Narrating from landscape in Andean archaeology: The problem with the suni natural region

Introduction

Technological advances have made it ever simpler to obtain accurate measurements of absolute geographic location (altitude, latitude, longitude). Yet to locate archaeological evidence and to narrate the precolonial past, establishing precise geographical coordinates is just the beginning. Ecological zonation models allow archaeologists to categorize sites, artifacts, and other remains, and to link them to narratives about past events, practices, and processes. However, uncritically “placing” sites within zonation models, without a clear understanding of how these were constructed and what data supports them, can result in misleading explanations of the past, an issue rarely discussed. In mountainous regions such as the Andes, there are at least two ways to define ecological zones: those that apply Western scientific criteria and those based on categories used by local people (Mayer, 2002: 244). This article deals with the latter, to consider how indigenous and local landscape knowledge may be incorporated into archaeological explanation. To do this, it examines a specific case within one of the most popular Peruvian zonation systems: the concept of suni, one of the “Eight Natural Regions” identified and publicized by the renowned Peruvian geographer Javier Pulgar Vidal (1911–2003) as part of his geographic model of Peru’s national space. This concept is frequently deployed in archaeological studies to explain land use and landscape organization in diverse parts of Peru and in various precolonial periods from a supposedly indigenous perspective.

Pulgar Vidal conceived of the Peruvian landscape as a series of vertically stacked zones, which he called “Natural Regions.” These zones are distinguishable mainly by altitude, but also by climatic features, presence of flora and fauna, and significantly, by cultural features: “traditional” human land uses (particularly related to cultivated crops) and toponyms in indigenous languages (Pulgar Vidal, 1987). Pulgar Vidal’s model explicitly presents the major subsistence and economic activities associated with each “Natural Region.” He argued that this landscape model was applicable to the entire national space and called it an example of “traditional indigenous knowledge” (Pulgar Vidal, 1941: 149, 160). Pulgar Vidal distinguished his model from earlier simplistic coast-highlands-jungle models, which he later argued to be a legacy of the Spanish conquest, created without incorporating local knowledge and traditions (Pulgar Vidal, 2014). The term suni has several meanings in Aymara and Quechua, but Pulgar Vidal defines suni as the zone located 3500–4000/4100 m above sea level (hereafter m.a.s.l.), just below the puna and just above the quechua. ¹ Suni is characterized as cold and dry, with strong diurnal temperature variation and seasonal precipitation. High-altitude crops like quinua, oca, and olluco may be cultivated in this zone, and pastoralism is also common (Pulgar Vidal, 1941: 155–156). Until recently, this model was used in territorial planning and included in national education curricula (Zimmerer and Bell, 2013).

Preliminary exploration of the definition of the term suni alerts us to complications surrounding Pulgar Vidal’s use of the term. In an Aymara-Spanish dictionary from 1612, the Jesuit Ludovico Bertonio included the following equivalence: “puna [=] suni” (Bertonio, 1612: 388), and other associated definitions, such as “suni [=] tierra desploblada (uninhabited land)” (Bertonio, 1612: 327). Four centuries later, Arnold, Jiménez, and Yapita suggested a definition of the Aymara term: “Suni. The highest part of the altiplano. Puna in the Quechua language” (Arnold et al., 2014: 255). ² These two definitions summarize one of the conclusions of this article: that suni is equivalent to puna and not an independent zone or “Natural Region.” In addition, both definitions are related to Bolivia and to the Aymara language, while Pulgar Vidal’s model pertains to Peru, begging the question of whether it is possible to generalize Bolivian Andean evidence for the rest of the Andes, or at least for Peru. Pulgar Vidal’s model thus exemplifies the complications that ensue when detailed ethnographic research on specific places is subsumed into a grand regional- or national-scale narrative.

This article explores the case study of the original creation and later application(s) of the suni concept within Andean(ist) archaeology, to address two major theoretical issues. The first draws on literature on landscape ethnoecology to highlight the challenges of understanding, recording, and using indigenous and local knowledge about landscape. The second relates to the idea of intra-regional variability. An interdisciplinary field such archaeology, which assimilates data and theories from ethnography, geography, and history, among others, confronts the persistent dilemma of attempting to generalize without erasing difference. The archaeologist endeavoring to understand landscapes of the past confronts both difficulties. After presenting these broader issues, this article considers the historical context of the original creation of the suni concept through analysis of the milieu in which Pulgar Vidal worked. It then analyzes major patterns in the uses of suni in both the ethnographic and archaeological literature, presents critiques of Pulgar Vidal’s model by ethnographers, and discusses recent archaeological applications of this concept. Finally, the conclusions summarize some of the challenges faced by archaeologists narrating the precolonial landscape and present suggestions for better ways forward.

