For a discrete symmetry: At the minimum value of the potential, $V$, in the Lagrangian density, why do we take $\phi= \langle v\rangle + \eta$? Aren't we deliberately breaking the symmetry? If we don't do this, the symmetry is intact. On the other hand, if we replace $\phi$ by $\phi= \langle v\rangle + \eta$, even when not in a ground state, then also the symmetry will be broken in $\eta$ field. Please explain.

  • $\begingroup$ How is this question specifically related to discrete symmetries? $\endgroup$
    – innisfree
    Sep 11, 2014 at 7:44
  • $\begingroup$ no it's not. i chose this because it is simplest example of spontaneous symmetry breaking $\endgroup$
    – 23rduser
    Sep 11, 2014 at 13:01

1 Answer 1


In canonical quantization, a quantum field is a linear combination of so-called "creation and annihilation operators". That means that the field $\phi$ creates and destroys particles of type $\phi$.

The state $|0\rangle$ is the vacuum: the state with no particles. If $\phi$ is a quantum field that creates and destroys particles, it must be that $\langle 0 | \phi | 0 \rangle=0$, because the state with particles created/destroyed, $\phi|0\rangle$ must be orthogonal to the empty state $|0\rangle$.

We have no choice, then, but to write $\phi=v+\eta$, such that $\langle \eta \rangle=0$. Th new field $\eta$ now has a good particle interpretation, whereas the original field $\phi$ did not, because $\langle\phi\rangle = v$.

  • $\begingroup$ Thank you. Yes, we should write phi(x) = <vev> = eta(x) not only at ground state but always. Why is it that we write it only for ground state? a field once defined is defined for all states. $\endgroup$
    – 23rduser
    Sep 12, 2014 at 8:32
  • $\begingroup$ @innisfree I want to make a comment. I don't think $\langle 0|\phi|0\rangle=0$ only when $\phi$ can be written as a linear combination of $a$ and $a^\dagger$. For example, in self-interacting $\lambda\phi^4$-theory, one has $\langle 0|\phi|0\rangle=0$. The actual criterion is, if $\mathcal{L}(-\phi)=\mathcal{L}(\phi)$, all $n$-point functions with odd $n$ will vanish even if $\phi$ is non-exapandable as a linear combination of $a$ and $a^\dagger$. $\endgroup$
    – SRS
    Jun 19, 2017 at 14:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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