A horizon, by definition, is characterized by a widespread culture that lasted a relatively short time (Willey and Phillips 1958). The Chavin culture, recognized and promoted by Tello over 60 years ago, was believed to be that widespread culture (Tello 1930, 1942, 1943, 1960). Since that time, however, questions have arisen concerning the Chavin culture and the Early Horizon in general. In the 1940s, Rafael Larco Hoyle defined and named the coastal Cupisnique culture, which he promoted as evidence for an earlier coastal manifestation of Chavin culture, thereby questioning Tello’s ideas of a highland origin (Larco 1941, 1946: 149-150, 1948). As early as 1951, Gordon Willey recognized a problem with the definition of the Chavin culture in the archaeological record (Willey 1951). According to the strict definition of that phenomenon, only objects (ceramics, stonework, bone and shell artifacts) decorated in the style manifest at the type site of Chavin de Huantar should be considered
Part of Chavin culture that included fantastic renderings of felines, snakes, and raptorial birds among other beings. This definition, however, seemed overly restrictive to numerous scholars who expanded the archaeological characteristics of Chavin culture to include:
1) specific types of incised pottery found with other artifacts decorated in the Chavin style;
2) additional types of pottery associated with the incised pottery; and/or 3) stone or adobe-brick mounds characterized by a U-shaped architectural plan, particularly mounds adorned with mud friezes or painted murals (Bennett and Bird 1960: 91-102; Bushnell 1963: 43-56;
Mason 1969: 43-59). Most commonly, the Chavin culture was interpreted as a religious cult spread by peaceful means throughout north and central Peru, perhaps along established trade routes (Patterson 1971: 41-46). An implicit or sometimes explicit assumption was that the religious cult began at, and expanded from the type site of Chavin de Huantar in the north-central highlands (Bird and Bennett 1960; Bushnell 1963; Tello 1960), although some scholars favored a coastal origin for the cult (Larco Hoyle 1941).
In the highlands, Chavin de Huantar was partially excavated by Tello in the 1940s, but then its burial by a landslide, retarded research for several decades. In the 1960s, a Japanese team conducted extensive excavations at Kotosh, a deeply stratified site that yielded valuable ceramic and architectural information relating Chavin to other archaeologically defined cultures dating both earlier and later (Izumi 1971; Izumi and Terada 1972). On the coast, the major excavated evidence believed at the time to pertain to the Chavin culture was obtained by the partial clearing in the 1930s of the Punkuri and Cerro Blanco mounds in the Nepena Valley and the Moxeke and Cerro Sechin mounds in the neighboring Casma Valley (Tello 1943). Small stratigraphic excavations in midden deposits on the north and central coasts helped flesh out chronological details (Strong and Evans 1952), but these data were usually overshadowed by more spectacular artifacts, that almost invariably came from looted contexts (Bushnell 1963; Roe 1974; Willey 1971).
The Rowe Seriation
Scholarly attempts were made to establish both a time frame for the Chavin culture, and to understand its spread over a wide area of Peru. John Rowe (1960) first established the horizon/ period chronological framework that is still used today by many Andean scholars, replacing more interpretive terms like “Cultist Period” (Bennett and Bird 1960: 91-102). According to this scheme the Early Horizon was the period of pan-Andean Chavin cultural influence. Rowe then turned his attention to the Chavin culture itself. Through careful examination of architecture at the site of Chavin de Huantar he proposed a construction sequence that he used to support a long stylistic chronology of Chavin stone carvings and designs at that site. He then used stylistic comparison with hundreds of ceramic vessels from burials in the Ocucaje area of the Ica Valley on the south coast to show that stylistic changes followed developments in sculpture at Chavin de Huantar, documenting continuous interaction (Rowe 1962, 1967). An assumption behind this analysis was that Chavin de Huantar was the source and that changes began there and spread outward. Most scholars embraced Rowe’s stylistic sequence for the clarity and order it appeared to bring to a complex phenomenon; some expanded upon his conclusions, culminating in an extreme extrapolation of the sequence by Peter Roe (1974).
