Physics Stack Exchange is a question and answer site for active researchers, academics and students of physics. It's 100% free.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I wanted to gain a better understanding of the Born rule to make my class on quantum mechanic feel less ad hoc. To do so I attempted to show that the version (1) given in my book is equivalent to the version (2) on Wikipedia.

  1. The version in my book:

    $P_x = \frac{\langle\psi_x\mid\psi_x\rangle}{\langle\psi\mid\psi\rangle}$ Where $P_x$ is the probability of getting a state $\psi_x$ when measuring.

  2. Version on Wikipedia:

    the probability of measuring a given eigenvalue $\lambda_i$ will equal $\langle\psi\mid P_i\mid\psi\rangle$, where $P_i$ is the projection onto the eigenspace of A corresponding to $P_i$.

There is a short explanation but I would be glad if someone could put it into better words. Specifically I don't understand what $P_i$ is. I do know about the Hermitian operator, its eigenvalues and eigenspaces. But for example, why is $P_i = \mid\psi_i\rangle\langle\psi_i\mid$ when the eigenspaces are one-dimensional. And how do I go from there to the form in my book?

share|cite|improve this question
up vote 2 down vote accepted

$P_i = \mid\psi_i\rangle\langle\psi_i\mid$ is the one-dimensional "projection" operator. By "one-dimensional" it means this projection operator projects $\psi$ onto a single dimension in Hilbert space.

Firstly, any wavefunction $\psi$ can be written as a linear combination of orthogonal components. That is, $\psi = \sum a_i\mid\psi_i\rangle$ where $a_i$ is some coefficient. If there are $n$ such non-zero coefficients, $\psi$ can be thought of as a vector in $n$ dimensions, having components in each direction of length $a_i$ in this $n-dimensional$ Hilbert space. $a_i$ is also the amplitude that the result $\psi_i$ will be obtained if the wave-function is measured in this basis. The probability is amplitude^2.

The projection measurement essentially "projects" the state $\psi$ onto one of these components.

It is easiest to demonstrate why $P_i = \mid\psi_i\rangle\langle\psi_i\mid$ by applying it to the state $\psi$.

$P_i\psi = P_i\sum a_k\mid\psi_k\rangle = \mid\psi_i\rangle\langle\psi_i\mid\sum a_k\mid\psi_k\rangle = \mid\psi_i\rangle\sum a_k\langle\psi_i\mid\psi_k\rangle = a_i\mid\psi_i\rangle$

Therefore, the operator $P_i$ acting on some arbitrary state $\psi$, projects $\psi$ onto its i-th component vector. (This is analogous to projecting a 2D vector onto say its x-component in Euclidean geometry. For instance if a vector $V = ax + by$, then the projection onto x-axis would yield $V_x = ax$)

So since, $P_i\mid\psi\rangle=a_i\mid\psi_i\rangle$ we can easily show that $\langle\psi\mid P_i\mid\psi\rangle=\sum \langle\psi_k\mid a_k^*a_i\mid\psi_i\rangle = \sum a_k^*a_i\langle\psi_k\mid\psi_i\rangle = a_ia_i^* = \mid a_i\mid^2$

Therefore we have shown that $\langle\psi\mid P_i\mid\psi\rangle$ gives the probability of the wavefunction being in the eigen-state $\psi_i$.

The next step is to show how the one in your book is also the probability. Note your book's use of $P_x$ is to represent probability and is not the projection operator. First consider the denominator $\langle\psi\mid\psi\rangle = \sum^j\sum^k\langle\psi_j\mid a_j^* a_k\mid\psi_k\rangle$. The only terms that survive is when i=j. Therefore we arrive at:

$\langle\psi\mid\psi\rangle = \sum^j\langle\psi_j\mid a_j^* a_j\mid\psi_j\rangle = \sum^j a_j^*a_j = \sum^j \mid a_j\mid ^2 $ This is the total probability of any state which, if this is normalized, should be 1. Therefore $\langle\psi\mid\psi\rangle = 1$ for normalized states, otherwise it is the sum of all possible amplitudes^2.

Next we consider the numerator. This is the dot product of the x-components of the state, which will yield $a_x^* a_x = \mid a_x \mid ^2$ Therefore, numerator over denominator gives $\frac{\mid a_x \mid ^2}{\sum^j \mid a_j\mid ^2}$. This is the probability for a particular state x to occur divided by the probability that any of the possible states will occur (which should be 1 for normalized states).

share|cite|improve this answer
Great explanation, thank you! – Althalos Sep 21 '12 at 6:55

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.