Linguistic Stratigraphy: Recovering Traces of Lost Languages in the Central Andes 3031421019, 9783031421013

Citation preview

Matthias Urban

Linguistic Stratigraphy Recovering Traces of Lost Languages in the Central Andes

Linguistic Stratigraphy

Matthias Urban

Linguistic Stratigraphy Recovering Traces of Lost Languages in the Central Andes

Matthias Urban Center for Advanced Studies ‘Words, Bones, Genes, Tools’ University of Tübingen Tübingen, Germany

ISBN 978-3-031-42101-3 ISBN 978-3-031-42102-0 (eBook) https://doi.org/10.1007/978-3-031-42102-0 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland Paper in this product is recyclable.

Acknowledgments

Work on this book was supported by grants number UR 310/1–1 and UR 310/1–2 from the Deutsche Forschungsgemeinschaft (German Research Foundation, DFG). My sincere thanks to an anonymous reviewer for providing detailed and helpful comments.

v

Contents

1 Linguistic Stratigraphy, Or How to Recover Traces of Lost Languages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 4

2 The Central Andean Linguistic Landscape Through Time and the Quechuan Language Family . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 The Ever-Changing Cultural and Linguistic Landscapes of the Central Andes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Quechuan Today . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 The Classification of Quechuan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 The Quechuan Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 The Non-Quechuan Linguistic Diversity . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5 7 11 12 14 20

3 Evidence and Methods for Investigating Substratal Languages . . . . . 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Toponyms and Anthroponyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 Substrate Influence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4 Substrate Vocabulary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5 Extralinguistic Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

25 25 28 36 38 40 40 42

4 The Aymara Presence in Southern Peru . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Toponymy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 Substrate Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4 Substrate Vocabulary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5 Perspectives from Outside Linguistics . . . . . . . . . . . . . . . . . . . . . . . . . 4.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

45 45 47 49 58 66 66 67

5

vii

viii

Contents

5 The Barbacoan Languages and the Southern Ecuadorian Highlands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 5.2 Toponymy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 5.2.1 Characteristic Toponymy of Cañar and Azuay . . . . . . . . . . . . 80 5.2.2 Continuous Northward Extensions of Cañar and Azuay Toponymic Endings . . . . . . . . . . . . . . . . . . . . . . . . 82 5.2.3 Noncontinuous Northward Extensions of Cañar and Azuay Toponymic Endings to the Ecuadorian-Colombian Border . . . . . . . . . . . . . . . . . . . 89 5.2.4 Northward Extensions of Phonological Characteristics of Cañar and Azuay Toponyms . . . . . . . . . . . . . . . . . . . . . . . . . 94 5.2.5 Evaluation of the Toponymic Evidence . . . . . . . . . . . . . . . . . . 96 5.3 Substrate Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 5.4 Substrate Vocabulary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 5.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 6 Chachapoyas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 Toponymy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3 Substrate Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4 Substrate Vocabulary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5 Perspectives from Outside Linguistics . . . . . . . . . . . . . . . . . . . . . . . . . 6.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

115 115 120 127 133 135 136 137

7 Synopsis and Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

About the Author

Matthias Urban is principal investigator of the Junior Research Group “The Language Dynamics of the Ancient Central Andes”, hosted by the University of Tübingen and funded by the German Research Foundation (DFG)’s Emmy Noether Program. Having held prior appointments at the universities of Leiden, Marburg, and the Max Planck Institute for Evolutionary Anthropology in Leipzig, his research interests include historical linguistics, in particular of the Andes, language contact, and linguistic typology.

ix

Chapter 1

Linguistic Stratigraphy, Or How to Recover Traces of Lost Languages

Abstract This opening chapter introduces the notion of linguistic stratigraphy. This is understood as the study of traces of languages, including undocumented ones, that are no longer spoken because at some earlier point of time, its speakers have shifted to another language that has newly arrived in the relevant part of the world. The chapter argues that such episodes are likely a common phenomenon in human linguistic history. Maintaining the consistent metaphoric comparison with archaeology implied by the term stratigraphy, the chapter also introduces the notion of “Sprachdenkmäler,” literally “language monuments,” as a cover term for three types of traces that such languages may leave and that can be studied even after they have disappeared: toponyms (placenames), contact-induced changes in the sound structure and grammar of the expanding language that can be attributed to imperfect language learning in the context of language shift, and vocabulary items that have been transferred to the expanding language and that tend to survive in certain semantic domains. Historical linguistics and contact linguistics have developed pieces of the analytic machinery that is necessary to study these phenomena, but these have not been put together to build a consistent framework. The chapter introduces such a framework as a general methodological contribution and anticipates how that framework will be applied to case studies concerning the Central Andes. Keywords Linguistic stratigraphy · Sprachdenkmäler · Language shift · Toponyms · Substrate · Contact-induced language change

Language shift and replacement have been a major factor in known language history, and what is known is likely just the tip of the iceberg of a much deeper and protracted history of language replacement in human prehistory. This is significant for a number of reasons: for one, we cannot assume that—as linguistic typology has only recently been coming to realize—observed presentday linguistic diversity, which is only a minute fraction of all languages that have ever existed, is representative of possible human languages writ large. And for the linguistic history of individual regions, it means that we must reckon with the existence of previous layers of linguistic diversity that are not readily accessible, but that included languages that, regionally or on larger scales, may have played a significant © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Urban, Linguistic Stratigraphy, https://doi.org/10.1007/978-3-031-42102-0_1

1

2

1 Linguistic Stratigraphy, Or How to Recover Traces of Lost Languages

role earlier. These lost languages may be related to languages that are still spoken in the same region or elsewhere, or their genealogical history may be completely unrelated to them. Language shift and replacement are not nearly as frequently theorized as the expansions and diversification of language families. Reconstructing such processes, especially when the identity of the replaced languages is not known, is fraught with methodological difficulties (cf. Ross, 2003): undocumented languages that are no longer spoken are not accessible by standard, time-tested methods of historical or contact linguistics, even though both offer the analytical seeds to attempt to make them so. Still, for various parts of the world, interest in understanding the dynamics of language expansion and language replacement in more complex, multifaceted, and realistic ways than just theorizing language expansions is increasing in recent years. In this vein, this book aims to provide a principled framework to make aspects of lost languages accessible through three distinct types of traces they may leave and that may remain visible long after they have ceased to be spoken. As elsewhere in historical linguistics, work on Indo-European linguistic prehistory can furnish some precedents and methodological orientations. In fact, it is IndoEuropeanists who have been most active in exploring contributions of now lost languages to the Indo-European newcomers whose languages ultimately replaced them. For instance, Schrijver (2007 [2001]) explores a set of bird names that show up in different European branches of Indo-European, that do not have cognates in languages further to the east, and that suggest to Schrijver a common pre-IndoEuropean source. Salmons (1992: 267) provides criteria for the identification of such “substrate” vocabulary in the context of Indo-European more generally. Stratigraphy provides an apt metaphor for the resulting distributions. In archeology, deeper strata in the record of anthropogenic remains may, other things being equal, be assumed to represent remains of earlier times. By comparing them with later ones, cultural continuities, as well as discontinuities, become visible. This book, in contrast, is about the layered stratigraphy of culture as it pertains to language. Like in archeology, underneath the present-day distribution of languages one can glimpse traces of earlier strata: On the one hand, the lexicon and grammatical structure of languages may bear the imprint of a pre-existing language that was spoken in the same area at an earlier—or, to stay in the metaphor of stratigraphy, deeper—point of time. As Lass (1997: 184–185) puts it beautifully, “[t]he palimpsest that makes up the observable surface of a language is rarely (if ever) entirely the result of its own internal history. As least in part, either superficial like lexis, or ‘deeper’ in structure will likely be the scars of encounters with other languages.” On the other hand, perhaps comparable to surface deposits in archeology, place names around the world have a tendency to be conservative in situations of language shift and often betray the former presence of a language in an area—inscribed in the landscape and the names of its landmarks and features—even when it is no longer spoken (Urban, 2018). Adopting a venerable term from the German linguistic tradition, in this book I will call the three types of remnants of disappeared languages Sprachdenkmäler (singular Sprachdenkmal), literally “language monuments.” This term is formidably

1 Linguistic Stratigraphy, Or How to Recover Traces of Lost Languages

3

compatible with the metaphor of linguistic stratigraphy in particular when one thinks of archeologically established cultural stratigraphies as the modeling analogon. Like archeological monuments are manifest witnesses of past cultures, these linguistic monuments are manifest, if sometimes hard to find, testimonies of the presence of past languages which make their presence known through them. The book is not only a methodological contribution, however. Beginning already as part of the methodological explorations of Chap. 3, I explore how Sprachdenkmäler allow us to explore aspects of the linguistic prehistory of one part of the world where language shift must have played a particularly decisive role in shaping linguistic history: the Central Andes of South America. Chapter 2 introduces the Central Andes and its linguistic landscape generally, but with particular reference to the Quechuan language family, whose expansion has given the region a treacherous impression of relative linguistic homogeneity. Chapter 3 then returns to methodological explorations of the three types of Sprachdenkmäler, treating toponyms and anthroponyms in Sect. 3.2, substrate effects in phonology and grammar in Sect. 3.3, and substrate vocabulary in Sect. 3.4. Chapters 4, 5, and 6 then explore how, and to what extent, the different types of evidence and criteria can create hypotheses as to the identity of the replaced languages in Southern Peru and Bolivia, the southern Ecuadorian highlands, and the Chachapoyas on the eastern slopes of Northern Peru, and thus to (partially) reconstruct the linguistic landscape of the Central Andes prior to the late phases of Quechuan spread. Pertinent issues associated with the situation in one of these areas, Southern Peru and Bolivia, are well-known and much discussed in the historical linguistics of the Quechuan family and the Central Andes at large. The other areas have only recently begun to be explored. It is thus opportune to begin the assessment with the most well-studied region of the three that this book will investigate, and to explore how the different types of evidence have been employed to uncover earlier pre-Quechuan language distributions there. To anticipate a major conclusion of the book, the success with which this can be done varies. I believe that this is no accident, but that it is a general characteristic of this type of work that it can be more successfully carried out in some parts of the world than in others. Still, analyses of Sprachdenkmäler with the framework proposed here should also be applicable to other regions of the Central Andes and indeed, possibly with some adaptations to make them align with the specifics of the respective language ecologies, also other parts of the world. What could be learned, and what problematic aspects remain unresolved, is summarized in the synopsis in the final Chap. 7 of this book that closes the discussion.

4

1 Linguistic Stratigraphy, Or How to Recover Traces of Lost Languages

References Lass, R. (1997). Historical linguistics and language change. Cambridge University Press. Ross, M. (2003). Diagnosing prehistoric language contact. In R. Hickey (Ed.), Motives for language change (pp. 174–198). Cambridge University Press. https://doi.org/10.1017/CBO978051148 6937.012 Salmons, J. (1992). Northwest Indo-European vocabulary and substrate phonology. In Perspectives on Indo-European language, culture and religion: Studies in honor of Edgar C. Polomé (vol. 2, pp. 265–279). Institute for the Study of Man. Schrijver, P. (2007 [2001]). Lost languages in northern Europe. In C. Carpelan, A. Parpola, & P. Koskikallio (Eds.), Early contacts between Uralic and Indo-European: Linguistic and archaeological considerations. Papers presented at an international symposium held at the Tvärminne Research Station of the University of Helsinki 8–10 January, 1999 (pp. 417–425). Suomalais-Ugrilainen Seura. Urban, M. (2018). Sprachlandschaften: Über die Rolle von Sprache in der Beziehung zwischen Mensch und Umwelt. Königshausen & Neumann.

Chapter 2

The Central Andean Linguistic Landscape Through Time and the Quechuan Language Family

Abstract This chapter provides a brief introduction to the culture history of the Central Andes, which is chequered and multifaceted. It emphasizes that there is a cultural palimpsest in which old and new elements coexist along each other, and links that to the observation that the same is true linguistically. The chapter then introduces Quechuan, the Andean language family with the largest geographical spread and the greatest number of speakers, discusses its internal classification, and traces the developments that led to its remarkable scope. The chapter shows that in many of the regions in which Quechuan is today spoken as the principal or only Indigenous language there is evidence that, sometimes until quite recently, other languages were once spoken, and that the Quechuan presence is but a relatively thin layer under which now lost diversity can be found. The chapter also discusses what is known about that non-Quechuan diversity, much of which has been further reduced in the meantime, and the state of affairs that can be reconstructed for earliest historical times. Keywords Quechuan · Language expansion · Language classification · Language geography

2.1 The Ever-Changing Cultural and Linguistic Landscapes of the Central Andes The Central Andes have entered a unique cultural trajectory toward societal complexity starting almost 5000 years ago. Since then, they have seen the rise and fall of pre-Columbian civilizations and empires again and again; they have then endured the horrors of the conquista and seen the establishment of the colonial Spanish regime; and its people are now more and more absorbed into the global economy and culture. It is no surprise that, as a result of this chequered and multifarious culture history, a truly complex layering of cultural patterns has come into being in this culturally precocious region. Some inherited cultural traditions have been, and still are, slowly replaced by more recently established practices. Yet, other old cultural traditions © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Urban, Linguistic Stratigraphy, https://doi.org/10.1007/978-3-031-42102-0_2

5

6

2 The Central Andean Linguistic Landscape Through Time …

continue to coexist with more recently established ones and, often in a transformed or reinterpreted form, survive until the present day. The ever-changing nature of this cultural fabric is visible everywhere. It can be seen in the adoption of bowler hats as part of the typical outfit of the Indigenous women of the Central Andes. It can be seen when despachos, the traditional offerings made to Andean deities, today include, alongside coca leafs and llama fat, dollar notes and store-bought candy. In immaterial culture, it can be seen when the story of Juan el Oso, which has European origins, is transformed and acquires distinctly Andean overtones and meanings (e.g., Marín-Dale, 2016; Morote Best, 1988 [1957]). But the layering of old and new in Central Andean cultural traditions is not restricted to the colonial period, and in fact is a consistent hallmark of the culture history of the region since prehistoric times. Thus, it can equally well be seen when, in spite of cataclysmic periods of natural disaster and societal collapse, elements of cultural traditions, though often transformed and imbued with new meaning, survived on the North Coast of Peru as one pre-Columbian culture succeeded another (see, e.g., Bawden, 1995; Malpass, 2016: 196; Shimada, 1994: 261 for some examples). In the southern part of the Central Andes, the iconography of the so-called Staff God is a persistent expression of religious beliefs and practices across cultural ruptures (Isbell & Knobloch, 2008; Quilter, 2012). In fact, the strong cultural continuity of certain archeologically visible expressions of culture through millennia, in spite of drastic societal changes and upheavals (see Isbell, 2018), is a striking characteristic of the Central Andes. In terms of language, arguably the most important Indigenous language family of the Andes is Quechuan. Linguistically, the differences between local varieties of Quechuan speech are not great. This indicates that the family is of relatively recent age, and that it has not been long ago that the regional variation developed from a common ancestor that must have been much more localized than the family is today. This means that Quechuan is but a shallow stratigraphic layer that overlays a deeper, and likely more diverse and regionalized, linguistic history. Indeed, we can assume that the arrival of Quechuan in some (but not all!) parts of the Central Andes probably occurred only in times of the Inca empire, which is visible in the archeological stratigraphy of the Central Andes as the so-called Late Horizon. Therefore, we have good reason to suspect that Quechuan superseded other languages that were spoken formerly in these regions and that were, either rapidly or in a more protracted process of language shift, replaced by expanding Quechuan speech. In its empirical part, the present book presents an attempt to use regional features in local Quechuan varieties and local toponymy to infer something about language distributions that were in place before the arrival of Quechuan. Given the particularly close entanglement between language and culture history that obtains in the Central Andes, by establishing the stratigraphy of the linguistic record, also the culture history of this center of pristine civilization is enriched by more facets: the stratigraphic dynamics of language expansion and shift through time is, like the succession of archeological cultures as established by archeological stratigraphy, informative of demographic and/or cultural changes that the Central Andes underwent. In particular, using this methodology, we can hope to learn more on the linguistic aspect of

2.2 Quechuan Today

7

cultural diversity in the Central Andes as it existed before the politically and culturally unifying force of Inca hegemony that correspond to the times immediately prior to the so-called Late Horizon, the Late Intermediate Period. Importantly, like the coexistence between old and new in the culture history of the Central Andes, the Quechuan case is likely only really special in that it is the most recent episode, or layer, in the linguistic palimpsest of the Central Andes, and in being a particularly large-scale one, but not otherwise remarkable qualitatively. In fact, there is evidence that not much earlier, another language family, Aymaran, spread through parts of the Central Andes and in the process absorbed the Puquina language and probably other languages of Southern Peru and the altiplano (some aspects of this are covered in Chap. 3). The analysis of Quechuan language data is, in the present context, not a goal in itself, but rather a tool that is applied in the service of elucidating the identity of the replaced languages. At its essence, it is these languages which the book is about. However, given that the analysis of Quechuan material is central to the goal, the remainder of this introductory section is dedicated to a more detailed overview of the Quechuan language family as it exists today (Sect. 2.2), its classification (Sect. 2.3), its expansion (Sect. 2.4), and how that expansion relates to the linguistic landscape of the Central Andes of the earliest colonial times (Sect. 2.5).

2.2 Quechuan Today When looking at a map of Indigenous languages like that in Fig. 2.1, it is easily verifiable that characterizations of Quechuan as the most important Indigenous language family of the Andes are well-founded. Stretching more than 3000 kms from southern Colombia to northern Argentina, it covers a more extensive area than any other Andean language family. A significant percentage of the Indigenous people of the Andes as a whole is thus Quechuan-speaking. While counts as to the number of speakers are fraught with problems and must be taken with more than just one grain of salt (Cerrón-Palomino, 2003 [1987]: 75; Adelaar, 2014: 4–5), rough estimates reach as high as seven or eight and a half million (Adelaar, 2014: 3; Adelaar & Muysken, 2004: 619–620; Cerrón-Palomino, 2003 [1987]: 76). About half of these speakers live in Peru, which, as we will see shortly, is the historic and current center of gravity of the Quechuan family; the remainder are distributed over the countries to the north and the south, with Ecuador and Bolivia having higher speaker numbers than the still more peripheral countries Colombia and Argentina (Adelaar, 2014: 4). The Quechuan lects of the department of Cuzco in Southern Peru and those of Bolivia in particular boast high number of speakers, and these closely related lects actually form one of the Indigenous languages with most speakers in South America at large. As one can see from the map in Fig. 2.1, too, in spite of the formidable geographic spread, Quechuan-speaking areas are not always continuous. Likewise in a more scattered than continuous manner, its distribution is interspersed by areas in which other,

8

2 The Central Andean Linguistic Landscape Through Time …

Fig. 2.1 Approximate present-day distribution of Indigenous languages of the Andes from Ecuador to Argentina. Quechuan is shown in dark gray, non-Quechuan languages in light gray. Based on Lewis et al. (2016), with the Quechuan and Aymara distribution in southern Peru and Bolivia refined according to Landerman (1994: maps 1 and 2). Map is for illustration only and does not fully reflect the demographic reality of Quechuan, which, through migration, also has a strong presence, e.g., in the Peruvian capital Lima and the eastern lowlands. Note also that Indigenous languages to the east of the Andes, of which there are many, are not shown on this map at all

2.2 Quechuan Today

9

unrelated, Indigenous languages are still spoken. In southern Colombia, the northernmost outpost of the Quechuan family, Inga, as the Quechua of the area is called, is separated from Ecuadorian Quichua through a gap in the Colombian–Ecuadorian border region. In fact, Inga in itself disintegrates into several small communities east of Pasto and along the courses of the Putumayo, Caquetá, and Mandur rivers (Levinsohn, 1976). At some distance, Inga-speaking areas are surrounded by areas in which other Indigenous languages are spoken: Cofán, a language isolate, to the south, and Awa Pit and Guambiano (now preferably called Nam Trik, together with the Totoró dialect) to the west and north. Awa Pit and Guambiano belong to the Barbacoan family of languages, as do Cha’palaa and Tsafiqui, which are spoken in Ecuadorian territory (more on these languages in Sect. 5.1); still further northwest, Epena Pedee, a language belonging to the Chocoan family, is found. In the highlands of Ecuador, in contrast, there is a relatively continuous zone in which Quichua, as the language is known there reflecting a sound change, is spoken, with phonological (Orr, 1978; Stark, 1985a) as well as lexical differences (Orr, 1978; Stark & Muysken, 1977) within the zone. It runs from the province of Imbabura in the north to northern Loja in the south, although also here the strength of the speaker base varies from region to region and is sometimes limited to certain communities (Stark, 1985a), a fact from which the homogeneous representation on the map should not distract. Forms of Quichua are also spoken in the Ecuadorian oriente; these are in use along the shores of the Napo and Pastaza rivers, and, following their course toward the Amazon, also in the Peruvian lowlands (Cerrón-Palomino, 2003 [1987]: 58). As in the Colombian–Ecuadorian border area, there also is a gap in the distribution of Quechuan between southern Ecuador and Northern Peru. In fact, in the North of Peru, Quechuan is again spoken only in discontinuous communities. It is customary to distinguish between Lambayeque (or Ferreñafe), Cajamarca, Chachapoyas, and San Martín (or Amazonas) Quechua, though also in this case, in reality, each of these speech areas is composed of several individual nuclei that correspond to individual communities (Landerman, 1991: 14–23). The Chachapoyas region is located at the intersection between Andes and Amazonia, and San Martín Quechua, still further east, is already spoken in a low-altitude forest environment. Apparently with Quechuan speakers who migrated still further east from there, the language also made its way to the shores of the Ucayali river (Cerrón-Palomino, 2003 [1987]: 58; Landerman, 1991: 21). Together with the extension of lowland varieties from Ecuador along the Napo and Pastaza rivers into Peru, this creates a Quechuan presence also in the Peruvian lowlands. Likewise, even more recently, the promise of economic opportunities has brought substantial numbers of Quechuan speakers also to the lowlands of Junín and Cuzco, creating a new and highly dynamic setting for linguistic interaction (Emlen, 2020). The frequent portrayal of Quechuan as a prototypical highland language family, therefore, obscures a more complex reality (see Emlen, 2017 for discussion). As one moves further south, the pattern of alternation between more dispersed enclaves and more continuous blocks in which Quechuan speech can be heard continues. In the Peruvian department of Ancash (but excluding the northernmost

10

2 The Central Andean Linguistic Landscape Through Time …

province, Pallasca), the northern border of what may be called the heartland of the Quechuan family begins. This nuclear area is also called the continuous zone (see Torero, 1968 for the term). It extends southward, always along the course of the Andean mountain chains and its fringes, into Southern Peru. Only a number of isolated small communities in the highlands of the Lima department speak another, unrelated, Indigenous language known as Jaqaru. The continuous zone extends down to the northern margins of the greater Titicaca basin, and from there further into much of highland Bolivia. In this area, Quechuan may be heard alongside another major language, Aymara (the sister language of Jaqaru), which for a long time has slowly given way to Quechuan (we will trace this process in detail in Chap. 3). Actually, on the map it appears as if Quechuan leapfrogged over an Aymara-speaking stronghold that radiates out from the southern shores of Lake Titicaca into Southern Peru and the northwestern highlands of Bolivia and Chile, to catch foot again in the southern part of the Bolivian highlands, again yielding a discontinuity within its latitudinal distribution (Stark, 1985b: 516). The real-world processes that led to this distribution were actually very likely a lot more complex than what this description of the distribution of the languages on a map may suggest. In fact, small Aymara enclaves in what otherwise are more or less continuous Quechuan-speaking areas (Landerman, 1991: 31–32) hint at the protracted and complex demographic processes underlying the present-day distribution of the languages. Another complication is that there were once three further Indigenous languages represented in the area: the isolate Puquina and the closely related Uru and Chipaya languages (see further Sect. 2.4). In fact, nested in the otherwise Aymara-speaking area of the northern Bolivian highlands, in turn sandwiched between Quechuan, is one community, Santa Ana de Chipaya, where Chipaya is still spoken. Finally, at the very southern end of the Quechuan-speaking area, the pattern shifts again from a more block-like situation as in the highlands of Ecuador and the continuous zone to a more dispersed, fragmented distribution. Indeed, a wide gap separates the southernmost Quechuan-speaking communities in Bolivia from the Argentinian province of Santiago del Estero, which also has a Quechuan-speaking population; in this area, unlike in the neighboring province of Jujuy, the language is still relatively widely spoken (Cerrón-Palomino, 2003 [1987]: 71). However, its wide distribution does not mean that Quechuan is not endangered by extinction, as Adelaar (2014) emphasizes. The language has already disappeared from many places of which we definitely know that they were Quechuan-speaking in earlier historic times (Cerrón-Palomino, 2003 [1987]: 58). In fact, as these lines are written, many regional varieties, such as that of Tarma, that of Chachapoyas, and others are facing extinction as the younger generation shifts to Spanish, fragmenting the Quechuan-speaking territory further and further even in the continuous zone (Cerrón-Palomino, 2003 [1987]: 58). Likewise, migration to urban centers like Sucre in the highlands of Bolivia (Eversole, 2005) or Lima on the coast of Peru (Marr, 2011), which acts as a magnet of migration from the entire country, is thinning out the rural speaker population and accelerating language loss as in the urban contexts the Indigenous languages are very rarely passed on. “Meanwhile,” as Adelaar (2014: 5) says, also in the home communities “the newly Hispanicized population remains

2.3 The Classification of Quechuan

11

typically Andean in culture and traditions, and the Spanish language as used by them is a variety of Andean Spanish (castellano andino) with an undeniable Quechua substratum” (substratum, or substrate, will be a central concept also for the present book; more on what it is and how to establish it in Sects. 2.4, 3.3 and 3.4). The literature on Andean Spanish and the role of Quechuan in the genesis of its features is vast (e.g., Andrade Ciudad to appear; Cerrón-Palomino, 2003; Escobar, 1978, 1994, 1997, 2000, 2011; Haboud, 1998; Zavala, 2001).

2.3 The Classification of Quechuan Quechuan is a language family, not a single language. However, the distinction between language and dialect is a particularly problematic one in this case. For large parts of its range, especially in the continuous zone, Quechuan has the characteristics of a dialect continuum with crisscrossing isoglosses that make it problematic to reify different types of speech as distinct languages or even distinct dialects. On the other hand, very divergent varieties are clearly beyond the threshold of mutual intelligibility so that for practical purposes it is best to think of Quechuan as a set of different languages with its respective dialects (Adelaar, 2014: 6). Here, I will use the neutral term “lect” when referring to particular local types of Quechuan speech to avoid the impression that I am capable to distinguish clearly between language and dialect. In the context of Quechuan more specifically, another reason to avoid the term “dialect” is that it has derogatory overtones in the specific social fabric of the Central Andes (Adelaar, 2012b: 578, 2014: 6–7).1 At the highest level of classification, it is customary to distinguish two main types of Quechuan lects, called most frequently Quechua I and Quechua II (following Torero, 1964) or sometimes Quechua B and Quechua A (following Parker, 1963). Quechua I lects are restricted to the continuous zone in Central Peru, especially its northern half. Within this area, further lects can be distinguished on the basis of isogloss lines that, however, often crisscross in geographical space. In spite of the ensuing difficulties in distinguishing distinct lects on the basis of a consistent number of features, Quechua I has been subclassified internally (see Adelaar & Muysken, 2004: 185; Cerrón-Palomino, 2003 [1987]: 229–237). At the southern margin of the Quechua I area, there are a number of lects, especially in the provinces of Cajatambo, Chancay, Huaral, and Yauyos, which show particularly high diversity within a restricted geographical area (Landerman, 1991: 24; Taylor, 1984; Shimelman, 2017). That notwithstanding, it is clear that generally Quechuan lects from Huancavelica and Huancayo southward differ rather sharply by Quechuan standards from Quechua I. These are accordingly classified as Quechua II. The genesis of the Quechua I-Quechua II divide, however, is disputed (Adelaar, 2013; Pearce & Heggarty, 2011).

1

Another solution is to use the equally neutral term “variety” (e.g., Adelaar, 2014).

12

2 The Central Andean Linguistic Landscape Through Time …

The lects of Huancavelica, Ayacucho, Cuzco, and Bolivia are all closely related and form a subgroup of Quechua II which is called Quechua IIC.2 This means that Quechua IIC alone already covers a much wider area than Quechua I. However, these lects do not exhaust the Quechua II branch. In fact, all the Quechuan lects to the north of the continuous zone are thought to belong to the Quechua II branch, too. So-called Quechua IIB lects are those spoken in Colombia, highland and lowland Ecuador, the eastern slopes of Northern Peru at Chachapoyas, and also parts of the lowlands to the east. Quechua IIA is, and always has been, problematic as a subgroup. It was set up to comprise the other two northern Peruvian varieties, i.e., those of Lambayeque and Cajamarca, as well as lects of the highlands of Lima such as those of Yauyos and that of Pacaraos. The highly discontinuous distribution of this putative branch was explained by Torero (1968) through linguistic expansions from Central to Northern Peru in prehistory. However, the historical signal that Quechua IIA lects present is mixed and not easily resolved according to a tree model of language diversification, as there is little unequivocal evidence for shared innovations that would justify the postulation of a subgroup (see Landerman, 1991: 247–253). It was exactly this problem which forced Torero (1968) to admit the great difficulties in applying a dendritic classification to Quechuan lects. Quechua IIA to the present day is poorly supported. If one accepts the validity of more general division between Quechua I and Quechua II, however, it is clear that the Quechua I varieties are a relatively local phenomenon of Central Peru, whereas the Quechua II varieties are those that spread particularly widely in the Andes, both to the north and the south of the continuous zone and the Quechua I nucleus.

2.4 The Quechuan Expansion Following the venerable principle that the area from which a language family originally expanded is typically that in which it shows the greatest internal diversity (Sapir, 1916), the homeland of the Quechuan family is often sought in Central Peru, close to where the Quechua I-Quechua II vault can be observed and where the Yauyos “microcosm” (Taylor, 1984) of high diversity is found. In sum, the lects of this area— as well as the extinct variety documented in the colonial grammar of de Santo Tomás (1560a, 1560b), which is usually though not uncontroversially identified as one once spoken on the Central coast of Lima (see Urban, 2021)—can be thought of as prefiguring much of the internal diversity of the entire family (Torero, 1975 [1970]: 243).

2

The Argentinan lects also have a strong Quechua IIC imprint, but a more detailed analysis of phonology, grammar, and lexicon show that they result from the leveling of several distinct lects from different parts of the Quechuan-speaking Andes (Adelaar, 1995).

2.4 The Quechuan Expansion

13

Torero (1975 [1970]) also makes reference to the coast of Central Peru, which in his writings is portrayed as a particularly early focal area of the Quechuan family.3 If one accepts Central Peru as the likely Where of the proto-Quechuan homeland and the nucleus from which it expanded both north- and southward through the Andes, what about the When? When the differences between Quechuan lects are impressionistically compared to those of other genealogical units in the world, the internal diversity is found to be similar to that of the Romance languages, perhaps minus the divergent French (Heggarty, 2007: 329). More reliable absolute dating of language families is not (yet) possible, which is why statements such as that of Torero (1975 [1970]: 243), who dates the split of the Quechuan family to at least eleven centuries ago, must be treated with considerable caution. The method of glottochronology, which is used to calculate such dates, has recently been modified in order to achieve more sophisticated approximations (Holman et al., 2011). For Quechuan, this refined version of the method yields a date of 1717 years before present (Holman et al., 2011, Table 7). While also this date must be treated as an approximation rather than taken at face value, it corresponds fairly well to the impressionistic comparison with the Romance languages. The split of Romance can be calibrated on the basis of the spread of vernacular Latin throughout Europe in the context of the expansion of the Roman Empire. Hence, the split of proto-Quechuan can very roughly be placed in the vicinity of the beginning of the Common Era, perhaps the first centuries CE, consistent with the refined glottochronology date. Yet, in many regions, Quechuan must have arrived much later than that.4 Fascinatingly, the spread of Quechuan into the current peripheries can be inferred to have not occurred much earlier than the Spanish invasion in the sixteenth century. The timing seems consistent with the Late Horizon of the Central Andean culture chronology, a time when the Inca empire, the largest polity ever witnessed in preColumbian South America, held sway in a territory that is strikingly similar in extent to the distribution of the Quechuan language family. The initial spread of Quechuan into Bolivia is confidently brought into the context of the Inca expansion during the Late Horizon, which means that it did not commence much earlier than the arrival of the Spaniards in the early sixteenth century (Stark, 1985b: 520). This confidence is not only based on inferences from the known expansive trajectory of the Inca empire, but the distribution of archaisms and innovations within Bolivian Quechua likewise suggests a later introduction of Quechuan in the south, consistent with an arrival of the language with the Inca from the north (Stark, 1985b: 530). The connection with the Southern Peruvian Quechua of the Cuzco region, in the heartland of the Inca, is also seen in that Cuzco and Bolivian Quechua shares some important characteristics, prominently the presence of glottalized and aspirated stops. Where there are differences between Cuzco Quechua in Peru and the Bolivian Quechua lects, they are small, and consistent with a time period of approximately 500 years since 3

Independently from one another, Itier (2013) and Urban (2021) offer critical appraisals of the role of the coast in Quechuan prehistory. 4 The breakup of proto-Quechua II is dated to only 974 years before present by Holman et al. (2011, Table 1).

14

2 The Central Andean Linguistic Landscape Through Time …

the onset of the divergence (Heggarty, 2007: 328). The emergence of Ecuadorian Quichua in the north may be that late too (Hocquenghem, 2012), but certainly it is unlikely to be much older (see Torero 1984 for the idea that Quechuan began to spread in Ecuador around the beginning of the second millennium CE as a trade language; see also discussion in Hartmann, 1979).5 Somewhat earlier is the origin of Cajamarca Quechua under the heuristic assumption of a causal relationship of its spread with the expansion of the Huari state (Adelaar, 2012a). This shows that Quechuan in all likelihood was already widespread in the Andes when the Inca rose to power. In particular, it is relatively safe to say that the entire continuous zone was already Quechuan-speaking then (Heggarty, 2007). In fact, it is a reasonable assumption that the Inca chose a form of Quechuan speech—in colonial Spanish documents often referred to with the term lengua general del Inca or similar designations—to facilitate the administration of the empire because it was already widespread (D’Altroy, 2014: 58). However, at the same time, the Inca continued to propel the spread of Quechuan into the peripheries of its present-day distribution: into Bolivia, Argentina, and Chile in the south and quite plausibly also to Ecuador in the north. From this state of affairs, Quechuan would spread even further. In fact, the arrival of Quechuan in the lowlands of Ecuador and Peru likely occurred in historic times, driven by the role which Christian missionaries assigned the language as the vehicle to bring the Christian faith to the Indigenous people of this part of the Spanish New World empire (see Zariquiey Biondi (2004) and Itier (2022) for summaries of the protracted process and the changes which the language experienced as a result). Quechuan acted as a catalyst in Nichols and Rhodes’ (2017) terms—a language that rides on the crest of the wave of another major expansion event, in this case, that of Spanish.

2.5 The Non-Quechuan Linguistic Diversity Given that Quechuan is thus a relatively recent overlay especially in the peripheral Quechua II-speaking regions, and that the regions that are now Quechua II-speaking were not void of human populations before, it is clear that the recent linguistic prehistory of the Andes must have involved several events of language replacement and shift. To stay within the vocabulary of stratigraphy, under the shallow Quechuan layer must lie an older, “substratal,” layer of non-Quechuan Indigenous languages that were originally spoken in the areas to which Quechua II expanded. As was already alluded to in Chap. 1, still for early colonial times, the linguistic landscape of the Central Andes was much more diverse than it is today. The map in Fig. 2.2 presents a rough reconstruction of that landscape. I emphasize that, where inferences have been made on earlier language distributions, these are often highly approximate and to some extent even speculative. On the other hand, for 5

According to Ostler (2006 [2005]), however, trade languages strongly tend to disappear again once the trade connections in the context of which they were used cease to exist.

2.5 The Non-Quechuan Linguistic Diversity

15

some languages or regions the necessary work to reconstruct language distributions simply has not been broached. Where I have not found literature that reconstructs the early colonial distribution of a language, I have relied on present-day language range mapping. Since language distributions have rarely if ever remained stable in colonial times, this likely introduces anachronisms on the map, but I have felt that this is preferable to not plotting a language at all. In still other cases, the question as to the linguistic landscape in early colonial times cannot be broached because even the highly limited and patchy statements from early colonial observers, e.g., in Spanish chronicles, censuses, or travel diaries, from which such distributions can be inferred elsewhere, are absent or have not yet been located. Such a lacuna is particularly virulent for the southern coast of Ecuador. For this region, Anonymous (1868 [~1605]: 286) mentions a high diversity of languages, but the names, locations, and characteristics of the languages remain essentially in the dark. Even though the map in Fig. 2.2 is meant to be maximally comprehensive, there is a particular empirical filter at work that makes it seem likely that that reconstruction is incomplete. The map shows languages which have been documented or at least mentioned in historical records, but surely even more languages existed that disappeared so early that their existence is not even mentioned in the notebooks of European explorers and administrators. As traced in detail by Cook (1981), regions like the Peruvian coast experienced rapid demographic collapse in the wake of the European invasion, and it is likely that here and/or elsewhere, linguistic diversity waned as a result before it could even be fully reflected in the written records of the Spanish colony. As one can see, even though significant gaps in our understanding of the linguistic situation remain, it is clear that the linguistic landscape of the Central Andes into which Quechuan was embedded historically was massively more diverse than it is today. At the southern edge of the distribution of the family, unrelated and poorly documented languages were spoken: Kakán (or Diaguita) and Humahuaca were once found in northwestern Argentina, with Kakán partially also spoken in the area that is now covered by Santiago del Estero Quichua. Both of these languages are essentially undocumented (though see Nardi, 1979 on the former). On the Kunza (or Atacameño) language there is some documentation, although it is principally lexical (Vaïsse et al., 1896); some grammatical patterns and morphemes can be extracted from this material, however (Adelaar and Muysken, 2004: 375–385; Peyró García, 2005). Kunza likely also had a presence in northern Argentina, though a center of its use must have been to the west in the area of Chile surrounding the Atacama Desert. As one can see, the use of Quechuan, at the point of its major extension (at least as reconstructed by Cerrón-Palomino, 2010), extended far into northern Chile and into the area where Kunza was spoken. As the map in Fig. 2.2 also shows, the same is true of two other Andean languages according to the reconstruction of Cerrón-Palomino (2010): Aymara on the one hand and Puquina on the other. In Cerrón-Palomino’s (2010) account of the early colonial and late pre-Columbian linguistic landscape of the Central Andes, a vast part of the South-Central Andes was characterized by the use of these three languages in the same geographical space. Aymara, as we have seen in Sect. 2.2, is still spoken,

16

2 The Central Andean Linguistic Landscape Through Time …

Fig. 2.2 Approximate reconstruction of the linguistic situation in the Central Andes in the early sixteenth century. Quechuan is shown in dark gray, non-Quechuan languages in light gray. Based on Paz y Miño (1961), Torero (1986, 1989, 1993), Adelaar and Muysken (2004), Cerrón-Palomino (2010), Andrade Ciudad and Bell (2016), Lewis et al. (2016), Urban (2019), and Wikipedia user Eredebhel’s map of Uru populations. Map is for illustration only. Note that Indigenous languages to the east of the Andes, of which there are many, are not shown on this map

2.5 The Non-Quechuan Linguistic Diversity

17

and there is substantial documentation of the language with a 400-year long grammatographic and lexicographic tradition starting with Bertonio (1612a, 1612b) and extending to Coler (2014). However, when compared with the present-day situation as sketched in Sect. 2.2, the distribution has drastically changed. Aymara-speaking enclaves in Quechuan-speaking areas in Bolivia and the generally smaller geographical area in which Aymara can be heard today suggest a scenario of language replacement; in fact, Cerrón-Palomino’s (2010) reconstruction enshrines an account of how the present-day distribution came about. According to that, Aymara receded gradually and unevenly against the advance of Quechuan. We will trace these developments in much greater detail in Sect. 2.4. The Puquina language, while probably extending farther south in early colonial times as in the reconstruction of Cerrón-Palomino (2010), had its nucleus from the northern shores of Lake Titicaca to the Pacific coast of Southern Peru and Northern Chile. The north, still according to Cerrón-Palomino (2010), was the exclusive domain of this language, whereas between the southern shores of Lake Titicaca and the Pacific coast it overlapped with Aymara in its distribution. Still further north, what is today Southern Peru is reconstructed as having been an area in which Quechuan and Aymara coexisted (which also is consistent with a recession of Aymara at a later point of time). Again, we will trace these developments in greater detail in Sect. 2.4. Puquina was an important language in pre-Columbian times, and still in the colony it was ranked among the lenguas generales which the Spanish administrators elected as vehicles for Christian indoctrination. The colonial and pre-Columbian importance of Puquina is unfortunately not reflected in a sizeable amount of documentation of the language. The only dedicated materials are translations of Christian texts in Oré (1607), but with the additional complication that they were translated not directly from Spanish, but via an intermediary Quechua and/or Aymara version that was prepared first (Emlen et al. to appear). Nevertheless, a rudimentary understanding of Puquina grammar can be reconstructed on the basis of this text; the current state of research is summarized in Emlen et al. (to appear). In addition, especially Rodolfo Cerrón-Palomino has taken on the task of amending the available information on Puquina grammar and lexicon through detailed investigations of ancillary sources such as toponymy and loanwords in Quechua and Aymara (see, e.g., Cerrón-Palomino, 2014, 2016). Embedded into the area of Quechua-AymaraPuquina trilingualism which Cerrón-Palomino (2010) reconstructs were two further related languages: Chipaya, then spoken in a still larger area rather than only in the town of Santa Ana de Chipaya, and its sister language Uru, which is extinct today but was spoken widely in the riverine and lacustrine environment of the altiplano on an axis linking Lake Popoó and Lake Titicaca. Chipaya is now well-documented (Cerrón-Palomino, 2006; Cerrón-Palomino & Aguirre, 2011), and the extinct Uru language can be fairly well reconstructed on the basis of the available data (Hannß, 2008). An overview of both Uru and Chipaya is in Hannß (to appear). North of the Quechuan continuous zone, in Northern Peru, once again one can infer a drastic loss of linguistic diversity when comparing the maps in Figs. 2.1 and 2.2. In fact, Northern Peru once hosted a complex mosaic of non-Quechuan languages on the coast (Tallán, Sechura, Mochica, and Quingnam), in the highlands (Culli and one or more further languages that must be inferred to account for the

18

2 The Central Andean Linguistic Landscape Through Time …

toponymic record—we will assess the situation in more detail in Sect. 2.5), and the eastern slopes (Chirino, Patagón, Copallín, Bagua and Sácata, which latter all cramped in a tiny area around the town of Jaén; to the south, a further undocumented language conventionally known as Chacha that will be of central interest in Chap. 6, and still further south Hibito and Cholón were spoken). All these languages are poorly documented (see Urban to appear), with available documentation ranging from colonial grammars that have been evaluated in terms of modern linguistic descriptive practices (Mochica and Cholón: see de la Carrera, 1644 and CerrónPalomino, 1995; Hovdhaugen, 2004; Adelaar & Muysken, 2004; Urban, 2019 for Mochica and Mata, 2007 [1748] and Alexander-Bakkerus, 2005 for Cholón), short wordlists (Tallán, Sechura, Quingnam, Culli, and the Jaén languages: see Torero, 1986, 1993; Cerrón-Palomino, 2004; and Urban, 2019 for analysis and Urban to appear for a summary account), to no primary dedicated documentation at all. With the exception of the Hibito and Cholón languages, which are commonly thought to have formed a low-level linguistic family (see Alexander-Bakkerus to appear), the languages that were once spoken in Northern Peru are not demonstrably related to other languages within the area or beyond, though there is some evidence that Sácata and Patagón were Arawakan and Cariban languages respectively, and thus belonged to language families that are widely distributed in Amazonia (see Torero, 1993 for the evidence and Urban to appear for review). The general pattern of relatively localized languages, which contrasts with the more widely distributed languages in the southern half of the Central Andes, continues as one moves further north to what is today the Ecuadorian–Colombian border area, where we find virtually undocumented languages known as Tabancale and Xoroca. The identity of the former has so far remained elusive; the limited evidence available for the latter suggests that it belonged to a small language family of the Ecuadorian and Peruvian Upper Amazon known as Jivaroan or Chicham (Torero, 1993). However, as for Sácata and Patagón, the evidence is extremely limited. In Northern Peru, languages are distributed largely according to the trisection between western lowlands, Andean highlands, and eastern slopes. Interestingly, on a very general and impressionistic level at least, this pattern seems to continue into Ecuador. While we do not possess a shred of evidence for the languages of the south coast of Ecuador, we know that they existed, and they seem to have had an identity that was distinct from the Cañari language that was spoken in the southern highlands of Ecuador (though not so much the most southern and eastern parts of Loja, where toponymy again suggests a Chicham presence according to Gómez Rendón, 2016). In fact, there is an entire chain of pre-Columbian languages that closely follows the mountainous highlands further northward: these are Puruhuay, Panzaleo, Cara, and finally Pasto in what is today the Ecuadorian–Colombian border area. For hardly any of these languages, which must still have been spoken at least in colonial times, when Quechuan was already present in the Ecuadorian Andes, is there primary documentation, though there is some evidence that suggests an affiliation with the Barbacoan language family. A more extensive review of the evidence and of the hypotheses regarding the affiliation of the non-Quechuan languages of the Ecuadorian highlands is provided in Chap. 5 , which will be concerned with the

2.5 The Non-Quechuan Linguistic Diversity

19

identity of the Cañari language in the south. At more northerly latitudes, the language geography in the lowlands to the east of the highest mountain chains becomes clearer, too. As seen in Fig. 2.1, even today, the Barbacoan languages Cha’palaa an Tsafiqui are spoken in several communities in this area, and, according to Paz y Miño (1961)’s reconstruction of the situation in the sixteenth century, on which the map in Fig. 2.2 is based, these languages were once more widely spoken still. For both Cha’palaa and Tsafiqui, there are basic dictionaries (Lindskoog & Lindskoog, 1964; Moore, 1966) and grammar sketches (Dickinson, 2002; Vittadello, 1988). In the area in which the Pasto language was spoken—the only of the non-Quechuan highland languages of Ecuador on which there is a short vocabulary, which is published in Jijón y Caamaño (1940: 197)—there in fact is a Barbacoan language still spoken today in the immediate vicinity, Awa Pit. It is described in a grammar by Curnow (1997). In fact, as discussed in greater detail in Sect. 5.1, Pasto seems to have been a very close relative of Awa Pit (Adelaar & Muysken, 2004: 393), thus clearly showing a Barbacoan presence in the highlands. This, in fact, continues to the present day in Colombian territory, where Guambiano and Totoró (now preferably called Nam Trik) continue to be spoken further to the north. There is a grammatical description and a basic dictionary of this language (Gonzales Castaño, 2014, 2019; Vásquez de Ruiz et al. 2009). Remarkably, oral traditions of the Chachi (as the speakers of Cha’palaa are called) tell of a highland origin (Floyd, 2010: 4), and there are also authors such as Pérez (1962) who, on onomastic grounds, have questioned the separate identity of the Panzaleo language and have identified it rather with Tsafiqui. That, again, would point to a more widespread Barbacoan presence in the Ecuadorian highlands or at least blur the linguistic distinctiveness of highlands and western lowlands. The impoverishment of the linguistic diversity of the Central Andes is in large part due to language shift to Spanish in colonial times, of which Quechuan lects, locally at least, have been a victim. However, it is likely that Quechuan itself already in prehistory contributed to a reduction of linguistic diversity. As Adelaar (2007: 326) observes, today “[b]oth natives and outsiders often defend the view that the Andean Indians constitute conservative, well-preserved societies, where Quechua with all its local varieties is the native language par excellence. In reality, Quechua acted as a ‘killer language’ in many areas where it is spoken today (Ecuador, Bolivia, northwestern Argentina).” The regions mentioned by Adelaar are, notably, exactly those in which the Quechuan presence is rather shallow in terms of time depth; two of these—Ecuador and Bolivia—are among the regions studied in detail in the present work.6 The speakers of the languages that were replaced by expansive Quechuan lects may have either been expelled by Quechuan-speaking newcomers, 6

Another likely fruitful ground for exploring substrate influence in the Andes is the Argentinan Quechua lect of Santiago del Estero, to which Adelaar (2007: 326) alludes when mentioning northwestern Argentina. Having heterogeneous origins even as far as the native Quechuan lexicon is concerned (Adelaar 1995; de Granda 1999), Santiago del Estero Quichua also bears hallmarks of imperfect language learning on behalf of people originally speaking a local non-Quechuan language (Adelaar & Muysken, 2004: 178; Cerrón-Palomino 2003 [1987]: 347); these include but are not limited to unusual word order regularities in which adjectives follow rather than precede the substantives they modify (Adelaar & Muysken, 2004: 380fn174). Nardi (1986), what is more, mentions that

20

2 The Central Andean Linguistic Landscape Through Time …

or, more likely, have chosen to abandon their native language in favor of Quechuan. Accordingly, the perception of Quechuan as the epitome of Andean Indigenousness is not an eternal static given, but the present widespreadness is, to the contrary, the result of a highly dynamic expansive process that occurred in prehistory and continued into colonial times. In fact, “[a]ll varieties, especially those that owe their existence to Quechuan expansion during Inca and colonial times, contain interesting remnants of contact with languages that are now extinct and virtually unknown, and that were once spoken in areas where Quechuan became dominant. Since Quechuan displaced scores of other languages at different stages of its history, each of its varieties may retain particular traces of past language contact that can be helpful for a reconstruction of the linguistic prehistory of the area” (Adelaar, 2014: 7). History is thus repeating itself: just like Quechuan lects are more and more abandoned and replaced by Spanish, imprinting it with a distinctive character in the process, so, earlier, local Quechuan lects must have resulted from language shift of speakers of local non-Quechuan languages, likewise imprinting these lects to some extent with their respective properties.