Landscape ethnoecology and intra-regional variability in archaeological interpretation

The value of Indigenous and Local Knowledge (ILK) for landscape-based studies is of growing interest in diverse areas of territorial management and governance, with many studies focusing on South America (Adade Williams et al., 2020). While the majority of this work analyzes applications of ILK for current and future sustainability and conservation, applications for archaeological research are also evident (Stump, 2013). Landscape ethnoecology is an interdisciplinary framework useful for understanding ILK about landscapes, which specifically seeks to study “landscape as perceived and imagined by the people who live in it, the land seen, used and occupied by the members of a local community” (Johnson and Hunn, 2012: 1). This goes beyond study of toponyms and specific places (sometimes called ethnogeography), or study of specific species of plants and animals (ethnobiology), to identify “folk ecotopes,” understood as “elements distributed repetitively across the landscape, conceptual elements that constitute the biotic and physical space in which people live” (Johnson and Hunn, 2012: 2). Specifically highlighted are the “features of the landscape useful for people making a living off the land” (Johnson and Hunn, 2012: 3). In Peru, landscape ethnoecology studies have focused on the Amazon, with less attention to the Andes, although several studies examine landscape units or “ecosystem-like concepts” prevalent in Andean agropastoral spaces (Mayer, 2002: 244; see also Boillat et al., 2013; Zimmerer, 1999). While proponents of landscape ethnoecology recognize that landscape knowledge develops over time and has deep historical roots, how exactly current knowledge can be applied in archaeological contexts is a complicated issue (Johnson and Hunn, 2012: 6).

A second dilemma arises when site-specific local knowledge of landscape is generalized to broader regional or even national scales. Considerable effort has been devoted to identifying the peculiarities of the Andean through deep ethnographic research in remote communities, but few discuss the difficulties of subsuming this diversity within a grand narrative (e.g. Why can’t an example from Bolivia or northern Argentina be used to analyze a Peruvian case? How might we move from describing a specific case to describing traits that are pan-Andean?). Related to this issue of intra-Andean variability, we must recognize the enormous influence of anglophone ethnographies from southern Peru on archaeological interpretations for the northern part of the country. With what criteria can the Cuzco ethnographic experience be used to understand archaeological material from Ancash, Cajamarca, or La Libertad? (See Figure 1 for map of places mentioned.) The problem of understanding landscape knowledge (or folk ecotopes) is thus combined within the perpetual search for how to situate local or regional peculiarities in relation to the whole.

OPEN IN VIEWER

Figure 1.

Map displaying locations mentioned in the text.

While archaeological excavation continues in the Andes, new discussions on the possibilities and limitations of using local landscape knowledge and of regional comparison are lacking. Close examination of the suni concept, a term Zimmerer (1996: 138) found to be “inconsistently applied” and “difficult to define,” will illuminate some of the main obstacles faced by archaeologists in the attempt to understand, order, and narrate the precolonial landscape. To this end, this article explores the genesis and later application(s) of the suni concept within Andean(ist) archaeology and ethnography.

(Dis)playing (with) the ethnic voice: Tello, Pulgar Vidal, and the origins of suni

Julio C. Tello (1880–1947), the most famous Peruvian archaeologist of all time, was born in a campesino community called Suni, in the highlands outside of Lima (Huarochirí province), where his surname is still one of the most common (Lecaros, 1996: 290). Pulgar Vidal, one of the best-known Peruvian geographers, was born in the town of Panao, in the Pachitea province of Huánuco. Neither came from Lima, Peru’s capital, which helps explain their respective positions regarding the production of academic knowledge. The two met at the Catholic University in Lima, where Tello was Pulgar Vidal’s professor.

Much literature is devoted to Tello (e.g. Burger, 2009; Daggett, 2016; Padilla Deza, 2018), so here we only emphasize that one of the original features of his work was that it tried to incorporate the indigenous voice. This represented a dramatic shift in an academic setting dominated by Lima-born scholars. With respect to Andean landscapes, Tello’s approach was more sophisticated than that of most of his colleagues, probably given his personal experience in the highlands (e.g. Tello and Miranda, 1923). It would be an exaggeration to call his approach decolonial, but the issue was in the air: he wanted to interpret the precolonial past by drawing on ILK. This posture in a university professor impressed the Huánuco-born Pulgar Vidal, who accompanied Tello on fieldwork trips.

