Module 1: Mathematical Foundations · Theory · assumes Foundation 8
The techniques that make hard problems solvable — and the discipline of knowing when they stop working.
Mahavakya
There is no exact treatment of anything. Every problem you have solved was an approximation — the skill is knowing which one you made.
On this page
Learning objectives
After this reading you should be able to:
- Apply (1 + x)n ≈ 1 + nx and recognise the situations that call for it.
- Use the small-angle approximations, and say roughly where each stops being safe.
- Show how sin θ ≈ θ converts the pendulum into simple harmonic motion.
- Read the tangent line as a local predictor, not only as a rate.
- Make an order-of-magnitude estimate and use it to check a detailed calculation.
- State the approximation behind a result, and the condition under which it holds.
1. Why physics approximates
Approximation has a reputation problem. It sounds like the thing you do when you cannot do the real calculation — a concession, a rough answer, second best.
That is backwards. Consider what you have assumed in every mechanics problem so far. The block was a point with no size. The string did not stretch and had no mass. The pulley was frictionless. The surface was perfectly flat. Gravity was uniform, air resistance absent, the Earth motionless, the collision instantaneous. Not one of those is true. Every problem you have ever solved was an approximation, and you were not told, because the alternative was no problem at all.
This is not a defect of teaching. It is how physics works. An exact description of a falling apple would need every air molecule, the Earth’s rotation, the Moon’s pull, the elasticity of the stalk — and would tell you nothing. Physics advances by deciding what to ignore, and being right about it.
So the techniques in this module are not a lesser mode of working. They are the mode, made explicit for once.
2. The binomial approximation
If you learn one thing here, learn this one. It appears more often in JEE physics than any other approximation.
(1 + x)n ≈ 1 + nx for |x| « 1
Any power — positive, negative, fractional
The power n is unrestricted, and that is what makes it so useful. Square roots become 1 + x/2. Reciprocals become 1 − x. Inverse squares become 1 − 2x. All of them without a calculator.
Worked example — gravity at altitude
How much weaker is gravity at height h above the Earth’s surface, when h is small compared with the radius R?
Exactly, g(h) = GM/(R + h)². Pull out R:
g(h) = (GM/R²)(1 + h/R)−2 = g0(1 + h/R)−2
With h/R small, the binomial approximation gives (1 + h/R)−2 ≈ 1 − 2h/R, so
g(h) ≈ g0(1 − 2h/R)
At the top of Everest, h/R ≈ 8.8/6370 ≈ 0.0014, so gravity is weaker by about 0.28% — roughly 9.78 rather than 9.81 m/s². Notice what the approximation bought: a messy inverse square became a straight line in h, and the answer is now something you can see rather than compute.
The approximation works because the terms it discards are of order x² and smaller. When x is 0.01, x² is 0.0001 — a hundred times smaller, and beneath the precision of anything you measured. Throwing it away costs nothing real.
3. Small angles
For angles small and measured in radians:
sin θ ≈ θ tan θ ≈ θ cos θ ≈ 1 − θ²/2
Radians only — in degrees these are simply false
Approximations are not wrong then right. They degrade, and you need to know how fast.
The left panel shows how gently sin θ and θ separate. At 10° the error is about 0.5%; at 20° about 2%; at 30° about 4.5%. Whether that matters depends entirely on what you are doing — 4.5% is intolerable in a precision measurement and irrelevant in an estimate.
Notice also that cos θ behaves differently. Its leading correction is θ²/2, so for very small angles cos θ ≈ 1 is often enough — the change is second order, not first. This is why a pendulum’s height barely changes for small swings while its horizontal displacement changes a great deal.
Worked example — why the pendulum oscillates
A bob of mass m hangs on a string of length L, displaced by angle θ. The restoring force along the arc is the tangential component of gravity, −mg sin θ, and the displacement along the arc is s = Lθ. Newton’s second law gives:
mL d²θ/dt² = −mg sin θ
This equation cannot be solved with elementary functions. The sine of the unknown makes it intractable — and for two centuries after Newton the exact pendulum remained a hard problem.
Now apply sin θ ≈ θ:
d²θ/dt² = −(g/L) θ
Look at what appeared. That is exactly the signature from Foundation 8 — acceleration proportional to displacement, directed back toward the origin. So ω² = g/L, and the period is T = 2π√(L/g). The most famous result in elementary mechanics rests entirely on one approximation, and it is true only for small swings. Push the pendulum to 90° and the real period is about 18% longer than the formula claims.
4. The tangent as a predictor
Foundation 8 built the tangent line to answer a question about rates: how fast is this changing right now. The same line answers a second question it was never introduced for.
f(x) ≈ f(a) + f′(a)(x − a)
Near x = a, the curve and its tangent are nearly the same thing
Read it as an instruction: start from a value you know, and walk along the tangent to estimate a nearby one. The right panel of the diagram shows √(1 + x) beside its tangent 1 + x/2 — the two are indistinguishable near the origin and drift apart slowly.