References Adelaar, W. F. H. (1995). Raices lingüísticas del quichua de Santiago del Estero. In A. V. Fernández Garay, & J. P. Viegas Barros (Eds.), Actas de las Segundas Jornadas de Lingüística Aborigen 15 a 18 de Noviembre de 1994 (pp. 25–50). Facultad de Filosofía y Letras, Universidad de Buenos Aires. Adelaar, W. F. H. (2007). The importance of toponymy, family names and historical documentation for the study of disappearing and recently extinct languages in the Andean region. In L. Wetzels (Ed.), Language endangerment and endangered languages: Linguistic and anthropological studies with special emphasis on the languages and cultures of the Andean-Amazonian border area (pp. 325–331). CNWS. Adelaar, W. F. H. (2012a). Cajamarca Quechua and the expansion of the Huari state. In P. Heggarty, & D. Beresford-Jones (Eds.), Archaeology and language in the Andes: A cross-disciplinary exploration of prehistory (pp. 197–217). Oxford University Press. Adelaar, W. F. H. (2012b). Languages of the Middle Andes in areal-typological perspective: Emphasis on Quechuan and Aymaran. In L. Campbell, & V. Grondona (Eds.), The indigenous languages of South America: A comprehensive guide (pp. 575–624). Walter de Gruyter. https://doi.org/10.1515/9783110258035.575 Adelaar, W. F. H. (2013). Quechua I y quechua II: En defensa de una distinción establecida. Revista Brasileira de Linguística Antropológica, 5(1), 45–65. https://doi.org/10.26512/rbla.v5i1.16542 Adelaar, W. F. H. (2014). Endangered languages with millions of speakers: Focus on Quechua in Peru. In P. de Crignis, M. Eibl, K. Franke, & N. Koch (Eds.), Gefährdete Sprachen/Endangered languages (pp. 1–12). Special Issue of Journal LIPP 3. https://doi.org/10.5282/journalipp/393 Adelaar, W. F. H., & Muysken, P. C. (2004). The languages of the Andes. Cambridge University Press. Alexander-Bakkerus, A. (2005). Eighteenth-century Cholón. LOT.

between two and three percent of the Santiago de Estero lexicon is inherited from local substrate languages, above all Kakán.

References

21

Alexander-Bakkerus, A. (To appear). Hibito and Cholón. In M. Urban (Ed.), The Oxford Guide to the Languages of the Central Andes. Oxford University Press. Andrade Ciudad, L. (To appear). The Andean Spanish of Southern Peru and Bolivia. In M. Urban (Ed.), The Oxford Guide to the Languages of the Central Andes. Oxford University Press. Andrade Ciudad, L., & Bell, M. G. (2016). Mapping colonial Quechua through trial interpretations in 17th-century Cajamarca. Colonial Latin American Review, 25(4), 445–464. https://doi.org/ 10.1080/10609164.2016.1281006 Anonymous. (1868 [~1605]). 2.a parte de la descripción de Guayaquil, en que se trata de la ciudad de Puerto Viejo y su distrito. In D. L. Torres de Mendoza (Ed.), Colleción de documentos inéditos, relativos al descubrimiento, conquista, y organización de las antiguas posesiones españolas de América y Oceanía, sacados de los archivos del reino, y muy especialmente del de Indias (vol. ix, pp. 276–309). Frias y Compañía. Bawden, G. (1995). The structural paradox: Moche culture as political ideology. Latin American Antiquity, 6(3), 255–273. https://doi.org/10.2307/971675 Bertonio, L. (1612a). Arte de la lengua aymara, con vna silva de phrases de la misma lengua, y su declaracion en Romance. Francisco del Canto. Bertonio, L. (1612b). Vocabulario de la lengva Aymara. Primera parte, donde por abecedario se ponen en primer lugar los vocablos de la lengua spañola para buscar los que les corresponden en la lengua aymara. Francisco del Canto. Cerrón-Palomino, R. (1995). La lengua de Naimlap (reconstrucción y obsolescencia del mochica). Fondo Editorial de la Pontificia Universidad Católica del Perú. Cerrón-Palomino, R. (2003 [1987]). Lingüística Quechua (2nd ed). Centro de Estudios Regionales Andinos ‘Bartolomé de las Casas.’ Cerrón-Palomino, R. (2004). Lenguas de la costa norte peruana. In Z. Estrada Fernández, A. V. Fernández Garay, & A. Álvarez González (Eds.), Estudios en lenguas amerindias: Homenaje a Ken L. Hale (pp. 81–105). Editorial Unison. Cerrón-Palomino, R. (2006). El chipaya o la lengua de los hombres del agua. Fondo Editorial de la Pontificia Universidad Católica del Perú. Cerrón-Palomino, R. (2010). Contactos y desplazamientos lingüísticos en los Andes centro-sureños: El puquina, el aimara y el quechua. In P. Kaulicke, R. Cerrón-Palomino, P. Heggarty, & D. Beresford-Jones (Eds.), Lenguas y sociedades en el antiguo Perú: Hacia un enfoque interdisciplinario (pp. 255–282). Special issue of Boletín de Arqueología 14. Cerrón-Palomino, R. (2014). En pos del puquina: La tercera lengua general del antiguo Perú. In M. Malvestiti, & P. Dreidemie (Eds.), III encuentro de lenguas indígenas americanas: Libro de actas (pp. 143–157). Universidad Nacional de Río Negro. Cerrón-Palomino, R. (2016). Tras las huellas de la lengua primordial de los incas: Evidencia onomastica puquina. Revista Andina, 54, 169–208. Cerrón-Palomino, R., & Aguirre, E. B. (2011). Chipaya: Léxico—etnotaxonomía. Radboud Universiteit Nijmegen, Centre for Language Studies/Fondo Editorial de la Pontificia Universidad Católica del Perú. Coler, M. (2014). A grammar of Muylaq’ Aymara: Aymara as spoken in Southern Peru. Brill. Cook, N. D. (1981). Demographic collapse: Indian Peru, 1520–1620. Cambridge University Press. Curnow, T. J. (1997). A grammar of Awa Pit (Cuaiquer): An indigenous language of south-western Colombia. [Ph.D. Dissertation]. Australian National University. D’Altroy, T. N. (2014). The Incas (2nd ed.). Wiley Blackwell. De Granda, G. (1999). El contacto lingüístico como configurador dialectal: Estudio de un caso en el área andina suramericana. Estudios Filológicos, 34, 99–119. https://doi.org/10.4067/S007117131999003400008 de la Carrera, F. (1644). El arte de la lengva yvnga de los valles del obispado de Truxillo del Peru, con vn confessonario, y todas las oraciones christianas, traducidas en la lengua, y otras cosas. Joseph Contreras. de Santo Tomás, D. (1560a). Arte de la lengua general del Peru, llamada, quichua. Francisco Fernández de Cordoua.

22

2 The Central Andean Linguistic Landscape Through Time …

de Santo Tomás, D. (1560b). Lexicon, o vocabulario de la lengua general del Perú. Francisco Fernandez de Cordoua. Dickinson, C. (2002). Complex predicates in Tsafiki. [Ph. D. Dissertation]. University of Oregon. Emlen, N. Q. (2017). Multilingualism in the Andes and Amazonia: A view from in-between. Journal of Latin American and Caribbean Anthropology, 22(3), 556–577. https://doi.org/10.1111/jlca. 12250 Emlen, N. Q. (2020). Language, coffee, and migration on an Andean-Amazonian frontier. University of Arizona Press. Escobar, A. (1978). Variaciones sociolingüísticas del castellano en el Perú. Instituto de Estudios Peruanos. Escobar, A. M. (1994). Andean Spanish and bilingual Spanish: Linguistic characteristics. In P. Cole, G. Hermon, & M. D. Martín (Eds.), Language in the Andes (pp. 51–73). Latin American Studies, University of Delaware. Escobar, A. M. (1997). Contrastive and innovative uses of the present perfect and the preterite in Spanish in contact with Quechua. Hispania, 80(4), 859–870. https://doi.org/10.2307/345107 Escobar, A. M. (2000). Contacto social y lingüístico: El español en contacto con el quechua en el Perú. Fondo Editorial de la Pontificia Universidad Católica del Perú. Escobar, A. M. (2011). Spanish in contact with Quechua. In M. D. Campos (Ed.), The Handbook of Hispanic Sociolinguistics (pp. 323–352). Blackwell. Eversole, R. (2005). Migration and resource access: View from a Quechua barrio. Migration Letters, 2(2), 93–100. https://ssrn.com/abstract=839226 Floyd, S. I. (2010). Discourse forms and social categorization in Cha’palaa. [Ph.D. Dissertation]. University of Texas at Austin. Gómez Rendón, J. (2016). Aproximaciones a la onomástica indígena del Austro lojano. Antropología: Cuadernos de Investigación, 16, 115–129. Gonzales Castaño, G. (2014). Esquisse de description de la langue Nam Trik: Langue amerindienne parlée dans les Andes Colombiennes [M.A. Thesis]. Université Lumière Lyon 2. Gonzales Castaño, G. (2019). Una gramática de la lengua namtrik de Totoró: Lengua barbacoa hablada en los Andes colombianos [Ph.D. Dissertation]. Université Lumière Lyon 2. Haboud, M. (1998). Quichua y castellano en los Andes ecuatorianos: Los efectos de un contacto prolongado. Abya-Yala. Hannß, K. (2008). Uchumataqu, the lost language of the Urus of Bolivia: A grammatical description of the language as documented between 1894 and 1952. CNWS. Hannß, K. (To appear). Uru and Chipaya. In M. Urban (Ed.), The Oxford Guide to the Languages of the Central Andes. Oxford University Press. Hartmann, R. (1979). ¿’Quechuismo preincaico’ en el Ecuador? Ibero-Amerikanisches Archiv, 5(3), 267–299. Heggarty, P. (2007). Linguistics for archaeologists: Principles, methods and the case of the Incas. Cambridge Archaeological Journal, 17(3), 311–340. https://doi.org/10.1017/S09597743070 0039X Hocquenghem, A.-M. (2012). How did Quechua reach Ecuador? In P. Heggarty, & D. BeresfordJones (Eds.), Archaeology and language in the Andes: A cross-disciplinary exploration of prehistory (pp. 345–371). Oxford University Press. Holman, E. W., Brown, C. H., Wichmann, S., Müller, A., Velupillai, V., Hammarström, H., Sauppe, S., Jung, H., Bakker, D., Brown, P., Belyaev, O., Urban, M., Mailhammer, R., List, J.-M., & Egorov, D. (2011). Automated dating of the world’s language families based on lexical similarity. Current Anthropology, 52(6), 841–875. https://doi.org/10.1086/662127 Hovdhaugen, E. (2004). Mochica. Lincom Europa. Isbell, W. H. (2018). Radical changes in the development of Andean civilization. In L. J. Seligmann, & K. S. Fine-Dare (Eds.), The Andean world (pp. 63–79). Routledge. Isbell, W. H., & Knobloch, P. J. (2008). Missing links, imaginary links: Staff God imagery in the South Andean past. In W. H. Isbell, & H. Silverman (Eds.), Andean archaeology III: North and south (pp. 307–351). Springer.

References

23

Itier, C. (2013). Las bases geográficas de la lengua vehicular del imperio inca. Bulletin de L’institut Français d’Études Andines, 42(2), 237–260. https://doi.org/10.4000/bifea.8030 Itier, C. (2022). Quechua expansion in the Inca and colonial periods. In S. S. Mufwene, & A.-M. Escobar (Eds.), The Cambridge Handbook of Contact Linguistics, vol. 2: Multilingualism in Population Structure (pp. 526–554). Cambridge University Press. https://doi.org/10.1017/978 1009105965.026 Jijón y Caamaño, J. (1940). El Ecuador interandino y occidental antes de la conquista castellana (vol. 1). Editorial Ecuatoriana. Landerman, P. N. (1991). Quechua dialects and their classification. University of California, Los Angeles. Landerman, P. N. (1994). Glottalization and aspiration in Quechua and Aymara reconsidered. In P. Cole, G. Hermon, & M. D. Martín (Eds.), Language in the Andes (pp. 332–378). University of Delaware, Latin American Studies Program. Levinsohn, S. H. (1976). The Inga language. Mouton. Lewis, M. P., Simons, G. F., & Fennig, C. D. (Eds.) (2016). Ethnologue: Languages of the World (19th ed). SIL International. Lindskoog, J. N., & Lindskoog, C. A. (1964). Vocabulario cayapa. Instituto Lingüístico de Verano/ Ministerio de Educación Pública. Malpass, M. A. (2016). Ancient people of the Andes. Cornell University Press. Marín-Dale, M. B. (2016). Decoding Andean mythology. University of Utah Press. Marr, T. (2011). ‘Ya no podemos regresar al quechua’: Modernity, identity, and language choice among migrants in urban Peru. In P. Heggarty, & A. J. Pearce (Eds.), History and language in the Andes. Palgrave Macmillan. https://doi.org/10.1057/9780230370579_10 Mata, P. de la. (2007 [1748]). Arte de la lengua cholona (A. Alexander-Bakkerus, ed.)). Iberoamericana/Vervuert. Moore, B. R. (1966). Diccionario castellano-colorado colorado-castellano. Instituto Lingüístico de Verano. Morote Best, E. (1988 [1957]). El oso raptor: estudio de un cuento del folklore peruano. In Aldeas sumergidas: Cultura popular y sociedad en los Andes, 179–239. Centro de Estudios Rurales Andinos ‘Bartolomé de las Casas.’ Nardi, R. L. J. (1979). El Kakán, lengua de los Diaguitas. Sapiens, 3, 1–33. Nardi, R. L. J. (1986). Características dialectales del quichua santiagueño. El Liberal, November 3. Nichols, J., & Rhodes, R. A. (2017). Vectors of language spread at the Central Steppe periphery: Finno-Ugric as catalyst language. In R. Iverson, & G. Kroonen (Eds.), Digging for words: Archaeolinguistic case studies from the XV Nordic Tag Conference held at the University of Copenhagen, 16–18 April 2015 (pp. 58–68). BAR. Oré, L. H. (1607). Ritvale, sev manvale Pervanvm, et forma brevis administrandi apud Indos sacrosancta Baptismi, Pœnitentiæ, Eucharistiæ, Matrimonij, & Extremæ vnctionis Sacramenta. Io. Iacobus Carlinus & Constantinus Vitalis. Orr, C. D. (1978). Dialectos quichuas del Ecuador con respecto a lectores principiantes. Instituto Lingüístico de Verano/Ministerio de Educación y Cultura del Ecuador. Ostler, N. (2006 [2005]). Empires of the word: A language history of the world. Harper Perennial. Parker, G. J. (1963). La clasificación genética de los dialectos quechuas. Revista Del Museo Nacional, 32, 241–252. Paz y Miño, L. T. (1961). Las agrupaciones y lenguas indígenas del Ecuador, en 1500 y 1959. Boletín de la Academia Nacional de Historia, 43(97), 5–17. Pearce, A. J., & Heggarty, P. (2011). ‘Mining the Data’ on the Huancayo-Huancavelica Quechua frontier. In P. Heggarty, & A. J. Pearce (Eds.), History and language in the Andes (pp. 87–109). Palgrave Macmillan. Pérez, A. R. T. (1962). Los seudo-pantsaleos. Talleres Gráficos Nacionales. Peyró García, M. (2005). Estructuras gramaticales en el glosario de la lengua atacameña (1896). LIAMES: Línguas Indígenas Americanas, 5(1), 25–42.

24

2 The Central Andean Linguistic Landscape Through Time …

Quilter, J. (2012). The Staff God: Icon and image in Andean art. In L. Sundstrom, & W. DeBoer (Eds.), Enduring motives: The archaeology of tradition and religion in Native America (pp. 131– 141). University of Alabama Press. Sapir, E. (1916). Time perspective in aboriginal American culture, a study in method. Government Printing Bureau. Shimada, I. (1994). Pampa Grande and the Mochica culture. University of Texas Press. Shimelman, A. (2017). A grammar of Yauyos Quechua. Language Science Press. Stark, L. R. (1985a). Ecuadorian highland Quechua: History and current status. In H. E. Manelis Klein, & L. R. Stark (Eds.), South American Indian languages: Retrospect and prospect (pp. 443– 480). University of Texas Press. Stark, L. R. (1985b). The Quechua language in Bolivia. In H. E. Manelis Klein, & L. R. Stark (Eds.), South American Indian languages: Retrospect and prospect (pp. 516–545). University of Texas Press. Stark, L. R., & Muysken, P. C. (1977). Diccionario español quichua-quichua español. Quito: Museos del Banco Central del Ecuador. Taylor, G. (1984). Yauyos: Un microcosmo dialectal quechua. Revista Andina, 2(1), 121–146. Torero, A. (1964). Los dialectos quechuas. Anales Científicos De La Universidad Agraria, 2(4), 446–478. Torero, A. (1968). Procedencia geográfica de los dialectos quechuas de Ferreñafe y Cajamarca. Anales Científicos De La Universidad Agraria, 3(4), 291–316. Torero, A. (1975 [1970]). Lingüística e historia de la sociedad andina. In R. A. de Matos, & R. Ravines (Eds.), Lingüística e indigenismo moderno en América (Trabajos presentados al xxxix congreso internacional de Americanistas) (vol. 5, pp. 221–259). Instituto de Estudios Peruanos. Torero, A. (1984). El comercio lejano y la difusión del quechua: El caso del Ecuador. Revista Andina, 2(2), 367–402. Torero, A. (1986). Deslindes lingüísticos en la costa norte peruana. Revista Andina, 4(2), 523–548. Torero, A. (1989). Areas toponímicas e idiomas en la sierra norte peruana: Un trabajo de recuperación lingüística. Revista Andina, 7(1), 217–257. Torero, A. (1993). Lenguas del nororiente Peruano: La hoya de Jaén en el siglo xvi. Revista Andina, 11(2), 447–472. Urban, M. (2019). Lost languages of the Peruvian North Coast. Ibero-American Institute/Gebr. Mann. Urban, M. (2021). Terminología marítima en el Lexicon, o Vocabulario de la lengua general del Perú de Domingo de Santo Tomás (1560) y posibles implicaciones para la historia de la familia lingüística quechua. Boletín de la Academia Peruana de la Lengua, 70, 13–61. https://doi.org/ 10.46744/bapl.202102.001 Urban, M. (To appear). Small and extinct languages of Northern Peru. In M. Urban (Ed.), The Oxford Guide to the Languages of the Central Andes. Oxford University Press. Vaïsse, E. F., Hoyos, F., & Echeverría i Reyes, A. (1896). Glosario de la lengua atacameña. Impresa Cervantes. Vásquez de Ruiz, B., Rojas Curieux, T., González Castaño, G. K., & Diaz Montenegro, E.. (2009). Léxico de la lengua namtrik de Totoró. Cabildo indígena pueblo Totoró. Vittadello, A. (1988). Cha’palaachi: El idioma cayapa. Pontificia Universidad Católica del Ecuador, Sede de Esmeraldas. Zariquiey Biondi, R. (2004). Fonología del Quichua del Napo: Una aproximación a su sincronía y a su historia. Boletín del Instituto Riva-Agüero, 31, 291–320. Zavala, V. (2001). Borrowing evidential functions from Quechua: The role of pues as a discourse marker in Andean Spanish. Journal of Pragmatics, 33(7), 999–1023. https://doi.org/10.1016/ S0378-2166(00)00049-7

Chapter 3

Evidence and Methods for Investigating Substratal Languages

Abstract This chapter describes the methodological approach to lost languages pursued in this book. It discusses the three types of Sprachdenkmäler such languages may have left behind. These include, first, toponyms (placenames) and anthroponyms (personal names). Toponyms in particular are firmly linked to geography by designating places, and they are for this reason often very conservative elements in processes of language shift. The framework of this book emphasizes the importance on recurrent toponymic endings which often correspond to terms for generic landforms such as ‘river,’ ‘mountain,’ ‘valley,’ etc. It argues that attempts to etymologize toponymic endings should focus on comparisons with nouns with such meanings to constrain semantic leeway. Second, Sprachdenkmäler include changes in the sound system and the grammar of the regional varieties of expanding language families that can be attributed to imperfect language learning during language shift, so-called substrate influence. The third type of legacy of lost languages are vocabulary items, which usually cluster in particular semantic domains and which can furthermore often be identified by their particular phonological or morphological structure. The crux of the approach is the integration of all three types of evidence into a single framework, which allows to check whether the properties of toponyms, vocabulary items, and substrate effects show a consistent profile that points to a single source. Keywords Toponymy · Anthroponyms · Substrate · Contact-induced language change · Language shift · Lexical borrowing · Loanwords

3.1 Introduction The basic methodological problem for linguistic stratigraphy is that the successful demonstration of Sprachdenkmäler in the form of structural changes in the expansive language, so-called substrate influence, requires that the structural and lexical properties of the substratal language are sufficiently well known. This is the case if the replaced language is still spoken, or has been reasonably well studied before becoming extinct. As Thomason (2009: 319) states clearly, “it is not possible to

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Urban, Linguistic Stratigraphy, https://doi.org/10.1007/978-3-031-42102-0_3

25

26

3 Evidence and Methods for Investigating Substratal Languages

argue successfully for contact-induced change by appealing to an unidentified longvanished substrate language.” In many of the most interesting cases, however, including those from the Central Andes that are discussed in this book, we are indeed mostly concerned with situations in which the substratal language has disappeared completely. Lass (1997) calls the attempt to identify characteristics and affiliations of undocumented substratal languages “etymologia ex nihilo.” Thomason (2009) is correct in rejecting substrate-based explanations of language change that rest entirely on hypothetical languages introduced into the argumentation. These are vacuous, and, as I state commenting on Thomason (2009) elsewhere, “appealing to a ‘substrate language’ for otherwise unaccounted facts of historical linguistic analysis in a deus ex machina-manner is methodologically as illicit as appealing to the fact that a reconstructed language was spoken deep in prehistory in order to explain some anomalous aspect of the reconstruction that violates the uniformitarian principle” (Urban, 2018b: 436). At the same time, I would contend that sensible scenarios of language shift involving extinct and essentially undocumented substratal languages can be posited under certain circumstances. This is the case when relatives of these languages, including ones so closely related so as to be practically dialects, are still spoken near the region of interest, or have been spoken long enough for some amount of documentation to be produced. In addition, contrary to what Lass’s phrase “etymologia ex nihilo” would suggest, absence of documentation of a substratal language does not mean that no data are available for analysis, as the language is likely to have left traces in the form of placenames, and possibly also in the structure and vocabulary of the replacing language (see in the Andean context Adelaar, 2007 and Solís Fonseca, 2009). Rather than “etymologia ex nihilo,” such cases thus rather involve “etymologia ex monumenta.” If (i) substrate effects in a given intrusive language—here, Quechuan lects—can plausibly be derived from structures found in these neighboring languages, if (ii) it features a certain amount of non-native vocabulary in relevant semantic domains which other related lects lack and can be etymologized through the same languages, and, importantly, (iii) if the territory which the expanding language has gained showcases a toponymic fingerprint that cannot be interpreted through it, but through the same neighboring languages through which substrate effects and substrate vocabulary can be explained, then it is not unreasonable to assume that the substratal language was one that was affiliated areally or genealogically with the surviving autochthonous languages. The strength with which such an affiliation can be claimed depends on the strength of the available evidence. Sometimes, perhaps even in the majority of scenarios, the identity of the substrate language will remain a matter of probability rather than one of certainty. In the Andes, relevant work in this vein has been carried out for the highlands of Ecuador. As will be discussed in more detail in Sect. 5.1, it is suspected that several Barbacoan languages which are entirely undocumented and known only by name are substratal to Quichua in the highlands of central and northern Ecuador. For one, the toponymy of the regions where these languages have once been spoken are characterized by river names ending in -pí ~ -bí, which is interpreted (see, e.g., Adelaar & Muysken, 2004: 394) as a reflex of proto-Barbacoan *pii ‘river, water’

3.1 Introduction

27

(Curnow & Liddicoat, 1998). As far as the lexicon is concerned, Goméz Rendón and Adelaar (2009) note a few lexical items in Imbabura Quichua that appear to have a Barbacoan rather than Quechuan source. The authors are confident that a subset of these items can be attributed to the ethnohistorically attested Cara language, which was once spoken in the Imbabura region and which is thought to have been Barbacoan. More recently, Bruil (2011) identifies an innovative construction in Ecuadorian Quichua generally that appears to be modeled on a Barbacoan template, and Muysken (2021) provides a more general panoramic overview of possible substrate influence on different Ecuadorian Quichua varieties from local non-Quechuan languages, Barbacoan and other. However, such lines of argumentation are not generalizable. Languages related to the putative substrate language must exist and a number of salient and clear parallels between lexical and structural remnants of the substratal language and toponymy must be identifiable. As the short example above already suggests, it is absolutely crucial to consider the full set of possible traces—Sprachdenkmäler—of substratal languages in conjunction. Thomason (2009) cautions that researchers aiming to demonstrate substrate influence should not focus on individual features, but to sets of features. As far as extinct and undocumented substratal languages are concerned, one can say in a similar vein that it is crucial to consider the structural inferences in the replacing languages in concert with the substrate vocabulary and even more importantly with the toponymic record. This does not mean that Sprachdenkmäler of all three types can always be identified—the relevant data may be lacking, or indeed speakers of a substratal language may have shifted to the replacing language without substrate interference and retention of a set of lexical items, scenarios which are in fact completely plausible (Thomason & Kaufman, 1988). Strong proposals as to the identity and affiliation of vanished substratal languages, however, I contend, should investigate all three types of Sprachdenkmäler. They will be able to present a substantial amount of evidence from all three types; toponymy, with its direct connection to the region in which the language shift is hypothesized to have taken place, should be assigned particular importance. However, in the scenario of “etymologia ex monumenta,” the available evidence may be simply too sparse, or it may be incongruent, in which case no strong claims are admissible. Indeed, in the case studies considered here, we will encounter a quality cline, from one case where quite good evidence for the identity of the substratal language is available to one where that identity only becomes visible more weakly.

28

3 Evidence and Methods for Investigating Substratal Languages

3.2 Toponyms and Anthroponyms The German Andeanist Middendorf (1891: 6) observed: Aus Ortsnamen lässt sich in der Regel mit ziemlicher Sicherheit auf die frühere Bevölkerung eines Landes schliessen; denn wenn diese auch verdrängt oder vernichtet worden ist, so behalten deren Nachfolger im Besitze des Landes gewöhnlich die überkommenen Namen. Auch wenn neue Namen eingeführt werden, erweist sich die Macht der Gewohnheit meistens stärker als der Wille der Eroberer. From placenames one can make rather secure inferences regarding the earlier population of a country; because even if it has been disposed or destroyed, their successors in possession of the land usually keep the superseded names. Even if new names are introduced the force of habit turns out to be stronger than the will of the conquerors in most cases.

As the general scope of Middendorf’s quote suggests, in many areas of the world, toponyms (placenames) survive processes of language shift. This is only logical. By virtue of designating names for populated places and natural landmarks, they are firmly linked to specific places (Urban, 2018a). As the process of language shift takes place in a physical environment with features that are already named and familiar to the shifting population there is, other things being equal, no incentive or need to coin new terms for these landmarks. Andean toponyms are typically conservative in this way (see further Urban, 2018a). By virtue of their conservativeness in language shift situations, placenames can preserve aspects of the phonetics and phonology of the substratal language from which they derive long after that language has ceased to be spoken. Also, clues as to the literal meaning of toponyms that were associated with them in the substratal language can be astonishingly persistent after its extinction. A particularly striking example is that an individual living in the area where the Culli language was once spoken in Northern Peru was, decades after the languages’ extinction, still able to indicate that the place name Acogoñ is related to water (Adelaar, 1988: 126)—and indeed, from the extant wordlists we know that or was the Culli word for ‘water’ (Martínez Compañón, 1985 [1782–1790]; Rivet, 1949). As Middendorf (1891) knew, the conservativeness of toponyms imbues them with special relevance for the identification of linguistic stratigraphies, i.e., the layering of substratal languages. In fact, a range of studies have been conducted in the Central Andes with that goal in mind (e.g., Adelaar, 1988; Andrade Ciudad, 2010; CerrónPalomino, 2005, 2008; Krzanowski & Szemi´nski, 1978; Torero, 1986; Urban, 2021). The central place of the study of toponymy in the investigation of linguistic stratigraphies derives from the expectation that, if a language was spoken in a region prior to the arrival of Quechuan, and if that presence was not ephemeral but substantial and long-term, it would, other things being equal, have left a legacy in the form of placenames.1 Therefore, a claim that in a given region, Quechuan replaced an earlier substratal language should always be supported by the presence of toponymy that cannot be explained through Quechuan. Claims as to the identity and affiliation 1

“Other things” may not always be equal, however. For instance, expanding states or empires may legislate policies aiming to oust substratal toponymy as part of a more general effort to erase the cultural memory of regional ethnic identities.

3.2 Toponyms and Anthroponyms

29

of that substrate language can find a basis in the identification of the non-Quechuan toponymic material with lexical or morphological material in extant related languages that were documented and that may be living relatives of the replaced language (and as detailed further below, this identification must be plausible phonologically and semantically while constraining acceptable leeway in both regards). None of this means that all placenames in that region should be explainable through a single homogeneous linguistic entity. For one, in the best of cases, toponymic studies are capable to show dialectal variation in the substratal language, even in the complete absence of relevant linguistic documentation to that effect. This is the case for Culli (Adelaar, 1988). Cerrón-Palomino (2015) places emphasis on the stratigraphic nature of the toponymic record and the fact that the etymology of placenames need not necessarily be sought in the language currently spoken in the region where they occur. However, for another, of course we must also reckon with a certain number of Quechuan and Spanish placenames that have been coined after the arrival of these languages. In the Andes, hybrid toponyms, formed by elements from the pre-existing and eventually replaced substratal language and Quechuan or even Spanish are reported for several different regions (Andrade Ciudad, 2010; Cerrón-Palomino, 2005). Where they exist, such placenames bespeak a sufficiently long period of bilingualism before completion of the language shift for them to be coined.2 And, of course, the extremely complex and at least partially noncontiguous distribution of languages in geographical space that could be observed in the early sixteenth century in the Central Andes (on which see Urban to appear) must be reckoned with. Therefore, we must expect a certain amount of heterogeneity in the toponymic record as a reflection of general patterns of language distributions in the Central Andes. The toponymic evidence for a substratal language will accordingly be found in only a subset of the placenames of the region under study. Much the same remarks are also applicable to anthroponyms (personal names). Indeed, anthroponyms have been used in the Andean context for the identification of prehistoric linguistic boundaries (Salomon & Grosboll, 1986) and, in conjunction with the toponymic record, to elucidate structures of an otherwise undocumented substratal language in the Chachapoyas area (Taylor, 1990). Anthroponyms, however, reflect the mobility of people in ways that toponyms do not. This is visible, for instance, in the anthroponymic records from the North Coast of Peru (Urban, 2015, 2019a), where anthroponyms appear to have diffused across linguistic boundaries. This property is on the one hand problematic, because anthroponyms that appear to derive from a particular language are not necessarily indicative of a sustained presence of that language in the region where the anthroponyms are attested. Furthermore, in a way that is more acute than for toponymy, the anthroponymic record may be quite 2

This situation is not a characteristic of the Andes only. In fact, a cross-linguistically common type of hybrid toponym are the so-called tautoponyms, in which a new head from a chronologically later language is added to a preexisting toponym from an earlier, possibly substratal language, which (unbeknownst to the speakers of the later language), means the same thing as the newly added head. An example is the Val d’Aran in the Pyrenees, where Occitan val ‘valley’ has been added as a new head to the toponym Aran, which in Basque, the original language of the region, also simply means ‘valley’ (Vennemann 2006: 972–975).

30

3 Evidence and Methods for Investigating Substratal Languages

heterogeneous as far as the (ultimate) source languages of the names is concerned. On the other hand, it is precisely their connection to actual people and their movements in geographical space that makes anthroponyms interesting. For instance, the anthroponymic record of the Chachapoyas region shows the presence of people with an Aymara linguistic background (Rivarola, 2007), suggesting a demographic influx from the South-Central Andes at some point of time. The etymological study of placenames in the Andean region has a long history that goes back in time way beyond Krzanowski and Szemi´nski’s (1978) toponymic study of the Chicama valley. While also some earlier authors, such as Middendorf (1891), have furnished relatively careful analyses of the evidence available to them, one cannot say that that has always been the case: too often have past and even present authors forced formally and semantically implausible etymologies onto placenames of the Andes. As a result, the field has come into some disregard (Cerrón-Palomino, 2015; Salomon & Grosboll, 1986). A typical move that is especially found in the older literature is to declare that the toponym is a “corrupted” form of a proposed etymological source, its “corruptness” being the reason for any formal differences between both. Likewise, as long as some amount of resemblance on the formal side is found, there was often little concern with semantic plausibility. Another all too frequently encountered practice was to compare toponyms of the Andes lightheartedly to lexical material from languages that are spoken sometimes as far away as Mesoamerica. While not all comparisons earlier authors have made are that far-fetched, also those that involve linguistic material from more plausible candidates frequently show an unfortunate tendency to force local toponyms in the procrustean bed of the lexical stock of Quechuan and Aymaran, neglecting the possibility that the etymology must be sought in a local language (Cerrón-Palomino, 2015: 190). To this, one can add also a lack of constraints when it comes to the selection of languages from which toponyms are explained, which leads to implausible comparisons for instance with, e.g., Mayan or, at the other extreme, a too heavy constraint that rules out all possibilities beyond Quechuan and Aymaran, the major Andean language families. The absence of reasonable constraining factors and principles in all regards leads to a situation in which, theoretically, toponymic material could either be explained through almost any language or must be explained through Quechuan and Aymaran. A convincing etymology, however, should satisfy some methodological standards that significantly reduce the possibilities of postulating etymologies in all regards: differences between the toponym to be explained and its proposed etymological form should, in the ideal case, be completely accounted for and, where they are not, this should be explicitly noted as a weakness of the proposal. It should go without saying that the morphological structure of the proposed etymology must be consistent with that of the language from which it is theorized to derive. This has, however, not always been observed. Thus, the formal weaknesses aside, von Buchwald’s (1918: 233) etymology of the place name Chanduy (a village in coastal Ecuador) as Mochica ‘high plain’ fails because it would require the modifier to follow the head, whereas Mochica, like the other languages of the Central Andes, places the modifier before the head.

3.2 Toponyms and Anthroponyms

31

In addition to these, problems of earlier toponymic analyses pertain to the formal, but especially the semantic side. Constraining semantic latitude in toponymic analysis far more strictly than what has been allowed for in early work is necessary: proposed etymologies for toponyms should yield plausible names for the places they designate. Cerrón-Palomino (2015: 186) points out that placenames everywhere in the world are typically nominal and motivated by the physical characteristics of the places they designate and their surroundings. This, as he continues, renders dubious any suggestions that the etymology of a place name should be sought in questions, exclamatives, imperatives, inflected verbs, etc.3 Ideally, the proposed etymological connection would make direct reference to the properties of the place (this is another count on which von Buchwald’s reading of Chanduy, a village located on the coast at sea level, as Mochica ‘high plain’ would fail). To the contrary, for instance, an etymology of placenames in the Northern Peruvian Andes that begin in llaca- are plausibly related to Cholón ly aka ‘red’ (Alexander-Bakkerus, 2005) in the sense that is currently at stake in that all three place names occur near areas of reddish or ochre-colored land (Urban, 2021; more on such toponyms in Sect. 6.2). Some clues on the semantics of recurrent toponymic endings that derive from substratal languages may thus come from the physical characteristics of the places they designate. However, not every time can relevant characteristics of the physical environment be identified. A conservative assumption is that recurrent endings are most likely to reflect words for landform types and human-made landmarks that are also recurrent in the same way: plains, slopes, mountains, rivers, valleys, lakes, but also towns, waystations, fields, etc. Across languages, placenames of the Central Andes are frequently made up of heads with such meanings and a modifier. To illustrate, Table 3.1 provides an overview of recurrent toponymic endings (given in the form in which they appear orthographically in placenames) that go back to common nouns for frequently occurring landmarks in four different Central Andean language groups.4 Toponyms involving a generic head that designates some kind of landform and a modifier are so widespread that we can, in more than one case, observe toponyms that, when etymologized through the locally dominant language, turn out to mean the same thing. Such is the case for Conchucos and Omasuyo, Culli and Aymara respectively

3

As Cerrón-Palomino (2015: 186) notes, there actually are some remarkably curious Spanish placenames in the Peruvian Andes such as No creo (< Span. no creo ‘I don’t believe’) or Salsipuedes (< Spanish ¡sal si puedes! ‘leave if you can’) in Loreto and Tarma respectively. 4 In addition, toponyms that consist of a combination of modifier and a landmark-designating head may be truncated. Thus, Cerrón-Palomino (2006) argues (not uncontroversially, see Faucet Pareja n.d.) that the name of the Inca capital Cuzco is an elliptical version of originally more complex versions like *qusqu rumi ‘owl rock.’ However, there are also Andean toponyms which neither feature the bipartite structure mentioned above nor are formally derived, as is the case for the village of Tambo northwest of the city of Ayacucho, which appears to take its name simply from tampu, the Quechuan designation for a waystation along the Inca road system, or the mountain Sinqa north of Cuzco, whose name is simply the Quechuan word for ‘nose,’ sinqa.

