Classic Maya Urban Demography

Two factors conspire right now to emphasize to me the importance of ancient Maya urban demography. First, I finished reading David Webster’s new book, The Population of Tikal (Webster 2018). Second, today I returned from a Working Group of the Social Reactors Project at the Santa Fe Institute , where our focus was on Maya cities and settlements. Needless to say, we talked a lot about population.

One of the most important things one can know about a city or settlement is its size. How many people lived there? Over what area? (and, consequently, What was the density and how did it vary?) These are fundamental observations in many disciplines that study cities and settlements, from urban history to urban economics, to urban geography. In chapter 3 of my current book, Urban Life in the Distant Past, I make the case for the importance of knowing the sizes of cities in the past (Smith n.d.).

Webster’s book reviews past population estimates for Tikal (and other Maya sites and regions). These are judged to have been too high, largely on the basis of comparative studies of regional population densities of early states in other parts of the world. “By overestimating Classic Maya populations, we have created problems that do not exist” (p. 3):

1.       Why were Maya populations so large and dense compared to other early states

2.       How could the landscape support so many people,

3.       How can we explain the huge population drop with the Classic Maya collapse?

But wait, how can you compare an urban density at Tikal to a regional density for other regions? Isn’t that a comparison of apples and oranges? Those familiar with David Webster’s past work will get this right away: The Maya did not have cities, and therefore the concept of urban population density is moot!  “I prefer not to call any part of this landscape a ‘city’.” (p. 35). OK, maybe instead of “urban” we should call inner Tikal, “The area close to the pyramids that had a higher density than the outlying areas.” But this would still cast doubt on many of Webster’s comparisons.

Another problem I have with the new book is that much of the quantitative data are presented within the text, and not in tables. Yes, there are a few data tables with some pertinent information. But I get confused, for example, when he gives a population estimate for Tikal of 10,000 (p. 52, paragraph 2), an area estimate of 452 square km (p.52, paragraph 7), and an overall density of “100-175 people per sq. km” (same paragraph). Huh? 10,000 people divided by 452 sq km is a density of 22.1 persons per sq km. One of these figures must be off.

I must admit that this is a pet peeve: If you use quantitative data, put them in a table! I struggled throughout the book with this. Yes, there are a couple of tables with data. But where making detailed comparisons and analyses of population and density data throughout the book, the figures are only found in the text, not in tables. Putting quantitative data into the text but not in a table makes it difficult for readers to re-use or analyze the data.

I’ll stop here, this is not a book review. The book is fun to read, like most of David Webster’s work, and I learned a lot. But the basic quantitative treatment, and the lack of a good conceptual framework for urbanism, or for regional settlement variation, limits the value of the book.

Our working group on Maya settlement included my colleagues in the Social Reactors Project, plus Jerry Sabloff, Bernadette Cap, Adrian Chase, Julie Hoggarth, and Heather Richards-Rissetto. Sarah Klassen, who works on urban issues at Angkor, also participated. See a short description of the session here.

Instead of giving a report of our workshop on Maya settlement, I’ll just mention a few key points.

First, we have found data on several regional settlement systems where area scales with population in a superlinear fashion. That is, larger sites are LESS DENSE than smaller sites; This is the opposite of all other agricultural societies we have studied (see our papers here). So, Maya settlement differs from other agricultural societies in ways that remain hard to pin down. We have a paper in the works, and plans for more collaborations and research in the future.

Second, some Maya settlements are very much like settlements in other areas, and some are very different. Not too surprising, I guess, but we discussed some of the patterns.

Third, Angkor looks very different from the Maya cities. It had a very dense urban core, with planned orthogonal neighborhoods surrounding major temples, surrounded by a large sprawling area of temples, houses, and reservoirs, In comparison, Maya urban cores seem only slightly higher in density than their outer neighborhoods. I’m starting to wonder about the usefulness of Roland Fletcher’s (2009; 2012) concept of “low-density urbanism” as a category that includes the Classic Maya and Angkor.

