Tutorials › Calculus III › Reversing Orientation Changes the Sign

Work and Vector Line Integrals · Tutorial 756 of 999

Reversing Orientation Changes the Sign

Learn how reversing a curve changes its tangent direction and why the work integral along the reversed path is the negative of the forward work.

Intermediate 9 min read

What You'll Learn

  • Define the reverse of a parametrized curve on the same interval
  • Derive the sign change using the chain rule and a change of variable
  • Calculate forward and reverse work along a parabolic path
  • Compare work on a quarter-circle path in opposite directions
  • Track orientation explicitly when computing work in three dimensions
  • Identify and correct the error of keeping the original tangent direction

Introduction

A force does work as an object moves through space. The force at a point is the same whether the object is moving forward or backward through that point, but its direction of motion is not. Since work depends on the component of force in the direction of motion, reversing the direction of travel changes the sign of the work.

This change is a precise property of the vector line integral. It does not depend on the particular shape of the curve or force field, provided the curve and integral satisfy the conditions stated below. We will derive the sign change and then calculate work in both directions on several paths.

Curves, Force Fields, and Work

Let \(c:[a,b]\to\mathbb{R}^n\) be a continuously differentiable curve, and let \(\mathbf{F}\) be a continuous vector field defined along the curve. The vector \(c'(t)\) is the velocity of the parametrized motion at parameter value \(t\). The work of \(\mathbf{F}\) along \(c\) is the vector line integral formed by taking the dot product of the force and velocity and integrating over the parameter interval.

Definition: The work of \(\mathbf{F}\) along the oriented curve \(c\) is $$ W_c=\int_a^b \mathbf{F}(c(t))\cdot c'(t)\,dt. $$

The orientation is the direction in which the parameter traverses the curve, from \(c(a)\) to \(c(b)\). Reversing orientation means traversing the same points in the opposite order, from \(c(b)\) back to \(c(a)\). The force field is not changed; what changes is the velocity vector at each corresponding point.

To describe the reverse traversal on the same interval, define \(c_{\mathrm{rev}}:[a,b]\to\mathbb{R}^n\) by \(c_{\mathrm{rev}}(s)=c(a+b-s)\). At \(s=a\), this gives \(c(b)\); at \(s=b\), it gives \(c(a)\). Thus the reverse curve starts where \(c\) ended and ends where \(c\) began.

Definition: For a curve \(c:[a,b]\to\mathbb{R}^n\), its reverse on the same parameter interval is $$ c_{\mathrm{rev}}(s)=c(a+b-s),\qquad a\leq s\leq b. $$

Why Reversing the Curve Negates Work

The key observation follows from the chain rule. Differentiating the reverse curve gives \(c_{\mathrm{rev}}'(s)=-c'(a+b-s)\). The minus sign records that the direction of motion has been reversed. The force at the corresponding point remains \(\mathbf{F}(c(a+b-s))\).

Theorem: If \(c:[a,b]\to\mathbb{R}^n\) is continuously differentiable and \(\mathbf{F}\) is continuous along its image, then $$ \int_a^b \mathbf{F}(c_{\mathrm{rev}}(s))\cdot c_{\mathrm{rev}}'(s)\,ds = -\int_a^b \mathbf{F}(c(t))\cdot c'(t)\,dt. $$

Here is the complete derivation. Set \(g(t)=\mathbf{F}(c(t))\cdot c'(t)\). This is a continuous scalar function on \([a,b]\), so its integrals below are defined. Apply the chain rule to the reversed curve, then substitute \(u=a+b-s\) in the integral.