-llacta, -marca

‘town, village, hamlet’

-pampa

-pampa ~ -bamba

-urcu

‘plain’

‘mountain, rock’

-qota

-cocha

‘lake, sea’

-uma

-marca

-sirca

-raqui

-yacu

-pacha

‘water, river’

Aymara

‘land, earth’

Quechua

(Assumed) Meaning

-

-chata ~ -llata

-

-

-

-phara ~ -para ~ -huara

Puquina (Cerrón-Palomino, 2016)

-bal ~ -ball

-day

-

-quida ~ -queda ~ - guida ~ -gueda

-chuco ~ -chugo

-con ~ -coñ ~ -gon ~ goñ

Culli (Adelaar, 1988)

Table 3.1 Toponymic endings from the Central Andes that reflect common nouns for frequently occurring landmarks in four languages of the Central Andes

32 3 Evidence and Methods for Investigating Substratal Languages

3.2 Toponyms and Anthroponyms

33

for ‘land of water’ (Adelaar, 1988: 119), but also for Zalcot and Yanayacu, Cholón and Quechua respectively for ‘black water.’5 Of course, toponymic endings may also be of a grammatical nature, as for instance numerous Aymara toponyms ending in -ni, the possessive suffix of the language, indicate (see Sect. 4.2 for more on these toponyms). Another example for toponyms that are derivatives rather than compounds are those in the Northern Peruvian highlands which end in -sh. These can be identified with the nowadays nonproductive Quechuan derivational morpheme -sh (Cerrón-Palomino, 2015: 193); the name Ancash is itself a relevant token. Mochica toponyms are often of the derived type, too; one class, for instance, is formed by adding the locative marker to roots (Salas García, 2010: 107) to yield place names such as Cupisnique or Ñañanique.6 Note that toponyms may only be classifiable in such terms through etymological analysis. Lima, the name of Peru’s capital, for instance, is derived from the Quechuan root *rima- ‘to speak’ that is assumed for the local Quechuan lect of the coast (with a sound change */r/>/l/) by means of a nominalizing suffix -q, which has been lost in the Spanish pronunciation of the name (Cerrón-Palomino, 2000).7 This fact leads to another issue toponymic analysis, past and present, is faced with: constituents, and more generally, the morphological material toponyms consist of, are often short in terms of segments. This is a general issue, but one that becomes particularly visible in the case of affixes, which cross-linguistically tend to be shorter than lexical material. -ni, in fact, is a relevant case: as discussed in more detail in Sect. 4.2, the suffix is actually homophonous with an unrelated element in the Arawakan Campan languages that happens to also be frequent in toponyms. This fact greatly reduces the value of -ni as a diagnostic for an Aymaran presence. The chances for such random similarities to arise, of course, increase the shorter the morpheme in question is, and the more common the consonants and vowels it contains are. For this reason Mossel and Emlen (to appear), who pursue a novel quantitative and probabilistic approach to toponymic analysis in the Central Andes, attribute increased diagnosticity to longer as opposed to shorter forms. A parallel issue arises in the application of the comparative method to demonstrate genealogical relationships between languages, and individual proposals have been criticized for involving the comparison of short, monosyllabic items that are felt to be of reduced diagnostic value for precisely the same reasons. It is a natural reaction to wonder how many chance matches short forms, especially those with generic phonology, might generate when enough comparisons are carried out, and to question, as a result, the diagnostic value of short forms in toponymic analysis, including when carried out in the context of research on substrate languages. But, especially in a situation when we are dealing with languages that happen to prefer monosyllabic lexical elements, this is the material we have to work with. In 5

Zalcot is attested as the name of a Cholón village in an appendix to the manuscript of Pedro de la Mata’s (2007 [1748]) Cholón grammar (Alexander-Bakkerus, 2005: 35). 6 Earlier, Torero (1986: 541) thought that toponyms in -nique rather reflect Mochica ‘river’. 7 Note that the city is named after the river that is running through it, and that its name, Rímac, not only conserves the original phonology of the lexical root, but also the nominalizer.

34

3 Evidence and Methods for Investigating Substratal Languages

the most extreme imaginable case, we would have no material left to analyze if we insist on looking only at long forms. I would argue that the requirement for otherwise strict phonological correspondences that should be the basis of strong claims, when combined with semantic constraints that insist that recurrent endings should be etymologized principally through landscape generics and derivational morphology, actually limits the number of imaginable matches, also of short forms, significantly (while not entirely ruling them out in particular in the case of affixes, as the case of -ni shows). In the case of substratal languages, we do not have at our disposal any information on the derivational mechanisms that may be relevant for analyzing the structure of toponyms, nor do we have any a priori information on their lexicon. There can be special circumstances in which we can nevertheless arrive at sensible interpretations of toponymy. For instance, zhuta ‘bird’ is a lexical item that uniquely occurs in the Quichua of the southern Ecuadorian Andes, a fact that, together with its initial voiced fricative, betrays its non-Quechuan origin (more on this area in Sect. 3.5). The inherited Quechuan form pishku ‘bird’ is retained, but, as would be expected on general grounds (Kuryłowicz, 1964: 12), it survives in Cañar Quichua only with the secondary meaning ‘penis’ (Howard, 2010: 142). That zhuta ‘bird’ is among the substrate vocabulary items retained in the local Quichua lect allows to interpret Zhutabamba, a local place name, as ‘bird plain’ (Urban, 2018b: 440fn6). Similarly but differently, Adelaar (1988: 118) can interpret the Culli toponymic ending -chuco ~ chugo as meaning ‘land, earth’ on the basis of the fact that an anonymous text, written by Augustinian friars in the sixteenth century (Anonymous, 1865, 1560]), states that the Quechuan expression Pachamama ‘mother earth’ corresponds to Chucomama in the region. Also grammatical elements of substratal languages may under lucky circumstances be meaningfully interpretable. As far as derivation is concerned, some Culli toponyms have a characteristic ending -t which appears to correspond to “a nominalizing and adverbializing element -t that occurs in the Quechua lects bordering on the Culli area” (Adelaar, 1988: 119, citing Krzanowski & Szemi´nski, 1978: 42). Given the phonotactic restrictions against final /t/ in Quechua, “the -t ending is strongly suggestive of Culli influence or substratum” (Adelaar, 1988: 121). In this case, the function of the corresponding element in Quechua suggests that we may be dealing with a derivational marker also in Culli. Generally, however, we cannot decide a priori if a recurrent sequence at the beginning or end of toponyms corresponds to a lexical item, a clitic, or a suffix.8 Therefore, I will usually use the neutral term “toponymic ending” or just “ending” for recurring final sequences in toponyms. As should be evident from the above examples, all available evidence must be taken into account in the analysis of local toponymy. However, in the absence of retained substrate vocabulary (as in the case of zhuta), ancillary sources (such as the Augustinian relation), or other types of information that aid toponymic analysis, by

8

However, if the same sequence found as part of longer placenames also occurs by itself, this is evidence that we are dealing with a form that is capable of occurring on its own as a “minimum free form” in the sense of Bloomfield (1933: 178), i.e., a word.

3.2 Toponyms and Anthroponyms

35

default analyses of toponymic Sprachdenkmäler from substrate language should pay special attention to comparisons with words for landforms. In exploring toponymy practically in the present book, public domain data from the GEOnet Names Server of the US National Geospatial-Intelligence Agency (available at http://geonames.nga.mil/gns/html/) is used. Loaded into the QGIS Geographic Information System together with national and administrative boundaries of the modern Andean nation states from http://www.diva-gis.org/gdata, the toponymic database of GEOnet can be queried to identify, for instance, the distribution of placenames ending with certain final elements in orthographic representation. Using the toponymic dataset from the GEOnet Names Server offers the considerable advantage that it provides high-density toponymic data for all relevant Andean countries in a unified format that can immediately be used for analysis. There are, however, also disadvantages. For one, the dataset is based on maps compiled earlier by the Peruvian Instituto Geográfico Nacional (and used for linguistic research, e.g., in Torero, 1989) that divide the national territory into rectangular grid cells. The comprehensiveness with which toponyms are recorded in these grid cells varies somewhat and is quite dense for many, but rather sparse for others. More generally, even the most detailed grid cells that have fed into the GEOnet data are unlikely to cover the full toponymic record in minute detail (as already mentioned, in the Andes, even minor landmarks typically receive proper names). For another, the historical dimension in the development of Andean toponyms is disregarded. This concerns changes in pronunciation from the earliest attestations to the present day that may prove crucial in determining their origin. But it also pertains to the fact that toponyms still in use in earlier colonial times in a considerable number of cases have in the meantime ceased to be used and remembered, and are hence not part of the database. The query results usually require some manual postprocessing as they typically contain confounds, such as names of Spanish origin which happen to feature the relevant sequence, too, by sheer coincidence. Another reason why postprocessing is sometimes necessary is that some toponyms have doublets or even multiples. This appears to occur in two cases: on the one hand, when one and the same toponym belongs to two or more different kinds of landmarks recognized in the database, the same basic toponym can occur multiple times tightly clustered in geographical space, for instance as the name of a river and of the canyon through which that river flows (sometimes the eponymous nature of the toponyms is resolved by additional Spanish material in the database, e.g., Spanish río and caño). Another situation that makes postprocessing necessary is when, in the case of long rivers or valleys, their name appears to be repeated several times along the course of those rivers or valleys. Finally, toponyms may be doubled if a populated place and an administrative unit of the modern Andean nation states are eponymous. Postprocessing, i.e., removal of all but one of the doubled or multiplied toponyms, is carried out throughout in all these cases, unless there are doubts and really two or more distinct but eponymous toponyms could be involved.

36

3 Evidence and Methods for Investigating Substratal Languages

3.3 Substrate Influence The traditional terminology in historical and contact linguistics when studying former linguistic distributions and their lasting effect in the observable linguistic record is directly compatible with the metaphor of stratigraphy that was introduced in Sect. 2.1 in analogy to archeology. The terms “substrate” and “substratal” are commonly employed in this context. These are, however, polysemous (see also Urban, 2018b): on the one hand, they may point to a chronologically older linguistic layer in a given region of the world (here, parts of the Central Andes) that has later been superseded and indeed obliterated by the arrival of an intruding language (here, of the Quechuan family). On the other hand, there is a well-known related but different reading. Such languages may or may not have left so-called substrate influence on the chronologically more recent languages that replaced them. Substrate influence comes about when adults learn the replacing language imperfectly and their learner errors are incorporated as features of a newly emerging regional standard (Thomason & Kaufman, 1988).9 I shall use the term “substratal” exclusively with reference to pre-Quechuan languages of the Central Andes, never to their potential influence on the superseding Quechuan lects. On the other hand, I will reserve “substrate” to the possible effects of language shift from a substratal language. In this way, not only is the polysemy associated with the terminology resolved, but also unambiguous reference to the different entities and phenomena involved becomes possible. Substrate influence (Thomason & Kaufman’s, 1988 “interference through shift”), then, is a language contact effect that results from the imperfect learning of a language by adults. In the cases we will be concerned with, the learned language will be an intrusive Quechuan lect and the learners will have been the original speakers of the substratal languages. The learner errors of these adults, given the right demographic and sociolinguistic context, then conventionalize and become features of the local standard of the language they have shifted to, as their children and further generations acquire them during first language acquisition. Generalizations on substrate influences are difficult indeed. However, ever since Thomason and Kaufman (1988), it is assumed that in a prototypical scenario of substrate influence as a result of language shift, phonetics, phonology, and syntax are affected first. A methodologically sound demonstration of substrate influence requires a chain of evidence (Thomason 2009): first, it must be shown that the languages indeed were in contact. Thomason (2009: 332) suggests that this can be done by demonstrating the existence of loanwords in the replacing language (see Sect. 3.4).10 However, she cautions that “in a long-past contact situation” that led to the complete disappearance 9

Thomason and Kaufman (1988), in fact, use the term “substratum interference,” but prefer the broader if more cumbersome phrase “interference through shift” to involve a whole range of sociolinguistically different language contact situations that involve language shift (see Thomason & Kaufman, 1988: 116). 10 At the same time, Thomason (2009: 324) emphasizes that “the often-repeated claim that the absence of loanwords means the absence of contact-induced change is invalid, because shift-induced interference often occurs without significant amounts of lexical transfer.”

3.3 Substrate Influence

37

of the substratal language the lexical record may be very slim, possibly even nonexisting. Toponyms that cannot be etymologized through the replacing language could be assigned a similar role: arguably, also toponyms and anthroponyms (personal names) form part of the lexicon of the substratal language, although they belong to the proper names, a subdomain of the lexicon that is distinct from common nouns (appellatives). But toponymic evidence may, at least theoretically, be deceiving. If speakers or indeed languages (compare Mannheim’s 2018 dictum of “languages moving over the Andes”) later migrated to a different region, the relevant language contact may have taken place in a different geographical setting than that in which the replacing language is attested. Toponyms do not typically travel in this way, though settlements in newly colonized territories may receive commemorative names in deference to the home region of the settlers. In the Ecuadorian Andes, a series of places called Catequilla suggest that the cult of the Huaca Catequil of Northern Peru was brought to Ecuador during Inca rule (Topic et al., 2002: 324–328). Similarly, several Inca provincial centers, referred to as “New Cuzcos” for being modeled on the capital of Cuzco itself in layout and structure, have landmarks that seem to have been named with reference to those of Cuzco itself. For instance, both Cuzco and the “New Cuzco” of Tumipampa in Cañar territory in southern Ecuador featured two rivers that trisected the urban core. One of these was called Huatanay both at Cuzco and Tumipampa (D’Altroy, 2014: 387). For most practical purposes, however, I assume that a heterogeneous toponymic record, for which at least two stratigraphic layers corresponding to two distinct languages are identifiable, can be used as a proxy for a language contact scenario. Having established the existence of a language contact scenario, similarities between the two languages must be identified that could plausibly have spread from the substrate language to the replacing language through language shift. Thomason (2009: 322) suggests that evidence should come from at least two subsystems of language (phonology, morphology, syntax, semantics, and discourse pragmatics). “Etymologia ex nihilo” scenarios involving unknown prehistoric substratal languages, however, pose severe limitations on the feasibility of such a demonstration. Indeed, the absence of any principled documentation limits what can be known about the vanished substratal language to the extreme. Some remnants of the morphological structure of the replaced language may be preserved in the structure of toponyms and the vocabulary that survives in the replacing languages, however (see also Sect. 3.4). Furthermore, the same items may preserve phonological and phonotactic patterns that are alien to the target language. However, more in-depth accounts would again need to recur to data from presumably related, documented languages as proxies, as is done in relevant research on the Ecuadorian highlands (Bruil, 2011). The next step in the chain of argumentation according to Thomason (2009: 323– 324) would be to show that these features were present in the substratal language before it came into contact with the replacing language, and conversely, that they were not present in the replacing language before contact. Ideally, one would also be able to show that it is not likely that they would develop in the replacing language internally. As Thomason (2009: 323) notes, “[d]emonstrating that the features are

38

3 Evidence and Methods for Investigating Substratal Languages

old in [the substrate language, MU] will only be possible if [its history, MU] is well understood, at least with respect to this set of features.” In the case of undocumented substratal languages, where the relevant features are not even synchronically “well understood,” but merely hinted at by the Sprachdenkmäler that were left behind, this is impossible. Again, evidence from putative sister languages whose synchronic structure and diachronic history is known to some extent may be relevant. In this way, Bruil (2011) identifies a construction in Ecuadorian Quichua that appears to be modeled on a Barbacoan template.

3.4 Substrate Vocabulary Lexical influence, as we have seen, is not a necessary corollary of substrate influence. One might expect little to no lexical material in the replacing language especially when the shift and concomitant disappearance of the substrate language is rapid and happens within just one or a few generations. Where the lexicon does play a role, however, speakers shifting to the new language “will probably keep their own nativelanguage words only for things the [target language, MU] has no words for: foods and other cultural items, and (if the [target language, MU] speakers are invaders from elsewhere) names for local animals, plants, and so forth” (Thomason & Kaufman, 1988: 39). While influence through borrowing and shift can and should be kept conceptually distinct, “there is unfortunately no reason to expect these two types of interference to take place in mutually exclusive contexts” (Thomason & Kaufman, 1988: 45). Where the shift situation is more prolonged, and there is a phase of bilingualism for several centuries as the original substrate language is slowly replaced, it is expectable that more transfer of vocabulary can take place than in cases of rapid shift. The vocabulary that is transferred to the replacing language in situations of language shift tends to have typical features that are useful for its identification. Salmons (1992: 267), building on work by Edgar Polomé, mentions four criteria for the identification of substrate vocabulary in the context of Indo-European: (i) semantic plausibility, (ii) etymological status, (iii) distribution of the form, and (iv) phonological or morphological aberrancy. Semantically plausible according to criterion (i) would be exactly items in the semantic fields mentioned above by Thomason and Kaufman (1988: 39). For the Andes, this can be somewhat refined, in that, alongside flora, fauna, and cultural vocabulary, “words for physical characteristics and deformations and (poor) food preparation were also found to be relevant” in a previous study (Urban, 2018b: 455). Criterion (ii), etymological status, points to the need to show the absence of convincing etymologies for the items in question in the family to which the replacing languages belongs. For Salmons (1992), this is Indo-European, in the present context, this is the Quechuan family. Any possibility to derive the source of a lexical item internally immediately undermines the idea that that source must be sought outside

3.4 Substrate Vocabulary

39

of the family. This is, therefore, a clear and highly important criterion in the identification of substrate vocabulary; most, perhaps all, language families are not perfectly documented, and this surely is true for the Andean language families as well. Therefore, it can be hard to know if a particular term exists in only one place (and thus is a candidate for a substrate item), or if it is more widely attested. Therefore, to implement this criterion, it is important to cast the net very wide in the necessary family-level explorations in search of possible cognates. It is usually not enough to sample a selection of dictionaries of what are felt to be representative varieties; in fact, experience shows that the relevant evidence may be found only in few, or just one, source. Previously, in Urban (2018b), I have argued that the methodologically safest way to proceed in the identification of substrate vocabulary is by way of this criterion. The semantic criterion (i) is not unequivocal in the same way: while substrate vocabulary may cluster in some semantic domains such as fauna vocabulary and material culture, all languages have words covering these domains. As a consequence, only a small subset of relevant vocabulary will be inherited from a substratal language in the language that replaces it. Conversely, since semantics can only be a rough guideline, relevant vocabulary items from a substratal language may also be found outside these semantic domains. Criterion (iii), on the distribution of the form, can be thought of as working hand in hand with the cognacy criterion: not all lexical material whose history cannot be traced etymologically to an ancestor of the replacing language, of course, has a particular substratal source. As formulated by Salmons (1992), criterion (iii) makes particular reference to (repeated) borrowing among neighboring languages which can result in areally widespread forms (so-called Wanderwörter , see Haynie et al., 2014 for a recent treatment of the topic). If comparable forms are found in neighboring languages, in particular many neighboring languages, a substratal source for that form in the replacing language becomes difficult to postulate as garden variety processes of (repeated) borrowing may provide a more natural explanation. Thus, to make a case for a substratal origin, this possibility should, wherever possible, be excluded. Criterion (iv), finally, is concerned with phonological and morphological peculiarities as clues to the substratal origin of vocabulary items. Indeed, vocabulary incorporated into the replacing language from a substratal source might preserve aspects of the phonology and morphology of the replaced language, and we will see how, especially regarding phonology, this is also true in the cases of the Central Andes discussed here. However, such special phonological or phonotactic status can only be noted at all if the words in the substratal language contained phones that do not occur in the replacing languages or that are combined phonotactically in ways that is not normally possible in them. Items for which this is not the case could well be phonologically and phonotactically indistinguishable from native words in the replacing languages. Thus, criterion (iv), while useful, is not necessarily applicable in all cases. Criterion (ii), based on the absence of cognacy, however, in conjunction with the exclusion of wanderwort status as per criterion (iii), is workable also in the case of words that are phonologically, phototactically, morphologically and semantically inconspicuous.

40

3 Evidence and Methods for Investigating Substratal Languages

3.5 Extralinguistic Evidence Extralinguistic evidence, as we will see for the Southern Peruvian and Chachapoyas case studies in Sects. 3.4 and 3.6, can be compatible with certain scenarios of language spread and language shift more than with others. The historical record, in particular administrative and/or ecclesiastic documents from early colonial times that sometimes make reference to the distribution and nature of Indigenous languages, also may play an important role if available. While mostly concerned with the lenguas generales, occasionally also information on relatively small-scale languages can be found in such documents (see, e.g., Urban, 2019 for North Coast languages). Furthermore, distinct archeological cultures in the implied areas might be indicative of distinct regional identities and perhaps also languages. But extralinguistic evidence can be much more explicit than that. In particular, genomic studies of past and present populations in a given region can identify to what extent the Quechuan spread was or was not accompanied by migration of people from the Quechuan-speaking heartlands (e.g., Barbieri et al., 2017; Gayà-Vidal et al., 2010). As in the establishment of genealogical links between language families, however, it is methodologically best to not allow hypotheses pertaining to the linguistic history of a certain part of the Central Andes be based on nonlinguistic evidence, and to only allow linguistic evidence proper in the establishment of historical linguistic scenarios (compare Campbell & Poser, 2008: 205–206). It is rewarding, of course, to see how the linguistic evidence compares to what other disciplines have to say about the prehistory of the respective region—but in a second step, after the linguistic evidence has been investigated in its own right.

3.6 Summary For the Andes at least, three types of Sprachdenkmäler can be used to investigate substratal languages that were replaced by the later stages of the expansion of the Quechuan family. Toponymy typically conserves a strong imprint from the autochthonous language, even though the record will never be homogeneous and show influence from the substratal language (or languages) as well as the replacing language (or languages, e.g., in regions where Quechuan has been superseded by Spanish). In addition, further languages may have left their mark on the landscape as a result of the chequered history and/or complex sociocultural organization of the Central Andes. An example for such confounds would be places named for the Northern Peruvian Huaca Catequil in Ecuador (Topic et al., 2002: 324–328), in spite of the fact that the Culli language, with which the name Catequil is associated, was never spoken there. Anthroponyms can be a valuable source, too, though these are probably even more prone to diffusion that toponyms. The replacing languages, in addition, may enshrine further clues that can give an identity to the linguistic prehistory of the regions, both in their structure and in their vocabulary. Structural influence

3.6 Summary

41

from undocumented substrate languages is suggested if the replacing Quechuan lect has undergone salient changes vis-à-vis its immediate ancestor, and if these changes can be linked in a clear way to properties that can be assigned to the substratal language (on the basis of the phonological and phonotactic characteristics of nonQuechuan local toponymy and substrate vocabulary). It may also be suggested if there is strong evidence that the Quechuan lect shows structural properties that are also found in possible surviving relatives of the extinct substratal languages. As far as the lexical record is concerned, the identification of lexical items as Sprachdenkmäler from a substratal pre-Quechuan language must crucially involve the demonstration that cognates are absent in other Quechuan varieties and that we are not dealing with a Wanderwort that has been repeatedly borrowed from language to language in the relevant region.11 Semantic and phonological-phonotactic criteria can supplement these main guidelines, but they are neither necessary nor sufficient to identify substrate vocabulary. To repeat, it is crucial that the evidence is ideally evaluated in concert. This is because the evidence from the different types of Sprachdenkmäler only reveals its true strength when it is combined. As we have seen, toponymy alone has been used to establish the distribution of languages in prehistory in Northern Peru, and structural properties of Ecuadorian Quichua alone have been compared with Barbacoan languages to argue for a substrate affiliated with this group of languages in Bruil (2011). Yet, particularly strong cases are those in which the investigation of undocumented substratal languages relies on more than one type of evidence, as their triangulation can help to reduce uncertainties associated with each individual type of Sprachdenkmal. For instance, in spite of semantic criteria, the distinction between substrate vocabulary and garden variety lexical borrowing that did not involve language shift can be a difficult one to make in individual cases. And the demonstration of structural aspects of a particular Quechuan lect as substrate features is fraught with difficulties in the absence of documentation of the pre-Quechuan languages. The presence of toponymy that derives from a certain non-Quechuan group of languages in the area where that Quechuan lect is spoken makes the postulation of substrate influence on that lect from the same source considerably more acceptable. But a concerted evaluation is also desirable because at several levels the individual types of Sprachdenkmäler interlock: changes in phonological structure may begin with the introduction and nativization of lexical material from a substratal language. And, having the same putative substratal source, the phonological, phonotactic, and morphological characteristics of such substrate vocabulary should be reflected in the toponymic record (this, however, is a stipulation that explicitly needs to be confirmed empirically, not an interpretative assumption that can be taken for granted). The evidence should also be evaluated qualitatively, not just quantitatively. Each Sprachdenkmal has the potential to strongly suggest the presence of a (particular) 11

A special case is that in which similar or identical forms are present in other related languages outside the region of interest, but where these cluster together geographically in a way that suggests that they may have been acquired from the same or a closely related source language. Schrijver (2007 [2001]) reports some cases from Northern Europe. We will see in the case study on the Southern Ecuadorian highlands in Sect. 3.5 how this point is also relevant for the Andes.

42

3 Evidence and Methods for Investigating Substratal Languages

substratal language, or it may only provide weak hints. Such qualitative and quantitative differences in the evidence can have their origin in the nature of the contact situation or simply be due to the vicissitudes of the documentary situation. In any case, the confidence we can have in the veracity of hypotheses as to the identity of substratal languages is a function of the strength of the evidence. Pertinent claims can be strong where the evidence is consistent and unequivocally points to a substratal language with relatively clearly inferable characteristics, but need to be adequately cautious and modest where the available evidence does not allow to say more. Scenarios developed in this manner may receive support from extralinguistic (ethnohistorical, archeological, and genetic evidence), but should be able to stand by themselves on grounds of the linguistic evidence alone.

References Adelaar, W. F. H. (1988). Search for the Culli language. In M. Jansen, P. van der Loo, & R. Manning (Eds.), Continuity and identity in Native America: Essays in honor of Benedikt Hartmann (pp. 111–131). E.J. Brill. Adelaar, W. F. H. (2007). The importance of toponymy, family names and historical documentation for the study of disappearing and recently extinct languages in the Andean region. In L. Wetzels (Ed.), Language endangerment and endangered languages: Linguistic and anthropological studies with special emphasis on the languages and cultures of the Andean-Amazonian border area (pp. 325–331). CNWS. Adelaar, W. F. H., & Muysken, P. C. (2004). The languages of the Andes. Cambridge University Press. Alexander-Bakkerus, A. (2005). Eighteenth-century Cholón. LOT. Andrade Ciudad, L. (2010). Contactos y fronteras de lenguas en la Cajamarca prehispánica. In P. Kaulicke, R. Cerrón-Palomino, P. Heggarty, & D. Beresford-Jones (Eds.), Lenguas y sociedades en el antiguo Perú: Hacia un enfoque interdisciplinario (pp. 165–180). Special Issue of Boletín de Arqueología, 14. Anonymous. (1865 [~1560]). Relación de la religión y ritos del Perú, hecha por los primeros religiosos Agustinos que allí pasaron para la conversión de los naturales. In J. F. Pacheco, D. Francisco de Cárdenas, & D. Luis Torres de Mendoza (Eds.), Colleción de documentos inéditos relativos al descubrimiento, conquista y colonización de las posesiones españolas en América y Occeanía, sacados, en su major parte, del Real Archivo de Indias (vol. 3, pp. 5–58). Manuel B. de Quirós. Barbieri, C., Sandoval, J. R., Valqui, J., Shimelman, A., Ziemendorff, S., Schröder, R., Geppert, M., Roewer, L., Gray, R., Stoneking, M., Fujita, R., & Heggarty, P. (2017). Enclaves of genetic diversity resisted Inca impacts on population history. Scientific Reports, 7, 17411. https://doi. org/10.1038/s41598-017-17728-w Bloomfield, L. (1933). Language. George Allen & Unwin. Bruil, M. (2011). The emergence of future converbs in Imbabura Quichua: Pre-hispanic language contact as a possible explanation. In C. Hasselblatt, P. Houtzagers, & R. van Pareren (Eds.), Language contact in times of globalization (pp. 27–44). Rodopi. Campbell, L., & Poser, W. J. (2008). Language classification: History and method. Cambridge University Press. Cerrón-Palomino, R. (2000). Nota etimológica: El topónimo Lima. Lexis, 24(1), 151–162.

References

43

Cerrón-Palomino, R. (2005). La toponimia como fuente de usos gramaticales arcaicos y como registro cronológico de desplazamientos idiomáticos. Lexis, 29(1), 111–124. https://doi.org/10. 18800/lexis.200501.006 Cerrón-Palomino, R. (2006). Cuzco: La piedra donde se posó la lechuza. Historia de un nombre. Lexis, 30(1), 143–184. https://doi.org/10.18800/lexis.200601.005 Cerrón-Palomino, R. (2008). Voces del Ande: Ensayos sobre onomástica andina. Fondo Editorial de la Pontificia Universidad Católica del Perú. Cerrón-Palomino, R. (2015). Toponimia andina: Problemas y métodos. Lexis, 39(1), 183–197. https://doi.org/10.18800/lexis.201501.006 Cerrón-Palomino, R. (2016). Tras las huellas de la lengua primordial de los incas: Evidencia onomastica puquina. Revista Andina, 54, 169–208. Curnow, T. J., & Liddicoat, A. J. (1998). The Barbacoan languages of Colombia and Ecuador. Anthropological Linguistics, 40(3), 384–408. D’Altroy, T. N. (2014). The Incas (2nd ed.). Wiley Blackwell. de la Mata, P. (2007 [1748]). Arte de la lengua cholona (A. Alexander-Bakkerus, ed.). Vervuert. Faucet Pareja, C. A. (n.d.). Etimología del topónimo Cuzco. https://www.researchgate.net/public ation/304940367_Etimologia_del_toponimo_Cuzco. Accessed March 31, 2023. Gayà-Vidal, M., Dugoujon, J.-M., Esteban, E., Athanasiadis, G., Rodríguez, A., Villena, M., Vasquez, R., & Moral, P. (2010). Autosomal and X chromosome Alu insertions in Bolivian Aymaras and Quechuas: Two languages and one genetic pool. American Journal of Human Biology, 22(2), 154–162. https://doi.org/10.1002/ajhb.20967 Gómez Rendón, J., & Adelaar, W. (2009). Imbabura Quechua. In M. Haspelmath, & U. Tadmor (Eds.), Loanwords in the world’s languages: A comparative handbook (pp. 944–967). Walter de Gruyter. Haynie, H., Bowern, C., Epps, P., Hill, J., & McConvell, P. (2014). Wanderwörter in languages of the Americas and Australia. Ampersand, 1, 1–18. https://doi.org/10.1016/j.amper.2014.10.001 Howard, R. (2010). ‘Why do they steal our phonemes?’ Inventing the survival of the Cañari language (Ecuador). In E. B. Carlin, & S. van de Kerke (Eds.), Linguistics and archaeology in the Americas: The historization of language and society (pp. 123–145). Brill. Krzanowski, A., & Szemi´nski, J. (1978). La toponimia indígena en la cuenca del río Chicama (Perú). Estudios Latinoamericanos, 4, 11–51. https://doi.org/10.36447/Estudios1978.v4.art1 Kuryłowicz, J. (1964). The inflectional categories of Indo-European. Winter. Lass, R. (1997). Historical linguistics and language change. Cambridge University Press. Middendorf, E. W. (1891). Die Aimarà-Sprache. Mit einer Einleitung über die frühere Verbreitung der diese Sprache redenden Rasse und ihr Verhältnis zu den Inkas. F.A. Brockhaus. Mossel, A., & Emlen, N. Q. (To appear). Northern Aymara toponymy, revisited. International Journal of American Linguistics. Muysken, P. (2021). Substrate influence in Northern Quechua languages. In E. O. Aboh, & C. B. Vigouroux (Eds.), Variation rolls the dice: A worldwide collage in honour of Salikoko S. Mufwene (pp. 133–160). John Benjamins. https://doi.org/10.1075/coll.59.06muy Salas García, J. A. (2010). La lengua pescadora. Boletín de la Academia Peruana de la Lengua, 50, 83–128. https://doi.org/10.46744/bapl.201002.004 Salmons, J. (1992). Northwest Indo-European vocabulary and substrate phonology. In Perspectives on Indo-European language, culture and religion: Studies in honor of Edgar C. Polomé (vol. 2, pp. 265–279). Institute for the Study of Man. Schrijver, P. (2007 [2001]). Lost languages in northern Europe. In C. Carpelan, A. Parpola, & P. Koskikallio (Eds.), Early contacts between Uralic and Indo-European: Linguistic and archaeological considerations: Papers presented at an international symposium held at the Tvärminne Research Station of the University of Helsinki 8–10 January, 1999 (pp. 417–425). Suomalais-Ugrilainen Seura. Solís Fonseca, G. (2009). No todo desaparece cuando una lengua muere. Construyendo Nuestra Interculturalidad, 5(5), 1–11.

44

3 Evidence and Methods for Investigating Substratal Languages

Thomason, S. G. (2009). How to establish substratum interference. In Y. Nagano (Ed.), Issues in Tibeto-Burman historical linguistics (pp. 319–329). National Museum of Ethnology. Thomason, S. G., & Kaufman, T. (1988). Language contact, creolization, and genetic linguistics. University of California Press. Topic, J. R., Topic, T. L., & Cava, A. M. (2002). Catequil: Archaeology, ethnohistory, and ethnography of a major provincial Huaca. In W. H. Isbell, & H. Silverman (Eds.), Andean archaeology I: Variations in sociopolitical organization (pp. 303–336). Klouwer Academic/Plenum Publishers. Torero, A. (1986). Deslindes lingüísticos en la costa norte peruana. Revista Andina, 4(2), 523–548. Torero, A. (1989). Areas toponímicas e idiomas en la sierra norte peruana: Un trabajo de recuperación lingüística. Revista Andina, 7(1), 217–257. Urban, M. (2018a). Sprachlandschaften: Über die Rolle von Sprache in der Beziehung zwischen Mensch und Umwelt. Königshausen & Neumann. Urban, M. (2018b). The lexical legacy of substrate languages: A test case from the Southern Ecuadorian highlands. Transactions of the Philological Society, 116(3), 435–459. https://doi.org/10. 1111/1467-968X.12129 Urban, M. (2019). Lost languages of the Peruvian North Coast. Ibero-American Institute/Gebr. Mann. Urban, M. (2021). Cholón and the linguistic prehistory of northern Peru: Triangulating toponymy, substrate lexis, and areal typology. Linguistic Discovery, 17(1). https://doi.org/10.1349/PS1. 1537-0852.A.513 Urban, M. (To appear). Language ecologies and dynamics in the ancient Central Andes. In: M. Urban (Ed.), The Oxford Guide to the Languages of the Central Andes. Oxford University Press. Vennemann, T. (2006). ‘A satisfactory etymology has long been available:’ Notes on Vasconic names outside the Basque Country with particular reference to some British Arn- and Earnnames and to German Arnoldsweiler. In J. A. Lakarra, & J. I. Hualde (Eds.), Studies in Basque and historical linguistics in memory of R. L. Trask (pp. 969–992). Special Issue of Anuario del Seminario de Filología Vasca ‘Julio de Urquijo’, 40(1/2). https://doi.org/10.1387/asju.4446 von Buchwald, O. (1918). Migraciones sud-americanas. Boletín de la Sociedad Ecuatoriana de Estudios Históricos Americanos, 1(3), 227–236.

Chapter 4

The Aymara Presence in Southern Peru

Abstract This chapter presents the first of three case studies, applying the framework for studying lost languages described in the previous chapter to the Andes of Southern Peru. Significant amounts of work on the linguistic history of this part of the Central Andes have already been done, and the discussion can draw on these. The case of Southern Peru is thus particularly well-suited as a first application of the framework. In particular, the idea investigated here is that the Aymaran language family, still represented in the highlands of central Peru and strongly on the altiplano of Bolivia, was once spoken in a much wider region, including large parts of the region of Southern Peru that now separates the two areas where the family is represented. The toponymic fingerprint of the Aymaran language family has been shown to be strong in just that area, and contact-induced influence of Aymaran on the sound system of Cuzco and Bolivian Quechua has been argued forcefully but not entirely uncontroversially. The chapter also offers a new study of the lexical influence of Aymaran on the southern Peruvian Quechua variety of Cuzco in particular. Keywords Quechuan · Aymaran · Southern Peru · Cuzco · Bolivia · Toponymy · Anthroponyms · Substrate · Contact-induced language change · Language shift · Lexical borrowing · Loanwords

4.1 Introduction Among the regions of the Central Andes where Quechuan is a relatively late newcomer are Southern Peru and Bolivia. As mentioned in Sect. 2.3, the presence of the Quechuan family in Bolivia, where Aymara is still spoken in many places, is very likely the result of Inca imperial agency. In Southern Peru, where Aymara is also still spoken today, the presence of the Quechuan family is clearly older, but also there it is unlikely to have been entrenched for more than a few centuries before the conquest of Peru. However, there is direct evidence in ecclesiastic sources from the Third Lima Council (1582–1583) that there were Aymara-speaking communities in the Cuzco area still in the sixteenth century (Tercer Consilio Limense, 1984 [1584]: 78r), for © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Urban, Linguistic Stratigraphy, https://doi.org/10.1007/978-3-031-42102-0_4

45

46

4 The Aymara Presence in Southern Peru

instance, at Canas and Canchis; this observation is confirmed with almost identical phrasing, as Cerrón-Palomino (2000: 37) observes, by Bertonio (1612). Furthermore, the Relaciones Geográficas de Indias from the late sixteenth century, responses to standardized questionnaires on the newly acquired territories of Spain in the New World (Jiménez de la Espada, 1965), are an important source confirming the early colonial Aymara presence in Cuzco (Cerrón-Palomino, 1999, see references to further relevant documentation there). In addition, the Relaciones Geográficas mention the Chumbivilcas region in the southwest of the modern Cuzco department, and even parts of Ayacucho, as regions in which Aymara (sometimes referred to by the labels “Chunbivilcas” or “Hahua Simi”) is spoken, see de Acuña et al. (1965 [1586]), de Carabajal (1965 [1586]), de Monzón (1965 [1586]), de Monzón et al. (1965 [1956]a, b), and de Ulloa y Mogollón et al. (1965 [1586]) for the original texts. While Aymara may have been brought to Vilcashuamán by mitmaqkuna, in the borderlands between the present-day departments of Ayacucho and Apurímac and in the south and southwest of Ayacucho, Aymara appears to have been a native language alongside Quechua (Torero, 1975 [1970]: 232–236; see also Cerrón-Palomino, 1999: 142–143 and Mannheim, 1985a: 650–651 for further reports).1 Indeed, it is frequently assumed that Aymara was once much more widespread in Southern Peru (as well as Bolivia) than it is today. Some of the issues connected with this idea, in particular the question of structural influence from Aymara on Southern Peruvian and Bolivian Quechuan lects, have been a central and hotly debated topic in Andean historical linguistics. A general consensus view to the effect that such influence has taken place has emerged, but no total unanimity has been reached. Toponymic evidence has been brought into play both to support as well as to question the idea of Aymara structural influence on Southern Peruvian and Bolivian Quechua lects. In sum, two of the main types of Sprachdenkmäler relevant for the study of substratal languages in the Central Andes have already been invoked in the discussion, making this an ideal first case study from which to move on later, in Chaps. 5 and 6, to less well explored regions of the Andes. We will discuss the Southern Peruvian and Bolivian case with a particular view to the nature of the available evidence used to argue for a presence of Aymara in Southern Peru and its influence on certain Quechuan lects. At the same time, we will also discuss those aspects of the record which led a number of skeptics to dismiss them.

1

On the basis of the mere fact that Guaman Poma was from Sondondo in Ayacucho, Ferrell (1996: 415) considers that the Aymara material in Guaman Poma’s chronicle derives from “Ayacucho Aymara,” which he considers to be a variety of "Cuzco Aymara" (see also Adelaar & Muysken, 2004: 262).

4.2 Toponymy

47

4.2 Toponymy In the Quechuan-speaking areas of Southern Peru and Bolivia, Quechuan has left a clear imprint on the Andean landscape by lending names to its landmarks and the human settlements it harbors. Placenames like Vilcashuamán in the Ayacucho region, Curpahuasi in the Cuzco region, and Jatun Q’asa in the Quechuan-speaking Andes of Bolivia mostly consist of lexical items that reconstruct to proto-Quechua, but not to proto-Aymara (Emlen, 2017), showing from which idiomatic source they derive. Yet, Aymara toponyms are also found in Southern Peru, as has first been noted in basic outlines by Middendorf (1891: 7–18). Ever since then, the Aymara toponymic imprint has been frequently mentioned and is not controversial (compare Adelaar & Muysken, 2004: 262). Strikingly but not unexpectedly, it is areas like Canas and Canchis that stand out by their particularly pronounced Aymara toponymy (Adelaar & Muysken, 2004: 262). But also names of places in the very heartland of the Inca, including Ollantaytambo and the name of the Inca capital Cuzco itself, have been argued to have Aymara etymologies (see Cerrón-Palomino, 1999: 144–150, 2006a, 2006b, 2015: 46 for further discussion). But Aymara toponyms extend well beyond the Cuzco region in Southern Peru and are, for instance, also found in Arequipa (Adelaar & Muysken, 2004: 262). As far as the Quechuan-speaking parts of Bolivia are concerned, CerrónPalomino (2005) amply demonstrates Aymara toponyms in the departments of Cochabamba and Chuquisaca. As we have seen, Aymara’s presence was still strong there in early colonial times, and what little reference to the language that can be extracted from historical documents suggests that Aymara was spoken there at least until the end of the sixteenth century (Cerrón-Palomino, 2005: 117).2 The toponymic record actually offers evidence to confirm this: there are toponyms that are derived from Spanish nouns by means of the Aymara possessive suffix -ni in these regions (Cerrón-Palomino, 2005: 120), a fact that is incompatible with an extinction of Aymara in the implicated regions prior to Spanish conquest. Aymara toponymy, in fact, is found all over the Southern Peruvian and Bolivian Andes. That examples such as the above are not just isolated anecdotal cases, but only the tip of the iceberg of a massive Aymara toponymic layer can be shown by tracing precisely the Aymara possessive suffix -ni in the toponyms of the Southern Peruvian and Bolivian Andes in the GEOnet toponymic database. Using the example of -ni, Fig. 4.1 shows the ample presence of Aymara toponymy in the Southern Peruvian Andes and the Bolivian altiplano (as well as northernmost Chile). The suffix -ni at the same time also provides an example for the caution that must be exercised when interpreting toponymic maps without accompanying further analysis: the tight cluster of tokens to the northwest of Cuzco is actually very unlikely to be derived from Aymara but instead reflects the presence of the Kampan Arawak 2

Aymara’s ceding of territory to Quechuan is a process that is continuing into modern times (see Harris, 1975 for some contemporary perspectives). Conversely, however, Stark (1985: 535–536) reports Quechua slowly giving way to Aymara rather than the other way around in parts of Oruro.

48

4 The Aymara Presence in Southern Peru

languages that are spoken in this region. These languages feature a homophonous attributive suffix -ni ‘watercourse’ (Mossel & Emlen submitted). As far as Aymara is concerned, at any rate, the toponymic evidence is not restricted to -ni. In fact, toponyms including other Aymara endings such as -ri basically have an identical distribution. Furthermore, it is not just the distribution of toponymic endings, but also the material that precedes them that can in many cases be directly traced to an Aymara source. The suffix -ni is rather used here pars-pro-toto for the entire, much larger, toponymic evidence that points to a presence of the Aymara language in Southern Peru; it was selected among much more Aymara material because it allows to link the discussion here directly to that of Cerrón-Palomino (2005), who chose the same ending as an example for tracing the disappearance of Aymara in areas of Bolivia that are now Quechuan-speaking, and because of the methodological note the homophony with Kampan -ni affords. The toponymic record is thus singularly clear and, even taken by itself, strongly suggestive of an Aymara presence in much of Quechua IIC-speaking Southern Peru and Bolivia (as can be seen in Fig. 2.1, actually, Aymara toponyms are also found still further to the Northwest, see Mossel and Emlen submitted for discussion of this fact). What is more, as Middendorf (1891: 18) had already observed, the toponymic evidence also suggests that the Aymara presence in the Southern Peruvian Andes was significant and strong—the massiveness of the evidence is hard to reconcile with isolated, discontinuous Aymara-speaking islands. The inferred homeland of Quechuan in Central Peru (see Sect. 2.4), the toponymic evidence, and colonial refer-

Fig. 4.1 Toponyms in -ni in Southern Peru and Bolivia

4.3 Substrate Effects

49

ences to the Aymara language in places that are non Quechuan-speaking suggests that Aymara was replaced by Quechuan gradually, starting with the westernmost regions of Southern Peru where Ayacucho Quechua is spoken, i.e., Huancavelica and northwestern Ayacucho.3 For this area, we have no historical evidence for an Aymara presence in the colonial records, only a strong toponymic fingerprint. For the neighboring areas of southern Peru such as the Cuzco region, Aymara-speaking communities are still mentioned in records from early colonial times. Still further to the southeast, in Bolivian territory, Aymara is still present in many places today, and where it has been replaced by Quechuan, as Cerrón-Palomino (2005) has convincingly shown by analysis of the toponymic evidence, this happened at times in which Spanish already was firmly established in the area, i.e., later in colonial times.

