Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

I Basic bra-ket arithmetic

  1. Feb 20, 2017 #1
    Say I have a vector product |x+a⟩⟨x| and I multiplied it by a ket vector |x'⟩. Can I pull the |x'⟩ into the ket vector |x+a⟩? also could you split up the ket vector |x+a⟩ into two ket vectors added together?
     
  2. jcsd
  3. Feb 20, 2017 #2

    Orodruin

    User Avatar
    Staff Emeritus
    Science Advisor
    Homework Helper
    Gold Member

    No and no.
     
  4. Feb 20, 2017 #3
    Okay thank you
     
  5. Mar 9, 2017 #4
    Is every complete bra-ket result a scaler, no matter how complicated?

    By every bra-ket I mean every operation that begins < and ends with >.

    Also, how do we interpret something like this? What is that equation saying?

    <Φ|A|Ψ> or <Ψ2|A|Ψ1>
     
  6. Mar 9, 2017 #5
    the probability amplitude of passing from the state ##A| \Phi \rangle## to ## |\Psi\rangle##
     
  7. Mar 9, 2017 #6

    blue_leaf77

    User Avatar
    Science Advisor
    Homework Helper

    As some operators are not Hermitian, it's actually more quantum-mechanically intuitive to say instead "the probability amplitude of passing from the state ##A| \Psi \rangle## to ## |\Phi\rangle##".
     
  8. Mar 10, 2017 #7
    for what I understand, quantum operators are hermitian, if an operator is not Hermitian, it is not associated to an observable. So when we speak of states bra or ket, we make the implicit assumption that operators are hermitian
     
  9. Mar 10, 2017 #8

    Orodruin

    User Avatar
    Staff Emeritus
    Science Advisor
    Homework Helper
    Gold Member

    What makes you think A is an observable? There are many operators in QM that are not Hermitian. The time propagation operator ##S = e^{-iHt}## comes to mind, or any other unitary operator.
     
  10. Mar 10, 2017 #9
    operator S is an observable ?
     
  11. Mar 10, 2017 #10

    hilbert2

    User Avatar
    Science Advisor
    Gold Member

    No it's not, because it doesn't have real eigenvalues except if for some value of t you have ##e^{-iHt} = 1## or ##e^{-iHt}=-1## by coincident.
     
  12. Mar 10, 2017 #11
    you have to read carefully .. I did not say that any operator in QM is hermitian, I said that the operator of an observable is hermitian. why are you unnecessarily complicating to a question about the elementary bases on the bra and ket. Go ahead and answer the question...
     
  13. Mar 10, 2017 #12

    Orodruin

    User Avatar
    Staff Emeritus
    Science Advisor
    Homework Helper
    Gold Member

    Obviously not. Yet it is a perfectly viable operator that is very important in QM. Your statement:
    is therefore wrong. For example, ##\langle \phi \lvert S \rvert \psi \rangle## would represent the probability amplitude of the state ##\lvert \psi \rangle## evolving into the state ##\lvert \phi \rangle##, not the other way around. There is no implicit assumption that ##A## must be a Hermitian operator in an expression such as ##\langle \phi \lvert A \rvert \psi \rangle##.
     
  14. Mar 10, 2017 #13
    and that's what I wrote in post # 5
    ok right, not in all generally, all operators are hermitian, but only associates to observable,
     
  15. Mar 10, 2017 #14

    Orodruin

    User Avatar
    Staff Emeritus
    Science Advisor
    Homework Helper
    Gold Member

    You said:
    blue_leaf77 pointed out that:
    to which you replied
    The last assertion in this post is wrong. We do not make the implicit assumption that operators are Hermitian when speaking of bras and kets. It is unclear to me why you start involving observables at all, this was not part of the original question and the first statement that you made generally needs the correction made by blue_leaf77. The operator ##S## is not an observable, but this does not matter, the bra-ket notation is perfectly capable of handling non-Hermitian operators and so blue_leaf77's correction of your post is perfectly warranted.
     
  16. Mar 10, 2017 #15
    because the concept of observable is really essential in quantum mechanics, and the formalism of the bra and ket, was introduced not on a whim, or because Dirac had nothing better to do but to clarify and mathematically formalize the concept of observable , among other things. I strongly advise you to read carefully "The principles of quantum mechanics" by Dirac.
     
  17. Mar 10, 2017 #16

    Orodruin

    User Avatar
    Staff Emeritus
    Science Advisor
    Homework Helper
    Gold Member

    Regardless of what Dirac intended, the A does not need to describe an observable and the braket notation is perfectly well suited to handle any linear operator on the Hilbert space. It is also done regularly in QM.
     
  18. Mar 10, 2017 #17
    Thanks for the replies.

    I did ask two questions in my post. I would really like to know the answer to the first question. Maybe it is so simple you think I should already know the answer. I suspect it is true.

    Is every complete bra-ket result a scaler, no matter how complicated?

    By every bra-ket I mean every operation that begins < and ends with >.

    Basically, what I think I am asking is if every <...> is a statement of a probability result of some kind.

    Also, based on the replies so far, I am seeing a similarity between the <| |> notations and conditional probability notation. In conditional probability notation, P(A|B) is interpreted to mean "the probability of A given that B has already occurred". Would I be wrong if I interpret the braket notation of, <Φ|A|Ψ> for examples as "the probability amplitude of being in state Φ given that the system is already in the state AΨ? (I was going to say observing instead of being but based on the previous answers I know that would be wrong)
     
  19. Mar 10, 2017 #18

    blue_leaf77

    User Avatar
    Science Advisor
    Homework Helper

    Depending on how we interpret the term scalar, yes it is always a "scalar". But sometimes there are some cases where the use of the word scalar can be ambiguous, consider the expectation value of position operator ##\langle \psi | \mathbf r |\psi \rangle##. The operator ##\mathbf r## is a rank-1 tensor (vector) operator and thus its expectation value.
     
  20. Mar 10, 2017 #19
    Thanks.

    By scaler I mean the quantity can be treated algebraically as a scaler.
     
    Last edited: Mar 10, 2017
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted



Similar Discussions: Basic bra-ket arithmetic
  1. Basic bra-ket algebra (Replies: 2)

  2. Ket to bra (Replies: 1)

Loading...