upon historical types, and so the two methods were synergistic.

Chronology was not only the product of the application of the new ideas, but also the limit to the application of these new ideas. And only those who actually constructed chronologies (a minority of practitioners, most of whom simply used or tinkered with preexisting chronologies) seem to have participated in or understood this revolution. But being able to relatively date assemblages of objects had important ramifications in almost all archaeological endeavors. Geographic patterns in artifact similarities had been noted since the 1880s (e.g., Holmes 1886b). Broad patterns and the absence of chronology tended to reinforce ethnological interpretations (e.g., Souixan, Algonkin, Iroquoian, Muscogean) of such differences. With the advent of the historical type, independent archaeological connections could be identified, as Frank Setzler (1933) did with “Hopewell” designs over much of the East. So-called whole cultural units also benefited with William McKern’s (1939) Midwestern Taxonomic Method, in which ethnic assignments were replaced by phenetic similarity and a hierarchic ranking, a primitive numerical taxonomy well ahead of its time. McKern’s approach was strictly phenetic, however, so as chronologies began to become widely available over the next decade, the approach was gradually replaced by nonhierachial units that had distinct temporal connotations, such as gordon willey and Philip Phillips’s (1958) phase. These developments led very quickly to the sine qua non of culture-history—the time-space chart/ culture historical synthesis (e.g., Setzler [1940] in the upper Midwest; Griffin [1946] for the Northeast; Ford and Willey [1941] in the Southeast; Kidder [1924] in the Southwest; and Martin, Quimby, and Collier [1947] for the whole United States and Canada), culminating in Willey’s two-volume synthesis entitled An Introduction to American Archaeology (1966, 1971) as well as many other less ambitious syntheses.

The crucial thing to notice about this culture-historical success is that it was won not by overt theory construction but by accident, and it was saved only by the use of chronological methods, the results of which were empirically testable. Culture-historians displayed no interest in why the methods worked; their only concern was that they did work. Consequently, when pressed, culture-historians had to rely upon arguments of “discovery” and “reality,” things that were patently not true when the results were explicitly the product of archaeological creation (compare Rouse 1939 with Rouse 1960). This tendency was reinforced by the success of culture-history. Culture-historical classifications became the way to describe the record, even if you were criticizing culture-history (e.g., Spaulding 1960a). Even in matters of chronology, materialism was only skin-deep. To make the archaeological record intelligible in the absence of theory, culture-historians relied on common sense. This meant that they had to describe the record in everyday terms—a warrant for “reconstruction.” It also meant that their chronologies, or orders of assemblages, had to be chopped up into groups that could be treated as a kind (essentialism), the archaeological period. Phases were inconceivable without this transformation. The quantitative revolution was just as tenuous. Kroeber (1940), for example, had pointed out that there were quantitative methods for assessing similarity and thus creating units in the Midwestern Taxonomic Method, but these admonitions fell on deaf ears. In short, the chronological methods won by trial and error in the first quarter of the century were applied to create an organization for intuitive, common-sense-driven reconstructions of past lifeways that were no more testable than those of their predecessors, so much so that many preferred to omit the reconstruction in favor of the framework alone. In fact, little of methodological or theoretical significance happened between the first Pecos conference in 1927 and the mid-1900s. To be sure, there were occasional rumblings and grumblings (e.g., Steward and Setzler 1938; Spaulding 1953; Taylor 1948) and alternative models from Europe (Childe 1936; Tallgren 1937) that hindsight shows to have foreshadowed what was to come. By and large, however, this was a period of normal science (Kuhn 1962) during which the goals, methods, and language of observation