4.3 Substrate Effects As just seen, the toponymic record is unusually clear, so clear, in fact, that on its basis alone an Aymara presence in Southern Peru, Bolivia, and beyond has been taken for granted. Still, let us proceed to ask about structural influence in Southern Peruvian Quechuan varieties in the areas where the toponymic record indicates a prior Aymara presence. Indeed, changes in the semantics and functions of elements of the grammatical apparatus of Cuzco Quechua proper have been explained by Aymara influence. Cuzco Quechua features special verb morphology to indicate so-called associated motion, i.e., to express that an event involves a concomitant motion (see Guillaume, 2016 for the areal context in Western South America). Itier (2011: 74) argues that the semantics of the Quechuan associated motion suffix -pu has been altered in Cuzco Quechua as it took on the meaning of the completive suffix -xa of Aymara, which, according to the author, is “an important substrate of Cuzco Quechua.” Similarly, Muysken (2015: 279) considers plausible the idea that the use of the distal demonstrative chay as a subordinator is modeled on Aymara’s distal demonstrative uka, which behaves in the same way. Genealogically, all Southern Peruvian and Bolivian Quechua lects pertain to the Quechua IIC branch and have strong grammatical and lexical similarities. Key sound changes shared by Quechua IIC lects are the merger of the two affricates of protoQuechua and the shift of */∫/ to /s/ (Torero, 2002: 59). It is common practice to distinguish two main varieties within Quechua IIC. The lects of Huancavelica, Ayacucho, and western Apurímac may jointly be referred to as Ayacucho Quechua (compare Landerman, 1994: 332–333fn1), a practice that, however, should not gloss over the fact that these are not entirely homogeneous internally. Ayacucho Quechua as a whole contrasts in some salient regards with the 3

Ferrell (1996: 415, 421) claims that “Ayacucho Aymara” was spoken in Huancavelica in historical times, too, but the argumentation rests on the distribution of the variant inya for illa ‘fertitity amulet’ rather than direct historical evidence.

50

4 The Aymara Presence in Southern Peru

varieties of eastern Apurímac, Cuzco, Arequipa, Puno, and Moquega in Peru as well as those of Bolivia (Landerman, 1994: 332). These features are also central to the claim of Aymara influence on Southern Peruvian and Bolivian Quechua. Following earlier authors (e.g., Adelaar & Muyksen, 2004: 195), I shall refer to this second group of lects as “Cuzco-Bolivian.” When the Cuzco lect in a narrow sense is meant, I shall speak of “Cuzco Quechua proper.” One feature of Cuzco-Bolivian lects which sets them apart from Ayacucho is the tendency for stops to undergo lenition. In Ayacucho, this only happens for the uvular stop in all positions of the word (Adelaar & Muysken, 2004: 201; Parker, 1969b: 168– 169, 172). In contrast, in Cuzco-Bolivian Quechua, the phenomenon can in principle affect all stops, but only in coda position. The extent to which it occurs differs somewhat between individual lects (Landerman, 1998: 36–37); Cuzco Quechua proper is described as exhibiting this phenomenon generally (Mannheim, 1991: 218).4 Coda lenition is a recent innovation. Its onset can be traced to different times in different regions of Southern Peru, but the testimony of Gonçalez Holguin’s (1608) dictionary, in which the phenomenon is not yet systematically reflected, suggests that it dates to the historic period (Mannheim, 1991: 210). Mannheim (1991) argues that the introduction of nonmodal consonants into Cuzco-Bolivian Quechua is the trigger for coda lenition. Indeed, the main feature which distinguishes Cuzco-Bolivian lects as a whole from those of Ayacucho is the presence of aspirated and glottalized segments in the former, but not in the latter (see e.g., Mannheim, 1985b: 491; in the following I shall use “nonmodal” as a cover term for both glottalized and aspirated consonants, following Coler, 2014). This feature, as we will see, plays a central role in arguments in favor of Aymara influence on Cuzco-Bolivian Quechuan lects. However, just like the linguistic distinction between Ayacucho and CuzcoBolivian lects is grounded in linguistic reality and useful for practical purposes, but glosses over a more complex linguistic reality, so the dividing line between Cuzco-Bolivian and Ayacucho Quechua in Southern Peru itself is not neat. In Southern Peru, mainly in the Apurímac region, but also in the southeast of Ayacucho (Landerman’s, 1998 “third area”), lects show properties that makes their position ambiguous between Ayacucho and Cuzco-Bolivian. In the province of Aymaraes, for one, there are nonmodal consonants, but there is no coda stop lenition (Landerman, 1998: 35; Mannheim, 1985b: 496). The same is true of the lects of the neighboring ayacuchan province of Paucar del Sara Sara (Landerman, 1998: 35). The varieties of Abancay, to the northeast of Aymaraes and thus closer to the city of Cuzco, are reported to feature aspirated, but not glottalized consonants (Cerrón-Palomino, 2003 [1987]: 183; Mannheim, 1985b: 496). In sum, in Southern Peru, two distinctive areas of Quechua speech, labeled here Ayacucho and Cuzco-Bolivian Quechua, flow into another with regard to their phonological characteristics. It is a common idea that, alongside the changes in individual morphemes sketched above, a main Aymara influence on Cuzco-Bolivian lects was the introduction of 4

See Mannheim (1985b: 494–496) and Landerman (1998: 36–41) for further phonological differences between Ayacucho Quechua and Cuzco-Bolivian Quechua.

4.3 Substrate Effects

51

Table 4.1 Ayacucho Quechua phoneme inventory, adapted from Mannheim (1991, Table 1.2) Labial

Alveolar

Palatal

Velar

Plosives

P

t

ch

k

Nasals

m

n

ñ

s

Fricatives

Laryngeal

q R

Tap/Flap l

Laterals Approximants

Uvular

ll

w

y

h

Table 4.2 Cuzco Quechua proper phoneme inventory, adapted from Mannheim (1991, Table 1.3). Only distinctive segments reproduced Labial

Alveolar

Palatal

Velar

Uvular

Plosives

P

t

ch

k

q

Glottalized plosives

p’

t’

ch’

k’

q’

Aspirated plosives

ph

th

chh

kh

qh

Nasals

m

n

ñ

s

Fricatives Laterals Approximants

h R

Tap/Flap l W

ll y

nonmodal consonants. Such a scenario has been posited with relative ease early on (Torero, 1964: 463–464, who explicitly speaks of substrate influence) and been widely though not universally accepted as a consensus view (see also Emlen, 2017: 324–325, to appear for review).5 I will first present the line of reasoning underlying the assumption of an Aymara contact effect. Ayacucho Quechua has a fairly small phoneme inventory, as is typical of most Quechua varieties. It can be seen in Table 4.1. The inventory of Cuzco Quechua, in contrast, is enriched by a series of glottalized and aspirated consonants (Table 4.2). The lects of Apurímac, in the intermediate “third area” between Ayacucho and Cuzco-Bolivia-type lects, show an interesting behavior regarding nonmodal consonants. As we have seen, these lects can be considered as transitional between 5

In most recent times, however, evidence is mounting that something may be wrong with this contact-based scenario, because it may be insufficient to explain all of the complexities of these nonmodal consonants in the history of the lineages. In particular, it is not as clear anymore as it used to appear that ancestral proto-Quechua did not possess such segments (e.g., Floyd, 2021; Emlen to appear). These views are not necessarily completely incompatible, as there may indeed have been Aymaran influence on Cuzco-Bolivian Quechua that reinforced the entrenchment of these segments there while they have been mostly lost elsewhere by now.

52

4 The Aymara Presence in Southern Peru

Ayacucho and Cuzco-Bolivian Quechua, in that nonmodal consonants are more weakly established than in Cuzco-Bolivian Quechua. Figure 4.2 reproduces the distribution of the characteristic Aymara toponyms in -ni from Fig. 4.1, but this time in the context of a map that shows the areas where Ayacucho-type, Cuzco-Bolivian-type, and intermediate lects are spoken. The map also shows the present-day distribution of Aymara. Simply by contemplating the Cuzco-Bolivian system and comparing it with the Aymara consonant inventory, it is clear that it resembles the latter much more strongly than its close Quechuan relatives of Ayacucho (Table 4.3). The story of nonmodal consonants in Cuzco-Bolivian Quechua, however, is much more complex and cannot be grasped by simply inspecting consonant inventories. For one, the distribution of ejectives and aspirates across the lexicon differs from region to region (although the extent to which this is the case may have been overstated to some extent, Landerman, 1994: 355). Thus, pairs like Cuzco phusuqu versus Bolivian Quechua pusuqu ‘foam,’ which differ by the type of consonant that appears in the onset of the first syllable, can be found alongside pairs where the nature of the nonmodal segment varies, as in Cuzco wathiya versus Bolivian wat’iya ‘earth oven’ (Mannheim, 1991: 178). But even within Cuzco Quechua proper, there is variation, as in k’ullpi ~ khullpi ‘pieces of material used for fuel’ and llusk’a ~ lluskha ‘slippery’ (Mannheim, 1991: 178).

Fig. 4.2 Aymara toponyms in -ni and the distribution of Quechua IIC and Aymara varieties at present, following Landerman (1994: 335–336, maps 1 and 2)

4.3 Substrate Effects

53

Table 4.3 Southern Peruvian Aymara consonant inventory adapted from Coler (2014: 33) Labial

Alveolar

Palatal

Velar

Uvular

Plosives

P

t

ch

k

q

Glottalized plosives

p’

t’

ch’

k’

q’

Aspirated plosives

ph

th

chh

kh

qh

Nasals

m

n

ñ h

χ

s

Fricatives

R

Tap/Flap Laterals Approximants

l w

ll y

In Cuzco Quechua proper, furthermore, there are strong restrictions on the occurrence of nonmodal consonants. They can only occur in roots, not in suffixes, and only once per root, on the first eligible stop in onset position. Given these restrictions, “once it is known that a particular root contains a modified [i.e., nonmodal, MU] segment, where it actually appears in that root is entirely predictable” (Landerman, 1994: 343). In other words, laryngealization can be considered “a feature of the root as a whole and not just of the particular segment where it occurs…” (Landerman, 1994: 343). In the Quechua of Puno, closer to the Bolivian border where Aymara was spoken until more recently, nonmodal consonants have been introduced to the morphological apparatus, too. Adelaar (1987) notes in particular the borrowing of a substantial number of verbal suffixes from Aymara which feature nonmodal consonants. As a result, occurrence restrictions on nonmodal consonants in Puno Quechua are weaker than in Cuzco Quechua proper. On the basis of the evidence discussed so far, a widespread assumption (compare Emlen, 2017: 324–325) is that Cuzco-Bolivian Quechua lects acquired nonmodal consonants through contact with one or more languages/lects of the Aymaran family.6 There are some additional arguments that make such a solution appear attractive. Within the Quechuan language family, nonmodal consonants are found in only two regions: alongside Cuzco Quechua with its series of nonmodal consonants, only Ecuadorian Quichua has aspirated, though not glottalized consonants. Ecuadorian Quichua has been argued to stand in a special contact relationship to Cuzco Quechua (perhaps related to the resettlement of imperial Inca bureaucrats from the Cuzco heartland to Ecuador) and therefore not to be an independent witness (Torero, 1964: 465–466). This position is even adopted by Landerman (1994: 340), a vocal critic of

6

While the characteristics of Cuzco Aymara are not a priori clear, Cerrón-Palomino (1999) concludes from the meager available evidence that it was close to Aymara as it is still spoken today, and that they are best considered dialects of one another. Ferrell (1996), to the contrary, shows that there are some significant differences to Aymara and affinities with present-day Jaqaru in the Aymara material of Guaman Poma. Note, however, that Ferrell considers this material to represent an “Ayacucho Aymara”.

54

4 The Aymara Presence in Southern Peru

the contact-based explanation.7 Thus, positing nonmodal consonants for the protoQuechua level, though argued for at least tentatively by Landerman (1994) and Campbell (1995) as per the orthodoxy of the comparative method, requires the costly assumption that they were lost independently at multiple nodes of the evolution of the Quechuan family (Torero, 2002: 152): first, if Cuzco-Bolivian nonmodal consonants are a conservative inherited trait, they need to have become extinct once as Quechua I evolved from proto-Quechua, they must have been lost a second time as Quechua II, which by definition must have had nonmodal consonants, diversified into the IIB and IIC branches, and then, thirdly, it must have been lost within Quechua IIC in Ayacucho Quechua.8 Furthermore, the nonsystematic occurrences of nonmodal consonants within and across Cuzco-Bolivian lects suggest a non-genetic origin. At the same time, so proponents of the contact-based explanation reason (e.g., Mannheim, 1991: 178–179), it is exactly Cuzco Quechua that borders on a region in which Aymara, a language where nonmodal consonants are deeply rooted, is spoken. The restrictions that nonmodal consonants have in Cuzco Quechua proper (less so, as we have seen, in the Quechua of Puno), which do not apply to Aymara, are often seen as evidence for their relative recency in Cuzco-Bolivian Quechua (Landerman, 1994: 337), too. Indeed, one source of nonmodal consonants in Cuzco-Bolivian Quechua is loanwords from Aymara (Mannheim, 1991: 180). But matters are more complicated, among other things, because nonmodal consonants also show up in vocabulary items which are inherited from proto-Quechua. Only a fraction of nonmodal consonants, however, has a possible internal source through the reinterpretation of palatality (Mannheim, 1991: 180). Such processes are insufficient to explain all tokens. Already Torero (1964: 464) noted that the nonmodal segments in Cuzco Quechua are imbued with certain iconic values. Thus, speakers of Cuzco Quechua proper insisted that q’asuy and p’anay ‘to beat somebody’ have ejectives because the club makes an abrupt noise when beating, whereas in qhasuy ‘to tear, to rip’, there is an aspirated sound because the tearing event makes a longer rustling noise. Mannheim (1991: 184–204) develops a model that relates to Torero’s observation on the iconic values of nonmodal consonants to account not only for the spread of nonmodal consonants to native Quechuan vocabulary, but also the unsystematic correspondences that sometimes occur. Mannheim argues that aside from cases of “primary sound symbolism” as in the observations made by Torero (1964), there is also “secondary sound symbolism,” which he also calls “associative lexical influence.” 7

Though see Cerrón-Palomino (2000: 324). There is some evidence that Ayacucho Quechua might have once featured nonmodal segments. Relevant is a non-inherited initial /h/ that is also present in vowel-initial roots with nonmodal segments in Cuzco Quechua (Campbell, 1995: 164–165). However, the distribution of initial /h/ within and across lects seems far from regular: one finds /h/-initial forms in Quechua that lack any nonmodal consonants in Ayacucho (Landerman, 1994: 344fn15, 1998: 38–39), and there are also inconsistencies in the distribution of initial /h/ across different Bolivian lects (Landerman 1998: 39). See also Cerrón-Palomino (2003 [1987]: 184–186) for discussion of the matter and see furthermore Proulx (1972, 1974) and now Cerrón-Palomino (2000: 320–321) for limited data from Quechua I, where aspiration in Cuzco-Bolivian appears reflected as vowel length.

8

4.3 Substrate Effects

55

In primary sound symbolism, ejectives act as markers of abruptly and forcefully executed actions and aspirates signal events that produce more prolonged, fricativelike noises, prominently events that involve the “expulsion of air” (Mannheim, 1991: 187), and among these especially bodily functions like belching, spitting, farting, and sneezing (Mannheim, 1991: 188, Fig. 8.3). In secondary sound symbolism, or “associative lexical influence,” laryngealization spreads from lexical items that fall in the relatively narrow categories involving primary sound symbolism to semantically related words, and thus through semantic domains across the entire lexicon. Mannheim (1991: 189) emphasizes that “[a]n important aspect of such associative lexical sets is the fluidity and variability of their membership,” and it is precisely this property that could have given rise to the inconsistencies in the distribution of the nonmodal segments in Cuzco-Bolivian Quechua. As a model, Mannheim’s (1991) account can thus explain how nonmodal consonants were introduced into Cuzco-Bolivian Quechua varieties and, imbued with semiotic functions, how they could have spread across the lexicon (even though, as critics like Landerman, 1994: 374 note, the relevant semiotic networks to explain all instances of nonmodal consonants have not been worked out). Mannheim’s (1991) model is also consistent with the restriction of nonmodal consonants to lexical roots and their ban from grammatical suffixes (with the exception of the borrowed suffixes in Puno Quechua, as we have seen): In Mannheim’s model, nonmodal consonants, through primary or secondary sound symbolism, act as an iconic device that signals aspects of the lexical semantics of roots. Given this function to create an image acoustique (de Saussure’s, 1916 alternative term for significant that seems particularly appropriate here) that embeds them into broader lexico-semantic networks, it is logical that nonmodal segments should indeed be restricted to roots. In another line of thinking which is consistent with this idea, proponents of the contact-based explanation have also argued that the distributional restrictions of Quechuan nonmodal consonants vis-à-vis the (supposedly) freer occurrence in Aymara are evidence for their recent introduction to Quechuan. Landerman (1994: 343–348), however, has shown that in Aymara, the distributional restrictions are actually quite similar. It is very typical for the first eligible segment in a word to be nonmodal. Items in which the second eligible segment is nonmodal, while the first one is not, are doubtful (see discussion in Landerman, 1994: 346). Furthermore, while more than one nonmodal segment can occur per root in Aymara, it is typically the case that the second one is a mere copy of the first, as in t’ant’a ‘bread’ or thantha ‘rag’ (Landerman, 1994: 346). In sum, the restrictions on nonmodal consonants in Cuzco-Bolivian Quechua and Aymara are actually more similar than it would appear at first glance.9 Landerman (1994) is correct when saying that the similarity of the

9

Jaqaru appears to behave somewhat differently with respect to the distribution of nonmodal segments (Torero, 2002: 151).

56

4 The Aymara Presence in Southern Peru

restrictions undermines the idea that they bespeak their recent introduction to CuzcoBolivian Quechua.10 However, it would be possible, and even logical, to at the same time argue that the striking similarity in the occurrence restrictions between Aymara and Cuzco-Bolivian Quechua that emerge after closer scrutiny strengthen the idea of a contact explanation.11 As alluded to, different lines of evidence are now coming forward that suggest a reevaluation of this scenario and that may require a reconsideration of the contactbased account of nonmodal consonants in Cuzco-Bolivian Quechua. These include but are not limited to the observation that nonmodal consonants were present in the earliest Quechua documented in the sixteenth century by Domingo de Santo Tomás (Floyd, 2021), and that they characterize words in proto-Aymara that on other grounds appear to have been borrowed from an early representative of the Quechuan lineage (Emlen to appear). These points touch upon much larger issues in Andean historical linguistics so they are not fully explored here. In addition to these, however, there is one major issue that remains and that is pertinent to the question of Aymaran substrate influence on Southern Peruvian Quechua. We have seen how the entire Southern Peruvian Andes where Quechua IIC is spoken is laced with Aymara toponymy. However, under the assumption of a contact-based scenario, only the Cuzco-Bolivian varieties acquired nonmodal consonants, while those of Ayacucho failed to do so. Now, as Landerman (1994: 343) observes, the toponymic record “argues for the presence of Aymara throughout most of the southern area, which only serves to weaken further the correlation between the occurrence of modified occlusives and Aymara influence.” In other words, if the Ayacucho area was once Aymara-speaking too (as also scattered colonial references suggest), why did Ayacucho Quechua not adopt nonmodal consonants, but Cuzco-Bolivian Quechua did (a question also raised in a different context by Heggarty & Beresford-Jones, 2012: 420)? There indeed is a mismatch between the geography of Aymara toponymy and the geographical distribution of nonmodal consonants within Quechua IIC. This is shown in Fig. 4.2. The map immediately underscores Landerman’s point: as far as the presence of Aymara toponymy in -ni is concerned, the areas where nonmodal-less Ayacucho lects and nonmodal-featuring Cuzco-Bolivian lects are spoken are indistinguishable from one another. On grounds of this, Landerman (1994) is skeptical of the contact-based explanation for the presence of nonmodal consonants in Cuzco-Bolivian Quechua, preferring to at least tentatively reconstruct nonmodal segments to the proto-Quechua level

10

Against the idea that the distributional restrictions of nonmodal consonants in Quechuan show that they are a recent introduction, Campbell (1995: 167) furthermore suggests that borrowed features are usually subject to less rather than more strong distributional restrictions in the receiving language. 11 One thing that would need to be ruled out, however, is that nonmodal consonants are subject to similar restrictions cross-linguistically. Grassmann’s law, a sound change in the history of IndoEuropean that eliminates aspirates in adjacent syllables, for instance, likewise produces occurrence restrictions on nonmodal consonants, and similar processes have been observed in the Americas for Ofo and Salishan languages (de Reuse, 1981; Thompson & Thompson, 1985).

4.3 Substrate Effects

57

(see also Campbell, 1995 and Cerrón-Palomino, 2000: 316–324).12 The nonmodal consonants in Aymara would then not be their source, but merely act as a “preservative agent” that can help explain why the varieties closest to Aymara did not lose them (Campbell, 1995: 165). Yet, the mismatch with the toponymic record would remain: if it were accepted that nonmodal segments need to be reconstructed for proto-Quechua, something which most Quechuanists would prefer not to do (compare Adelaar & Muysken, 2004: 195), and if it were accepted that Aymara could have exerted influence that promoted the preservation of these segments, why were nonmodal segments only preserved in Cuzco-Bolivian Quechua, but not in Ayacucho, where the toponymic record clearly betrays the presence of Aymara speakers, too? Torero (2002: 154) draws attention to the fact that the involved lects were not uniform, and that we must rather assume a scenario in which a diverse set of different Quechuan lects interacted with different Aymara varieties in different regions— indeed, the Third Lima Council has adopted a kind of Aymara for its purposes that would be most generally understandable to all speakers, including those that speak it “corruptedly” (see Cerrón-Palomino, 1997: 200–203). This can only be an indication of diatopic variation in Aymara, even though Cerrón-Palomino (1999) concludes that the Aymara lects of present-day Southern Peru and Bolivia and the extinct ones of Cuzco stood in a close relationship to one another. Yet, unfortunately, we do not know much about the properties of these lects, and hence the—entirely reasonable—assumption of internal variation within the Aymara family could be used to support both ideas: maybe the Aymara once spoken in the prehistoric Ayacucho region did not feature nonmodal segments, or the nonmodal consonants had properties that precluded any influence on neighboring Quechua lects. Whence an Ayacucho Quechua without nonmodal segments. Or alternatively, if one prefers a scenario in which nonmodal consonants hark back to the proto-Quechua level, a hypothetical nonmodal-less Aymara language in the Ayacucho area would not have been able to exert the preservative influence on Quechuan nonmodal consonants (see Emlen, 2017: 329–332 for the possible lack of nonmodal consonants in an early ancestor of Aymara). Thus, the reasonable assumption of diatopic variation adduced by Torero (2002: 154) does not help to decide between the two solutions in the absence of clarity on the characteristics of the extinct lects of Aymara in the Ayacucho area. As we have seen, the toponymic record can be interpreted temporally by combining assumptions about the location of the proto-Quechuan homeland, the distribution of Aymara toponymy, and the historical record of Aymara-speaking communities: expanding from a homeland in Central Peru, Quechuan would first have reached Ayacucho, then at a later point of time Cuzco and the other Southern Peruvian provinces, and only then penetrated even further south into Bolivian territory (the more conservative nature of northern Bolivian Quechua lects vis-à-vis southern ones which Stark, 1985: 530–532 notes is consistent with an expansion from north to south). Given the toponymic record, this expansion happened at the 12

However, even if nonmodal segments would have to be reconstructed, this would merely mean that the proto-language consonant inventories are similar in featuring three stop series, not necessarily that the language families are related (Campbell, 1995: 167; Mannheim, 1985a: 664).

58

4 The Aymara Presence in Southern Peru

expense of extinct Aymara lects that were, as Torero (2002: 154) reminds us, not necessarily identical to those still attested today but clearly identifiable as Aymara on grounds of the toponymic record nevertheless. The strength of Aymara influence on the expanding Quechua IIC varieties under the contact-based scenario is strikingly parallel to the inferred temporal dynamics of the Quechuan spread: Aymara influence becomes gradually stronger the further and the later in time Quechuan penetrated into originally Aymara-speaking areas. It is weakest for Ayacucho, stronger for Cuzco, where Aymara-speaking communities are attested ethnohistorically, and even stronger for Puno, where the shift may be assumed to be centuries more recent, well in the historical period. Heggarty and Beresford-Jones (2012: 421) make much the same point with the nonmodal segments in mind, but the particularly strong influence of Aymara on the morphological apparatus of the Quechua of Puno (Adelaar, 1987) points in the same direction independently. The records from toponymy and substrate influence are consistent with Aymara influence that accrues bit by bit the further Quechuan penetrates to the southwest. Under this interpretation, the lack of nonmodal consonants in Ayacucho Quechua fits into the overall picture very well. This overall view is not unique, although the linguistic variables involved and the sociolinguistic parameters, of course, are. Nevertheless, on the basis of data from Filppula (1986), Heine and Kuteva (2005: 64) are able to infer highly similar spatiotemporally layered patterns of Irish influence on the syntax on Irish English, which become gradually stronger as one moves from Dublin city, an area from which Irish has disappeared particularly early, to the city’s southern outskirts, where it ceased to be spoken around 150 years ago, to the west of Ireland, where English is under direct recent influence from Irish and where Irish influence in English syntax is strongest.

4.4 Substrate Vocabulary In contrast with the toponymic record and structural changes in Cuzco-Bolivian Quechua as discussed in Sects. 4.2 and 4.3, the actual lexical record of CuzcoBolivian Quechua has not been invoked systematically in the discussion of a substratal Aymara language in Southern Peruvian regions such as Cuzco. Nevertheless, Cerrón-Palomino (1999) finds a number of Aymara-derived lexical items in Gonçales Holguín’s (1608) Quechua dictionary from the early sixteenth century; these, however, are not necessarily indicative of the language in sixteenth century Cuzco, as they may also have entered Quechuan as loanwords from neighboring Aymara-speaking regions. More relevant are the lyrics of a hymn sung by the ladies of the Inca court at Cuzco on the occasion of the celebration of Tupac Inca Yupanqui’s victory over the Sora. These lyrics as analyzed by Cerrón-Palomino (1999: 150– 151) are clearly Aymara. While showing some typical Aymara phenomena such as vowel elisions, they betray some slight deviations from Aymara as it is known from Bertonio’s (1612) colonial work and suggest it was as a slightly deviant lect of the Aymara language (Cerrón-Palomino, 1999: 151–153).

4.4 Substrate Vocabulary Table 4.4 Borrowed Aymara suffixes in proto-Quechua IIC, from Parker (1969b: 169) with adapted orthography

59

proto-Quechua IIC Jaqaru

Aymara

*-cha ‘diminutive’ -cha ‘restrictive’ -t’a ‘diminutive’ -*ysi ‘benefactive’ -ishi ‘reflexive’

-si ‘reflexive-reciprocal’

-*puni ‘definitely’

–

-pini ‘always, certainly’

*-ra ‘un-, dis-’

–

*-ra ‘un-, dis-’

Perhaps one of the reasons why the lexical record has not been explored more so far is that speakers of Cuzco-Bolivian Quechua and Aymara remain in contact to the present day, a fact that makes it difficult to distinguish items borrowed in the context of that contact from items that could have been incorporated into the lexical stock of Cuzco-Bolivian Quechua in the context of prehistoric and/or historic language shift. At any rate, scholars have pointed their attention more to derivational and inflectional morphology. Indeed, Cerrón-Palomino (1999: 144), too, identifies bound Aymara morphology in phrases that are contained in Bertonio’s dictionary. Parker (1969b: 169), in addition, reconstructs for Quechua IIC “four suffixes which are shared with Aymara or with Jaqaru and Aymara, and which I consider borrowed from […; an Aymara, MU] substratum.” These suffixes are listed in Table 4.4 with present-day Jaqaru and Aymara forms for comparison. Another complication when investigating lexical transfer in this particular case is posed by the peculiar Southern Peruvian linguistic ecology. As discussed in Sect. 3.4, criteria for the identification of substrate vocabulary as per Salmons (1992) include (i) semantic plausibility, (ii) etymological status, (iii) distribution of the form, and (iv) phonological or morphological aberrancy. As for criterion (iv), there are phonological and phonotactic characteristics that can unambiguously betray the Aymara origin of some lexical items (Adelaar, 1986; Emlen, 2017). But this, taken by itself, does not yet say anything about the contact situation. Aymara lexical material could easily have entered Cuzco-Bolivian Quechua through contact without shift, too. The plausibility of this scenario, together with the possibility that nonmodal consonants in Cuzco-Bolivian Quechua spread to native material, annuls one of the most important of Salmon’s (1992) criteria for detecting substrate criteria, that of phonological aberrancy. A further methodological consequence of the still ongoing contact situation is that lack of Quechua-internal etymologies (criterion ii) is not diagnostic in itself: relevant items, too, could easily be the result of ongoing contact. What can be done, however, is to investigate, with criterion (i) in mind, in what semantic domains Aymara-derived items in Cuzco-Bolivian Quechua cluster (if they cluster at all) and if something on the nature of the contact scenario that gave rise to these borrowings can be learned from this. This requires, first, to actually compile a list of items with Aymara origin in Cuzco-Bolivian Quechua. To make a firm case for Aymara origin, following mutatis mutandis earlier work of a similar kind on other parts of the Central Andes (Urban, 2018, 2021), such items can be identified by applying a set of interlocking criteria: (i) they should not be widely distributed (excluding Wanderwort status), and, of course, (ii) they should have convincing Aymara etymologies (i.e., there should be a

60

4 The Aymara Presence in Southern Peru

cognate in Aymara’s sister language Jaqaru that suggests that the item was inherited from a common ancestor), and at the same time, (iii) they should not reconstruct to proto-Quechua. To make an even stronger case for truly localized Aymara origin, they should not even be attested in any other Quechuan lect outside Southern Peru. In practice, I have identified items shared by Cuzco Quechua as documented in the dictionary of the Academia Mayor de la Lengua Quechua (2005) and Aymara as documented in the pan-lectal dictionary by Huayhua Pari (2009). Given the multilayered interlacing of the Quechuan and Aymaran lineages (Emlen, 2017, to appear), attestation of an item in Cuzco Quechua and Aymara does not yet guarantee that the shared material is evidence for localized language contact between Cuzco Quechua and Aymara. Therefore, I have searched for cognates in other members of the Quechuan and Aymara families. For Quechuan, I have taken into account the varieties of Ayacucho (Quechua IIC, Parker, 1969a; Soto Ruiz, 1976), San Martín (Quechua IIB, Park et al., 1976), Chachapoyas (Quechua IIB, Taylor, 1979), Cañar (Quechua IIB, Cordero, 1992, 1895]), Inga (Quechua IIB, Tödter et al., 2002), Cajamarca (Quechua IIA, Quesada, 1976), Lambayeque (Quechua IIA, Taylor, 1996), Pacaraos (Quechua IIA, Adelaar, 1982), Yauyos (Quechua I, Shimelman, n.d.), Tarma (Quechua I, Adelaar, 1977), Huallaga (Quechua I, Weber et al., 1998), Huanca (Quechua I, Cerrón-Palomino, 1976), Ecuadorian Quichua (Quechua IIB, Yánez Cossío, 2007), and the colonial dictionary of de Santo Tomás (1560). For Jaqaru, I have relied on Belleza Castro’s (1993) dictionary. The likely non-exhaustive results of these comparisons can be seen, in alphabetical order, in Table 4.5 (I have standardized the original orthography in the sources for ease of comparison, representing / h/ by and not by also in Aymara and Jaqaru and not indicating allophonic vowel lowering in the context of uvulars in Cuzco Quechua). There is actually little vocabulary that can be associated with the semantic domains in which lexical items of replaced substratal languages usually tend to survive (though note that indeed these are only tendencies). Of the 23 items in Table 4.5, only one, (17.) qipu ‘thistle’ has to do with flora and fauna, another two, (8.) luqhe ~ lunqhe ~ lluqhe ‘lazy etc.’ and (20.) wallik’u ‘caruncle’ with undesirable personal qualities and physical deformations, and only item (24.), winku, in the reading ‘calabash used as a vessel,’ pertains to daily material culture. It is also notable that item (18.), tullka ‘son-in-law,’ figures on the list, suggesting that social relations at the time of borrowing were such that speakers of the two languages intermarried or were at least familiar with their respective kinship terminology; this suggests intimate and familiar relationships. Two semantic domains are, in fact, particularly well represented. These domains are weaving and spinning (items no. (1.), (7.), (13.), and (16.)), and animal breeding (items no. (2.), (3.), and (19.)). While under a certain interpretation these pertain to the material culture and the mode of subsistence of (some) speakers of Central Andean languages, the semantics suggests a transfer of know-how from speakers of Aymara to speakers of Quechua more than a situation in which Aymara speakers shifted to Quechua. Taken together, the lexical evidence does not add further support to a language shift scenario from Aymara to Quechua in the areas of Southern Peru that are now Quechuan-speaking. The main reason is that the continuing contact situation makes

hipi ‘to cut (with scissors or scythe), mow’

hip’i ‘stable or basic provisional pen to allow livestock herds to spend the night in open territory; rubbish remaining after threshing quinoa which consists of small twigs and perianths which fall like fine powder from the blade-side of leaves’

khaka(-) ‘thick, viscous, pasty, oversaturated; khakha ‘to sever completely, overproduction of a fruit tree; male duck; extend limbs’ nickname for fat people who walk slowly like ducks; to sever, break free from a “lodo” (?), to fit something together forcefully (Bolivia)’

k’aka(-) ‘crack; to crack, split, slash open’

k’ichi- ‘to pull small particles out of something k’ichi ‘to collect coca, to by pinching’ pinch’

(3.)

(4.)

(5.)

(6.)

khakha ‘to sever, tear down (wall)’

k’ichi ‘to pinch, pull out maize leaves, remove corn husk’

k’ak’a ‘to fissure, for the earth k’aki ‘split, crack, or other thing to split open’ opening; to split, to crack, have crack or split, be cracked’

hipi ‘shell of grain or cereals, quinoa saponine (Bolivia)’

haynu ‘husband’

haynu ‘male camelid selected for breeding in a haynu ~ aynu ‘husband’ special ceremony, generally the most sturdy animal with the best wool’

(2.)

Jaqaru cognate ts’uku ‘to hit, punch, sew; fuzz’

Aymara ch’uku ‘sewing; part of roof with sewing or central part of roof; to sew with thread’

Cuzco Quechua

ch’uku- ‘to curl, pleat, reduce the length of a textile while sewing, leaving curls or creases’

No.

(1.)

Table 4.5 Aymara borrowings in Cuzco Quechua, with Jaqaru cognates and annotations Comments

(continued)

See Cerrón-Palomino (2000: 169–171) for the correspondence /a/: /i/. Compare Tarma Quechua kaka ‘to cut open the throat of a freshly killed hen’ (Adelaar, 1977)?

Compare Tarma Quechua kaka-‘to cut open the throat of a freshly killed hen’ (Adelaar, 1977)?

See Cerrón-Palomino (2000: 136–137) for the correspondence /ch’/: /ts’/

4.4 Substrate Vocabulary 61

luqhe ~ lunqhe ~ lluqhe ‘lazy, negligent, luqhi ‘mad’ slovenly; braggart, dupe; irrational, mentally unstable, acting irresponsibly; nasty’ (Bolivia)

llamayu ‘to scratch, harvest tubers; season of harvesting tubers’

musi ‘to guard at night’

llaklla ‘sensitive, sensitive to touch; anxious, coward, discouraged, afraid’

llami(-) ‘touch, contact, beat; children’s game played in pairs with marbles or other small objects; to play, beat with the tact or the palm of hands’

(8.)

(9.)

(10.)

p’iti(-) ‘act of pulling; displacement of rope or p’ita ‘to weave (with stick or cord; piece of torn-away rope; to pull string or crocheting)’ band, separate or unlink spinneret or chain’

(13.)

uqu ‘to gulp down’

musi- ‘to guard, protect’

uqu- ‘to swallow, gulp down, to swallow rapidly without chewing well’

(11.)

(12.)

llahlla ‘fearful, timid’ (Bolivia)

k’ullk’u ‘narrow place’

k’ullku(-) ‘very thin, fine thread; to twist thread excessively; narrow’

Aymara

Cuzco Quechua

No.

(7.)

Table 4.5 (continued)

p’ityi ‘to weave (with stick or crocheting)’

uuqu ‘to swallow’

mushi ‘to guard’

ñami ‘to collect (grains from ground), harvest (wheat, beans), separate or arrange (small disordered things)’

ñahlla ‘timid’

luq’i ‘deaf’

k’uku ‘narrow, tight (corridor, gallery, gorge); narrow corridor, narrow gallery, narrow defile’

Jaqaru cognate

(continued)

See Cerrón-Palomino (2000: 171–172) for vowel length in Jaqaru; compare uqu(-) ‘swamp, wet ground; wet damp, moist; to become damp’ in other Quechuan lects such as Ayacucho (Parker, 1969a?)

See Cerrón-Palomino (2000: 157) for the correspondence /ll/: /ñ/

/qh/: /q’/ appears to be an irregular correspondence; note possible contamination with Span. loco ‘mad’

/ll/: /∅/ appears to be an irregular correspondence

Comments

62 4 The Aymara Presence in Southern Peru

phatu ‘to cast seed or seedlings densely’

phaqsi- ‘for the sun to emerge from behind the phaxsi ‘moon, month’ surrounding mountains’

qipu ‘small thorn of the pulla pulla prickly pear or cacti which contains allergogenic substances, causes warts’ tullqa ~ tulqa ‘son-in-law’ thunxu ‘to romp around, have faulty foot position’ (?)

phatu ‘thick, double, voluminous, of notable thickness’

phari(-) ‘loose cloth, thinly woven and not twisted a lot; squanderer, very lavish; to sew without twisting the wool or cotton too much’

qipo ‘thistle; fine thorns of thistle and other plant species’ (Bolivia)

tullka ‘son-in-law’

thunku(-) ‘shackle, rope to tie animals’ feet; the act of tying, leashing; knot of shackle; handcuff; tie animals together by the horns,’ thunku thunku ‘one walking with difficulty, as if in shackles or handcuffed’

(15.)

(16.)

(17.)

(18.)

(19.)

phala ‘twisted rope made from chillihua or paja brava grass; to twist with the palms of the hands, wind’

Aymara

Cuzco Quechua

No.

(14.)

Table 4.5 (continued) Comments

Compare Cerrón-Palomino (2000: 133) for the correspondence /q/: /qh/

(continued)

tyhunka ‘to make walk /u/: /a/ and /x/: /k/ appear to be irregular with force’ correspondences

tyullqa ‘son-in-law’

qhipu ‘a weed with thorns; swelling or wart [sic!] that forms by piercing of the thorns of that plant’

phari ‘to twist or twirl /l/: /r/ appears to be an irregular (rope)’ correspondence

pahshi ‘moon, month’ Compare Torero (2002: 138), who noted the presence of this item, e.g., in Gonçalez Holguin’s (1608) colonial Quechua dictionary, but did not note its presence—with altered semantics—in the modern language

phatru ‘thick (especially flat things)’

Jaqaru cognate

4.4 Substrate Vocabulary 63

Cuzco Quechua

wallik’u ‘caruncle on the body of humans and the neck and breast of cattle’

wat’atata(-) ‘to deliver kicks in self-defense, for animals to kick while taking their last breaths; convulsion or unexpected movements of worms’

wikha(-) ‘piece, part, rest, segment of something, separated forcefully; to break or separate part of a peg’

winku ‘small calabash, utensil (generally half of a calabash) used to scoop liquids; oblong, elongate’

No.

(20.)

(21.)

(22.)

(23.)

Table 4.5 (continued) walli ‘to disembowel, pull out bowels (of animal), incise or cut up finely (for dried meat or jerky)’

Jaqaru cognate

winku ‘to lie’

wikha- ‘to break tender objects without spoon’

See Cerrón-Palomino (2000: 169–171) for the correspondence /a/: /i/. The correspondence /kh/: /k’/, on the other hand, appears to be irregular.

Comments

winkuru ‘to crouch Compare winku(-) ‘twisted, wrinkled, down, bow, make warped; to twist slightly’ in Ayacucho oneself small for fear’ Quechua (Parker, 1969a, 1969b) and other varieties?

wik’i ‘to pull (root), push (detachable things)’

wat’a ‘to stretch legs; blanket’ wats’a ‘foot of the bed’

walli ‘varix, blood which coagulates in the vein’

Aymara

64 4 The Aymara Presence in Southern Peru

4.4 Substrate Vocabulary

65

it exceedingly difficult to exclude that relevant items in Table 4.5 were transferred in the context of sustained bilingualism rather than language shift. The semantics of borrowed items, furthermore, are not strongly suggestive of a language shift scenario – but given the continuing contact situation, they do not contradict such a scenario either. While Thomason and Kaufman (1988: 42–43) have emphasized that structural interference need not be accompanied by lexical borrowing, Thomason (2009) at least considers lexical borrowing as one possible way to demonstrate that the languages were in contact at all. What remains are the following observations: • ethnohistorical evidence from parts of the Ayacucho region, the Cuzco region, and Bolivia suggests a continuing expansion of Quechuan ever further to the south and the east into regions where Aymara had been spoken before • concomitantly, there are increasingly strong structural similarities of expanding Quechuan lects with Aymara • there is clear and ample Aymara toponymy throughout the region in which this expansion happened, as well as regions further west (Ayacucho, etc.). Given the apparently protracted process of the expansion of Quechuan, beginning in prehistory and continuing into the historical period, it is clear that contact must have been “long-term” and “intimate” (Mannheim, 1991: 178). But that specifies only its intensity, not its quality.13 Since the ethnohistorical evidence indicates clearly that Quechuan replaced Aymara in the colonial period in significant portions of Southern Peru over several centuries, the emerging overall picture is one of a long-term language contact situations where languages shift, which likely continued from prehistoric times rather seamlessly into colonial times, was accompanied or preceded by extensive periods of bilingualism over several generations. This is also consistent with the semantic domains in which lexical borrowings from Aymara into Cuzco Quechua occur. Such a scenario is also either posited explicitly or implied in the literature. Indeed, the discussion of pertinent issues is often framed explicitly in terms of substrate influence as when Torero (1964: 463), Cerrón-Palomino (2003 [1987]: 182), and Heggarty and Beresford-Jones (2013: 405) speak of “Aymara substrate traces” in Cuzco Quechua. Torero (2002: 108) also speaks of Aymara gradually imbuing Quechuan structurally as it lost territory and speakers to Quechuan. Adelaar and Muysken (2004: 262) speak of an “Aymara substratum” for the lects of Puno and Arequipa and then become more explicit in mentioning language shift as the responsible process for bringing about the morphological influence of Aymara on the Cuzco-Bolivian lects of those regions. The effects, they say, “can only be explained by assuming a relatively recent language shift from Aymara to Quechua after extensive bilingualism.” More generally, mentioning (in a footnote) the Aymara toponymic layer in areas that are now Quechuan-speaking, Durston (2007) says that the “historical and toponymic evidence points to multiple linguistic strata, which can

13

Adelaar and Muysken (2004: 195) merely speak of “language contact between these dialects and Aymara.”.

66

4 The Aymara Presence in Southern Peru

often be attributed to language shift, the result of a combination of rampant multilingualism and changing relations of power over time.” The strongest and farthestreaching statement which most explicitly refers to a scenario of substrate influence during language shift that I am aware of actually comes from non-linguists: Bird et al. (1983, 1984: 193) mention that “Aymara influence might derive from … faulty learning of Quechua by Aymara-speaking people learning from the descendants of earlier Wari period colonists [who are assumed by the authors to have been Quechua speaking, MU].”

