development. Furthermore, he identified most of the parameters now routinely used in ceramic analysis. His most enduring result, the geographic groups of ceramics (Holmes 1903), was largely a fluke, however, being accidental combinations of geography and the chronology of the custom of including vessels in graves. The first lasting accomplishment of U.S. archaeologists, beyond the recognition and accumulation of samples of the record itself, arose at this time as well—the appreciation and understanding of the technology of U.S. artifacts. Here, too, Holmes was a leader, as he “figured out” (via intuitive combinations of engineering, physics, and ethnographic accounts) how virtually all major classes of American artifacts were manufactured (e.g., 1886a, 1886b, 1890, 1897).

The First Paradigm: Culture History—Science at Last

Archaeology as an ad hoc assemblage of common sense and analogies with ethnographic observation was rather limited in its prospects despite the accumulation of descriptions and objects, even if they were systematic. Although trained as an artist, Holmes was a committed scientist in archaeological matters (Meltzer and Dunnell 1992), and his “philosophical” approach, a step in the right direction, still fell short of qualifying as a general theory. His guiding principles were clearly evolutionary, but, as is so often the case in archaeology, his evolution muddled scientific evolution (e.g., Holmes 1892) and with a more Spenserian evaluation, which was actually influenced by lewis henry morgan et al. (1877) progress-driven framework, little differentiated from Victorian optimism. As crucial as his work has proved to be, Holmes himself lived to see (but never really participated in) archaeology’s first and closest approach to becoming science, his lifelong goal.

Three developments proved fundamental to developing a discipline-wide consensus on method, language, and, most important, problem—in short, what Thomas Kuhn (1962) characterized as a paradigm. This consensus has since become known as culture history. The first element was the development of a language of observation and an approach to the classification of artifacts. Although critically important, this revolution was confined largely to potsherds, with minor extensions to lithics (stone tools) and bone tools, and it remained restricted to the scale of the portable, discrete object. The second element was adopting a materialist ontology. Like classification, this innovation was far from general in application, being restricted to a narrow methodological role. The final critical ingredient was quantification. Prior to this time, although objects, monuments, and mounds might be enumerated, there was no effort to generate archaeologically meaningful numbers. This new effort entailed wholesale changes in the ways in which archaeologists generated data. What made it all come together as a consensus was that it worked—and for the first time archaeologists could produce conclusions that could be tested empirically like those of other scientists, and they gained a measure of respect that had thus far eluded them.

The so-called stratigraphic revolution (Willey and Sabloff 1980, 84–93) had little to do with stratigraphy per se. It was not an epiphany on the relevance of the geological notion of superposition; U.S. archaeologists from Jefferson on had routinely understood and made stratigraphic observations. Rather, it depended on the coincidence, largely accidental, of the three innovations just named: quantification, materialism, and classification. The success of Europeans using stratigraphy (e.g., de Mortillet 1883; Lubbock 1865; Worsaae 1848) to create archaeological chronologies was not unnoticed by U.S. Archaeologists. They specifically sought out the kinds of deposits that had been key to the European success (especially caves and rock shelters—e.g., Nelson 1917). Even when expanded to include such obviously stratified deposits as shell middens (e.g., Dall 1877; Nelson 1910; Uhle 1907; Wyman 1875), these efforts were unrewarded because they continued to look at the record in qualitative, essentialist terms. European sequences, because of the vastly greater amounts of time involved, yielded coarse chronological kinds or “periods” under such assumptions but not so the much shorter U.S. record. Repeatedly, U.S. archaeologists would observe strata but then conclude they