So what happened here?  There are two main types of explanations, environmental and social, and their relative popularity has varied over the years.  The environmental explanation depends largely on the striking coincidence of the abandonment of Mesa Verde with the so-called “Great Drought” of AD 1276 to 1299, one of the earliest major climatic events to be identified in the tree-ring record.  The near-perfect alignment of the drought with the final abandonment of the area is indeed remarkable, and this explanation has been pretty popular and remains so today, but there is considerable evidence that there was more going on.  For one thing, while all of the Southwest is in some sense marginal for agriculture, within that context the northern San Juan is one of the most productive and reliable agricultural areas.  Indeed, much of southwestern Colorado is used today for commercial farming, largely using dry-farming methods not all that different from those used in antiquity.  The Mesa Verde area gets plenty of rainfall, and while a short growing season can be an issue at the higher elevations, throughout most of the region it is not generally problematic.  Models of agricultural potential based on tree-ring data have generally shown that the carrying capacity of the Northern San Juan greatly exceeded any plausible estimate of its overall population throughout prehistory, although that doesn’t necessarily mean that individual communities would always be able to support themselves on the land they happened to have.

Social factors, then, are probably involved along with the drought.  There is definite evidence for a considerable amount of violence during the thirteenth century in this area, and settlement patterns become increasingly defensive over time.  There is also an increasing diversity in public architecture among the various communities, suggesting that traditional religious or ideological structures may have been breaking down and being replaced by new ones.  A strong tendency toward settlement aggregation, perhaps due to defensive considerations, may have played a role in these religious trends.  Furthermore, all of this may have been influenced or set in motion by deteriorating environmental conditions; environmental and social factors were not necessarily separate things.

So where did the people go?  The general assumption is that they mostly went to the northern Rio Grande Valley, which sees a remarkable increase in its population right around AD 1300, just as Mesa Verde is emptying out.  This is a bit problematic, however, since there is relatively little evidence for people with obvious Mesa Verde cultural traits showing up in the Rio Grande at this time.  This may be because people were emigrating away from Mesa Verde in small groups and assimilating into existing Rio Grande communities, or it may have been because people were changing their cultures as they moved, perhaps abandoning the old social institutions that had been ineffective in preventing the abandonment and adopting new ones that seemed to work better.  It’s hard to say, really, and this is a topic of ongoing research.  One interesting effort recently has been the Village Ecodynamics Project by Washington State University and Crow Canyon Archaeological Center, which has used agent-based modeling and other innovative techniques to try to understand the culture history of the Northern San Juan.

Okay, so that’s more or less what happened.  Does it count as a “collapse”?  Let’s look back at Jared Diamond’s criteria for collapse:

1. Collapse involves a major loss of population.
2. Collapse involves a loss of complexity.
3. Collapse occurs over a large geographic area.
4. The changes brought about by collapse persist for a long time.

In this case 1, 3, and 4 are pretty obvious.  The Mesa Verde region was totally depopulated, which is about as major a loss of population as you can get.  It’s also very large, and the changes that resulted from the abandonment of the region and the influx of population to the Rio Grande have persisted to the present day; many aspects of Mesa Verde culture notable in the archaeological record were not brought to the Rio Grande, and are not present in the modern Pueblos there.  Since Diamond apparently considers only one of the first two criteria to be necessary, he probably would consider this a collapse, but most other collapse theorists, including Joseph Tainter, consider loss of complexity to be a more important consideration than loss of population, so let’s look at complexity in the Mesa Verde case.

There basically isn’t any evidence for significantly complexity in Mesa Verde before its abandonment or in the Rio Grande afterwards.  Unlike the Chaco case, the villages in the thirteenth-century Northern San Juan seem to have had relatively egalitarian social structures, at least economically and probably politically as well.  This is not to say that there were definitely no disparities in political power, but that they were likely masked and subverted by an egalitarian ideology that prevented massive accumulation of wealth and power.  This is the case in the modern Pueblos, where despite some possible inequalities in power and political influence among different clans or societies the overall ideology has enforced a general economic equality.  This seems to have been the case in the thirteenth-century Northern San Juan as well, and it could have been in part a reaction to the perceived excesses of the Chacoan era, although it’s noteworthy that a certain amount of Chacoan influence seems to have persisted, especially in the Totah area around Aztec, albeit without much evidence for the economic inequality that marked the Chacoan era itself.

