Glenn Reynolds links to Robert Bruegmann's Sprawl : A Compact History, commenting, "A very interesting book, reporting that people have been worrying about "sprawl" for centuries, and that most efforts to reduce it make things worse . . ."
Amazon.com's book description also has this:
Robert Bruegmann calls [sprawl] a logical consequence of economic growth and the democratization of society, with benefits that urban planners have failed to recognize . . . Taking a long view of urban development, he demonstrates that sprawl is neither recent nor particularly American but as old as cities themselves, just as characteristic of ancient Rome and eighteenth-century Paris as it is of Atlanta or Los Angeles.A few thoughts . . .
Frederick Law Olmsted, the 19th century's leading urban designer, thought that sprawl would be a great thing, and accurately (and optimistically) foresaw the explosive expansion of single-family residential suburbs in America. Like Bruegmann, Olmsted thought they were a democratizing and healthy development, because they provided light, air, and wholesome public places for Americans' moral development.
Olmsted did not, however, desire subdivisions lacking sidewalks, landscaped rights-of-way, large public parks, and greenways, and he did not foresee the rise of the automobile and its often desolating effects on urban spaces and design.
At any rate, the fact that Olmsted anticipated this kind of growth means that it did not then exist in the form that it does now, though some American cities were quite mature by the end of the 19th century.
As to Bruegmann's take on ancient Rome . . . I'm skeptical. The Roman upper classes certainly escaped to their rural villas whenever they could (you can read wonderful descriptions of some of the most lavish in Varro's De Agri Cultura), but I'm wary of drawing any facile analogies between ancient patterns of land use and what's going on now, if only because the Romans didn't have cars.
It's important to keep in mind that the kind of urban density which modern urban planners such as the New Urbanists promote is nothing like that which was found in ancient Rome, Victorian Liverpool, or modern Manhattan. Often it's only slightly more dense than the sort of suburbs Olmsted liked, and modern planners tend to encourage a mix of compatible uses (residences, small retail, offices) mainly to increase convenience and a sense of place for the people who live there.