Fourth, we discussed the use of LiDAR for analyzing the sizes and layouts of Maya cities.

We spent some time comparing Tikal and Caracol. My view is that Caracol was a large, integrated urban center, while Tikal was a much smaller city of ca. 16 sq km. (I know the Chases will agree about Caracol, but I’m not sure of their views of Tikal). A key question – indeed, one of our targeted themes at the working group – is the nature of settlement boundaries. I don’t see any evidence for clear boundaries at Tikal. However, there does seem to be a drop-off in density beyond the central 16 sq km mapped by Carr and Hazard, which favors bounding the city there (see table).

Whereas Caracol has various features that seem to signal a spatial and social integration over a large area (causeways, replicated patterns of civic architecture pointing to the integration of different areas, and a continuous stretch of agricultural terracing). Outside of the core area, Tikal lacks these things. All those areas beyond several km from downtown Tikal – these must have been rural areas, the hinterland of the city of Tikal. This is a provisional analysis. I want to analyze the density patterns objectively (kernel density analysis), to see whether there is indeed a zone of declining density (well, more precisely, I want a student to do this…). We should do such analyses for all of the Maya (and other!) sites that are well mapped. Intra-urban density distributions are a big unknown for ancient cities, but we now have the data and methods to work on this question.

This  table shows a bit of density data. I was having trouble finding the information, buried in Webster’s text, so I got the Maya figures from a nice table in a paper by Barbara Stark (Stark 2014). Note the drop of 50% in density beyond the central 16 sq km. Teotihuacan figures are from a paper now under review (Smith, et al. n.d.).

I think it would help if archaeologists could agree on a series of zones over which to calculate population density. Density is, in fact, a very complex phenomenon for modern cities, with lots of relevant measures (Campoli, and MacLean 2007; Dovey, and Pafka 2014). But for archaeology, it is best to concentrate on straight population density (persons per hectare), calculated over one of several spatial units (depending on available data and on research questions). I suggest five relevant units, as follows:

1.   Regional density, 1: Large area. For a large zone including several urban centers. E.g., the southern Maya lowlands. Useful for large-scale comparisons (of the type Webster makes with other regions).

2.   Regional density, 2: Urban hinterland. Density over the hinterland of a city. E.g., Copan Valley. Useful to compare settlement in, say, the Copan Valley with the Belize River Valley.

3.   Urban density: whole city (all of Tikal or Caracol, or Teotihuacan). This is probably the most useful measure for urban comparisons. Of course, one must first identify the boundaries of the urban settlement….. We discussed this at SFI, and I am working on some methodological guidelines.

4.   Epicenter density: The population density of residences clustered around the ceremonial core. Not sure if this should include the civic areas or not. I am less confident in this unit, compared to the others.

5.   Neighborhood density: Density of a delimited neighborhood or zone. This is useful for several reasons: (1) To compare neighborhoods within a city; (2) To generate density data where whole-site information is not available. For example, consider Mohenjo-daro. Several neighborhoods have been completely excavated, but the residential patterns of the entire site are not clear.

If you were looking for solid data or conclusions about Maya urban demography here, I am sorry to disappoint you. This field is in its infancy. Culbert and Rice (Culbert, and Rice 1990) was a crucial work, but there has been very little work since then. Webster’s new book is a step in the right direction, but we need many more such analyses (with lots of data in tables!). But I am more optimistic now than I’ve been for some time now. The session at SFI made some progress, and I see some good research coming out of these and other archaeologists in the near future. I am itching to work on settlement delineation at a variety of sites, using quantitative spatial methods. And, finally, I know that a number of prominent Mayanists agree that urban demography is important, and I look forward to more work in the near future.

References

 Campoli, Julie and Alex S. MacLean  (2007)  Visualizing Density. Lincoln Institute of Land Policy, Cambridge, MA.

