Abstract

It is difficult to pin down the definition of a weed; rather, the idea of a weed is constructed through a set of characteristics that are, for the most part, dependent on context and relative interactions. Following existing literature, we describe this dynamic, ongoing construction as a product of the agency of both people and plants. We interpret studies of ancient Maya agricultural techniques through the lens of plant agency and human-plant relations and aim to investigate the placemaking of garden landscapes through an analysis of both helpful and destructive “weed” agencies in traditional planting and weeding practices. Using new LiDAR data from the ancient Maya site of Coba, we examine the spaces in and around houselots, sometimes called “toft zones,” to look at time management and placemaking at different temporal scales. The result is a view of how weeds participated in ancient Maya agricultural landscapes, as well as a new appreciation for how certain plants impacted daily time management schemas and contributed to generational social memory.

Vegetation is usually considered to be the epitome of the “natural world.” But vegetative landscapes, like other landscapes, are culturally constructed. This becomes most apparent when focusing in on those plants commonly referred to as “weeds.” Weeds defy cultivation and maintenance, but they are brought into existence by an imposed cultural interpretation defined by these same processes, resulting in a paradox. As such, it is difficult to pin down the definition of a weed; rather, the idea of a weed is constructed through a set of characteristics that are, for the most part, dependent on context and relative interactions: place, time, and intention (e.g., Doody et al. 2014; Harlan and de Wet 1965:17, table 1).

Here, we consider the placemaking of agricultural landscapes through an analysis of how “weed” agencies would have contributed to everyday life in the houselots of ancient Coba, a large Classic Maya urban center in Quintana Roo, Mexico. Using new LiDAR data, we analyze two houselots at Coba, as both an experimental test of the technology’s utility in visualizing houselot terrain and as part of the larger intention of understanding weed agency as it would have been experienced in the houselot. LiDAR is found to be a useful predictor of exposed bedrock, and combined usage of these data with those collected through visual inspection allows us to assess the tableau of nature within which humans engaged with both purposeful and unplanned vegetation.

Weeds

Kristen J. Gremillion (1993:506), in a study of Chenopodium seed morphology in the Eastern Woodlands, emphasized that the distinction between wild and cultivated seeds is a spectrum involving a “crop/weed complex” in which weeds have actually affected the evolution of the cultivated plants (see also Baker 1974:12–13). This complex problematizes the distinction between uncultivated and cultivated plants, or weeds and crops, which becomes even more intricate when domestication is added to the equation (see also Alcorn 1984:324–327). Similarly, Robert A. Bye Jr. (1981:118) notes that weeds and domesticated plants are both “end products of genetic and ecological alterations mediated by human activities” (see also Casas et al. 2007). So, how can one define a “weed?” Although most people would probably agree that they understand the term, definitions of weeds can be extremely varied (see Harlan and de Wet 1965) depending on whether they are contingent upon plant part morphology (e.g., Gremillion 1993), plant behavior and/or characteristics (e.g., Baker 1974; Doody et al. 2014:126, table 1; Harlan and de Wet 1965; Radosevich and Holt 1984), or the continuously navigated interactions between people and plants that in themselves construct the plant’s identity as a weed (e.g., Doody et al. 2014; Radosevich and Holt 1984:1–2).

Botanist Herbert G. Baker (1974:1) wrote that “definitions of a weed are almost as numerous as the authors of papers dealing with them.” He employed his previously published definition, which outlined a weed as the following: “In any specified geographical area, its populations grow entirely or predominantly in situations markedly disturbed by man (without, of course, being deliberately cultivated plants)” (1). This biological definition is striking in that it logically lists contextual (and social) variables needed to define a weed: space, disturbance (e.g., building, clearing, or agriculture), and intention. Furthermore, because weeds enjoy disturbed environments, people (perhaps despite their strongest wishes) are constantly creating weeds’ ideal habitat (Harlan and de Wet 1965)—necessitating follow-up practices of “weeding” as a process of removal. Finally, the definition of a weed is especially difficult to standardize when we consider that weeds are also often useful plants (Bye 1981; Kawa 2016; Stepp 2004). Because of these factors, the term weed itself, as a supposedly standardized classification of plants, may be misleading. We use the term weeds here to refer to “volunteer” plants—thus not excluding those plants that are then used by people (e.g., Dine et al. 2019).