Pulgar Vidal applied Tello’s method to landscape classification. In the 1930-40s, several classifications for the Andean landscape were being used (see Pulgar Vidal, 1941); however, Pulgar Vidal insisted on a more detailed version, one that emphasized both the complexity of the Andean landscape and ILK (Zimmerer and Bell, 2013). He developed a model of the “Eight Natural Regions” of the Peruvian national territory (described above), conceived as series of stacked tiers, taking on the appearance of a layer cake. As part of this classification, he incorporated a “Natural Region” not previously mentioned as such in geographic texts: suni. In an early publication Pulgar Vidal included helpful “Necessary clarifications,” indicating that the nomenclature employed “… is, in part, that taught by traditional indigenous wisdom, that still survives, and it is customarily used by the Indians from the diverse regions of the country, that are the toponyms: yunga, quechua, suni, jalca or puna, janca and ruparupa” (Pulgar Vidal, 1941: 149). He also recognized that he originally identified the “Eight Natural Regions” along a single transect crossing the northern part of the Department of Lima (Chillón valley, province of Canta, La Viuda cordillera, etc.) (Pulgar Vidal, 1941). (See Figure 2.) With these clarifications he established the indigenous source of his nomenclature and recognized that his data came from a specific area of the central Andes. In the following 10 editions of the book the first point was emphasized, becoming its most attractive asset. Meanwhile, the second clarification became increasingly blurry.

OPEN IN VIEWER

Figure 2.

Cross section of Perú, from Lima to Pucallpa (west–east), showing the stacked or layer cake conception of the “Eight Natural Regions.” Redrawn and translated from Pulgar Vidal (1941).

The empirical basis of Pulgar Vidal’s model has not been explored in detail. However, it has been suggested that Tello proposed the use of the term suni to Pulgar Vidal, in order to establish a clear difference with previous nomenclatures (Córdova and Bernex, 1999: 16). Pulgar Vidal also followed Tello’s example when he directed the fieldwork of his Catholic University students. Like Tello, Pulgar Vidal encouraged students to research their hometowns and write descriptive geographical essays about them. The best essays were published in three edited volumes. Within this fascinating collection, the essays most relevant here are those that include the term suni, particularly Morocho (1938) on the Chillón valley (central Lima); Mercado (1941) on the district of El Mantaro (Junín); and Iturrizaga (1941) on Tupe (Yauyos, southern Lima). In none of these works was the category suni observed in the field; rather, it was incorporated later according to what the authors themselves termed “Pulgar Vidal’s theory.” The most intriguing article is Iturrizaga (1941), which was written in both Spanish and Cauqui-Jacaru, a variety of Aymara from Tupe. The version of the text written in the indigenous language includes the word suni several times, but it seems to be proposed by the author and not by the interviewees. On further examination of the use of suni in Jacaru, we find that Belleza’s dictionary (1995: 164) defines suni as long or elongated (preferably horizontally), but mentions nothing related to ecology or landscape. Similarly, Yolanda Payano Iturrizaga, an official translator of Jacaru and native of Tupe, confirms this definition of suni (long) and discards any use of the term as an ecological zone, mentioning instead that the term shallqa is used to refer to puna. In sum, the supposedly vernacular basis for suni as landscape nomenclature is absent.

Finally, how did Tello himself use the term suni? It is absent from his most ethnographic text about the Lima highlands (Tello and Miranda, 1923). Complicating the situation, in three of his most well-read archaeological publications he equates suni with puna. In his early synthesis on Ancient Peru, he wrote, “The aborigines distinguished these zones with special names. They called the cold mountains and plateaus Suni or Puna; Keshua to the warm lands of the ravines and valleys; and Yunka to the hot [zones] …” (Tello, 1929: 12). He later observed, “… there is an intermediate warm zone, that the Indians call keshwa, which is not as cold as the Puna or suni, nor hot like the deep dry ravines that they call Yunga” (Tello, 1942: 9). Finally, his book on the precolonial site and society of Chavín, posthumously co-authored with his collaborator Mejía, states: “A region like Pukcha, endowed with enormous natural resources in its three climatic zones—yunga, keshwa and suni or puna—, must have drawn the attention of the Andean man to establish his settlements …” (Tello and Mejía, 1960: 42). In these examples he always uses suni as an equivalent of puna (e.g. Arnold et al., 2014; Bertonio, 1612); however, he refers to “the aborigines” and “the Indians,” like Pulgar Vidal, to validate his findings. This validation is not only in reference to ecological zones in the present, but also in relation to the remote past, which was Tello’s actual target. Overall, it is surprising that, despite having suggested the term to his student (Córdova and Bernex, 1999: 16), Tello was still equating suni with puna, never considering it as an independent entity.

