ABSTRACT
This chapter addresses one of the more interesting conundrums faced by the Pakbeh Regional Economy Program (PREP), which investigated the predominantly Early Classic lowland Maya site of Chunchucmil, Yucatán. Why, when it was the most accessible stone-cutting medium, was the number of artifacts made out of chert so low in all contexts, including domestic ones? Very low-quality chert, or what we call “silicied limestone,” is seemingly everywhere on the karstic landscape, and high-quality chert can be found some 30 to 60 kilometers inland in the Puuc region. We propose that silicied limestone was the primary material used to make stone tools in northern Yucatán, but that preservation and other issues of archaeological collection techniques have left this material unrecognized. Members of the Pakbeh Regional Economy Program (PREP) have worked at the predominantly Early Classic site of Chunchucmil in northwest Yucatán for 13 years. At its peak, the city had a projected population of 35,000 to 42,500 (Hutson et al. 2008; Magnoni 2008; Magnoni et al. 2007). As a result of our extensive horizontal and test excavation programs, PREP has collected just 624 pieces of chert and chert-like material, including debitage. Of these artifacts, 299 came from our program of test pitting in residential groups that moved approximately 500 cubic-meters of earth. Thus, the density of chert artifacts is just 0.6 per cubic meter of excavated material. Another index of scarcity is the average number of chert implements consumed per household. This amounts to 0.41 chert artifacts per tested context. Of the total 624 pieces of chert we recovered at Chunchucmil, only 48 were used as tools. These include utilized akes as well as formal and informal tools. Thus, the number of chert implements seems far too low to meet household needs, even if we cannot precisely specify what those needs were. Most of these tools were made of good-quality chert, but good-quality chert is not locally available and had to be imported. Tool-quality chert sources are located 30 to 60 kilometers inland in the Puuc region, so material could have been brought to the site fairly easily and inexpensively. The Chunchucmil Maya were not averse to importing cutting materials, because, as we have argued elsewhere, it was a specialized trade center with an active marketplace (Dahlin 2003; Dahlin and Ardren 2002; Dahlin et al. 1998; Dahlin et al. 2007; Hutson et al. 2005). As if to prove the point, our excavations produced 2,716 pieces of obsidian, which is more than four times the total number of chert artifacts. Therefore, if we add the amount of chert and obsidian, we arrive at a total of 3,340 pieces (including tools and debitage). A portion (N=107) of the obsidian collection has been
chemically sourced by XRF at the University of California at Berkeley, and Mazeau has visually sourced the entire collection. Ninety-two percent of the obsidian came from El Chayal, Guatemala, roughly 670 kilometers south of Chunchucmil (Figure 1.1), or nine to 13 times farther than the nearest high-quality chert sources (Mazeau 2002a and 2002b). Obviously, obsidian was far more costly to import than chert. Signicantly, none of the obsidian came from ritual deposits, as is the case at some other sites (see Moholy-Nagy 1997; Rice 1987). In fact, PREP has not excavated many caches or burials. Almost all of the Chunchucmil obsidian is in the form of blades recovered from domestic or crafting contexts. In fact, more than 25 percent of the obsidian from Chunchucmil was recovered in a single residential group that was engaged in some sort of specialized craft. Nor can obsidian be considered simply as an item of wealth, because it was found equitably distributed among both elite and commoner households within greater Chunchucmil, and obsidian artifacts typically show considerable use wear (Mazeau and Forde 2003). Finally, obsidian cannot be considered to be an essential cutting implement needed by all, because many households, particularly outside of greater Chunchucmil, probably did not have access to it (Hutson et al. 2008; Mazeau and Hixson 2005). As far as we are able to determine, all or most of the Chunchucmil obsidian was intended for practical use and should be considered together with the chert artifacts in order to understand how many cutting implements were consumed per household. Even when combined, that total still seems quite small. Obsidian, a volcanic glass, is fragile and cannot withstand great impact, especially when used in the form of prismatic blades. The edges of obsidian tools are sharp but delicate, so this material was probably reserved for making precise cuts into soft materials such as foods, fabrics, cordage, and some wood. In contrast, chert was reserved for heavier