Vector-Valued Function Derivatives — Derivatives of Vector-Valued Functions (Velocity and Acceleration) - Calculus 3

SECTION 1

Historical Context & Motivation

The study of motion along curves in space is one of the oldest motivating problems of mathematics and physics, dating back to the planetary trajectories that captivated astronomers for millennia. When Newton and Leibniz independently developed calculus in the late seventeenth century, their tools were primarily scalar — functions of one variable yielding a single output. Yet the physical world rarely confines motion to a single dimension: planets sweep out ellipses, projectiles trace parabolas, and charged particles spiral through magnetic fields. The need to describe velocity and acceleration as quantities possessing both magnitude and direction pushed mathematicians toward a calculus of vectors, ultimately giving rise to the derivative of a vector-valued function.

1687 • Newton's Principia

Isaac Newton formulated the laws of motion using geometric arguments about directed quantities — effectively proto-vectors — to describe planetary orbits and projectile trajectories, laying the conceptual groundwork for vector calculus.

1843 • Hamilton's Quaternions

William Rowan Hamilton introduced quaternions as an algebraic system for three-dimensional rotations, inspiring later formalization of vector algebra and providing a language in which differentiation of multi-component quantities could be rigorously defined.

1881 • Gibbs & Heaviside Formalize Vectors

Josiah Willard Gibbs and Oliver Heaviside independently extracted the modern vector notation from Hamilton's quaternions, establishing the component-wise framework r(t) = ⟨x(t), y(t), z(t)⟩ that made differentiation of vector-valued functions straightforward.

1901–1910 • Frenet–Serret Apparatus

Building on earlier work by Frenet and Serret, the moving-frame formalism for space curves matured, connecting vector derivatives to curvature and torsion and establishing the geometric interpretation of velocity and acceleration that is standard today.

The central question this lesson addresses is deceptively simple: if a particle's position in space at time t is given by a vector-valued function r(t), how do we rigorously define and compute its instantaneous velocity and acceleration? Answering this question requires extending the limit definition of the derivative to vector outputs, and the results unify the kinematics you studied in single-variable calculus with the richer geometry of three-dimensional space.

SECTION 2

Core Principles & Definitions

At its core, differentiating a vector-valued function proceeds component by component. A vector-valued function maps a scalar parameter (often time) to a vector in ℝ² or ℝ³. Its derivative, when it exists, is itself a vector whose components are the ordinary derivatives of the component functions. This simple principle carries deep geometric and physical meaning: the derivative vector is tangent to the curve, its magnitude encodes speed, and the second derivative encodes how the velocity is changing — the acceleration.

Vector-Valued Function

A function r(t) = ⟨f(t), g(t), h(t)⟩ whose output is a vector. The parameter t traces a curve C in space as it varies over an interval.

Velocity Vector

The first derivative v(t) = r′(t) is tangent to the curve at r(t). Its magnitude ‖v(t)‖ equals the speed.

Acceleration Vector

The second derivative a(t) = r″(t) measures how the velocity vector itself changes, capturing both speeding up/slowing down and turning.

Component-wise Differentiation

If r(t) = ⟨f(t), g(t), h(t)⟩, then r′(t) = ⟨f ′(t), g′(t), h′(t)⟩. Each component is differentiated independently using standard single-variable rules.

Speed (Scalar)

Speed is the magnitude of the velocity vector: ‖v(t)‖ = √[f ′(t)² + g′(t)² + h′(t)²]. It is always non-negative and tells how fast the position moves without regard to direction.

✦

KEY TAKEAWAY

Think of a drone flying through the air with a GPS logger recording its coordinates every millisecond. The position vector is the list of (x, y, z) readings; the velocity vector is the instantaneous rate of change of those readings — telling you the direction the drone is heading and how fast. The acceleration vector captures how that heading and speed are evolving: is the drone turning? speeding up? diving? Differentiation component by component is the mathematical mechanism that extracts this information from the position data.