Redefining the Extent of Chavin
During the mid-1970s through the early 1980s extensive excavations at Huaca de los Reyes (T. Pozorski 1975, 1980, 1982), Garagay (Ravines and Isbell 1975), Cerro Sechin (Samaniego
Et al. 1985), and Chavin de Huantar itself (Burger 1981) uncovered an array of data that did not fit the Rowe Chavin stylistic sequence. Coastal sites yielded radiocarbon dates that were much older than dates from Chavin de Huantar, although investigators were initially reluctant to accept these dates and their implications. A case in point is the site of Huaca de los Reyes, in the Moche Valley. Huaca de los Reyes has a U-shaped plan, believed at the time of discovery to be diagnostic of a classic Chavin center. When the first mud friezes were uncovered and published with radiocarbon dates in the early 1970s (Moseley and Watanabe 1974; T. Pozorski 1975), the immediate reaction of the academic community was to attempt to fit the frieze designs into Rowe’s Chavin stylistic sequence. The consensus at the time was that the site dated to the latter part of the Rowe sequence despite the fact that three of the four radiocarbon assays predated 1300 cal BC, well before any published date for the site of Chavin de Huantar. It was only with the publication of Burger’s (1981) seminal article on the late dating of Chavin de Huantar, relative to the coastal sites of Huaca de los Reyes, Garagay, and Cardal, that scholars accepted the idea that most coastal U-shaped mounds decorated with mud friezes were constructed during the Initial Period, before 900 or 1000 cal BC, rather than during the Early Horizon, so new interpretations of cultural dynamics were required. More than a quarter-century of further investigations has continued altering and improving our understandings of Chavin culture, the Early Horizon, and the Initial Period.
Fieldwork at Chavin de Huantar since 1975
Based on his excavations at Chavin de Huantar and especially at the site of Huaricoto in the Callejon de Huaylas, Burger proposed a much narrower time frame (400-200 cal BC) for the spread of the Chavin religious cult (a Chavin horizon) over a fairly large area of the central and north-central highlands as well as along much of the Peruvian coast (Burger 1981, 1985, 1992; Burger and Salazar-Burger 1980, 1985). Archaeologically, Burger defined this spread as coinciding with the dissemination of Janabarriu style pottery (black ware decorated with impressed circles, circles and dots, S-shapes) along with other artifacts bearing Chavin-style iconography. This proposal was met with enthusiasm by many scholars because it seemed to clarify a complex situation and reinterpret the Chavin religious cult as a true horizon, i. e., a short time period dominated by a short-lived but widespread culture (Willey and Phillips 1958). The Early Horizon, from its earliest formulation, had encompassed an unrealistically long span of time, usually 700-1000 years (Lanning 1967: 25; Rowe and Menzel 1967: ii). Thus, Burger’s proposal seemed more reasonable, at least regarding the spread of the Chavin religious cult.
Unfortunately, by collapsing the duration of the Early Horizon to 200-300 years, the earlier Initial Period increases; its length expanding from 900 to more than 1500 years. This lengthened Initial Period is problematic for scholars, especially the time span between 1000 and 500 cal BC, because by about 1000 cal BC, many of the Initial Period coastal mound centers were abandoned, creating a time gap from about 1000-600 cal BC, at least along the coast. However, data from coastal sites in the Nepena Valley (Cerro Blanco, Huaca Partida) and the Lurin Valley (Cardal, Manchay Bajo) reveal occupations lasting until 800-600 cal BC. And some sites were reoccupied by 500 cal BC or later by local cultures only marginally affiliated with the Chavin religious cult. Perhaps the gap is already disappearing (Burger 1987, 2003; Burger and Salazar-Burger 1991; S. Pozorski and
T. Pozorski 1987; T. Pozorski and S. Pozorski 2005; Shibata 2006).
In the early 1990s, building on his earlier work on Chavin, Burger painted an eloquent picture of how the Chavin religious cult spread during the late Early Horizon (Burger
1992: 225-227). This phenomenon was characterized by a stratified society where power and prestige were gained through religious ideology, although elites appropriated control of long-distance trade and exotic goods. Archaeologically, the Chavin religious cult (or Chavin horizon or civilization) is differentiated from earlier Initial Period cultures by the appearance of burials containing wealth items such as Spondylus shell, obsidian, and, most importantly, gold and silver artifacts (Burger 1992: 225-227). According to Burger, complex society started in the highlands with the Chavin civilization. We take issue with Burger’s characterization of coastal Initial Period societies as lacking societal complexity (see Initial Period section above), as well as his inferences about the extent and influence of the Chavin religious cult.
