Cowgill (1988) has cautioned that civilizational collapse is a concept fraught with imprecision. It is necessary, he argues, to distinguish between the level at which “collapse” is operating (state, society, or civilizational), whether decline involves wholesale social collapse or merely political fragmentation, and whether decline is rapid or gradual, complete or partial. We would add that arguments should distinguish locally contingent events from broader processual or systemic trends, although the distinction between the two is not always clear-cut.
With respect to the lowland Maya case, one of the difficulties is a terminological one: the Terminal Classic (and increasingly the Early Postclassic) has simply meant too many things to too many people. Even its initial formulations demonstrated an uneasy balance between developmental and chronological senses. In an early definition of the term, the Terminal Classic was
The period during which the processes of the downfall worked out their course. There are some difficulties in choosing a period term by which to designate the Tepeu 3 horizon. Traditionally, there has been a tendency to include the horizon as the last part of the Late Classic Period because of clear ceramic continuities with earlier horizons. On the other hand, most of the patterns that gave rise to the idea of Classicism had ceased by this time.
We have, therefore, adopted the designation “Terminal Classic” in this volume with the hope that it will connote both the continuity and the destruction of previous patterns demonstrated in the archaeological record. (Culbert 1973b: 16-17)
“Terminal Classic” was thus utilized as the Tepeu 3 period designator (thus privileging ceramic data, more specifically the ceramic sequence of the central southern lowlands) and as the period of the “Maya collapse.” The two senses were made compatible by the assumption that it represented a “horizon,” briefly but widely experienced throughout the lowlands. Containment of the events during this time span within a single ceramic period, reinforced by the relatively short span over which hieroglyphic dates ceased to be erected, encouraged their treatment as part of a single, relatively brief process. (Note “downfall” in the singular in the passage above, although Culbert recognized the challenge the Dzibilchaltun data potentially presented.) Culbert’s (1988) more recent consideration is very much in this vein, with virtually no consideration of the northern lowlands.
Data collected over the past twenty-five years suggest the collapse was a protracted and far from monolithic process. Demarest and colleagues argue for the abandonment of the Petexbatun area well before the dates traditionally assigned to the Terminal Classic, while Chichen, Kabah, and perhaps Uxmal seem to have hung on until a. d. 1000-1050, thus opening up nearly three centuries for the “collapse.” Other sites, such as Lamanai, seem to have weathered whatever crises occurred elsewhere. We must therefore be wary of assuming that the failure of particular polities was the result of general systemic processes, rather than of restricted, contingent factors, such as military defeat or local environmental degradation.
Another problem with the concept of the Terminal Classic has been our choice of chronological markers. In many cases, late occupations are defined on the basis of markers of limited areal distribution, even within particular sites, including fine paste wares, particularly Silho (X) Fine Orange and Tohil Plumbate, or in northern Yucatan, Sotuta slate wares. These markers were either imports or produced in limited regions, and were hence dependent upon particular distribution networks. In addition to status and ethnicity, ideology may also have affected the composition of such networks. We have argued, for instance, that Plumbates and fine paste imports, particularly Silho Fine Orange, were distributed chiefly by sites participating in the cult of Quetzalcoatl (Ringle, Gallareta, and Bey 1998: 216-218) and may have been involved in related ritual, feasting, or pilgrimage activities. As such, they may well have been excluded from other markets because of ideological differences. Whether or not this eventually proves true, the restrictions on the distribution of these ceramics render them of limited value in dating individual structures, particularly of the humbler sort. This has resulted in the separation of what were actually contemporaneous deposits, producing rather dramatic changes between Late and Terminal Classic settlement patterns (or Pure and Modified Florescent in the north), whereas in reality the transition may have been much more gradual.
Cobos (Chapter 22, this volume; Winemiller and Cobos 1999) has recently argued for a subdivision in the Sotuta complex. Developing observations by Lin-coin (1990), he argues that Silho Fine Orange and Thin Slate are characteristic of the earlier facet and Plumbate of the later, and that sites such as Ek Balam, Dzibilchaltun, and Coba were contemporaneous only with the earlier facet. During the later period, a. d. 900-1050, Chichen, Uxmal, and perhaps Tiho rose to regional prominence. This framework has much merit, since it is supported also by architectural superimpositions and style changes at Chichen itself and is close to our own thinking about when Ek Balam declined. But again the diagnostic ceramics are specialty wares of limited distribution whose absence can be attributed to a variety of explanations besides chronology. Thus, the absence of Plumbate at Ek Balam is not necessarily confirmatory, since the earlier diagnostics. Fine Orange and Thin Slate, are also in short supply there.
