Predation, Propitiation and Performance: Ethnographic Analogy in the Study of Rock Paintings from the Lower Parguaza River Basin, Bolivar State, Venezuela

Abstract

Rock art sites located in areas inhabited by indigenous peoples offer extraordinary opportunities for interpretation using ethnographic analogy. Nonetheless, we must examine the pertinence of a direct historical approach when dealing with sequences of rock art that may extend back several millennia. Recent decades have witnessed increasingly sophisticated ethnographic analyses that reveal the intimate relations between human and non-human entities and the generative role of myth, music, dance, artifacts, and physical settings in the enactment of creative contexts of lowland South America. This literature has led to a reassessment of the meaning of rock art images, the significance of context, and the place of sites in the landscape. In the pictographs found in several rock shelters on the lower Parguaza River of Venezuela, depictions of a wide variety of human and non-human figures offer insight into the relations between predation, propitiation, food, illness, and the different paths to spiritual knowledge that prevail in the myths and practices of local indigenous populations to this day. In this contribution we explore the promise and limits of ethnographic analogy in the study of sites from this area and offer an analysis of the development of the sites through time, with an eye on both disruption and continuity.

1. Introduction

Rock art sites located in areas inhabited by indigenous peoples in the lowlands of northern South America offer extraordinary opportunities for study using ethnographic analogy. Recent decades have witnessed increasingly sophisticated analyses that reveal the intimate relations between human and non-human entities often depicted in rock art (Descola 1994, 2012; Fausto 2000; Stahl 2014; Turner 2009; Viveiros de Castro 1996, 1998) and the generative role of myth, music, dance, artifacts, and physical settings in the enactment of creative contexts (Guss 1989; Mansutti Rodríguez 2006, 2019; Santos-Granero 2004; Severi 2014). This literature has led to a reassessment of the meaning of images, the significance of context, and the place of rock art sites in the landscape. In the Northwest Amazon and neighboring Orinoco, ethnographers and indigenous scholars have led the way in analyzing the role of petroglyphs and other geographical features in the construction and maintenance of the sacred landscape (González Ñáñez 1980, 2020; Hill 2002; Hugh-Jones 2016; Reichel-Dolmatoff 1971; Santos-Granero 1998, 2009; Vidal 2000, 2002; Wright and Hill 1986; Wright et al. 2017). Archaeologists working in the area have benefitted from these ethnographic studies, as well as from their own personal experiences informed by visits and conversations with community members who share their perspectives on the rock art sites and images found in them (Castaño-Uribe 2019; Greer 1995; Hampson et al. 2024; Iriarte et al. 2022; Lozada Mendieta et al. 2024; Riris and Oliver 2019; Riris et al. 2024; Scaramelli 1992; Scaramelli and Tarble de Scaramelli 2018; Tarble de Scaramelli et al. 2021). Although we have found no evidence for current or recent production of rock art in the Middle Orinoco area, some sites have been modified by indigenous communities through the reengraving of petroglyphs and the painting of new messages.1 These activities underscore the importance that rock art continues to play in the cultural practices of the local communities and further support the utilization of ethnographic analogy in the attempt to comprehend indigenous ontologies and their expression in rock art.

At the same time, other sources of evidence can complement the ethnographic gaze. Archaeology offers chronological depth in a region known to have been occupied for millennia, stretching back to the late Pleistocene and early Holocene (Aceituno et al. 2024; Barse 1990, 2009; Tarble de Scaramelli and Scaramelli 2017). In the Middle Orinoco, evidence for multiple, culturally distinct occupations has been observed in the excavated sequences of numerous sites (Barse 2009; Cruxent and Rouse 1958–1959; Lozada Mendieta et al. 2016; Roosevelt 1980; Vargas 1981; Zucchi 2002; Zucchi et al. 1984), leading to different proposals for population dispersal through time (Antczak et al. 2017; Hornborg and Hill 2011b; Lathrap 1970; Oliver 1989; Rouse 1985; Tarble 1985; Zucchi 2002). At the time of European contact, a wide variety of ethnic groups with different modes of production (farmers, fishermen, hunter/gatherers) occupied the area. Many of these have disappeared or were absorbed by other ethnic groups following the impact of colonization (Mansutti Rodríguez and Bonneuill 1994–1996; Morey and Morey 1975). Linguistic research points to both migration and the creation and maintenance of long-distance ritual and trade networks involving different language-bearing groups that resulted in a dense mosaic of language families in the Orinoco region including Arawakan, Salivan, Cariban, Guajiboan, and other less easily classified languages, such as Pumé (Yaruro), and the now extinct Otomaco and Guamo (Durbin 1985; Heckenberger 2002; Hornborg and Eriksen 2011; Hornborg and Hill 2011a). The rock art of the area is also varied in both technique and context (Cruxent 1946; De Valencia and Sujo-Volsky 1987; Greer 2001; Tarble de Scaramelli and Scaramelli 2010), suggesting the presence of several local traditions of both petroglyphs and rock paintings. In one of the most extensive surveys of rock painting in the Middle Orinoco, Greer (1995) has defined a chronology that extends from pre-ceramic to post-contact times.

Considering this complex occupational history, the analysis of the rock art in different contexts offers the opportunity to discern changes through time, as well as long-term continuities that may affect the relevance of ethnographic analogy (Pearce 2012). Notable shifts in technique, content, and placement of rock art point to the introduction of new practices associated with distinctive horizons of cultural change (Skinner and Challis 2022, p. 9). Continuity is inferred in the reutilization of previously painted panels and the replication of motifs or in the superposition or intentional superimposition of new figures on older ones (Gunn et al. 2022; Motta et al. 2020). Additional information on the utilization of the sites is obtained through the documentation of cultural materials present, such as cupules, polishing grooves, burials, and surface artifacts. When available, stratigraphic excavations in or near sites with rock art reveal further information on the use and chronology of the site.

The analysis of spatial dimensions at different resolutions also contributes to the understanding of change and continuity through time. By situating rock art sites in their context, both geological and as related to other geographic and cultural features, inferences can be made as to the factors involved in the choice of sites and their possible temporal variation in the construction of the landscape. The appearance of new contexts may indicate changes in the function of sites, whereas the maintenance and reutilization of old sites inform the continuity of activity linking different spaces. At a more intimate spatial scale, the size, capacity, orientation, and location of sites are important indicators of preferences through time, as the criteria for rock art site creation may vary with the activities associated with their use. Whitley (2023) has described different types of sites associated with various types of shamanic practice in the far western United States during the same period, related to the social group and the type of ritual being performed.2

In the following sections, we explore the promise and limits of ethnographic analogy in the study of sites from the Parguaza area of the Middle Orinoco and offer an interpretation of the development of the sites through time, with an eye on both disruption and continuity. We will focus on the rock paintings found in the large rock shelters unique to the area, where long-term cultural practice is manifest in the use of these sites for ceremonial and funerary activities and in the multiple styles of paintings therein. We will demonstrate significant shifts in the content of the rock art at different moments in the creation of the complex murals visible today. We propose that following a prolonged but sporadic pre-ceramic presence at the sites, a major change occurs with the introduction of gendered, propitiatory cults, possibly related to the expansion of Maipuran Arawakan expansion throughout the Orinoco Basin. We will focus on the ethnographic evidence that supports this proposal. A later significant shift may reflect a Cariban influx into the area, bringing with it a notable change in the content and technique of the rock paintings. At the same time, we maintain that rock art related to predation, propitiation, and performance is a fundamental aspect of the Northwest Amazonian ethos manifest in the longue-durée (Hill and Santos-Granero 2002), even while recognizing the impact of colonization, demographic loss, and other factors that have resulted in localized responses and innovations (Hill 2002; Hill and Wright 1988; Vidal 2002; Wright and Hill 1986).

