Abstract: This paper presents and explores names of places pertaining to the southern Yunga languages – Muchik or Quingnam – from the valley of Nepeña (Ancash, Peru). Toponyms include possible Quechua-Yunga compounds and, possibly Muchik-Quingnam hybrids. Their regional distribution is described and their temporal placement discussed. Archaeological data patterning, the location of sacred waka places, routes of interregional interaction and political developments are described. Enduring multilingualism – coupled with established oracular shrines – is put forward as an alternative to language replacement theories.
Keywords multilingualism, historical linguistics, toponyms, archaeology, Muchik, Quingnam, Quechua, Ancash, Peru.
Resumen: Este trabajo presenta y explora topónimos en las lenguas yunga del sur, quingnam y muchik, en el valle de Nepeña (Ancash, Perú). Los nombres identificados incluyen posibles voces compuestas quechua-yunga y, posiblemente, muchik-quingnam. Se expone su distribución especial y se discuten las implicancias cronológicas. Con base en una comparación con la distribución espacio-temporal de datos arqueológicos, la ubicación de lugares sagrados waka y rutas de interacción interregional, así como con dinámicas políticas, se propone el multilingüismo duradero, vinculado a templos oraculares establecidos, como modelo alterno a teorías de sustitución de lenguas.
Palabras clave: multilingüismo, lingüística histórica, topónimos, arqueología, muchik, quingnam, quechua, Ancash, Perú.
Multilingualism on the North Coast of Peru: An Archaeological Perspective on Quingnam, Muchik, and Quechua Toponyms from the Nepeña Valley and its Headwaters
Multilingüismo en la costa norte del Perú: una perspectiva arqueológica sobre los topónimos quingnam, muchik y quechua del valle de Nepeña y sus cabeceras (Ancash, Perú)
As the medium for social intercourse par excellence , language plays a crucial part in human interaction and social memory. It is one of the most pervasive indicators of social identity at spatial scales ranging from the local to the regional, and at social scales articulated by relations as kinship, ethnicity, class and beliefs. At the scale of the ethnic group, unity of language or dialect has been posited as one of the four main constitutive characteristics of ethnic identity in the 16th century central Andes, along with unity of origin and beliefs, dress, and socio-political leadership (Rostworowski 1991; Topic 1992, 1998). There is ample historical evidence from the central Andes of people ‘following’ different caciques yet co-residing in early 16th century villages – from Conchucos (Cook 1976-1977, 1981; Espinoza Soriano 1964; León Gómez 1994, 2003) as well as from the eastern escarpments of Huánuco (Hastings 1985, 1987; Ortiz de Zúñiga 1972), the Titicaca Altiplano (Diez de San Miguel 1964; Hyslop 1976) and the northern coast of modern Peru (Ramirez 1985, 1999).
Archaeological evidence strongly suggests that the ample regional variation in the configuration of the specific relationship established between the Inka and local ethnic groups discerned from historical sources (Pease 1982) was not an Inka prerogative. Instead, the establishment of settler enclaves – similar to Inka mitmaq colonists – as well as multi-ethnic settlements may have been the norm, rather than the exception, at least as far back as the 6th century AD (Goldstein 2000; Ryan Williams, Isla & Nash 2002). Arguments for high seasonal mobility in pre-Colonial times have also been brought forward for the Inka and his entourage (Ramirez 2005) as well as for ceramic specialists from across the central Andes (Ramón 2013). These and other findings indicative of high mobility are at odds with the distinct and bounded “glottogeographical” (Cerrón-Palomino 1995: 38) language areas which suggest relatively homogenous, and mutually exclusive language areas, offered as reconstruction by linguists (e.g. Torero 2002: Fig. 5, Fig. 6; Cerrón-Palomino 1995: mapa II). Despite the distinct probability of pre-Columbian multilingualism being widely spread, however, the issue of multilingualism and its implications has scarcely been addressed by scholars working on the central Andean coast.
