- #31
Jonathan Scott
Gold Member
- 2,346
- 1,192
It's fine accepting them within their established scope of validity.ZapperZ said:
GR is known to be incredibly precisely accurate in weak fields at the scale of the solar system and quite a long way beyond that (including the creation and propagation of gravitational waves), but that's basically just a small relativistic correction to Newtonian theory. It's also known to be fairly accurate in stronger fields, but as we move through neutron stars towards collapse the results become more qualitative than quantitative, and so far we have no direct confirmation of the existence of event horizons, nor of other related predictions such as black holes being unable to retain any significant magnetic field.
What concerns me is that the theoretical predictions of GR beyond that strength and scale diverge from the current experimental observations and come into conflict with QM, but the fact that it works so well in the solar system is often assumed to mean that it is valid at all scales, so we end up for example with "dark matter" being assumed, and with predictions of physically meaningless singularities.
When the first LIGO GW (gravitational wave) observation occurred, the Fermi observatory reported an apparent GRB (gamma ray burst) at the same time, which appeared to be statistically significant, although not totally convincing. To someone who is open to GR only being an approximation, this suggests the very interesting possibility that objects of that mass don't actually have an event horizon, so the collision could emit electromagnetic radiation. However, the general reaction was that GR says no significant electromagnetic radiation can be produced by a collision of black holes, so it must be a mistake or require an exotic explanation (such as a collision inside a star). The theory, pushed well beyond its previously tested scope, was immediately assumed to take precedence over experiment. Some re-analysis concluded that the probability of the Fermi observation being a false signal in background noise was actually higher than originally calculated (although still small), and no other observatories detected anything, so the result was apparently comfortably dismissed. This may well have been the correct conclusion, but I was quite disturbed by the way in which the result was apparently immediately assumed by the "establishment" to be wrong based on theory which had never before been tested near that limit.
This suggests to me that although GR is an attractively neat theory with a wide range of successful applicability, it is probably being accepted more strongly outside that range than the evidence warrants, which could make it more difficult than necessary to make progress in finding the next better theory. Athough I know this isn't preventing some people from trying, I still consider it a "bit of a roadblock".