Culbert, T. Patrick and Don S. Rice (editors)  (1990)  Precolombian Population History in the Maya Lowlands. University of New Mexico Press, Albuquerque.

Dovey, Kim and Elek Pafka  (2014)  The Urban Density Assemblage: Modelling Multiple Measures. Urban Design International 19:66-76.

 Fletcher, Roland  (2009)  Low-Density, Agrarian-Based Urbanism: A Comparative View. Insights (University of Durham) 2:article 4.

Fletcher, Roland  (2012)  Low-Density, Agrarian-Based Urbanism: Scale, Power and Ecology. In The Comparative Archaeology of Complex Societies, edited by Michael E. Smith, pp. 285-320. Cambridge University Press, New York.

Smith, Michael E.  (n.d.)  Urban Life in the Distant Past: Archaeology and Comparative Urbanism. (book in progress).

 Smith, Michael E., Abhishek Chatterjee, Sierra Stewart, Angela Huster and Marion Forest  (n.d.)  Apartment compounds, households, and population at Teotihuacan  (paper under review).

Stark, Barbara L.  (2014)  Ancient Open Space, Gardens, and Parks: A Comparative Discussion of Mesoamerican Urbanism. In Making Ancient Cities: Space and Place in Early Urban Societies, edited by Kevin D. Fisher and Andy Creekmore, pp. 370-406. University Press of Colorado, Boulder.

 Webster, David  (2018)  The Population of Tikal: Implications for Maya Demography. Archaeopress, Oxford.

Urban toilet at Mohenjo-daro

Did ancient cities have a middle class?

The middle clas

Last night I watched a National Geographic Special on PBS, "Quest for the Lost Maya," Like most of these shows, it combined interesting research material, first-hand images of fieldwork, overly dramatic music and prose, and silly media-speak. Here I want to explore briefly one aspect of the show, the claim that archaeologists (George Bey and William Ringle) have now identified the Maya "middle class." As quoted from the show's website (link above), "Could this be evidence of the Americas' first ever middle class, emerging a full a millennium before 1776?" My answer is "no."

The Maya middle class ??

This is not the first time that Maya archaeologists have claimed to find the Maya middle class; Arlen and Diane Chase made this claim (in print) for the Classic period city of Caracol some time ago (Chase and Chase 1992; 1996).

I am deeply skeptical of such claims, for two reasons. First, comparative analysis of early cities and states by many scholars has shown that in nearly all cases, these societies had a two-class system. Classes are categories of people who have a specific common relationship to wealth, power, and resources (or, in Karl Marx's terms, a specific relationship to the means of production). In ancient urban societies, the elites who controlled wealth and power typically made up somewhere between 5% and 10% of the overall population, and everyone else were commoners. Each of the two social classes contained variation in wealth and power, of course (e.g., slaves, serfs, free commoners, wealthy merchants among the commoner class). Here are some of the authors who provide the data and theoretical concepts to support the two-class model Trigger (2003), Sjoberg (1960), Williamson (2010).

My second objection to the "Maya middle class" argument is that the concept of "middle class" has a specific meaning in comparative social history. This is a class that arose in the context of medieval cities as an urban, commercial class differentiated from the peasants and the nobility by the creation of a new category of wealth: income from commerce. These men were the burghers, and the new commercial middle class was (and is) called the bourgeoisie. As stated by the great historian Henri Pirenne,

·      “Never before had there existed, it seems, a class of men so specifically and strictly urban as was the medieval bourgeoisie.”(Pirenne 1925:132). (see also Kocka 1995).

I doubt any of the Maya experts would claim that their "Maya middle class" was an urban commercial class, equivalent to the medieval Burghers. No one ever claimed for a Maya city that "city air makes one free" (an expression for medieval cities). Medieval European cities represented a radical break with other pre-modern cities in their commercial role, their legal independence from the king, and growing power of a new social class (the bourgeoisie). In these and other traits, the medieval European city  is not a good model for other ancient or premodern cities around the world, such as the Maya.