tasks, such as chopping and cutting wood, limestone, and other dense materials. Households at Chunchucmil required both ne and crude cutting implements, but they had a greater need for more heavy-duty cutting tools than the ratio of obsidian to chert artifacts implies. In short, we should expect to nd more chert than obsidian artifacts, even though obsidian tools could be recycled for more heavy-duty tasks after they became dull from delicate cutting. One possible explanation for the apparent shortage of cutting implements is that they were so expensive and difcult to obtain that they were used to the point that little was left to enter the waste stream (McAnany 1993:82). Nevertheless, most of the cutting implements at Chunchucmil were not used to the point of total exhaustion. Yet another possibility is that the Maya used perishable cutting materials, which, given the high rate of organic deterioration in the subtropical environment, have long since decomposed (see Dahlin 2006). For example, slivers of hardwoods like ziracote (Central American ironwood) could have been honed to a ne edge by sharpening them with abrasive stones and leaves. Landa (Tozzer 1941:121) mentions that arrow points were sometimes made from sharp sh teeth, and although the occupation of Chunchucmil predates the time when bows and arrows were common, perhaps sharp sh teeth were used for spear points and other piercing and crude cutting implements. Some soft materials like cotton bers could have been cut with readily available obsidian substitutes like the sharp edges of grasses, which grow locally. There is no doubt the Maya used organic cutting implements for some purposes, but since we cannot identify these activities, there is no way to test this hypothesis. In any event, organic tools probably would not have competed with chert as heavy-duty cutting implements. Silicied Limestone and Stone-Tool Production
Another possibility, one that we pursue here, is that archaeologists are not recognizing some lithic tools recovered during their excavations and surveys. Heavy, dense, and durable silicied and capstone limestones, although they cannot be reliably fractured in a predictable manner, could have
been shaped into rough chopping and pounding implements. Silicied limestone akes can be used with some success as cutting and scraping tools, and fashioned into expedient tools with an edge that is serviceable for a limited duration (Hruby et al. 2007). Given that silicied limestones are ubiquitous at the site and that silica is a common element in regional soils (Beach 1998)—even present in architectural blocks that likely were shaped with the aid of denser capstone and silicied limestone implements-the Maya must have known of their useful cutting potential. Silicied limestone is porous and less compact than chert, oxidizes easily, is highly soluble, and thus is susceptible to the elements. It is, therefore, highly subject to weathering and other diagenetic processes. Post-depositional processes, specically exposure to rainfall (which is moderately acidic and has a pH about 5.6) and contact with plants or other organisms that release acidic compounds into the soil, dissolve silicied limestone. Such dissolution ultimately can obliterate cutting edges, fracture patterns, and original surfaces. Moreover, many of the soil environments, especially near architecture, are rich in calcite and dolomite, which provide ready sources of calcium and magnesium to patinate and obscure surfaces. Silicied limestone does not fracture in a conchoidal manner, so it does not feature the diagnostic traits found on tools and debitage made of chert and obsidian. Most detachment scars on silicied limestone appear to have a sheared cone fracture that does not produce a signicant bulb of force. Compounded with the greater frequency of inclusions in silicied limestones, worked pieces often do not look like artifacts. Moreover, the major component of the debitage produced from working silicied limestone is non-diagnostic shatter, which is rendered even less recognizable by diagenesis. Archaeologists in the northern plains, and we suspect elsewhere in karstic terrains, constantly nd pieces of limestone in survey and excavation. Some may once have had cutting edges. Such nebulous objects lack concrete diagnostic traits and are routinely discarded as “pot-lid” akes or “dog stones.” We almost certainly committed this error at Chunchucmil. Sampling Methodology We explore the hypothesis that households at Chunchucmil, and by extension within the northern Maya lowlands in general, routinely used silicied limestone to make cutting implements. At some sites, silicied limestone artifacts may have constituted the majority of the toolkit, and we are simply not recognizing these “phantom