SECTION 3

Visual Explanation — Velocity & Acceleration on a Curve

The curve C (violet) is the path traced by r(t). At the point r(t₀) (amber dot), the velocity vector is tangent to the curve, while the acceleration vector points toward the concave side of the curve. A second point r(t₁) (emerald) shows how both vectors change along the path.

Several observations emerge from the diagram. First, the velocity vector v(t₀) is always tangent to the curve at the point of evaluation — this is a direct consequence of the limit definition, since the difference quotient [r(t₀ + Δt) − r(t₀)] / Δt points from one position on the curve toward a nearby position and, in the limit, aligns with the tangent direction. Second, the acceleration vector a(t₀) generally does not point along the curve; it tilts toward the concave side because it encodes how the direction of motion is turning. Notice that at different points (t₀ versus t₁) both vectors change in both direction and magnitude, reflecting the evolving geometry of the path.

SECTION 4

Mathematical Framework

We now formalize the derivative of a vector-valued function using the limit definition, derive the velocity and acceleration vectors, and establish the product rules that extend from scalar calculus.

DERIVATIVE DEFINITION

r′(t) = lim [r(t + Δt) − r(t)] / Δt as Δt → 0

r(t) = ⟨f(t), g(t), h(t)⟩ is a vector-valued function of the scalar parameter t. The limit is taken component-wise, so r′(t) exists if and only if each component function is differentiable.

COMPONENT-WISE DERIVATIVE

r′(t) = ⟨ f ′(t), g′(t), h′(t) ⟩

Each component is differentiated independently with respect to t. This follows directly from the linearity of the limit.

VELOCITY AND SPEED

v(t) = r′(t), speed = ‖v(t)‖ = √[ f ′(t)² + g′(t)² + h′(t)² ]

The velocity vector v(t) is the derivative of the position vector. Speed is the scalar magnitude of velocity, always ≥ 0.

ACCELERATION

a(t) = v′(t) = r″(t) = ⟨ f ″(t), g″(t), h″(t) ⟩

The acceleration vector is the derivative of velocity (or the second derivative of position). It captures changes in both speed and direction.

Differentiation Rules for Vector-Valued Functions

All the familiar rules from single-variable calculus have vector-valued analogues. If u(t) and w(t) are differentiable vector-valued functions and c(t) is a differentiable scalar function, then (d/dt)[u + w] = u′ + w′ (sum rule), (d/dt)[c(t)u(t)] = c′(t)u(t) + c(t)u′(t) (scalar product rule), and analogous product rules hold for the dot product and cross product — with the critical caveat that order matters in the cross product because it is not commutative.

SECTION 5

Tangential & Normal Components of Acceleration

A deeper understanding of the acceleration vector comes from decomposing it into components parallel and perpendicular to the velocity. The tangential component a_T measures how quickly the particle's speed is changing, while the normal component a_N measures how quickly the particle's direction is changing. Together they satisfy a = a_TT + a_NN, where T is the unit tangent vector and N is the principal unit normal vector. This decomposition is especially illuminating for circular and helical motion.

At point P on the curve C, the total acceleration a(t) (pink, dashed) decomposes into a tangential component along the direction of motion and a normal component perpendicular to it, pointing toward the center of curvature. The dashed construction lines show the parallelogram law.

TANGENTIAL & NORMAL COMPONENTS

a_T = (v(t) · a(t)) / ‖v(t)‖, a_N = ‖v(t) × a(t)‖ / ‖v(t)‖

Here · denotes the dot product and × denotes the cross product. These formulas let you compute a_T and a_N directly from v and a without explicitly finding T or N.

Consider the special case of uniform circular motion. The speed is constant, so a_T = 0: the acceleration is entirely normal, pointing inward toward the center of the circle. This is precisely the centripetal acceleration from physics. Conversely, for straight-line motion there is no curvature, so a_N = 0 and all acceleration is tangential. Most real trajectories exhibit a mix of both components, and understanding this decomposition is essential for applications ranging from roller-coaster design to satellite orbit analysis.

