Module 1: Mathematical Foundations · Theory · assumes Foundations 3 and 7
Computing what Foundation 7 taught you to see — the exact rate of change at a single instant.
Mahavakya
A derivative is a rate measured over no interval at all. Nothing changes in an instant — and yet the change at that instant is the most useful number in mechanics.
On this page
Learning objectives
After this reading you should be able to:
- State the limit definition and explain how a chord becomes a tangent.
- Differentiate using the power, product, quotient and chain rules, and know which one a problem is asking for.
- Move up and down the kinematic ladder — position, velocity, acceleration, jerk.
- Locate equilibria from U(x) and classify their stability, by graph first and by the second derivative when there is no graph.
- Recognise a = −ω²x as the signature of simple harmonic motion and read off the period immediately.
- Use the chain rule on time-dependent constraints to connect rates that change together.
Notation. Leibniz: dy/dx. Lagrange: f′(x), f″(x). Newton’s dots, used only for time derivatives: ẋ = dx/dt, ẍ = d²x/dt². Δx is a finite change; dx is the vanishing one.
1. From average to instantaneous
Take a particle moving along a line and two moments, t and t + Δt. The average velocity between them is the slope of the chord joining the two points on the position–time graph:
vavg = Δx / Δt = [x(t + Δt) − x(t)] / Δt
Foundation 7 described what happens next in words: shrink the interval and the chord pivots until it settles on the tangent. Here is that process made exact.
The chord never becomes the tangent. It approaches it, and the limit is what we mean by arriving.
f′(x) = df/dx = limΔx→0 [f(x + Δx) − f(x)] / Δx
The derivative, where this limit exists
The word “limit” is doing careful work. We never set Δx = 0 — that would make the fraction 0/0, which says nothing. We ask instead what value the fraction approaches as Δx gets small without ever arriving, and if it settles on a single number, that number is the derivative.
Contemplation
Consider an arrow in flight, and freeze it at one instant. In that instant it occupies exactly one position. It does not move — there is no duration in which movement could occur. The same is true at every other instant. The flight is made entirely of moments in which nothing happens, and yet the arrow crosses the field. Zeno posed this twenty-four centuries ago, and it is not a trick.
Notice what calculus does with it. It does not answer the objection. It declines to define instantaneous velocity as motion during an instant, and defines it instead as the value that average velocities approach as the interval closes. Velocity at a moment is not something the arrow is doing. It is a statement about the whole family of intervals surrounding that moment, compressed into a single number and attached to a point.
This is Foundation 1’s theme in its most concentrated form. Nature does not hand us instantaneous velocity; the arrow simply flies. The derivative is our construction — and it is so good a construction that after a while it becomes almost impossible to remember it was built rather than found.
When the limit fails. A function must be continuous to be differentiable, but continuity is not enough. At a corner — the graph of |x| at the origin — the chords approach one slope from the left and a different one from the right, so there is no single limit and no derivative. Mechanics usually assumes motion is smooth, which is a modelling choice rather than a fact about the world. A ball striking a wall changes velocity so fast that treating it as instantaneous makes the acceleration undefined, which is precisely why collisions are handled with impulse instead of force.
2. The toolbox
Working from the limit definition every time would be intolerable. These rules follow from it once, and are then simply used.
| Rule | Form | Where you meet it |
|---|---|---|
| Power | d(xn)/dx = nxn−1 | Differentiating ½at² to get at |
| Constant multiple | d(cf)/dx = c f′ | Pulling mass out of d(mv)/dt |
| Sum | (f ± g)′ = f′ ± g′ | Term-by-term on any polynomial |
| Product | (uv)′ = uv′ + vu′ | dp/dt when mass varies, as in a rocket |
| Quotient | (u/v)′ = (vu′ − uv′) / v² | Any ratio of changing quantities |
| Chain | df/dx = (df/du)(du/dx) | Almost everywhere — see below |
The chain rule deserves a sentence of its own, because it is the one students under-use. It says that if a quantity depends on something that depends on time, the rates multiply. Energy depends on position, position depends on time, so dU/dt = (dU/dx)(dx/dt). Written that way it looks like cancelling fractions, which is not what is happening — but the notation was chosen so that it would look that way, and the resemblance is a reliable guide.
Derivatives worth knowing cold
d(sin x)/dx = cos x d(cos x)/dx = −sin x d(tan x)/dx = sec²x
d(ekx)/dx = k ekx d(ln x)/dx = 1/x (x > 0)
Radians only. The trigonometric derivatives above are false in degrees. They rest on sin θ ≈ θ for small θ, which holds only in radian measure — and a calculator left in degree mode will silently produce wrong answers all through a paper.
One property of ex is worth noticing: it is its own derivative. A quantity whose rate of growth equals its current size grows exponentially, which is why this function turns up in decay, in charging capacitors, and in a body approaching terminal speed.