The Maya archaeologists base their claims for a middle class on evidence that some commoners were better off economically than others. The Chases use burial data (some tombs have fancier offerings than others, but they do not seem to be elite burials), and Bey and Ringle use evidence from houses and architecture at the site of Kiuic. I agree completely with them that uncovering evidence for prosperous commoners much earlier than anyone had expected (ca. 500 BC) is an exciting find. But this is not an example of the "middle class" as typically used by scholars outside the realm of Maya archaeology.

So, why claim to have found a Maya "middle class" - why not just say that some Maya commoners were prosperous, or that the commoner social class contained a lot of variation in wealth? (this is the way I see the evidence). I offer three possible explanations:

1. Perhaps the archaeologists want to link their findings to modern society. That is, they want to show that the ancient Maya had similar institutions to modern society, or that they are relevant to modern concerns. Any reader of this blog knows that I sympathize with this urge, but I don't think it justifies misusing a standard social science concept.
2. Perhaps the archaeologists find their new evidence incongruous with traditional models of the Maya and other ancient societies in which commoners are seen as exploited, downtrodden, and miserable, with no wealth or freedom. If you separate out a group of prosperous commoners as a new category, a "middle class," this can preserve the downtrodden commoners view for the rest of the commoners. I am sure that none of the archaeologists would admit to this motive, but the downtrodden commoners view is quite strong for many ancient societies.
3. Perhaps the archaeologists are unfamiliar that the concept "middle class" has a specific usage in social history and social science that does not match the ancient Maya.
4. Perhaps they really think they have identified a new social class, something not present in most other ancient civilizations. That would be an exciting finding, but it would require an explicit theoretical and comparative argument, backed up by considerable data. No one has done this yet, and I don't think it is feasible.

I don't know whether these four suggestions account for why archaeologists want to promote a "Maya middle class." Maya society was fascinating and complex, but there are better ways to describe it.

References:

Chase, Arlen F. and Diane Z. Chase
1996    A Mighty Maya Nation: How Caracol Built an Empire by Cultivating its "Middle Class". Archaeology 49(5):66-72.

Chase, Arlen F. and Diane Z. Chase
1992    Mesoamerican Elites: Assumptions, Definitions, and Models. In Mesoamerican Elites: An Archaeological Assessment, edited by Diane Z. Chase and Arlen F. Chase, pp. 3-17. University of Oklahoma Press, Norman.

Kocka, Jürgen
1995    The Middle Classes in Europe. The Journal of Modern History 67(4):783-806.

Pirenne, Henri
1925    Medieval Cities. Princeton University Press, Princeton.

Sjoberg, Gideon
1960    The Preindustrial City: Past and Present. The Free Press, New York.

Trigger, Bruce G.
2003    Understanding Early Civilizations: A Comparative Study. Cambridge University Press, New York.

Williamson, Jeffrey G.
2010    Five Centuries of Latin American Income Inequality. Revista de Historia Económica 28(Special Issue 02):227-252.

Frank Lloyd Wright and the Maya

Taliesen West facade

Taliesen West facade

Cindy and I visited the Frank Lloyd Wright workshop/school Taliesen West (in Scottsdale) over the Christmas holiday. Our nephew, Los Angeles architect James Diewald, was in town, as were Cindy's parents. I had heard that Wright was influenced by ancient Maya architecture, so we looked for evidence of this at Taliesen West. It didn't take long to find. Several of the buildings exhibit a sloping exterior wall in a form common in the architecture of ancient Mesoamerica. The outward-sloping panel is called a "talud" by Mesoamericanists. It is most famous at Teotihuacan, where the sloping panels alternate with vertical framed panels called "tableros." But Wright used the talud without the tablero.

