The Personal Blog of Stephen Sekula

When Hypotheses Collide

Toward the late afternoon, I left my office and went down to Stanford’s main campus from SLAC. I had been at SLAC since 7:30 that morning, with a morning spent in meetings and my afternoon spent doing actual, honest-to-God research. Jodi’s book club met tonight, so I decided to just go to Stanford early and work from her office until she was ready to leave.

When I showed up, Jodi printed off a document, handed it to me quietly, and went back to her work. The article she printed appeared online in The New Scientist, and was entitled “Gravity theory dispenses with dark matter”:http://www.newscientistspace.com/article.ns?id=dn8631. I sighed. Such sensationalist headlines were sure to hide the facts, and I decided to read the short article and see what the *real* story was.

Here’s the gist. A suite of independent observations suggest that there is something that affects our original view that the universe is made or ordinary matter, and that its large structure is due to Newtonian/Einsteinian gravity (that is, good old 1/r^2 for the dependence ofthe gravitational force on distance at large distances). If you hypothesize that there is a *dark matter* component to the universe, something particulate in nature but not a component of ordinary matter, you made progress. If dark matter is gravitationally attracted to ordinary matter, this explains galactic rotation velocities. Dark matter would affect gravitational lensing, by lensing more light than ordinary matter itself can – this is born out by observation. Dark matter would play a critical role in the formation of structure in the very early universe – within the first 300,000 years. Observations of the Cosmic Microwave Background (CMB) by such satellite-based experiments as COBE and WMAP demonstrate that the fine structure of the CMB is **extremely** well explained by a dark matter component to the universe, making up about 25% of the energy of the universe.

There are hypotheses that have been proposed that don’t require dark matter to explain the galactic rotation. MOND (Modified Newtonian Dynamics) is one such hypothesis, wherein gravity’s strength is modified when the acceleration is very, very small. This is successful in explaining the rotation curves of a limited number of galaxies. It fails, however, to explain the masses of clusters of galaxies. Therefore, that makes it less likely that it’s a correct description of Nature. Dark matter, however, passes this particular test.

Other ideas have been proposed. One among them is scalar-tensor-vector gravity (STVG) [gr-qc/0506021], wherein there are space- and time-varying scalar, tensor, and vector fields are able to generate effects that can explain some of the above observations. It has yet to demonstrate that it can explain the structure formation in the CMB, and this is apparently what the authors are intending to study next.

MOND, STVG, etc. are somewhat unsatisfying hypotheses because instead of being a mathematically rigorous framework, they are models. That is, they are mathematics built on top of observations, with their parameters fixed from existing data rather than bootstrapped from within the mathematical framework itself. The difference between a model and a mathematical theory is that a theory gives some clues as to “why”, where are a model simply describes “what is” with few, if any, statements about “why”. In the above examples, gravity is allowed to have a bunch of extra components, but no explanation is provided as to why this is allowed in Nature. The particle hypotheses is a bit more elegant in that everything we’ve ever observed as a fundamental component of Nature is a particle. It’s natural, then, to hypothesize that a particle is responsible for the observations of galaxies, the CMB, etc.

Now, arguements about who’s more natural or rigorous really all must bow to data. So far, the particle (dark matter) hypothesis has withstood all data, whereas MOND is not broadly applicable and STVG has to pass the most precise test: the CMB. What bothers me about the “The New Scientist” article is that the title suggests that just because STVG can describe a bit more than MOND, it’s giving the particle hypothesis a run for its money. It also implies that we should somehow abandon the particle hypothesis in favor of STVG.

The scientific thing to do is to follow the evidence to wherever it leads. Experiments now running, and soon to start running, will all test aspects of the dark matter particle hypothesis. The experiment Jodi works on, CDMS-II, is a direct test of the particle hypothesis which looks for galactic dark matter to collide with the CDMS detector. THe Fermilab Tevatron, and soon the Large Hadron Collider, directly search for the dark matter particle or particles by trying to produce them from proton collisions at tremendous energy. Even BaBar and Belle have something to say about dark matter – if it’s the result of supersymmetry, then precision measurements from the B-factories will constrain many parameters of that hypothesis.

MOND, STVG, and other non-particle theories have a lot of work to do. What’s more, they need to make predictions that would provide one answer for an observation, whereas the dark matter hypothesis would provide a different answer. Only by looking for questions that are differently addressed by these frameworks can we hope to narrow in on the real scientific theory that explains Nature, and weed out the dead-ends.


.. [gr-qc/0506021] “http://arxiv.org/abs/gr-qc?papernum=0506021”:http://arxiv.org/abs/gr-qc?papernum=0506021