3. Higher derivatives and the kinematic ladder
Differentiate a derivative and you get the next rung. In one-dimensional motion the ladder has names:
- Position x(t) — where the particle is.
- Velocity v = ẋ = dx/dt — how fast the position changes.
- Acceleration a = ẍ = d²x/dt² — how fast the velocity changes.
- Jerk j = d³x/dt³ — how fast the acceleration changes. Rarely examined, but it is what makes a lift feel unpleasant: constant acceleration is comfortable, sudden changes in it are not.
Each step divides by another second, so the units go m, m/s, m/s², m/s³. That is a free check on any answer — if a derivative comes out with the units of the original quantity, something has gone wrong.
4. Force, equilibrium and stability
For a conservative force, the connection to potential energy is a derivative:
F(x) = −dU/dx
Force is the negative of the slope of the potential energy curve
The minus sign says something physical: things are pushed downhill in energy. Where the curve slopes up to the right, the force points left. Where the curve is flat, the force is zero — and that is an equilibrium.
Read the graph first. Foundation 7 already gave you the classification, and it needs no calculus: a dip is stable, because a nudge produces a force pushing the particle back; a hump is unstable, because a nudge produces a force driving it away. Marble in a bowl, marble on a dome. The right-hand panel of the diagram above shows both, and if you are handed a curve this is all you need.
The second derivative is for when there is no graph. Given U(x) as a formula, curvature tells you which kind of turning point you have:
- d²U/dx² > 0 — a minimum, so stable.
- d²U/dx² < 0 — a maximum, so unstable.
- d²U/dx² = 0 — inconclusive. The test simply fails to decide, and you must look at the function itself.
That third case is worth a moment, because it is routinely misreported as meaning neutral equilibrium. It does not. Take U = x⁴: at the origin the second derivative is zero, yet the curve is a genuine minimum and the equilibrium is perfectly stable — merely very flat near the bottom. Neutral equilibrium is something else entirely: U constant over a whole region, so the force is zero not at a point but everywhere in it, and a displaced particle simply stays where you put it. A single zero of the second derivative tells you neither.
Worked example
A particle of mass 2 kg moves in one dimension with potential energy U(x) = x⁴ − 8x² + 12, with x in metres and U in joules. Find the force, locate the equilibria, classify them, and find the largest restoring force between x = −2 and x = 2.
Force. F = −dU/dx = −(4x³ − 16x) = 16x − 4x³ N.
Equilibria. Set F = 0: 4x(4 − x²) = 0, so x = 0, +2, −2 m.
Stability. d²U/dx² = 12x² − 16. At x = 0 this is −16, negative, so a maximum — unstable. At x = ±2 it is 48 − 16 = +32, positive, so minima — stable. The particle sits happily in either well and is pushed away from the hump between them.
Largest force. To maximise F we differentiate it: dF/dx = 16 − 12x² = 0 gives x² = 4/3, so x = ±2/√3 ≈ ±1.155 m. Substituting: F = 64/(3√3) = 64√3/9 ≈ 12.32 N.
Notice the last step used differentiation twice over, for two unrelated purposes — once to get force from energy, and once to find where a function is largest. The same operation, asked two different questions.
5. The signature of simple harmonic motion
One pattern is worth recognising instantly, because JEE returns to it constantly and often in disguise. If the acceleration of a body is proportional to its displacement and directed back toward the origin, then
d²x/dt² = −ω²x
The defining condition of simple harmonic motion
and the motion is simple harmonic, with x = A cos(ωt + φ) and period T = 2π/ω. You are almost never asked to solve that equation; you are asked to spot it. The work in a typical problem is to find the net force on a displaced body, show it comes out proportional to −x, read ω² off the coefficient, and quote the period.
Confirming the solution takes two differentiations. Starting from x = A cos(ωt), the chain rule gives v = −Aω sin(ωt), and differentiating again gives a = −Aω² cos(ωt) = −ω²x. The function returns to itself, negated and scaled — which is exactly what the equation demanded.
This also closes a loop from section 4. Near the bottom of any potential well, the curve looks like a parabola, the force looks like −kx, and the motion is simple harmonic. That is why oscillation is so common in physics: it is what almost anything does when disturbed slightly from a stable equilibrium.
6. Rotating basis vectors — the promise from Foundation 3
Foundation 3 left something unfinished. In Cartesian coordinates the basis vectors are constant, so dr/dt differentiates only the components. In polar coordinates r̂ and θ̂ turn as the particle moves, so they cannot be treated as constants — and extra terms appear. Here is where they come from.