1
Differentiate the reversed parametrization.
Since \(c_{\mathrm{rev}}(s)=c(a+b-s)\), the chain rule gives $$ c_{\mathrm{rev}}'(s)=-c'(a+b-s). $$
2
Write the reverse work using that derivative.
The force is evaluated at the point \(c(a+b-s)\), so $$ W_{c_{\mathrm{rev}}} =-\int_a^b \mathbf{F}(c(a+b-s))\cdot c'(a+b-s)\,ds =-\int_a^b g(a+b-s)\,ds. $$
3
Change the integration variable.
Set \(u=a+b-s\), so \(ds=-du\). When \(s=a\), \(u=b\); when \(s=b\), \(u=a\). Therefore $$ \int_a^b g(a+b-s)\,ds =\int_b^a g(u)(-du) =\int_a^b g(u)\,du. $$ It follows that \(W_{c_{\mathrm{rev}}}=-\int_a^b g(u)\,du=-W_c\).

This proof identifies exactly where the sign comes from: the reverse velocity is the negative of the forward velocity at the corresponding point. The change of variable changes the order of the endpoints as well, and the two minus signs from that substitution cancel each other. The minus sign from the reversed velocity remains.

Worked Example: A Parabolic Path

Worked Example: A Parabolic Path

Let \(c(t)=(t,t^2)\), \(0\leq t\leq 1\), and let \(\mathbf{F}(x,y)=(x+y,x-y)\). Find the work in the forward and reverse directions.

For the forward path, $$ c'(t)=(1,2t),\qquad \mathbf{F}(c(t))=(t+t^2,t-t^2). $$ Thus the work integrand is $$ \mathbf{F}(c(t))\cdot c'(t) =(t+t^2)(1)+(t-t^2)(2t) =t+t^2+2t^2-2t^3 =t+3t^2-2t^3. $$ Integrating gives $$ W_c=\int_0^1(t+3t^2-2t^3)\,dt =\left[\frac{t^2}{2}+t^3-\frac{t^4}{2}\right]_0^1 =\frac12+1-\frac12=1. $$ Differentiating the displayed antiderivative gives \(t+3t^2-2t^3\), so the integration is verified.

The reverse path is $$ c_{\mathrm{rev}}(s)=(1-s,(1-s)^2),\qquad 0\leq s\leq1, $$ with velocity \(c_{\mathrm{rev}}'(s)=(-1,-2(1-s))\). Put \(u=1-s\). Then the force at this point is \((u+u^2,u-u^2)\), and its dot product with the reverse velocity is $$ -(u+u^2)-2u(u-u^2)=-u-3u^2+2u^3. $$ Substituting \(u=1-s\) and expanding gives the reverse integrand $$ -(1-s)-3(1-s)^2+2(1-s)^3=-2+s+3s^2-2s^3. $$ Consequently, $$ W_{c_{\mathrm{rev}}} =\int_0^1(-2+s+3s^2-2s^3)\,ds =\left[-2s+\frac{s^2}{2}+s^3-\frac{s^4}{2}\right]_0^1 =-2+\frac12+1-\frac12=-1. $$ The derivative of the antiderivative is the displayed integrand. The result also checks against the theorem: the reverse work is \(-1\), the negative of the forward work \(1\).

Worked Example: A Quarter-Circle Path

Worked Example: A Quarter-Circle Path

Consider \(c(t)=(\cos t,\sin t)\), \(0\leq t\leq\frac{\pi}{2}\), in the force field \(\mathbf{F}(x,y)=(-y,x)\). Calculate work in each direction.

On the forward path, $$ c'(t)=(-\sin t,\cos t),\qquad \mathbf{F}(c(t))=(-\sin t,\cos t). $$ Their dot product is \(\sin^2 t+\cos^2 t=1\). Therefore $$ W_c=\int_0^{\pi/2}1\,dt=\frac{\pi}{2}. $$ The integrand is constant \(1\), so this evaluation is direct.

The reverse curve on the same interval is $$ c_{\mathrm{rev}}(s)=(\sin s,\cos s), \qquad c_{\mathrm{rev}}'(s)=(\cos s,-\sin s). $$ The force at the reverse point is \(\mathbf{F}(c_{\mathrm{rev}}(s))=(-\cos s,\sin s)\). Hence $$ \mathbf{F}(c_{\mathrm{rev}}(s))\cdot c_{\mathrm{rev}}'(s) =-\cos^2 s-\sin^2 s=-1, $$ and $$ W_{c_{\mathrm{rev}}}=\int_0^{\pi/2}(-1)\,ds=-\frac{\pi}{2}. $$ Both dot products use the appropriate velocity: \(c'(t)\) forward and \(c_{\mathrm{rev}}'(s)\) in reverse. The results agree with the sign-change theorem.