lithics,” that is, pieces that are not obvious stone tools or debitage. Our argument is built on carefully recorded data from a sample of 115 distinct operations. This sample was drawn from a much larger one that was stratied according to the 16 types of architectural groups that we recognized at the site. The 115 operations include test excavations at 110 residential groups, but we also include horizontal excavations at ve households in order to slightly amplify our sample. All operations met certain minimum criteria. To be included in our sample, an operation at a single residential unit had to yield a minimum of four kilograms of ceramics or, failing that, it had to have been tested by a minimum of eight one-by-one meter testpits. The operations discussed here represent only those residential units whose primary period of occupancy was the height of Chunchucmil during the late Early Classic and early Late Classic periods, ca. A.D. 500-700 (Mansell and Bond-Freeman 2001). Architectural groups that probably served civic-religious rather than domestic functions were excluded. We found evidence of specialized craft production at only two of these habitation groups, and activities conducted there required obsidian artifacts (Hutson 2004). We turn now to a brief description of the chert-artifact sample recovered at Chunchucmil from 1993 to 2006. We follow with discussions of the karstic geology of the Chunchucmil region and the processes that made chert and limestone formations in the Maya lowlands. We focus in this section on two deposits of chert-like silicied limestone that might have been easily accessible to the
inhabitants of Chunchucmil. Attention is then drawn to relatively well-preserved silicied limestone tools that were found inside sascab (gravel, sometimes incorrectly called marl) quarries at Chunchucmil. Minimal labor investments were required to make these tools, and they likely were discarded quickly as they became dull or broke. Moreover, the preservation of silicied limestone tools in sascaberas implies that at least some such tools were used at the site, and suggests that, over the centuries, those deposited in less protected environments might have lost their diagnostic traits, making them all but unrecognizable as tools. Chert and Chert-Like Silicied Limestone Tools
The density of non-obsidian lithic artifacts that we calculate for Chunchucmil, a mere 0.6 per cubicmeter, is quite low for an important Maya site. These non-obsidian lithic materials run the gamut from high-quality chert to silicied limestone with barely observable cryptocrystalline properties. Most of the tools are made of high-quality chert, but some are heavily patinated and many suffered an indeterminate amount of dissolution of their surfaces. Moreover, with the exception of several rather nely retouched bifacial implements (e.g., projectile points), the vast majority (91.8 percent) of the artifacts in the non-obsidian lithic collection are unretouched akes, ake fragments, nondiagnostic debris, and a few complete and fragmentary ake cores. The collection also exhibits great diversity in both the morphology and size of complete akes, with a standard deviation greater than 50 percent in all measured dimensions. The complete and fragmentary cores exhibit multidirectional knapping scars of various sizes, and present no signicant aking patterning. The diverse techniques and methods of production, reected in non-standardized artifact morphology, lead us to conclude that the dominant lithic industry was ake-tool production practiced by people who lacked a high degree of knapping skill. That is, non-obsidian chipped-stone tools were typically fashioned and curated as needed by amateur knappers who also used the tools. By contrast, obsidian tool-use required professionally made prismatic blades. We suspect that material and morphological variability, as well as the apparent paucity of cutting implements in domestic contexts at Chunchucmil, created a situation where our workers and staff were good at nding obsidian and high-quality chert, but not as skilled at identifying lithic artifacts made of silicied limestone. This could have been especially true for artifacts that had few identifying morphological features and resembled naturally occurring rock fragments. In all likelihood, even more questionable tools were discarded during washing in the lab and were never cataloged or analyzed. Through discussions with our colleagues working at other sites, we strongly suspect that the tendency to discard highly questionable tools is common on other large, long-term projects. This suspicion was given considerable support in 2006 when we explored two modern quarries that contained a large number of freshly exposed nodules of silicied limestone, and simultaneously discovered some well-preserved silicied limestone tools inside ancient sascaberas. Karstic Geology, Silicied Limestone, and Dolostine Nodules in Quarries Near Chunchucmil