SECTION 6

Worked Example — Helix in 3D

A particle moves along a circular helix with position r(t) = ⟨3 cos t, 3 sin t, 4t⟩. Find the velocity, acceleration, speed, and the tangential and normal components of acceleration at t = π/2.

Finding Velocity, Acceleration, and Speed for a Helix

1
Step 1 — Differentiate to Find Velocity

Differentiate each component of r(t) = ⟨3 cos t, 3 sin t, 4t⟩ with respect to t. The derivative of 3 cos t is −3 sin t, the derivative of 3 sin t is 3 cos t, and the derivative of 4t is 4.

v(t) = r′(t) = ⟨−3 sin t, 3 cos t, 4⟩
2
Step 2 — Differentiate Again for Acceleration

Differentiate each component of v(t). The derivative of −3 sin t is −3 cos t, the derivative of 3 cos t is −3 sin t, and the derivative of 4 is 0.

a(t) = r″(t) = ⟨−3 cos t, −3 sin t, 0⟩
3
Step 3 — Evaluate at t = π/2

Substitute t = π/2 into both expressions. At t = π/2: sin(π/2) = 1, cos(π/2) = 0.

v(π/2) = ⟨−3, 0, 4⟩, a(π/2) = ⟨0, −3, 0⟩
4
Step 4 — Compute Speed

The speed is the magnitude of the velocity vector: ‖v(π/2)‖ = √[(−3)² + 0² + 4²] = √[9 + 0 + 16] = √25.

Speed = 5
5
Step 5 — Tangential Component $a_T$

Compute a_T = (v · a) / ‖v‖. The dot product is (−3)(0) + (0)(−3) + (4)(0) = 0. Therefore a_T = 0/5 = 0.

a_T = 0 (constant speed on a helix)
6
Step 6 — Normal Component $a_N$

Since a_T = 0, all acceleration is normal. Alternatively, compute ‖a(π/2)‖ = √[0 + 9 + 0] = 3. Since ‖a‖² = a_T² + a_N², we get a_N = √[9 − 0] = 3.

a_N = 3 (centripetal acceleration toward helix axis)

💡

Physical Insight

The tangential component being zero confirms that the helix has constant speed — the particle never speeds up or slows down. All acceleration goes into bending the path, pointing radially inward toward the helix axis. This is analogous to uniform circular motion (a_N = v²/r = 25/r), and indeed the radius of this helix is 3, giving a_N = 9/3 = 3. ✓

SECTION 7

Common Pitfalls & Key Distinctions

Common Mistake	Why It's Wrong	Correct Approach
Confusing speed with velocity	Velocity is a vector (direction + magnitude); speed is a scalar (magnitude only). Writing speed = v(t) is incorrect.	Speed = ‖v(t)‖. Always take the magnitude of the velocity vector.
Assuming acceleration is tangent to the curve	Only the tangential component $a_T$ is along the curve. The full acceleration vector generally has a normal component too.	Decompose a(t) = $a_T$ T̂ + $a_N$ N̂ to separate the contributions.
Differentiating the magnitude instead of the vector	d/dt ‖r(t)‖ ≠ ‖r′(t)‖ in general. The derivative of a magnitude is not the magnitude of the derivative.	Differentiate component by component first, then take magnitudes if needed.
Swapping order in cross-product differentiation	The cross product is anti-commutative: u × w = −(w × u). Swapping factors changes the sign.	Maintain left-to-right order when applying the product rule to cross products.
Forgetting that constant-magnitude vectors have perpendicular derivatives	If ‖u(t)‖ = c (constant), then u(t) · u′(t) = 0 — a key geometric fact often overlooked.	Use this property to verify results, e.g., the unit tangent vector T̂ satisfies T̂ · T̂′ = 0.