The definition and equivalences of suni varied throughout re-editions of Pulgar Vidal’s textbook. He started by equating jalca and puna (Pulgar Vidal, 1941: 156), then switched to jalca or suni (Pulgar Vidal, 1967: 12, 14). Next, he proposed a combined solution trying to adapt to latitude, a feature that he never handled especially well: “jalca is used in the north [of Perú] instead of suni” (Pulgar Vidal, 1973: 244, n. 1), and, “Sometimes, the term ‘jalca’ and also ‘jallca’ is used in central [Peru] as a synonym of ‘puna’” (Pulgar Vidal, 1973: 244, n. 2). Finally, he used suni as equivalent to jalca (Pulgar Vidal, 1987: 79). Unfortunately, previous studies of Pulgar Vidal’s works have not explored the potential significance of these terminological short-circuits (e.g. Zimmerer and Bell, 2013). ³

The presence of suni in archaeological and ethnographic publications

In the Andes, ethnography and archaeology are historically connected: data collected by ethnographers has been used as a source to propose analogies for precolonial targets. In that sense, it may be highly illustrative to compare differences in how both groups treat specific topics. In the case of suni, we may start by examining the uses of this concept in both fields. In ethnographic work, understood as research conducted through close observation and dialogue with local populations, the term suni is absent. Systematic review of ethnographic publications has not revealed the use of this term among Andean populations of the 20^th–21^st century. (See Supplemental Material for a list of authors/works reviewed.) There are rare exceptions to this trend, (e.g. Bolin, 1998: 17 [Cuzco]), but in these cases suni is never reported as a word said by interviewees, but as reference to other authors or through Pulgar Vidal’s influence. Personal communication with three well-known scholars from different academic backgrounds with extensive Andean experience—César Itier (Andahuaylas, Ayacucho, Cuzco), Juan Javier Rivera (Lambayeque, Lima), and Joshua Shapero (Ancash)—confirm that the term suni is not used by local populations to refer to an ecological zone. Finally, in our own experience of almost three decades working in northern Peru we have never heard the term. The term suni appears in several ethnographies from Bolivia, but as a synonym of puna. Meanwhile, in the case of archaeologists, who also travel to remote Andean places and interact with local populations but focus on precolonial evidence, the term suni appears frequently, as an ecological zone useful for classifying sites (see Supplemental Material for references including suni).

Additional information can be derived from bibliometric analysis of the presence of three terms related to Andean ecological zones (puna, punas, and suni) in the most emblematic journal of Peruvian humanities, the Revista del Museo Nacional. This journal of 51 volumes (1932–2016) was created and directed for several decades by the ethnohistorian Valcárcel, a scholar who fundamentally shaped Andean studies. While this journal is currently out of circulation, during its golden era (1930s–1980s) it published texts written by the most renowned archaeologists, ethnographers, and historians of the Andes. It thus serves as a database to explore trends in landscape terminology across fields. Figure 3 summarizes analysis of this publication, showing the massive presence of the terms puna and punas, always as synonyms, and the rare presence of the word suni.

OPEN IN VIEWER

Figure 3.

Presence of the terms puna, punas, and suni in articles in the Revista del Museo Nacional (1932–2016). From 1932–1964 the Revista del Museo Nacional published one issue per year (except for the double issue 19–20 published 1950–1951). After number 34 the publication becomes increasingly inconsistent.

Considering the rare examples of suni, we can see that the term is used in four ways. First, suni appears as equivalent of “long” or “big” (1941, 1942, 1955, 1970). Second, it is included as a synonym of puna (1946, 1971) (as described above). Third, it serves as part of a place name or a surname (1950–1, 1953, 1954, 1963, 1974). In this third group the term appears 18 times as the name of Tello’s hometown. Fourth, suni appears for the first and only time as an ecological zone in the last issue of the journal (2016), in an ethnographic article on Ayacucho (Muñoz, 2016). The same author published a similar report on the same topic and village (Muñoz, 2014) where suni does not appear. It is possible that this represents Pulgar Vidal’s influence and not evidence directly collected by Muñoz. Several final points about the Revista del Museo Nacional are also relevant. This journal included multiple texts by the linguist José Farfán, who collected, transcribed, and published in Aymara, Jacaru, and Quechua from different parts of Peru, none of which include the term suni (Farfán, 1944, 1946, 1947, inter alia). In addition, the special issue devoted to Andean ethnography (1971), which includes nine works by well-known scholars about different villages, does not include suni as an independent zone distinct from puna. Lastly, Duviols’ (1973) landmark article published in this journal, which combines ethnographic and historical information from the central Peruvian Andes, and which has had enormous impact on precolonial archaeology to explain the different lifeways of the inhabitants of the quechua and puna zones, does not include the term suni. This final detail adds an additional facet to the discussion: suni is not only absent in the ethnographic literature, but it has also not been recorded in colonial sources (Zimmerer and Bell, 2015).

Several hypotheses may be drawn from this preliminary analysis of the ethnographic and archaeological literature. First, the model that supports the existence of a suni “Natural Region” (hereafter the suni model) seems to be popular among archaeologists, but not ethnographers. Secondly, the term does not appear to exist in the field, or at least the term has not been in common use by Andean populations during the last half century. The third hypothesis is that, with rare exceptions, there is limited dialogue between local populations and archaeologists regarding landscape. Fourth, perhaps archaeologists do not closely read ethnographic publications in Spanish when considering these kinds of topics.