The limestone bedrock in the northwestern Yucatán peninsula is of Miocene to Pliocene origin (SGM 2005). Below the dense, petrocalcic surface (Isphording and Wilson 1973; Perry et al. 1989), the lithology is soft and porous with zones of sascab (saprolitic limestone) and pockets of clay sandwiched between the ground surface and the water table (Figure 7.1). The petrocalcic surface forms irregularly in the upper meter because seasonally rising groundwater, saturated in calcium, precipitates calcium carbonate near the surface. This reduces the porosity from 15 percent to just one percent (Perry et al. 1989). Sascab forms as percolating water leaches calcium carbonate
Sa sc
ab er
M ou
nd Li
m es
to ne
Pa ve
m en
t W
el l
U itz
upward and downward, increasing porosity and weathering the less resistant minerals. But this leaching also leaves behind more resistant, silicied, dolomitic, and crystalline limestone, and, where it occurs, chert. Limestones near Chunchucmil are generally low in silicate minerals, but diatoms and volcanic and Saharan dusts have added some silica to these lithologies. Karst processes also create pockets in the limestone and sascab by dissolution along joints and bedding plains. Such pockets may continue to grow into sinkholes, which become reservoirs for eroding surface clays. Modern Quarries Near Chunchucmil We inspected a recently opened, ve-meter deep road-ll quarry located 15 kilometers east of Chunchucmil and near the village of Granada. The quarry has a number of these zones and pockets that can be seen in cross section (Figure 7.2). There are many dissolution cavities that begin at the petrocalcic surface and continue into the sascab at a depth of more than three meters. Most of these cavities are lled with red clays that washed down from the surface. We do not know to what degree the clays represent anthropogenic activity, but they do suggest a signicant amount of erosion, a conclusion that Sedov et al. (2008) also reached for the area near Yalahau, northern Quintana Roo. The sascab zones in the Granada Quarry contain an abundance of nodules of silicied and denser limestone, as well as dolostone. We found more such nodules strewn about the quarry oor. At rst glance, the faces of these recently broken nodules easily can be mistaken for chert. On closer inspection, however, it is apparent that the nodules tend to be softer, coarse to medium in texture, and dull: not unlike chalk. Clay and mineral inclusions are common. Some fossils were detected macroscopically. No nodules were found that, upon inspection, approached the high quality of the chert purportedly exported from sources in the Puuc region to the south and east. The water table in the Granada quarry is ve to seven meters below ground surface, depending on local microtopography, and only two to three meters below the ground surface around Chunchucmil, which limits the range for ancient quarrying. Nodules of harder carbonate rocks litter the oor of a two-meter deep quarry located on the northern edge of the village of Chunchucmil and just two kilometers from the site center. These nodules were even more coarse and chalky than those found at the Granada Quarry, but we were able to make serviceable bifaces from them. The nodules were found just one or two meters below the surface, the usual depth of ancient quarries. The material was mostly tan in color but generally resembled the Granada nodules in all other respects. These quarries demonstrate that workable silicied limestone was readily available and locally accessible. The Testimony of Sascaberas
The pockets containing clay, sascab, and nodules of silicied limestone-such as those noted in the quarries at Granada and north of Chunchucmil-were exploited anciently for many purposes. One important feature of these pockets, for example, is that they are located on low microtopographic surfaces, and serve as natural drains for the otherwise at terrain. We observed that huge ponds form in these topographic lows several days after heavy rains. Not infrequently, such ponds measure several hectares in area, and inundate houselots and even whole neighborhoods. Thus, although pockets with drains may not have been observable to inexperienced surveyors like us, especially during dry periods when most of our mapping was accomplished, it is inconceivable that anyone living on this landscape would not know precisely where they were located. Furthermore, the Maya could have used them for planting deeply rooted fruit-and nut-bearing trees, like ramon, that normally do not thrive in these thin soils (Dahlin et al. 2005).