2. Rock Art in the Parguaza River Basin

2.1. The Parguaza Basin

The Parguaza River Basin comprises one of the most spectacular settings of the western portion of the Guayana Shield (Figure 1). It is a land of savannah and forests, countless rivers, waterfalls and rapids, deep canyons, and steep hills. The region is dominated by dome-shaped granite hills of Mesoproterozoic origin, technically called inselbergs (Bonilla-Pérez and Frantz 2013; Urbani and Szczerban 1975). These huge outcrops, referred to in the literature as part of the Parguaza Granites, belong to a large batholith protruding through the Amazon Tertiary plains in the form of isolated inselbergs, the largest anorogenic granites worldwide (Figure 2). They are characterized by vertical, or near vertical, walls that present fractures, boulder concentrations, and crevices that, in some cases, have been eroded through the action of wind and rain to form large rock shelters, both at the base of the hills and on shelves eroded out of the face of the gradients. Although a blackish, organic patina covers the surface of the granite, unexposed surfaces, fractures, and cracks reveal the internal colors, ranging from yellowish oranges and pinks to light grey, and it is on these surfaces that rock art is found.

The Parguaza River carries fresh, clear waters from the upper Guayana highlands toward the sediment-laden waters of the Orinoco. The mountains, hills, and plains offer diverse niches characterized by dense forest vegetation, gallery forests, savanna grass, and scrub woodlands. High canopy forests are ubiquitous to the eastern highlands and to the evergreen galleries located along the main streams, creeks and rivers, as well as at the base of the granite boulders and mountains. Resources vary widely between the forested uplands and the savannas. Aquatic and terrestrial fauna are abundant along the rivers and gallery forests. These areas offer a wide variety of resources, including fruits, nuts, woods, resins, honey, palm leaves, and, of course, fish and game. In the savanna, resources are more limited in number and variety but still provide several species of game. In addition, the savanna provides open spaces for human habitation, which today tend to be widely dispersed. The central valley is drained by the Parguaza River and by numerous smaller intermittent streams, all of them plentiful in vegetal and faunal resources. The yearly weather cycle can be roughly divided into a rainy (May–December) and a dryer season (January–April), and the landscape changes noticeably with this cycle, as does its utilization for hunting, fishing, and horticultural practices.

Currently, in addition to the Criollo (Venezuelan nationals) population, the Parguaza Basin is home to disperse Huotthüha or Piaroa (Salivan) communities, but according to local history, the Mapoyo (Cariban) occupied the lower reaches of the river up to the early decades of the 20th century (Perera 1992). Mining activities and deforestation pose increasing threats to the area today (Fundaredes 2023).

2.2. Rock Art Sites in the Parguaza

Rock art in the form of petroglyphs is found in nearly every Venezuelan state where suitable stone support is available (Sujo-Volsky 1975; De Valencia and Sujo-Volsky 1987); nonetheless, the southern states of Bolívar and Amazonas are the home to most of the sites with rock paintings in Venezuela. Dozens of rock art sites have been found in the area, some of which provided shelter for human groups, dating to the initial peopling of the region over 10,000 years ago (Tarble de Scaramelli and Scaramelli 2017). In previous work, sites have been classified based on location, relation to habitation sites, accessibility, visibility of the painted motifs, superposition of styles, and associated features, such as burials, grinding surfaces, and the presence of archaeological remains, including ceramic, lithic, and more recently manufactured goods (Greer 1995; Perera 1983, 1992; Perera and Moreno 1984; Scaramelli 1992; Scaramelli and Tarble 1996; Tarble and Scaramelli 1999; Tarble de Scaramelli and Scaramelli 2010, 2017). Two types of sites are present in the Parguaza area. (1) Large rock shelters are made up of horizontal apertures in the sides of local granitic hills, where ceilings serve to protect open spaces extending between 35 and 150 m along the drip line. Grinding facets and cupules are also common on the bedrock of the shelters (Figure 3). The back walls and ceilings exhibit panels of rock paintings in different styles (Figure 4). Many of these sites were also used for funerary activities in the past and are still the repository for burial bundles and wooden coffins today (Brites 1994; Galarraga et al. 2003). Recent utilization of the sites can also be inferred from the remains of cans, bottles, and ceramic wares dated to the late 19th and early 20th centuries. (2) Small caves are found with limited numbers of paintings on the ceilings. These smaller sites often serve as cemeteries for the Huotthüha who currently inhabit the region.

Based on his research in the southwestern area of Venezuela, Greer defined seven styles/periods based on features, including the technique of application, care of application, color shade and thickness of the paint, line width, figure size, and content (Greer 1995). He used superposition to support his relative chronology for the paintings and integrated archaeological data to provide a provisional scheme for the time periods and geographical branches. The Parguaza region pertains to the northernmost Cedeño branch (Greer 2001, pp. 691–92). The proposed sequence of periods ranges from early pre-ceramic settlement through various ceramic occupations and finalizes in the early colonial period (Greer 2001, p. 689).

Our excavations in the Gavilán 2 (BO-126) rock shelter on the Lower Parguaza exposed a pre-ceramic occupation characterized by a simple, unifacial flaked, crystalline quartz industry associated with well-preserved botanical and faunal remains and a large panel of rock art. AMS ¹⁴C dates obtained on charred materials ranging between 10,580 to 10,250 cal. BP (Beta-252625) and 3830 to 3600 cal. BP (Beta-252621) support the presence of an early to middle archaic occupation with a wide spectrum economy that included hunting, fishing, and plant collection like those found in other Venezuelan and Colombian sites (Aceituno Bocanegra 2010; Barse 1990). Fragments of red ochre with striations found in the excavations at levels dated to 3830–3600 BP to 7730–7600 cal. BP may have been used to prepare the pigments for some of the rock paintings found at the site, but no association can be made directly to the paintings. Surface collections revealed ceramic sherds related to Saladoid, Barrancoid, Cedeñoid, Arauquinoid, Valloid, and Nericagua materials, spanning the entire ceramic sequence for the region. Unfortunately, despite decades of investigations, the archaeological chronology for the ceramic sequence of the Middle Orinoco is still disputed, with proposals that range between 6000 BP to 2000 BP for the introduction of the earliest ceramic occupations (see discussions in Antczak et al. 2017; Gassón 2002; Greer 1995). Correlations between ceramic style and rock art images have been suggested, but no direct dating is available for the rock art (Greer 1995; Lozada Mendieta et al. 2022; Páez 2021; Riris et al. 2024; Rivas 1993; Tarble 1991; Tarble and Scaramelli 1995).

