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

Dyson's 1951 text on arXiv

  1. Sep 18, 2006 #1
    Just started reading this online text (arxiv 0608140) and already ran into trouble in the second equation (see quoted below).

    The energy product of the wave function is negative. (Am I saying this correctly?)

    Dyson says "NR wave-mechanics tells you to take the equation [tex]E=\frac {1}{2m}p^2[/tex] of classical mechanics, and write

    [tex]E \rightarrow i \hbar \frac {\partial} {\partial t}[/tex] [tex]p_x \rightarrow -i \hbar \frac {\partial} {\partial x}[/tex] eqn. 1
    to get the wave-equation
    [tex]i \hbar \frac {\partial} {\partial t} \psi = -\frac {\hbar^2}{2m}(\nabla^2)\psi[/tex] eqn. 2

    satisfied by the wave-equation [tex]\psi[/tex]"

    I make the substitution of p(x) into the classical mechanics eqn. and so I have to square [tex]-i \hbar \frac {\partial} {\partial x}[/tex], which seems to me now to result in a positive quantity, contrary to the negative value given on the right in the wave equation. How should I evaluate the square so that it comes out positive?

    ok I got it. Factor the -1, squares to 1, and i squares to -1, easy.

    Well at least I got some practice using latex!

    Now what is that j vector in eqn 3? Does that little arrow over the j mean it is a unit vector? I think not, it is just an ordinary vector.

    [tex]\frac {\partial \rho} {\partial t} + \nabla \cdot \vec{j}=0[/tex] eqn. 3

    here rho is probability, [tex]\rho=\psi^* \psi[/tex] of finding the particle at xyzt.

    Dyson says probability is conserved because of eqn. 3, where

    [tex]\vec{j}=\frac {\hbar}{2mi}(\psi^*\nabla\psi-\psi \nabla\psi^*)[/tex] eqn.4

    How does this formula show that probability is conserved?

    Equation 3 tells me that [tex]\frac {\partial \rho}{\partial t}[/tex] and [tex]\nabla \cdot \vec{j}[/tex] are equal absolute value and opposite sign, or both zero. [tex]\frac {\partial \rho}{\partial t}[/tex] is the partial change in probability over time, which, as zero, is just the math statement that probability is conserved. So the meaning has to be encoded in the equality with [tex]-\nabla \cdot \vec{j}[/tex], the inner product of the del operator with j.

    Substituting eqn 4 into eqn 3 gives:[tex]\frac {\partial \rho} {\partial t} + \nabla \cdot \frac {\hbar}{2mi}(\psi^*\nabla\psi-\psi \nabla\psi^*)=0[/tex]


    [tex]\frac {\partial \rho} {\partial t} =- \nabla \cdot \frac {\hbar}{2mi}(\psi^*\nabla\psi-\psi \nabla\psi^*)[/tex]

    Now I am not getting much meaning from the right side of the eqn. IIRC [tex]\nabla^2[/tex] is zero, so the right side equals zero, as it should. But where is the meaning of all the rest of the formula? I am sure it isn't there just to be folded back into zero.

    Maybe it is in the meaning of the complex conjugation, whose rules I will have to go look up. Or maybe it is in the fraction, which I now interpret as the number of Planck constants in one cycle (2 pi) of the rather mysterious term, 2mi. At least I know, now, that mi is not the abbreviation for miles. But what is the significance ofthe product of twice the mass and the imaginary number i?
    Last edited: Sep 19, 2006
  2. jcsd
  3. Sep 19, 2006 #2


    User Avatar
    Gold Member

    Yep, j is the so called probability current, what as you see in the electromagnetic case becomes proportional to the electrical current, so they took the letter from classical electromagnetism or something so. The point is that this equation shows you that any change in rho does not happen from nothing, but comes from a current, j, so nobody is lost. You could fix a time t0 and integrate to calculate the total probability, then fix t1 and do the same and find it is the same. They do not worry about stating the fixed surfaces at t0 t1 (it is actually a bit tricky for relativistic mechanics), they just think on lines in the same way than gauss theorem (or Stokes theorem) and they are so used to do it that it is never mentioned, just the differential requeriments are stressed. A fast view of it, if you remember your vector integral calculus, is to check that Stokes (of Gauss) theorem can be applied to the 4 dim vector composed of (rho, j) just because of eqn 3. All these rotational of gradient, divergence of rotational and blah blah, do you remember?
  4. Sep 19, 2006 #3
    Yes, but you may have noticed I have been on a steep learning curve and as I remember it there was a lot of rubble at the base of the slope. As an autopedant, many things that I cover are only partly understood, and then partly retained. I get to see the connections later.