Naming plays a significant role in place-making and toponyms do not infrequently make reference to events – real, imagined or a mixture of both. In the 19th century Ernst Middendorf (1894-1895: 22-23) noted that different sections of the Nepeña River had specific names. Like elsewhere in the Andes and beyond, each name is traditionally associated with a different section of the valley. Today, circumscription of named (toponymic) areas is rooted in a keen awareness by campesino peasants of their social memory about territorial claims as well as shared dependence on irrigation water from particular sources, be these lakes, springs or canal intakes. The presence of major archaeological sites at major hydrological junctures indicates that substantial parts of the canal system have remained unchanged over the centuries. In the lower south section of the Nepeña Valley, for instance, the name ‘Rio Viejo’ suggests a geomorphological alteration in river channels that may have been precipitated by human intervention, natural processes or a combination of both. Place names may also express measures of attachment, reified in social interaction and stabilised through practice, that provide indications of the deployment of power relations in space. When durable relations about places sediment in history, toponyms can provide valuable insights into the meanings of places, although combined archaeological, historical, ethnographic and linguistic study may be required to ascertain what lies behind each name (e.g. Chávez 2003).
Like other forms of expressing identity overtly – ethnic or otherwise – the command of multiple languages can be expected to have played a more prominent role in areas of intense contact between speakers of different languages, not only along boundaries and roads but also in multi-ethnic settlements. The nature, spread, relationships and change through time of languages spoken across the 16th century Andes (Torero 2002) has recently received increased attention thanks to the publication of a handbook (Adelaar 2004), as well as major publications on Muchik (Brüning 2004; Cerrón-Palomino 1995; Hovdaugen 2004), a new debate on the origins, dispersal and relationship between Quechua and Aymara (Heggarty 2005 and comments therein; Heggarty & Beresford-Jones 2010 and comments therein), and the search for archaeological correlates of competing language dispersal models in the archaeological record (Cerrón-Palomino et al. 2010; Heggarty & Beresford-Jones 2012). Here I begin to develop the argument that multilingualism decidedly shaped the development of the complex and dynamic language mosaic encountered in the Andes at the time of conquest.
The necessary point of departure for reviewing changing relationships between people speaking different languages and the spatial distribution of socio-political formations in the Andes, is their distribution at the time of contact. The details of this picture I have been able to reconstruct, based on historical linguistics and the distribution of toponyms in the landscape, suggest a complex linguistic history of wide-spread multilingualism, stronger in some areas and places than others, as discussed in following.
Linguistic readings of colonial sources and toponymic study pertaining to the broader research area (Adelaar 1989, 2004; Torero 2002: 203-273) indicates that Quingnam was the predominant language spoken in the coastal valley of Nepeña at the time of conquest; Quechua in the neighbouring inter-Andean Santa and upper Yanamayo valleys to the East; and Culle to the north and further east, possibly including a corridor along the coastal section of the Santa Valley (Figure 1). This interpretation is consistent with the encomienda grant signed by Francisco Pizarro in January of 1535, partially published by Schaedel (1985: 453, 463), which makes reference to a cacique and settlement with Quingnam names “[...] Çuy Çuy lord of the town of Nasapac [...]” (Schaedel 1985: 474; author’s translation), a likely forebear of Don García Suy Suy, the cacique of Nepeña named in an unpublished document from 1585 cited by Soriano Infante (1947: 269; Zevallos Quiñones 1993b: 44).
The coastal Yunga language complex includes Muchik, also known as Mochica, as well as Tallán, Olmos and Sechura and Quingnam, a set of poorly known distinct languages which were spoken along the central and northern coast of modern-day Peru (Cerrón-Palomino 1995; Hovdaugen 2004; Lehmann 1929-1937; Middendorf 1892, 1894-1895; Torero 2002: 203-223). 1 According to Cerrón-Palomino (1995; cf. Adelaar 2004), Quingnam – as mentioned by Agustinian friar Antonio de la Calancha (1653, cited in Torero 2002: 216) – was the southernmost and most widely spread language of the Yunga complex. Quingnam has been emphatically argued to be the same language which other early colonial sources refer to as Pescadora (Adelaar 2004: 173; Cerrón-Palomino 1995: 36-41, 181; Rostworowksi 1977), possibly representing a southern variant of the former (Torero 2002: 217).
At the time of conquest, Cerrón-Palomino (1995: mapa II) has argued, Quingnam was spoken from the Chicama Valley in the north, where it overlapped with Muchik, to the Rimac Valley in the south. Similarly based on Calancha’s account Torero had earlier mapped a similar area of distribution to the north yet limited to the Santa Valley in the South, albeit with one main difference: it was not homogenous (2002: Fig. 5). Based on toponym distribution Torero suggested the existence of a Culle speaking corridor in the lower Santa Valley, a westward extension of Culle speakers otherwise restricted to the northern highlands - which would have effectively split the Quingnam area in two.