Within biological literature, weeds have been discussed in terms of their relationships with humans (Harlan and de Wet 1965; Radosevich and Holt 1984:v, 1–2). Since people use weeds for medicinal and other purposes (Kawa 2016; Stepp 2004), both parties benefit from the relationship, though in different ways. Through use of the descriptors “obligatory” versus “facultative,” Jack R. Harlan and J. M. J. de Wet (1965:20) outline a spectrum of how closely tied different plants might be to human-made spaces and places. It is useful here to consider commensal organisms, which Elizabeth Matisoo-Smith (2009:152) defines as “animals living in close association to humans,” providing as examples species such as pigs and rats. Commensal animals are often used to track human migration or the movement of domesticated species (e.g., Fuller and Boivin 2009; Matisoo-Smith 2009; Storey et al. 2013; see also Ammerman, chapter 10; Quintus et al., chapter 9; and Tomášková, chapter 11; all this volume). Dorian Q. Fuller and Nicole Boivin (2009:10) in particular look at weeds as an analogous way to investigate agricultural plant species.

A closer look at the weed/non-weed relationship demonstrates that it is less straightforward than it might immediately seem, especially with respect to the attribution of agency to one party or the other. A commensal organism is not necessarily in a commensal symbiotic relationship, which is defined as a relationship that benefits one party and does not affect the other (Radosevich and Holt 1984:126). Commensal organisms most certainly benefit from spaces they share with humans, but they also actively influence the very nature of that relationship in ways both positive and negative—for example, rats carry diseases that can be dangerous to humans. There is no way to neatly categorize or quantify this entanglement. Similarly, assuming that weeds are the only party affected by their relationship with humans may be denying them their agency or at least the effects they have on the lives of people. These plants are an important resource and serve as or for food, medicine, animal forage, soil nutrient deposition, and erosion prevention (e.g., Alcorn 1984:327–336; Bye 1981; Casas et al. 1996; Fujisaka et al. 2000; Kawa 2016; Vieyra-Odilon and Vibrans 2001). In fact, edible or useful weeds are a category of plants referred to as quelites, originating from the Nahuatl word quilitl (Bye 1981; Casas et al. 1996; Vieyra-Odilon and Vibrans 2001).

The study of ancient weeds proceeds according to the same methodological protocols as the study of ancient domesticated plants, as part of the range of plant remains recovered from excavations (Fuller and Boivin 2009; see also Slotten 2015:127). Paleoethnobotanists—archaeologists who study plant remains—examine seeds, charcoal, plant parts, starch grains, phytoliths, and pollen as categories of macro-botanical and micro-botanical evidence for various types of ancient plant use (e.g., Dussol et al. 2017; Marston et al. 2014; Pearsall 2015; Piperno 2006). Christopher T. Morehart and Shanti Morell-Hart (2015) have called for a more social approach where plants can be interpreted in a light similar to a potsherd or an obsidian flake, in which relative perceptions of utility or value are context-dependent.

Vegetative Agency

One of the most common areas in which discussions of vegetative agency come into play in archaeology is within studies of domestication. Melinda A. Zeder (2012:162) has reviewed some of these approaches, noting that “the primary difference between different definitions of domestication lies in the degree of emphasis placed on either the human or the plant/animal side of the equation.” BrieAnna S. Langlie and colleagues (2014) provide a useful review of some of the ways archaeologists can access questions of plant domestication, including assessments of morphological evolution in plant parts, documentation of micro-botanical plant remains, and considerations of taphonomic processes (see also Smith 2001). They note that domestication is often conceptualized as “coevolution,” but their discussions of agency revolve around humans (Langlie et al. 2014:1611–1612). Marijke van der Veen (2014:801) challenges such approaches to domestication, asserting that “while the process of domestication has often been regarded as a process brought about by people, we might equally consider it something that plants and animals have done to us.”