4.5 Perspectives from Outside Linguistics As stated in Sect. 3.5, scenarios for language shift and language replacement and hypotheses as to the structure and affiliation of substratal languages should, for methodological reasons, rest on linguistic evidence alone. However, in a separate second step, it can be rewarding to see how the language-based scenarios fit with what can be inferred on demographic changes in prehistory from other disciplines. As far as the substratal Aymara layer in Southern Peru and Bolivia is concerned, there is excellent evidence from genetics that is directly relevant for the linguistic scenario. For Bolivia, Gayà-Vidal et al. (2010) investigate the genome (specifically, Alu insertions in the autosome and the X chromosome) of 192 Aymara and Quechua speakers from Bolivia. Their specific goal is to test two competing hypotheses of Quechuan language spread: one in which there is a considerable influx of Quechuan speakers from Peru that bring with them their language, and another in which such migrations play at best a subordinate role, and rather speakers of Aymara, with roots in Bolivia, shift to Quechuan. They find that Bolivian speakers of Quechua and Aymara are much more similar to one another with respect to the investigated part of the genome than with Quechuan-speaking people of Peru, suggesting that “the Quechua language expansion in Bolivia took place without any important demographic contribution” (Gayà-Vidal et al., 2010: 154), i.e., by language shift.

4.6 Summary Despite some mismatches between the stories that the different types of Sprachdenkmäler of Southern Peru tell, a relatively strong case for recent language shift from Aymara to Quechua can be made for this region. This shift appears to have taken place first in the Ayacucho area, then at a slightly later point of time in the region of Cuzco, and still later in most parts of Bolivia. The strength of the traces that Aymara has left on the Quechua spoken by the descendants of the shifting speakers appears to have accumulated to some extent, as they are generally stronger in those areas in which the shift has occurred most recently. Note, however, that Quechua IIC, including Ayacucho Quechua, as a whole can be thought of as exhibiting Aymara

References

67

influence: we have seen how Parker (1969b: 169) relates some suffixes that are innovated in Quechua IIC to an Aymara source. In addition, it would be a curious coincidence that the Quechua IIC subgroup is mainly defined by the sound change */∫/ to /s/, and that the Aymara language is distinguished from its northern relative Jaqaru by exactly the same sound change (Cerrón-Palomino, 2000: 146–147). Rather, as already argued by Adelaar (1986), these parallel changes are more likely to reflect periods of prolonged and intense interaction. The toponymic record does not show these chronologically sequenced layers of Aymara influence, but rather presents, as a kind of baseline to the contact events that are responsible for the effects discussed in Sects. 4.3 and 4.4, uniformly strong evidence for an Aymara presence in Southern Peru and the Quechuan-speaking parts of Bolivia.

References Academia Mayor de la Lengua Quechua. (2005). Diccionario Quechua—español—quechua, qheswa—español—qheswa simi taqe. Gobierno Regional Cusco. Adelaar, W. F. H. (1982). Léxico del quechua de Pacaraos. Universidad Nacional Mayor de San Marcos, Centro de Investigación de Lingüística Aplicada. Adelaar, W. F. H. (1986). La relación quechua-aru: Perspectivas para la separación del léxico. Revista Andina, 4(2), 379–426. Adelaar, W. F. H. (1987). Aymarismos en el quechua de Puno. Indiana, 11: 223–231. https://doi. org/10.18441/ind.v11i0.223-231 Adelaar, W. F. H. (1977). Tarma Quechua: Grammar, texts, dictionary. Peter de Ridder Press. Adelaar, W. F. H., & Muysken, P. C. (2004). The languages of the Andes. Cambridge University Press. Belleza Castro, N. (1993). Vocabulario jaqaru-español. https://de.scribd.com/doc/161208137/ Diccionario-Jaqaru-1, https://de.scribd.com/doc/161207982/Diccionario-Jaqaru-2. Accessed March 30, 2023. Bertonio, L. (1612). Arte de la lengua aymara, con vna silva de phrases de la misma lengua, y su declaracion en Romance. Francisco del Canto. Bird, R. McK., Browman, D. L., & Durbin, M. E. (1983/1984). Quechua and maize: mirrors of Central Andean culture history. Journal of the Steward Anthropological Society, 15(1/2), 187– 240. Campbell, L. (1995). The Quechumaran hypothesis and lessons for distant genetic comparison. Diachronica, 12(2), 157–200. https://doi.org/10.1075/dia.12.2.02cam Cerrón-Palomino, R. (1976). Diccionario quechua: Junín-Huanca. Ministerio de Educación/ Instituto de Estudios Peruanos. Cerrón-Palomino, R. (1997). La primera documentación del aimara. In Zimmermann, K. (Ed.), La descripción de las lenguas amerindias en la época colonial (pp. 195–257). Iberoamericana/ Vervuert. Cerrón-Palomino, R. (1999). Tras las huellas del aimara cuzqueño. Revista Andina, 17(1), 137–161. Cerrón-Palomino, R. (2000). Lingüística aimara. Centro de Estudios Regionales Andinos ‘Bartolomé de las Casas.’ Cerrón-Palomino, R. (2003 [1987]). Lingüística Quechua (2nd ed.). Centro de Estudios Regionales Andinos ‘Bartolomé de las Casas.’ Cerrón-Palomino, R. (2005). La toponimia como fuente de usos gramaticales arcaicos y como registro cronológico de desplazamientos idiomáticos. Lexis, 29(1), 111–124. https://doi.org/10. 18800/lexis.200501.006

68

4 The Aymara Presence in Southern Peru

Cerrón-Palomino, R. (2006a). Ollantay: Topónimo antes que antropónimo. Lexis, 30(2), 323–340. https://doi.org/10.18800/lexis.200602.007 Cerrón-Palomino, R. (2006b). Cuzco: La piedra donde se posó la lechuza. Historia de un nombre. Lexis, 30(1), 143–184. https://doi.org/10.18800/lexis.200601.005 Cerrón-Palomino, R. (2015). The languages of the Inkas. In I. Shimada (Ed.), The Inka empire: A multidisciplinary approach (pp. 39–53). University of Texas Press. Coler, M. (2014). A grammar of Muylaq’ Aymara: Aymara as spoken in Southern Peru. Brill. Cordero, L. (1992 [1895]). Diccionario quichua-castellano y castellano-quichua (5th ed.). Proyecto Educación Bilingüe Intercultural/Corporación Editora Nacional. de Acuña, F., de la Cueva, L. S., Quispe, C., Goliardo, F. S., Rigón, M., Chaguaya, D. N., de Medina, L., & Supanta, S. (1965 [1586]). Relación fecha por el corregidor de los Chunbibilcas Don Francisco de Acuña, por mandado de su Ex.o del señor Don Fernando de Torres y Portugal, visorrey destos reynos, para la discrepción de las Indias que su Majestad manda hacer. In M. Jiménez de la Espada (Ed.), Relaciones geográficas de Indias—Peru (vol. 1, pp. 310–325). Atlas. de Carabajal, P. (1965 [1586]). Descripción fecha de la provincia de Vilcas Guaman por el ilustre señor Don Pedro de Carabajal, corregidor y justicia mayor della, ante Xristobal de Gamboa, escribano de su juzgado, en el año de 1586. In M. Jiménez de la Espada (Ed.), Relaciones geográficas de Indias—Peru (vol. 1, pp. 205–220). Atlas. de Monzón, L. (1965 [1586]). Descripción de la tierra del repartimiento de los Rucanas Antamarcas de la corona real, jurisdicción de la ciudad de Guamanga, año de 1586. In M. Jiménez de la Espada (Ed.), Relaciones geográficas de Indias—Peru (vol. 1, pp. 237–248). Atlas. de Monzón, L., Saravia, B., de Frías, P., Taypimarca, P. (1965ª [1586]). Descripción de la tierra del repartimiento de Atunsora, encomendado en Hernando Palomino, jurisdición de la ciudad de Guamanga, año de 1586. In M. Jiménez de la Espada (Ed.), Relaciones geográficas de Indias—Peru (vol. 1, pp. 220–225). Atlas. De Reuse, W. J. (1981). Grassman’s law in Ofo. International Journal of American Linguistics, 47(3), 243–244. https://doi.org/10.1086/465693 de Santo Tomás, D. (1560). Lexicon, o vocabulario de la lengua general del Perú. Francisco Fernandez de Cordoua. de Saussure, F. (1916). Cours de linguistique générale (C. Bally, A. Sechehaye, with the collaboration of Albert Riedlinger (Eds.).) Payot. de Ulloa y Mogollón, J., Hernández Talavera, D., de la Feria, H. M., González, A., Gómez de Butrón, G., Taypi, M. N., Caquia, J., Pacta, F. I., Chacha, D., & Anco, D. C. ([1586] 1965). Relación de la provincia de los Collaguas para la discrepción de las Indias que su Magestad manda hacer. In M. Jiménez de la Espada (Ed.), Relaciones geográficas de Indias—Peru (vol. 1, pp. 326–333). Atlas. Durston, A. (2007). Pastoral Quechua: The history of Christian translation in colonial Peru, 1550– 1650. University of Notre Dame Press. Emlen, N. Q. (2017). Perspectives on the Quechua-Aymara contact relationship and the lexicon and phonology of pre-proto-Aymara. International Journal of American Linguistics, 83(2), 307–340. https://doi.org/10.1086/689911 Emlen, N. Q. (To appear). The Quechua-Aymara relationship. In M. Urban (Ed.), The Oxford Guide to the Languages of the Central Andes. Oxford University Press. Ferrell, R. M. A. (1996). Textos aimaras en Guaman Poma. Revista Andina, 14(2), 413–455. Filppula, M. (1986). Some aspects of Hiberno-English in a functional sentence perspective. University of Joensuu. Floyd, S. (2021). Oclusivas complejas en el quechua de Domingo de Santo Tomás. Letras, 92(136), 115–140. https://doi.org/10.30920/letras.92.136.10 Gayà-Vidal, M., Dugoujon, J.-M., Esteban, E., Athanasiadis, G., Rodríguez, A., Villena, M., Vasquez, R., & Moral, P. (2010). Autosomal and X chromosome Alu insertions in Bolivian Aymaras and Quechuas: Two languages and one genetic pool. American Journal of Human Biology, 22(2), 154–162. https://doi.org/10.1002/ajhb.20967

References

69

Gonçalez Holguin, D. (1608). Vocabvlario de la lengva general de todo el Perv llamada lengua Qquichua, o del Inca. Francisco del Canto. Guillaume, A. (2016). Associated motion in South America: Typological and areal perspectives. Linguistic Typology, 20(1), 81–177. https://doi.org/10.1515/lingty-2016-0003 Harris, O. (1975). Laymis y Machas: Temas culturales del norte Potosí. In Dualismo o pluralismo cultural en Bolivia: Trabajos presentados a la Mesa Redonda sobre expresiones de la cultura boliviana en el lapso, 1925–1974 (vol. 1, pp. 73–83). Casa Municipal de la Cultura Franz Tamayo. Heggarty, P., & Beresford-Jones, D. (2012). A cross-disciplinary prehistory for the Andes? Surveying the state of the art. In P. Heggarty, & D. Beresford-Jones (Eds.), Archaeology and language in the Andes: A cross-disciplinary exploration of prehistory (pp. 407–434). Oxford University Press. Heggarty, P., & Beresford-Jones, D. (2013). Andes: Linguistic history. In P. Bellwood (Ed.), The global prehistory of human migration (pp. 401–409). Wiley Blackwell. Heine, B., & Kuteva, T. (2005). Language contact and grammatical change. Cambridge University Press. Huayhua Pari, F. (2009). Diccionario bilingüe polilectal aimara—castellano castellano—aimara. Fondo Editorial de la Universidad Nacional Mayor de San Marcos. Itier, C. (2011). What was the Lengua general of colonial Peru? In P. Heggarty, & A. J. Pearce (Eds.), History and language in the Andes (pp. 63–85). Palgrave Macmillan. Jiménez de la Espada, M. (Ed.). (1965). Relaciones geográficas de Indias—Perú. Atlas. Landerman, P. N. (1994). Glottalization and aspiration in Quechua and Aymara reconsidered. In P. Cole, G. Hermon, & M. D. Martín (Eds.), Language in the Andes (pp. 332–378). University of Delaware, Latin American Studies Program. Landerman, P. (1998). Internal reconstruction in Aymara and Quechua. In J. H. Hill, P. J. Mistry, & L. Campbell (Eds.), The life of languages: Papers in linguistics in honor of William Bright (pp. 35–57). Mouton de Gruyter. Mannheim, B. (1985a). Contact and Quechua-external genetic relationships. In H. E. Manelis Klein, & L. R. Stark (Eds.), South American Indian languages: Retrospect and prospect (pp. 644–688). University of Texas Press. Mannheim, B. (1985b). Southern Peruvian Quechua. In H. E. Manelis Klein, & L. R. Stark (Eds.), South American Indian languages: Retrospect and prospect (pp. 481–515). Austin: University of Texas Press. Mannheim, B. (1991). The language of the Inka since the European invasion. University of Texas Press. Middendorf, E.W. (1891). Die Aimarà-Sprache: Mit einer Einleitung über die frühere Verbreitung der diese Sprache redenden Rasse und ihr Verhältnis zu den Inkas. F.A. Brockhaus. Mossel, A., & Emlen, N. Q. (To appear). Northern Aymara toponymy, revisited. International Journal of American Linguistics. Muysken, P. (2015). From nominal predicate to deictic clausal highlighter: The development of hina ‘like.’ In M. S. Manley, & A. Muntendam (Eds.), Quechua expressions of stance and deixis (pp. 259–286). Brill. https://doi.org/10.1163/9789004290105_008 Park, M., Weber, N., & Sangama, V. C. (1976). Diccionario quechua: San Martín. Ministerio de Educación/Instituto de Estudios Peruanos. Parker, G. J. (1969a). Ayacucho uechua grammar and dictionary. Mouton. Parker, G. J. (1969b). Comparative Quechua phonology and grammar III: The Evolution of Quechua A. University of Hawaii Working Papers in Linguistics, 1(9), 149–204. Proulx, P. (1972). Proto-Quechua/ph/. International Journal of American Linguistics, 38(2), 142– 145. https://doi.org/10.1086/465195 Proulx, P. (1974). Certain aspirated stops in Quechua. International Journal of American Linguistics, 40(3), 257–262. https://doi.org/10.1086/465319 Quesada C., F. (1976). Diccionario quechua: Cajarmarca-Cañaris. Ministerio de Educación/ Instituto de Estudios Peruanos.

70

4 The Aymara Presence in Southern Peru

Salmons, J. (1992). Northwest Indo-European vocabulary and substrate phonology. In Perspectives on Indo-European language, culture and religion: Studies in honor of Edgar C. Polomé (vol. 2, pp. 265–279). Institute for the Study of Man. Shimelman, A. (n.d.). Yauyos Quechua Lexicon/Léxico del Quechua de Yauyos. https://www.ailla. utexas.org/islandora/object/ailla%3A242751. Accessed April 6, 2023 Soto Ruiz, C. (1976). Diccionario quechua: Ayacucho-Chanca. Ministerio de Educación/Instituto de Estudios Peruanos. Stark, L. R. (1985). The Quechua language in Bolivia. In H. E. Manelis Klein, & L. R. Stark (Eds.), South American Indian languages: Retrospect and prospect (pp. 516–545). University of Texas Press. Taylor, G. (1979). Diccionario normalizado y comparativo quechua: Chachapoyas-Lamas. L’Harmattan. Taylor, G. (1996). El quechua de Ferreñafe: Fonología, morfología, léxico. Acku Quinde. Tercer Consilio Limense. (1984 [1584]). Doctrina christiana y catecismo para instrvcción de los Indios, y de las de mas personas, que han de ser enseñadas en nuestra santca Fé. Con vn confessionario, y otras cosas necessarias para los que doctrinan, que se con tienen en la pagina siguiente. Petroperú. Thomason, S. G. (2009). How to establish substratum interference. In Y. Nagano (Ed.), Issues in Tibeto-Burman historical linguistics (pp. 319–329). National Museum of Ethnology. Thomason, S. G., & Kaufman, T. (1988). Language contact, creolization, and genetic linguistics. University of California Press. Thompson, L. C., & Thompson, M. T. (1985). A Grassman’s Law for Salish. Oceanic Linguistics Special Publications, 20, 134–147. Tödter, C., Waters, W., & Zahn, C. (2002). Shimikunata asirtachik killka inka-kaste-llanu: Diccionario inga-castellano (Quechua del Pastaza). Instituto Lingüístico de Verano. Torero, A. (1964). Los dialectos quechuas. Anales Científicos De La Universidad Agraria, 2(4), 446–478. Torero, A. (1975 [1970]). Lingüística e historia de la sociedad andina. In R. A. de Matos, & R. Ravines (Eds.), Lingüística e indigenismo moderno en América (Trabajos presentados al xxxix congreso internacional de Americanistas) (vol. 5, pp. 221–259). Instituto de Estudios Peruanos. Torero, A. (2002). Idiomas de los Andes: Lingüística e historia. Instituto Francés de Estudios Andinos/Editorial Horizonte. Urban, M. (2018). The lexical legacy of substrate languages: A test case from the Southern Ecuadorian highlands. Transactions of the Philological Society, 116(3), 435–459. https://doi.org/10. 1111/1467-968X.12129 Urban, M. (2021). Cholón and the linguistic prehistory of northern Peru: Triangulating toponymy, substrate lexis, and areal typology. Linguistic Discovery, 17(1). https://doi.org/10.1349/PS1. 1537-0852.A.513 Weber, D. J., Cayco Zambrano, F., Cayco Villar, T., & Ballena Dávila, M. (1998). Rimaycuna: Quechua de Huánuco. Diccionario del quechua del Huallaga con índices castellano e inglés. Instituto Lingüístico de Verano. Yánez Cossío, C. (2007). Lexico ampliado quichua-español, español-quichua. Corporación Educativa Macac.

Chapter 5

The Barbacoan Languages and the Southern Ecuadorian Highlands

Abstract This chapter presents the second of three case studies, applying the framework for studying lost languages to the Andes of the southern Ecuadorian highlands. While both (from the Quechuan point of view) “exotic” toponymy and a layer of lexical items with similarly “exotic” sound structure have been noted in the Quechuan of Cañar and Azuay at the outset of modern Quechuan historical linguistics, this case is still much less explored than the one of the previous chapter. This chapter analyzes the toponymy of the Ecuadorian provinces of Cañar and Azuay in the greatest detail so far, and not only carves out its phonological characteristics, but also shows continuities of certain toponymic endings of the region with the north, in particular the Ecuadorian-Colombian border area. Drawing on an earlier study, the chapter shows that the local Quechuan lect has been imbued with a lexical layer of foreign origin that at the same has altered its segmental phonology. The different types of Sprachdenkmäler are suggestive of a Barbacoan presence throughout the Ecuadorian highlands, though there are some loose threads and open question that invite further research. Keywords Quechuan · Barbacoan · Southern Ecuador · Cañar · Azuay · Toponymy · Anthroponyms · Substrate · Contact-induced language change · Language shift · Lexical borrowing · Loanwords

5.1 Introduction From its inferred homeland in Central Peru, Quechuan expanded both in a northward and southward direction to reach places as far apart as the Sibundoy valley in southern Colombia and the plains of Santiago del Estero in Northwestern Argentina.1 In the previous Section, we have discussed some of the latest phases of the southward spread of Quechua IIC into Southern Peru and Bolivia. The case study in this section puts the spotlight on the late phase of Quechuan spread in a northward direction, to Ecuador. 1

The chronology of this expansion is a matter of current debate. Alongisde coincidental spread both to the north and south, it is also possible that the spread initially was to the south only, and that the north was then “colonized” from there.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Urban, Linguistic Stratigraphy, https://doi.org/10.1007/978-3-031-42102-0_5

71

72

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

While Torero (1984) has sought to make plausible that Quechuan spread through Ecuador in pre-Inca times in the context of commercial activities of people from Chincha in Central Peru, more recently, Hocquenghem (2012) suggests a shallower time-depth for Quechuan in Ecuador.2 According to Hocquenghem, it is more likely that Quechuan arrived with the Inca and continued to spread through Ecuador in later colonial times (see also Stark, 1985: 444–446 for further review of proposals). Today, Ecuadorian Quichua—the difference in spelling is conventionalized, but reflects the fact that uvulars have merged with velars in Ecuador so that the high vowel is not lowered—is spoken throughout the Ecuadorian highlands. The presence is particularly strong in the provinces of Imbabura and Chimborazo (see Stark, 1985: 458–473 for an extended sociolinguistic survey of the situation in the 1980s). Ecuadorian Quichua has been classified as belonging to the Quechua IIB group, which also includes Chachapoyas and San Martín Quechua on the eastern slopes of the Northern Peruvian Andes, as well as the lowland varieties spoken on the shores of the Pastaza and Napo rivers (see Sect. 2.3). As already alluded to, these varieties merge uvulars with velars and tend to voice occlusives after nasals (CerrónPalomino, 2003 [1987]: 239).3 Ecuadorian Quichua is mainly distinguished by the simplification of the morphological apparatus. With the exception of the Pastaza lect, morphological marking of possession on nouns is replaced by analytic possessive constructions, the distinction between inclusive and exclusive possession is lost, and there are various changes in the function of preserved morphological makers (Cerrón-Palomino, 2003 [1987]: 239–240). Ecuadorian Highland Quichua distinguishes itself from the lowland varieties by the presence of aspirated consonants (or sounds that clearly developed from aspirated consonants). There are two competing hypotheses as to the historical background of these aspirated consonants. On the one hand, as we have already seen in the discussion of nonmodal consonants in Cuzco-Bolivian Quechua in Sect. 4.3, there are those, including Campbell (1995) and Orr and Longacre (1968), who consider them an inheritance from a protoQuechua language that featured nonmodal consonants. On the other hand, it has been argued that Ecuadorian Quichua aspirates could have arisen through the influence of Cuzco-type lects spoken by Inca elites that were present to consolidate imperial control in the newly conquered provinces of Ecuador (Stark, 1985: 453–454, see also Landerman, 1994: 340). Note that the correspondences between aspirates in Ecuadorian Quichua and Cuzco-Bolivian Quechua have been considered irregular by Torero (1984), see also Cerrón-Palomino (2003 [1987]: 187–188). If one accepts this conclusion (which may require further empirical support), this may suggest that aspirated consonants reflect a IIC overlay in what are essentially IIB-type lects (compare Cerrón-Palomino, 2003 [1987]: 344). The further development of the aspirated consonants allows to distinguish two different types of lects in the Ecuadorian highlands (Cerrón-Palomino, 2003 [1987]: 240): Quichua as spoken in the Imbabura 2

See Stark (1985: 453–458) for another proposal for a “deep chronology” of Quechuan in Ecuador. Note that the presumed coastal Quechua described by de Santo Tomás (1560a, 1560b) has also been considered to belong to the IIB branch, in spite of strong evidence that it retains the velar-uvular distinction (Cerrón-Palomino, 1990) as well as featured nonmodal consonants (Floyd, 2021).

3

5.1 Introduction

73

and Pichincha provinces maintains /th/, but the labial and velar aspirated consonants /ph/ and /kh/ develop into fricatives /f/ and /χ/, respectively. The Quichua lects of Tungurahua, Chimborazo, Cañar, and Azuay maintain the aspirated consonants unchanged. With the exception of Loja and the lowland lects, proto-Quechua */ll/ develops into a voiced fricative (Cerrón-Palomino, 2003 [1987]: 163–164). Morphological evidence suggests to Stark (1985: 447) a division between northern lects of Pinchincha, Cotopaxi, and Tungurahua versus southern lects in Imbabura, Chimborazo, Cañar, Azuay, and Loja. Comparing this picture with the phonological changes shows that, as is true elsewhere, Ecuadorian lects showcase a complex web of isoglosses that do not necessarily bundle, and this makes it difficult to associate particular innovations with particular substrate languages (see Muysken, 2021). The only non-Quechuan language family that is still represented in the Ecuadorian highlands today is the small Barbacoan family, consisting of four languages spoken in Ecuador as well as Colombia (see also Sect. 2.2). In spite of its limited geographical spread, Barbacoan does not appear to be a shallow language family. Cognacy rates are low even in basic vocabulary (Curnow & Liddicoat, 1998: 391), suggesting a significant period of in situ diversification. Curnow and Liddicoat (1998: 404–405) tentatively propose a binary family tree with two main branches: One would consist of Guambiano-Totoró (a.k.a. Nam Trik), two closely related lects spoken in the highlands of southern Colombia and another language, Awa Pit, which is actually spoken on both sides of the Ecuadorian-Colombian border in a relatively inaccessible region on the rugged western slopes of the Andes (Curnow, 1997: 2–4). Research reported in preliminary form in Norcliffe et al. (2015) suggests a flatter structure, with Awa Pit forming a subgroup of its own. Barbacoan languages are, however, not exclusively found in the highlands. The less controversial so-called southern branch, consisting of the Cha’palaa and Tsafiqui languages, is found in the Pacific-facing lowlands of Ecuador west of the Andean mountain chains. The hallmark of a Barbacoan presence in the toponymic record is the ending -pí and its variant -bí, which reflects a pan-Barbacoan root *pii ‘river, water’ (Curnow & Liddicoat, 1998: 392). Corroborating evidence for this is that the initial constituent of such river names is in some cases also identifiable as Barbacoan in origin (e.g., the case of Manabí, where mana is identified as the word for ‘deer’ by Gómez Rendón, 2017: 64fn8). Toponyms with this ending, which unsurprisingly frequently denote rivers, are widely distributed not only in the Andes, but also in large parts of the northwestern lowlands of the Nariño department (where the suggestively named town of Barbacoas is located); Beuchat and Rivet (1907: 69) already noted the particularly dense concentration on the left shore of the Río Patía in Colombia. This area, once inhabited by a group called Sindagua, may accordingly have been Barbacoan-speaking at some point, too. Jijón y Caamaño (1939: 171) claims that the Río San Juan originally separated Chocoan-speaking people in the north from the Barbacoan-speaking Barbacoas people, to whom the name of the town of Barbacoas also goes back, in the south. As one can see in Fig. 5.1, relevant toponyms in pí ~ -bí ~ -pi ~ -bi are also found in the highlands, suggesting the possibility of an essentially undocumented extension of Barbacoan languages into the Ecuadorian Andes. Wolf (1879: 58) already noted that -pí and -bí occur widely and suggested

74

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

that tracing this element might provide an opportunity to identify prehistoric people who spoke-related languages. The ethnohistorical record for the highlands of Ecuador allows to catch a glimpse at a highly diverse linguistic landscape, with a number of now extinct small-scale local languages succeeding each other as one moves from north to south (Fig. 5.2). The northernmost one, once spoken “around the border towns of Tulcán [in the Ecuadorian province of Carchi, MU] and Ipiales, striding the Colombian-Ecuadorian border” (Adelaar & Muysken, 2004: 393), is known as Pasto. There is a short wordlist

Fig. 5.1 Toponyms ending in -pí ~ -bí ~ -pi ~ -bi in Ecuador and Colombia

5.1 Introduction

75

Fig. 5.2 Ecuadorian highland provinces and pre-Inca language areas. The map also shows towns mentioned in the text

from the nineteenth century describing the Colombian lect of Muellamués (the data are in Jijón y Caamaño 1940: 197). This lect “may have represented one of the last surviving dialects of the Pasto language” (Adelaar & Muysken, 2004: 393). This word list allows to clearly identify Muellamúes/Pasto as a Barbacoan language, more specifically, as “clearly related, though not identical, to Awa Pit” (Adelaar & Muysken, 2004: 393).

76

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

Cara, Pasto’s and Awa Pit’s former southern neighbor, was spoken from the Mira and Chota valleys in northern Ecuador to the northern limits of the Ecuadorian capital of Quito (Fig. 5.2). These areas pertain to the present-day Ecuadorian provinces of Imbabura and Pichincha, known in the present as centers of Indigenous culture and as strongholds of Quechuan. Language shift from Cara to Quechuan may only have been completed in the eighteenth century (Gómez Rendón & Adelaar, 2009: 947–948), suggesting the coexistence of Cara, Quichua, and Spanish for several centuries in a likely multifarious situation of multilingualism. Cara, unlike Pasto, is entirely undocumented (according to von Buchwald, 1908: 123, Cara rulers were called shiri, a title that may or may not go back to the Cara language). However, the fact that the toponymic areas that can be assigned to the two languages showcase some commonalities (e.g., the endings -mued and -pud) suggests “some kind of relationship” (Adelaar & Muysken, 2004: 394) between them. This and the fact that also the former Cara-speaking areas features river names in -pí and -bí (as can be seen in Fig. 5.1, see also Jijon y Caamaño, 1940: 239–240) suggest that Cara was a Barbacoan language, too (Adelaar & Muysken, 2004: 394).4 Gómez Rendón (2017: 62–63) asserts that the Ecuadorian highlands from Carchi to Cotopaxi showcase Barbacoan toponymy, thereby affirming the Barbacoan affiliation of at least Pasto and Cara, while noting that the toponymic fingerprint becomes less clear in the central highlands.5 In addition, Gómez Rendón and Adelaar (2009: 954) identify four lexical items in Imbabura Quechua to which they attribute a Barbacoan, specifically Cara, origin. They note that likely more items would come to light if the study were extended to further semantic domains that are not strongly represented in their study, particularly flora and fauna vocabulary. Salomon and Grosboll (1986) have identified two linguistic source populations for personal names of the area of Quito through statistical analysis; the database is a large set of personal names recorded in a 1559 inspection of six villages in the Quito basin. Names from villages in the northern part cluster together and also the names from villages in the southern part show communalities not found in the north, including the presence of orthographic clusters , , and (Salomon & Grosboll, 1986: 394). More systematically, the language underlying names of the northern villages showcases evidence for postnasal voicing, whereas the one of the southern villages does not (Salomon & Grosboll, 1986: 394). It makes sense to associate the northern population Salomon and Grosboll have identified with Cara and the southern one with Panzaleo, the language mentioned to the south of Cara in colonial documents (Salomon & Grosboll, 1986: 396; the precise location of the linguistic frontier is debated, cf. Gómez Rendón, 2017: 60–61). That the signature of the two languages can be identified in personal names suggests that they were notably different from one another. Indeed, the Panzaleo area, including the south of the present-day province of Pichincha as well as Cotopaxi and Tungurahua, showcases 4

Paz y Miño (1941: 46) already suggested the Barbacoan affiliation of Cara (though he denied that there were any similarities with Pasto and Panzaleo). 5 Jijon y Camaaño (e.g., 1943: 181) posits a particularly close relationship between Cara and the Southern Barbacoan languages Tsafiqui and Cha’palaa.

5.1 Introduction

77

a toponymic fingerprint that is different from Cara. Characteristic endings are -leo, which also occurs in the placename Panzaleo itself, -haló ~ -aló, and -lagua ~ -ragua (Adelaar & Muysken, 2004: 394); the latter is represented for instance in the name of the Tungurahua province. Panzaleo toponymy has always appeared more difficult to interpret than that associated with Cara and Pasto; Jijón y Caamaño (1943: 181– 389), for instance, is at great pains to determine the nature of and influences on the language. To the south of Panzaleo, in most of today’s Chimborazo province, and centered around the province’s capital Riobamba, a language known as Puruhá or Puruguay was spoken. Puruhá placenames frequently end in -shi, -tus, -bug, -cahuan, -calpi, and -tactu (Adelaar & Muysken, 2004: 395). There are some placenames in -pí in the area in which Puruhá was once spoken, but not many (Jijón y Caamaño 1940: 402; the map in Fig. 5.1 confirms this impression). There is no known documentation of Puruhá. In spite of this, Puruhá is “generally believed to be relate to Cañar” (Adelaar & Muysken, 2004: 395), the language once spoken to the south in the Cañar and Azuay provinces with which this chapter is concerned.6 On the one hand, this is because there are some endings in toponyms that are common in the areas associated with both; Adelaar and Muysken (2004: 396) mention -pala, -pud, -big, and-shi (these and specifically Cañar toponymic endings will be discussed further in the following Sect. 5.2). On the other hand, toponyms also share some phonological characteristics, including initial and final voiced obstruents and a voiced palatal fricative written . Tokens do not always reflect proto-Quechua *ll as in Ecuadorian Quichua, but appear to derive from a pre-Quechuan language given that they are found in nonQuechuan placenames (the same is the case, as we will see in Sect. 5.4, for substrate vocabulary in Cañar Quichua). Cañar and Puruhá also share an unfortunate history of misguided classification attempts. Unlike the bona fide Barbacoan association for more northerly languages like Cara, Cañar, and Puruhá have been related, inter alia, to Mochica (Jijón y Caamaño, 1940: 412–413, 1941: 35). However, there is no reason for any faith in these ideas, as they rest on very few and poor comparisons (Urban, 2019). To the south of the former Cañar-speaking area, the province of Loja was characterized by a widely distributed language known as Palta, which also extended in the west to Zaruma in the province of El Oro, in the east to parts of Zamora Chinchipe province, and to the area of Jaén in present-day Peru. Another language, known as Malacato, was spoken just to the south of the town of Loja (both are not plotted on the map in Fig. 2.1). Palta data are restricted to four words from the Relación de la Tierra de Jaén (Anonymous, 1965 [~1570]), which is contained in the Relaciones Geográficas. These suggest that the language belonged to the Chicham (Jivaroan) family (Gnerre, 1975; Torero, 1993). Toponymy also strongly supports a Chicham presence in the southernmost regions of the Ecuadorian Andes (Gnerre, 1975; Gómez Rendón, 2016).

6

Gómez Rendón (2017: 61), however, says, that Cañari, Puruhá, and their southern neighbor Palta were all different from one another.

78

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

This case study concerns Quichua as spoken in the southern Ecuadorian provinces of Cañar and Azuay and expands on an earlier study (Urban, 2018). The population history and thereby the sociolinguistics of language contact and language replacement in the Cañar-Azuay region has some commonalities with what happened in other parts of Ecuador, but also some particular local characteristics. Local people of the Cañar-Azuay region and “small parts of El Oro, Loja, Guayas, and Chimborazo” (Hirschkind, 1995: 316), thus including parts of the western and eastern slopes of the Andes (Renard-Casevitz et al., 1988: 213), were, before the impact of the Inca, linked by cultural traditions and language but were organized into a set of relatively small-scale chiefdoms without higher levels of political integration. On the basis of archeological evidence, Renard-Casevitz et al. (1988: 213) suggest ethnic and/or dialectal differences between people of the Cañar valley in the north and people of the Cuenca depression, including the environs of Gualaceo, Chordeleg, and Sigsig in the south. As other substratal languages of the Ecuadorian highlands, the Cañari language is undocumented. However, the Relaciones Geográficas for the relevant parts of the Andes describe the region and provide etymologies for several placenames and one personal name. These include Guapdondelic ‘great plain like the sky’ (Pablos, 1965 [1582]: 265), Paute ‘stone,’ Maras ‘rocky hill’ (de Pereira et al., 1965 [1582]: 271–272), Copsi ‘hill of rough and sandy rock,’ P(u)eleusi ‘yellow field,’ Guichannauto ‘heavy head’ (de Gallegos, 1965 [1582]: 274–275), Leoquina ‘lagoon of the snake,’ Tamalaycha ‘river which eats the Indians’ (a variant spelling found elsewhere is Tamalannecha) (Arias Dáuila, 1965 [1582]: 278, 280), Chunchi ‘burns’ (de Gaviría, 1965 [1582]: 285), and Alusi ‘cherished and valuable thing’ (Italiano, 1965 [1582]: 288). As Adelaar and Muysken (2004: 396) note, the explanation for Guapdondelic is particularly interesting, since it allows an interpretation of other toponyms in -del(eg), which are characteristic of the Cañar-Azuay region. The Cañari people were affected particularly badly by the expansion of the Inca empire to Ecuador. The Inca struggled hard to incorporate the Cañari region into their realm involving “a series of advances, failures, and reconquests” (Hirschkind, 1995: 320). After they achieved control at least in the most densely populated regions, the Inca resettled Cañari people to other parts of the empire and also established loyalist foreigners in Cañari territory; both measures served to pacify and consolidate the region (Hirschkind, 1995: 321–323). Later on, the Cañari became entangled in the internal struggle between Huascar and Atahuallpa for succession of their father, Inca Huayna Capac, to the Inca throne. Huascar had established a power base in the newly conquered northern territories of the empire, in what is now Ecuador. Having forged an alliance with the Cañari people specifically, he established his seat at the town of Tumipampa in Cañari territory (Hirschkind, 1995: 314).7 After defeating Huascar, Atahuallpa, remembering the prior alliance of the Cañari and 7

Pablos (1965 [1592]: 265) actually says that Tomebamba is the Quechuan translation of Guapdondelic. This could only be true if his explanation for the Cañari placename is inaccurate: tumi is the name for a ceremonial knife used in the pre-Colombian cultures of the Central Andes. It appears in Carranza Romero’s (2003) dictionary of Ancash Quechua and also in the Cuzco Quechua dictionary of the Academia Mayor de la Lengua Quechua (2005), but with the additional information that it is a historic term. Given that tumi appears in Bertonio’s (1612) dictionary of Aymara and is still in

5.1 Introduction

79

his half-brother and competitor, “killed as many of the men and children among them as he could and destroyed Tumipampa” (Hirschkind, 1995: 314). The Cañari population thus collapsed, and the chronicler Cieza de León, when passing through Cañari territory, reports a relationship of women to men in the order of fifteen to one (Hirschkind, 1995: 315). Given the catastrophic impact of the conquista on the Indigenous population generally (now confirmed also by genetic evidence reported in O’Fallon & Fehren-Schmitz, 2011), population decline is likely to have continued after the Spanish conquest from the already extremely low levels that resulted from the bloodshed which Atahuallpa reportedly instigated; reasons include epidemics, conscription to Spanish military campaigns, and excessive labor. By the end of the sixteenth century, the population may have declined by 90%, and only 5000 Indigenous people may have survived (Alcina Franch, 1980: 7–8). At the same time, because of the need for laborers in the Spanish mines and the already depleted local population, the influx of people from elsewhere continued, too. Hirschkind (1995) argues that Cañari as an ethnic identity only emerged in response to the unifying pressures brought onto the chiefdoms of the Cañar-Azuay region by the Inca threat, an identity that survived the population collapse in late prehistoric and early historic times and that, in fact, later also appears to have functioned as a banner under which the diverse resettled Indigenous people brought to the region by Inca and Spaniards could rally. Present-day ideas among people of Indigenous heritage to the effect that “they are the descendants and cultural heirs of a fierce, rebellious people who fought against Inca imperialists may appeal to the imagination but does not stand up under historical scrutiny” (Hirschkind, 1995: 314). One of the emblems of present-day Cañari identity are some particular characteristics of Cañar Quichua. It is somewhat ironic that characteristics of the language that was brought to Cañar and Azuay by a hostile empire serves as a condensation point for a regional Indigenous identity. However, it readily finds its explanation in the fact that the characteristics that are relevant for the assertion of local identity in Quichua are in fact attributed to the agency of the extinct original language of the Cañar region (Howard, 2010). This ideology actually has a factual basis: right at the onset of modern Quechuan studies, Torero (1964: 464–465) noted phonological and lexical peculiarities in Cañar Quichua which he attributed to the influence of an unidentified pre-Quechuan language, and it is the same features which (some) present-day speakers of Cañar Quichua refer to when asserting their heritage. These will be discussed in detail in Sect. 5.3.

use today as the word for ‘knife’ in Bolivian Aymara (Huayhua Pari, 2009), an Aymara origin is likely.

80

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

5.2 Toponymy 5.2.1 Characteristic Toponymy of Cañar and Azuay The toponymy of Cañar and Azuay is very salient and characteristic. In the first place, toponyms in -pí ~ -bí ~ -pi ~ -bi, the hallmark of a former Barbacoan presence in more northerly region of the Ecuadorian Andes, are virtually absent here (Fig. 5.1).8 At the same time, the region features a number of endings that characterize it positively. Most notable is an ending that has been postulated to have the form -cay (Adelaar & Muysken, 2004: 396). This extremely frequent sequence (Fig. 5.3) is often featured in names for rivers or canyons, so that it has been assumed to mean ‘river’ or ‘water’ (Paz y Miño, 1961: 200). Under this assumption, one can say that -cay replaces -pí and variants in the Cañar-Azuay region (see already Wolf, 1892: 504 to this effect).9 Note, however, that there are also some Colombian tokens of toponyms ending in -cay (Fig. 5.3). Also here, all denote rivers or river canyons in an area where -pí ~ -bí ~ -pi ~ -bi occurs, too. We will return to this fact later.10 Adelaar and Muyksen (2004: 396) mention -del(eg), -copte, -huiña, -turo, -zhuma, and -zol as characteristic Cañar toponymic endings in addition to -cay. The sequences -del(eg), -zhuma, and -zol showcase voiced obstruents, which appear spelled as . These segments characterize Cañar and Azuay toponyms generally. As -del(eg) shows, voiced obstruents can occur at the beginning as well as end of toponymic endings, and further placenames like Zhud show that the same is true also at the level of the word. None of the mentioned endings have Quechuan etymologies, and neither do Zhud and other placenames with the mentioned characteristics. There are some hybrid toponyms, though, which feature a Quechuan ending. An example is Zhutapampa, in which zhuta corresponds to a local (probably Cañari -derived) word for ‘bird’ (Howard, 2010: 142) and -pampa reflects Quechua pampa ‘plain.’ In the absence of phonetic analysis, it would be adventurous to assign precise phonetic realities to segments spelled in the toponymic record. Recordings of such names as pronounced by speakers of Cañar Quichua would be a desideratum, as their pronunciation, perhaps even by speakers of Spanish, may retain phonetic characteristics of the substratal pre-Inca language rather faithfully. We may tentatively assign labial, alveolar, and velar places of articulation for ; the phonology of Cañar Quichua, with which we will deal in some more 8

Relevant Cañar and Azuay tokens include Quebrada Mishquiapi, which, however, seems to be of a curious Quechuan derivation, meaning “sweet porridge.” Hacienda Cachipambi, the second token, might perhaps be a typo for Cachipamba, which would be interpretable as Quechuan ‘salt plain.’ A third token is Murucu Limpi. 9 For reasons of brevity and ease of reference I will use the phrase “Cañar -Azuay region” as a shorthand for the area in which the substratal Cañari language was once spoken. Note however, that in reality, this area also includes adjacent parts of eastern El Oro and northern Loja, which should be thought of as forming part of the “Cañar Azuay region,” too. 10 A rather large number of toponyms involving Spanish material in combination with the indigenous placename Micay were presented in Colombia. Such doublets are not shown on the map.

5.2 Toponymy

81

Fig. 5.3 Ecuadorian and Southern Colombian toponyms in -cay ~ -cáy

detail in Sect. 5.3, furthermore suggests that they are stops. However, it would be somewhat premature to assume that this manner of articulation maps directly onto the articulation of placenames, not least because of the allomorphic relationship between voiced stops and fricatives in Spanish. As far as and are concerned, it is probably safe to interpret as a voiced alveolar fricative and as a voiced postalveolar or palatal fricative [ʝ ] or [Z]. Cañar toponymy has been further investigated together with other Sprachdenkmäler, in particular the few words that can be extracted from the Relaciones Geográficas, by Paz y Miño (1961). This author identified recurrent sequences in (orthographically represented) Cañar and Azuay placenames and grouped them into 282 sets according to common initial or final sequences.11 However, it is not clear if sequences identified by Paz y Miño actually correspond to morphs in the preQuechua Cañari language (compare Salomon & Grosboll, 1986: 388). For instance, there seems to be no principled reason why the toponym Ayacáña is identified as featuring an ending -áña, while Dágña, Núgña, and Tugña are grouped into another set on grounds of their ending -ña. Nevertheless, Paz y Miño (1961)’s work can still be thought of as a starting point for further investigation of the wider distributions of toponymic endings which occur in the Cañar-Azuay region. 11

Paz y Miño (1961)’s sets feature some elements that can clearly be attributed to Quechuan, e.g., -cocha or -pirca, or even Spanish, e.g., -loma, rather than the pre-Quechuan language of the region.