Cerrón-Palomino (1995: 42-46) also follows Calancha in linking the distribution to the Chimú expansion, whereas the spread of Muchik is suggested to have begun earlier, and is attributed to the polities of Moche and Sicán. If Calancha (1653: cited in Torero 2002: 216) was correct in stating that Chimú conquests were directly responsible for the spread of Quingnam, and the spread of Muchik is accepted as a phenomenon similarly linked to the territorial expansion of a state-level polity, the spread of Muchik would have preceded that of Quingnam by seven or eight centuries. Yet even if the highly contentious coupling of language spread and state expansion were to be accepted – e.g. as part of a purported policy of subjugation – questions regarding what language or languages, was or were spoken before the purported territorial expansions loom large.
Assertions regarding the importance and prestige of the Quingnam language need not only be based on Calancha’s account; there are indications that Domingo de Santo Tomás – author of the first Quechua dictionary and grammar – had a working knowledge of it (Juan Meléndez 1681: I, libro 4, capítulo 2, pág. 325-326, cited in Cerrón Palomino 1995). Unlike Mochica, which remained in use until the early 20th century, however, Quingnam appears to have become extinct by the middle of the 18th century (Torero 2002: 222, see also: Quilter et al. 2010). No grammars or dictionaries have survived, however, leaving lists of names of people and places (Zevallos Quiñones 1993a, 1993b), and numbers (Quilter et al. 2010) as the main sources of study.
The overwhelming number of Quechua toponyms in Nepeña supports the view that Quechua generally preceded Spanish and Quingnam and Muchik generally preceded Quechua in this area’s history. The traces of Quingnam in local toponymy, arguably stronger than those of Muchik, have been suggested by Calancha (see above) to correlate with Chimú colonisation since the colonial period. Yet much like past archaeological interpretations, the above mentioned maps of linguistic distribution appear to assume valleys as relatively homogenous linguistic and cultural units. This is an unlikely scenario for most of the prehistory of the Nepeña Valley. Instead, I will argue that multilingualism was probably widely spread in space and time, and language dispersals possibly coupled to the changing fortunes of specific oracular shrines.
Based on published cartography and complementary sources, I have compiled a list of 22 toponyms likely to be closely related to the Yunga language complex (Table 1). Most have clear Quingnam elements, although a few show elements that are likely to be Mochica or shared by Quingnam and Mochica, raising the likelihood of Quingnam-Mochica compounds. Two place names appear to be prima facie examples of Quechua-Yunga language compounds.
The first group includes the names used to refer to the Valley as a whole in colonial and republican times: Nepeña ( Nepeiñ ?)2 and Huambacho ( Guambách ?), names shared by the main body of the river and one of the most prominent lower valley mountains. A third non-Quechua toponym is Caylán, one of the most extensive archaeological sites in the valley (Chicoine & Ikehara 2010; Helmer, Chicoine & Ikehara 2012; Kosok 1965: Figs. 20-21; Proulx 1968, 1973), which stands by the substantial reservoir of the same name and dates back to the second millennium BC. The suffix -an means house in Muchik (e.g. Sipán, Sicán) (Brüning 2004: 5). It may have also pertained to Quingnam, but further likely Muchik phonemes can be found in the names of other archaeological sites as well. Quisque is a place name of Yunga origin, but it could either be Quingnam or related to kišmik , which means ‘old’ in Muchik (Brüning 2004: 25). The hilltop site of Captuy likewise includes the Muchik root kap- in its name, which means ‘to carry’ (Brüning 2004: 23). Other place names characterised by ending in -pon or -pong – which means ‘rock’ and, by extension, ‘rocky mountain’ in Mochica (Torero 2002: 267) – pertain to the rocky cerros named Popo ( Poo-pon(g) ?), Pimpón ( Pim-pong ) and Huancarpón (see below). This suggests the possibility of compound Quingnam-Muchik toponyms.
Imperfect as the distinction between Quingnam and Muchik still is, the distribution of toponyms pertaining to the Yunga language complex Nepeña opens the distinct possibility that Muchik-speaking populations resided alongside Quingnam-speaking populations. Indications to this effect appear to be stronger in the lower valley, where the valley’s most extensive irrigation areas are found. A converse situation, a predominantly Quingnam-speaking village within a Muchik-speaking area, appears to be what Toribio de Mogrovejo encountered near the end of the 16th century at the seaside village of Eten, near modern Chiclayo (Cerrón-Palomino 1995: 31, fn. 7). A set of questions that follows on from here revolves around the processes and interactions through which Quingnam-Mochica multilingualism might have come about.