Xochicalco, Feathered Serpent Temple

Contrary to various books about Wright's influences, the closest parallels of this talud form are not to the Maya, but to ancient central Mexican architecture, such as the Feathered Serpent Temple at Xochicalco. ((NOTE: I am not providing links for Xochicalco, since the readily available websites (e.g., the Wikipedia entry for Xochicalco) are pretty bad and filled with nonsense. Xochicalco was an urban center southwest of Cuernavaca that flourished from the sixth to ninth centuries AD; I worked at the site as a graduate student. Major recent fieldwork projects were directed by Kenneth Hirth and Norberto González; see references below)). A number of Maya cities did use the talud form, though.

At Taliesen West I asked our guide and some employees at the (very nice) bookstore about Mayan influence on Wright's architecture, but they didn't know much. One person said that this was one aspect of Wright's life that had not been researched yet. That didn't sound correct. I skimmed through various books on Wright's architecture in the bookstore, and they mentioned his explicit use of Maya models as a matter of course, mostly in reference to a set of houses he designed in the 1920s in Los Angeles.

Hollyhock House, Los Angeles

Hollyhock House, Los Angeles

The Hollyhock House was built for Aline Barnsdall between 1919 and 1921, and shows a general formal similarity to buildings and complexes (the so-called "Nunnery  Quadrange") at the Maya city of Uxmal. This is a distinctive and attractive house; see more photos and information at the Hollyhock House website.

Ennis House, Los Angeles

Ennis House, Los Angeles

The Ennis House (built in 1924) uses similar forms and techniques, but has a greater number of specific Maya items in its architecture and decoration.Wright's client evidently had an affinity for Mayan art. Like the Hollyhock House, this is a gorgeous and fascinating structure; see more at the Ennis House website.
This house was used as a set in a number of films and television shows, including  Blade Runner and Buffy the Vampire Slayer.

What is Maya about these structures? Two features stand out to me; there are probably others. First, the overall form of the individual structures and their configuration resembles building in the so-called "Puuc Style" of the Yucatan Peninsula. Uxmal is the best-known city with predominantly Puuc architecture, and the well-visited site of Chichén Itzá has much architecture in the Puuc style:

Uxmal

Chichén Itzá

The second Mayan feature of Wright's Los Angeles houses is the use of individual blocks to produce walls with a rich textured surface. Wright called these "textile blocks." The Puuc Maya used varying kinds of blocks to produce textured walls, some depicting the rain god and others geometric in design.

One of Wright's "textile blocks"

Mosaic facade at Kabah (Puuc Maya)

Compare the Kabah facade to both interior and exterior walls at the Hollyhock and Ennis houses. There are other Maya parallels that turn up in Wright's work over a period of many years. They were not at all limited to the Los Angeles houses.

A bit of library research turned up much information about Frank Lloyd Wright's Mayan (and more general Mesoamerican) influences.  By far the best account is Barbara Braun's excellent book, Pre-Columbian Art and the Post-Columbian World: Ancient American Sources of Modern Art, which has a chapter called "Frank Lloyd Wright: A Vision of Maya Temples." In a 1930 lecture, Wright said, "I remember how, as a boy, primitive American architecture, Toltec, Aztec, Mayan, Inca, stirred my wonder, excited my wishful imagination" (quoted in Braun, p.138). Braun goes on to chronicle Wright's use of Mayan architecture. She does not seem to have a good grasp of non-Mayan Mesoamerican architecture, however, and Wright's use of elements from sites like Xochicalco, Tula, and other non-Mayan cities is a topic that could stand some additional research. Additional information can be found in Ingle (1984) and Tselos (1969), a semi-rigorous article. The 1920s and 1930s were a period when ancient Mesoamerican art and Mesoamerican traditional culture more generally were very popular in the U.S., and Wright was in the midst of this movement (see works by Braun, Delpar, and Park below).

I was particularly interested in the role of the Chicago fair of 1893, the Worlds Columbian Exposition, in the possible development of Wright's appreciation for Mayan architecture. Wright was working in the office of Chicago architect Louis Sullivan at the time, and participated in the design of several structures at the fair. The fair also included full-size replicas for several Puuc Maya structures (from Uxmal, Chichén Itzá, and Labná). When the fair was dismantled, these were later assembled at the Field Museum of Natural History (in Chicago). It was not clear from the sources I consulted (see below), however, how much of an impression these made on Wright, or the specific nature of their possible influence on his ideas.