82

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

Following up on Paz y Miño’s work reveals that it indeed correctly identified a number of toponymic endings that are, judging from the GEOnet database, restricted to the Cañar-Azuay region or exceedingly rare and/or scattered elsewhere. This distribution makes them plausible candidates for Sprachdenkmäler of the Cañari language. Since some of these endings do not enjoy a high token frequency in the GEOnet database, however, it is sometimes difficult to say to what extent they can be considered characteristic or even diagnostic, in particular if there are scattered— presumably unrelated—occurrences outside the Cañar-Azuay region. I follow the rule of thumb that tokens outside the area in which the Cañari language is thought to have been formerly spoken should not cluster together in any obvious way, and their number should generally not be higher than 50% of the tokens inside that area. I caution, however, that spatial statistical analysis should be carried out to double-check the validity of the identified patterns statistically. Endings that appear to be characteristic of the region include -car, -cay, -ceo (which may or may not be related to -eo as in Panzaleo), -chún (with an outlier near the Pacific coast), -del ~ -deleg, -lel, -llapa ~ -zhapa, -nar ~ -ñar, -pali (which is most strongly represented in El Oro), -quir, -say, -sho ~ zhó, -uro (which is most frequent in Northern Loja), -zhun, and -zhuma.

5.2.2 Continuous Northward Extensions of Cañar and Azuay Toponymic Endings While sequences like -cay, as Fig. 5.3 shows, are very strongly associated with Cañar and Azuay, some toponymic endings that occur there extend to the north into the area where colonial sources report Puruhá. Adelaar and Muysken (2004: 396) mention -pala, -pud, -bug, and -shi as endings that are shared between the two areas. Paz y Miño (1961) also discussed these.12 In the GEONet database, -bug occurs only four times, and that in a tight cluster in the former Puruhá stronghold in the Riobamba region (Fig. 5.4). The ending -pala has a similar distribution, though there is one token in the province of Cañar, and there are also outliers on Puná island, in Loja, and in Colombia (Fig. 5.5). The proper interpretation of this observation is not clear; it may have something to do with de Gaviría’s (1965 [1592]: 286) reports for the Relaciones Geográficas to the effect the people of Chunchi (in the present-day Chimborazo province) “speak the general Quichua language of the Inka, others their particular language, which is the Cañar language of the province of Cuenca, and partly mixed up with the 12

Paz y Miño (1942) has made similar observations and offers a complicated interpretation of the facts. He claims not only to be able to infer that Puruhá was notably different from Cara and Panzaleo, but also that there is another language, “Makají,” underlying Puruhá, whose speakers were related to the people of Azuay before the immigration of the Cañari. This he deduces from the observation that Puruhá and Cañari share toponyms with a certain structure, namely ones ending in -ud which is sometimes preceded by a set of different consonants and sometimes is not. These toponyms he assigns to this hypothetical older “Makají” linguistic stratum.

5.2 Toponymy

83

Fig. 5.4 Toponyms in -bug ~ -búg

language of the Puruhuayes of the province of Riobamba.”13 He also mentions other languages being spoken by the same people. In this case, the simple coexistence of two languages in geographical space on a local level may be part of the explanation for the toponymic overlap (though, as we will see shortly, this cannot be the whole story). It is also relevant that the well-known call for materials in the languages of the diocesis of Quito from the late sixteenth century assigns one and the same cleric, Gabriel de Minaya, the task to provide translations of the catechism and confessionary for both Cañari and Puruhá (López de Solís, 1596). Were they very closely related languages, so that the same individual could easily master both? But there are wider questions. According to Adelaar and Muysken (2004: 396), on the one hand, the ending -pud characterizes the Cañar-Azuay and Puruhá regions in southern Ecuador, but the same authors (2004: 394) note on the other hand that 13 “Hablan la lengua general quichua del Inga los más la lengua particular dellos, ques la cañar de la provincia de Cuenca, y en partes revuelta con la de los purguays de la provincia de Ríobamba.” Renard-Casevitz et al. (1988: 213) suggest that, west of the Nudo del Azuay, the Alausí basin may have hosted a mixed bilingual population. The same region may also have been a junction of interregional trade.

84

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

Fig. 5.5 Toponyms in -pala

it also suggests “some kind of relationship” between the areas in which Pasto and Cara were once spoken in Northern Ecuador. As Adelaar and Muysken (2004: 397) suggest further on, such broader similarities show that the toponymic record of Cañar and Azuay can be integrated into a broader Ecuadorian context. Toponyms in -shi, as Fig. 5.6 reveals, indeed occur in the Cañar-Azuay region as well as to the north in the area associated with the Puruhá language. However, like a chain following the course of the Andean cordillera, they also extent further northward into northern Ecuador.14 Other toponymic endings can be described as following a distribution similar to that of -shi, showcasing representative tokens in the Cañar-Azuay region, but with further tokens extending northward into the central and northern Ecuadorian Andes. Among the particularly interesting sets in this regard is -qui ~ -quí. As Fig. 5.7 shows, this ending is present in Loja and has a high density in eastern Azuay, but there are some tokens in the more central parts of the Ecuadorian Andes, namely in 14

There is also a strong presence of this ending on the Guajiro peninsula of which, however, it seems safe to say that it is unrelated to the Ecuadorian toponymic set and more likely associated with the Arawakan Wayuu (Guajiro) language spoken there.

5.2 Toponymy

85

Fig. 5.6 Toponyms in -shi

Chimborazo, Tungurahua, Morona Santiago, Cotopaxi, and Napo. However, there is a particularly dense cluster of relevant toponyms in Pichincha and Imbabura that extends eastward into Sucumbíos. The toponym Tarqui, which occurs in Azuay and Morona Santiago, also pops up on the coast of Santa Elena. In southern Colombia, tokens are more scattered again and include another Tarqui (note that there are also Colombian tokens further north—I again follow the rule of thumb that tokens in central and northern Colombia should not cluster together in any obvious way, and their number should generally not be higher than 50% of the token in Ecuador and southern Colombia). Another relatively clear instance of the “-shi distribution” is observed for the ending -pata (Fig. 5.8), which has one representative token in northern Loja, many in Azuay and Cañar, and also some in Tungurahua, Cotopaxi, and Pichincha. There are no tokens in Imbabura, Carchi, and southern Colombia. Note that hybrid toponyms involving linguistic material from Quechuan (Cochapata, Cachipata) as well as from Spanish (Naranjapata, Manazanapata) shows that the language that gave rise to these toponyms survived the Quechuan spread and onwards into colonial times.

86

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

Fig. 5.7 Toponyms in -qui ~ -quí

Another quite salient ending with a distribution akin to that of -shi is that of -ig or -eg (seen in Fig. 5.9). This ending occurs once in El Oro, frequently in Azuay, especially in the northeast, and further north in Cañar, Chimborazo, Bolivár, Cotopaxi, and Pichincha. As is obvious these include tokens of the salient ending -deleg, which, however, is perhaps composite. Also very prominent, especially again in eastern Azuay, is the ending -apa (Fig. 5.10). Tokens occur in Loja and adjacent parts of El Oro, Azuay, Tungurahua, and Pichincha. In Ecuador, outliers are found in Santa Elena, Morona Santiago, and Sucumbíos. The southernmost Colombian token occurs in the department of Huila, and there are rather many tokens scatted throughout central and northern Colombia. A particular cluster is found in the Guajira Peninsula, which is likely associated with the Wayuu language. It is interesting to note that three of the mentioned toponymic sets, those in del, -qui ~ quí, -and apa, show a skewed distribution within the Cañar-Azuay area. They cluster in western and southern Azuay (mainly the Girón, Nabón, Oña, Sigsig, Chordeleg, and Gualaceo cantons). This clustering may reflect dialectal differences

5.2 Toponymy

87

Fig. 5.8 Toponyms in -pata

already suggested by Renard-Casevitz et al. (1988: 213) on the basis of archeological evidence. Other toponymic endings with a distribution that is similar to that of -shi include -tus ~ -tús, which occurs eight times from southern Loja via Azuay, Chimborazo, and Pichincha into the Colombian department of Nariño15 ; -rin ~ rin (18 tokens), which is strongly represented in Azuay, but also found in Cañar, Tungurahua, Pichincha, and Imbabura16 ; -run ~ -rún, with four tokens, one in Cañar, one in Chimborazo, one in Imbabura, and one on the coast of Nariño; -quil, with seven tokens, occurring in Loja, El Oro, Azuay, Guayas, Chimborazo, and Pichincha; -ñag, with five tokens, represented in Cañar, Chimborazo, Tungurahua, and Sucumbíos; -lan ~ lán ~ -llán ~ -llan, which is found 27 times all in all, in Loja, Zamora Chinchipe, Azuay, Cañar, Chimborazo, Tungurahua, Cotopaxi, Pichincha, and Imbabura (there 15

There also is a Río Piatus in Napo, though this appears to be a clerical error for intended “Río Piatua”. 16 In the Colombian department of Tolima, the ending occurs in the names of gorges, rivers, and mountain crests called Cambrín and Sincerín. There also is a place called El Jadrín, which may be a clerical error for Spanish jardín ‘garden.’ There are scattered tokens of Sincerín and also one Circerín in Northern Colombia; these may have a Spanish origin, too, and there is a Río Tzurin in the Ecuadorian lowland province of Morona Santiago, a name which may have a Chicham origin.

88

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

Fig. 5.9 Toponyms in -ig ~ -eg

are also tokens in Guayas, Santa Elena, Manabí, and Pastaza); -iña, with 14 tokens in Eastern and Northern El Oro, Northern Loja, and a dense hotspot of occurrences in Azuay, Chimborazo, Zamora Chinchipe, and western Napo17 ; -huín, with five tokens distributed over northernmost Azuay, Cañar, Chimborazo, and Tungurahua; -gan ~ gán, represented five times in Azuay, Cañar, Chimborazo, Imbabura, and southern Nariño; and finally -aló, with 42 tokens overall, which occur in Loja, Azuay, Chimborazo, Morona Santiago, and then densely in Tungurahua and Cotopaxi. Further, tokens are found in Los Ríos, Pichincha, Imbabura, and Sucumbíos. In Colombia, there is a token in western Nariño, and further ones in Cauca, Valle del Cauca, Tolima, and Chocó.

17

There is one token in Colombia’s Huila department, and another one much further to the east in Southern Colombia, in Caquetá; there are further scattered tokens still further east and north.

5.2 Toponymy

89

Fig. 5.10 Toponyms in -apa

5.2.3 Noncontinuous Northward Extensions of Cañar and Azuay Toponymic Endings to the Ecuadorian-Colombian Border There are also toponyms which show a different distribution outside of Cañar and Azuay. The most prominent representative is the fourth ending which Adelaar and Muysken (2004: 396) mention as shared between Cañar and Puruhá: -pud. As Adelaar and Muysken (2004: 397) and Paz y Miño (1942: 54) before them have noted, -pud is an ending that recurs, skipping the intermediate Panzaleo-area, in the once Pasto and Cara-speaking region in the northern Ecuadorian highlands, and also in the area where Awa Pit is spoken until today. Data from the GEOnet Names Server (Fig. 5.11) confirm this (there is one relevant toponym between these two areas, near Riobamba, and in fact, Jijón y Caamaño, 1940: 445 is able to cite some more tokens from the Puruhá area).18

18

Pichipud, Chachupud, Chillapud, and Sasaput.

90

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

Fig. 5.11 Toponyms in -pud

The dense cluster of occurrences of -pud in the Awa Pit-Pasto area suggests an association of this toponymic ending with these languages specifically, and thereby also the Barbacoan languages more broadly.19 We will return to this shortly. There are also toponyms with a similar distribution, i.e., which likewise display discontinuous occurrences in Ecuador, being found in the Cañar-Azuay provinces on the one hand and in the highlands of the Ecuadorian-Colombian border region on the other. Also here, the cautionary remark that the density of most of the mentioned endings is usually rather low is in order; sometimes, there are just few occurrences. This makes it difficult to assess the significance of the distribution, especially if the picture is confounded by scattered tokens outside of the areas of interest. Note also that the distinction between -shi and -pud toponyms is not always clear-cut. In the discussion 19

The apparent association of this ending with the Pasto language has also been noted by Jijón y Camaaño (1940: 199–217). He mentions Cayalapud, Cualapud, Cuaspud, Chalipud, Muelambud, Nispud, Palampued, Pialapud, Pispud, Puchimbud, Tanyarbud, Tembud, Tepud, Ujampud, Yanguibud, Yurambud as placenames (or names of huacas, which likely are natural landmarks) that go back to the Pasto language (as well as Casbuepud, Coambued, Cuasapud, Chalapud, Chalpued, Guachagpud, Guelpud, Ipialpud, Malambud, Malapud, Malpud, Puergualpud, and Taimbud, which are personal names).

5.2 Toponymy

91

of toponyms with “-shi-like” distribution, for instance, we have noted that between the Cañar-Azuay and the Carchi-southern Nariño area in Northern Ecuador and southern Colombia, the central Ecuadorian highlands sometimes feature more isolated occurrences of the toponymic endings. Also here, low token frequencies exacerbate the problems. Take the ending -gan ~ -gán, whose distribution is plotted in Fig. 5.12. We have two tokens in the Cañar-Azuay region and also two tokens in Northern Ecuador and southern Colombia. The ending -gan ~ -gán can therefore be thought of as having a distribution similar to that of -pud. However, there is one token, a mountain called Tungán, in between, in the province of Chimborazo. For this reason, the distribution of -gan ~ -gán can also be thought of as having a continuous distribution, more similar to that of -shi. The assignment of the ending to one of the major patterns here, thus, hinges on just one placename. Another set with even lower token frequency is that characterized by the ending -tur ~ -dur. This occurs in Azuay but is particularly interesting because Mandur—which is the name of a river there—has a doublet in the Río Mandur, which flows in the lowlands to the east of Pasto through the Putumayo department to feed into the Caquetá (another possible Colombian token of -tur ~ -dur is Verdur, a place in the Meta department). Even more frequently attested toponymic endings, such as -qui ~ quí, have a lower density of occurrences in central Ecuador, where the Panzaleo language that proved so difficult to classify for early researchers such as Jijón y Camaaño (1943) was once spoken. Viewed against this background, difficulties in differentiating between “-pud-like” and “-shi-like” distributions may not only be a practical nuisance, but symptomatic and reveal something important about the distribution of toponyms in the Ecuadorian highlands and, by proxy, preQuechuan substratal languages. At any rate, they highlight the fact that the categories of “-pud-like” and “-shi-like” distributions of toponymic endings in the Ecuadorian highlands are artificial constructs in the first place and only serve heuristic purposes in that they help to order a much more complex reality (in which, of course, the distribution of each toponymic ending is different from any other). A particularly interesting ending with a “-pud-like” distribution is -quer. As has frequently been noted, this is a recurrent ending in the Colombian-Ecuadorian border area (Fig. 5.13). Wolf (1892: 504) cites “Cuaiquer, Yaguanquer, Mayasquer, Pupusquer, Piquer, Chunquer, Túqueres, Ipiales, Pupiales, Imues, Supuyes, Funes, Iles, Males, etc.,” for the Pasto region, while Jijón y Caamaño (1919: 342–343) adds Huaquer and Tulcanquer. From the Panzaleo area, Jijón y Caamaño (1940: 343) mentions the isolated case of Buborker. Cuaiquer, in fact, has a special connection with the Awa Pit language in that it is a name used in older literature for the language and its speakers (Curnow, 1997: 1). For this reason, the ending -quer has figured prominently in earlier discussions of the Pasto and Awa Pit languages and their (former) speakers. What is interesting in the present context is that the ending -quer recurs sparsely in Azuay, where a Río Chúquer is found, as well as Northern El Oro, very close to the border with Loja, where a place called Sequer is attested in the GEOnet database (Fig. 5.13). The ending -quir, identified as characteristic of Cañar-Azuay, may be a local variant. Further cases of toponymic sets with distributions similar to that of -pud include that defined by the ending -cán, which is found in the Cañar-Azuay and Puruhá area

92

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

Fig. 5.12 Toponyms in -gan ~ -gán

eight times, but also in the name of Tulcán, the name of a town in the EcuadorianColombian border area that was already mentioned in Sect. 5.1 as formerly Pastospeaking, as well as three other names of that region (note, however, that there are some toponyms in -cán scattered in Ecuador and Colombia as a whole, which include other instances of Tulcán). Very much the same distribution is, as we have already seen, observed for -gan ~ gán, though the low token frequency and the occurrence of Tungán in Chimborazo makes it difficult to interpret this set. The distribution of -dán is somewhat narrower in Ecuador, restricted to the Cañar-Azuay area, where it cooccurs with clear markers of Cañari heritage such as initial in the name Zhordan (as well as in three further toponyms). In Colombia, conversely, the distribution is somewhat broader: While there is a Hacienda Indan in the area where Awa Pit is spoken, there are places called Puldán and Quendán further west, and there is an extension into the Pacific-facing lowlands, where the typical Barbacoan ending -pí ~ -pi ~ -bí ~ -bí figures strongly, with the name of the Laguna Pilcaran.20 Leonan and Ganán are toponyms found in Cañar and Azuay, respectively, but the ending -nán ~ -nan also occurs in the names of the Paramo de Pinan in the north of Ecuador’s 20

There is also a Quebrada Albadán in the Huila department and El Sudan and Topundan in Bolívar.

5.2 Toponymy

93

Fig. 5.13 Toponyms in -quer

Imbabura province as well as the names Panán and Chavísnan in the Colombian department of Nariño near the Ecuadorian border (toponyms ending in -nan also occur four times in other parts of Colombia). The ending -tan ~ -tán is present twice in the Cañar-Azuay area, and four times in the Carchi-Nariño border area associated with the Pasto and Awa Pit languages. In between these, one finds three tokens in the central Ecuadorian highlands of Cotopaxi and Tungurahua. The frequency with which toponymic endings of the shape -Cán or -Can (where C is any consonant) recur may suggest that we are rather dealing with a more generalized ending of the shape -án. Exploration of Barbacoan lexical material would be necessary to assess to what extent different sequences of the shape -Cán ~ -Can are etymologizable; it is possible that some of them turn out to be phantoms. But there are also endings with different but similar shapes that are found both in the Cañar-Azuay area and the Ecuadorian-Colombian border area where Awa Pit is spoken. The ending -cún, for instance, occurs both in the Cañar-Azuay and Puruhá areas, but there also is one token, Piacún, just to the north of the EcuadorianColombian border.21 The ending -ñag has a thin distribution extending from the Cañar 21

Note the placename Suescún in an area much further north, in which the Duit language, apparently a close sister of Muisca, was once spoken.

94

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

province into Tungurahuá, but a Quebrada Simiñag is found on the Ecuadorian side of the border with Columbia in Sucumbios province, north of Nueva Loja. In Cañar, furthermore, one finds two placenames characterized by the ending -tar, and there is also a token in Zamora Chinchipe. There is one potentially relevant placename, Guastar, in Nariño (as well as a Güíntar west of Medellín, much further north). A final potentially relevant toponymic set is that characterized by the ending -uay, as in Azuay and three other toponyms of the Cañar-Azuay area, with an eastern outlier, Camaguay, in coastal Guayas. Outside southern Ecuador, such placenames recur in the name of the Quebrada Cachiuay (with a possible Quechua modifier kachi ‘salt’) as well as in the name of Maranguay in Nariño. There are relatively few further scattered tokens in Colombia.

5.2.4 Northward Extensions of Phonological Characteristics of Cañar and Azuay Toponyms So far, we have seen how a number of toponymic endings that figure in the CañarAzuay area recur elsewhere in the Ecuadorian and southern Colombian Andes, sometimes showing a continuous distribution, as does the ending -shi and others, sometimes skipping over the Central Andes of Ecuador, as does -pud and others. These distributions appear to link the Cañar-Azuay area with regions that are known or are thought to have once been inhabited by speakers of Barbacoan languages. What is more, the reappearance of the ending -quer, which is strongly associated with the Ecuadorian-Colombian border zone and the Awa Pit language, in CañarAzuay reinforces the idea of a Barbacoan extension to the southern Ecuadorian highlands. However, as discussed in Sect. 2.2, the clearest cases for the link of a toponymic set with a particular language or language family rest not only on the geographical distribution of that set, but require convincing etymologies for the relevant toponymic ending through that language or language family. In the case of recurrent endings, these should correspond to landscape generics to constrain semantic latitude. However, even this requires in depth-knowledge of Barbacoan lexis and morphosyntax that is unlikely to be achievable on the basis of the relatively restricted lexical and grammatical documentation, so that I shall here focus on another avenue of exploration. The geographical patterning of endings like -pud or phonological characteristics such as initial voiced obstruents may give some idea as to the affiliations of the preQuechuan Cañari language. However, unless plausible etymologies for toponymic endings—as in the case of river names in -pí ~ -bí ~ -pi ~ -bi through proto-Barbacoan *pii ‘river water’—can be proposed, their distribution may be suggestive, but they do not yet allow to make a strong case for a particular affiliation of the substratal language to which they go back. Likewise, salient phonological characteristics in toponyms are most suggestive if they can be linked to phonological and phonotactic patterns

5.2 Toponymy

95

in languages that are hypothesized to have been close relatives to the substratal languages on etymological grounds. In addition to tracing individual elements in toponyms etymologically, there is another way in which Cañar-Azuay toponymy can be subjected to comparative analysis: this is by way of its outstanding phonological characteristics, i.e., the highly frequent initial and final voiced obstruents. We have already seen that toponyms with such characteristics also occur in the Puruhá area. Voiced fricatives represented by , however, appear to peter out already here (Fig. 5.14). Once placenames like Cruzhuaicu (and orthographic variants) are identified as consisting of the Spanish word cruz ‘cross’ and an indigenous element huaicu (most likely Quechua wayku ‘valley, mudslide’) and discarded accordingly, placenames spelled with north of the Cañar-Azuay region are reduced to Cutzhualó in Cotopaxi and San Pablo de Uzhpayacu in the lowlands of the oriente, which latter apparently featuring Quechua yaku ‘water’ as its head. Other voiced obstruents continue into Central Ecuador to the area where the Panzaleo language was once spoken (Jijón y Caamaño, 1940: 295–386). Relevant toponyms to support this assertion include Buto, Buborker, Duizela, Dum, Dakilema, Gawan, Gilka, Guya, Kañad, Pintag, Yakchig, and Bungachig. As the examples show, voiced obstruents can occur in both initial and final position. Upon reaching the Cara Fig. 5.14 Ecuadorian toponyms with the orthographic sequence

96

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

area, we can observe similar characteristics, though here voiced obstruents appear to become more restricted to internal or final position, as in Cantacuced, Cuchumued, Pintag, Abagag (compare data in Jijón y Caamaño, 1940: 243–259). While in the GEOnet toponymic database, there are few tokens of initial in western Sucumbíos (Dantupamba, Dashiño), and there is a Gültig Alto in Pichincha, the majority pattern also here is that voiced obstruents occur finally, as in Purugulromed, Parogad, or Intag. And of course, we have seen placenames ending in -pud in the Pasto region, showing that these extend up to the Colombian border. In sum, while highly salient and distinctive, phonological characteristics of Cañar-Azuay toponyms are to varying degrees found further to the north but gradually become more restricted: voiced obstruents except are still found in the Puruhá region as well as, to a smaller extent, in the Panzaleo area, and final voiced obstruents are still encountered in Cara and Pasto. All highland Barbacoan languages—Awa Pit, Guambiano and Totoró—have rules that either voice or voice and spirantize stops. In Awa Pit, voicing occurs if the stop is found after voiced consonants and before voiced phonemes, and voicing and spirantization between voiced vowels or glides (Curnow, 1997: 23); however, sandhi phenomena can also convert stops in initial position into voiced fricatives (Curnow, 1997: 28). In Guambiano, /p, tʂ , k/ become voiced fricatives between vowels and word finally (but devoice again in utterance-final position) (Vásquez de Ruiz, 1988: 45). A language with such a rule of word-final voicing and spirantization could plausibly yield placenames with final . As demonstrated in Urban (2018: 440), a comparison of the non-phonemic transcription of Guambiano in Otero (1952) with the IPA transcriptions in Huber and Reed (1992) shows that the letters are typically used to represent the Guambiano stops in contexts in which they surface as voiced fricatives. Barbacoan languages also have relatively rich systems of fricatives. Awa Pit as described by Curnow (1997: 31) has a voicing opposition in fricatives, contrasting /s/ with /z/ and /∫/ (with an allophone [dZ]) with /Z/. In Guambiano, in contrast, voiced fricative phonemes are lacking, though /t, ts, t∫/ become voiced after voiced consonants (Vásquez de Ruiz, 1988: 46). However, while the phonology of Awa Pit and Guambiano opens intriguing avenues for understanding the phonological peculiarities of the toponymy of the Ecuadorian Andes, it is also clear that the languages that left the toponymic Sprachdenkmäler must have obeyed voicing and/or fricativization rules that are different from the two extant highland Barbacoan languages. In particular, Cañari must have featured rules that could yield voiced fricatives in any position of the word (or it must have featured such sounds as phonemes).

5.2.5 Evaluation of the Toponymic Evidence Adelaar and Muysken (2004: 397) observe that the pre-Quechuan languages of highland Ecuador “with the exception of the languages of Loja and possibly also Panzaleo … share a number of distinctive elements that seem to point at a common origin or a

5.2 Toponymy

97

long period of interaction at the least.” Features they mention are final -d and -g, the ending -pud, and “syllables with a complex labial onset” represented orthographically by a consonant followed by . Since Pasto shares these characteristics “a Barbacoan connection for this whole group of languages” is suggested. This study yields considerable support for these observations, and the statements can now be refined: for one, there is no need to exclude Panzaleo from the implicated set of languages. While toponyms with a distribution like -pud indeed show a characteristic gap in the central Ecuadorian highlands, others—those toponyms with a distribution like -shi— as well as voiced obstruents in initial and final position allow to integrate Panzaleo into the general picture. The distribution of -shi-toponyms, in fact, shows a continuity in the highlands of Ecuador more or less from north to south. Since relevant toponyms and phonological characteristics extend to the Cañar region, linguistic connections of the pre-Inca Cañari language with the highlands of the Ecuadorian Andes and with southern Colombia are suggested. The toponymic evidence, on the other hand, clearly also has regional peculiarities. This is shown by the more extensive presence of voiced fricatives in the Cañar-Azuay region, and also by toponyms that are peculiar to this area (e.g., -cay in hydronyms of the Cañar-Azuay region). Beyond the more or less generalized distribution of -shi-type toponyms, there are also more restricted similarities that point to closer connections of the southern Ecuadorian highlands of Cañar and Azuay with the Carchi-Nariño border area where the Awa Pit language is spoken, and Pasto was spoken. This is suggested by the distribution of toponyms of the -pud-type. Indeed, as Adelaar and Muysken (2004: 397) say, the fact that Pasto likely was a Barbacoan languages suggests a Barbacoan affiliation of the substratal languages of large parts of the pre-Inca Ecuadorian highlands. Since relevant toponymic evidence, however, is not restricted only to the area where Pasto was once spoken but also extends to the area of Awa Pit, the suggestion of a Barbacoan connection need not rely only on the sparse material available for Pasto. Rather, there are toponymic continuities with areas that are clearly associated with Awa Pit. Also the distribution of the ending -pí ~ -pi ~ -bí ~ -bi, which extends to large parts of the Ecuadorian highlands, but not significantly to the Cañar area, is good toponymic evidence for a presence of Barbacoan languages in an area that was once much wider than today, including regions that are now Quechuan-speaking. This toponymic set is particularly significant, because it can be linked plausibly to a lexical item that reconstructs to proto-Barbacoan and that is inherited in all known daughter languages. The Achilles heel of the other toponymic sets, including those that extend to Cañar and Azuay, is that, although their distribution extends to regions in which Barbacoan languages are still spoken today, they are still in need of Barbacoan etymologies. It is in fact conceivable that toponymic endings of the Ecuadorian highlands—with the probable exception of -pí ~ -pi ~ -bí ~ -bi—partially or entirely go back to languages that were not genealogically related to Barbacoan (note that this has in fact been claimed by Paz y Miño, 1942 for some endings, namely those he associates with “Makají”). But there are also other open questions. Why, for instance, is the toponymic record that suggests Barbacoan affiliations not stronger in the areas where the Southern Barbacoan languages Cha’palaa and Tsafiqui are spoken today, or around Popayán,

98

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

where the Nam Trik varieties of Guambiano and Totoró are spoken? Such imbalances may find their explanation in different toponymic “fingerprints” of different branches of Barbacoan—as noted in the introduction, this is, in spite of the rather modest geographical extension, not a particularly shallow language family. If the substratal languages of the highlands of Ecuador were particularly closely related to Awa Pit, something that would make sense on geographical grounds, the toponymy may most closely be associated with that of the area where Awa Pit is spoken, and where Pasto was spoken. And this is what actually is suggested by the distribution of individual toponymic endings that follow the -pud distribution. However, as just noted, the allophonic patterns of present-day Awa Pit are less likely to have yielded toponyms with the phonological characteristics observed throughout the Ecuadorian highlands. Guambiano as described by Vásquez de Ruiz (1988) comes closer. Another issue is that languages like Awa Pit as described by Curnow (1997) feature a rich array of voiceless and voiced fricatives, including /Z/. Yet, completely unlike what is observed in Cañar and Azuay, toponyms involving letters such as or are strangely absent from the area where Awa Pit is spoken. Could reflect a regional transcription habit that was only conventionalized in the spelling of names in the southern Ecuadorian highlands?

5.3 Substrate Effects As briefly mentioned in Sect. 5.1, Ecuadorian Quichua has to some extent simplified and/or reorganized the morphological apparatus it has inherited from proto-Quechua. For instance, the double-marked possessive construction, as, e.g., in Ayacucho Quechua qan-pa mama-yki ‘your mother,’ where the genitive marker -pa is attached to the pronoun root qan- and mama ‘mother’ bears the 2nd person possessive suffix -yki, has been replaced by kan-pak mama, where the possessive relationship is only marked on the pronoun (Muysken, 2009: 80, Table 1). Muysken (2009: 85) states that “[t]he question of possible substrate languages…influencing Ecuadorian Quechua remains open.” He notes that “the Barbacoan family … and the Jivaroan family … would be the most likely candidates,” but also that “[a] detailed comparison of the structural features of these languages and the specific traits of Ecuadorian Quechua still needs to be made.” While Zúñiga (2015: 33) believes that no strong case for Barbacoan agency in bringing about the special characteristics of Ecuadorian Quichua can be made, he notes at the same time that the nature of the changes “mostly point toward Barbacoan,” and a similar observation is made by Muysken (2021), who offers some of the more detailed comparisons he had called for earlier. Indeed, already Bruil (2011: 35) draws attention to the fact that the suffix -chun, “used throughout the Quechua family to express a wish or a third person command,” in Imbabura Quechua has evolved into a converb marker which indicates a difference of subjects:

5.3 Substrate Effects (1.)

99

yapu-chun

saki-wa-rka

plough-converb.ds

leave-1sg.obj- pst

‘(S)he left me so that I could plough.’ (Bruil, 2011: 33, glosses and spelling adapted)

This development, as Bruil (2011: 39) notes, could be motivated by similar constructions in Barbacoan languages which, as in the Imbabura Quechua example above, can also have the function of purposives. The following example is from Tsafiqui: (2.)

sona

mera-sa

ta-yo-é

woman

listen-converb.ds.fut

possess-conjunct- decl

‘I have it (radio) so that my wife can listen to it.’ (Bruil, 2011: 39, glosses and spelling adapted)

A similar construction, Bruil hypothesizes, could have existed in the substratal Cara language, which, as we have seen, could have pertained to the Barbacoan family. However, Bruil (2011: 41) also notes that future converbs are found in other varieties of Ecuadorian Quichua, too, including Cañar Quichua as described by Cordero (1992 [1895]). Bruil (2011: 41) ponders two different explanations for such possibly generalized contact-induced innovations: one relates to a possible influence of Imbabura Quichua on other varieties of Ecuadorian Quichua. The other is that “[i]t is possible that the Barbacoan languages were spoken throughout a larger area in Ecuador or that the indigenous languages in Ecuador were in rather close contact,” in which case Ecuadorian Quichua may be cast upon a structurally similar (Barbacoan) substratum throughout the Ecuadorian Andes. We know that the Inca expansion in Ecuador had three distinguishable phases: first, they annexed the lands up to the Nudo del Azuay in the north and consolidated their control there before attacking the central highlands and finally the north (RenardCasevitz et al., 1988: 206). The complete incorporation of Ecuador into the realm was a protracted process, and the region of Quito, for instance, was only pacified completely in the 1490s (Renard-Casevitz et al., 1988: 206). Therefore, the peoples of the northern region only lived approximately thirty or forty years under Inca control, whereas southern Ecuadorian peoples like the Cañari were subjugated almost an entire century, a difference that is reflected by different degrees of Inca acculturation in the north and south (Renard-Casevitz et al., 1988: 206). If, as Hocquenghem (2012) suggests, the spread of Quechuan to Ecuador was mediated entirely or partially by the imperial presence of the Inca, we may presume that Palta- and Cañari-speaking people were also acculturated linguistically and began learning Quechua as a second language first. Ecuadorian Quichua may accordingly have first experienced substrate features from these languages, which then later spread northward together with an already transformed Quichua (perhaps in a scenario that is very similar to the scenario for the gradual imbueing of Quechuan with Aymara features as it moved south). It is difficult to say if Cañar Quichua specifically showcases special morphosyntactic characteristics that could be explained by a Barbacoan substratum beyond those

100

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

found in Ecuadorian Quechua generally. This is because the grammatical sketches in Cordero (1992 [1895]) and Paris (1961) are quite basic and would not allow to easily pinpoint in any detail possible areas of the grammar that may have been remodeled on the basis of the template of a substratal language. While the grammatical peculiarities of Cañar Quichua remain little known, the Cañar Quichua lexicon is relatively extensively documented thanks to the full-scale dictionary elaborated by Luis Cordero (1992 [1895]). Cordero was the son of wealthy landowners of southern Ecuador and learned to speak Quichua from the peasants that were employed on his parents’ estate. In spite of shortcomings, Cordero’s dictionary is a valuable source because it extensively describes what clearly is a local Quichua lect of the Cañar-Azuay region. Another dictionary, elaborated by José María Lévesque and Gilberto Mejía and published as an appendix to Paris’s (1961) Quichua grammar, makes reference to Azuay and could contain additional data. However, these dictionaries do not employ a consistent orthography based on phonemic principles. For instance, the postnasal voicing characteristic of Quechua IIB is often reflected in the spelling of dictionary entries of Cordero (1992 [1895]), e.g., in ‘mushroom,’ but not consistently, cf. e.g., ‘sun’ (Torero, 1964: 465). Howard (2010: 134), however, clarifies that postnasal voicing is systematic and regular in Cañar Quichua, a fact that highlights the unsystematic nature of Cordero’s (1992 [1895]) orthography. On the other hand, Cordero (1992 [1895]: xxi) opted to generally prefer the letter over to reflect the perceived “sonorous” character of Cañar Quichua. Cordero’s (1992 [1895]) Cañar Quichua dictionary also makes use of the letters which are so salient in Cañar-Azuayan toponymy, i.e., . From Cordero’s (1992 [1895]: xxi–xxii) discussion, their phonetic value scarcely becomes clear, though we learn that is like Spanish but pronounced more softly, whereas corresponds to the “French” . These letters, in other words, represent [z] and [Z], respectively. Generally, thus, are by default best interpreted as [b, d, g, z, Z], respectively (Urban, 2018: 443), though recordings of Cañar Quichua speech from different localities would be desirable to establish the phonetic realities more firmly. That the characterization does not go completely wrong, however, is confirmed by Howard (2010), who reports on modern fieldwork on Cañar Quichua and makes special reference to these sounds. Howard (2010: 124) in particular mentions a voiced palatal fricative with phonemic status in Cañar Quichua which she describes as “foreign to the Quechua sound system.” Howard (2010: 131) also mentions “the voiced stops /b/, /d/, and /g/ and the voiced fricatives /zh/ and /z/, [which are] not considered to be phonemes in Quechua and which can therefore be assumed to belong to the sound system of a language other than Quechua.” As was briefly mentioned in Sect. 5.1, [Z] does have a potential internal source in proto-Quechua *ll (though Cerrón-Palomino, 2003 [1987]: 164 characterizes the resulting sound as alveolar rather than palatal). Furthermore, as Torero (1964: 465) observes, the affricate resulting from the merger of proto-Quechua */ch/ and */úʂ / undergoes sporadic postnasal voicing too, thereby yielding another source for orthographic in Cordero ’s (1992 [1895) dictionary. Furthermore, as Torero (1964: 466) observes, in Cordero’s (1992 [1895]) dictionary corresponds to an ejective aspirate in Cuzco Quechua in four items. How exactly these correspondences

5.3 Substrate Effects

101

arose—and why they arose in these instances, but not in others—is an unresolved matter, but it seems fair to say that possible internal sources for voiced stops and fricatives in Cañar Quichua appear to be relatively minor. As far as the other sounds are concerned, Howard (2010) is right in that they are not normally found in the phonological inventories of Quechua lects. Cañar Quichua appears to have acquired most of its voiced stops and fricatives through the adoption of lexical items from a substratal language (Torero, 1964: 464– 465). Parker (1969b: 155) likewise notes that these segments are “found mostly in nouns which name local flora and fauna and which have no cognates outside Ecuadorian [Quichua]” and adds that “evidently we are dealing with loanwords from a substratum language.” We will deal with this lexical layer in more detail in Sect. 5.4. Relevant for this Section is that by the introduction of voiced obstruents through non-Quechuan vocabulary, Cañar Quichua also acquired new phonemic voicing contrasts. Thus, for instance, piñan ‘kind of bush’ is distinguished from biñan ‘larva of vermin,’ tucu ‘kind of wall cupboard to store household appliances’ from ducu ‘a kind of tree,’ cuzu ‘larva of some insects’ from guzu ‘pool, swamp,’ sipi ‘hanging, strangling, conflict, pressure, compulsion’ from zipi ‘cracked,’ and shuru ‘crude basket made from flexible sticks to transport fruit’ from zhuru ‘pockmarked face’ only by the voicing contrast in the initial sounds, which hence must be assigned phonemic status (Torero, 1964: 465). Voiced obstruents have thus become part of the phoneme inventory of Cañar Quichua (as Howard, 2010 in fact states).22 Another phoneme that was added to the phonological inventory of Cañar Quichua is /l/, compare mulu ‘rustic clay plate,’ which has no Quechua cognates, versus mullu ‘bead, pellets of clay, glass, etc.’ (Torero, 1964: 465). As already noted in Urban (2018: 443), “[l]etters from the set in fact also show up, without apparent rules for their presence becoming obvious, in further items with Quechua etymologies, e.g., in ‘rough, coarse,’ which corresponds to, for example, Cuzco Quechua phatu ‘rough, double, voluminous’ (Academia Mayor de la Lengua Quechua, 2005).” Such phenomena could have two conceivable motivations: on the one hand, Cordero (1992 [1895]) may simply have misperceived the item and interpreted some ancillary phonetic detail of the word as pronounced by Cañar Quichua speakers as voicing. On the other hand, such tokens could also indicate that voiced phonemes spread from borrowed lexical items which lack a Quechuan etymology to native vocabulary, in a manner that is comparable to the spread of nonmodal consonants in Cuzco-Bolivian Quechua (see Blevins, 2017 for theoretical background). Recordings of actual Cañar Quichua speech and phonetic measurements would be as crucial to investigate this further as would be a more thorough examination of relevant tokens in Cordero’s (1992 [1895]) dictionary. In any case, it is noteworthy that the phonological characteristics of the local toponymy matches the characteristics of the items that are responsible for introducing the voicing contrasts. Ostensibly at least, they form a phonologically and 22

Parker (1969b: 155) notes that there is another way by which Ecuadorian Quichua dialects can acquire a voicing distinction, namely when voiceless stops after nasals re-enter “by pronunciation borrowing,” but this doesn’t seem to be involved here.

102

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

phonotactically homogeneous set with the same properties, so that it is plausible to assume that both derive from the same undocumented source. In the following section, we will have a closer look at these lexical items and what can be learned on their origins. Thus, Cañar Quichua words like biñan, ducu, guzu, zipi, zhuru, and mulu are mirrored by toponyms like Bulcay, Ducur, Gonzol, Zuñac, Zhuya, and Suscal. In fact, since toponyms and anthroponyms as proper nouns are part of the lexicon just like common nouns are, one can say that relevant phonemic contrasts in Cañar Quichua do not just rest on words such as biñan, ducu, guzu, zipi, zhuru, and mulu, but also on the names of the places in which Cañar Quichua is spoken. And given that toponyms and common nouns with voiced obstruents but without Quechua etymologies together form a phonologically and phonotactically homogeneous set with the same properties, it is highly plausible to assume that both derive from the same source: the undocumented substratal Cañari language. In the following section, we will have a closer look at these lexical items and what can be learned on their origins.

5.4 Substrate Vocabulary In the previous section, we have observed how new phonological contrasts were introduced in Cañar Quichua as the result of the influx of lexical items from a local language. Drawing on Urban (2018), this section provides a more detailed analysis of lexical items that do not have a Quechuan origin, together with an attempt to determine their etymologies. The database consists of the data from Cordero’s (1992 [1895]) pioneering dictionary, but also includes data from Cordero Palacios (1981 [1923]), Paris (1961), and Howard (2010). Cognates for items in these sources were sought mainly in Cuzco Quechua (Quechua IIC, Academia Mayor de la Lengua Quechua, 2005), Ayacucho Quechua (Quechua IIC, Soto Ruiz, 1976), San Martín Quechua (Quechua IIB, Park et al., 1976), Huallaga Quechua (Quechua I, Weber et al., 1998), and Inga (Quechua IIB, Tödter et al., 2002). Any Cañar Quichua item for which no cognates could be found in other varieties (and which thus satisfies Salmons’s, 1992 second criterion for the identification of substrate vocabulary) was extracted and subsequently compared with lexical sources of Barbacoan languages. These were Moore’s (1966) dictionary of Tsafiqui, Lindskoog and Lindskoog’s (1964) dictionary of Cha’palaa, and the wordlists contained in Huber and Reed (1992), which latter is the default source for the remaining languages, i.e., Nam Trik (Guambiano and Totoró dialects) and Awa Pit. Vásquez de Ruiz et al.’s (2009) basic dictionary of Guambiano and Otero’s (1952) wordlist of Totoró are further sources for these languages, as is Curnow’s (1997) grammar for Awa Pit. Some, but not all, of these vocabulary items feature voiced obstruent phonemes that have been argued to betray their non-Quechuan origin (compare Salmons’s, 1992 criterion (iv) as reviewed in Sect. 3.3). Also at this stage, no semantic criteria (compare Salmons’s criterion i) have been taken into account yet, a decision that aims to avoid unduly restricting the search space. The results of these comparisons are

5.4 Substrate Vocabulary

103

repeated here without change, apart from the correction of typos, from Urban (2018: Table 1) as Table 5.1. Items are ordered roughly according to the persuasiveness of the comparisons, strong ones appearing earlier in the Table than later ones. The Table features separate columns for all three consulted sources for Cañar Quichua so that, where an item is listed in more than one of the sources, different spellings and glosses become apparent. There is also a separate column which is used to indicate if the item in question is also attested in other sources of Ecuadorian Quichua, namely Stark and Muysken (1977), Ministerio de Educación (2009), and the appendices with flora and fauna terminology in Orr and Wrisley’s (1965) lowland Ecuadorian Quichua dictionary. Then follows a column with comparisons that can be made for the proto-Barbacoan level, with data from Curnow and Liddicoat (1998) unless otherwise noted, and finally five columns for the individual Barbacoan languages (or, in the case of Guambiano and Totoró, Barbacoan varieties). The rightmost column is reserved for comments. Explainable material in the cited forms that does not enter comparisons is placed between (parentheses). In contrast, extra material that does not enter comparison, but which cannot be plausibly explained, is presented in [brackets]. All in all, as Table 5.1 shows, there were 21 items for which Barbacoan comparisons could be found that were judged formally and semantically plausible, though not always without problems.23 There are six comparisons, items no. (1.) to (6.), that are particularly noteworthy because the corresponding forms in Barbacoan reconstruct to the proto-language. The semantic correspondence between ‘fog, drizzle’ and ‘smoke’ in item (1.) is highly plausible, as is evidenced by frequent cross-linguistic associations of the relevant meanings (Urban, 2012). Items (3.) to (6.) all appear to reflect proto-Barbacoan *tsik ‘tree, stick,’ which is combined with different modifiers to yield names for different types of trees and bushes. Note that there is a general tendency in Barbacoan languages toward the loss of final stops (Gonzáles Castaño 2014: 78–79). The process is completed in Southern Barbacoan, but the unreleased allophone in final position in Awa Pit, and the realization of /p/ in Nam Trik, which oscillates between a like unreleased allophone, a voiceless fricative, and indeed already zero (Gonzáles Castaño 2014: 58–59), shows that there is a family-wide drift. The consistency with which the names of different trees and shrubs in Cañar Quichua are built on the basis of the same recurrent morpheme -chi, which does not have a Quechuan etymology, but for which a plausible Barbacoan origin is readily available, suggests strongly that Barbacoan influence is at work also here in the southern Ecuadorian Andes (note, though, that the name pilchi is also attested as far as the Ecuadorian oriente according to information from Orr & Wrisley, 1965). How do the items relate to the voiced obstruents of Cañar Quichua and CañarAzuay toponymy? Item (1.) is interesting in that, to yield the Cañar Quichua forms, the vowel of the reconstructed form *i∫ has apparently been copied at the end of the word to fit the typically disyllabic Quechuan root canon. Something similar appears 23

In Urban (2018), the enumeration jumped from (6.) to (8.), leaving out (7.), and Vásquez de Ruiz et al. (2009) were erroneously cited as Vásquez de Ruiz (2009). This has been corrected here.