Contemporary studies of weeds and weeding provide insights into the ways people engage with intentional and unintentional plants. Working in Christchurch, New Zealand, Brendan J. Doody and colleagues (2014:125, original emphasis) use Judith Butler’s work on performativity to consider how “rather than having a pre-figured meaning, weeds are performed together by people and plants.” The implication of considering a category of plants in this manner is that their meanings at any given moment are heavily based on context. However, people in Christchurch do not simply decide which plants will become “weeds” but instead react to qualities of the plants themselves in association with other spatial factors (Doody et al. 2014). Doody and colleagues’ conceptualization of people-plant interactions reflects a give and take that defines the identity of weeds. Somewhat differently, Guntra A. Aistara (2013) found that a socially (and institutionally) ingrained definition of weeds caused Latvian farmers to be wary of permaculture strategies, which embrace those plants as part of the landscape.

Alternatively, Nicholas C. Kawa (2016:89–91), working in rural Brazil, has observed the “weedy” characteristics of many medicinal plants in people’s house gardens. Interestingly, these weeds are caught up in cultural and religious practices in the area—an important example of how being forced to react to the materiality of a plant is in some ways a manifestation of its agency (Kawa 2016; see also Jones and Cloke 2008). In another example, ethnobotanical research with the Teenek of northern Mexico led Janis B. Alcorn (1984:324–327) to reject classification by imposed terms such as “domesticated” or “cultivated” and to question spatial designations that leave no room for the fluidity and messiness of everyday life among plants. The category of “spontaneous vegetation” is used rather than the word weeds (327–328). Finally, as demonstrated in John Charles Ryan’s (2012:113–115) case study of the West Australian Christmas Tree, a “human-plant studies” approach draws on a variety of scholarship, including Western science as well as Indigenous knowledge and practices, to conceptualize a network of interaction with plants as agentive beings. What these examples emphasize are the effects of (1) materiality and (2) social context on local interpretations of weeds.

Thinking about weeds raises the question of what it means for a plant to belong—or not belong—in a certain place and the spatial and environmental politics implicated in attempting to answer this question (Besky and Padwe 2016; Head et al. 2014). In considering Tim Ingold’s (1997:250) approach to the theorized dichotomy between what is “natural” and what is “social,” human agency is not eliminated; we instead embrace a foundational premise that the way we experience “nature” is also social. Ingold’s evaluation of the problem provides necessary nuance as he writes that “one cannot get rid of a troublesome dichotomy, such as that between nature and society, simply by collapsing one side into the other” (250). Following Ingold, we evade the dangers involved in approaches such as actor network theory, which Robert W. Preucel (2012:17) has warned involve a move away from “the social.” Furthermore, regardless of whether we conceptualize nature socially, the materiality of natural processes is real (Jones and Cloke 2008:94). When agentive decisions made by humans are in reaction to environmental phenomena, those phenomena must be considered important forces (see Smith, chapter 1, this volume). This is not to deny the profound effects of human agency; rather, it is to place human agency in context and acknowledge real influences on our decision-making processes (Ingold 1997; Jones and Cloke 2008).

