Starting from an answer you already have
Imagine you have perfectly solved some clean problem — say the hydrogen atom, whose energy levels you know exactly. Now reality intrudes: you switch on a weak external electric field, or there is some small extra force you neglected. The problem is now slightly different, and the old exact solution no longer applies. But it is *almost* right. Perturbation theory is the systematic way to start from the answer you already have and gently correct it to account for the small new effect.
The everyday analogy is paying off a near-correct estimate in installments. Suppose a friend asks the price of a meal and you say "about thirty." The true bill is 31.40. Rather than recompute from scratch, you announce a first correction ("plus a bit more than a dollar"), then a smaller second correction for the cents, and so on. Each round nudges you closer. Perturbation theory does precisely this for energies and wave-shapes: a known starting value, plus a first correction, plus a smaller second correction, plus tinier ones still.
Splitting the problem into known plus small
The whole method begins with one honest bookkeeping move. You split the system's total energy operator — its Hamiltonian — into two parts: a familiar piece you can solve exactly, and a small leftover piece called the perturbation. The familiar piece gives you the energy eigenstates and energies you start from. The perturbation is the small new term whose effect you want to fold in. The key word is *small*: the method only works when that extra term is genuinely a minor disturbance, not a dominant one.
H = H0 + (small) V
(total) (solvable) (perturbation)
energy = E0 + E1 + E2 + ...
(known) (1st) (2nd) (smaller)First order: a beautifully simple idea
The leading correction — the first-order correction to the energy — turns out to have a wonderfully intuitive meaning. It is simply the average value of the perturbation, measured in the *original*, unperturbed state. In words: "how much does the new effect cost, on average, if the system keeps doing exactly what it was already doing?" You do not even need to know how the state changes yet; you just take the old state and ask what the new term contributes on average within it.
A real example makes it vivid. Drop a hydrogen atom into a weak electric field and its sharp energy levels shift and split apart slightly — an effect you can actually see in a spectrometer, called the Stark effect. Perturbation theory predicts those shifts to terrific accuracy without ever solving the full, field-included problem from scratch. That is the everyday power of the method: real, measurable consequences from a few terms of correction.
Going deeper, and where it breaks
If first order is not accurate enough, you go to the second-order correction, which captures something subtler: how the perturbation slightly *mixes* the original state with other states. For the lowest state this always drags the energy down a little further, since every other state lies above it. Second order is more work, but it carries a reliable rule of thumb — the second correction is usually much smaller than the first, which is why the whole series settles toward the true answer.
There is one important trap, worth naming honestly. The simple recipe quietly assumes each starting energy level is distinct. When two or more states happen to share the same energy — a situation called degeneracy — the naive formulas blow up, dividing by zero. The fix, degenerate perturbation theory, is to first let the perturbation pick out the "right" combinations of those tied states before correcting them. You do not need the machinery now; just remember the warning sign: equal starting energies demand extra care.