>That's not a mechanism -- that's an improved
>predictive model. The question is "why does matter
>cause space to 'curve'?", i.e., "why is there
>gravity?", "why does matter have this occult
>property?"
>
i'd split the difference here. no previous theory i know of ever suggested a clock should run more slowly deeper in a gravitational field (a quick result from the equivalence principle), though Gauss DID wonder if large scale triangulation would show interior angles adding up to 180 degrees. a theory which -- on top of making clock synchronization and simultaneity practical realities -- ties our clocks and rulers to matter and energy is a more down-to-earth theory than we had previously.
although it doesnt qualify as a mechanism in the way i think Chris means, its nice to have a dynamic, field oriented model for gravity, even if gravity waves arent exactly of blinding intensity in this part of space-time. finding wave equations -- be it for the photon (electromagnetism), the electron, the graviton, and so on -- seems always to advance our understanding. such a theory for gravity suggests that binary star systems will evolve due to emission of gravitational radiation via such waves. (google: Taylor Hulse PSR B1913+16 <http://en.wikipedia.org/w/index.php?title=PSR_B1913_plus_16&action=edit> )
as to general relativity and the notion of a force, many physicists are fond of the expression "tidal interaction" or "tidal force" for gravity's action. this in honor of the observation that although free fall in space makes us feel weightless, body parts closer to earth are pulled away from body parts farther away from the earth's center. in a similar fashion, body parts seperated spatially but equidistant from earth's center are pulled closer together. the ocean's tides are a prime example of this effect. it is this tidal effect which is central to GTR. the notion that there is no gravity force because everything just falls freely in curved space-time is only one piece of the puzzle. a much clearer way to visualize GTR is to take an initially spherical cloud of mases and let them free-fall. watching the distortion of the spherical geometry gives a nice picture of the curved spacetime near massive objects. [if jks took relativity with Wheeler, he's heard all about this.]
by the way, Nature on Wednesday published a paper that verifed E=mc^2 to several parts in 10 million (a factor of 55x over the previous accuracy winner, electron/positron annhilation experiments). in the same issue, Freeman Dyson discusses S. Chandrasekhar theory of the death of "heavy" stars (stellar astrophyics plus general relativity plus QM) and Ed Witten waxes prophetic on string theory, with, however, a caveat on unified theories:
In fact, there are ample reasons why one might doubt whether
Einstein's vision is achievable, or at least achievable in the
foreseeable future. Crucial clues may be hopelessly out of reach.
When looking back at Einstein's own work, most physicists would say
that many of the most important clues for a unified field theory —
involving strong and weak nuclear interactions, the role of gauge
theory and the world of elementary particles — were simply not known
in Einstein's day.
Moreover, even if we could somehow find the unified field theory, it
is not at all clear whether we could determine that it is right.
From a simple combination of Planck's constant, the speed of light,
and Newton's gravitational constant, one can construct a natural
unit of length — the Planck length. First defined by Max Planck a
century ago, this length is so fantastically small that if it, or
something close to it, is fundamental in
physics, then some of the most important phenomena may be
permanently beyond our experimental reach.
in my opinion, though superstring/brane theory offers a possible escape from "failure to renormalize", the nicer results right now at the intersection of quantum field theory and general relativity are work on black hole energetics and thermodynamics (hawking et al, the penrose process, Wikipedia not bad here) -- and yes, string theory has a take on these topics -- along with first order quantum field theory in curved spacetime (my favorite result is the Unruh effect: http://en.wikipedia.org/wiki/Unruh_effect ), and quantum geometries (smolin et al, is area quantized? http://www.phys.lsu.edu/mog/mog21/node11.html ). though not yet (and maybe never) amenable to experimental verification, they have the kind of down and dirty feel to them that reminds me of einstein's speculations on time, simultaneity, and equivalence of mass and energy, all of which took years to really gel into a cohesive, experimentally verified whole.
les schaffer