Furthermore, Clark L. Erickson (2010 [2008]) has emphasized that landscapes seen as “natural” have long been managed by people; the idea of wilderness as separate from and opposed to human activity can even hinder conservation efforts (Ardren et al. 2015; Gómez-Pompa and Kaus 1992). Humans and crops similarly exist in the same environment as weeds (see also Ingold 1997:244). There is also a philosophical-botanical avenue of approaching the agency of plants from plants’ points of view, encompassed in theoretical notions of consciousness as well as quantitative assessments of communication (Calvo 2017; Gagliano 2013; Marder 2012). We find this particular realm of engagement with plant agency useful to think with, but we turn to more context-specific understandings of plants to conceptualize how people might have engaged, at a macroscopic level, with weeds at ancient Coba. As Stephen Houston (2014:78) notes, Classic Maya understandings of animacy operate within a contextualized worldview or cosmology given that “Maya evidence . . . pushes and concentrates energy in all manner of ways.”

Plants and People in the Maya World

Paleoethnobotanical studies of ancient Maya agriculture and plant foods have yielded information about the use of a wide variety of species, including the staples of maize, beans, and squash, as well as garden plants like chile and bush spinach and tree crops such as mamey, avocado, and nance (Lentz 1999; Ross-Ibarra and Molina-Cruz 2002). In addition to house gardens and orchards or areas designated for plant cultivation, the Maya would have maintained what Arturo Gómez-Pompa (1987:6, table 1) calls “ ‘Natural’ forest ecosystems,” again highlighting the fluidity between the “natural” landscape and the constructed one. Such engagement with and embeddedness in the surrounding natural world (or socially constituted nature) would have given the ancient residents of Coba direct and extended experience with the relentless, creeping growth of plants (e.g., Besky and Padwe 2016:21) as well as garden constituents that required both care and caution (Ardren and Miller 2020). In fact, imagery from elite Maya ceramics highlights the active role of many culturally important plants by depicting them in a “personified” manner (Houston and Scherer 2020).

Archaeological data from the site of Joya de Cerén, destroyed and preserved by a volcanic eruption, show that the ancient population grew a wealth of plant resources in their house gardens and, significantly, heavily relied on plants commonly described as so-called alternative resources (Lentz and Ramírez-Sosa 2002; Sheets et al. 2011; see also Slotten 2015). This perspective is bolstered by the work of Venicia M. Slotten (2015:116–121), who found significant amounts of weed seeds in paleoethnobotanical samples from agricultural contexts at Cerén and suggested that these plants may have been useful to residents. Were these plants simply left to grow, or were they carefully managed as part of the houselot landscape (Slotten, Lentz, and Sheets 2020)?

For the Classic Maya, spirit and personhood could be situated in various materials, objects, or entities (Harrison-Buck 2020; Houston 2014; Jackson 2019). Stone, as monumental fragmentation of the earth itself, was animate, a quality that could be more explicitly revealed by careful carving (Stuart 2010). Plants were sometimes portrayed as the living embodiment of past souls (Houston 2014:13). Materials imbued with life force were, “in their working . . . magical and revelatory, releasing by human craft the potentials promised by myth” (Houston 2014:98). Perhaps cultivation and farming over generations produced similar results in agentive plants. Today, some individuals among the Tzotzil Maya–speaking people of Chiapas define weeds by the fact that they “do not have good souls” (Laughlin 1993:105). The agentive properties of these plants are evident in the anger they feel upon being removed and the jealousy they feel toward those plants that were spared (106). In fact, an ethnobotanical work by Dennis E. Breedlove and Robert M. Laughlin (1993:463) has a section dedicated to “Plant Emotions.” Karl Taube (2003:469) has cited some of these negative conceptions of weeds as supporting evidence for a notion among the Maya of an untamed and dangerous forest as distinct from civilization, where the “constant encroachment of weeds” is an ever-present reminder of what lies outside. However, it is not only “weeds” that show agency (Laughlin 1993:106); and Taube (2003:465, 485-486), too, recognizes the ambiguities that arise from attempts to apply a rigid separation of nature and culture to Maya conceptualizations of the environment, including the important role the “forest” played in daily life for the ancient Maya. Here, we consider previously published mapping data on houselots and analyze how LiDAR data may provide new insight into how certain plants—those classified as “weeds”—would have factored into daily life and time management in the northern Yucatán.