Common errors when differentiating vector-valued functions

✦

KEY TAKEAWAY

A useful engineering analogy: imagine you're a pilot in a banking turn. The tangential acceleration is what you feel when the throttle pushes you back in your seat (changing speed). The normal acceleration is what pushes you sideways into the seat during the turn (changing direction). In level flight at constant speed, there is no tangential acceleration — just the normal component keeping you on the curved flight path. The total force you feel is the vector sum of both, and that's precisely what r″(t) encodes.

SECTION 8

Connections to Curvature & Advanced Theory

The derivatives of vector-valued functions open the door to the rich differential geometry of curves. The concepts of curvature κ and torsion τ are defined entirely in terms of vector derivatives, and the Frenet–Serret frame {T, N, B} provides a moving coordinate system that rides along the curve. These ideas extend naturally to surfaces and manifolds in differential geometry and to dynamics in physics via Newton's second law F = ma(t).

This Lesson	Advanced Extension
r′(t) — velocity vector	Unit tangent vector T̂ = r′(t) / ‖r′(t)‖; arc-length parametrization r(s)
r″(t) — acceleration vector	Curvature κ = ‖r′ × r″‖ / ‖r′‖³; radius of curvature ρ = 1/κ
$a_T$ and $a_N$ decomposition	Full Frenet–Serret formulas: T̂′ = κN̂, N̂′ = −κT̂ + τB̂, B̂′ = −τN̂
Speed ‖v(t)‖	Arc-length integral s(t) = ∫₀ᵗ ‖r′(u)‖ du; line integrals over vector fields
Particle kinematics in ℝ³	Dynamics: F = ma applied with position-dependent forces; Kepler orbits; Coriolis effects

From vector derivatives to advanced topics

Understanding the velocity and acceleration vectors as derivatives of r(t) is therefore not an endpoint but a launching point. Curvature, torsion, arc-length parametrization, and the Frenet–Serret apparatus all grow organically from the foundations established here. In physics, the equation of motion F = ma(t) directly identifies acceleration with applied force per unit mass, making the second derivative of the position vector the bridge between geometry and dynamics.

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

A particle moves along a curve with constant speed. What can you conclude about the direction of its acceleration vector? Justify your answer using the tangential and normal decomposition.

Show answer

If the speed ‖v(t)‖ is constant, then d/dt[‖v(t)‖] = 0, which means the tangential component

a_T

= (v · a)/‖v‖ = 0. Consequently, the entire acceleration is given by the normal component: a(t) =

a_N

N̂. This means the acceleration vector is always perpendicular to the velocity and points toward the center of curvature. Physically, the force only changes the direction of motion, not the speed — exactly what happens in uniform circular motion or along any constant-speed curved path.

Question 2: BASIC CALCULATION

Let r(t) = ⟨t², 2t, ln(t + 1)⟩. Find the velocity v(t) and the speed at t = 0.

Show answer

Differentiate component by component: v(t) = r′(t) = ⟨2t, 2, 1/(t + 1)⟩. At t = 0, we get v(0) = ⟨0, 2, 1⟩. The speed is ‖v(0)‖ = √(0² + 2² + 1²) = √(4 + 1) = √5. The velocity vector at t = 0 has no component in the x-direction and lies in the y-z plane, with a magnitude of √5 ≈ 2.236.

Question 3: INTERMEDIATE

A particle has position r(t) = ⟨eᵗ cos t, eᵗ sin t, eᵗ⟩. Find the velocity and acceleration vectors, and determine whether the speed is constant.