Burger (1992) correctly distinguished the numerous Initial Period sites as separate from the Chavin phenomenon; and by doing so, he eliminated the U-shaped mounds that for decades had been the assumed diagnostic of Chavin presence along the coast. When this happened, much of the purported coastal evidence for Chavin disappears. What remains of the archaeological evidence for the spread of the Chavin religious cult are an assortment of ceramics, shell and bone artifacts, metal artifacts, stone artifacts, textiles, petro-glyphs, and some decorated architecture tied to the Chavin religious cult primarily based on artistic style. Ceramics are by far the most common artifacts of the Chavin religious cult. Fragmentary Janabarriu type sherds have been found in secure, often dated contexts, well beyond Chavin de Huantar (e. g., Burger and Matos Mendieta 2002). Most spectacular whole ceramic vessels portraying feline heads or other fantastic beings come from looted contexts (Burger 1992: 214-216) and are attributed to the Chavin horizon based on motifs that they share with Chavin stone sculpture or with more securely dated sherds from stratigraphic contexts or architectural clearings. Many of the other classes of artifacts have little or no context, having been looted or otherwise unscientifically recovered from, often unknown, regions of the coast and highlands.
This brings us to current investigations at Chavin de Huantar by John Rick and associates (Kembel 2001; Kembel and Rick 2004; Rick 2005). Intensive mapping and excavation reveal that the architectural and sculptural history of the site is much more complex and long-lived than Rowe’s simple Old Temple/New Temple seriation in the 1960s. Construction at Chavin de Huantar expanded both horizontally and vertically during 700 years, from 1300 to 600 cal BC (Rick 2005). Throughout this time the Chavin elite were manipulating the Chavin cult and its accouterments to their own ends as a means to consolidate power and status (Rick 2005). The ramifications of these new data, including the radiocarbon dates, have yet to be determined for a definitive understanding of the Chavin phenomenon.
A Scenario for Chavin Consolidation and Expansion
We believe that Chavin was an amalgamation of ideas derived from outside the highland site proper. Tello (1942) recognized this when he noted elements in the iconography at Chavin de Huantar that suggested a tropical forest source. As the chronological placement of coastal sites has been revealed by radiocarbon dating, elements of these sites’ ceramics, architecture, and iconography can be recognized as the probable influences on Chavin, as ideas moved from coast to highlands. For instance, the stirrup-spout bottle has a long history on the northern Peruvian coast and greater antiquity further north in Ecuador. It is likely that the bottle form with a narrow, trapezoidal-shaped stirrup spout that is typically described as Cupisnique provided the inspiration for Chavin stirrup-spout bottles. Once integrated into the Chavin culture, the bottle changed: the spout became more rounded
In arc, fatter in cross-section, and often decorated (Burger 1991:189; Elera 1993:252; S. Pozorski and T. Pozorski 1987:58-59, 68-69). Especially noteworthy examples of architectural traits absorbed by Chavin are the U-shaped mound form and the often-associated sunken circular plaza that seem to have originated in the Fortaleza-Huara zone of the north-central coast (Shady 2006). Ideas, and perhaps belief systems, reflected in the coastal iconography were also incorporated by Chavin: profile felines depicted in friezes at Huaca de los Reyes and Pampa de las Llamas-Moxeke appear in the stone sculpture bordering Chavin de Huantar’s sunken circular plaza, and minor elements of other Huaca de los Reyes friezes (pendant snake heads, agnathic jaws) are also duplicated in the more ornate Chavin reliefs (T. Pozorski 1975, 1980, 1982; S. Pozorski and T. Pozorski 1986).