Several features serve to distinguish the rock art sites of the Parguaza Basin from other sites in Venezuela. Sites chosen for rock art in early periods tend to be reutilized through time, often for millennia. This is particularly true for the large rock shelters that offer evidence for long-term use. Unique to the area is the variety of painted styles, including monochrome, bi-chrome, and polychrome motifs, often superimposed in complex panels (Greer 1995; Scaramelli 1992). Painted areas within sites appear to determine the areas of repainting, with multiple styles superimposed in such a way as to create increasingly complex murals (Figure 5). This practice appears to have been conceived as an enhancement or reaffirmation of the imagery rather than an iconoclastic practice. Also notable is the prominence of zoomorphic motifs, including realistic and stylized renditions of humans, fish, deer, rodents, birds, lizards, turtles, snakes, and other recognizable elements, some of which show interaction in the form of scenes (Figure 6). Finally, although motifs and styles can be seen to change through time, there is also a tendency to duplicate, imitate, and restate earlier images. Therefore, despite disruption, there is also continuity in the selection of content. This would seem to indicate a shared worldview of great time depth. These characteristics confirm the idea that once chosen as a site to be painted, the same sites were prone to visitation and symbolic revalidation through time, thus serving as anchors to the landscape and the maintenance of the cultural code determining use and meaning (Scaramelli and Tarble de Scaramelli 2018).

3. Methods: Defining the Stylistic Sequence of Rock Paintings in the Parguaza Region

The sequence outlined in

Table 1. Motifs and diagnostic characteristics of the periods/styles defined for the sites referred to in this paper (after Greer (1995), with some modifications).

Style/Period	Figure/Motif	Sites	Diagnostic Characteristics
2 Pre-ceramic	Lizard, bird, armadillo	BO-27-B BO-29 BO-35	Fine lines; thin, purplish paint
	Fish		Vertical interior lines; thin, purplish paint; light red paint, vertical placement, often aligned in series
	Anthropomorphs, zoomorphs		Broad lines; light red paint
3 Early Ceramic	Anthropomorphs, “ghost figures”, quadrupeds (especially deer), possible flute player	BO-26-B BO-27-B BO-29 BO-35 BO-126	Large figures (some life-sized or larger) Solid white or yellow paint outlined with red paint Some action between figures
	Fish (various identifiable species), stingray, turtle, feline, bottlenose dolphin, armadillo, tapir, capybara, snake, amphibians (realistic) 1		Solid white paint, outlined with red paint, some interior lines Vertical, lateral, or dorsal
	String of “gourds”
	Round, flat “basket” designs; large, rectangular “textile/mat” designs		Red/white bichrome; interior geometric designs
	Unidentified figures		Yellow paint
4 Early Ceramic	Small, monochrome zoomorphic figures: fish, deer, tapir, birds, lizard, turtle, feline, etc.	BO-27-B BO-29 BO-35 BO-126	Red paint predominates with variations between light red, dark red, and orange Figures tend to be small to medium size
	Anthropomorphic figures with
	headdresses and rectangular or triangular body coverings or long masks Theriomorphs combining human and non-human features: snake/bird/human; bipedal deer
	Small “scenes” of human/non-human figures interacting, hunting with throwing stick or spear thrower/atl atl
	Falling anthropomorphic figures
	Anthropomorphic figures of various linear types
	Zoomorphic and anthropomorphic figures placed on horizontal lines
	Cultural objects: baskets, maracas, nets or fishing traps, woven mats
	Phytomorphic figures (rare)
	Outlined crosses, sometimes multiple
	Concentric circles and rhombuses Segmented box or square
	Parallel dotted lines “Netting”, open, connected lines
5 Late Ceramic	Paired zoomorphic figures Circle with internal lines (turtle shell?) Winged circles, circle grids, and chains, “textile” or “basket” designs   Crosses and outlined crosses	BO-26-B BO-29 BO-35	Dark red paint with thick lines Black paint Bichrome red and white paint, less symmetrical Bichrome red and black paint Polychrome red, black, and white paint
6 Late Ceramic	Anthropomorphs with rectangular, solid bodies Paired figures in contrasting colors	BO-26-B BO-29 BO-35 BO-126	Thick, white, pink, and yellow paint predominates Sloppy execution
	Rabbit, lizard, fish “Maracas” Concentric circles with interior divisions Meandering lines Smears
Period 7 Colonial To present	“COMUNIDAD” “KH” “Kiloin”	BO-126	Text in cream-colored paint Very recent; painted after our first visit

¹ See identification in Tarble de Scaramelli et al. (2021).

is based on data retrieved from five rock shelters located in the Lower Parguaza Basin (BO26-B, BO-27-B, BO-29, BO-35, BO-126)3 (Figure 7). This is part of an ongoing analysis of the motifs associated with different styles at each site. In the three sites located in indigenous territory, we obtained permission and were accompanied to the rock shelters by several members of the local communities. At the other sites, we also obtained permission from the local Criollo community who facilitated guides to the sites. Each site was mapped using a compass and tape measure, and associated cultural remains and grinding surfaces were logged; in some cases, diagnostic ceramic sherds and lithic artifacts were collected for analysis. During fieldwork, drawings and some tracings of the superpositions were made to enable later analysis of the photos. As noted above, limited excavations were carried out at BO-126, a site that is not currently located on indigenous land. Burials and funerary deposits were left undisturbed. The sites were photographed on successive short visits since 1992, using film and digital cameras. The late Miguel Ángel Perera donated photos from one of the sites (BO-29), taken during his visit in 1987. We have used the D-Stretch application developed by Jon Harman (2009) and Adobe Photoshop enhancement to analyze the photos.

A database in Excel is currently under construction to record details of the motifs for each site. Although the photographic coverage is incomplete, we have endeavored to follow the layout of the sites, creating an ordered sequence from one end of the site to the other. Figures were divided into two broad categories: anthropomorphic/zoomorphic, identified by numerals, and abstract/geometric/artifactual, labeled alphabetically. Further details of the type of paint, color, possible identification, and tentative period placement are also noted in the database.4

Both in the field and, later, in the study of photographs, special attention was given to the occurrence of superpositioning. BO-35 was the key site to determine the sequence throughout the Parguaza, as it offered the clearest evidence for the superpositioning of the largest number of styles/periods in one site. We compared our data with those supplied in Greer’s analysis (1995). Based on the sequence developed for this site, it was possible to observe similar sequences of superpositioning at the other sites, and modify and add to Greer’s observations, although none of the other sites offered as complete a sequence as that found at BO-35. The sites vary considerably in the relative frequency of the different styles/periods and in the presence or absence of certain types of motifs. For example, BO-35 contains the greatest concentration of style/period 3 fish and other aquatic zoomorphs and the only appearance of polychrome style/period 5 figures. BO-29 is distinguished by a higher number of masked figures and style/period 4 zoomorphs and interactive scenes. BO-26-B, the smallest cave in the sample, has a distinctive style/period 3 painting technique, while BO-126 is notable for the predominance of style/period 6 figures.

As seen in Table 1, there are considerable differences in the content and technique of the painted styles. Greer (1995) has argued that it may be useful to contemplate parallel developments for the styles and suggests that a continuous monochrome tradition of styles 1-2-4 may be interrupted by, but parallel to, the intrusion of style 3 and its further development into style 5. The increasing variety in stylistic representation during the later periods (5–6) probably reflects the ethnic diversity mentioned earlier for the area by the time of European colonization. We will discuss these proposals in greater detail in the following paragraphs.