Coba

Coba is a Classic Maya site in the northern lowlands (Folan et al. 1983), located in the modern-day municipality of Tulum, Quintana Roo, Mexico (figure 7.1). The site played an important role in the Classic period (300–900 CE) as a royal dynastic polity that may have engulfed surrounding cities such as Yaxuna (e.g., Loya González and Stanton 2013), but it was also occupied in the Postclassic (1100–1500 CE) when many building additions were constructed (Andrews 1981:7–8). As an urban center, it contained centralized elite households and spaces that intersected with a radiating pattern of raised causeways (Folan et al. 2009; see also Andrews 1981:5). Plant resources were unevenly distributed within the city (Folan et al. 1979). William J. Folan and colleagues (2009:65) have suggested that Coba may have been a “garden city,” citing the importance of “in-between” areas amid structures. An important resource at the site was sascab, or powdered limestone, extracted from sascaberas, which may have eventually become spaces used as gardens (Folan 1978; see also Folan 1983:24–25).

Case Study: Coba Group 1

Group 1 at Coba is a cluster of nine small to medium-size structures surrounded by a stone boundary wall. It is located 890 m north of the center of urban Coba, as designated by the tallest pyramid (known by its Maya name, Nohoch Mul), but within the broader urban core and adjacent to the longest ancient Maya road, which runs for 100 km from the center of Coba west to the smaller urban center of Yaxuna. Group 1 has a primary cluster of four larger structures arranged around a central interior plaza, with five smaller domestic structures and adjacent open areas to the east. Sacbe 1, the ancient road, forms the boundary of this group to the south, while the dry stone wall (albarrada) encircles the other three sides, for a total enclosed area of 8,372 m² (figure 7.2). The group is located in a part of the site that incorporates mid-level residential structures with administrative and ritual architectural features. Some of the structures within the main cluster of buildings are over 4 m tall and exhibit elements that suggest they are funerary or ritual monuments.

Ground verification of LiDAR data covering Group 1 was performed during the 2018 field season. The area in which Group 1 is located is currently covered by secondary forest growth. According to local officials, in the early 1980s this area was under cultivation, as milpa and cattle were allowed to graze there; since then, it has seen only periodic visitation by local residents to collect firewood or other forest products on a very small scale. Group 1 was intensively excavated during the 2018 field season, which necessitated clearing much of the underbrush—revealing additional structures not visible in the LiDAR, specifically the low domestic structures in the eastern part of the walled group (see also Hutson et al. 2016). A small rectangular depression visible in LiDAR was determined to have been a stone quarry, and bedrock was visible in approximately 15 percent of the ground surface in Group 1. There were no cavities visible in the bedrock within Group 1.

Given the relatively small presence of bedrock on the ground surface of Group 1, as well as the lack of solution holes or other depressions in the bedrock, intensive weeding would likely have been directed toward the entire houselot rather than concentrated in one space. It is also likely that weeds’ agency went less contested on the open, eastern side, as the performative delineation between the houselot and surrounding areas was farther away from the structures themselves. The stone quarry may have provided an advantageous location for tree crops or other plants that needed both deeper and wetter soils (see Folan 1978). Given the large amount of open area adjacent to the structures of Group 1 and within the boundary wall, weeding could have consumed a significant amount of time regardless of whether this space was used for cultivation or kept clear for household activities. The location of Group 1, along one of the most important ancient roads emanating from the city center, also suggests that humans may have spent significant time addressing the agency of plants that sought to enter the road, activity areas adjacent to the road, or ritual structures visible from the road. It is also possible that daily traffic on the road had the effect of inhibiting weed growth, in which case the definition of such plants would need to be reconsidered, as people may not have been interacting with them intentionally at all.