Why then is the suni model so popular among Andean archaeologists? It appears that archaeologists need to deal with landscape names and categories, but either do not possess the necessary tools for this project or do not devote sufficient time to it. Pulgar Vidal’s proposal, which is simple in comparison with others (e.g. Tosi, 1960), and lists a series of stacked zones identifiable by a diverse suite of environmental and cultural features, provides a key for directly linking altitudinal ranges with more complicated narratives about landscape, resources, and land use activities. For instance, if we take the site Wañucoto, which is located within the altitudes associated with the suni level in Pulgar Vidal’s textbook (3500–4000/4100 m. a.s.l.), we can use Pulgar Vidal’s model to add a whole series of features to the narrative about precolonial Wañucoto even without excavation (e.g. Wañucoto inhabitants grew quinua, herded llamas, etc.). However, before discussing the potential problems with this kind of operation, we will present an early ethnographic critique of the suni model.

Ethnographic critique of Pulgar Vidal through a dialogic approach to landscape

Anthropologists and ethnographers working in the Andes have rarely used the concept of suni and have also critiqued Pulgar Vidal’s model in direct and indirect ways. An interesting example is found in the work of anthropologist John Murra. While Pulgar Vidal was cementing his model, Murra focused on the Inca empire. Murra placed great emphasis on the relation between people and landscape in both his doctoral dissertation and his multidisciplinary team project studying Inca provincial life at the site of Huanucopampa and the region of Huánuco as a whole (Murra, 1972, 1977). Here, he developed the concept for which he is best known: “The vertical control of a maximum of ecological levels” (Murra, 1972). This work was the result of a spectacular methodological innovation in Andean studies: the systematic comparison of early colonial documents on land use with ethnographic and archaeological evidence. Through this process Murra’s team noticed the value and limitations of Pulgar Vidal’s model.

Murra provided his collaborators in Huánuco with a theoretical frame for collecting detailed information on the structure of indigenous communities (Fonseca, 1966: 22). One of the members of the team was ethnographer César Fonseca, from Llata, Huánuco, who published two reports based on fieldwork in the community of Cauri and the Chaupiwaranga area (1966, 1972). Fonseca spoke with people from both areas about local names for the different ecological zones, among many other topics. Comparing Fonseca’s reports with those published by Pulgar Vidal’s students, it is clear that Fonseca carried out more sophisticated interviews, asking the same sets of questions to people living at different altitudes within the same area. In these interviews he asked about local landscape terms instead of imposing his own. Fonseca never included the names of the interviewees, but he always identified where they were from, and he also incorporated direct quotations from local testimonies in his publications.

Fonseca began his first report by situating his study area in relation to Pulgar Vidal’s model, identifying it as kechwa. He then performed a deeper multivocal analysis: “For the people of Cauri, Chaupiwaranga is in a kechwa zone, and its inhabitants are called kechwaruna or kebrada runa [people of the quechua or people of the valley], meanwhile, those from the valley know the people of Cauri as jalga runa [people of the mountain]” (Fonseca, 1966: 26). He was able to further distinguish two sub-zones within the kechwa and the jalga, thus questioning the supposedly diagnostic agricultural products used by Pulgar Vidal to define the “Natural Regions.” For instance, with respect to the tuna cactus (probably Opuntia streptacantha), which Pulgar Vidal identified as typical of the quechua, Fonseca registered two different varieties, a kechwa tuna and a jalga tuna (Fonseca, 1966: 28). In his second report, Fonseca expanded his critique of Pulgar Vidal’s model:

During my fieldwork in the “indigenous communities” of the different provinces of the Central-Northern part of the country I have observed that the campesinos use concepts in runasimi [the quechua language]: kichwa and jalka to refer to their home base. In other words, it is not about two geographical regions arranged parallel to the coastline, with exact and rigidly determined limits. These concepts appear to be rather elastic. (Fonseca, 1972: 318)

Here Fonseca is questioning the diagrams presented by Pulgar Vidal, where the “Eight Natural Regions” are drawn as a perfectly defined layer cake (these diagrams are repeated in textbooks referenced below: Burger, 1992; Quilter, 2014). Fonseca never registered the term suni among his interviewees. The only time the word appears is in a footnote comparing suni and jalka, although his table on ecological nomenclature seriously questions any potential equivalence (Fonseca, 1972: 319). In his study area the categories kichwa and jalka seem to work as relational and complementary elements: every community has a lower part (kichwa) and a higher part (jalka), and several crops have a kichwa and jalka variety (Fonseca, 1972: 322; see also Mayer, 1974: 31–32 for Tangor, Pasco Department).