Style/period 2, the earliest clearly identifiable style found at the sites included in this analysis, underlies all other styles in superimposed contexts (Figure 8). Fine-lined purplish paint is used to create predominantly zoomorphic figures such as interior-lined reptiles and large birds with open wings, although some anthropomorphic figures painted with faint, broad strokes are also present. Very few geometric motifs are associated with the style (Figure 9). The rather rare figures pertaining to this style at different sites do not overlap one another on the walls. It is, nonetheless, common to find period 3 paintings superimposed on period 2 motifs. Outside the Parguaza area, Greer (1995) has reported other sites with period 2 motifs to the south, and Lozada Mendieta et al. (2024) report interior-lined fish for sites on the western side of the Orinoco that they associate with period 2. We have also documented interior-lined fish motifs at sites near Puerto Ayacucho that support a distribution of the style along the Middle Orinoco. Greer has suggested that this style may be associated with pre-ceramic occupations in the area and shows continuity from his period 1.

Style/period 3 appears dramatically in the study area. Large, ghost-like anthropomorphs with singular headpieces, smaller, bow-legged figures, and square-bodied figures that seem to be wearing woven costumes are characteristic of the style, along with multitudes of carefully detailed fish, amphibians, turtles, and other terrestrial and aquatic species. Life-sized deer portrayed in active postures are present at several of the sites, portrayed in such a similar manner that it would be possible to infer that they were painted by the same person. At one of the sites (BO-27-B), three deer appear to be running as if being chased (Figure 10).

Other less common figures include iguana, feline, bottlenose dolphin, armadillo, tapir, capybara, and snake (Figure 8 and Figure 11). Phytomorphs are very rare. Fish are more common at the sites closer to the Orinoco (BO-35), whereas anthropomorphs and mammals are more frequent inland (BO-26-B, BO-27-B, and BO-29). The use of white or yellow figures outlined with red contrasts sharply with the underlying monochrome red figures, suggesting an intrusive style that is almost exclusively limited in geographic extent to the sites in the Parguaza River Basin. Period 3 figures are painted over period 2 figures in several sites analyzed for this paper, even when ample empty space was available for painting (Figure 8). This suggests an intentional superpositioning and continued use of previously painted sites. Greer (1995) has suggested that this style may be linked to the appearance of the Saladoid occupation in the Middle Orinoco due to the emphasis on bichrome white-on-red painting. Alternatively, it may be a localized expression, contemporaneous with the widespread expansion of style/period 4, which we will discuss in the next paragraph. Notably, geometric/abstract figures are infrequent in this style, indicating that human and non-human encounters continue to dominate the subject matter of the rock art.

Style/period 4 literally dances onto the scene with a proliferation of monochrome red, masked figures with decorated squarish or triangular bodies, with and without headdresses and other paraphernalia, such as rattles or leg bands (Greer and Greer 2006) (Figure 12).

Many different styles of stick figures representing humans in profile and frontal views are present. Some of these show male genitals, but most are not easily identified. Some figures are lined up as if dancing; others have raised arms in an attitude of supplication. Other anthropomorphic figures appear to be falling (Figure 13).

Numerous zoomorphic figures are outlined, solid, or with spotted/reticular-filled interiors, and appear both individually and lined up horizontally on a line. Deer, fish, birds, reptiles, frogs, felines, canines, and others are portrayed in pairs, rows, and individually (Figure 14). Several geometric figures appear in this style period: outlined crosses (single, double, and multiple), concentric circles (some segmented and others rayed), segmented boxes, and lines of dots and dashes (Figure 15).

Although most figures show little interaction, several “scenes” are portrayed in this period. Especially notable are scenes in BO-29 portraying humans hunting deer or other quadrupeds, using what appear to be spear throwers (atlatls) or throwing sticks (Figure 16a,c,h). No bows and arrows, blowguns, or firearms are pictured. Human figures with long poles or ropes are depicted with lines of quadrupeds, often shown on horizontal lines, as if herding or lassoing.5 Other scenes show interactions between birds and fish and possible fish traps associated with schools of fish. At BO-29, the superimposition of some style 4 hunting scenes, zoomorphs, and geometric figures within the body of large style 3 figures points to the intentional framing of these scenes. In other sites, such as BO-35, style 3 zoomorphs are frequently overlaid by style 4 geometrics, such as multiple outlined crosses (Figure 8).

In several of the rock shelters (BO-27, BO-29, BO-35), there are sectors defined by crevices, or openings in the base of the walls, from which multitudes of figures appear to emerge, often superimposed on one another, blurring the individual motifs (Figure 17).

Style 4 is much more widespread than any of the earlier styles, with motifs reminiscent of figures found as far west as Chiribiquete and La Lindosa in Colombia (Castaño-Uribe 2019; Muñoz Castiblanco 2020)6 and to the east, at sites in the vicinity of Caicara del Orinoco (Tarble de Scaramelli and Scaramelli 2010, 2021) and Puerto Ordaz (Sanoja and Vargas 1970). To the north, rock paintings found in rock shelters in Aruba show stylistic similarity to styles 3 and 4, with the presence of outlined crosses, concentric circles, circle and bar, as well as zoomorphic and anthropomorphic motifs (Kelly 2009, p. 184). Kelly attributes the similarity to influence from the Middle Orinoco. There is considerable diversity in style 4, both intersite and intrasite, that Greer suggests may indicate “somewhat autonomous communities” within a larger region where interaction seems to have been frequent (Greer 1995, p. 98).

Style/period 5 is marked by a proliferation of abstract monochrome and polychrome designs, with an emphasis on figures such as connected circles, dots and lines, concentric arcs, and chains of concentric circles with “wings”. Once again, many style 5 figures are placed on top of earlier paintings, creating dense concentrations in complex panels (Figure 18). This is the only style with polychrome figures executed in red, black, and white.

Style/period 6 marks a pronounced shift in technique, color combinations, type of paint, and, to an extent, content of the rock art. While earlier styles were generally well-executed, with an eye to detail and careful execution, style 6 motifs are somewhat sloppy and poorly executed. The clay-based paint is often runny or thick, with a dull, chalky appearance, found in white, pink, yellow, and dull reds. Caraña (a type of dark resin) is also present (Greer 1995). Quadrupeds, anthropomorphs, and “maracas” are found, but zig-zag lines, meanders, hatching, parallel lines, etc., are more frequent. The best representation of this style is found in BO-126 (Tarble de Scaramelli and Scaramelli 2017), where the main panel is dominated by style 6 paintings. Paired representations are characteristic of the style, and these are often painted in 2 different colors, such as dark yellow and white (Figure 19).

