Modeling Andean Agricultural Technology: The Chaquitaclla

Introduction: Project Scope and Goals

Depictions of the past frequently neglect the lives of ordinary people in favor of displaying the flashy, ornate lives of the elite and their palaces, temples, gardens, and tombs. As reconstructions of the past carry both objective information and implicit assumptions, they reveal historical knowledge and contemporary values (Solometo and Moss 2013). Unfortunately, depictions of the past are rarely accompanied by disclaimers. Furthermore, many popular media represent a major source of historical knowledge for the public. As the public obtains historical knowledge from these films, documentaries, television shows, and publications, they often learn primarily about the elite. Overall, these visualizations of the past in these popular media demonstrate a preoccupation with treasure, personal wealth, and riches (Membury 2002; Russell 2002). Although popular media often claim to minimize inaccuracies, biased depictions of the past in National Geographic and other publications containing such preoccupations cannot be considered accurate (Solometo and Moss 2013). In addition, these depictions boldly assume viewers find the lives of ordinary people disinteresting. Even if this were true, focus on creating “pretty” visualizations detracts from the ability to provide historical information (Favro 2006). Artists, publications, and archaeologists alike should work towards addressing the excessive focus on the elite in visualizations the past.

When given the opportunity to depict the Peruvian past, I opted to focus on the lives and material culture of the non-elite. This focus, combined with my own research interests in crop domestication and agriculture, motivated my choice to depict Andean agricultural technology. In this paper, I provide a brief introduction to the Peruvian ceremonial and urban center which my class aims to model. In addition, I describe the chaquitaclla and its contemporary importance in Andean agriculture. Given this contextual knowledge, I report on my 3D modeling process, evaluate its success, and suggest future improvements.

Pachacamac

Pachacamac continuously served as an important ceremonial and urban center beginning in the Early Horizon and until Spanish occupation (Tiballi 2010). The vast coastal center, representing one of the largest sites of the ancient Andes, served important religious functions for more than 1,000 years (Eeckhout 2013). Since its earliest occupations in 200 CE until the Spanish Conquest in 1533, the site witnessed immense changes in religious beliefs and practices. Pachacamac initially served as the capital of the Ychsma chiefdom and as the sanctuary of Pacha, the creator god for which the site was named. Thousands of pilgrims traveled to Pachacamac, often facing harsh conditions and treacherous terrain, to consult the oracular idol. After Incan conquest during the fifteenth century, Pachacamac’s religious and political importance grew beyond its local and regional significance. Under Inca rule, Pachacamac became a pan-Andean sanctuary and ceremonial center, now dedicated to the coastal god. This newfound prominence persisted until the Spanish conquest.

Archaeologist Max Uhle excavated Pachacamac in 1896 and found a vast number of burials, palaces, public plazas, “convent,” and many temples (Uhle 1903; Shimada 1991). Since then, archaeologists have continued to excavate the site (Eeckhout 2013). Pachacamac’s long-term local, regional, and pan-regional importance has captivated researchers for the last century. More recent research focuses on understanding the function, development, and influence of Pachacamac in pre-Hispanic periods (Tiballi 2010). After over a century of research, archaeologists studying Pachacamac continue to focus on the rich cemeteries, ornate temples, and elite palaces. While providing insight into the Pachacamac’s religious and political functions, these studies do not adequately examine the lives of ordinary individuals.

Supporting the large resident population and significant number of pilgrims and support staff arriving to the ceremonial center probably required intense agricultural production of food. In addition, agricultural products such as maize held large symbolic value. The excavations of Uhle and others do reveal large quantities of food remains, including corn cobs, in burial contexts (Uhle 1903; Eeckhout 2013). Therefore, agricultural labor including preparing soil, planting seeds, irrigating and tending crops, harvesting mature crops, and processing and storing harvested crops probably occupied many Pachacamac residents. Lower-class residential areas can be found north of the site’s core between the inner and outer walls. However, most of the people supporting Pachacamac probably lived dispersed throughout the valley in hamlets and small towns. Although archaeological excavations did not reveal agricultural tools from Pachacamac, agricultural activities certainly took place within and nearby. Residents of Pachacamac likely cultivated the valley floor to the east and south of the site (Tiballi 2010). Unfortunately, though, archaeological excavations at Pachacamac reveal little about agricultural practices within and near the site. However, some modern agricultural practices can be traced back several centuries and even millennia. Such conserved agricultural traditions and tools can inform reconstructions of agricultural activities at Pachacamac. With this in mind, I chose to model a critical component of the Andean farmer’s toolkit: the chaquitaclla.