Mesa Verde doesn’t get mentioned in the collapse literature as much as Chaco, although sometimes the two are kind of muddled together incoherently, and for good reason.  What we seem to be seeing at Mesa Verde is a period of societal difficulty that resulted in depopulation and migration, a common pattern in Southwestern prehistory.  While there were some changes in society during the abandonment and migration that make it difficult to tell exactly where the people ended up, these changes don’t seem to have been related to any change in the overall complexity of the society, which remained about as complex as it had been before.

The human occupation of Chaco Canyon goes back a very long way, but the key developments that made it an important regional center seem to have begun in the AD 800s with the initial building of a few “great houses,” which in that period were large masonry structures similar in layout and construction techniques to the “small houses” in which most Southwestern people lived at the time but much on a much larger scale.  These early great houses, including Pueblo Bonito and Una Vida, show considerable signs of residential use in their earliest parts, and it seems that they were at least initially residential structures.  It’s not at all clear what inspired their construction, but there were similar structures being built in other parts of the region at the time, so Chaco may not have been particularly special at first.  Over the course of the next hundred years, however, something seems to have happened to make Chaco a major regional center, and starting around AD 1030 a building boom in the canyon in which the existing great houses were expanded using much more elaborate techniques and an even larger scale of construction coincided with the construction of entirely new great houses both in the canyon and throughout the San Juan Basin of northwestern New Mexico and beyond into Colorado, Utah, and Arizona.  These “outlying” great houses were mostly placed in existing small house communities, which continued to be occupied, and were connected to the canyon via an elaborate road system.

Over the next hundred years, construction both inside and outside the canyon continued almost without pause, and at the same time a vast amount of material of all kinds was brought into Chaco from a vast surrounding region: turquoise, shell, copper bells, macaws, and other exotic materials, as well as more quotidian items such as pottery, construction timbers, and corn.  Then, around AD 1130, everything seems to have come to an abrupt halt.  Construction of great houses, after a shift around AD 1100 toward a different type of architecture, seems to stop entirely by around 1125, and activity in the canyon slowed to a crawl at that point.  There was probably at least a small population remaining until the depopulation of the whole region in the late 1200s, but it was nowhere near as large as the apparent population at the system’s height.  It is this decline in activity that collapse theorists seek to explain when they look at Chaco as a case study.

So what happened?  There are various theories out there.  Many point to a prolonged period of drought from around 1130 to 1180, which coincides suspiciously closely with the end of major activity at Chaco, as having somehow led to the collapse, although this explanation is somewhat problematic given that earlier droughts, especially a short but severe one in the 1090s, didn’t have nearly the same effects on the system.  Others argue that political, social, or economic instability within the Chaco system itself, whatever its nature, was the main cause of the collapse, with drought perhaps playing a subsidiary role.  Most people agree, however, that Chaco is indeed an example of societal collapse.

But is it?  Let’s look at some of the criteria for defining collapse, using Diamond’s list:

1. Collapse involves a major loss of population.
2. Collapse involves a loss of complexity.
3. Collapse occurs over a large geographic area.
4. The changes brought about by collapse persist for a long time.

Measuring the population of Chaco at any time is surprisingly difficult, but given the much lower level of activity after 1130 I think it’s safe to say that there was a major decline of some sort.  The extent to which Chaco was a complex society at all is disputed, but I find the arguments for complexity more convincing than the arguments against it, so let’s take relative complexity as a starting point and see if there’s evidence for a loss of it.  Recall Ben Nelson’s definition of complexity:

Social systems are considered complex if they are comparatively large demographically and spatially, encompass multiple settlements in an integrated political structure, and exhibit horizontal and vertical social differentiation. Other properties associated with complexity are hereditary ranking, production of surplus and its appropriation by an elite, craft specialization, and long-distance exchange.

Large demographic scale is basically the same as population, so that one’s covered.  We’ll get back to spatial scale and settlement pattern later.  There isn’t much evidence for horizontal social differentiation at any point in the Chacoan archaeological record, but the vertical differentiation implies by the elaborate burials in Pueblo Bonito does seem to end around 1130.  Ranking goes along with vertical differentiation, and surplus and its appropriation are controversial and hard to find in the archaeological record, as is craft specialization.  That leaves us with long-distance exchange, which does continue to go on at Chaco, but at a much lower level than before.  So yes, I think it’s fair to say that Chaco became less complex according to most of the criteria that can be used to assess complexity there.