Following the standard hypothesis sketched above Quingnam-speaking Chimú colonists would have moved into the valley roughly between the 12th and the 16th century. Muchik speakers would have moved in with Moche’s southward expansion, probably occurring around the 4th - 5th century AD. What language would have been spoken in Nepeña beforehand is unclear. It seems very likely to have been a Yunga language, however. A look at the archaeological evidence for Moche and Chimú administration (i.e. architecture) indicates that it is not homogeneously spread across the whole valley. Instead it appears strictly coupled to the water and land available within distinct hydromorphological units (Herrera Wassilowsky in prep.). Little if any of the archaeological evidence at Caylán, Captuy or Kiske – toponyms which strongly suggest Muchik ancestry – show any evidence of Moche settlement. What they do have in common, however, is their strategic location for transit along, across and out of the valley into the highlands.
In the upper Nepeña Valley and the Cordillera Negra, the distribution of non-Quechua toponyms suggests that Quingnam and Quechua were both spoken in the pre-Colonial past under conditions that made convergence possible. It is thus not possible to reject the possibility of three-way multilingualism in the Nepeña, although discussions and conclusions must be confined to the Yunga language complex as a whole until understanding of the formal differences between Quingnam and Muchik can be refined.
The distribution of toponyms pertaining to the Yunga language complex in the Nepeña Valley suggests enduring patterns of interaction between people speaking different languages that settled in relative close proximity to each other (Figure 2). Such toponyms are not found only along the coast. Their distribution ranges from Huin Huin Lake (ca. 4,500 m.a.s.l.) in the high Cordillera Negra (headwaters of the Huiri/Colcap tributary of the Jimbe River), to locations within both main tributaries’ valleys and the port of Samanco (Table 1 and Figure 2). The intriguing presence in the Huin Huin Lake area of Cerro Huin Huin and other locations prefixed Huin- , including a toponym followed by the likely locative -co ( Huin-co ; also Coishco, Samanco, Huanchaco) may suggest the presence of a local waka , but it also begs the question of the root. It could pertain to Quechua (as in winchu ) but the repetition (Huin Huin) is typically Quingnam (as in muymuy, maymay, Suy Suy, Puquio Pipí, Cerro Popo). The name of the southern tributary of the Nepeña, Loco, is generally associated with the erratic behaviour of a high energy mountain stream with a large catchment area (Gambini Escudero 1984). It may, however, also be a hispanisation of Lloc ( Lloc-co , as in San Pedro de Lloc, Pacasmayo) that also belies a very old and deep history for Quingnam in Nepeña.
In the upper Nepeña Valley and its tributaries compound Quechua-Quingnam or Quechua-Muchik toponyms suggest parallel settlement by both Yunga language and Quechua speakers. The name of Huancarpón, a major archaeological site, mountain and major irrigation water intake north of the town of Moro, juxtaposes wanka , Quechua for ‘large standing stone’ – often a lithified ancestral waka (Duviols 1979; Falcón 2004) – with the Quingnam/Muchik suffix -pon or -pong meaning rock, or mountain. As discussed above, Quechua is likely to be the younger language substrate, but unlike Muchik (Salas 2004) too little is known about Quingnam to distinguish it from toponymic evidence alone. The likelihood that Quingnam or Muchik (or both) were spoken on the Ancash coast before Quechua spread to the area is also raised by the Yunga language names of valleys to the south, Casma ( Cash-man / Cash-mak ?) and Sechín, as well as the town of Coishco, on the outskirts of Chimbote. In this area particular the possibility Quingnam / Muchik, Culle and Quechua multilingualism is raised by Torero’s interpretation (see above). Scores of likely Quingnam toponyms extending along the coast to Lima, and beyond (e.g. Chancay, Chillón, Ancón, Callao, Cantolao), are further taken to suggest that multilingualism shaped a complex and dynamic language history across much of the later prehistory of the central Andes.