The Chaquitaclla

Andean farmers have used the chaquitaclla, or footplough, for thousands of years (Erickson 2000; Morlon et al. 1996; Denevan 2001). Farmers at Pachacamac also probably used the chaquitaclla. The chaquitaclla remains in use today, although the introduction of draft animals and farming machinery caused its popularity to diminish (Morlon et al. 1996; Gade and Rios 1972). Andean farmers employed the chaquitaclla, along with the rawkana (hoe) and waqtana (clodbuster), to completely transform the landscape through agriculture (Erickson 2000). Various forms of digging sticks, including the chaquitaclla, were also used for turning soil, planting, and harvesting crops (Morlon et al. 1996; Denevan 2001). The chaquitaclla, unlike any other tool developed in the Americas, employs pressure from the foot onto the tool’s taquilpo (footrest) to efficiently penetrate the ground with a blade. In doing so, the tool facilitates the process of breaking up compacted earth after periodic intervals of fallow. Because farmers permit fields to rest, the soil regains nutrients leading to improved crop production years later. However, fallowed soil hardens and becomes colonized by native grasses, impeding future planting. Farmers equipped with the chaquitaclla loosen the earth relatively easily. Today, as in the past, the use of the chaquitaclla typically requires a team effort (Morlon et al. 1996; Gade and Rios 1972). Groups of two to five individuals work together to dig up, loosen, and turn over sod (Figures 1-3). Men typically employ the chaquitaclla to loosen the earth while women following them to turn over the dirt clods or cut sod (Donkin 1970). In addition to using the chaquitaclla to process fallowed land, they also use the tool for planting, harvesting, cutting sod, and general digging purposes (Figure 4) (Denevan 2001).

The relatively simple tool consists of a shaft (chanchaca), handle (kcumo), footrest (taquilpo), and blade (chanka) all bound with leather (yauriga) (Figure 5). Although Andean blacksmiths fashion blades from the steel of a broken truck leaf spring today, in the past, farmers used ground and/or chipped stone or fire-hardened wood blades (Erickson 2000). The blade on some contemporary chaquitacllas can be reversed to adapt the tool for different soil types. While llama or deer leather traditionally provided the binding during prehistory, cattle leather is used today (Gade and Rios 1972). Andean farmers fashion the three wooden parts of the chaquitaclla from local trees specifically cultivated or collected for this purpose (Erickson 2000). Despite its apparent simplicity, the chaquitaclla can vary widely in shaft length, blade size, and style depending on local traditions or its intended use (Figures 6-7).

Process: 3D Modeling in Maya

Given the absence of pre-Columbian agricultural tools at Pachacamac, I could not model an authentic coastal chaquitaclla. However, Dr. Erickson possesses a modern chaquitaclla which resembles those used by farmers living at the site in the past. I visited the Penn Museum to study and photograph the chaquitaclla to better inform my modeling process. There, I analyzed the object from multiple perspectives, taking note of how the object was constructed (Figure 8). I particularly focused on the complicated leather strapping which holds the chaquitaclla together (Figure 9). In addition, I observed the highly variable wood textures found throughout the chaquitaclla. The shaft and handle contained regions of polished wood produced by hands rubbing the surface over years of use (Figure 10). Aging had also darkened the wood and raised wood grains, resulting in a roughened appearance in other regions. I also observed several areas with relatively deep cracks. Throughout this process, I took dozens of photographs to use as reference while modeling the chaquitaclla. After this initial study, I began to produce my 3D model in Maya.

The basic structure of the chaquitaclla consists of the main shaft, the curved handle, the footrest made of wood, and the steel blade which are bound with leather straps. I imported reference photographs depicting the entire chaquitaclla and reference photos focusing on the individual components. I modeled each of the four components separately, beginning with the curved handle. First, I created a polygon-based cylinder in Maya. I then extruded the face of the cylinder with strategic rotations of the faces to produce a curved appearance. However, my initial attempt produced an unnatural and unrealistic curved handle (Figure 11). I had mistakenly selected the entire object before extruding, which disturbed the rest of my manipulations of the cylinder. I restarted with a new cylinder and extruded with soft select until I achieved a more natural looking handle (Figure 12). While possibly creating a similar or better effect with the Bezier Curve Tool, I chose to focus on producing the other parts of the tool before perfecting the handle with this approach.