[chuchip]chi ‘kind of small tree’

[pil]chi

‘kind of tree’

‘kind of tree’

‘calabash tree’

(3.)

(4.)

(5.)

izhi ‘fog, very light drizzle’

mulu ‘rustic plate, made of clay and without adornment’

chiru ‘orangutan [sic!]‘

‘plate’

(8.)

nunchi

‘skin disease’

(15.)

palu

piri ‘light scabies’

pu[nya]-‘stink excessively, emit a nauseating smell’ piri ‘vile, despicable, mangy’

palu

palu

‘lizard’

‘smell, stink’

(13.)

(14.)

zzuyu

‘dusk’

(12.)

cuilan ~ cullan ‘small lizard’

‘knot’

(11.)

piri ‘light scabies, grain mold’

sutu, zzutu

‘monkey’ cuilan

chipu ‘locust’

‘lizard’

cuylan ‘small lizard’

Present in Ecuadorian Quichua?

yes

yes

yes

yes, with the meaning ‘cricket, grasshopper’

[pil]chi ‘vessel, junk made of yes coco, calabash’

izhi

Cañar Quichua (Paris)

(9.)

mulu

[pil]chi ‘kind of plant, Crescentia Cujete’

[chuchip]chi ‘kind of plant, Abatia Verbascifolia’

izhin ‘drizzle’

Cañar Quichua (Cordero Palacios)

(10.)

chipu ‘a kind of insect that jumps’

‘kind of shrub’

‘a kind of insect that jumps’

(6.)

(7.)

[pa]chi ‘kind of tree of the eastern highlands’

putu[l]

‘fog/smoke’

‘rotten’

(1.)

Cañar Quichua (Cordero)

(2.)

Gloss

No

Table 5.1 Cañar Quichua items with Barbacoan comparanda, reproduced from Urban (2018: 448–450, Table 1)

(continued)

*tsik ‘tree, stick’

*tsik ‘tree, stick’

*tsik ‘tree, stick’

*tsik ‘tree, stick’

*pɨ t(ɨ ) (own reconstruction)

*i∫ ‘smoke’

Proto-Barbacoan

104 5 The Barbacoan Languages and the Southern Ecuadorian Highlands

Totoró

pu[ba] ‘smelly’

(14.) , ‘lepra’ (Otero, 1952: 317, 310)

[lán]palo ‘common lizard’

(15.)

suyun ‘rainbow’

(13.)

pu[dyu] ‘smelling badly’

su’tú ‘inserted, put between’

(12.)

milán ‘plate, junk’

chi’pín ‘flea’

pe’ten ‘rotten’

Tsafiqui

(11.)

(10.)

(9.)

churi ‘monkey’

chijpi ‘flea’

(8.)

ñinchi ‘kind of tree’

[pa’]chi ‘kind of tree’

pete ‘rotten’

Cha’palaa

(7.)

kalun[c’i] ‘lizard’ (Vásquez de Ruiz et al., 2009)

(na’k)i ‘smoke’

(6.)

(5.)

(4.)

py[a]-‘to smell (bad)’ (Curnow, 1997: 149)

pi̵ t(i̵ t), compare pɨ t-‘to rot’ (Curnow, 1997: 98)

(2.)

(3.)

Guambiano

(nak)i∫ ‘smoke’

AwaPit

i∫ ‘smoke’

No

(1.)

Table 5.1 (continued)

(continued)

lan is ‘laziness,’ alluding to the lizard’s habit of basking in sunlight to thermoregulate?

Also in lowland EcuadorianQuichua(Orr & Wrisley, 1965);Tsafiquihas bilí ‘a kind of tree’

Comments

5.4 Substrate Vocabulary 105

tulu

cuzu ‘larva of some insects’

punzu

‘sack or bag’

‘vegetable waste/ grass’

‘larva/worm’

‘fine straw, tow-like rubbish’

‘small leaves, fibrous rubbish’

(17.)

(18.)

(19.)

(20.)

(21.)

pu[zha]

taba ‘rubbish consisting of branches and leaves which covers rocky regions’

palti

‘platform built in high place’

(16.)

Cañar Quichua (Cordero)

Gloss

No

Table 5.1 (continued)

tulu

Cañar Quichua (Paris)

pu[zha] ‘stalk, rubbish’

cussu ‘larva of beetle (?!)’ punzu ‘rubbish, tow, fine straw’

cuzu ‘thick worm, found especially in potatoes’

taba ‘scrub, place full taba ‘mosquito, rubbish, of scrub’ tangle’

tulu

palti ‘height, lookout (?)’

Cañar Quichua (Cordero Palacios)

yes

yes

yes

Present in Ecuadorian Quichua?

(continued)

*tɨ ‘firewood’

Proto-Barbacoan

106 5 The Barbacoan Languages and the Southern Ecuadorian Highlands

(Otero, 1952: 314), po (Vásquez de Ruiz et al., 2009) ‘straw’

(Otero, 1952: 314), po (Vásquez de Ruiz et al., 2009) ‘straw’

(20.)

(21.)

Guambiano

ku∫i ‘worm’

[tu̥ ] ‘higra (a shoulder bag)’ (Curnow, 1997: 43)

AwaPit

(19.)

(18.)

(17.)

(16.)

No

Table 5.1 (continued) Totoró

ku’∫i ‘worm’

pala ‘high, up’ (Vasquez de Ruiz et al., 2009)

tape ‘grass’

Cha’palaa

pu’chú ‘rest’

ta’pé ‘grass, bush’

toló ‘bag’

Tsafiqui

Comments

TheTsafiquiform could be a direct borrowing from Quechuan, with the donor form cognate withCuzcoQuechua puchu ‘rest’, etc. TheCañarQuichua item could reflect a related item with an additional nasal akin to Huallaga punsha ’trash, rubbish, debris, dried brush’ (Weber et al., 1998) orIngapunshu ’pedazo de machete’ (Tödter et al., 2002)

Possibly is from Proto-Quechua *kuru ‘worm’ (Emlen, 2017) . The phonological development of *r would be irregular, but the : /i/ correspondence would match forms in (7.) and (8.)

Note Aymaratapa ‘bird’s nest’ (Huayhua Pari, 2009), also inAyacucho(Apurímac) Quechua tapa ‘nest’ (Parker, 1969a)

palti is attested inLojawith the meaning ‘bed’ (Stark & Muysken, 1977)

5.4 Substrate Vocabulary 107

108

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

to have happened in item (15.). It is also interesting that the proto-Barbacoan fricative *∫ is reflected as the voiced fricative [Z] in the intervocalic environment that is thus created. In item (19), Nam Trik /∫/ corresponds to in Cordero’s (1992 [1895]) data, and to in Cordero Palacios’s (1981 [1923]) dictionary. Unfortunately, the items for which Barbacoan comparanda were found do not frequently feature voiced obstruents so that their origins are not clarified by the data. Item (18.), ‘rubbish consisting of branches and leaves which covers rocky regions,’ corresponds to tape ‘grass’ in Cha’palaa and ta’pé ‘grass, bush’ in Tsafiqui. Southern Barbacoan languages have innovative voiced stops, but their origins are largely unclear (Curnow & Liddicoat, 1998: 401). All this suggests complicated histories for the voiced obstruents in Cañar Quichua. The source of Cañar Quichua /l/, however, appears to be Barbacoan /l/, as per items (8.), (10.), (13.), (16.), and (17.). As far as vowels are concerned, proto-Barbacoan *ɨ appears to be adapted as /u/ in Cañar Quichua as per item (2.), but as /i/ in item (16). Otherwise, there are some interesting cases of vowel metathesis in the context of liquids (items (9.), and (10).) and backing of Barbacoan /i/ to /u/ in word-final position (items (7.) and (19.). Semantically, the lexical items largely conform to expectations as to likely lexical semantics of substrate vocabulary, certainly more so than in the Aymara case study reported in Sect. 4.4. In this particular case study, food preparation and physical deformations were also found to be relevant; the elevated occurrence of voiced fricatives /z/ and /zh/ in particular in items denoting “unusual, perhaps defective, physical or moral attributes” was already noted by Howard (2010: 140). There likely is some iconic value with which these sounds are imbued, just as is the case, interestingly, with the new phonemes which Cuzco Quechua possibly acquired through contact with Aymara (see Sect. 4.3).

5.5 Summary For the southern Ecuadorian highlands, there are to my knowledge no extralinguistic studies that could shed additional light on the sociolinguistic characteristics of the Quechuan spread and the displacement of the substratal language. Therefore, we can only evaluate the linguistic evidence as presented in the preceding sections. The common phonological characteristics of local toponymy of the Cañar-Azuay region and the lexical items for flora, fauna, daily material culture, food preparation, etc., which have not only enriched the Cañar Quichua lexicon, but also its phonological inventory, are strongly suggestive of the “fingerprint” of a substratal pre-Quechuan language of the region. This is in all likelihood the Cañari language, which was still spoken in this region in early colonial times, as the reports for the Relaciones Geográficas indicate. A detailed comparison of Cañar-Azuay toponymy with the rest of the Ecuadorian highlands and also the adjacent regions of southern Colombia has shown that, in spite of its distinctive character, the toponymy allows to draw some connections between Cañar, Azuay, and the remainder of the Ecuadorian highlands. For one, voiced stops and obstruents in toponyms extend to the area where

5.5 Summary

109

colonial sources mention the Puruhá language, and voiced stops in final position can be found in toponyms that extend all the way to and across the Colombian border. An especially relevant toponymic ending is -pud, which is clearly associated with the Cañar-Azuay region, but has a secondary center of density on both sides of the Ecuadorian-Colombian border in the area where the Barbacoan language Awa Pit is and its likely sister Pasto was once spoken. Other toponyms show a similar distribution, suggesting a linguistic connection of the pre-Quechuan Cañari language of the southern highlands of Cañar and Azuay with Awa Pit specifically or a language that had similar relevant characteristics. On the other hand, a different set of toponymic endings, including but not limited to -shi, connect the Cañar-Azuay region with the remainder of the Ecuadorian Andes more broadly. Since in these regions, the characteristic Barbacoan toponymy in -pí ~ -bí ~ -pi ~ -bi is found, this observation would support the idea of an extension of the Barbacoan languages as far south as Cañar and Azuay. However, the analysis of the toponymic record, at the present state of research, also leaves many loose threads and open questions: issues include the etymologization of toponymic endings through Barbacoan, the distribution of putatively relevant endings which are strangely thin on the ground in some currently Barbacoan-speaking areas, and the mismatch between the phonetic or even phonemic presence of voiced fricatives in some Barbacoan languages and the distributions of such sounds in the toponymy. As far as changes that can be attributed to the agency of a substrate language are concerned, evidence is restricted to the phonologization of voiced stops and fricatives in Cañar Quichua (the innovation of a future converb construction with different subjects in subordinate and matrix clause in Ecuadorian Quechua as a whole aside). While there are some internal sources for the innovative sound, it is clear that a mayor contributing factor was the addition of lexical items from a non-Quechuan language. Just like in the case of toponymy, a Barbacoan affiliation for this language is suggested by some notable similarities between the substratal layer in the Cañar Quichua lexicon with the lexicon of Barbacoan languages, including some forms with pan-Barbacoan distribution. However, on the other hand, it is just a small subset of the total amount of the uniquiely Cañar Quichua lexical items that can compared with Barbacoan in this way. And, just like Bruil (2011) has noted that there appears to be Barbacoan influence on the innovative future converb in Ecuadorian Quichua generally, some of the relevant items in Cañar Quichua, such as pilchi, recur in the Ecuadorian lowland varieties of Quichua, suggesting that they have undergone a period of lexical influence from Barbacoan prior to the spread of Quichua to the oriente. In sum, as Torero (1964) already noted, there is strong evidence for a substratal language that has left its mark on Cañar Quichua. But, unlike Torero (1964), who considered this language to be unknown, we can suggest that the Cañari language that was responsible for these changes was a southern representative of the Barbacoan language family. While several independent lines of evidence point into this direction, however, the case cannot be considered proven until toponymic endings with a suggestive distribution can be etymologized more firmly and more lexical items in Cañar Quichua can successfully compared with Barbacoan. In the course of

110

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

this, ideally also the emergence of the voiced stops and fricatives in those items and in toponymy will be clarified.

References Academia Mayor de la Lengua Quechua. (2005). Diccionario quechua – español – quechua, qheswa – español – qheswa simi taqe. Gobierno Regional Cusco. Adelaar, W. F. H., & Muysken, P. C. (2004). The languages of the Andes. Cambridge University Press. Alcina Franch, J. (1980). Tomebamba y el problema de los indios cañaris de la sierra sur del Ecuador. Anuario de Estudios Americanos, 37, 403–433. Anonymous. (1965) [~1570]. Relación de la tierra de Jaén. In M. Jimenez de la Espada (Ed.), Relaciones geográficas de Indias – Peru (Vol. 2, pp. 143–147). Atlas. Arias Dáuila, P. (1965) [1582]. Pacaibamba o Leoquina. In M. Jiménez de la Espada (Ed.), Relaciones geográficas de Indias – Peru (Vol. 2, pp. 274–281). Atlas. Bertonio, L. (1612). Vocabulario de la lengva Aymara. Primera parte, donde por abecedario se ponen en primer lugar los vocablos de la lengua Española para buscar los que les corresponden en la lengua Aymara. Iuli: Francisco del Canto. Beuchat, H., & Rivet, P. (1907). Contribution a l’étude des langues Colorado et Cayapa (republique de l’Équateur). Journal de la Société des Americanistes, 4, 31–70. Blevins, J. (2017). Areal sound patterns: From perceptual magnets to stone soup. In R. Hickey (Ed.), The Cambridge Handbook of Areal Linguistics (pp. 88–121). Cambridge University Press. https://doi.org/10.1017/9781107279872.006 Bruil, M. (2011). The emergence of future converbs in Imbabura Quichua: Pre-hispanic language contact as a possible explanation. In C. Hasselblatt, P. Houtzagers, & R. van Pareren (Eds.), Language contact in times of globalization (pp. 27–44). Rodopi. Carranza Romero, F. (2003). Diccionario quechua ancashino – castellano (W. Lustig (Ed.)). Iberoamericana/Vervuert. Cerrón-Palomino, R. (1990). Reconsideracion del llamado ‘quechua costeño.’ Revista Andina, 8(2), 335–386. Cerrón-Palomino, R. (2003) [1987]. Lingüística quechua (2nd ed.). Centro de Estudios Regionales Andinos ‘Bartolomé de las Casas.’ Cordero, L. (1992) [1895]. Diccionario quichua-castellano y castellano-quichua (5th ed.). Proyecto Educación Bilingüe Intercultural/Corporación Editora Nacional. Cordero Palacios, O. (1981) [1923]. El quechua y el cañari (contribución para la historia precuencana de las provincias azuayas) (2nd ed.). Departamento de Difusión Cultural, Universidad de Cuenca. Curnow, T. J. (1997). A grammar of Awa Pit (Cuaiquer): An indigenous language of south-western Colombia [Ph.D. dissertation]. Australian National University. Curnow, T. J., & Liddicoat, A. J. (1998). The Barbacoan languages of Colombia and Ecuador. Anthropological Linguistics, 40(3), 384–408. de Gallegos, G. (1965) [1592]. Sant Francisco Pueleusi del Azogue. In M. Jiménez de la Espada (Ed.), Relaciones geográficas de Indias – Peru (Vol. 2, pp. 274–278). Atlas. de Santo Tomás, D. (1560a). Arte de la lengua general del Peru, llamada, quichua. Francisco Fernández de Córdova. de Santo Tomás, D. (1560b). Lexicon, o vocabulario de la lengua general del Perú. Francisco Fernandez de Cordoua. Paz y Miño, L. T. (1942). Lenguas indígenas del Ecuador. IV – La lengua puruguay. Boletín de la Academia Nacional de Historia, 21(59), 42–74.

References

111

Salomon, F., & Grosboll, S. (1986). Names and peoples in Incaic Quito: Retrieving undocumented historic processes through anthroponymy and statistics. American Anthropologist, 88(2), 387– 399. https://doi.org/10.1525/aa.1986.88.2.02a00060 Soto Ruiz, C. (1976). Diccionario quechua: Ayacucho-Chanca. Ministerio de Educación/Instituto de Estudios Peruanos. de Gaviría, M. (1965) [1592]. Chunchi. In M. Jiménez de la Espada (Ed.), Relaciones geográficas de Indias – Peru (Vol. 2, pp. 285–287). Atlas. de Pereira, M., Tostado, J., & López, P. (1965) [1582]. Paute. In M. Jiménez de la Espada (Ed.), Relaciones geográficas de Indias – Peru (Vol. 2, pp. 271–274). Atlas. Emlen, N. Q. (2017). Perspectives on the Quechua-Aymara contact relationship and the lexicon and phonology of pre-proto-Aymara. International Journal of American Linguistics, 83(2), 307–340. https://doi.org/10.1086/689911 Floyd, S. (2021). Oclusivas complejas en el quechua de Domingo de Santo Tomás. Letras, 92(136), 115–140. https://doi.org/10.30920/letras.92.136.10 Gnerre, M. (1975). L‘utilizzazione delle fonti documentarie dei secoli xvi e xvii per la storia linguistica Jíbaro. In E. Cerulli, & G. D. Ragione (Eds.), Atti del xl Congresso Internazionale degli Americanisti, Roma – Genova, 3 – 10 Settembre 1972 (Vol. 3, pp. 79–86). Linguistica – folklore – storia americana – sociologia. Tilgher. Gómez Rendón, J. (2016). Aproximaciones a la onomástica indígena del Austro lojano. Antropología: Cuadernos de Investigación, 16, 115–129. Gómez Rendón, J. (2017). Las lenguas barbacoanas meridionales y el quechua. Revista Pucará, 28, 55–97. Gómez Rendón, J., & Adelaar, W. (2009). Imbabura Quechua. In M. Haspelmath, & U. Tadmor (Eds.), Loanwords in the world’s languages: A comparative handbook (pp. 944–967). Walter de Gruyter. Gonzales Castaño, G. (2014). Esquisse de description de la langue Nam Trik: Langue amerindienne parlée dans les Andes Colombiennes [M.A. thesis]. Université Lumière Lyon 2. Hirschkind, L. (1995). History of the Indian population of Cañar. Colonial Latin American Historical Review, 4(3), 311–342. Hocquenghem, A.-M. (2012). How did Quechua reach Ecuador? In P. Heggarty & D. BeresfordJones (Eds.), Archaeology and language in the Andes: A cross-disciplinary exploration of prehistory (pp. 345–371). Oxford University Press. Howard, R. (2010). ‘Why do they steal our phonemes?’ Inventing the survival of the Cañari language (Ecuador). In E. B. Carlin, & S. van de Kerke (Eds.), Linguistics and archaeology in the Americas: The historization of language and society (pp. 123–145). Brill. Huayhua Pari, F. (2009). Diccionario bilingüe polilectal aimara - castellano castellano - aimara. Fondo Editorial de la Universidad Nacional Mayor de San Marcos. Huber, R. Q., & Reed, R. B. (1992). Vocabulario comparativo: Palabras selectas de lenguas indígenas de Colombia/Comparative vocabulary: Selected words in indigenous languages of Colombia. Instituto Lingüístico de Verano. Italiano, H. (1965) [1592]. Alusi. In M. Jiménez de la Espada (Ed.), Relaciones geográficas de Indias – Peru (Vol. 2, pp. 287–289). Atlas. Jijón y Caamaño, J. (1919). Contribución al conocimiento de las lenguas indígenas, que se hablaron en el Ecuador interandino y occidental con anterioridad a la conquista española (ensayo provisional). Boletín de la Sociedad Ecuatoriana de Estudios Históricos Americanos, 2(6), 340–413. Jijón y Caamaño, J. (1939). Materiales para el mapa lingüístico del occidente de Colombia. Boletín de Estudios Historicos, 8(90), 167–175. Jijón y Caamaño, J. (1940). El Ecuador interandino y occidental antes de la conquista castellana (Vol. 1). Editorial Ecuatoriana. Jijón y Caamaño, J. (1941). El Ecuador interandino y occidental antes de la conquista castellana (Vol. 2). Editorial Ecuatoriana.

112

5 The Barbacoan Languages and the Southern Ecuadorian Highlands

Jijón y Caamaño, J. (1943). El Ecuador interandino y occidental antes de la conquista castellana (Vol. 3). Editorial Ecuatoriana. Landerman, P. N. (1994). Glottalization and aspiration in Quechua and Aymara reconsidered. In P. Cole, G. Hermon, & M. Daniel Martín (Eds.), Language in the Andes (pp. 332–378). University of Delaware, Latin American Studies Program. Lindskoog, J. N., & Lindskoog, C. A. (1964). Vocabulario cayapa. Instituto Lingüístico de Verano/ Ministerio de Educación Pública. López de Solís, L. (1596). Sinodales del obispado de Qvito celebradas en la civdad de Loxa por el ill mo , y r mo , sr ., d, fr, Lvis Lopez de Solis obispo de la Santa Yglesia de Qvito el año de 1596. Manuscript kept at the Archivo Arzobispal de Quito, Gobierno Eclesiástico, 2. Ministerio de Educación. (2009). Kichwa yachakukkunapa shimiyuk kamu: runa shimi – mishu shimi, Mishu shimi – runa shimi. http://www.illa-a.org/cd/diccionarios/diccionarioKichwaEcua dor.pdf. Accessed 31 March 2023. Moore, B. R. (1966). Diccionario castellano-colorado colorado-castellano. Instituto Lingüístico de Verano. Muysken, P. (2009). Gradual restructuring in Ecuadorian Quechua. In R. Selbach, H. C. Cardoso, & M. van den Berg (Eds.), Gradual creolization: Studies celebrating Jacques Arends (pp. 77–100). John Benjamins. https://doi.org/10.1075/cll.34.09muy Muysken, P. (2021). Substrate influence in Northern Quechua languages. In E. Oladé Aboh, & C. B. Vigouroux (Eds.), Variation rolls the dice: A worldwide collage in honour of Salikoko S. Mufwene (pp. 133–160). John Benjamins. https://doi.org/10.1075/coll.59.06muy O’Fallon, B. D., & Fehren-Schmitz, L. (2011). Native Americans experienced a strong population bottleneck coincident with European contact. Proceedings of the National Academy of Sciences of the United States of America, 108(51), 20444–20448. https://doi.org/10.1073/pnas.111256 3108 Norcliffe, E., Floyd, S., & Hammarström, H. (2015). The reconstruction and classification of the Barbacoan family of languages. Talk delivered at the 48th annual meeting of the Societas Linguistica Europaea 2015. Leiden University, The Netherlands. Orr, C., & Longacre, R. E. (1968). Proto-Quechumaran. Language, 44(3), 528–555. https://doi.org/ 10.2307/411720 Orr, C., & Wrisley, B. (1965). Vocabulario quichua del Oriente del Ecuador. Instituto Lingüístico de Verano/Ministerio de Educación Pública. Otero, J. M. (1952). Etnología caucana: estudio sobre los orígenes, vida, costumbres y dialectos de las tribus indígenas del Departamento del Cauca. Editorial Universidad del Cauca. Pablos, H. (1965) [1582]. Relación que enbió a mandar su magestad se hiziese desta ciudad de Cuenca y de toda su provincia. In M. Jiménez de la Espada (Ed.), Relaciones geográficas de Indias – Peru (Vol. 2, pp. 265–270). Atlas. Park, M., Weber, N., & Cenepo Sangama, V. (1976). Diccionario quechua: San Martín. Ministerio de Educación/Instituto de Estudios Peruanos. Parker, G. J. (1969a). Ayacucho quechua grammar and dictionary. Mouton. Parker, G. J. (1969b). Comparative Quechua phonology and grammar III: The Evolution of Quechua A. University of Hawaii Working Papers in Linguistics, 1(9), 149–204. Paris, J. (1961). Gramática de la lengua quichua actualmente en uso entre los indígenas del Ecuador (3rd ed., Rev. José María Lévenque and Gilberto Mejía). Editorial Santo Domingo. Paz y Miño, L. T. (1941). Lenguas indígenas del Ecuador. II – La lengua kára. Boletín de la Academia Nacional de Historia, 21(57), 28–52. Paz y Miño, L. T. (1961). Lenguas indígenas del Ecuador: La lengua kañar. Boletín de la Academia, Nacional de Historia, 43, 193–229. Renard-Casevitz, F. M., Saignes, T., & Taylor, A. C. (1988). Al este de los Andes: Relaciones entre las sociedades amazónicas y andinas entre los siglos xv y xvii (Trans. Juan Carrera Colin and rev. Gonzalo Flores and Olinda Celestino). Abya-Yala.

References

113

Salmons, J. (1992). Northwest Indo-European vocabulary and substrate phonology. In Perspectives on Indo-European language, culture and religion: Studies in honor of Edgar C. Polomé (Vol. 2, pp. 265–279). Institute for the Study of Man. Stark, L. R. (1985). Ecuadorian highland Quechua: History and current status. In H. E. Manelis Klein, & L. R. Stark (Eds.), South American Indian languages: Retrospect and prospect (pp. 443– 480). University of Texas Press. Stark, L. R., & Muysken, P. C. (1977). Diccionario español quichua, quichua español. Museos del Banco Central del Ecuador. Tödter, C., Waters, W., & Zahn, C. (2002). Shimikunata asirtachik killka inka-kaste-llanu: Diccionario inga-castellano (Quechua del Pastaza). Instituto Lingüístico de Verano. Torero, A. (1964). Los dialectos quechuas. Anales Científicos De La Universidad Agraria, 2(4), 446–478. Torero, A. (1984). El comercio lejano y la difusión del quechua: El caso del Ecuador. Revista Andina, 2, 367–402. Torero, A. (1993). Lenguas del nororiente Peruano: La hoya de Jaén en el siglo xvi. Revista Andina, 11(2), 447–472. Urban, M. (2012). Analyzability and semantic associations in referring expressions: A study in comparative lexicology [Ph. D. dissertation]. Max Planck Institute for Evolutionary Anthropology and Leiden University. Urban, M. (2018). The lexical legacy of substrate languages: A test case from the Southern Ecuadorian highlands. Transactions of the Philological Society, 116(3), 435–459. https://doi.org/10. 1111/1467-968X.12129 Urban, M. (2019). Lost languages of the Peruvian North Coast. Ibero-American Institute/Gebr. Mann. Vásquez de Ruiz, B. (1988). La predicación en guambiano. CCELA. Vásquez de Ruiz, B. V., Rojas Curieux, T., Gonzáles Castaño, G. K., & Diaz Montenegro, E. (2009). Léxico de la lengua namtrik de Totoró. Cabildo indígena pueblo Totoró. von Buchwald, O. (1908). Die Kara. Globus, 94, 123–125. Campbell, L. (1995). The Quechumaran hypothesis and lessons for distant genetic comparison. Diachronica, 12(2), 157–200. https://doi.org/10.1075/dia.12.2.02cam Weber, D. J., Cayco Zambrano, F., Cayco Villar, T., & Ballena Dávila, M. (1998). Rimaycuna: Quechua de Huánuco. Diccionario del quechua del Huallaga con índices castellano e inglés. Instituto Lingüístico de Verano. Wolf, T. (1879). Viajes cientificos por la República del Ecuador, verificados y publicados por órden del supremo gobierno de la misma República (Vol. 3). Memoria sobre la geografía y geología de la provincial de Esmeraldas, con una carta geográfica. Imprenta del Comercio. Wolf, T. (1892). Geografía y geología del Ecuador, publicada por órden del supremo gobierno de la república. F.A. Brockhaus. Zúñiga, F. (2015). How strong is the case for contact-induced grammatical restructuring in Quechuan? Linguistic Discovery, 13(2), 23–36. https://doi.org/10.1349/PS1.1537-0852.A.461

Chapter 6

Chachapoyas

Abstract This chapter presents the third and final case study of this book, which is concerned with the Chachapoyas area on the eastern slopes of the Andes in northern Peru. The site of a distinctive pre-Columbian culture, this region, like the southern Ecuadorian highlands, has a highly distinctive toponymy that is decidedly un-Quechuan and that has been attributed to a pre-Quechuan language conventionally called “Chacha”. As another parallel to southern Ecuador, also the Quechua of the Chachapoyas region has undergone some drastic changes, including vowel reduction processes that create surface syllables that are much heavier than is common in Quechuan. Here, I analyze the toponymy and a small set of vocabulary items that do not have Quechuan etymologies, to suggest that the so-called “Chacha” language may have been a relative of Cholón, a language that was still spoken until recently to the south in an area that was also part of the Chachapoyas archeological culture. However, in this case, it turns out that the structural changes of Chachapoyas Quechua, with the effects of stress assignment investigated by a new corpus study, cannot be argued to have been induced by a language like Cholón. Keywords Quechuan · Cholón · Chachapoyas · Toponymy · Anthroponyms · Substrate · Contact-induced language change · Stress · Vowel reduction · Language shift · Lexical borrowing · Loanwords

6.1 Introduction In the preceding two chapters, we have considered two areas of the Central Andes, in the north in Ecuador and in the south in Southern Peru and Bolivia, where Quechuan in all likelihood arrived relatively late in prehistory. In both cases, the successful expansion of the Inca empire may have been partly or entirely responsible for spreading the language. Attempts by the Inca to expand their domain to the densely forested lowlands to the east of the Andes had more chequered outcomes (see Renard-Casevitz et al., 1988: 43–55 for an overview and Renard-Casevitz et al., 1988: 59–116 for detailed and multifaceted case studies on the relations the Inca entertained with the adjacent lowlands). In the lowlands of Bolivia, an Inca fortress © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Urban, Linguistic Stratigraphy, https://doi.org/10.1007/978-3-031-42102-0_6

115

116

6 Chachapoyas

is reported near Ixiamas, which probably formed part of a chain of fortifications that extended further to the southeast (Pärssinen et al., 2003: 45). This demonstrates that the Inca did expand into the lowlands immediately adjacent to the Andes (Ixiamas has an altitude of just approx. 300 m above sea level) and secured control there. In other regions, however, campaigns to penetrate into the lowland jungles had catastrophic results. Cieza de León as paraphrased by Seitz (2017) relates, for instance, that the “[a]ttempts of the armies of the Inca Huayna Capac to dominate the Jívaro [in northwestern Amazonia, MU], for instance, were unsuccessful due to the rugged terrain, the presence of rivers that were difficult to cross, the heavy rains, and the warlike attitude of the Jívaro.” As far as language is concerned, in spite of the limited success of Inca expansion to the east, Quechuan eventually did reach the lowlands of Peru and Ecuador, but at a later point of time in the context of a complex mosaic of missionary activity and populations movements during the rubber boom (see e.g., Zariquiey Biondi, 2004 for a sketch of the sociolinguistic history of Napo Quichua). The same Huayna Capac who, according to Cieza de León, failed to subdue the Jívaro, successfully managed to conquer the region of Chachapoyas, located on the eastern slopes of the Central Andes in northern Peru, a cloud forest ecozone at the intersection between Andean highlands and Amazonian lowlands.1 And indeed, in this region, a Quechuan lect known as Chachapoyas Quechua or Amazonas Quechua is spoken. In fact, the earliest colonial sources do not mention any language in the Chachapoyas except the lengua del Inga, i.e., some form of Quechua (Valqui & Ziemendorff, 2016: 6). Nevertheless, this Quechua, as we have seen in Sect. 2.5, must originally have been embedded in a highly diverse linguistic landscape: further up the Marañón, on both sides of its steep valley, the linguistic mosaic of Jaén would have been found. Seven distinct non-Quechuan Indigenous languages were spoken in an area smaller than the Netherlands; some of these languages appear to have pertained to Amazonian language families (Torero, 1993). To the south, the closely related Hibito and Cholón languages were spoken on the interfluve of the Marañón and Huallaga rivers, Cholón likely also further east (Torero, 2002: 161). Still today, Chachapoyas is one of the Quechuan-speaking areas of the Andes, even though Chachapoyas Quechua is heavily endangered and is likely to become extinct soon as all remaining speakers are of advanced age, and the language has not successfully been transmitted to younger generations. Based on fieldwork in 1

The etymology of the name Chachapoyas is not clear. Blas Valera, a sixteenth century mestizo cleric hailing from Chachapoyas, suggests it means ‘place of strong men’ (Garcilaso de la Vega, 1829 [1609]: 109). Middendorf (1895: 233) derived it from Aymara chacha and Quechua puyu ‘cloud.’ Given the modifier-head order in Andean languages, as Middendorf was well aware, such a term could not have denoted to ‘men from the clouds,’ which would have been appropriate for the cloud forest environment, but must be interpreted as ‘cloud of men.’ This, in a move that is typical for folk-etymology, he justifies because the term could have referred to the masses of Chachapoyas warriors that appear like clouds. A third interpretation accepts the identification of the sequence -poya with Quechua puyu ‘cloud,’ but sees Chacha as deriving from Quechuan sacha ‘forest.’ Again, the reverse order would yield the meaning ‘cloud forest,’ but this is not available, so that one, rather implausibly, would have to say that Chachapoyas means ‘forest cloud.’ As “Chachas seems to have been the name of one of the ayllus or parcialidades in the colonial documents” from the Chachapoyas region (Schjellerup 1997: 30), such etymological attempts are likely misguided.

6.1 Introduction

117

the 1960s and 1970s, Taylor (1979: 5) reports Chachapoyas Quechua spoken at the communities of Huancas, Levanto, Yambajalca-Diosán (now Granada, Landerman, 1991: 19), Olleros, Quinjalca, and Goncha in the province of Chachapoyas, one of the southern provinces of Peru’s Amazonas department, and Colcamar, Lonya Chico, Conila-Cohechán, Luya Viejo, Trita, Cuémal, Olto, and Paclas on the upper Imaza in the Province of Luya, also part of the Amazonas department. The map in Fig. 6.1 plots the location of these communities within the Amazonas department. Further to the east, on the shores of the Huallaga river and its tributaries, people speak a Quechuan lect which is closely related to that of Chachapoyas. It is most frequently referred to as San Martín Quechua, named after the Peruvian department of which the relevant region is part. However, San Martín Quechua is not spoken along the entire course of the Huallaga river in the San Martín department, but mostly in the province of Lamas and adjacent parts of the province of San Martín proper (Landerman, 1991: map 9). The Quechuan-speaking areas of Chachapoyas and San Martín, according to Taylor (1979: 5), once formed a larger whole, but became separated by the hispanization of the intermediate provinces of Moyobamba and Rioja. Chachapoyas and San Martín Quechua, like the Quichua of Ecuador, show characteristics typically associated with Quechua IIB, such as the merger of velar and uvular stops (Taylor, 1996: 7) and some degree of postnasal voicing. Chachapoyas

Fig. 6.1 Approximate extent of the archeological Chachapoyas culture within Northern Peru (based on Church & von Hagen, 2008: Fig. 45.1) and Chachapoyas Quechua-speaking communities in the 1970s (based on Taylor, 1979: 5)

118

6 Chachapoyas

Quechua, though, is conservative in retaining the retroflex affricate */úʂ / that was inherited from proto-Quechua, although it turns into /ch/ in a significant portion of the vocabulary (Taylor, 1996: 7). Like the two other Quechuan varieties of northern Peru, Lambayeque and Cajamarca, Chachapoyas and San Martín Quechua have grammaticalized *llapa ‘all’ (Parker, 1969: 22), which does duty as a verbal plural marker (Landerman, 1991: 72; Taylor, 1999: 26). A further development that Chachapoyas Quechua shares with its northern Quechuan neighbors is the delateralization of */ll/ in most communal lects (Landerman, 1991: 72, 251); as Taylor (1996: 18) notes, the local realization of Spanish /ll/ also is [dZ]. On the other hand, Chachapoyas Quechua also has some highly characteristic features that sets it apart from its sister lect in San Martín. These include the assignment of primary stress to the initial syllable, the reduction of unstressed vowels, and the monophthongization of diphthongs / ay/ and /uy/ with compensatory lengthening to /e:/ and /o:/, respectively (Taylor, 1999: 23, 27). These changes play a central role in suggestions of influence from a substratal language on Chachapoyas Quechua, so we will discuss these in more detail in Sect. 6.3. Chachapoyas is a multiply polysemous term.2 We have already seen that it is the name of a province in the Peruvian Amazonas department, and that the local Quechuan lect is also known as Chachapoyas Quechua. More prominently, Chachapoyas is the name of a local pre-Columbian culture that was at home in the cloud forests of this part of Peru. Finally, the name can also refer to the bearers of that ancient Chachapoyas culture. As an archeological culture, Chachapoyas comprised larger areas of the eastern slopes of Peru than those in which today Quechuanspeaking communities are found. It corresponds roughly to the southern part of the Peruvian department of Amazonas rather than just the provinces of Luya and Chachapoyas (Fig. 6.1). However, the area once affiliated with the Chachapoyas culture extends further to the south on the interfluve between the Marañón and Huallaga rivers, which corresponds to the Bolívar and western Pataz provinces of the La Libertad region, roughly up to the border with Húanuco (Church & von Hagen, 2008: 904). In this region, the Cholón language was attested in historical times. One of the southernmost Chachapoyas sites is Nunamarca, to the northwest of the town of Tayabamba (Church & von Hagen, 2008: Fig. 45.1). Like the department, the Chachapoyas culture area is limited in the west by the valley of the Marañón; the limits of the Chachapoyas culture in the north and east “remain unknown and are likely to have fluctuated” (Church & von Hagen, 2008: 904, compare Schjellerup, 1997: 27). The map in Fig. 6.1 shows the extent of the Chachapoyas archeological culture based on the map of Chachapoyas archeological sites in Church and von Hagen (2008: 905, Fig. 45.1).3 2

Church and von Hagen (2008: 904) distinguish between Chachapoya as the name of the preInca society that bore the archeological culture and Chachapoyas as “the pre-Hispanic geographic domain surrounding the modern city of Chachapoyas.” 3 The Chachapoyas culture may, however, have extended further northward, roughly to the latitude of the town of Bagua (Schjellerup 1997: 212) and would then have included the entire region in which Chachapoyas Quechua is attested.

6.1 Introduction

119

Chachapoyas is an archeological culture that flourished in the Late Intermediate Period, from approximately 900 AD until the Inca conquest around 1470. Characteristic for the Chachapoyas culture are round houses with friezes that are decorated with geometric patterns (Church & von Hagen, 2008: 914). The environs of Leymebamba mark an artistic watershed: triangle and rhombus-shaped patterns dominate in the north, whereas in the south meandering designs prevail (Schjellerup, 1997: 238). One of the most prominent sites of the Chachapoyas culture is the massive fortress of Kuelap, overlooking the Utcubamba river in the present-day province of Luya. Another notable archeological site associated with the Chachapoyas culture, Gran Pajatén, is located much further south on the Marañón-Huallaga interfluve (Bolivar province, La Libertad department), near the southern limit of the archeological culture. The topography of this terrain is rugged and deeply incised by the river valleys. Ecosystems range from barren grass and shrubland in the heights of the Cordillera Oriental, which reaches an altitude above 4000 m locally, to humid tropical cloud forest ecozones down to 400 m of altitude near the Huallaga (Church & von Hagen, 2008: 906). These gradual changes in the natural surroundings appear to be reflected by different adaptations of the peoples of this area. Schjellerup (1997: 54–55), thus, says that “there was a cultural difference between the highland Chachapoyas group who, although they dominated all the ecological zones from warm yunga coca-growing lands to the highest tuber-growing areas and lived in the ceja de montaña region, developed their distinctive features in contrast to the upper lowland montaña groups, the Xibitos and Cholones, who seem to have developed their characteristic culture based on more typical lowland cultural traits but who were still different from the lowland selva tribes, the former representing an intermediate adaptation.” As in the case of pre-Inca Cañari, the ethnohistorical and archeological evidence suggests that Chachapoyas society, while internally stratified, was organized into several chiefdoms that were not united permanently by an overarching political authority or leader (see review in Church & von Hagen, 2008: 904). Also as in the Cañari case, a regional identity that is evidenced archeologically by “similarities in pottery styles, archeological details, and settlement patterns” (Church & von Hagen, 2008: 904) is likely to have emerged in the early Late Intermediate Period, but Chachapoyas as an overarching ethnic category may have been the result of the practical needs of Inca administrators, and strengthened only when the opposition to the Inca dominance constituted a unifying force (Church & von Hagen, 2008: 904; Schjellerup, 1997: 36). The Inca impact on the Chachapoyas was intense and transformative with respect to sociopolitical organization. In yet another parallel to the Cañari of Ecuador, the Inca also resettled a significant part of the population from the Chachapoyas region to other parts of the empire and installed administrators from outside. Rebellions ensued also in the Chachapoyas (Schjellerup, 1997: 65), and the Inca had to invest significant energy into securing control of this strategically important region that allowed access to exchange systems that reached into Amazonia (Church & von Hagen, 2008: 917). Warfare and violent conflict that the Chachapoyas people became involved in after contact with the Inca led to a collapse of the population, so that the Cabildo of Chachapoyas remarked in 1544 that the

120

6 Chachapoyas

region was almost depopulated; from these low levels the population would further decline well into the seventeenth century (Schjellerup, 1997: 39–40).