Which reveals something about section 2. The binomial approximation is this formula applied to (1 + x)n at x = 0, and the small-angle results are the same formula applied to sine and cosine. They are not three separate tricks to memorise. They are one idea — every smooth curve looks like a straight line if you stand close enough — wearing three costumes.
5. Order of magnitude
Sometimes the right answer is a power of ten. Estimating before calculating tells you what scale to expect, and an exact answer that disagrees with a sound estimate by a factor of a thousand is announcing an error.
The method is to keep only powers of ten, round aggressively, and let the errors cancel — round one quantity up and another down and the estimate stays honest. What matters is not being right to three figures but landing in the correct decade.
This is a habit rather than a formula, and it is the fastest error-catcher available in an exam. If a question about a car yields a speed of 10⁵ m/s, you do not need to find the mistake to know there is one. Foundation 4’s bound on a resultant did the same job for vectors; this generalises it to everything.
6. Knowing when it breaks
An approximation without a stated condition is not an approximation — it is a mistake waiting for the right question. Every result in this module carries one:
- (1 + x)n ≈ 1 + nx — needs |x| « 1. At x = 0.5 it is badly wrong.
- sin θ ≈ θ — needs θ small in radians. Beyond about 15° the error starts to matter.
- Tangent-line estimates — need x close to a. How close depends on how sharply the curve bends.
- The pendulum formula — small swings only, and it is not a law but a limiting case.
How large is “small”? There is no universal answer, and that is the honest position rather than an evasion. It depends on how much error the situation tolerates. The professional habit is to name the neglected term and estimate its size — if the next term is of order x² and x is 0.01, you are discarding one part in ten thousand, and you can say so.
Contemplation
Look back at what you have carried through ten modules. Foundation 1 said an equation is a compressed record of an observed regularity, valid only inside its assumptions. Foundation 3 said a coordinate system belongs to our method, not to nature. Foundation 4 said perpendicular components separate because the equations are linear, and that real projectiles do not oblige. Foundation 9 said the drag in a worked example was chosen to make the integral tractable.
Each was the same observation arriving in different clothing, and here it can be said plainly. There is no exact physics. There are models, each true within a region, and the whole art is in choosing which one and knowing where its region ends. Newton’s mechanics is an approximation that fails near light speed. General relativity is an approximation that fails at the smallest scales. Nobody has ever written down the exact description of anything, and there is no strong reason to expect they will.
This ought to be unsettling and is not. A student who believes physics delivers exact truths will one day be disillusioned. A student who understands that it delivers models with stated domains has understood the subject, and has something that will not be taken away by anything they learn later.
Key insight
Every approximation in this module is the same one: near enough to a point, a smooth curve is a straight line. The binomial expansion, the small-angle results and the tangent estimate are three faces of that single fact. Learn it once and the three stop being separate things to remember.
Common misconceptions
1. Approximating means settling for a worse answer
Usually it means getting an answer at all. The pendulum has no elementary exact solution; without sin θ ≈ θ there is no period formula to teach. And an answer good to 0.3% is not worse than an exact one when your measurement was good to 2%.
2. Small-angle formulas work in degrees
sin 10° is about 0.174, not 10. The approximation says sin θ ≈ θ with θ in radians, and 10° is 0.175 rad. A calculator in degree mode will not warn you.
3. The binomial approximation needs a whole number power
Any n works — 1/2, −2, 3.7. The restriction is on x being small, never on n. Students who learned binomial expansion as an algebra topic often carry the integer assumption across, and lose the most useful cases.
4. An approximation is either valid or invalid
It degrades continuously. sin θ ≈ θ does not fail at some threshold; the error grows smoothly, and the question is always whether it is small enough for the purpose at hand. Asking “is this allowed” is the wrong question. Ask how large the error is.
5. More decimal places means a better answer
Precision is not accuracy. A result quoted to six figures from a model that ignored friction is a precise answer to a question nobody asked. The limiting error is nearly always the model, not the arithmetic.
Reflection
- Take the last problem you solved. List three things you assumed without being asked to. Which would break first if the situation changed?
- The pendulum formula holds only for small swings, yet it is taught as though it were exact. Is that dishonest, or is it what teaching necessarily involves?
- If every physical theory is a model with a domain, what would it even mean for one to be finally true?
Key takeaways
- (1 + x)n ≈ 1 + nx for small x, with any power at all — the most used approximation in JEE physics.
- sin θ ≈ θ, tan θ ≈ θ, cos θ ≈ 1 − θ²/2, in radians only.
- sin θ ≈ θ is what turns the pendulum equation into simple harmonic motion and produces T = 2π√(L/g).
- All of these are the tangent line in disguise: close enough, every smooth curve is straight.
- Order-of-magnitude estimates catch errors that detailed calculation hides.
- An approximation without a stated condition is not an approximation.
- Precision is not accuracy. The model is usually the limiting error, not the arithmetic.
Prerequisite: Foundation 8 · Reading: 16 min · Practice: 20 min · Difficulty: Intermediate