    I am belaboring this point because I need to get a firm understanding of the mathematical formulation of some really rather simple ideas. IIRC Gauss' law just says, in plain English, that what goes into a closed curve must come out again. You add up all the vectors and get zero.

    But what about the physical meaning that seems to me now to have to lie in the fraction or in the complex conjugation? I don't have my notes on these things to refer to, as I am not at my own desk. So I have to rely on Wiki.

    You see Dyson says the probability is conserved, because of eqn 3, and if this statement is meaningful it cannot merely be tautology. The use of because seems to me now to imply causation, not the mere equality it would be if the two statements only imply each other.

    Last edited: Sep 19, 2006
  5. Sep 19, 2006 #4


    User Avatar
    Gold Member

    what goes into a closed curve must come out again... if the field is conservative. There is a diffential geometric formulation for this "conservativity", coming from Stokes theorem and vectorial analysis, and formulae 3 is just the statement that of it.

    BTW, I hope now it is available more people will enjoy to peruse Dyson book (can you give the exact link to the arxiv numer?). If you get bored with the quantum mechanic, jump to read about Bethe estimation of Lamb shift, about the middle of the book, and compare with the final rigurous one, about the end of the book.
  6. Sep 19, 2006 #5
  7. Sep 19, 2006 #6

    Hans de Vries

    User Avatar
    Science Advisor


    A tip: Evaluate the right side with the plane wave solution,

    [tex] \psi\ =\ \exp{(-iEt/\hbar+ip_xx/\hbar+ip_yy/\hbar+ip_zz/\hbar)}[/tex]

    and see how you first get the momentum/current vector:

    [tex] {p_x \over m},\ \ {p_y \over m},\ \ {p_z \over m} \ \ =\ \ j_x,\ j_y,\ j_z[/tex]

    Note that the subtraction eliminates any "imaginary momentum"
    if p is complex instead of real. The divergence [itex]-\nabla \cdot [/itex] then gives you
    the amount of current flowing into the little cube dxdydz and
    hence how much rho changes per unit of time.

    Regards, Hans
    Last edited: Sep 19, 2006
  8. Sep 19, 2006 #7
    Hi Hans. Thanks.

    That exp operator has confused me in the past. I think I remember now that it means that the phrase following it is the exponent of e. Am I correct in imagining it could be written


    And then to evaluate the right side using the plane wave solution, do I just substitue the phrase in for psi? And then about that complex conjugate. Drat that I forgot to bring my notes. Something about (uv-vu)=?

    By the way, about a month ago I found my copy of The Cambridge Handbook of Physics Formulas in a box from when I moved Spring of 2005, and it was a joy to have in hand. But then I moved again the beginning of this month, and guess what, I can't find it again!

    However, I now think

    [tex]- \nabla \cdot \frac {\hbar}{2mi}(e^{(iEt/\hbar-ip_xx/\hbar-ip_yy/\hbar-ip_zz/\hbar)}
    \nabla e^{(-iEt/\hbar+ip_xx/\hbar+ip_yy/\hbar+ip_zz/\hbar)})
    \nabla e^{(iEt/\hbar-ip_xx/\hbar-ip_yy/\hbar-ip_zz/\hbar)})[/tex]

    Last edited: Sep 19, 2006
  9. Sep 19, 2006 #8


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    For what it's worth, I think you are going about this the wrong way. You need to learn stuff from the ground up, not the top down. And the Dirac equation is several steps up from being able to write down the complex conjugate of a number.
  10. Sep 19, 2006 #9
    Hi Gokul43201

    I am sure you are right, but I don't have a study guide to this stuff. I am trying to make progress wherever I can. I started the Dyson text because it seemed to me I could almost understand it, and my only method so far has been to try to understand, then fail, then analyse my failure, and try to understand what it was I didn't understand.

    Anyway thanks for the input.

    I suppose I might explain my method a bit.

    You have noticed the popularity of certain authors who interpret string theory for the masses in a rather non-mathematical way. Some of them made significant contributions, such as Einstein, Feynman, Hawking and others. Then they turned to explaining their insights to ordinary human beings. It is partially their fault that I have come to have a feeling for "beyond the standard model" kinds of theoretical physics. My math preperation is not sufficient to allow me to follow the more subtle and current arguments, so I set out to learn as much of the math as I could, by teaching myself from any text or conversant I could find. Unfortunately I am not talented in mathematics, and understanding it has been a struggle. Still, I have made progress.