After making the handle, I moved on to modeling the footrest. I first created a cube which I scaled to produce a rectangular prism. Then, I added edge loops and moved some vertices such that I could extrude parts of the footrest (Figure 13). This approach mimics the concave section in the footrest where the leather strap is located (Figure 14). Then I used soft select to extrude the lower part of the footrest to achieve the more detailed structures in this section (Figure 15). Although apparently simple, achieving the natural appearance of the footrest produced over years of use was difficult.

Making the main shaft and blade shapes were simpler. I began modeling the main shaft by creating and elongating a cylinder. However, I struggled to achieve the naturally worn, curved ends of the shaft. To simulate this, I target welded half of a sphere to the cylinder. For the other end, I soft selected the central vertex of the cylinder face and translated it to produce a less dramatic curved appearance. To model the blade, I first I elongated a cube and added edge loops. With the additional edge loops, I formed the triangular portion of the blade by translating carefully selected vertices. I then used soft select to create a curved appearance for the edge of the blade.

After producing the basic shapes, I began to apply the proper material, texture and bump mapping. Today and in the past, the chaquitaclla is made of wood from local trees. After many years, natural aging and the effects of wear alter the appearance of wood. Throughout the tool, aging darkens the wood over time. Where the tool contacts other surfaces, such as human hands or soil, the wood acquires a polished appearance. The chaquitaclla developed this polished appearance on the curved handle, upper portions of the stalk, footrest, and blade. Unfortunately, modeling wood and other organic materials is a difficult task. Tileable textures, which regularly repeat one image, cannot easily mimic natural variations in texture, color, and pattern. I assumed I could apply a sample texture from my reference photographs to my model. Successfully creating a tileable texture from a reference image proved to be too difficult. Therefore, I tried a variety of wood textures, including Maya’s native wood texture along with others from the internet, before finding one which sufficiently matched that of the chaquitaclla (Figures 16). I considered color as well as grain size, shape, and density when making my selection. I created a material using the wood texture that best matched the chaquitaclla and added bump mapping to produce the 3D appearance (Figures 17-18). I used this material for the handle, footrest, and stalk components (Figure 19).

After selecting the wood texture and manipulating the color, I emulated the worn appearance of the tool. I chose to model the weathered and worn wood rather than new construction material (Figure 5). When creating my wood material, I initially used the basic Lambert material which produces a matte appearance. However, I decided I wanted a shinier effect to more effectively mimic the wear on the tool. Therefore, I used a Blinn material. Interestingly, this texture is more commonly used for metal materials.

The chaquitaclla blade employs a different material than the rest of the tool. In my modeled chaquitaclla, the blade was made from a broken metal truck leaf spring. However, in the past, farmers fashioned blades out of ground stone, gourd fragments, or fire-hardened wood (Gade and Rios 1972). I chose to model my blade as if made of fire-hardened wood. Therefore, I applied a manipulated version of the original wood texture I use for the rest of the model. I lightened the color and increased the shininess, an effect which a TA showed me, to mimic the effect of fire-hardening (Figure 20).

Having sufficiently modeled the shape and texture of the main chaquitaclla components, I positioned the pieces together in Maya with the Translate and Rotate tools (Figure 21). Then, I began modeling the leather straps. As expected, this difficult process required substantial effort. Although possible to model ropes and straps with various sculpting tools, I opted to use another, less time-intensive approach. To mimic the appearance of straps without actually modeling them as a single component, I created many small, thin cylinders (Figure 22). Each cylinder mimics the appearance of part of a strap wrapped around the components of the chaquitaclla. I used the Duplicate and Translate tools strategically place many of these small cylinders (Figure 23). Then, I used soft select to tilt some of the cylinders to produce the appearance of a strap naturally wrapped around the tool (Figure 24). I also variably scaled the cylinders to give them a less uniform appearance. Satisfied with this approach, I applied a similar procedure with the straps on the footrest (Figure 25). After duplicating and placing several cylinders on the footrest, I realized these poorly represented the more complicated strapping on the footrest. Furthermore, when I attempted to apply a texture, the texture was applied to the circular face of the cylinder rather than the sides. As a result, the texture would not be visible on the “strap” side of the cylinder. Therefore, I decided to start over.

