Theories concerning the structure and history of the whole universe have assumed an increasingly empirical aspect in the 20th century. Beginning in the 1960s, particularly, a combination of new observation techniques, new discoveries, and applications of special and general relativity has resulted in considerable progress. The most important techniques added to those of observations by means of visible light were radio astronomy; infrared, ultraviolet, X-ray, and gamma-ray astronomy from extraterrestrial platforms; cosmic-ray investigations; neutrino astronomy; and examination of the Moon and other astronomical bodies by unmanned and manned space exploration.

Edwin Powell Hubble, a U.S. astronomer, had discovered that the more distant astronomical objects exhibited a shift of spectral lines toward the red (long wavelength) end of the spectrum, the extent of the shift increasing the greater their distance from Earth. This cosmological red shift has been generally interpreted as evidence of rapid recession of these distant objects in an expanding universe. The present rate of expansion is expressed as the amount of recession per unit distance and is known as the Hubble constant. It amounts to about a mile per second recessional velocity for a distance of 10⁵ light-years. Equivalently, if the expansion has been occurring at a constant rate, it must have started about 2 x10¹⁰ years ago.

Quasars, also called quasi-stellar objects (QSO's), appear to be structures that combine extreme luminosity (100 times that of a bright galaxy) with great compactness, taking up less space than the distance between the Sun and its nearest neighbour star. Wherever a spectral analysis of a quasar's emitted light has been possible, the spectral lines have been found considerably red shifted. If these red shifts are cosmological (an interpretation now accepted by most astronomers), some quasars are more distant from the Galaxy than any other known objects. As such they may provide indications of the large-scale structure of the universe, which could not be obtained from investigations confined to "close" surroundings. The term close is to be understood in relation to distance in which Hubble's red shift becomes large ("cosmological distances"), distances amounting to thousands of millions of light-years.

Finally, the term primeval fireball refers to the discovery of an all-pervasive background of radiation whose frequencies lie in the border region between microwave radio frequencies and infrared, corresponding to wavelengths of the order of millimetres and centimetres. In the early 1970s this radiation was interpreted as a remnant of the original intensive radiation that must have been associated with the early history of the universe, when matter was both extremely dense and extremely hot; hence its name. Its spectral composition, which has been the object of intensive investigation, might provide some clues to the early history of the universe.

General relativity contributes to a theoretical discussion of cosmology the idea that the universe as a whole need not be flat even on the average and that it probably is not. Even if one were to assume that on a very large (cosmological) scale the universe is homogeneous and isotropic (i.e., having the same properties in all directions), which appears a reasonable working hypothesis in the absence of any evidence to the contrary, there are a number of different possibilities. The universe might be spatially open (as a flat one surely is), or it might be closed, somewhat as the surface of a sphere is closed, without boundaries. Likewise, in the time direction the universe might be either open or closed; it is a little difficult to visualize a time-wise closed universe, which appears to be inconsistent with ordinary notions of cause and effect. But because these notions are distilled out of normal experience, they might be inapplicable on the scale of billions of years. In brief, many different cosmological models have been proposed and investigated theoretically, but observational information does not seem to favour one particular type. The information appears to favour types that expand from an early stage involving fireball conditions.

An outgrowth of a unified field theory of the early 1920s has been the development of a class of theories based on the hypothesis that underlying the four-dimensional space-time of our experience is a manifold having a higher dimensionality, whose geometric structure can accommodate all known force fields, including those associated with stable and unstable subatomic particles. Though these concepts remained highly speculative, they offered much promise and occupied many investigators.

Apart from the attempts to devise unified field theories, several modifications of general relativity have been proposed during the late 20th century. One of these was presented by the British scientist Fred Hoyle, whose results, together with the proposals of the astronomers Hermann Bondi and Thomas Gold, became the basis of the so-called steady-state cosmological theory. Bondi, Gold, and Hoyle opposed the "big-bang" theory of the origin of the universe, arguing instead that matter is being created continuously at a very low rate, just sufficient to maintain the constant average density of the universe in spite of the observed expansion. Though the steady-state hypothesis evoked much interest for some years, the existence of the cosmic background radiation (established in the 1960s) has been generally accepted as proof that the universe has in fact passed through a highly dense stage.