    My premise is that the ideas themselves are not very complicated, but the mathmatical language used to describe the ideas is complicated. Even elegant mathematical statements such as Maxwell's equations are difficult or impossible to understand, if you have not first learned the machinery of calculus. It is very like trying to read a novel composed in a language you are unfamiliar with. The author may be discussing something as simple as cooking oatmeal, but you must first learn the symbols, the grammer, the definitions of seemingly endless numbers of words and phrases. You must do all this work, even if you already know how to cook oatmeal.

    Although I am not talented at math, I am a pretty good writer and do have pretty good mechanical aptitude and ability to visualize three dimensional objects. I have even come to think I can imagine the projection of the three dimensional objects in several dimensions of time. I have discussed my projections, non-mathematically, for some years now, and have come to the conclusion that I can only really challenge my perceptions by testing them against the ideas of people whose primary language is mathematics. So I study.

    As an outsider, even an outlander in the physics community, I am sometimes greeted by derision, ridicule, and outright rejection. Sometimes I think I understand how the barbarian felt when confronted by the streets of Imperial Rome. However, I have no interest in conquering the senate and sacking the Imperium. I don't want your honors and positions, so I don't care if you reject me personally. I am here to learn what I can of the truth. Several people here have been kind enough to help me along the way, and I am grateful. If the only way I can repay is by way of my laughable shambling street dance, and the only use you find for my attempt at self-education is the dull amusement of mockery and the feeble sense of superiority that it gives you, so be it. I have other things to think about.

    I mean no offense to you, Gokul, or to anyone, but I am sure you can find a better use for your wit than to use it attacking honest pilgrems. If I have misunderstood your approach and you are in fact trying to provide useful direction, I am sorry. However I have spent a sleep deprived and unproductive night after what I now think is another case of disparagement.

    As the poet Kerouac said, "Useless! Useless! Useless! Rain falls on the open sea." I hope by this to be able to regather my focus and direction.


    Last edited: Sep 20, 2006
  11. Sep 20, 2006 #10
    [tex]- \nabla \cdot \frac {\hbar}{2mi}(e^{(iEt/\hbar-ip_xx/\hbar-ip_yy/\hbar-ip_zz/\hbar)}\nabla e^ {(-iEt/\hbar+ip_xx/\hbar+ip_yy/\hbar+ip_zz/\hbar)})-(e^ {(-iEt/\hbar+ip_xx/\hbar+ip_yy/\hbar+ip_zz/\hbar)} \nabla e^ {(iEt/\hbar-ip_xx/\hbar-ip_yy/\hbar-ip_zz/\hbar)})[/tex]

    You see Dyson says the probability is conserved, because of eqn 3, and if this statement is meaningful it cannot merely be tautology. The use of because seems to me now to imply causation, not the mere equality it would be if the two statements only imply each other.

    My host, who works in industry as a physicist, has tried to help me.

    He started out with the definition of psi Hans gave in post 6,
    [tex]\psi = e^{(iEt/\hbar-ip_xx/\hbar-ip_yy/\hbar-ip_zz/\hbar[/tex]

    Then he began by calculating [tex]\nabla \psi[/tex] and [tex]\nabla \psi^* [/tex]


    [tex]\nabla\psi=\hat i \frac{\partial}{\partial x} (e^{(iEt/\hbar-ip_xx/\hbar-ip_yy/\hbar-ip_zz/\hbar)}) + \hat j \frac{\partial}{\partial y} (e^{(iEt/\hbar-ip_xx/\hbar-ip_yy/\hbar-ip_zz/\hbar)}) + \hat k \frac{\partial}{\partial z }(e^{(iEt/\hbar-ip_xx/\hbar-ip_yy/\hbar-ip_zz/\hbar)})[/tex]


    [tex]\nabla\psi^* =\hat i \frac{\partial}{\partial x} (e^{-(iEt/\hbar-ip_xx/\hbar-ip_yy/\hbar-ip_zz/\hbar)}) + \hat j \frac{\partial}{\partial y} (e^{-(iEt/\hbar-ip_xx/\hbar-ip_yy/\hbar-ip_zz/\hbar)}) + \hat k \frac{\partial}{\partial z }(e^{-(iEt/\hbar-ip_xx/\hbar-ip_yy/\hbar-ip_zz/\hbar)})[/tex].