Considering the information collected by his team, it is interesting to examine Murra’s opinion of Pulgar Vidal’s model. While Murra was fond of Tello’s work (Murra, 1982), his opinion of Pulgar Vidal evolved. He recognized the geographer as a pioneer of ethnoscience in the Andes, calling attention to the importance of the research that led to the initial model. However, he also observed:

Unfortunately, the various editions that have come out since then include no further work. I frequently urged Professor Pulgar to continue, because it seems to me, if you think about, for the late 30s, for somebody to do what very much later was done, the whole idea of an ethnoecology, of an ethnogeography, appeared only in the 50s in this country [USA]. (Murra, [1969] 2017: 7)

Pulgar Vidal never followed Murra’s advice and continued publishing new editions of his textbook without making significant additions to its empirical base. Murra never used the term suni in his publications. Interestingly, some scholars use the two models in parallel, seemingly uncritically, to discuss precolonial vertical socioeconomic organization using the “Natural Regions” identified by Pulgar Vidal (e.g. Quilter, 2014: 18, 32, 43–44).

Shades of suni in recent archaeology

Even given the above-mentioned complexities in ethnographic information, the concept of suni is frequently employed by archaeologists (see Supplemental Material). In most cases, it is applied simply as an altitudinal equivalence: a site located between 3500–4000/4100 m.a.s.l. is said to be found in the suni, supposedly adding ILK to altitude. Identifying the “Natural Region” in which a site is located means that Pulgar Vidal’s descriptions of human activities can be applied to the archaeological remains and hypotheses can be formulated about precolonial livelihoods and land uses. The problem is that the equation of altitude with “Natural Region” is often made uncritically. For starters, one of the most curious outcomes is that the altitude range attributed to suni varies between authors without strictly following Pulgar Vidal (see representative sample in Table 1).

OPEN IN VIEWER

Table 1.

Representative sample of publications that list altitude ranges attributed to suni, in chronological order.

suni (m.a.s.l.)	Reference
3500–4100	Pulgar Vidal (1941: 18, 11) [Perú]
3500–4100	Pulgar Vidal (1946: 105) [Perú]
3500–4000	Pulgar Vidal (1973: 90) [Perú]
3500–4000 (suni or jalca)	Pulgar Vidal (1987: 25, 79) [Perú]
3500–4000 (suni or jalca)	Pulgar Vidal (2014) [Perú]
3200–4000	Burger (1992: 20–1) [mostly Ancash]
3300–3910	Bolin (1998: 17) [Cuzco]
3500–3850/3500–4000	Parsons et al. (2000: 14, 32) [Junín]
3800–4000	Stanish (2003: 34) [Puno]
3800–4000	Janusek (2008: 38) [Puno]
3000–4000	Lane (2009: 178) [Ancash]
3600–3800	Wernke and Withmore (2009: 428) [Arequipa]
3000–4000/3500–4000	Lau (2011: 24, 44) [Ancash]
3000–4000/3200–4000	Lau (2013: 13, 14) [Ancash]
3000 and 3500–4000	Quilter (2014: 32) [all Perú]
3600–3850	Parsons et al. (2013: 10) [Junín]
3800–4000	Stanish et al. (2014: 10) [Puno]
3000–4000	Lau (2016: 29, 224) [Ancash]
3600–3900/3100–3600, 3600–4100/3000–4000	Lane and Grant (2016: 140 “upper suni,” 141 “Suni/Jalka,” 144 “Suni/Páramo”) [Ancash]

Given the general confusion represented in Table 1, in the following we examine several specific cases of the use of suni among influential archaeologists, to better understand why this concept has proven attractive for the narration of the precolonial past, and to show how quickly complications arise when this term is applied. These four cases, presented in chronological order, were selected because they each link the usage of suni to broader and more complex dilemmas relating to the interpretation of precolonial landscapes.

The first is from one of the members of Murra’s team, the archaeologist Matos, from Acobamba, Huancavelica. Like Fonseca, Matos was raised speaking Quechua, and was also curious about landscape nomenclature. In his book on Pumpu, Matos defines puna as the zone between 3700–4500 m.a.s.l. (Matos, 1994: 22, 29), and later in the text defines specific levels: llampu-jalca (3900–4200 m.a.s.l.) and jatun jalca, also called sallca, salga or jalca (4200–4500 m.a.s.l.). He assumes jalca and puna as equivalent while recognizing that the term puna is not used in Huancavelica (Matos, 1994: 42–51, also like Itier and Ortiz, 2019: 168). He does not mention the term suni. In a subsequent publication—the comprehensive two-volume work on the archaeology of the upper Mantaro and Tarma drainages of Junín, published by Parsons, Hastings, and Matos—the word suni is basically absent. While in the first volume the word puna appears 179 times and kichwa 515, suni is included just three times and is mostly used when referring to works of other people (Parsons et al., 2000: 14, 32). In volume two, puna appears 192 times, kichwa 349, and suni just once, as an equivalence of upper kichwa, “sometimes referred to as the suni” (Parsons et al., 2013: 10). It is notable that three authors with so much experience in the central highlands do not use the term.

