home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!ogicse!uwm.edu!caen!kuhub.cc.ukans.edu!husc-news.harvard.edu!husc8!mcirvin
- From: mcirvin@husc8.harvard.edu (Mcirvin)
- Newsgroups: sci.physics
- Subject: Re: Is Space a continuum?
- Message-ID: <mcirvin.713815149@husc8>
- Date: 14 Aug 92 17:59:09 GMT
- Article-I.D.: husc8.mcirvin.713815149
- References: <1992Aug14.154321.794@sinix.UUCP>
- Lines: 89
- Nntp-Posting-Host: husc8.harvard.edu
-
- dhg@sinix.UUCP (David Griffith) writes:
-
- >Is space a continuum? In other words, does General Relativity suitably
- >adjusted by Quantum Mechanics, treat space-time as infinitely
- >sub-divisible? If it's not, how does this affect the mathematics, eg
- >calculus which assumes deltas that tend to zero?
-
- Nobody really knows, but it's hoped that quantum mechanics will
- have some effect on these sorts of questions. Presumably spacetime
- doesn't behave like a smooth manifold on the Planck scale, but what
- it does instead is not obvious.
-
- The reason why such modification is to be hoped for has to do
- with quantum field theory. In classical field theory the components
- of the field are treated as functions of spacetime. In quantum
- field theory the functions are replaced by function-like things
- whose values are operators something like the ones encountered
- in nonrelativistic QM, except that the commutation (or anticommutation)
- relations don't involve finite [anti]commutators, but rather have
- things like Dirac delta functions (not the kind of delta
- you're talking about) in them.
-
- A Dirac delta "function" is the sort of thing that physicists
- use far too cavalierly for the tastes of mathematicians. It's
- zero everywhere except at one point, where it's infinite; and
- this infinitely tall, infinitely thin spike has a finite
- integral. This isn't really a function, it's a distribution. It
- is really only defined in terms of what you get when you
- integrate it over various intervals, and it's unclear what
- happens when you multiply, say, two Dirac deltas on the real
- line that have their spikes in the same place.
-
- Likewise, since these operator functions' commutation relations
- have Dirac deltas in them, they're not really operator-valued
- functions, but operator-valued distributions. You only get
- a legitimate quantum operator with finite values by integrating
- these things over some volume of space. The smaller the scale
- over which you integrate them, the more outrageously the numbers
- mount. Yet interacting field theories involve multiplying
- various numbers of these fields together at the same spacetime
- point. It's no wonder that hideous divergences show up in
- the process of calculating things like transition probabilities.
-
- Now, the means of getting finite answers out of quantum
- field theory (though this is not all it's good for) is
- called renormalization, and it typically involves two
- steps. The first step is to somehow modify the physics
- at short distances by applying an arbitrary short-wavelength
- cutoff, or making the space discrete, or some such thing;
- this is called regularization. Then, all results are expressed
- in terms of quantities having only to do with physics on the
- distance scale actually being dealt with. For some quantum
- field theories (a family that fortunately includes the theories
- of strong, weak, and electromagnetic interactions),
- this procedure will produce results that are actually
- independent of the arbitrary means of regularization, which
- can be removed from the final result. But the regularization
- was necessary to make things make sense in the first place.
- It's treated as a kind of fiction that gets removed after
- the renormalization is done, but you can't do calculations
- without it.
-
- One means of justifying renormalization as a legitimate
- procedure is to say that some form of regularization actually
- physically happens at very short distances, say, on the
- Planck scale, where it no longer becomes correct to
- speak of spacetime as a smooth manifold. Then, as long as
- the theory you're looking at is "renormalizable" -- that
- is, the values of physical quantities are actually
- independent of things that happen at arbitrarily short
- distances-- you can use any regularization method you
- find convenient and get correct results, knowing that
- the true method of regularization is somehow buried in
- quantum gravity.
-
- More importantly, if you try in any straightforward way to
- construct a theory of quantum gravity by turning general
- relativity into a quantum theory, you find that it is
- *not* renormalizable! Physical quantities depend very
- much on the method used to cut off divergences, if you
- try to calculate them using the means normally used in
- particle physics. So the hope is that quantum gravity
- itself will modify spacetime at short distances in such
- a way as to provide a natural cutoff for these divergences.
- Otherwise, it doesn't seem possible to make the theory
- make sense.
-
- --
- Matt McIrvin, professional gradgrind, amateur Usenet drifter
-