Climatic Change
Environmental crisis and warfare continue to be two of the major systemic explanations for the “Maya collapse.” However, environmental crisis theories emphasizing overpopulation, pressure on agricultural resources, and consequent ecosystem degradation may be faulted for ignoring what would seem to be the far more fragile northern lowlands. Northern Yucatan receives less rainfall and has thinner soils, yet on present evidence seems to have been more densely settled than most areas of the south during the Late Terminal Classic and persisted for an addition 100 to 200 years. One alternative to local overexploitation models has been to identify factors that disrupted otherwise relatively balanced adaptations over wide areas. Several authors (e. g., Folan, Gunn, et al. 1983; Dahlin 1983) have noted the correlation of the end of the Classic period with the onset of the Medieval Climatic Optimum (a. d. 900-1200/1300). Over the past decade additional studies have attempted to verify the relevance of global climatic change using data specific to Yucatan. Most publicized are the results from Lake Chichancanab, southeast of Ek Balam (Hodell, Curtis, and Brenner 1995). Particularly dry conditions were noted in a core at a position judged to date to ca. 1140 B. P., suggesting a dry period lasting from a. d. 800 to 1000 with peak aridity ca. A. D. 922 that the authors cautiously note correlates with the end of the Terminal Classic period.
These authors and their colleagues have also conducted similar but less publicized studies at Lakes Cobd, San Jose Chulchaca, and Sayaucil, in effect providing a transect across the peninsula (Whitmore et al. 1996; Leyden et al. 1994, 1998). These sediment cores do not reflect a similar severe period of desiccation, though human influence is evident, as are what may have been more extreme east-west differences in precipitation. The lack of drought “spikes” are perhaps attributable in part to poor resolution in the upper sections of the cores, but it is also interesting to note that these lakes reflect much less colluvium resulting from agricultural erosion than do sites in the south, such as Lake Peten Itza. On the other hand, data from Punta Laguna, located not far north of Lake Coba, do indicate dry conditions for an extended period, from 1785 to 943 b. p. (Curtis, Hodell, and Brenner 1996), with an end to drought conditions ca. 930 b. p. (ca. a. d. 1100). Peaks were registered at a. d. 862 and 986. Note, however, that this core indicates that dry conditions lasted for most of the Classic period, indicating they must somehow have been generally favorable for social expansion.
All in all, drought seems unlikely to have been the proximate cause of cultural disruption at Ek Balam, which occupies what would seem to be a well-buffered section of the peninsula. The area around Valladolid has a relatively high level of rainfall and some of the highest average maize yields in Yucatan (Duch Gary 1988: 227-228; Ringle 1985: Figure 6.1). Furthermore, rainfall is better distributed throughout the year than in areas farther east. The potable water supply was also well protected, being derived almost exclusively from the water table. Locally this is approximately twenty meters below the surface and can be reached via the numerous cenotes that dot the region (Figure 21.3). Efforts by Houck and project member John McCall located fifty cenotes in the greater Ek Balam region, and more undoubtedly exist. The floors of the deeper sinkholes {rejolladas) that pepper the landscape could also be quarried for a few meters until the water table was reached. Twenty-six dzadzes (partially or seasonally inundated rejolladas) and five wells in sinkholes were found in our mapping. These sinks were almost certainly used agriculturally as well, providing somewhat more humid soil for specialty crops. Such groundwater-based resources might be expected to be more resilient than rainfall or runoff-based systems.
Most crops were dependent on rainfall, however, and so may have felt the effects of increasing aridity while drinking water remained available. Considerable evidence exists for global climatic changes during the Medieval Climatic Optimum. High-resolution data from Quelccaya, Peru, glacier cores (Thompson et al. 1985; Kolata 1993: 284-291) and from Lake Marcacocha, Peru (Chepstow-Lusty et al. 1996), reflect warm, dry conditions after a. d. 1000/1050 to 1400, the Quelccaya data suggesting such activity may have begun as early as a. d. 920950. Comparable evidence exists for California (Stine 1994) and Europe. The tenth-century drought spikes in the studies above may be a local manifestation of these trends. If so, this would suggest climatic factors were most important about a century after many northern centers began to unravel, and may have provided a final blow to remaining population concentrations. Nevertheless, the Yucatan cores suggest that the severity of climatic deterioration may have differed significantly from region to region.
World History