The second case can be found in the work of Burger, where the term suni appears frequently. In his book on Chavin (Burger, 1992), he presents the layer cake diagram and performs the validation step:

The complex highland setting is economically described in the indigenous Quechua language of Peru by a series of terms which correspond to distinct zones of agricultural production. This native nomenclature has been adopted by Peruvian geographer Javier Pulgar Vidal and will be employed here. (Burger, 1992: 20, emphasis added)

Nevertheless, Burger includes an insightful observation about the layer cake model, something few of his colleagues notice, which relates to the potential shifting of the limits between these ecological zones due to climate changes during the Initial Period and Early Horizon (Burger, 1992: 24). While Burger does not explore this interesting point in more detail, upslope shifting of vegetation in Andean landscapes as a response to recent anthropogenic climate change has been documented (e.g. Morueta-Holme et al., 2015), and the impact of cooler temperatures during the Little Ice Age has been shown to result in glacial expansion and downslope shifting of ecological zones (Jomelli et al., 2009; see also Zimmerer and Bell, 2015). Clearly these observations indicate that the elevation ranges of ecological zones fluctuate over time, calling into question the direct application of the ranges published by Pulgar Vidal in other time periods.

The third case comes from Stanish’s research on Tiahuanaco, near the Bolivia-Peru border. Even citing the Aymara equivalence of suni and puna suggested by Bertonio (Stanish, 2003: 297), Stanish states: “I generally follow the work of Pulgar Vidal (n.d.), who divides the Titicaca Basin into two broad agricultural and ecological regions called the suni and puna” (Stanish, 2003: 33). He proposes limits for the suni, different to those used by Burger (see Table 1), and states that suni is the “upper limit of plant agriculture,” although shortly thereafter adds that potatoes are cultivated in the suni and puna. Later in the text, he declares:

The classifications of Pulgar Vidal, Troll, and so forth were not designed for anthropological research. The basic puna/suni distinction of Pulgar Vidal is a good first approximation of the broad agricultural/pastoral zones and is the most useful of the existing classifications. From a cultural perspective, however, a variety of distinct geographical zones are important in understanding the region’s prehistory. A very useful alternative to the ecological classifications is the indigenous categories of land. (Stanish, 2003: 35, emphasis added)

This text summarizes several misunderstandings. Pulgar Vidal explicitly proposed his model as based on “indigenous categories of land,” so it is not that this disciple of Tello did not intend his model to be used in anthropological research. As Murra stated, Pulgar Vidal was proposing ethnoscience. Even with this discussion, as in Burger’s case, Stanish does not go beyond the application of suni as an altitudinal marker (Stanish, 2003: 35–36, 38, 135). Moreover, in a book about Puno, an area with a large number of Aymara speakers, it is extremely confusing to refer to suni as an ecological zone that is not puna (see also Janusek, 2008: 38). Overall, this example suggests the (potential) incompatibility of general national-level landscape models with local, specific cases.

Finally, more recent developments in the archaeological application of Pulgar Vidal’s model can be found in Lau’s work on Ancash. Lau uses suni to describe the Ancash landscape, presenting a different altitudinal range than the aforementioned authors and locating many of the precolonial sites that he surveyed in that zone (Lau, 2011: 52, cf. with Lane and Grant, 2016: 140–141, 144). Unlike Burger and Stanish, he does not introduce the term as indigenous, although this is assumed. Ancash is an interesting area since it is close to the region where Pulgar Vidal proposed his concepts and to where Fonseca demonstrated its limitations. Like other authors, Lau (2011, 2013) classifies sites as suni based on altitude, but he also goes a step further. In his valuable book on ancient alterity, Lau explores Duviols’ (1973) proposed dichotomy huari/llacuaz (described above). This concept presents two different and complementary lifeways associated with the quechua and puna ecological zones. These zones are home to two distinct social groups, maize cultivators and herders, respectively (Duviols, 1997: 34). The huari/llacuaz dichotomy has provided an extremely powerful source to establish comparisons with the precolonial past: it associates landscape with economy, thus with ways of living. However, in an unexpected switch, Lau (2013: 84) located the llacuaz in the puna but placed the huari in the suni (contra Duviols, 1997: 34 and Parsons et al., 1997: 322, among others). Lau’s move raises several questions, especially given that the term suni is not currently used locally in Ancash (personal communications, Giner Aranda, Elvis Crisóstomo, George Lau; see also Shapero, 2019). Given that suni, according to Pulgar Vidal, is also called jalca/jallqa/hallqa (Pulgar Vidal 1973: 244, 1987: 79), a term which in Ancash is equivalent to puna, Lau unintentionally locates both the huari and the llacuaz in the puna. Quilter later refers to Lau’s work to explain that the Recuay people living in Ancash were “highly successful” at exploiting “the high altitudes between the suni and puna life zones” (Quilter, 2014: 174). Again, the application of Pulgar Vidal’s model causes enlightening short-circuits.

