component and divide each component by the same scalar $h$:

\frac{\mathbf r(t_0+h)-\mathbf r(t_0)}{h} = \left\langle \frac{f_1(t_0+h)-f_1(t_0)}{h}, \ldots, \frac{f_n(t_0+h)-f_n(t_0)}{h} \right\rangle.

The $i$th component on the right is precisely the scalar difference quotient for $f_i$ at $t_0$. By the componentwise limit property, the vector quotient has a limit if and only if every one of these scalar quotients has a limit. Those scalar limits, when they exist, are $f_1'(t_0),\ldots,f_n'(t_0)$. This proves both the existence condition and the stated formula.

The theorem concerns derivatives at a particular parameter value; it does not require the components to have derivatives everywhere in the interval. To compute at $t_0$, find each scalar derivative at $t_0$, then place the results in the matching component positions.

Worked Examples

Worked Example: Differentiating Three Smooth Components

Let $\mathbf r(t)=\langle t^2+1,\,2t^3-t,\,\sin t\rangle$. Find $\mathbf r'(1)$.

First, the component functions are $f_1(t)=t^2+1$, $f_2(t)=2t^3-t$, and $f_3(t)=\sin t$. Their derivatives at $t=1$ are

f_1'(1)=2(1)=2,\qquad f_2'(1)=6(1)^2-1=5,\qquad f_3'(1)=\cos(1).

Each scalar derivative exists, so the componentwise theorem gives

\mathbf r'(1)=\langle 2,5,\cos(1)\rangle.

To check directly from the difference quotient, the first component quotient at $t=1$ is $\frac{(1+h)^2+1-2}{h}=2+h$, which tends to $2$. The second is $\frac{2(1+h)^3-(1+h)-1}{h}=5+6h+2h^2$, which tends to $5$. The third is $\frac{\sin(1+h)-\sin(1)}{h}$, which tends to $\cos(1)$ by the scalar derivative of sine. The three limits reproduce the stated vector.

Worked Example: Using Difference Quotients for Each Component

Let $\mathbf r(t)=\left\langle\sqrt{t+4},\,\ln(1+t),\,\frac{t}{t+2}\right\rangle$, defined for $t$ near $0$. Find $\mathbf r'(0)$ directly from component difference quotients.

The value at zero is $\mathbf r(0)=\langle 2,0,0\rangle$. For the first component, rationalizing the numerator gives

\lim_{h\to0}\frac{\sqrt{h+4}-2}{h} = \lim_{h\to0}\frac{(\sqrt{h+4}-2)(\sqrt{h+4}+2)}{h(\sqrt{h+4}+2)} = \lim_{h\to0}\frac{1}{\sqrt{h+4}+2} = \frac14.

For the second component, the standard scalar logarithm limit gives

\lim_{h\to0}\frac{\ln(1+h)-0}{h}=1.

For the third component, simplify the quotient before taking the limit:

\lim_{h\to0} \frac{\frac{h}{h+2}-0}{h} = \lim_{h\to0}\frac{1}{h+2} = \frac12.

All three component limits exist. Therefore,

\mathbf r'(0)=\left\langle\frac14,1,\frac12\right\rangle.

As a check, multiplying the proposed derivative by the componentwise difference quotient interpretation gives exactly the three limits just computed: $1/4$, $1$, and $1/2$. Thus no component has been omitted or placed in the wrong position.

Worked Example: One Component Prevents Differentiability

Let $\mathbf r(t)=\langle t^2,\,|t|,\,\sin t\rangle$. Determine whether $\mathbf r'(0)$ exists.

The vector value at zero is $\mathbf r(0)=\langle 0,0,0\rangle$. The difference quotient is

\frac{\mathbf r(h)-\mathbf r(0)}{h} = \left\langle h,\frac{|h|}{h},\frac{\sin h}{h}\right\rangle.

As $h\to0$, the first component tends to $0$, and the third tends to $1$. The middle component does not have a two-sided limit: it equals $1$ when $h>0$ and $-1$ when $h<0$. Hence the vector difference quotient has no limit, and $\mathbf r'(0)$ does not exist.

The componentwise test gives the same conclusion: the scalar function $|t|$ has no derivative at zero, so not all component derivatives exist. The existence of derivatives for the first and third components cannot compensate for that failure.

A Reliable Calculation Process

For a derivative at a specified parameter, the componentwise method can be organized into a short sequence. The important check is to establish existence for every component, rather than assuming that a vector derivative exists because some components behave well.

Identify the components.
Write the vector function as a list of scalar component functions in their original order.

Differentiate each component at the parameter.
Use scalar differentiation or, when needed, evaluate the scalar difference quotient limit.

Check existence component by component.
If any scalar derivative does not exist, the vector derivative does not exist.

Assemble and verify.
Place the scalar derivative values in the corresponding component positions and compare with the limits of the vector difference quotient.

Common Mistake: Using a Component’s Value Instead of Its Derivative

A component’s value at $t_0$ is not its derivative there. For instance, in the example with $|t|$, the value $|0|=0$ does not imply that the derivative of $|t|$ at zero is $0$. The derivative depends on how the function changes on both sides of the point.

A mistaken calculation might replace the middle component quotient by zero just because the component has value zero at the point:

\frac{|h|-|0|}{h} \ \text{is incorrectly treated as}\ \frac{0-0}{h}=0.

This substitutes the value at $h=0$ for the value at a nearby parameter $h$, even though the difference quotient requires $h\ne0$. The correct calculation keeps the nearby value:

\frac{|h|-|0|}{h} = \frac{|h|}{h} = \begin{cases} 1,&h>0,\\ -1,&h<0. \end{cases}

The one-sided limits disagree, so the scalar derivative, and therefore the vector derivative, does not exist at zero. Always form the quotient using the actual function values before taking the limit.

Why the Component Test Matters

The theorem turns a vector differentiation problem into a finite list of scalar problems without changing the definition. It is useful both for finding derivatives and for diagnosing failure: a single component with a missing derivative is enough to rule out a vector derivative. Conversely, when all component derivatives exist, their ordered list gives the vector derivative immediately.

This does not mean that the derivative is merely a collection of unrelated calculations. The components are the coordinates of one vector difference quotient, and their limits must occur together to produce its vector limit. The componentwise theorem is what makes these scalar calculations equivalent to the original vector limit.

Check Your Understanding

Use the componentwise definition and theorem to answer the following questions.

If $\mathbf r(t)=\langle f(t),g(t)\rangle$, what must be true of $f$ and $g$ at $t_0$ for $\mathbf r'(t_0)$ to exist?
For $\mathbf r(t)=\langle t^3,\cos t\rangle$, find $\mathbf r'(0)$.
Why does the failure of one component derivative imply failure of the vector derivative?
For the component $|t|$ at zero, what are the right-hand and left-hand limits of its difference quotient?
In the componentwise derivative formula, why must the derivative values remain in the same order as the original components?