Puuc Maya replicas at the Chicago Worlds Fair, 1893

I highly recommend this outstanding account of the Chicago fair, The Devil in the White City: Murder, Magic and Madness in the Fair that Changed America by Erik Larson. I really enjoyed this book a few years ago (although I can't recall now whether Larson discusses the Maya buildings).

Sources on Maya influences on Frank Lloyd Wright:

Braun, Barbara  (1993)  Pre-Columbian Art and the Post-Columbian World: Ancient American Sources of Modern Art. Abrams, New York.

Delpar, Helen  (1992)  The Enormous Vogue of Things Mexican: Cultural Relations between the United States and Mexico, 1920-1935. University of Alabama Press, Tuscaloosa.

Heinz, Thomas  (1979)  Historic Architecture: Frank Lloyd Wright, Ennis-Brown House. Architectural Digest (October):104-111, 160.

Ingle, Marjorie  (1984)  Mayan Revival Style: Art Deco Fantasy. University of New Mexico Press, Albuquerque.

Park, Stephen M.  (2011)  Mesoamerican Modernism: William Carlos Williams and the Archaeological Imagination. Journal of Modern Literature 34(4):21-47.

Steele, James  (1992)  Barnsdall House: Frank Lloyd Wright. Phaidon, London.

Tselos, Dimitri  (1969)  Frank Lloyd Wright and World Architecture. Journal of the Society of Architectural Historians 28(1):58-72.


On Puuc Maya architecture:

Andrews, George F.  (1995)  Architecture of the Puuc Region and the Northern Plains Area. Labyrinthos, Lancaster, CA.

Gendrop, Paul  (1998)  Río Bec, Chenes, and Puuc Styles in Maya Architecture. Translated by Robert D. Wood. Edited and with a forward by George F. Andrews. Labyrinthos, Lancaster, CA.

Kowalski, Jeffrey K.  (1987)  The House of the Governor: A Maya Palace of Uxmal, Yucatan, Mexico. University of Oklahoma Press, Norman.

Pollock, Harry E. D.  (1980)  The Puuc: An Architectural Survey of the Hill Country of Yucatan and Northern Campeche, Mexico. Memoirs vol. 19. Peabody Museum of Archaeology and Ethnology, Harvard University, Cambridge.

Proskouriakoff, Tatiana  (1963)  An Album of Maya Architecture. University of Oklahoma Press, Norman.


On Xochicalco:

de la Fuente, Beatriz, Silvia Garza Tarazona, Norberto González Crespo, Arnold Leboef, Miguel León Portilla and Javier Wimer  (1995)  La Acrópolis de Xochicalco. Instituto de Cultura de Morelos, Cuernavaca.

González Crespo, Norberto, Silvia Garza Tarazona, Hortensia de Vega Nova, Pablo Mayer Guala and Giselle Canto Aguilar  (1995)  Archaeological Investigations at Xochicalco, Morelos: 1984 and 1986. Ancient Mesoamerica 6:223-236.

Hirth, Kenneth G. (editor)  (2000)  Archaeological Research at Xochicalco. Volume 1, Ancient Urbanism at Xochicalco: The Evolution and Organization of a Pre-Hispanic Society. Volume 2, The Xochicalco Mapping Project. 2 vols. University of Utah Press, Salt Lake City.

Hirth, Kenneth G. (editor)  (2006)  Obsidian Craft Production in Ancient Central Mexico: Archaeological Research at Xochicalco. University of Utah Press, Salt Lake City.

López Luján, Leonardo, Robert H. Cobean and Alba Guadalupe Mastache  (2001)  Xochicalco y Tula. CONACULTA, Mexico City.