That brings us back to spatial scale, and here’s where things get tricky.  It turns out that the evidence for reduced activity at Chaco Canyon after 1130 doesn’t correspond to a similar reduction in activity in most other parts of the Chaco system at the same time.  Indeed, some areas, such as Aztec Ruins on the Animas River to the north, see a marked increase in activity after 1130, and both Aztec and the Mesa Verde area further north see continued activity on a large scale, indicative of a large population, until the depopulation of the whole area in the late 1200s.  The area to the west doesn’t see such dramatic growth, but it does seem to keep on going without much change after 1130.  Similarly, while the area immediately south of Chaco seems to have been largely depopulated even earlier than the canyon itself, the area further south continued to see activity long after, indeed up to the present day at Zuni Pueblo.  And in many of these areas, especially at Aztec and at the northern and southern extremes of the original Chaco system, the outlying Chacoan great houses seem to have continued to be used, though perhaps not the same way as they were originally intended to be used, long after the cessation of great house construction in Chaco itself.

So it seems that the Chaco “collapse” really only applies to a single location, Chaco Canyon itself, and not to the society as a whole.  Indeed, some archaeologists have interpreted these data as showing not so much the collapse of the system centered on Chaco but a series of changes in it, possibly including a shift in emphasis away from Chaco itself toward Aztec, which replaced it as the center of the system.  Whether or not some form of the system that developed at Chaco continued at Aztec, it’s clear that there were a lot of changes going on in the region during the 1100s, including an apparent movement of population away from Chaco, probably at least in part to Aztec and Mesa Verde.  The lack of continued construction on the scale seen from 1030 to 1130 and the reduced level of trade do seem to suggest that the Chacoan system declined in power and influence after 1130 whether or not it moved to Aztec, but there turns out to be very little evidence of a “collapse” occurring over a large spatial scale, although the changes do seem to persist for a long time.

So what are the implications of this for studies of collapse in general?  It’s hard to say, but I think one lesson is that it’s important to look at these things on the societal level rather than on individual sites or localities, no matter how important or central they seem.  Some Southwestern archaeologists now prefer the term “reorganization” to “collapse” for situations like the changes at Chaco after 1130 and the contemporaneous events in the Mimbres Valley of southwestern New Mexico.  It’s certainly quit different from the massive depopulation of the whole Four Corners region in the late 1200s, which however doesn’t fit well into “collapse” models either because there’s little evidence of a system on any level larger than the individual community during this period, with the possible exception of a rump Chacoan system operating on a small scale out of Aztec.  That event, which corresponds to another prolonged drought, is of interest in its own right, but this post is long enough already.

The accompanying web-only piece illustrates how Navajo oral history deeply shapes the views and beliefs of Taft Blackhorse, a Navajo archaeologist who I spent time with while reporting on these stories. I will say that I grew quite fond of Taft and his colleague, John Stein. They were generous hosts and there’s a part of me rooting for them to continue their maverick ways and quixotic quest. That said, I have no doubt that many archaeologists will be shaking their heads in disbelief at some of the statements they make.

Combined, the two stories reveal an interesting dilemma for archaeologists who strive to reconcile data-driven science with information gleaned from a culture’s oral tradition.

I’ll have more to say on all this shortly, as I suspect others will offer their own commentary, some who I know have already read the print story. I look forward to a spirited exchange.

One final thought: while writing these stories, I was reminded of something I once read in an essay by geographer D.W. Meinig, in this classic book:

Any landscape is composed not only of what lies before our eyes, but what lies in our heads.

In Girl With Skirt of Stars, Jennifer Kitchell draws a sharp contrast between modern society and a culture that has occupied the southwest of North America for thousands of years.

The Reslience ecologists, who I believe have much to contribute to the sustainability debate, might also do well to question the resilience of American Indian romanticization.  We’ve already been down that road with the Anasazi.

All the worldwide headlines on this latest story about Angkor, the ancient Cambodian city, mention drought. And for good reason. As the AP reports, new tree ring evidence by scientists show

that Southeast Asia was hit by a severe and prolonged drought from 1415 until 1439, coinciding with the period during which many archeologists believe Angkor collapsed.

But as this previous research published in 2007 suggests, population pressure, deforestation and soil erosion had already started to stress the sprawling settlement.

Then there is the 1431 invasion of Angkor from Siam (now Thailand) to keep in mind.

Put it all together and you have, as one scientist interviewed in the current AP story explains, a knockout blow delivered by climate change:

We have these droughts occurring on top of preexisting pressures…It’s like pouring petrol on a fire. It makes social and economic pressures that may have been endurable disastrous.

American archaeologists studying the social chaos and eventual depopulation of the Four Corners region in the Southwest during the 13th century are often reluctant to put too much emphasis on environmental factors–despite evidence of similar mega-droughts.

But there seems to be an emerging pattern to the rise and fall of the Anasazi and Angkor, and other famous examples, such as  the ancient Maya, that is worth paying attention to today, given our current ecological and climate challenges.