For my second approach, I first created a cube and extruded it with rotations to form a blocky, oblong torus shape. My first attempt produced an odd shape which would certainly not work for my straps (Figures 26). I started over and created a U-shape by extruding a cube with rotations (Figure 27). Then, I duplicated and rotated the shape to mirror my original shape. I then removed the faces of each shape and used the Target-Weld tool to join the two shapes together. This produced a more symmetrical shape that I could use to mimic straps more effectively. I used a similar approach of duplicating, translating, and rotating many of these fake straps along the chaquitaclla (Figure 28). I carefully achieved the crossed-over strap appearance near the footrest by rotating the straps (Figure 29). While I chose not to model the knot which secures the end of the long leather strap, this might be attempted using an approach employed by classmate Davies Lumumba. He produced a simulated rope appearance by twisting several polygons together, which might allow me to simulate a knotted appearance. I also found obtained a more natural appearance by soft selecting faces in the center of the strap and scaling them down. This replicated the appearance of tension which would have been produced when wrapping the straps around the tool.

Having effectively simulated the shape of the straps, I applied the correct material, texture, and color. Although the reference chaquitaclla employs cattle leather for the straps, rather than the historically accurate llama leather, the two types of leather probably appear similar when aged. Like the issues faced with wood, texturing organic materials such as leather proved difficult. First, I attempted to use Maya’s leather texture which permits some degree of customizability. However, Maya’s leather appeared more like the terrain of Mars than leather (Figure 30). Therefore, I again turned to the internet to find leather textures. Unfortunately, many did not sufficiently match the leather straps. Online leather textures probably failed because they were created to model dyed leather clothing or shoes. I also tried to use other textures, such as skin and even wood, in the hope that these would be similar. Eventually, I did use a skin texture, although I manipulated its color (Figure 31). With the final texturing process finished, the chaquitaclla model was complete.

Finally, I decided to begin positioning a human model with the chaquitaclla. Simply modeling the chaquitaclla itself does little to inform the viewer about its use and importance in the daily lives of individuals. In addition, because I hoped to provide insight into the lives of ordinary individuals, I decided to model a person actively partaking in their daily activities. In doing so, my visualization of the past not only provides contexts for objects but prioritizes people.

First, I used the combine function to join the various components into one object in Maya. This facilitated future manipulations because I could scale, move, and rotate a single object. Then, I imported the male human rig and 3D digital model provided by the TAs into the scene with the chaquitaclla. I then scaled the chaquitaclla to what I thought was an appropriate size for the male model (Figure 32). I struggled to manipulate the human rig effectively, despite hiding the model’s clothing to reveal the rig components. After asking a TA for help, I realized I was translating and rotating the bones within the human rig. This manipulated the structure, whereas the blue arrows move the rig without distorting its structure. They showed me settings I could change in Maya which facilitate rig manipulations.

After learning this, I began positioning the human model with the chaquitaclla based on images of farmers using a chaquitaclla. First, I moved and rotated the foot to rest on the footrest as if the human model was about to push the blade into soil (Figure 33). When trying to position the hands, I ran into several problems. As I attempted to move one hand to hold the handle while the other grasped the main stalk, I achieved unnatural positions. A TA showed me that I could change a setting in Maya which moved the rest of the arm automatically when I moved the hand. This helped me achieve more natural, accurate hand positions on the chaquitaclla. Then, they showed me settings to move the fingers, rather than individually select each one with the mouse (Figure 34). I used this feature to increase the spread and curl of the fingers to better emulate a grasp on the tool. Overall, the model’s grasp and foot position begin to emulate how Andean farmers use the chaquitaclla in their everyday life.

Results and Future Directions

Despite facing anticipated and unexpected difficulties during the 3D modeling process, I successfully created a 3D model of a chaquitaclla which begins to approach realism (Figures 35-36). As expected, one of my most significant challenges involved texturing my 3D model. However, positioning my human model with the chaquitaclla proved much more difficult than originally anticipated.