Show answer

Using the product rule on each component: v(t) = ⟨eᵗ(cos t − sin t), eᵗ(sin t + cos t), eᵗ⟩. The speed is ‖v(t)‖ = eᵗ√[(cos t − sin t)² + (sin t + cos t)² + 1]. Expanding the squares: (cos t − sin t)² + (sin t + cos t)² = cos²t − 2sin t cos t + sin²t + sin²t + 2sin t cos t + cos²t = 2. So ‖v(t)‖ = eᵗ√(2 + 1) = eᵗ√3. Since eᵗ is not constant, the speed is not constant — it grows exponentially. For acceleration, differentiate v(t) again: a(t) = ⟨eᵗ(cos t − sin t − sin t − cos t), eᵗ(sin t + cos t + cos t − sin t), eᵗ⟩ = ⟨−2eᵗ sin t, 2eᵗ cos t, eᵗ⟩.

Question 4: APPLIED

A projectile is launched from the origin with initial velocity v₀ = ⟨20, 30, 50⟩ m/s under gravity (g = 9.8 m/s² downward in the z-direction). Write the position vector r(t), find the velocity and acceleration vectors, and determine the speed at t = 3 seconds.

Show answer

The acceleration is constant: a(t) = ⟨0, 0, −9.8⟩ m/s². Integrating once gives v(t) = ⟨20, 30, 50 − 9.8t⟩ m/s. Integrating again (with r(0) = 0): r(t) = ⟨20t, 30t, 50t − 4.9t²⟩ m. At t = 3: v(3) = ⟨20, 30, 50 − 29.4⟩ = ⟨20, 30, 20.6⟩ m/s. The speed is ‖v(3)‖ = √(400 + 900 + 424.36) = √1724.36 ≈ 41.53 m/s. Note that the acceleration vector remains constant throughout the flight — only gravity acts — but the velocity vector changes because its z-component decreases linearly with time.

Question 5: CRITICAL THINKING

Let r(t) = ⟨a cos(ωt), a sin(ωt), bt⟩ describe a circular helix with radius a, angular frequency ω, and pitch parameter b. Prove that

a_T

= 0 for all t and express

a_N

in terms of a and ω. Then explain geometrically why the result is independent of b.

Show answer

Differentiate: v(t) = ⟨−aω sin(ωt), aω cos(ωt), b⟩ and a(t) = ⟨−aω² cos(ωt), −aω² sin(ωt), 0⟩. The speed is ‖v‖ = √(a²ω² + b²), which is constant — hence

a_T

= 0 for all t. For

a_N

, compute v · a = (−aω sin(ωt))(−aω² cos(ωt)) + (aω cos(ωt))(−aω² sin(ωt)) + 0 = a²ω³ sin(ωt)cos(ωt) − a²ω³ sin(ωt)cos(ωt) = 0, confirming

a_T

= 0. Then

a_N

= ‖a(t)‖ = aω² (since ‖a‖ = aω²√(cos²(ωt) + sin²(ωt)) = aω²). The result is independent of b because the vertical (z) motion is uniform and linear; it contributes to speed but not to the change in direction. The turning of the path is entirely due to the circular projection in the x-y plane, governed by a and ω alone. Geometrically, adding vertical drift 'stretches' the circle into a helix without curving the path any differently in the transverse direction.

SUMMARY

Summary — Derivatives of Vector-Valued Functions

Summary

The derivative of a vector-valued function r(t) = ⟨f(t), g(t), h(t)⟩ is computed component by component: r′(t) = ⟨f ′(t), g′(t), h′(t)⟩. The first derivative gives the velocity vector v(t), which is tangent to the curve and whose magnitude equals the speed. The second derivative yields the acceleration vector a(t) = r″(t), which decomposes into a tangential component (measuring change in speed) and a normal component (measuring change in direction).

The standard differentiation rules — sum, scalar product, dot product, and cross product — all extend to vector-valued functions, with the important caveat that order matters for cross products. Key formulas include a_T = (v · a)/‖v‖ and a_N = ‖v × a‖/‖v‖. These concepts form the foundation for studying curvature, torsion, the Frenet–Serret frame, arc-length parametrization, and the application of Newton's laws to curved-path dynamics.