Case Study: Coba Group 28

Group 28 at Coba is a cluster of five small structures around a central patio area surrounded by a rectilinear dry stone boundary wall, located 3.2 km south of Nohoch Mul in an area of dense residential settlement (figure 7.3). The five small structures are distributed in a single cluster of domestic platforms on three sides of a central patio in addition to sixteen ancillary structures within or adjacent to the walled group. Between the dry rubble wall surrounding the domestic structures and a quadrangular grid system of larger and more substantial masonry walls that define this “neighborhood,” or region of the city, is an extensive open area that would have provided garden and activity space for the group’s inhabitants. The quadrangular design of the broader neighborhood of modest residential settlement that contains Group 28 is distinctive from other residential zones, even some that are adjacent to the north and east. Starting in the mid- to late 2000s, the area in which Group 28 is located has been under periodic milpa cultivation and is also used as cattle pasture. It was last cleared for planting in 2017, so there is minimal tree growth or other large vegetation, but the area nonetheless was covered with bushy undergrowth that was removed in 2018 to allow for horizontal excavation of living and activity areas. Local Maya residents saw the bushy undergrowth as economically useful cattle fodder, while our archaeological team viewed the same plants as weedy impediments to research.

The LiDAR survey was performed in 2017, with ground verification in 2018 that involved clearing vegetation—an activity that revealed details of the very small ancillary structures and a previously undetected stone walkway that connects a circular structure in the northwestern part of the group. However, for the most part, ground-truthing did not reveal new structures, and it was demonstrated that the LiDAR data were extremely accurate for this group. Bedrock was present at ground surface in only one very small area of the group, along the northern boundary wall; and the vast majority of the open areas was covered in soil. This small patch of bedrock likely comprised only approximately 1–2 percent of the walled area. Open areas with soil cover comprised the majority of Group 28 and would have provided ample space for gardening and its attendant responsibilities of weeding, without apparent specific focus in any one space. The presence of two wall features, the circular one around core residential structures and the more substantial rectilinear one outside the houselot area, suggests that a greater effort was made to prevent plants from entering the domestic compound. The human occupants of Group 28 spent significant time and resources to secure their living space—not from the forest, as this group is located in an area of dense settlement, but perhaps from the smaller weedy intrusions that might have impeded full (human) use of the landscape. It is also useful to consider how the rectilinear boundary wall would have made a sharp delineation of labor investment compared to the usual development of a toft zone resulting from a gradual tapering application of labor as one moved farther away from buildings.

Visual examination of the cleared ground surface at Groups 1 and 28 of Coba allowed us to return to the LiDAR imagery and refine our processing in order to attempt to correlate bedrock visible from the ground surface with LiDAR visual data. With minimal modification of the digital elevation model files to enhance color we noticed open areas of bedrock were visible in LiDAR data as areas of smooth, texture-less background. This pattern held true for both residential groups, one of which was much more heavily forested than the other. Thus both case studies suggest that LiDAR data provide useful information about the houselot landscape and thus time spent in negotiation with weeds.

Conclusion

Although the phenomenon of weeds is found globally, the interaction between weeds and cultigens has a distinct mutualism in Mesoamerica. As biologists have found for weed growth in milpas, weed biomass increases every year a plot is cultivated until, by the third year, weed growth significantly decreases agricultural yields by out-competing maize plants for critical resources (Lambert and Arnason 1980, 1986). Soil quality does not decline; in fact, weeds lock soil nutrients into place by preventing leaching. The subsequent stages of vegetative regeneration in an abandoned milpa or any land that has been managed for cultivation involve predictable sequences of secondary growth (e.g., Ford and Nigh 2015; Ford et al. 2012). For example, the length of fallow time can eventually affect fallow plant communities within a milpa plot (Johnston 2003). Local knowledge of these secondary vegetation patterns could therefore be used to infer past land-use patterns in a particular place.