4. Placing the Rock Art in a Cultural Context

In the presence of rock art, one is often overwhelmed by sensations that go far beyond “seeing”. The access to many sites in the Parguaza basin involves a hot climb up a steep, granitic surface or a hike through dense, tangled vegetation. The immediate sensation on arrival is the cool relief of the shelter and then the musky smell of decaying plant matter, termite residue, and, at times, burial remains. Wasps and bees may be excited by a foreign presence, setting off a menacing buzz. As one’s eyes adjust to the muted light, figures start to come into focus, and on what was once a wall of undefined shapes and colors, certain images begin to emerge, sometimes singly, sometimes in a mass of superimposed figures. It is easy to comprehend that these are special places, filled with elusive meanings and vestiges of past performances. On several occasions, our guides and companions felt inspired to narrate ancient stories as we set about our recording. They did not claim to have made the paintings; that was the work of deities who “thought” them into place or “cried” them into being in ancient times. Currently, the sites are considered to hold special significance and demand respect. Our attempts to understand these sites in the context of the peoples who created them and used them must take into consideration the range of actions and senses that were involved in these practices. Although we are limited in our means to access the meanings that prevailed in different times and under varying circumstances, we can draw on ethnographic and historical sources to approach the challenge these sites present.

Several aspects of the rock art described above suggest that the sites were used for ceremonial purposes in different ways through time. A pervading theme throughout is the predominance of anthropomorphic and zoomorphic figures, although the emphasis on these diminishes in the latter part of the sequence. The sites are all located some distance from habitation sites, except for BO-126, a site found at ground level, where a lengthy pre-ceramic occupation was found, apparently predating most of the paintings. Each site included in the study shows a prolonged sequence, supporting the idea that the caves with rock paintings continued to be utilized by groups that may or may not have been affiliated culturally with the original painters and their communities. This implies that these sites came to acquire a special status in the landscape in accordance with widespread belief systems that grant agency to rocks, caves, waterfalls, lagoons, and other geographic features (Halbmayer 2012; Overing and Kaplan 1988; Santos-Granero 2009; Wright et al. 2017). It could be argued, then, that the painted walls of the caves became their “social skin”, much as body painting “dressed” the person (Turner 1995).7

The use of the caves for funerary rites appears to have increased as time passed, although earlier burials may have been removed or decayed through time. Both Huotthüha and the Mapoyo are known to use caves as repositories for burials at the present, although both deny painting the sites, nor choosing them because of the presence of paintings.8 The association of burials with rock shelters points to the role of caves as conduits between the underworld and the world of daily activity. In light of these features, we propose that the painted shelters were the setting for diverse ritual activities related to worldviews that incorporated shamanic intervention in the propitiation of well-being, the negotiation of predation, and the protection against harm.

4.1. Ethnographic Analogy and the Search for Meaning

Numerous attempts to find meaning in the rock art of the area have referred to perspectivist and animist belief systems found throughout the Northwest Amazon and the need to go beyond the cultural/natural dichotomy (Castaño-Uribe and Hammen 1998; Castaño-Uribe 2019; González Ñáñez 1980; Iriarte et al. 2022; Riris et al. 2024; Wright et al. 2017). In their analysis of the rock art found in la Serranía La Lindosa, in Colombia, Hampson et al. include a detailed discussion of the ethnographic references to the fluidity between the natural and supernatural, the “complex and many-layered social and ideological relationships between humans and animals beyond utilitarian exploitation” (Hampson et al. 2024, p. 11), and the role of shamans in the negotiation with the Master of Animals to assure the release of game prior to predatory activity. In discussing South American lowland indigenous cosmologies, Halbmayer stresses that they “do not create encompassing totalities or an integrated universe, but […] a multiverse of co-existing and multiply connected worlds relying on a specific form of non-totalizing partial encompassment” (Halbmayer 2012, p. 120). The multiverse may include persons, spirits, houses, tools, musical instruments, animals, plants, and villages (of humans or non-humans). Brabec de Mori and Seeger (2013, p. 277) further this discussion, maintaining that the parts are related to each other and, “through the distribution of agency in various aspects of the multiverse, these relations as well as the entities themselves are never stable but have to be constantly negotiated, maintained or transformed” (Brabec de Mori and Seeger 2013, p. 277). Recent work has built on the ontological turn to explore different approaches to the comprehension of the relation between humans and non-humans. Hill (1993, 2013) criticizes the perspectivist approach for overly focusing on the visual elements, arguing that recent ethnomusicological analysis is revealing the fundamental role of the sonic to communicate and interrelate between spirits and humans (see also Halbmayer 2012; Severi 2014). Others seek to overcome the overemphasis on predation and aggression, while neglecting aspects of conviviality and equilibrium (Overing and Passes 2000; Wright 2011). In line with this position, Hill (2002) seeks in his work to show how collective ritual performance serves to “musicalize the other” to maintain a balance between life-giving and life-taking forces (Hill 2002).9

As we look for ways to comprehend the depictions of humans and non-humans in rock art, it becomes clear that they cannot be taken simply at face value, nor can we interpret zoomorphic and phytomorphic elements as mere illustrations of the natural habitat or potential foodstuffs.10 Likewise, geometric or abstract patterns, such as those found in rock art, houses, baskets, textiles, gourds, pottery, body paint and stamps, tattoos, etc., are best seen as parts of a wider system of memory that includes music, speech, chants, and songs (Guss 1989; Hugh-Jones 2016; Reichel-Dolmatoff 1987; Santos-Granero 1998; Severi 2014). By this, we do not mean that rock art figures can be interpreted as illustrations of myths, but rather that myths, chants, and iconographies can be understood as “‘variations’ of the same ‘conceptual imagination’” (Severi 2014, pp. 45–46). For example, among the Yekwana, a Carib-speaking group of the Upper Orinoco, baskets decorated with designs “incorporate a complex system of symbols that act as an index and key to the rest of the culture” (Severi 2014, pp. 48–49). The baskets are actual embodiments of mythical beings, have agency, and their designs are linked to the chants that accompany their weaving (Guss 1989, p. 36 in Severi 2014, p. 49). We argue that a similar relation is likely to underly the designs found in rock art, where the design itself embodies multiple layers of meaning, and the context of the graphisms and sites can provide important clues regarding different types of ritual performance and their variation through time.

4.2. Continuity and Disruption in the Parguaza Rock Art Sequence

As discussed above, there is clear evidence for some continuity in the conception of rock art that permeates the entire sequence of the Parguaza region. At the same time, certain shifts in emphasis suggest that there were major disjunctions that may have been associated with the influx of populations bringing with them innovations in language, productive strategies, and cosmology that are manifest in the rock art of the area. These can be summarized as follows:

• Pre-ceramic occupations. These may have lasted several millennia, characterized by sporadic painting at rock shelters in the Parguaza area and beyond, suggesting a widespread sharing of symbolic representation. Zoomorphic elements include the detailed rendering of reptiles, large, long-legged birds with outspread wings, interior-lined fish, and depictions of small quadrupeds. A few representations of highly stylized anthropomorphs are also present. No interaction is seen between the figures, and no superpositioning is noted. It is difficult to assess the types of activity associated with these manifestations, but apparently, there was no emphasis on large, organized ritual celebrations, just as you would expect from small-scale populations.
• Early ceramic occupations. The sudden advent of new painting technology, content, and frequency associated with style/periods 3 and 4 signals a rupture with previous traditions. The emphasis on carefully executed, costumed anthropomorphs and scenes of interacting figures portraying an array of activities, including flute-playing, dancing, hunting, fishing, and trapping, signals the appearance in the region of more sedentary societies, with a novel cosmovision that may have coincided with the arrival of ceramic-bearing Arawakan peoples and the hierarchical, ritual complex we know today as the Yuruparí or Kuwai, widely distributed throughout the Northwest Amazon and beyond. This shift has its counterpart in many of the petroglyphs found along the Orinoco and tributaries, with similar motifs, but with far greater visibility that suggests a different function for the sites. We will discuss this proposal further below.
• Late ceramic occupations. The final shift in the rock art of the Parguaza entails a notable change in the conception and execution of the paintings, where the planning and carrying out of carefully rendered anthropomorphic and zoomorphic figures and scenes are replaced by an increasing emphasis on hastily painted geometric and abstract designs, and the appearance of paired figures painted in contrasting colors. Ghost figures, hunting/trapping scenes, and massive displays of zoomorphs are no longer present. The emphasis on duality indicates a shift in cosmology that may coincide with the influx of Cariban-speaking communities into the area.11 It may also be associated with the appearance of ceramics of the Valloid tradition, in which little attention is paid to decorative technique, in contrast to earlier Saladoid and Arauquinoid traditions. It is not clear whether the more frequent use of the sites as burial repositories dates to this shift or whether it is a more recent reutilization of the caves, both with and without rock art, by present-day indigenous communities, such as the Huotthüha and Mapoyo.