Write the two polar basis vectors in Cartesian components. If the particle is at angle θ, then
r̂ = cos θ î + sin θ ĵ
θ̂ = −sin θ î + cos θ ĵ
Now differentiate with respect to time, remembering that θ itself changes with time — so every trigonometric function needs the chain rule, contributing a factor of θ̇:
dr̂/dt = (−sin θ î + cos θ ĵ) θ̇ = θ̇ θ̂
dθ̂/dt = (−cos θ î − sin θ ĵ) θ̇ = −θ̇ r̂
A satisfying result: each basis vector’s rate of change points along the other one. Now differentiate the position r = r r̂ using the product rule, since both r and r̂ can change:
v = ṙ r̂ + rθ̇ θ̂
The first term is motion outward; the second is motion around. Differentiate once more — product rule again on both terms — and the accelerations appear:
a = (r̈ − rθ̇²) r̂ + (rθ̈ + 2ṙθ̇) θ̂
Set r constant, as it is for circular motion, and everything with a ṙ or r̈ vanishes, leaving a = −rθ̇² r̂ + rθ̈ θ̂. With ω = θ̇, the first term is the familiar centripetal acceleration rω² pointing inward, and the second is the tangential acceleration produced by angular speeding-up. Neither was put in by hand. Both fell out of differentiating a rotating frame — which is what Foundation 3 promised and could not yet show.
Key insight
Centripetal acceleration is not an extra force or a special rule. It is what differentiation produces when the coordinate frame turns with the particle. Choose a frame that rotates, and the mathematics generates the terms the physics requires — which is Foundation 3’s claim about representation, now demonstrated rather than asserted.
7. Related rates
When several quantities are tied together and all changing, differentiating the relationship with respect to time connects their rates. The tool is the chain rule; the discipline is remembering that every variable depends on t.
A ladder of length L leans against a wall, its foot sliding out. The constraint is fixed: x² + y² = L². Differentiate with respect to time, applying the chain rule to each squared term:
2xẋ + 2yẏ = 0
ẏ = −(x/y) ẋ
Read what that says. The top of the ladder falls faster and faster as y shrinks, and as the ladder approaches flat, the predicted speed grows without bound — a signal that the model has left its domain, since a real ladder loses contact with the wall first. The mathematics is telling you where it stops being about a ladder.
The procedure generalises: write the constraint, differentiate everything with respect to time, then substitute the values at the instant in question. Substituting first is the standard mistake — it freezes the quantities you were supposed to differentiate.
Common misconceptions
1. d²x/dt² is (dx/dt)²
The first is acceleration, in m/s². The second is velocity squared, in m²/s². The notation is unfortunate but the units settle it instantly — and checking units is the habit that catches this before it costs anything.
2. Zero velocity means the object stays put
It means momentarily at rest. A ball at the top of its flight has v = 0 and a = −g, and it does not linger. Rest lasting for an interval requires the velocity to be zero throughout it, not at one instant.
3. The derivative of a product is the product of derivatives
It is uv′ + vu′. Picture a rectangle with both sides growing: the new area comes from a strip along each edge, one contributed by each side’s growth. The tiny corner square is what vanishes in the limit, leaving two terms rather than one.
4. Negative acceleration means slowing down
It means the acceleration points in the negative direction. A body speeds up whenever velocity and acceleration share a sign — including when both are negative. Deceleration is about agreement of signs, not about the sign itself.
5. Equilibrium is where U = 0
Force depends on the slope of U, not its value, so equilibrium is where dU/dx = 0. The value of U there is entirely up to where you chose to measure energy from — which, as Foundation 7’s transformations showed, is a free choice that changes no physics.
6. A zero second derivative means neutral equilibrium
It means the test has failed to tell you anything. U = x⁴ is stable at the origin despite U″ = 0 there. When the test is inconclusive, look at the function.
Reflection
- If U has units of joules and x of metres, what are the units of dU/dx and d²U/dx²? What does each of those quantities turn out to be?
- The derivative is defined as a limit that is never actually reached. Does that make instantaneous velocity less real than average velocity, or differently real?
- Near the bottom of any smooth well, motion is simple harmonic. Why should oscillation be the generic behaviour of disturbed systems, rather than a special case?
Key takeaways
- The derivative is the limit of chord slopes — a rate attached to an instant, built rather than found.
- Power, product, quotient and chain rules replace the limit definition in practice; the chain rule is the one to reach for most often.
- Position, velocity, acceleration, jerk — each rung divides by another second, which checks your units for free.
- F = −dU/dx. Equilibria are flat points; classify by graph when you have one, by curvature when you do not.
- A zero second derivative is inconclusive, not neutral.
- a = −ω²x is the signature of simple harmonic motion; recognise it and the period is 2π/ω.
- Differentiating a rotating frame produces centripetal and tangential terms without anyone inserting them.
- Smoothness is an assumption. Collisions break it, which is why they are handled with impulse.
Prerequisites: Foundations 3 and 7 · Reading: 22 min · Practice: 30 min · Difficulty: Intermediate