In the 4th. Century BCE, Epicurus taught that there are an infinite number of worlds like (and unlike) ours, while Aristotle taught that there is only one. Neither hypothesis can currently be falsified, and indeed we may see the continuation of this metaphysical battle in the so-called inflationary cosmological models.
Philosophers since Anaximander (Kahn (1994)) have long debated the true nature of the Universe, presenting often remarkably prescient ideas notwithstanding the lack of any real data. Given the lack of data, the only basis for constructing a Universe was symmetry and simplicity or some more profound cosmological principle.
The ancients saw nested crystalline spheres fitting neatly into one another: this was a part of the then culture of thinking of mathematics (i.e. geometry in those days) as being somehow a fundamental part of nature 2. Later thinkers such as Swedenborg, Kant and Descartes envisioned hierarchies of nested whirls. While these ideas generally exploited the scientific trends and notions of their time, none of them were formulated in terms of physics. Many are reviewed in Jones (1976) where detailed references to the classical works are given.
Perhaps the first detailed presentation of cosmogonic ideas in the modern vein was due to Poincaré in his Leçons sur les Hypothèses Cosmogoniques (Poincaré, 1894), some of which was to be echoed by Jeans in his texts on Astronomy and Cosmogony (Jeans, 1928). Jeans' work is said to have had a profound effect on Hubble's own thoughts about galaxy evolution and structure formation (Christianson, 1995).
2 Einstein's great intellectual coup was to geometrize the force of gravity: we are governed on large scales by the geometry of space-time manifesting itself as the force of gravity. Back.