How to Use Derivations

A derivation is not a performance of algebra. It is an argument that turns assumptions into a result. To learn from it, track what is assumed, what is proved, what is approximated, and what would fail if the assumptions changed.

Before Reading a Derivation

Answer these questions first:

1. What result is being derived?
2. What system or model is assumed?
3. What mathematical objects are being used?
4. Is the result exact, approximate, perturbative, variational, or conventional?
5. What prerequisites are needed to follow the steps?

If you cannot state the target result, the derivation will feel like a sequence of tricks.

Identify the Inputs

Most quantum derivations start with a small set of inputs:

• a Hilbert space;
• a Hamiltonian or Lagrangian;
• a basis or representation;
• boundary conditions;
• commutation relations;
• symmetry assumptions;
• normalization conventions;
• an approximation or limiting regime.

Write those inputs explicitly. For example, a harmonic-oscillator derivation needs the Hamiltonian, the domain or representation being used, and the condition $\omega>0$. A variational derivation needs a normalized trial state and a Hamiltonian bounded below in the relevant setting.

Track the Type of Each Step

Not every line has the same status.

Step type	What to ask
Definition	Is this introducing notation or a new object?
Algebra	Does the manipulation preserve operator order and domains?
Theorem	What hypotheses are required?
Approximation	What is being neglected, and why?
Convention	Would another source use a different sign, phase, or normalization?
Physical interpretation	Which measurement or observable would test the result?

This habit prevents exact statements, approximations, and conventions from being mixed together.

Reproduce the Skeleton

After reading once, close the page and reproduce the skeleton:

1. starting assumptions;
2. main transformation or method;
3. key intermediate identity;
4. final result;
5. validity conditions;
6. common failure modes.

Do not try to reproduce every line at first. A derivation whose skeleton you understand is easier to rebuild than one you memorized.

Check Dimensions, Limits, and Special Cases

A derived formula should survive sanity checks:

• Are units consistent?
• Does the result reduce to a known case?
• Does it behave sensibly when a coupling goes to zero?
• Does it respect normalization?
• Does it preserve Hermiticity, unitarity, or positivity when required?
• Does it respect symmetry?

For example, the harmonic-oscillator energy

$$E_n=\hbar\omega\left(n+\frac12\right)$$

has units of energy, is positive for $\omega>0$, and has evenly spaced levels. Those checks do not prove the derivation, but they catch many mistakes.

Compare Methods

Some results have multiple derivations. The harmonic oscillator can be solved by differential equations or ladder operators. Perturbative transition rates can be motivated by time-dependent amplitudes or density-of-states reasoning. Scattering formulas can be derived through differential equations, Green functions, or partial waves.

Different derivations teach different structure:

• differential-equation derivations teach boundary conditions and special functions;
• operator derivations teach algebra and spectra;
• variational derivations teach bounds and optimization;
• symmetry derivations teach selection rules and degeneracy;
• path-integral derivations teach stationary phase and histories.

For comparison, see Differential-Equation Solution and Ladder-Operator Solution.

Turn a Derivation into Problems

After reading, create small variants:

• change a sign and see where the proof fails;
• add a degeneracy;
• use a different basis;
• check a limiting case;
• repeat the derivation for a two-level example;
• identify the first step that uses a specific assumption.

This is often more valuable than copying the derivation again.

When a Derivation Feels Opaque

If the derivation feels opaque, locate the obstruction:

Symptom	Likely issue
You do not know what is being solved	Missing model or target result.
The algebra changes order mysteriously	Operator noncommutativity is being used.
Boundary terms vanish without explanation	Boundary conditions or domains matter.
A sum becomes an integral	A continuum or density-of-states limit is being used.
A term is dropped	An approximation or rotating-frame argument is being used.
A phase disappears	It may be a global phase or a convention.

Follow the prerequisite that matches the obstruction rather than rereading the whole page mechanically.

Common Mistakes

• Copying a derivation without listing assumptions.
• Treating an approximation as an exact theorem.
• Ignoring operator order.
• Forgetting boundary conditions in wave-mechanics derivations.
• Checking only algebra and not dimensions or limits.
• Reading a derivation once and assuming recognition equals understanding.

References

• R. Shankar, Principles of Quantum Mechanics, 2nd ed., Springer, 1994.
• J. J. Sakurai and J. Napolitano, Modern Quantum Mechanics, 3rd ed., Cambridge University Press, 2020.
• C. Cohen-Tannoudji, B. Diu, and F. Laloë, Quantum Mechanics, Wiley, 1977.

Exercises

1. Take a derivation you recently read and identify the first line where a physical assumption enters.

Solution

The answer depends on the derivation. In a particle-in-a-box derivation, the boundary condition is a physical input. In a perturbation derivation, the smallness of the perturbation is a physical and mathematical assumption. In a symmetry derivation, invariance under a transformation is the key assumption.

2. Why is reproducing a derivation skeleton often better than copying every line?

Solution

The skeleton records the logical structure: assumptions, method, key identity, result, and validity conditions. Once that structure is clear, missing algebraic details can be rebuilt. Copying every line can produce recognition without understanding where the result comes from or when it applies.