5. The Yuruparí Ritual Complex and Its Expression in the Rock Art of the Parguaza

5.1. The Yuruparí Ritual Complex

The Yuruparí ritual complex is found among multiple ethnic groups of the Río Negro, Vaupés, and Middle Orinoco (Hugh-Jones 2016; Mansutti Rodríguez 2006; Wright 2011). The Yuruparí, and others like it, such as the Warime ceremony found among the Huotthüha and the Kuwai, found among the Arawak-speaking groups, can be understood as a ceremonial complex involving aerophone instruments and, in many cases, masks, that act as agents for the reproduction of human and non-human beings.12 The chants, songs, myths, puberty rites, and festivals associated with the complex comprise a system of beliefs and practical codes, imparted to humans by the primordial being, master of the visible and invisible, and taught to children at their initiation rites (Vidal 2000, p. 641). Intertribal Yuruparí ceremonies often last for extended times and usually involve the exchange of food and drink between kin and affines, dances, singing, and initiation rites. More intimate intratribal rituals tend to be focused on healing. In some cases, the Yuruparí ceremonies also involve ritual whippings or other corporal penances. Although there are important differences between the rituals celebrated between different groups, often responding to particular demographic, social, and political circumstances, there are basic elements that they share: the aerophones are played exclusively by men and utilized in initiation rites; in mythic times the aerophones were stolen by the women and then stolen back by the men; spaces are created for the cult, segregated by gender, such as a men’s house and related secret societies; among the Arawak and Tukano, the cult has a hierarchical function, with particular instruments and songs belonging to patrisibs who form part of exogamous societies (González Ñáñez 1986; Hill 1993, 2013; Hugh-Jones 2016; Santos-Granero 2004; Vidal 2000, 2002; Wright 2011; Hill and Wright 1988). Women and uninitiated children are prohibited from seeing the aerophones and bearers of the masks but actively participate in the ceremonies from within special buildings or from the forest beyond the village.

The flutes and trumpets are said to be the body of the first ancestral being from which all peoples descended. Based on a three-part mythical cycle, oral tradition deals with the primordial origins, which for the Arawak is found in Hípana, the “universal navel” that connects Earth with the sky (Wright et al. 2017). This is an inchoate, small “universe” filled with cannibalistic beings feeding on one another. It was a time when there were no distinctions between dead and living, male and female, and human and nonhuman. A second cycle revolves around the birth of Kuwai/Yuruparí, “whose body consists of all worldly elements and whose humming and singing…causes the world to expand and brings into being all living species and natural elements” (Hill and Wright 1988, p. 86). It is during this cycle that Kuwai taught the first ancestors agriculture and the sacred rituals of initiation (Vidal 2000, p. 642). Kuwai is killed in a fire, and the world shrinks again, but out of his ashes grow the plant materials used to make the aerophones played in initiation rituals and other sacred ceremonies. The women steal the sacred instruments and lead a chase to the four ends of the earth, expanding the world once again, making the world a fertile place, and leaving evidence of their stopping places in the petroglyphs, rock formations, and other geographic features on the landscape (Hugh-Jones 2016, p. 159, Vidal 2000; Wright et al. 2017, p. 189)13. Following the repossession of the sacred instruments by the men, the third cycle of the narrative refers to the creation of humans and their division into named ethnic units linked by marriage and exchange, as well as many accounts of migrations and warfare (Hugh-Jones 2016, p. 159). The third cycle also narrates the relationship between humans and powerful spirits and how this relates to concepts of illness, death, cures, and prevention in the hands of shamans, some of whom can cross boundaries between the living and dead through the ritual language of chants and the musical language of Kuwai (Hill and Wright 1988, p. 88). Hill explains the central role of wind instruments in the ritual and ceremonial practices of Amazonian communities, asserting that

… it is necessary to explore the importance of indigenous understandings of breath and breathing as expressions of life-force and the associated meanings of wind instruments as cultural tools for channelling this life-force into activities designed to collectivise shamanic abilities to return to life from death and to ensure the continued fertility of animal nature as well as the regeneration of human social worlds. Collective performances of wind instruments are closely related to shamanic powers of killing, dying and returning to life, which invoke natural processes that are simultaneously concerned with the reproductive behaviours of ‘animal’ species—fish, birds, game animals and various spirits—and images of predatory violence (Hill 2013, p. 324)

Throughout the Northwest Amazon, an essential aspect of the relationship between humans, non-humans, and the role of shamans in the propitiation of well-being has to do with the notions of disease, hunting, the ingestion of food, and chanting or blowing. Among the Huotthüha, for example, there is a category of terrestrial spirits and gods that send disease to the animals. These are known as the guardians or “grandparents” of the disease. The most important of these are the grandfather/owner of the animals of the land and air and the grandfather/owner of the water and aquatic animals (Mansutti Rodríguez 2006, p. 25). When a hunter kills an animal, these gods send the spirit of the dead animal into the body of the hunter to cause him to become ill. By means of singing and blowing, the shaman sends other gods to combat the illness in the body of the patient (Overing and Kaplan 1988, p. 396). Meat can only be consumed following strict protocols in which the shaman, through his singing and blowing, transforms the meat into vegetal food, free of the danger of contamination and resulting illness (Overing and Kaplan 1988, p. 382). Humans are conceived as related to animals through affinal kinship, a relationship that is cannibalistic by nature. When a person eats an animal, it will cannibalize him in turn, causing illness. Both animals and affinal kin are considered to be transmitters of disease and, therefore, potential cannibals (Overing and Kaplan 1988, p. 402). In the mythology of the Arawakan Baniwa, the figure of Kuwai, “who is all predatory animal spirits in one… gave rise to all predatory spirits of nature who give sickness to humans, called Iupinai”. (Wright 2011, p. 14) It is seen, then, that aerophones play an ambivalent role in the ritual context, both as the source of sickness, in the form of predatory animal spirits, and as the cure for the illness through their music. The essence of the Yuruparí ritual complex and its associated shamanic interventions is its ability to domesticate the threat of predation and transform it into harmonious conviviality (Wright 2011; Hill 2013). Male initiation rites serve to instruct the initiates in the sacred knowledge of the chants and instruments and to prepare them as successful hunters and members of society. Among the Huotthüha, for example, knowledge is visualized metaphorically as the internal “pathway of beads” that a person continues to expand throughout his lifetime by participating in further ceremonies. The designs used in their face paintings and as designs on their baskets are an external expression of the beads they have accumulated in their bodies. The beads contain the words of the sacred songs, and the designs are conceived as their pathways (Overing and Kaplan 1988, p. 339).

