|Annu. Rev. Astron. Astrophys. 1993. 31:
Copyright © 1993 by . All rights reserved
An analysis of theories for the origin of large-scale cosmic structure by Peebles & Silk (1988) opened with the question, "Will our feeble minds ever comprehend the evolution of the Universe?" A good question. It would be easy to be pessimistic about this matter. When we understand a phenomenon well, we can make testable predictions. We are confident that we understand the solar system, in good part because we can predict eclipses and cometary apparitions. But our record of predicting what new studies of large-scale structure will find has been abysmal. In addition, the theories have seemed, to "outsiders" not working actively in the field, fantastic in the extreme, highly dependent on physics for which there was no laboratory evidence and often seemingly at odds with existing astronomical knowledge.
But I believe that there are strong grounds for optimism as well. First, increasingly, the theories proposed are falsifiable in the classic Popperian sense: They make definite predictions which, if tested and found wanting, provide evidence against the theory. Second, the range and depth of observational constraints are growing rapidly. The standards by which theories may, in principle, be measured grow stricter by the day. Three examples of new observational constraints come to mind: First, fluctuations in the CBR (Cosmic Background Radiation) have now been reliably detected on the 10° angular scale by the COBE satellite, thus normalizing all theories on that scale (Smoot et al 1992); second, from spectral information, COBE has set strict limits on the mean Sunyaev-Zeldovich y parameter of the CBR (Smoot et al 1992); and third, better distance indicators have enabled us recently to map out the local velocity field (e.g. Faber et al 1989) to an accuracy of better than 102.5 km s-1. Thus, within the last few years, quite new and important measures of cosmic structure have become available; they constrain the theories in different ways from prior observations and are perhaps less dependent on the uncertain relationship between matter density fluctuations and galaxy density fluctuations than were prior tests.
The other development has been in our rapidly improving ability to quantify what the theories predict. The Cold Dark Matter (CDM) scenario has been justly praised as honestly falsifiable in the sense that it is a quite definite theory with (essentially) only one free parameter, the amplitude of perturbations, the predictions of which should be equally definite and testable. The difficulty has been that, once perturbations have grown past the linear regime, our methods of computation were so poor that we could not really say what the (definite) predictions were and so could not compare theory to reality. This fault is rapidly being remedied as advances in computer hardware and software are beginning to permit three-dimensional simulations with good physics to be made at resolutions approaching adequacy.
The Cold Dark Matter scenario, because it has been the most popular and best studied picture, is the testing ground for these advances. New observations are often made to test its predictions, and new simulations are made for it, in preference to its rivals, to make its predictions more definite. So, paradoxically, if in this review the CDM scenario is found to be wanting in certain essential aspects (although excellent in others), that then is "the good news." It means that cosmology and the origin of large-scale structure are becoming proper quantitative scientific subjects, losing their theological aura, and moving into the realm of quantitative modeling and testing that we are accustomed to in "normal science".