home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!cis.ohio-state.edu!pacific.mps.ohio-state.edu!linac!uwm.edu!ogicse!das-news.harvard.edu!husc-news.harvard.edu!husc8!mcirvin
- From: mcirvin@husc8.harvard.edu (Mcirvin)
- Newsgroups: sci.physics
- Subject: Re: FTL
- Message-ID: <mcirvin.715059853@husc8>
- Date: 29 Aug 92 03:44:13 GMT
- Article-I.D.: husc8.mcirvin.715059853
- References: <1830.2A8E1404@catpe.alt.za> <1992Aug18.085010.24910@syma.sussex.ac.uk>
- <17mel5INNc1n@agate.berkeley.edu> <mcirvin.715053036@husc8> <mcirvin.715057081@husc8>
- Lines: 89
- Nntp-Posting-Host: husc8.harvard.edu
-
- mcirvin@husc8.harvard.edu (Mcirvin) writes:
-
- [more puzzlement about why the non-disappearance of the propagator
- at spacelike separation doesn't transmit information]
-
- I think I've got it. Here's an explanation, complete with bad
- ascii graphics (I hope the noisy phone line doesn't blitz this
- completely):
-
- In the covariant formulation of QFT (let's use the theory of
- electrons for familiarity), we can "create an electron" with a
- current at a point in spacetime and then "destroy it" elsewhere,
- with a nonzero probability at spacelike separation:
-
- *----------<-----------*
-
- The asterisks are the current. But this isn't the whole process;
- the current might also create multiple particle-antiparticle pairs,
- which could do all manner of things.
-
- (This "current" isn't the electric current; it's a fictional e
- function that has been inserted into the Lagrangian to give the
- electron field a handle we can grab it with. In practice this could
- be replaced by any physical process that creates or destroys electrons.)
-
- This Feynman diagram is an expression in covariant perturbation theory (PT).
- Now, let's go back to the pre-World War Two means of doing things,
- with non-covariant PT of the sort that is still usually used in
- nonrelativistic QM. This works just as well for relativistic QM,
- though it's harder to use. In covariant PT energy and momentum
- are always conserved, but virtual particles can violate laws of
- motion to do things like what I pictured above. In non-covariant
- PT the particles (or their wave functions) always obey the laws
- of motion, but energy conservation can be violated for short periods
- of time, in accordance with the time-energy uncertainty principle.
-
- In non-covariant PT the process looks like this:
-
- ^
- / \
- / \
- / \
- v ^
- / \
- / \
- * *
-
- Time is increasing upwards.
- The line is supposed to be going up from the asterisk on the right,
- turning a corner, and then going back down to the asterisk on the left.
- The angle is less than 90 degrees. The process is the creation of an
- electron on the right and a positron on the left, which annihilate.
- Energy is temporarily nonconserved, since the asterisk on the left is
- supposed to be absorbing energy rather than creating it, but everything
- balances out in the end, since the annihilation of the virtual pair
- creates no photons. If the distance between the asterisks
- is d, then the process takes a time t > d/(2c).
-
- Now, you might say that since the probability for the whole process
- depends on the creation of the electron on the right, you could detect
- a superluminal signal on the left by measuring probabilities for repeated
- trials. Not so! This isn't the amplitude for creating an electron
- on the right and a positron on the left; it's (part of) the amplitude for
- creating that initial state and *ending up with nothing.* Like Sidney
- said, the same current can produce different numbers of particles if
- it's localized to any degree, by the same Fourier transform that brought
- you the Uncertainty Principle in wave mechanics (although for Fermi
- particles this is kind of hard to imagine, since both the field and the
- current I'm using to manipulate it are made of Grassmann variables, but
- that's immaterial; I could just as easily have used bosons.) The only way
- someone standing at the position of the asterisk on the left can even
- *tell* that this is the process in question, and therefore begin to
- calculate the measured probability, is to wait until it can be seen
- that there aren't any particles left, which is a time t' of at least
- t' = t + d/(2c), which is greater than d/c.
-
- Really, having imagined this noncovariant picture in which the particles
- behave nicely, you can see it in the covariant picture too. What
- doesn't vanish at spacelike separation is the amplitude for creating
- a particle at point a, seeing it at point b, and having nothing left
- over. You can't tell that more than one particle wasn't created in the
- first place until you can see the creation event at point a, so you
- can't extract information from probabilities until such a time as you
- could have received light-speed signals anyway.
-
- It's quite pretty.
-
- --
- Matt McIrvin, Cambridge, Massachusetts, USA
-