    At this point he wasn't sure if the differentials for Px and X ...(and similarly for Py and Pz) have to be done separately as independent variables:

    [tex]\frac{\partial}{\partial x} \frac{PxX}{\hbar}= \frac{x}{\hbar} \frac{\partial Px}{\partial x} + \frac{Px}{\hbar}.....[/tex]

    or if they could be done together as

    [tex]\hat i i \frac{Px}{\hbar}....[/tex]

    Is the momentum at x dependent on x? That is, in the wave, does the momentum change as the particle position changes?

    I should think that in the three dimensional representation, where one spatial axis, the direction of motion, is replaced by the time axis, the particle would be seen to follow a path of constant velocity in a sort of corkscrew trajectory. Hence, I should think that momentum is not dependent on position in the four dimensional system, but would remain constant. However, my friend was not sure about this interpretation. Any advice?

    I have a job interview as a mad scientist and have to go now, but will return later today.

    Last edited: Sep 20, 2006
  12. Sep 20, 2006 #11
    We would all like to help you but the more questions you ask because you didn’t start from “the ground up” the less people are going to be willing to help you. Don't let your ego get in the way of learning physics and math in a manner that would be easier not only for you but, for everyone else.

    If you need a study guide check the requirements from any university for a undergraduate physics degree, they will list all the classes you need to get a solid understanding of basic physics.

    The concepts in physics are simple but, knowing the laws of physics in general is very different from knowning how the laws of physics apply to a specific situation and how they work to give a specific result, that's why we need the complicated maths.

    If you wish to continue, I think the “Quantum Physics” forum would be a better fit for you. The reason why physicsforums has separate forums is to organize all of our efforts to learn physics, help others learn physics, and explore the boundaries of physics, the most efficient way to do that is to separate the forums along several lines in physics.
    Last edited: Sep 20, 2006
  13. Sep 20, 2006 #12

    Hans de Vries

    User Avatar
    Science Advisor

    The E and p's are all constants in a plane wave, so:

    [tex]\frac{\partial}{\partial x} e^{-(iEt/\hbar-ip_xx/\hbar-ip_yy/\hbar-ip_zz/\hbar)} =\ i\frac{p_x}{\hbar}\ e^{-(iEt/\hbar-ip_xx/\hbar-ip_yy/\hbar-ip_zz/\hbar)}[/tex]

    Regards, Hans
    Last edited: Sep 21, 2006
  14. Sep 20, 2006 #13
    That seems right to me now, except shouldn't the derivative on the left be over x?


  15. Sep 21, 2006 #14

    Hans de Vries

    User Avatar
    Science Advisor


    You now can see how the exponents cancel each other in the usual way
    via the product with its complex conjugate:

    [tex]\psi^*\frac{\partial \psi}{\partial x}-\psi\frac{\partial \psi^*}{\partial x}[/tex]

    The subtraction merely doubles the result since the signs are opposite as
    long as p is real. In the case that p is complex however, It removes the
    imaginary part so the end result is always real.

    The divergence [itex]-\nabla\cdot[/itex], as you might remember, takes the derivative of
    each current component and sums them. If you envision an infinitesimal
    cube dx, dy, dz then the derivative gives you the difference of the
    current going in and the current going out. The minus sign takes care
    that it's the net current entering the cube.

    The net current entering the cube now gives the rate of change of rho,
    which is as it should be according to the requirement of conservation.
    The division by m just normalizes the mass-currents and turns them
    into probability-currents where the total probability is 1.

    The law thus describes the conservation of probability (=unitarity), the
    total chance to find the particle remains always 1.

    Regards, Hans
    Last edited: Sep 21, 2006
  16. Sep 21, 2006 #15


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Guess that's what you get for trying to help. :confused:

  17. Sep 21, 2006 #16
    I am reading on in the Dyson book. I have found that Wiki has grown since I last used it to study the del and the d'Alembert operators. I find their explanations clear and insightful, at my current level of understanding. I also feel I have a better understanding now of the Laplacian. At the end of the article is a list of vector identities for well-behaved scalar and vector functions. I noticed that the fourth one listed seems incomplete....it does not contain an equivalance relation. I put a note about that in the discussion section.

    Tomorrow I have appointments all day, so I will probably not have much time here. I look forward to resuming this study as soon as possible.

    Thanks all,

Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?

Similar Discussions: Dyson's 1951 text on arXiv
  1. Arxiv trackback (Replies: 24)

  2. Arivero paper on arxiv (Replies: 14)

  3. New Arxiv stats (Replies: 4)