On 24/05/12 23:09, Jos Bergervoet wrote:
> On 5/23/2012 6:01 PM, Hendrik van Hees wrote:
>> On 23/05/12 03:05, Jos Bergervoet wrote:
> ...
>> If you could describe time as an observable in quantum theory, then it
>> must fulfill Eq. (4) since by definition the Hamiltonian generates time
>> translations. Of course, then you could define H as -i \partial_t and T
>> as the multiplication with t,
>
> Well, we could call H just p_0. No need to give it a
> special name in that case!
It doesn't help to rename things. You can denote your Hamiltonian with
whatever letter you like.
>
>> .. but it is bad for (at
>> least) two reasons.
>>
>> If you assume time to be an observable and obeying the above commutation
>> relation then by the same arguments as for position and momentum
>> operators in the Heisenberg algebra [x_i,p_j]=i delta_{ij} one has whole
>> R as the spectrum of H, and thus energy wouldn't be bounded from below,
>> and there was no stable matter, contrary to everyday experience.
>
> That is not completely obvious.. The assumption seems to be
> that the energy would tend to go to -infinity. But (since
> we're comparing the coordinates) we do not see that all
> particles in our universe tend to get p_x -> -infinity!
> So why do you claim that something will go wrong with time,
> where it clearly is no problem with the space coordinates?!
The reason is that H is the generator for time translations, i.e., it
governs the time evolution of the system. If the energy is not bounded
from below, there is no stable ground state. Applied to the description
of particles this would mean that matter wouldn't be stable, contrary to
everyday experience.
Momenta generate translations in space, and the Cartesian coordinates of
the position vector are indeed unbounded, and there is no problem with
stability whatsoever.
>
>> Moreover, I do not see, how you could get any real dynamical description
>> of quantum systems, since it is not clear, which equation you would
>> solve. The equation of motion i\partial_t psi(t,x)=H psi(t,x) would be a
>> tautology (despite the fact that with our sign choice of (4) I'd have to
>> write -i\partial_t on the left-hand side).
>
> We could make *another* choice for the evolution operator!
> For instance a "z-Hamitonian", H_z, which describes the
> evolution in the spatial z-direction. It would be like
> using the Dirac equation but not isolating the d/dt term
> (by factoring out a gamma_0) but instead isolating a d/dz.
Let's discuss non-relativistic quantum mechanics here. In relativistic
quantum theory things are more complicated, particularly with respect to
the definition of position operators. Of course also there, the above
argument against the interpretation of time as an observable holds. The
existence of a stable ground state is an important constraint for model
building also in relativistic quantum theory in terms of a local causal
quantum field theory. Together with locality it determines the general
structure of physically sensible QFTs, leading to quite general
predictions as the connection between spin and statistics and CPT
invariance.
>
> Obviously, the result would have positive and negative p_z
> solutions (as with the original Dirac energy spectrum!) but
> this need not bother us now, so this choice is much better!
Sure, also the Cartesian components of momentum have an unbounded
spectrum, which is no problem either.
>
> Conceptually, it comes down to the question whether we can
> describe an "evolution" mechanism in any direction of space-
> time, or only in the t-direction. The latter is of course
> special because of the signature of the metric, but it is
> not immediately clear (to me) what goes wrong if you take
> the other choice!
Of course, time is fundamentally different from space since it
parametrizes the "causality flow" of events, and that's why the metric
in relativity has its specific structure (1,3) or (3,1) in the
west-coast or east-cost notation, respectively.