It is still only possible to speculate on the timing of the different multilingual scenarios on the precolonial North Coast of Peru: Muchik-Quingnam, Muchik-Quechua and Quingnam-Quechua. The later introduction of Quechua would have led to a bilingual Quechua/Yunga language situation in which Quechua eventually came to be preeminent. Dating this scenario is as yet uncertain, however. Quechua in Ancash predates the Inka expansion (Cerrón-Palomino 1987), and Heggarty & Beresford-Jones (2010) have linked the northward spread of southern Quechua to purported Wari militarism during the 6th century AD Middle Horizon. The Ancash highlands are usually regarded as a likely origin hearth for (proto-) Quechua,3 however, the hypothetical ur-language that would have given rise to both the northern and southern later variants. It is thus tempting to suggest linkages between the Chavín cult of the second and first millennium BC, the site of Chavín de Huántar and the spread of Quechua (Heggarty 2005).
Without firm bases for dating many of the above hypotheses must remain speculative but they do raise the distinct possibility that multilingual situations in the Cordillera Negra extended over millennia. The spatial distribution of toponyms pertaining to the Yunga language complex, moreover, allows pin-pointing the areas in which such conjunctions are most likely to have unfolded. It is in these areas that further archaeological and linguistic scrutiny is likely to be particularly rewarding.
As my review of non-Quechua place-names in the Nepeña Valley and its Cordillera Negra hinterland shows, the vast majority of non-Quechua toponyms make reference to lakes, rivers, ravines springs and mountains, that is, to landscape referents linked to water availability. Such places are crucial for agro-pastoral ways of life. Many are likely to have been considered animate places in the 16th century, the living and speaking entities commonly referred to as waka .
That first settlers to South America spoke the same language is an unwarranted assumption (Heggarty & Beresford-Jones 2010: 177) and, as Lucy (2005: 92) reminds us, it is most unlikely that the multilingualism shared by most of humanity today is a recent development. Shared language, like shared material culture, may correspond with ethnicity but the conditions under which this occurs are only rarely clear-cut (Hodder 1982). Languages also have much to do with people’s choices, both in forms of self-expression and the directionality of interactions, choices coupled to their perceived prestige and other causes which may have no material correlates, but are very much part of the real world. The adoption and intensification of particular farming practices or crops probably helped enhance the prestige of the particular temples, oracles or polities that succeeded in leading their adoption and adaptation to local conditions.
Yet it remains unclear which particular language or languages was or were spoken at the temple of Wampukayán, better known as Cerro Blanco in the archaeological literature (Tello 2005). Here the transition between the brewing of beer from Maniok and Maize has been dated to the first half of the first millennium BC in Nepeña Valley (ca. 1100-450 BC cal.) (Ikehara, Paipay & Shibata 2013; Tykott, Burger & van der Merwe 2006). Wampu, however, is the name of a deity described by Tello (2005: 39-41) as ‘floating’ deity that was as important locally as Wallallo (Karwa Anchi Wallallo) in the highlands of Lima. It is also the name of an imposing section of the Yungay Ignimbrite rock formation at the juncture of the Casma and Nepeña Valley (Coldwell, Clemens & Petford 2011), a location that probably played a prominent role in the Inka struggle against the Chimú. While Quechua is likely to have been the main language spoken by the Inca, and Quingnam by the Chimú, the spread of both languages in Nepeña is most likely to pre-date the Late Intermediate Period/Late Horizon interphase. Yet perhaps more than the age or velocity of language spread, the languages spoken by the deities, and their oracles, need to be carefully considered in any historical reconstructions (Herrera Wassilowsky in prep.).
Multilingualism is likely to have formed an integral part of a complex and dynamic history across the Andes. It was probably the norm rather than an exception occurring along the boundaries of otherwise homogenous culture areas or political units. A mechanism through which multilingualism could have been brought in the Andes is population resettlement – voluntary or enforced – and the coupled transposition of local waka deities speaking non-local languages. Waka veneration legitimated the presence of the original speakers of one language in areas in which other languages were dominant. Over centuries, as the oracles were consulted, grew in prestige or fell from favour, complex multilingual scenarios would have arisen, such as those whose traces I have aimed to reconstruct. Crucially such a model may help explain the perseverance of certain languages over small areas. As the voice of the ancestors and gods as well as ‘speaking places’ in their own right, I suggest that waka were central to linguistic dynamics.
2 Father Soriano Infante’s (1947) linking of Nepeña to the Spanish language (‘Ni peña’) fits his forcing of a Christian foundational myth.
3 Personal communication, Paul Heggarty 2004.