Worked Example: Work in Three Dimensions

Worked Example: Work in Three Dimensions

Let \(c(t)=(t,t^2,t)\), \(0\leq t\leq1\), and take the force field \(\mathbf{F}(x,y,z)=(0,0,x)\). Find the work for both orientations.

The forward velocity and force along the curve are $$ c'(t)=(1,2t,1),\qquad \mathbf{F}(c(t))=(0,0,t). $$ Their dot product is \(t\), so $$ W_c=\int_0^1t\,dt=\left[\frac{t^2}{2}\right]_0^1=\frac12. $$ The derivative of \(t^2/2\) is \(t\), confirming the integral.

The reverse curve and its velocity are $$ c_{\mathrm{rev}}(s)=(1-s,(1-s)^2,1-s),\qquad c_{\mathrm{rev}}'(s)=(-1,-2(1-s),-1). $$ At that point the force is \((0,0,1-s)\). The dot product is $$ (0,0,1-s)\cdot(-1,-2(1-s),-1)=-(1-s)=-1+s. $$ Thus $$ W_{c_{\mathrm{rev}}}=\int_0^1(-1+s)\,ds =\left[-s+\frac{s^2}{2}\right]_0^1 =-1+\frac12=-\frac12. $$ The antiderivative differentiates to \(-1+s\), and the result is the negative of the forward work.

A Common Sign Error

A frequent mistake is to reverse the order of the endpoints mentally but then keep the forward tangent vector in the work integrand. In the quarter-circle example, this mistake would use the forward dot product \(1\) while integrating over the reverse parameter interval, and incorrectly report $$ \int_0^{\pi/2}1\,ds=\frac{\pi}{2} $$ for the reverse work. The reverse velocity is not the forward velocity: it is \(c_{\mathrm{rev}}'(s)=(\cos s,-\sin s)\), not \((-\sin s,\cos s)\). The correct reverse dot product is \(-1\), giving work \(-\frac{\pi}{2}\).

When computing reverse work directly, write down the reverse curve and differentiate it before forming the dot product. Do not change the sign of the final answer by intuition alone while leaving the integrand unchanged; the reversed velocity is the source of the sign change and should appear explicitly in the calculation.

Why Orientation Matters

Work measures the accumulation of force in the direction of motion. Along a forward traversal, a force component aligned with the tangent contributes positively; along the same path in reverse, the tangent points oppositely, so that same force component contributes negatively. A zero work value remains zero after reversal, since its negative is still zero.

This sign rule applies to the vector line integral \(\int \mathbf{F}\cdot d\mathbf{r}\), whose integrand includes the directed tangent. It is not a rule that every line integral changes sign on reversal. For example, a scalar line integral with respect to arc length uses the nonnegative speed rather than a directed tangent, so reversing direction does not introduce this tangent sign. The important question is which quantity is being integrated and whether direction enters its definition.

Key takeaway: For a vector line integral of a continuous force field along a continuously differentiable curve, reversing the direction of traversal preserves the geometric path but negates the work: $$ W_{c_{\mathrm{rev}}}=-W_c. $$

Check Your Understanding

Use the definitions and sign rule above to answer these questions.

  1. For \(c:[a,b]\to\mathbb{R}^n\), write the formula for its reverse on the same interval and state its starting and ending points.
  2. Differentiate \(c_{\mathrm{rev}}(s)=c(a+b-s)\). What does the resulting sign indicate geometrically?
  3. If work along a curve is \(7\) joules, what is the work along the reversed curve under the same force field?
  4. For the quarter-circle example, which velocity vector must be used to compute reverse work?
  5. Why does reversing orientation negate a vector line integral but not necessarily a scalar line integral with respect to arc length?