Chavin de Huantar probably consolidated its power as elites gathered artifacts, architecture, iconography, and ritual from diverse arenas to solidify new social positions. With this power base a Chavin polity aggressively colonized certain areas. We interpret new data from the lower Nepena Valley mound sites of Cerro Blanco and Huaca Partida as indicating they were Chavin outposts, based on coincident iconography, ceramics, and radiocarbon dates (Koichiro Shibata 2006). Development and maintenance of the Sechin Alto polity, centered in the Casma Valley, likely left a near-vacuum in the neighboring Nepena Valley to the north as the population of this smaller valley was drawn to Casma. If so, the Nepena Valley was wide open to penetration by the Chavin migrants, and it is in this valley that we find Chavin outposts reaching furthest toward the coast.
Pallka, located in the upper Casma Valley, also was probably a Chavin colony, based on its mound form and abundant Chavin ceramics (S. Pozorski and T. Pozorski 1987: 88-89; Tello 1956: 32-48). Colonization was more difficult in the Casma Valley and hence limited to its upper reaches.
Possibly a third and small Chavin colony is Huaca Guavalito in the Caballo Muerto Complex of the Moche Valley where several Janabarriu ceramic types are associated with the small mound (T. Pozorski 1983:10-19).
Other sites that show significant Chavin influence and presence are the highland sites of Kuntur Wasi, Kotosh, and Huaricoto (Burger 1985; Burger and Salazar-Burger 1980, 1985; Izumi 1971; Izumi and Terada 1972; Kato 1993). At Kuntur Wasi, Chavin influence is reflected in ceramics and gold artifacts found in elite burials. Both Kotosh and Huaricoto had a long tradition, extending from the Late Pre-ceramic, of ritual hearth use. Perhaps this enabled the settlements to resist Chavin influence more effectively, limiting its presence to ceramic sherds from site refuse. Significantly, the Chavin polity does not seem to have adopted the ventilated hearth chamber and associated rituals into its amalgamation of architecture and belief systems even though several coastal sites apparently participated in variants of this Kotosh Religious Tradition (T. Pozorski and S. Pozorski 1996).
In contrast to Chavin colonies and sites exhibiting significant influence clearly traceable to Chavin de Huantar, there are numerous sites of local cultures only marginally impacted in the Chavin phenomenon. These include sites such as Cardal and Manchay Bajo in the Lurin Valley, where radiocarbon dates and ceramics document occupation that continued into the Early Horizon (Burger 1987, 2003; Burger and Salazar-Burger 1991). However, the local large, asymmetrical mound form continues to be used in accord with the architectural tenets of the Initial Period. In the lower Casma Valley, marked changes are evident between the Initial Period and the Early Horizon; however, these later Casma activities reflect primarily local developments (S. Pozorski and T. Pozorski 1987: 51-70). Small platform and room complexes with diverse orientations replaced mound-oriented sites with central axes of the Initial Period. This late settlement pattern is typical of Pampa Rosario, San Diego, and
The reoccupation of Sechin Alto site in the Casma Valley (S. Pozorski and T. Pozorski 1987:
51-79; T. Pozorski and S. Pozorski 2005) as well as Caylan and Huambacho Viejo in the Nepena Valley and Macabalaca in the Huarmey Valley. Connections with Chavln are represented only by occasional use of Chavln decorative motifs on ceramics (e. g., stamped circles and dots) and shared artifact types, most notably ground slate points.
This reconstruction is tentative. New data from John Rick’s ongoing work at Chavln de Huantar and research by other scholars at other Early Horizon sites will surely lead to refinement and revision. Nevertheless, we think the data show that Chavln was a developing and expanding polity, albeit one whose strength and appeal seem especially strongly grounded in ritual and religious symbolism (Burger 1992; Rick 2005). Chavln’s origins lie earlier, in Late Pre-ceramic and Initial Period cultures that we think achieved state government, and perhaps even city life—at least during the Initial Period. However, we need not assume that Chavln dominated the entire Early Horizon Period. There were certainly numerous and varied local developments. In Chapter 35 of this volume, Richard Burger presents an alternative interpretation of Chavln.
World History