A final aspect of the Yuruparí Complex that is relevant to our discussion of rock art is the interplay between vertical and horizontal axes of ritual chants. Chants can be structured in a sequence of named sacred sites, many of which are located along rivers, following a lineal course (Hugh-Jones 2016, pp. 159–61). Hill refers to this aspect of chants as the “chasing after names”. The vertical axis of a chant involves the “heaping up of names” (Hill 1993, 1996, 2011), or lists of names, species, ritual objects, etc., that convey the multiple meanings imbued in the songs and dances that evoke episodes of the myth corresponding to that specific site. The vertical and horizontal axes, at the same time, correspond to two aspects of the cosmos that, according to Wright (1998), are “antithetical yet complementary and correspond in reality to two important principles of social organization: hierarchy, which defines social relations among sibs of the same phratries; and egalitarianism, which defines relations among peoples of different phratries or language groups with whom marital exchanges may be made”. Both axes are essential to an understanding of the role of the ritual complex in social reproduction.14

5.2. Rock Art Associations with the Yuruparí Complex

When analyzing the paintings associated with styles/periods 3 and 4 in the Parguaza sequence, several aspects lead us to propose a relation to a ritual complex similar to the Yuruparí/Kuwai tradition.

• These styles coincide with the proliferation of motifs associated with masked figures and musical instruments, including flutes and maracas, and lines of dancers that signal more organized ritual practices compared to those portrayed in earlier periods. (see Figure 12)
• The variety of designs on the body of the masked figures and the assortment of headdresses may point to a hierarchy of participants, each with specific roles in the ceremonies, such as those illustrated in the ethnographic literature (Goulard and Karadimas 2011; Koch-Grünberg [1925] 2008, 2010; Mansutti Rodríguez 2006, 2019) (Figure 20).
• The hunting and fishing scenes and the numerous depictions of anthropomorphic, zoomorphic, and theriomorphic figures speak to the preoccupation with the complex relations between humans and non-humans and the capacity for transformation and transmutation in a multiverse populated by visible and invisible agents (Severi 2014) (see Figure 16).
• An almost exclusive representation of male or non-gendered anthropomorphs points to the exclusivity of the ritual practices carried out at the sites on the Parguaza, possibly associated with male initiations. As we saw in the previous section, the sight of aerophones (and the identity of masked dancers) was prohibited for women and non-initiated members of the community. The depiction of falling figures may refer to the ritual suicides enacted because of the violation of the proscriptions (see Figure 13d).
• The seeming emergence of multiple figures out of crevices or openings in the rock suggests the ‘calling out’ of beings by the owners of species from their otherworldly spaces. The placement of paintings at several of the Parguaza sites takes advantage of holes, cracks, or other features of the rock to show interaction with the supporting medium. There are frequent references in the ethnographic literature to caves and mountains as the “maloca” or home of spirit beings and gods (Overing and Kaplan 1988, pp. 393–94; Mansutti Rodríguez 2012; Scaramelli 1992, pp. 47–49) (see Figure 17). This is a common feature of rock art associated with shamanism worldwide (see, for example, Rozwadowski 2017; Stoffle et al. 2024).
• The careful superpositioning of figures within the bodies of existing paintings and the displays of massive numbers of certain species, such as fish, suggests the “piling on of names” or vertical axis of chanting referred to by Hill and Hugh-Jones above. The multiple repetition of figures and the layering of motifs could emulate the chanting of a series of names, spirits, species, etc., in the context of rituals carried out in the rock shelters (Figure 21).
• Geometric motifs, such as concentric circles or rhombuses, outlined crosses, and other designs like those found on baskets and body paint may be translational depictions of mythical elements, fragments of chants and songs, that accompany the rituals. Although we are limited to the visual aspects of the past ritual practices in the rock art sites, this interpretation of these motifs reminds us that music and chanting played a fundamental role in ceremonies.15
• Many motifs of the style 4 paintings have a broad distribution throughout the Northwest Amazon and Orinoco, suggesting an interaction area associated with the initial Arawakan expansion that, at the same time, may have had local expressions, such as that found in style 3 of the Parguaza. Common to all these areas are the abundant and diverse zoomorphic, anthropomorphic, and theriomorphic depictions, hunting and fishing scenes, masked figures, groups of dancers and figures in attitudes of supplication, geometric and abstract designs like those used in body paint, basketry, and textiles. Red monochrome figures predominate this style at sites throughout the region (Greer 1995; Iriarte et al. 2022; Hampson et al. 2024; Lozada Mendieta et al. 2024). Vidal (2000, Figures 2 and 3) has reconstructed the sacred routes associated with the Kuwai oral traditions that include a vast area covering the NW Amazon and Guayana (see also Hill 2011, Figure 13.1 and 13.2). This area coincides with the predominantly fluvial route of the northern part of the Arawakan exchange system around AD 900 illustrated by Hornborg and Eriksen (2011, Map 6.1) (Figure 22).
• The similarity of certain motifs at different sites in the Parguaza region suggests the visitation of sites by some of the same artists. At the same time, each site has specific content that points to a specialized function or local tradition. This recalls the axes of the Yuruparí ritual, in which the horizontal axis entails physical or shamanic visions of visitation to the different sites named in the chants (Hill 2011; Hugh-Jones 2016; Wright et al. 2017; Vidal 2000, p. 644). This is most clearly expressed in the numerous petroglyph sites that are found along the course of the Orinoco (González Ñáñez 2020; Wright et al. 2017). Recent research on giant petroglyphs in the Middle and Upper Orinoco (Riris et al. 2024) shows a stylistic similarity between motifs found in the monumental petroglyph panels and ceramics of the Arauquinoid and Valloid series. Several of the motifs found in the petroglyphs are also found in the style 4 paintings of the Parguaza sequence: masked figures, concentric circles, stylized birds and anthropomorphic figures, mammals, and lizards, suggesting broad contemporaneity. Nonetheless, the emphasis on snakes in the monumental art supports the association of these sites with the myths of Tukano and Arawak-speaking groups who tie their origins to the emergence, travels, and resting places of ancestral serpent-canoes (Riris 2017; Riris et al. 2024, p. 733). This suggests two rock art traditions, one expressed as highly visible petroglyphs in a riverine orientation that commemorate the cosmic origins and horizontal axis of the Yuruparí, and the other, expressed in paintings, associated with more restricted ritual performance in rock shelters of difficult access.