Conclusions

Through this exploration of the genesis and application of the suni model, we may define three major challenges faced by archaeologists when attempting to incorporate their sites and materials into landscape narratives that consider ILK. In the Andes, these difficulties are related to using Pulgar Vidal’s model analogically as a source to better understand the targets, which in this case are the landscapes and land uses of the precolonial past. We present these challenges or difficulties as a caution to archaeologists and as a provocation for additional research.

The first complication relates to intra-regional variability, more specifically ecological variation with respect to latitude. Despite his cautious beginnings, it is evident that Pulgar Vidal’s main objective was to generalize his model for the entire nation. However, one of the main dilemmas that haunted his scheme was Andean latitudinal variation, specifically the difference between paramo and puna Andes (Troll, 1968). Looking closely at the terminological complications in Pulgar Vidal’s publications over time with respect to jalca/suni/puna, it becomes clear that latitudinal variation is a major weakness in this model, which has resulted in a terminological conundrum (as evident in the work of archaeologists presented above). Any application of Pulgar Vidal’s model to archaeological interpretations must incorporate latitude: the Andes should not be considered a series of uniformly stacked zones in a layer cake formation across the entire range from south to north. At a micro level, such layer cake models of ecological zones do not necessarily reflect local realities, which may be quite complex (Itier and Ortiz, 2019; Zimmerer, 1999), and which may in fact be relational (Fonseca, 1966).

The second difficulty is that the limits between Andean ecological zones have been shown to vary over time, as mentioned by Burger (1992). Up- and downslope shifts of flora and fauna, and consequently human land uses, respond to changing climatic conditions. Historically important shifts occurred during the Little Ice Age (downslope with colder conditions) and are ongoing because of anthropogenic global warming (upslope in response to warmer conditions). The boundaries between Pulgar Vidal’s “Natural Regions” correspond to the period of his field observations (mid-20^th century) and are not directly applicable to other periods. In this sense, archaeologists should review climate models and reconstructions for the periods in which they are interested before attempting to equate an altitude measurement with precolonial activities (e.g. agricultural practices).

Finally, it is not an exaggeration to say that the human activities described by Pulgar Vidal for each zone have been understudied and grossly overgeneralized. While his descriptions can provide an initial hypothesis for human use of the landscape in the precolonial past, they cannot be the justification for more advanced explanations. Pulgar Vidal’s observations were historically contingent and must be applied with this qualification. Further complicating the scenario, the identification of types of productive activities does not necessarily tell us much about the historicity of the economy associated with varying prestige of ecological zones, the exchange value of the products from each, economic and power relations, or social organization over time.

Analysis of Pulgar Vidal’s model also allows us to further the discussion of how landscape ethnoecology may be pursued in archaeological contexts. Seemingly one of the most attractive features of Pulgar Vidal’s model for archaeologists is that it apparently recuperates and systematizes ILK (presumably of precolonial origin). The origins and applications of suni synthesize all of this but represent the weakest link in Pulgar Vidal’s model. New field studies and further systematization of data already collected will allow us to refine understandings of the relation between ecological zones and local categories or folk ecotopes. If the dozens of archaeological projects that annually explore diverse areas across the Andes collected local names and definitions instead of imposing Pulgar Vidal’s model, a valuable database could be collectively constructed. This would actually be more in accordance with Pulgar Vidal’s initial proposal, prior to its stagnation. While we are still far from having a detailed map of local names of the ecological zones from which we might establish pan-Andean comparisons, it is never too late to dialogue and take seriously the voices of the people living close to archaeological sites. Decolonization, ontology, and phenomenology must never avoid dialogue with those who know more about their own landscapes. This will also help to advance the discussion on regional difference in the Andes.

In conclusion, Pulgar Vidal’s model of the “Eight Natural Regions of Peru,” identified in the 20^th-century landscape, has been used as a direct analogy for diverse precolonial moments without significant discussion or critique. The application of this model to the precolonial past is limiting in terms of the potential to recognize difference and the model incorporates ILK into landscape analysis in an oversimplified manner. While this model may serve to formulate introductory hypotheses for precolonial land use and landscapes, archaeologists must also improve dialogue with local populations about their understandings and uses of landscape. This case shows that despite increasing interdisciplinarity in Andean studies, there is still much room for improvement in terms of understanding and narrating past landscapes.

Supplemental Material