In addition, my model remains far from perfect. After spending a significant amount of time positioning the chaquitaclla with the human model, I realized I modeled the stalk component of the tool to be far larger than the stalk component in the reference images. The chaquitaclla I modeled only extends to about one meter including the blade. Therefore, the chaquitaclla stands only slightly taller than half the model’s height. Similarly, the shaft thickness appears too wide and does not taper towards the end. When I attempted to resolve this issue, I realized that not only would I have to manually adjust several components in my original model, but I would also have to re-position the model with the amended chaquitaclla. This small change requires a large amount of time. Furthermore, several examples of chaquitaclla with longer stalk components exist. Modifying the length of the stalk component can change the tool’s utility in specific tasks or help adapt the tool to human users with variable statures.

Furthermore, I had hoped to model an entire 3-person team plowing a field. Given the challenges I faced positioning one human model, this quickly became unfeasible. To model a field, I would need to learn terrain mapping and a variety of other skills. Focusing on the chaquitaclla itself, I chose not to model a field. Unfortunately, this limits the ability of the viewer to really understand the chaquitaclla in its context.

Not only had I planned on modeling one team working on a field, I had also intended to depict multiple teams working on a field communally. To do this, I would vary my original chaquitaclla model using free-form deformation such that each team used a slightly different tool. Unfortunately, the method employed to simulate the straps would make this difficult and time-consuming to achieve. After manipulating any component of the chaquitaclla, I would have had to adjust all twenty solid polygons used to simulate straps. If known in advance, I might have learned the more time-intensive sculpting method instead.

Although adequate, the textures remain imperfect. After listening to other student presentations, I became aware of UV mapping approaches to texturing as well as a site where students obtained free textures. In the future, I would attempt to implement these approaches in my own model. After browsing the PBR website for wood textures, I failed to find a texture which improves upon the one I already selected. In addition, the site only provides leather textures for use on leather furniture which do not work for the leather straps. Given more time, I would learn how to create my own tileable textures using the textures from my model as references.

In addition to these adjustments of my 3D model, several ambitious future projects could incorporate or improve my model. Inspired by another student’s project which depicts pottery making from clay to finished product, I could imagine a similar project illustrating agricultural production. To highlight the chaquitaclla, I would begin the visualization with a field of mature crops and their harvest. Then, I would depict the passage of several years as the field remains in its resting period. Afterwards, I would model the use of the chaquitaclla to till the now grass-colonized soil after the fallow period. This approach could include an animation depicting the chaquitaclla piercing the ground, plying up and flipping sod, and exiting the soil. Alternatively, one could model the construction process of the chaquitaclla step by step including chopping down a tree or collecting sufficiently large branches, cutting and carving the wood components, fire-hardening the blade, and so on. While time-consuming, these approaches would provide viewers with a better glimpse into the daily life of individuals. A similarly ambitious project could illustrate the chaquitaclla’s efficiency when compared with draft animals in different environments. Viewers could select different terrains and soil types, then observe the rates and effectiveness with which the two methods process a field. This visualization conveys the importance and utility of the chaquitaclla in the past, when no draft animals existed in the Andes, and today, when some Andean farmers still choose to use the chaquitaclla.

On a less ambitious note, I hope my model will be incorporated with other student’s projects and with the larger Pachacamac project. Stacey Fox’s project, which modeled the rawkana or Andean hoe, also helps illustrate the traditional agricultural technology of the Andes. In addition, the chaquitaclla could be imported to populate a farmer’s storeroom or house if the dwellings of ordinary individuals in Pachacamac are depicted in the future.

Conclusion

Through this project, I effectively 3D modeled an important component of Andean agricultural technology. The development of 3D modeling skills throughout this process involved a steep learning curve, but I am satisfied with the results. Viewers of my model may better visualize the daily activities of ordinary individuals at Pachacamac. The activities of ordinary individuals, who provided the bulk of agricultural and other labor, supported Pachacamac as a prominent ritual and ceremonial center. By including agricultural technology along with the monumental architecture, colorful textiles, and intricate pottery, viewers will receive a more complete picture of life at Pachacamac. Furthermore, viewers will gain an appreciation for ingenuity and practical knowledge of Andean farmers. Despite being faced with a difficult environment, their technological innovations transformed the landscape of the Andes and continues to support large populations of individuals.

References Cited