Nielsen and Chuang, Chapter 2 (Box 2.6):

Suppose $M$ is any observable on a system $A$, and we have some measuring device which is capable of realizing measurements of $M$. Let $\tilde M$ denote the corresponding observable for the same measurement, performed on the composite system $AB$. Our immediate goal is to argue that $\tilde M$ is necessarily equal to $M \otimes I_B$. Note that if the system $AB$ is prepared in the state $|m\rangle |\psi\rangle$, where $|m\rangle$ is an eigenstate with an eigenvalue $m$ and $|\psi\rangle$ is any state of $B$, then the measuring device must yield the result $m$ for the measurement, with probability one. Thus, if $P_m$ is the projector onto the $m$ eigenspace of the observable $M$, then the corresponding projector for $\tilde M$ is $P_m \otimes I_B$. We therefore have

$\tilde M = \sum_{m} m P_m\otimes I_B = M \otimes I_B$.

Our professor mentioned that for a system of two qubits $A$ and $B$, if we consider the wavefunction to be $|\phi\rangle = a |00\rangle + b |11\rangle$, then the expectation value of the observable $\tilde M$ will be:

$$\langle \phi| M \otimes I_B |\phi \rangle$$

$$= (a^*\langle00|+b^*\langle 11|) M \otimes I_B (a |00\rangle + b |11\rangle)$$

The expression for the observable $\tilde M = M\otimes I_B$ is only valid if the system $AB$ was initially prepared in the state $|m\rangle |\psi\rangle$ where $|m\rangle$ is an eigenstate with eigenvalue $m$. That means it is necessary that both the eigenstates $|0\rangle$ and $|1\rangle$ of $A$ had the same eigenvalue $m$. Otherwise the professors' expression for the expectation value is simply not true! I don't understand why the professor mention it explicitly. Am I missing something?

Or, are eigenvalues for the states $|0\rangle$ and $|1\rangle$ (for measurement operator $M$) always same, by default?

  • $\begingroup$ I don't understand what you are asking. Different measurement operator will have different eigenvalues and eigenvectors, and generally speaking the eigenvalues corresponding to $|0\rangle$ and $|1\rangle$ (provided that these states are eigenvectors) will be different. Also you wrote the equation for the expectation value of $\tilde M$ over $|\phi\rangle$, but then you mention $|m\rangle|\psi\rangle$. What is the relation between these states? $\endgroup$
    – glS
    Commented Jan 26, 2018 at 17:19

1 Answer 1


The text says, if you start in an arbitrary eigenvector $|m \rangle$, then you get the eigenvalue $m$. However, $m$ has no meaning, it's just a dummy variable. In your specific case, this means that if you start in $|0 \rangle$ you get $0$ and if you start in $|1 \rangle$ you get $1$. The equivalence $\tilde{M} = M \otimes I_B$ is true for all vectors since it's true for all eigenvectors .

  • $\begingroup$ Okay, but I don't think you addressed my question! I was asking about the eigenvalues for the two states...are they equal? If not, is the professor's expression for expectation value legitimate? $\endgroup$
    – user182786
    Commented Jan 26, 2018 at 16:40
  • $\begingroup$ @user182786 Whoops, I read too fast. New answer! $\endgroup$
    – knzhou
    Commented Jan 26, 2018 at 16:48
  • $\begingroup$ So, I understand that the eigenvalue is not same for $|0\rangle$ and $|1\rangle$. They are $0$ and $1$ respectively. OK. Then, I don't think the professor's expression for the expectation value of $\tilde M$ is correct if $\phi = a|00\rangle + b|11\rangle$ as the eigenvalue for $|0\rangle_A$ and $|1\rangle_A$ are not same. It would be correct if it were something like $\phi = a|00\rangle + b |01\rangle$ or $\phi = a|10\rangle + b|11\rangle$ where the eigenvalues for the first member's (state of $A$) are same. Am I right? $\endgroup$
    – user182786
    Commented Jan 26, 2018 at 16:56
  • $\begingroup$ I mean for this $\tilde M = \sum_{m} m P_M\otimes I_B = M \otimes I_B$ to be valid, it is necessary that the system of two qubits was prepared such that $A's$ state was an eigenstate in itself with some eigenvalue $m$. $a|0\rangle + b|1\rangle$ is clearly not an eigenstate with a certain eigenvalue $m$. So the expression cannot be valid for the original $\phi = a|00\rangle + b|11\rangle$ $\endgroup$
    – user182786
    Commented Jan 26, 2018 at 16:59
  • $\begingroup$ @user182786 The derivation is more general than that. It says, $\tilde{M}$ and $M \otimes I_B$ are equivalent for any eigenvector with an arbitrary eigenvalue, which we might call $m$. So they're equivalent, period. $\endgroup$
    – knzhou
    Commented Jan 26, 2018 at 17:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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