There is an under-considered nuance to human-weed interactions: in cases of long-term habitational changes, weeds act as conveyors of localized ecological knowledge of a particular place. In cases where households continuously occupy a landholding over generations, older household members act as the repositories of this ecological knowledge (Netting 1993); but when breaks occur in the history of a houselot’s occupation, new occupants have to learn the localized landscape without the benefit of the accumulated wisdom of previous generations. Perhaps the existence of weed communities was able to communicate a story of past land-use practice and agricultural strategies to the newcomers in a particular houselot landscape, taking over the role of repository of localized ecological knowledge in the houselot and instructing inhabitants on localized ecological potential: how long had it been since people last occupied this space? Which parts of the houselot had been used for cultivation, and where did the greatest amount of water accumulate? Had certain species been allowed to proliferate where others were not, indicating their utility? Weed communities contained the answers to these questions.

Weeds would also have contributed to social memory. Rosemary Joyce (2003:112) has described the role of movement around a city, including within houselots, as part of the memory-making process (see also Magnoni et al. 2012). Patterns of movement produced by realization of daily chores in the houselot would have relationally altered the landscape, connecting it and those within it to a multitude of meanings and memories (e.g., Ingold 2010 [1993]). One can imagine that weeding would have been an activity that allowed ancient residents of Coba to shape the landscape, but at the same time, through habitual motion, it left its own mark on their bodies (e.g., Bourdieu 1990). In studies of ancient identities, the domestic world has been identified as a primary arena where culture is reproduced. Due to the importance of household gardens in the subsistence success of ancient families, tending plants in the household compound of ancient Maya centers was a highly charged repetitive activity that helped convey cultural values and solidify the social imaginary while providing nutritional content.

Coba has already played a groundbreaking role in our understanding of the highly localized interactions between Maya people and plant communities over generations in houselots. As demonstrated here, LiDAR data were useful and reliable predictors of both domestic architecture and natural features such as bedrock at ground surface, a significant factor in calculating work hours for estimations of weeding, gardening, and other household tasks. Archaeological fieldwork has allowed us to evaluate LiDAR’s potential to predict terrain differences that can inform our understanding of how weeds and plant communities more generally were part of the lived experience at Coba on a daily, seasonal, generational, and centennial scale. By calculating the ratios of bedrock to soil cover and observing the nuances and idiosyncrasies of the bedrock itself, we can better understand how people would have engaged with the landscape and contended with vegetation at the level of the houselot. LiDAR becomes a way not to reconstruct architecture but to reconstruct human-plant relations, understanding where, spatially, people would have spent most of their time and in what ways their interactions would have co-constructed the category of weeds.

We have discussed and applied this under-utilized approach to LiDAR technology by considering its productivity in predicting bedrock-to-soil ratios and thus factoring into interpretations of human-plant relationships. By considering these results through the lens of vegetative agency, we have presented an archaeological approach to the study of weeds as a relational category that is an essential and inescapable counterpart to domesticated plants. For the ancient inhabitants of Coba, weeds would have been a set of agentive beings that both contributed to the construction of and were defined by spatial boundaries. As pests, medicines, and archives of knowledge, depending on the context, weeds would have affirmed their own agency in the creation and maintenance of social memory in ancient Maya houselots.

Acknowledgments. We wish to thank Monica Smith for the invitation to participate in this exciting volume as well as for her productive comments on earlier drafts, as well as the two anonymous reviewers whose comments improved this chapter. We thank the Consejo de Arqueología, Instituto Nacional de Antropología e Historia, for permission to conduct research at Coba as part of the Proyecto Sacbe Yaxuna-Coba. We thank all the PSYC project members but especially Travis Stanton, Stephanie Miller, Ashuni Romero Butrón, and Patrick Rohrer for their generous assistance with the 2018 excavation data, and we thank Stephen Houston and Robert Preucel for helpful comments on earlier versions of this chapter. Dine would also like to thank John (Mac) Marston, whose classes at Boston University provided space to think about vegetative agency. This research was generously supported by the National Science Foundation (award #1623603).

References