6. Animals, Plants, and the Longue-Durée

In closing, there is one striking aspect of the rock paintings in the Parguaza region that merits further discussion. How can we explain the emphasis on anthropomorphic and zoomorphic figures in contexts that were associated with productive strategies that included the gathering and eventual cultivation of a wide variety of plants? Although it has become increasingly clear that the processes of familiarization, domestication, and cultivation of an immense number of plants and trees took place over millennia in the tropical lands of South America (Fausto and Neves 2018), depictions of flora are rare throughout the rock art sequence (Figure 23). This is particularly noteworthy when we analyze the rock art of styles/periods 3 and 4, which we have argued are associated with the Arawakan expansion throughout northern South America and, eventually, the Caribbean. This diaspora has been correlated with settled agricultural village life (complemented with aquatic resources), social hierarchy, regional social organization (Heckenberger 2002)16, and an “ethos” that, among other things, assigns religion a key role in personal, social, and political life (Santos-Granero 2002, p. 45).

There is, perhaps, a clue to be found in the Yuruparí myth cited above. As we recall, in the second cycle of the myth, Kuwai taught the first ancestors agriculture and the sacred initiation rituals. Following his burning, out of the ashes grew the plant materials used to make the aerophones played in initiation rituals and other sacred ceremonies. The women stole the sacred instruments and led a chase, expanding the world for a second time, and through the music of Kuwai, created the cultural and geographic landscape inhabited by humans and other creatures today (Hill 2011, pp. 261–62). Nonetheless, the men chased them and stole the flutes back, and the current third cycle of the myth commences. The control of Kuwai’s voice and its expression through the different aerophones made from his body parts and other natural species, were kept in the hands of shamans and chant-owners. Imbued with mythic powers, they recreate the ritual landscape and commemorate the place-naming of the ritual cycle during the initiation rituals held for boys and girls. Nonetheless, only in the male initiation rites are the secrets of the sacred aerophones revealed during the boys’ lengthy training in adult male activities.

It is possible that the expansion of rituals such as the Yuruparí, with its origin derived from the struggle for shamanic power between men and women, may reveal a shift in power relations between genders that coincides precisely with the moment when agricultural production, an activity controlled mainly by women, began to play a central role in the productive practices of Orinocan societies. Mansutti Rodríguez comments on this aspect of the Yuruparí/Warime ceremony, where masked figures representing the masters of agricultural products, the master of the waters and all the aquatic animals, and the master of all the animals of the earth and sky are present:

This ceremony consolidates alliances between the shamans and their communities with the masters who control animals and plants exploited by the Piaroa people, facilitates men’s expropriations and limits women’s power to the domestic sphere, and shifts a society with no major hierarchical organizations into a highly hierarchical one as long as the ceremony is occurring (Mansutti Rodríguez 2012, p. 46).

Therefore, although women exercise the day-to-day control of the domestic and agricultural world (which they created in mythical times), they are displaced from the shamanic control of power (through the stealing back of the flutes), from the space of fertility where they have strength (procreation), and from the space whose control could allow them to be autonomous from men (agriculture) (Mansutti Rodríguez 2012, p. 69). It could be argued, then, that the myths serve to justify the male role in the ritual sphere as the promoters of fertility and protectors of the well-being of the community. Ritual practice continues to focus on the shamanic negotiations with the masters of the animals and plants in spaces dominated by men and prohibited to women and the uninitiated. This may help us to comprehend the use of rock shelters for ritual purposes, with the frequent depictions of (mostly male) anthropomorphic and zoomorphic figures characteristic of styles/periods 3 and 4, even in societies increasingly dependent on agriculture and forest management. This ritual affirmation of gendered power relations in situations of subsistence change can be found in rock art contexts outside the study area. As Whitley has argued for the prehistoric Coso Range in the Great Basin, a shift toward intensive female seed gathering was not reflected in the rock art, which, to the contrary, increasingly emphasized males and their activities, especially males’ importance in rainmaking (Whitley 2023, p. 295). “Recognizing that culture in part serves an ideological purpose by masking the real nature of social relations and thereby supports the status quo, it is then apparent that the articulation of art and social relations occurs within a larger cultural system of ideological symbols” (Whitley 1994, p. 367).

As we argued earlier, the later styles/periods of rock paintings (5 and 6), with their greater emphasis on geometric/abstract motifs, may be associated with newcomers to the region. As the Arawakan groups expanded northward and eastward along major rivers, they were apparently replaced by Cariban and Salivan occupations in the Parguaza region. The shift in content we have described for the later periods of rock art may indicate a more typically Guianese form of ritual practice, less focused on hierarchy, specialty roles, and gender segregation typical of the Northwest Amazon, and more concerned with the relationships of alterity and insider/outsider dealings (Henley 2001). This suggestion merits further comparative research.

7. Final Remarks

Given the complexity of the rock art sequence in the Parguaza region and neighboring areas, we have only begun to touch on the potential for analysis of the images and their context. We have considered here significant shifts in emphasis over time in the larger rock shelters, where superpositioning has allowed us to discriminate between different periods, and the styles associated with them. The lengthy sequence ranges from the early, sporadic paintings of pre-ceramic occupations, to the extraordinary murals that we argue are associated with more sedentary, hierarchical societies centered on elaborate ritual performances, and finalizes with a swing toward more abstract representations, executed with less of an eye to realism, that may coincide with the arrival of new occupations to the region from the Guianas to the east. Continuity can be found throughout the sequence in (1) the use of the same sites and surfaces for paintings, often with intentional superpositioning, (2) the overall emphasis on the depiction of easily recognizable human and non-human figures, (3) the rarity of phytomorphic motifs, (4) the general lack of framing of panels, (5) infrequent interaction between figures, and (6) sufficient similarity between certain motifs as to suggest visitation between sites. Nonetheless, important differences between styles/periods can be found in (1) the type, color, and technique of painting typical of each style/period as described in Section 4, (2) the appearance of bichrome “ghost” figures, massive displays of fish and other zoomorphs, masked figures and interactive hunting/fishing scenes associated with styles/periods 3 and 4, and (3) the increase in geometric, abstract figures in styles/periods 5 and 6. These shifts in visual representation may imply different ontological perspectives and forms of practice.

The turn to perspectivism and animism in the ethnographic analogies employed in the recent interpretation of rock art sites in the Northwest Amazon has offered insights into the images and their possible meanings, especially in areas where indigenous peoples continue to ascribe meaning to the sites and figures they contain. As Whitley has observed for the North American Great Basin, “The ethnographic comments…reflect widespread cultural beliefs that are based on specific kinds of knowledge, and a non-Western/scientific ontological system” (Whitley 2013, p. 84). In this sense, the use of ethnographic analogy offers valid insights into shared belief systems in the longue durée. We have explored the relevance of the ethnographic information that characterizes the Northwest Amazon, with emphasis on the Arawakan traditions to interpret the shift associated with styles/periods 3 and 4. We must proceed with caution, nonetheless, the further back in time we go, as in the case of style/period 2, and in situations for which there is insufficient ethnographic information, such as those associated with styles/periods 5 and 6.

Variations that are specific to different time periods and types of sites, and how they relate to geographical features and resources, visibility, and accessibility, must ideally be taken into consideration when applying ethnographic analogy. Social relations and productive modes, while difficult to infer from the current archaeological record, most assuredly impacted the conceptions and practices associated with the rock art and need to be considered when seeking to understand the significance of the images and their context. To this end, in the future, it would be fruitful to expand the analysis we have presented in this study to include the paintings documented in smaller rock shelters and open-air sites and petroglyphs found both in riverine settings and on outcrops and rock shelters inland. Further investigations at rock art sites and related habitation sites would enable us to refine our knowledge of the contexts.