6.2 Toponymy It has been noted early on that the region where Chachapoyas Quechua is spoken today is sprinkled with non-Quechuan toponymy. Bandelier (1907: 9) remarks that near Chachapoyas, one can find some ruins called Aymarabamba, a fact that becomes significant when noting that the anthroponymic record of the Chachapoyas also appears to contain a small set of names with apparent Aymara etymologies (Rivarola, 2007). Given the ease with which the toponym and the anthroponyms can be identified, the presence of the Aymara language in the Chachapoyas would appear to be more likely recent and faint rather than ancient and well-entrenched. But Bandelier (1907: 9) also noted that, beyond the case of Aymarabamba, the Chachapoyas region displayed “names of localities, derived from a tongue of which no trace is left. Kuélap, Camdshian, Macro, are neither Quichua nor Aymará.” Making similar observations also on personal names, Bandelier (1907: 10) concludes that “the country of Chachapoyas was once inhabited by a tribe or tribes, that belonged to a stock different from the [Quechuan-speaking, MU] Peruvian mountaineers, their western neighbors.” Since Torero (1989: 238), the hypothetical pre-Quechuan language of Chachapoyas, has been known as Chacha, a term that “evidently referred to a major ethnos occupying the upper Utcubamba Valley during the sixteenth century” (Church & von Hagen, 2008: 904). Valqui Culqui (2004) situates the extinction of the Chacha language to the end of the fifteenth or the beginning of the sixteenth century, i.e., in the time shortly after the Inca conquest. Taylor (2000: 31), however, opines that the language survived as an expression of popular culture until the eighteenth century (though Taylor, 1990: 123 himself notes that in that case, it is odd that the language does not figure in Martínez Compañón’s, 1985 [1782–1790] late eighteenth century comparative wordlists of northern Peruvian languages). There is, in any case, no direct evidence for such datings, since already the earliest Spanish accounts already only mention Quechua in the Chachapoyas. It is quite possible that the Chacha language became extinct earlier, and it is equally possible that it survived, unnoticed by the Spanish observers, into colonial times. What is clear is that there are some hybrid toponyms with Quechuan initial elements and non-Quechuan local toponymic endings, e.g., Mashumal, Cochamal, Chuquimal. Their existence suggests a period of bilingualism (Valqui & Ziemendorff, 2016: 28–29). More recent authors have investigated the toponymic record of the Chachapoyas area more rigorously than Bandelier. Torero (1989: 238) notes a concentration of toponyms in -mal, -lon, and -lap in the provinces of Chachapoyas, Rodriguez de Mendoza, Luya, and the south of Bagua and Bongará. Figure 6.2, reproduced from

6.2 Toponymy

121

Fig. 6.2 Toponyms in -cat, -den, -mal, -lap, and -lon in Northern Peru, from Urban (2021a)

Urban (2021a), shows that these toponyms indeed cluster significantly in the southern part of the Amazonas department.4 The area thus delimited includes the regions in which Chachapoyas Quechua is spoken, but is larger. It corresponds almost to the entire southern part of the Amazonas department and thereby covers a significant part of the area that was once associated with the Chachapoyas archeological culture. As has been noted before, too, the distribution of these endings sometimes transcends the boundaries of the southern Amazonas department in which Chachapoyas Quechua is spoken. Indeed, we can observe a pattern that is similar to what was found in the Cañari case study in Sect. 5.2: the local toponymy of the Chachapoyas region in the southern Amazonas department has some very characteristic endings that are exceedingly rare outside (-lap, -lon), but there are also endings which are more widely distributed in the adjacent regions of the Andes. Torero (1989: 237) draws

4

Taylor (1990: 133) also investigates the toponymy of Chachapoyas and suggests that -mal could have meant ‘plain’ and -lap (with variant -lape) possibly ‘fortress, fortified settlement.’ Taylor (1990: 133) also discusses a somewhat less frequent ending -huala, of which he says that it could have had the meaning ‘hill.’ The same author also discusses -rán, which is unusual among the toponymic endings of the region in that it attracts stress; the element could be related to Ancash Quechua -urán ‘slope,’ itself possibly related to the Quechua verb ura- ‘to lower’ (Torero 1989: 237–238). Valqui Culqui (2004) furthermore adds -oc, which could have referred to rocky or sandy terrain, to the roster of Chachapoyas toponymic endings, and Valqui Culqui (2003: 70) in addition discusses the ending -sul.

122

6 Chachapoyas

attention to the fact that Toribio de Mogrovejo (2006 [1593–1605]), the widely traveled sixteenth century archbishop of Lima, mentions a village called Puymal further south and, interestingly, in the “provincia de Cholón,” a place called Chamal. The extension of -mal toponyms to the Bolívar province of La Libertad can also be gleaned from the map in Fig. 6.2. Indeed, the distribution of -mal corresponds to the extension of the prehistoric Chachapoya culture as established by archeological research (Valqui & Ziemendorff, 2016: 14). There are yet wider connections. Taylor (1990: 133) furthermore mentions among the toponymic endings of Chachapoyas -gat ~-gate ~-gache ~-cache ~-gote, to which, being frequent in river names, he assigns the meaning ‘river’ or ‘water,’ an association which is confirmed by Valqui and Ziemendorff (2016: 16).5 Now, Torero (1989: 234) observes that such toponyms, with almost exactly the same formal variation and the same association with bodies of water, occur not only in the provinces of Utcubamba, Bagua, Luya, and Chachapoyas, but, crucially, have a much wider distribution in Northern Peru. The area in which they are attested comprises the entire highlands of Cajamarca and adjacent parts of Lambayeque, and also, with a more scattered distribution, the highland provinces Ayabaca and Huancabamba of Piura and the Pallasca and Huaylas provinces of Ancash. In large parts of its area, toponyms in -cat and variants overlap with another toponymic set, called -den by Torero (1989: 229) after the most frequent allomorphs (others are -don, -ten, and -ton). All these observations are confirmed by the map in Fig. 6.2. One can add that, like -mal, there are also tokens of -den and -cat toponyms on the Marañón-Huallaga interfluve where Cholón is historically attested. Indeed, as Torero (1989: 237) mentions, an addendum to the manuscript of Pedro de la Mata’s (2007 [1748]) grammar of the language, not written by Mata himself, provides a list of Hibito and Cholón villages, some with translation. Among these, there is a village named Zalcot, accurately translated as ‘black river’ (compare Cholón t s al ‘black,’ kot ‘water,’ Alexander-Bakkerus, 2005). The fact that the Cholón word for ‘water,’ kot, is so similar to the toponymic ending cat and variants is suggestive. However, similar words for ‘water’ are not only attested for Cholón’s sister Hibito, which has (Martínez Compañón, 1985 [1782– 1790]), but also in the unclassified Copallín language of the Jaén area still further north, where one finds (Torero, 1989: 236, 1993: 459). Also Culli quida ‘sea,’ a form also found in Martínez Compañón (1985 [1782–1790]), is not far phonetically and semantically (Torero, 1989: 236). However, a privileged connection between the toponyms with Cholón is suggested by the remarkable fact that Zalcot has several “twins,” spelled Salcot, in the Cat-area delimited by Torero (1989). Torero (1989: 236–237) accordingly calls for further research on the possible relation between the Den and Cat areas with Cholón, whose speakers, as he says, did not live far away at all. 5

As far as the formal variation is concerned, Valqui and Ziemendorff (2016: 17–18) make the interesting observation that a preceding nasal always triggers the variant with voiced [g], and the same authors note that the variants with palatalization of [t] to [t∫] correlates with the presence of paragogic -e. They also note that initial gach-, which occurs for instance in the name of Gachmal, the site of various springs, is likely identical in origin with the ending -gat ~ -gate ~ -gache ~ -cache ~ -gote.

6.2 Toponymy Table 6.1 Chachapoyas toponymic endings and Cholón comparanda

123

Chachapoyas toponymic endings

Cholón comparison

-cat

kot ‘water’

-mal

mol ‘ground’

-puy

pey ‘earth’

In previous work reported in Urban (2021a) , I have taken some steps in this direction, paying, as Torero (1989) suggests, particular attention to Cholón, to which the hitherto available evidence points. In fact, in the meantime Jolkesky (2016, Table 10) observed that the Chachapoyas ending -mal could be relatable to Cholón mol ‘ground, day,’ which is certainly plausible semantically, especially when one remembers that -mal could have designated plains (Taylor, 1990: 133). A further toponymic ending of the Chachapoyas region that can plausibly be explained through Cholón is -puy. This ending, in fact, also recurs much further south, for instance in Culpuy, a hamlet in La Libertad’s Bolívar province that was traditionally Cholón-speaking. Once again, the list of Hibito and Cholón villages that was passed down to modern scholars together with Mata’s Cholón grammar is relevant. The name of the town of Jopeytè is translated as ‘land like blood,’ an explanation that is confirmed by the contents of Mata’s Cholón grammar itself (ho ‘blood,’ pey ‘land,’ Alexander-Bakkerus, 2005). Again, not only would it be natural for pey to recur as the head of placenames, but we also have direct evidence that this is the way it indeed acts in an incontrovertible Cholón toponym. For the Chachapoyas region, Schjellerup (1997: 84) mentions among others an ayllu and site called Ampui, which, once -pey ‘earth’ is identified as its head, can be further illuminated through Cholón am ‘food, place’ (AlexanderBakkerus, 2005; compare also Amami, a name for a field according to Schjellerup, 1997: 340). To summarize, a whole set of Chachapoyas toponymic endings can, in a way that is plausible formally and semantically, be related to Cholón common nouns, as Table 6.1 summarizes. However, while the consonant correspondences are exact, the vowel qualities differ between the toponymic endings and the Cholón comparanda (though note that -cat is only a label for a toponymic set that involves a whole range of allomorphs, some of which with a mid-back vowel). These deviations could be due to several factors. For one, the Chacha language which left the toponymic record of Chachapoyas as a Sprachdenkmal need not have been identical to Cholón, but could be a closely related sister language that has evolved in a different direction than Cholón as far as its sound system is concerned. This, however, is an ad hoc explanation, and not in itself convincing. The reason for the discrepancies may also well be more mundane: as explained in Urban (2021a), there is evidence from sparse modern data in Latorre (2014) that Cholón featured a phonemic high-central vowel that is not faithfully distinguished in Mata’s (2007 [1748]) grammar. One can also observe other mismatches in the transcription of vowels in Latorre’s phonemically transcribed data and Mata’s colonial grammar. For these reasons, there is some doubt on the accuracy of the representation of vowels in this pre-modern source generally, and issues with

124

6 Chachapoyas

vowel qualities and their orthographic representation of vowels may partially underlie the divergences. Also other toponyms of the Chachapoyas area may have Cholón etymologies. Urban (2021a) mentions that the name of the Río Shocol in the Chachapoyas region may be identical etymologically to the word for ‘river’ in Mata’s Cholón grammar, šokot, literally ‘flowing water’ (compare šo(h) ‘to pour,’ kot ‘water,’ AlexanderBakkerus, 2005). However, it remains unexplained how and why the final obstruent became a lateral. Limabamba, furthermore, the name of a village built on a plane surrounded by a horseshoe-shaped mountain chain to the southeast of the town of Chachapoyas, is plausibly explained as a hybrid of Quechua pampa ‘plain’ and Cholón limaŋ ‘mountains, highland’ (Alexander-Bakkerus, 2005). The possibility that the initial component goes back to Spanish lima ‘lime’ should be excluded before accepting this etymology, but the fact that the list of Hibito and Cholón villages in the appendix to Mata’s Cholón grammar includes one called Lima (without a translation being provided) shows that Lima is an element that figures in Cholón toponymy, making the idea that the initial sequence Lima- in Limabamba has a Cholón etymology, too, more plausible. And, given its location on a highland plain, the name Limabamba thus etymologized would befit the place superbly well.6 Personal names from the Chachapoyas region (Rivarola, 2007; Zevallos Quiñones, 1966) can also partially be reconciled with Cholón. A set of personal names ends in pol or more frequently -pul, namely Bapol, Chupul, Copul, Mapul, and Llancapul. In fact, Pul alone also occurs as a name, as does Llanca, which forms the initial element in Llancapul. Together, these occurrences virtually guarantee that -pul represents a separable morpheme (compare Salomon & Grosboll, 1986). The fact that it can be combined with other personal names suggests that personal names ending in -pul reflect some sort of connection between different individuals, concretely, that the individual Llancapul stands in some relation to the individual called Llanca, with -pul specifying what relation this is. These observations make it quite attractive to connect the ending -pul with Cholón pul(up) ‘son’ (Alexander-Bakkerus, 2005). In a similar vein, frequent personal names ending in -pa (Calampa, Chamcapa, Chancapa, Chipa, Cupa, Gochcapa, Guanpa, Guispa, Hoccheba, Hocchipa, Hocchupa, Hocñapa, Hopchipa, Hopchupa, Llanchipa, Llauchipa, Muchuypa, Ñaupa, Ompa, Opa, Osapa, Osumpa, Rapa, Tutpa, Zanpa) could be related to Cholón pa ‘father’ as documented by Mata (2007 [1748]) (Alexander-Bakkerus, 2005).7 Consistent with 6

Going beyond Urban (2021a), one can note that an anonymous sixteenth century account on the native religion of the peoples of Peru (Anonymous, 1865 [~1560]: 34) mentions a huaca in the form of a large stone that was driven into the ground in the center of villages. While the account does explicitly mention Chachapoyas in the text, it is unclear what part of Peru the above passage is referring to. However, the fact that Schjellerup (1997: 168–173) identified such a huaca at Papamarca confirms that the Chachapoyas region is relevant. This stone, the account explains, was considered the eye of the village, and it was worshipped and adored so that it may protect the village and the fields. It was known as according to the Augustinians. Now, the fact that these huacas were considered to be “eyes” makes it interesting to note that the sequence strongly resembles Cholón nyaˇce ‘eye’ (Alexander-Bakkerus, 2005). 7 The list does not mention names that might more likely reflect Quechua -sapa.

6.2 Toponymy

125

this interpretation, -pa appears to establish a relationship to other people, as the names Hocchepa ~ Hocchipa ~ Hocchupa and Hocñapa appear to contain variants of the frequent Chachapoyas name Oc (also spelled Occ, Hoc, etc.); similarly, the name Muchuypa is apparently related to the independently attested name Muchu. Note, however, that there does not appear to be a correlation between these names with sex, which one might expect if -pul -and -pa are identified with Cholón pul(up) ‘son’ and pa ‘father’, respectively: names ending in -pul can also be borne by females in the 1782 document published by Rivarola (2007), and the same appears true of names ending in -pa. If the Cholón connection suggested above should be accurate, this would mean that by the late eighteenth century, perhaps hundreds of years after the extinction of the Chacha language from which they must derive, the original connection of the last names with kinship would have been lost from memory (as happened with, for instance, Scandinavian names in -sen or -sson). Returning to the toponymic record, we have seen how Chachapoyas toponymy shares some characteristics with Northern Peru more broadly. This saliently includes the ending -cat with variants, which strongly suggests a Cholón connection. However, also the ending -den is sparsely present in the Chachapoyas area (see Fig. 6.2). If -cat is the signature of a language related to Cholón (note, incidentally, Cholón hydronyms like Axuacot ‘fish water,’ where the relevant item is clearly present), the toponymic evidence would suggest that Cholón or closely related lects were once spoken widely in Northern Peru. Indeed, the distribution of -puy suggests something similar. Beyond the historically known Cholón-speaking areas and the Chachapoyas region, we find Calipuy, the name of a wildlife reserve on the La Libertad-Ancash boundary, just about 50 km from the Pacific coast. Adelaar (2012: 580 and p.c.) has added an observation that corroborates such a scenario. He notes that the placenames Llacanora and Llagadén could partially be explained through Cholón ly aka ‘red’ (Alexander-Bakkerus, 2005). In Urban (2021a), I observe that there is another relevant item, Llacamate, a place near the Calipuy wildlife reserve just mentioned. All three placenames are strikingly located near places with reddish or ochre land (probably red clay soils), lending significant extralinguistic support to the suggestion of a relation to Cholón ly aka ‘red.’ Note also that by way of the placename Llagadén, the -den toponyms are also implicated. In spite of their partial geographic overlap, previous research has tended to treat the Den and Cat-toponymic areas as the signatures of two distinct languages (although Torero, 1989: 236 does explicitly not exclude the possibility that placenames in -cat and variants go back to the same linguistic source as -den). Andrade Ciudad (2010) proposes an elaborate chronological scenario that assigns these toponymic layers to different temporal layers, with the goal to explain the partial overlap of -cat with -den and the areas where other languages such as Culli were once spoken. However, contrary to what Torero (1989: 235) suggested, Urban (2021a, 2021b) observes that toponyms in -den on the one hand and toponyms in -cat on the other share part of their modifying elements and, by proxy, part of their lexicon (compare Malcat, Melcot, and Molecote with Molladén or Shúngote with Shuendén), which suggests some sort of connection between the languages. Importantly mal- ~ mel- ~ mol~ molle- may have no other etymology than the toponymic ending -mal. Shún- ~

126

6 Chachapoyas

shuen-, on the other hand, can plausibly be explained through Cholón šuŋ ‘village’ (Alexander-Bakkerus, 2005, note also the classifier -šuŋ ‘accumulation’). In sum, some limited aspects of the evidence suggest that toponyms in -den and -cat in northern Peru as a whole are the signature of one or more languages that were related to Cholón. The fact that these toponyms have different centers of density does not constitute strong evidence against this idea. Regional differences at the dialectal level or possibly even beyond may be involved. But this cannot be proven. More to the point, a comparison with the toponymic record in other parts of the world shows that some toponymic endings can show remarkable clustering although the linguistic landscape is relatively homogeneous. Examples from this come from the continental Germanic dialect continuum. For instance, although all derive from closely related dialects, the ending -leben is particularly common in the states of Sachsen-Anhalt and Thuringia in Germany, -broich is common in the Lower Rhine region, and -büll has a strong center of gravity in Schleswig–Holstein in Northern Germany. However, there are complicating factors. Hybrid Quechua-Den and even two apparent Spanish-Den toponyms (Andrade Ciudad, 2010: 174) suggest that the language that left the toponyms in -den as Sprachdenkmäler was spoken until relatively recently, possibly into colonial times. And in the testament of Sebastian Ninalingón, the sixteenth century indigenous ruler of Xaxadén, a place ostensibly associated with the -den toponymic areas, one finds three words that have been considered non-Quechuan: nus ‘principal wife,’ losque, and mizo ‘mistress’ (Espinoza Soriano, 1977: 410; note, however, that nus may be a heavily adapted form of ñusta, a Quechuan term for Inca nobility usually translated as “princess;” this idea is made more plausible formally by the fact that the relevant area of northern Peru, including Cholón, seems to have undergone shared sound change involving depalatalization, and semantically by the fact that terms for ranks and roles in Inca society have also been borrowed into Yanesha’, an Arawakan language of the eastern slopes of Northern Peru). The possibility of a survival of the language from which -den toponyms derive make it logical to assume that these words, or a subset of them, are associated with that language (Andrade Ciudad, 2010). However, they cannot be etymologized through Cholón (which has t s ala ‘married woman,’ Alexander-Bakkerus, 2005). A more general complicating factor is one which we have already encountered in the Cañari case study in Sect. 5.2. We have seen instances of relevant toponymic endings that are without any doubt associated with the Cholón language directly, including, from the appendix to Mata’s Cholón grammar, Zalcot, in which the -cat ending is represented, Jopeytè ‘land like blood,’ in which we find a direct instance of the ending -puy, and the place called Chamal in the “provincia de Cholón” which Mogrovejo (2006 [1593–1605]) reports. However, as a look at the map in Fig. 6.2 also shows, while occurring in the historically Cholón speaking area, these toponymic endings are actually not strongly represented there. In fact, the distribution of all three endings is much more scattered in the area where Cholón is known to have been spoken in historical times than in other parts of Northern Peru.

6.3 Substrate Effects

127

Table 6.2 Assumed proto-Chachapoyas-San Martín Quechua phoneme inventory, adapted from Taylor (1999: 21) Plosives

Bilabial

Apical

p

t

m

Retroflex

ch

ts

n

ñ

Fricatives

s

sh

Vibrant

r

Lateral Approximants

ll w Front

High

Velar k

Affricates Nasals

Palatal

y Central

i

Low

Back u

a

6.3 Substrate Effects Taylor (1999: 21) reconstructs the common phonological system of the immediate ancestor of Chachapoyas and San Martín Quechua as shown in Table 6.2. Taylor (1990: 128) says that what can be learned on Chacha phonology from personal names and placenames does not warrant the conclusion that substrate influence is responsible for the changes which this common ancestor has undergone: the merger of the uvular and velar stops could be motivated by the absence of such a distinction in Chacha (see Cerrón-Palomino, 1990 for a more general argument that this distinction tends to be lost in contact situations), but the loss of initial */h/, postnasal voicing, and the delateralization of */ll/ seem hardly motivatable by Chacha substrate. In Chachapoyas Quechua itself, however, this system undergoes some drastic changes. /i/ and /u/ have mid-allophones [e] and [o] without apparent contextual motivation, and in La Jalca, /a/ can also surface as [e] (Taylor, 1999: 27–28). Combinations of /a/ with the semivowels /y/ and /w/ tend to be pronounced [e:] and [o:], respectively (Taylor, 1999: 27; the latter realization is characteristic of La Jalca and the Upper Imaza lects; in Luya and Huancas, the pronunciation is [oy]). The vowels of unstressed syllables can reduce to shwa (Taylor, 1999: 28). Taylor (1999: 29) represents the Chachapoyas Quechua consonant system orthographically as in Table 6.3, noting that “[o]n a certain linguistic level, all the graphical symbols indicated above function as phonemes”.8 As one can see in Table 6.3, the Chachapoyas system is significantly more complex than the inherited one (/ɟ ʝ / is from proto-Quechua */ll/ except in the Upper Imaza, which retains /ll/). According to Taylor (1996: 7), this is due to the combination of substrate influence from the Chacha language, the Spanish superstrate, and the “idiosyncrasies of the dialect” (which he does not specify). Landerman (1991: 103) 8

“A cierto nivel de lenguaje, todos los símbolos gráficos indicados arriba funcionan como fonemas.”

128

6 Chachapoyas

Table 6.3 Chachapoyas Quechua consonant phoneme inventory, adapted from Taylor (1999: 29) Bilabial

Dental

Apical

Voiceless plosives

p

t

Voiced plosives

b

d

Palatal

g ch

Voiced affricates

ɟʝ m

Voiceless fricatives

f

Voiced fricatives

β

Vibrant

tʂ ŋ

n

ñ

s

sh

ʂ

(Z)

ʐ

r

Lateral Approximants

n̪

Velar k

Voiceless Affricates Nasals

Retroflex

l w

ll y

observes succinctly that “[t]he dialects of the Amazonas area are characterized by a complicated series of phonological changes consisting of vowel dropping, monophthongization, and assimilation of voicing and point of articulation.” He goes on to say that “[t]he overall effect is so radical that it is difficult at first to believe that the language being spoken has anything to do with Quechua.” The phonemes /Z/, /f/, and /β/ are restricted to loans from Spanish (Taylor, 1999: 35–36; [Z] is the realization of Spanish /ll/ in the Spanish of the Chachapoyas region). Voiced stops /b/, /d/, /g/ as well as /l/ occur in words borrowed from Spanish, but also in words with “local” (i.e., substratal) origin; there are also some marginal internal sources for the sounds (Taylor, 1999: 31, 34). Chachapoyas Quechua flora and fauna terminology investigated by Aviva Shimelman in 2018 includes names such as bequkillu, dukup, itil, or gulu’gulu. These do not seem to have cognates in other Quechuan lects, thereby demonstrating that indeed a substrate origin of (some) voiced stops in Chachapoyas Quechua is plausible. We will have a closer look at the non-Quechuan vocabulary in Sect. 6.4. As in the Cañari area, we find initial voiced stops represented by in the toponymic record as well (e.g., Bocumal, Danjamal, Dondomal, Gomal, Gamgache, Gopioc, Valqui Culqui, 2004). Final voiced obstruents are much less frequent, though Martínez Compañón (1985 [1782–1790]), for instance, mentions a place called Giyib with final , and a final is attested in Lamud. Much the same picture is found in the anthroponymic record; here, we even find an instance of final in the name Bisig (Zevallos Quiñones, 1966: 31). As in the Cañari case study, therefore, the lexical record is consistent with the toponymic one, in that both have common characteristics. Indeed, Taylor (1990: 127) posits voiced stops /b/, /d/, /g/, the alveolar lateral /l/, and also voiceless and voiced fricatives /f/ and /v/ ([β]?) for the substratal Chacha language. The split which the inherited nasals undergo and also the emergence of retroflex fricatives is due to some characteristic and salient phonological developments of Chachapoyas Quechua. Specifically, the retroflex fricatives are the results of contact with /r/ after the elision of unstressed vowels (Taylor, 1999: 35). The nasal splits are

6.3 Substrate Effects

129

conditioned by the loss of entire suffixes which only leave a trace of their former presence by altering the quality of preceding nasals (Taylor, 1999: 32). Taylor (1999: 24) links the suppression or weakening of vowels to another salient characteristic of Chachapoyas Quechua, namely the assignment of primary stress to the initial syllable. These changes have been suggested as possible substrate features that emerged after the split of the last common ancestor of San Martín and Chachapoyas Quechua, which is why they merit some discussion. But first, to give an impression of the magnitude of the cumulative changes that give Chachapoyas Quechua its distinctive character, the example in (3.) is an excerpt from a story told to Taylor (1996: 98–99) by Santiago Puerta from La Jalca. Taylor has reconstructed the morphophonologically underlying forms which follows the surface forms in a separate line. (3.)

che:bna

ko:shk,

ko:shpak

kuyéshka,

kuyeshna

‘kunam

{chay-pi-na

kawa-shka

kawa-shpa-ka

kuya-shka

kuya-sh[pa]-na

kunan-mi}

shutpk

imashnapshna

kachun mana

ñukk

wasitak

rishach,

{shuti-pi-ka

ima-shina-pish-na

ka-chun mana

ñuka-ka

wasi-ta-ka

ri-sha-chu}

ke:bna

ñuk

tiyasha’

nishna (-)

vaya

warmtak

kuyashpak

{kay-pi-na

ñuka

tiya-sha

ni-sh[pa]-na

[vaya]

warmi-ta-ka

kuya-shpa-ka}

‘There he saw her, and seeing her he loved her, loving her he said ‘now, truly, may happen what will happen, I won’t go to my house, I will stay her.’ Oh my, loving the woman…’.9

The conditions which trigger vowel reduction and suppression in Chachapoyas Quechua are only beginning to be understood. According to Taylor (1999: 24), vowel suppression mostly affects the high vowels /i/ and /u/ and only rarely /a/, namely in (i) the ultima of a disyllabic word, (ii) the middle syllable in trisyllabic words, (iii) the penultimate of a trisyllabic or quadrasyllabic word, and (iv) the vowels of suffixes, including in this case /a/. Suppressed vowels may resurface in adjacent syllables. There also is a tendency to delete entire non-initial syllables with /p/ in the onset, sometimes leaving a trace in the form of bilabialization of preceding nasals (Taylor, 1999: 26). Hintz (2006: 498) reports for Southern Conchucos (Ancash) Quechua devoiced vowels in unstressed syllables, too, especially in word-final position, but not consistently (and only in spontaneous speech, not in the careful pronunciation of controlled elicitation). However, devoicing of unstressed medial vowels is much less frequent, and complete vowel elision is not reported by Hintz (2006). In Lambayeque Quechua as described by Escribens Trisano (1977: 56–57), vowel elision goes further: in extremely careful wordlist pronunciation, final vowels are retained, whereas already in somewhat faster speech, the final vowels are devoiced when preceded by a voiceless consonant, and in the normal speech tempo as well 9

Taylor (1996: 98–99) points out that the characteristic monophthongization of Chachapoyas Quechua has interesting stylistic effects (note also that the above passage is characterized by tail-head linkage, a discourse strategy that is frequent in Amazonia).

130

6 Chachapoyas

as in allegro forms they are entirely lost. Vowels in non-initial unstressed syllables are also elided entirely between voiceless consonants (Escribens Trisano, 1977: 57). Chachapoyas Quechua is thus not alone in devoicing or eliding unstressed vowels, it just takes it further than its Conchucos and Lambayeque sister lects. According to preliminary acoustic measurements summarized by Valqui (2018: 96), Chachapoyas Quechua stress assignment functions very similarly to that described by Hintz (2006) for Conchucos Quechua, with the primary stress on the initial syllable and secondary prosodic peaks on alternating syllables counting leftward from the penultimate (see also Cerrón-Palomino, 2003 [1987]: 259). Taylor’s description would also be consistent with the construction of feet from the right edge of the word. However, Valqui (2018) notes that secondary penultimate stress is more frequently observed for verbs than for nouns and, contrary to what is observed for Conchucos, there are also cases of primary stress on the penultimate in Chachapoyas Quechua. Together with comparative evidence from other Quechuan lects, including that of Conchucos, Valqui (2018) considers the Chachapoyas stress pattern as a retention from proto-Quechua rather than a substrate-induced innovation (a possibility already alluded to by Cerrón-Palomino, 2003 [1987]: 259). Figure 6.3 from Valqui (2018: 88, Fig. 5), appears to illustrate this. However, in Fig. 6.3, we observe a reduction, rather than emphasis, of the vowel of the penultimate. Indeed, Valqui’s (2018: 84, Table 1) acoustic measurements show that unstressed vowels (in the second or third syllable of a trisyllabic word) are shorter when compared to the ones in the stressed initial syllable, or else are devoiced or completely elided. The waveform and spectrogram in Fig. 6.4 shows a strongly reduced vowel /u/ in the plural suffix -kuna. However, this finding is inconsistent with the statement that secondary stress is assigned to the penultimate, spreading leftward to alternating syllables. In sum, the rules for stress assignment and vowel reduction and elision appear not to be completely understood at the present state of research. There is a whole set of factors that may interfere and blur the picture. One is morphological. Taylor (2000: 79), for instance, says that the plural morpheme -kuna generally reduces to -kna in Luya and La Jalca and to -kun in the lects of the Upper Imaza. In the example waveforms in Figs. 6.3 and 6.4, one can observe reduction of the high vowel in -kuna regardless of its position within the word, confirming that vowel reduction is to some extent Fig. 6.3 Waveform for the word runakunata ‘these men (acc.),’ from Valqui (2018: 88, Fig. 5)

6.3 Substrate Effects

131

Fig. 6.4 Chachapoyas Quechua waveform and spectrogram for the phrase che: pirkakuna ‘these walls,’ from Valqui (2018: 87, Fig. 4)

associated firmly with certain suffixes. In Taylor’s (1996) texts, also -shpa regularly seems to reduce to -sh. A further potentially interfering factor is resolution of the so-called “binary plus clash” condition which Hintz (2006) reports for Conchucos Quechua, in which two adjacent syllables at the beginning of the word bear stress because the binary leftward-counting rule of secondary stress placement assigns prominence to the second syllable of the word. Finally, bearing in mind the ongoing obsolescence of Chachapoyas Quechua and the continually shrinking number of speakers, idiolectal factors would also have to be considered. Cholón did feature vowel suppression rules, too. According to AlexanderBakkerus (2005: 121), “[i]n a form consisting of three syllables or more, the vowel of the second syllable is often suppressed whenever the syllable structure permits it,” which is the case if the resulting consonant sequences can be syllabified in such a way that no illicit syllable structures arise. Alexander-Bakkerus (2005: 121) notes further that “[i]n a noun + verb compound, the vowel of the second and the fourth syllable can be suppressed.” Thus, in Cholón, *a-ts ala ‘my wife’ surfaces as a-t s la, and *a-manayaˇc-aŋ ‘I watch the road’ surfaces as amnayˇcaŋ (Alexander-Bakkerus, 2005: 121). This suggests the presence of some sort of rhythmic alternation of syllable prominence that is related to stress reduction, though probably with different properties than Conchucos and likely also than Chachapoyas Quechua. To investigate the factors that trigger vowel suppression and vowel elision in Chachapoyas Quechua further, and to assess if a case for influence of a system of stress assignment like that in Cholón on Chachapoyas Quechua can be made, I have analyzed the passages of text for which Taylor (1996: 98–101) provides morphophonologically underlying representations. Note that the only variable that can be evaluated on the basis of these materials is whether a morphophonologically underlying vowel is represented orthographically also on the surface or not. It is not possible to tell if vowels that are not were completely elided, devoiced, weakly articulated, or had some other phonetic property that led Taylor (1996) to not represent them in the surface forms. I have ignored cases in which the underlying sequence / awa/ reduces to [o:] because it is impossible to decide if the surfacing long vowel reflects the vowel of the first or the second syllable of the underlying form. I have also ignored Spanish insertions such as vaya ‘oh my!’ which appear to be instances of code-switching rather than integrated loanwords, but I have taken into account vowel

132

6 Chachapoyas

elisions and retentions in Spanish-derived roots that are more firmly integrated as evidenced by the fact that they receive Quechua suffixes. I have counted haplologies as instances of vowel elision. I have first investigated if the consonantal environments have an influence on vowel reduction. Table 6.4 shows the counts for different environments: word-initially, word-finally, between voiced consonants (represented by Cv in Table 6.4), between voiceless consonants (represented by Co ), between a voiceless and a voiced consonant, and between a voiced and voiceless consonant. I have not found an effect of the voicing or voicelessness of the preceding consonant on vowel reduction (χ2 = 2.0171, p ≈ 0.16) nor the following consonants (χ2 = 0.266, p ≈ 0.61). Then, I have assessed whether the vowel elisions in the texts can be predicted if one assumes a leftward assignment of (secondary) stress due to foot construction from the penultimate with an extrametricality condition like in Conchucos Quechua (as discussed, Valqui’s, 2018 data do not seem to be consistent in showing this). Table 6.5 presents the relevant counts. Consistent with the general typological tendency for stress-timed languages to reduce vowels in unstressed syllables (Auer, 1993), there was a significantly higher frequency of stress reduction in unstressed syllables (ultima, antepenultimate, 5th-tolast) than in stressed ones (penultimate, 4th-to-last, 6th-to-last). This is significant at p < 0.01 (χ2 = 0.0045). This result may be influenced by the fact that the antepenult, 4th-to-last, 5th-to-last, and 6th-to-last syllables in a significant number of cases bear primary stress. However, the correlation stays significant and even becomes stronger when only the ultima and penult are considered. The ultima is significantly more likely to be elided than the antepenult (χ2 = 10.9399, p < 0.001). I have then conversely counted vowel elision and retention rates from the left edge of the word to the right, because vowel reduction in Cholón as described by Alexander-Bakkerus (2005) appears to operate moving rightward from the left edge. Table 6.4 Vowel elision and retention in some Chachapoyas Quechua texts depending on environment #_

_#

Cv _Cv

Co _Co

Co _Cv

Cv_Co

Elision

0

29

2

10

6

3

Retention

9

15

12

31

16

28

Table 6.5 Vowel elision and retention in some Chachapoyas texts depending on stress patterns (expected values in italics in parentheses) Assumed unstressed syllables (ultima, antepenult, 5th-to-last)

Assumed stressed syllables (penultimate, 4th-to-last, 6th-to-last)

Elision

38 (29.81)

12 (20.19)

Retainment

58 (66.19)

53 (44.81)

6.4 Substrate Vocabulary

133

Table 6.6 Vowel elision and retention in some Chachapoyas texts depending on stress patterns First

Second

Third

Fourth

Fifth

Sixth

Elision

0

18

19

11

5

0

Retention

50

31

21

7

2

2

The distribution is, as can be verified in Table 6.6, strongly skewed in that— unsurprisingly given the uncontroversial and consistent prominence peak on the initial syllable—, the vowel of the first syllable is never ever deleted. To assess whether there is a significant difference between vowel elision and retention rates across syllables counting from the left, one must disregard the initial syllable. Then, there is no appreciable statistically difference between syllables (χ2 = 6.913, p = 0.14). Bearing in mind the lack of complete clarity of the nature of stress assignment and vowel reduction in both Chachapoyas Quechua and Cholón, at present, no case for a substrate influence from Cholón on Chachapoyas Quechua can be made in this domain. In addition, devoicing and elision of unstressed vowels, as Escribens Trisano (1977: 58) has already observed in the context of his discussion of the phenomenon in Lambayeque Quechua, are phonetically natural processes of assimilation and weakening, which makes any attempt to explain them through substrate interference a particularly difficult task as per Thomason (2009). At the same time, it is notable that it is exactly two of the Northern Peruvian Quechua varieties, embedded into a linguistic landscape of languages with an increased incidence of monosyllabic roots and closed syllables vis-à-vis Quechuan (Urban, 2019a, 2019b), that showcase strong tendencies toward vowel reduction. The adaptation of Quechuan numerals to Quingnam, which assimilate the disyllabic Quechuan pattern to the monosyllabic “heavy” structure of the Northern Peruvian non-Quechuan languages (Urban, 2019a), is also relevant here. It is, of course, possible, even plausible, that contact with Northern Peruvian languages has strengthened the natural tendency to weaken unstressed vowels in Lambayeque and Chachapoyas Quechua to the point of their disappearance. But a case for substrate influence in the narrow sense from a concrete language cannot be made.

6.4 Substrate Vocabulary Taylor (1990: 129) says that a reconstruction of the Chacha lexicon is an even more difficult challenge than the reconstitution of its phonological system. Indeed, just like colonial sources do not mention any language other than Quechua, colonial sources also rarely mention non-Quechuan words in the context of the Chachapoyas.10 Uniquely, D’Avalos y Figueroa (1602: 124v) mentions that the Chachapoyas chiefs were called , a strange form given the absence of onset clusters in the 10

This happens occasionally elsewhere, e.g., in sources on the North Coast (Urban 2019a).

134

6 Chachapoyas

personal names from the Chachapoyas area (on which see Taylor, 1990: 126). Also local lore can only provide weak suggestions as to the possible meaning of last names. Taylor (1990: 132) reports a legend according to which people with the last name Occ (which also appears as Hoc or Oc in colonial documentation, and which is pronounced [oh]) trace their descent to the puma. Accordingly, ‘puma’ is a possible meaning of Occ (or however this item is represented orthographically in the early sources) in the original Chacha language. Bandelier (1907: 19), furthermore, relates that “[a]n Indian from Tingo, a very aged man, told me, that the inhabitants of ‘Malca’ were sorcerers called ‘Ualqui Shaos,’ and that from the pueblo of Quemia, situated fifteen miles to the west of south near the Marañón, there came a powerful wizard, a ‘Chimal Ualqui,’ who in one bound leaped to the top of a steep height called Incupuy, a short distance south of Kuélap and with ruins of ancient buildings.” Since the sequence —almost certainly identical to the common Chachapoyas last name Valqui—occurs twice in this short passage in the context of sorcerers and wizards, it invites speculation that its meaning is associated with the powerful religious specialists for which the Chachapoyas were (in)famous in colonial times (see Schjellerup, 1997: 33–34). Otherwise, one must, as in the Cañari case, draw mainly on a closer analysis of the non-Quechuan component of the lexicon. In this case, Taylor (1979) has already provided a list of items which, within Quechuan, appear to be restricted to the Chachapoyas. Among them are two items that may be related to Cholón (Urban, 2021a). Shalla ‘basketry’ can be related to Cholón šala ‘basket’ (Alexander-Bakkerus, 2005). There are at least two factors that could have contributed to the difference in the quality of the lateral. First, as Taylor (1979: 22) says, in Chachapoyas Quechua, the palatal lateral depalatalizes sporadically to [l]. On the other hand, there is evidence for dialectal differences within Cholón. In the manuscript through which Mata’s (2007 [1748]) grammar is known to us, palatal affricates have been emended by a hand that is not Mata’s and replaced by alveolars, something that occasionally also happens with laterals (Alexander-Bakkerus, 2005: 88). “In sum, the place of articulation of the laterals apparently varies both in Chachapoyas Quechua and in Cholón diatopically, diachronically, or both” (Urban, 2021a). However, a Quechuan origin (possible with subsequent borrowing into Cholón) for shalla ‘basketry’ cannot be entirely ruled out: Santo Tomás (1560) has ‘mat,’ and the Quechua I lect of the Antonio Raimondi province has shalla ‘thicket, weed’ (Parker & Chávez, 1976). A Quechuan origin can more securely be ruled out for Chachapoyas Quechua musha. This term is used to refer to the unusually light-skinned blond or redhaired individuals for which the Chachapoyas province is known and which appear to have been noted even by the earliest colonial observers such as Cieza de León (2005 [1553]: 212). I suggest to interpret musha as related to Cholón mušak ‘sun’ (AlexanderBakkerus, 2005) on the basis of similar equations in other languages of the Americas, including the Mochica language of coastal Peru (Brüning, 2004), where a female albinotic person was designated by the same word as the ‘sun’ (and a male albinotic person by the same word as the ‘moon’). While also shalla ‘basketry’ is plausible as a substrate item according to semantic criteria, there are even stronger grounds

6.5 Perspectives from Outside Linguistics

135

to believe that musha is a retention from a substratal language: after all, it is a designation for a highly salient locally occurring phenotype for which the Quechuan languages are extremely unlikely to have had a designation before its arrival in the Chachapoyas. A third potentially relevant item is enshrined in a story on a war between Lámud and Luya, told to Gerald Taylor (1996: 56–57) by Buenaventura Santillán in Lámud in the year 1968. In this story, a character is given two objects called kalu by a shaman, one containing fog and another containing a snake. Taylor (1996: 57fn27) explains that years later, in 1975, kalu was explained to be a kind of reed from which baskets can be made. In the story, however, the kalu clearly appear to be receptables, making a comparison with Cholón kaloˇc ‘plate’ not completely implausible. In the Chachapoyas case, like in the Cañari case study in Sect. 5.4, the possible search space for lexical comparisons is significantly restricted by the available documentation on the compared language(s). The total known lexicon of Cholón is restricted to approx. 600 items (Urban, 2021b), and semantically many of these reflect the fact that they come from a missionary grammar. Taylor (1990: 133) notes that Chachapoyas flora and fauna terminology comprises some plant and animal names that correspond in phonological structure to what can be observed in local placenames and personal names, suggesting that they come from the same source. Valqui Culqui (2004) mentions the plant names llep, gulgul, supalín, shashkibo, tonsho, and wísul, the bird names kepkin and talakua ~ delakua as well as solpe, the name of a carrying net, as possible Chacha lexical items. Aviva Shimelman has collected a significantly larger portion of Chachapoyas Quechua flora and fauna terminology which, however, cannot be fruitfully compared to Cholón because that portion of the lexicon is insufficiently described. Furthermore, Valqui Culqui (2003: 70) reports from his fieldwork in the Chachapoyas the “sounds” (“sonidos”) llip llip ‘fluttering eyelashes,’ llot llot ‘footsteps in mud,’ uish uish ‘frying, growing, something gushing,’ pólchok ‘something half-cooked,’ juicho juicho ‘leaving early,’ juip juip ~ uip uip ‘walking quickly,’ sorop sorop ‘drinking hot water,’ jop ~ ope ‘arriving at a house,’ and lacla ‘not having eaten well, hungry, thin, malnourished.’ Valqui Culqui (2004) adds soclón soclón ‘shaky, flabby’ and shec shec ‘more shaky than soclón soclón.’ These would appear to be ideophones or ideophone-like expressions, which are in this form not found in highland Quechua lects. Ideophones do appear, however, prominently in the lowlands, for instance in Pastaza Quichua (Nuckolls, 1996).

6.5 Perspectives from Outside Linguistics The manner by which Quechuan spread to the Chachapoyas was until very recently entirely unclear. At the same time, the arrival of people who the Incas resettled in the region has been taken as a plausible relevant factor (Heggarty & BeresfordHones, 2013: 404). However, a recent genetic study (Barbieri et al., 2017: 1) indicates that that present-day speakers of Chachapoyas Quechua show “no particular genetic

136

6 Chachapoyas

similarity or gene-flow with Quechua speakers elsewhere, suggesting that here the language spread primarily by cultural diffusion, not migration.” In other words, in the Chachapoyas, as in Bolivia, Quechuan became entrenched because people who had before spoken a local substratal language learned it as a second language, and a process of language shift that eventually displaced the original language was set in motion.

6.6 Summary The genetic evidence thus suggests a picture of language shift and replacement without a major population turnover in the Chachapoyas region. As we have seen in this case study, the toponymy of the Chachapoyas region, with its strong presence of non-Quechuan elements, is consistent with such a scenario, and relevant toponyms can safely be assumed to reflect the original pre-Quechuan substratal language of the Chachapoyas. Some toponyms and a number of recurrent endings can be analyzed through Cholón. However, in this case, the toponymic evidence also points to continuities with Northern Peru more broadly, a fact that suggests that a related language was also once spoken much more widely in Northern Peru. As far as the phonological characteristics of the pre-Quechuan toponymic layer of the Chachapoyas is concerned, it matches those Chachapoyas Quechua lexical items that are apparently restricted to this lect within the Quechuan family. This, and the fact that these lexical items cluster in relevant semantic domains, including prominently flora and fauna terminology, suggests that they may stem from the same substratal source, possibly a relative of Cholón. However, other Quechuan varieties of Northern Peru, i.e., Lambayeque and Cajamarca Quechua, do not show the same relation with Cholón in substrate vocabulary (Urban, 2021b) that can be suggested for Chachapoyas on the basis of items such as shalla, musha, and possibly laku. This could have something to do with the closer geographical proximity of the Chachapoyas to the areas where Cholón is attested historically and was documented, or else it could be an indication of different contact scenarios. A closer examination of the transformations which Chachapoyas Quechua phonology has undergone since this lect and that of San Martín diverged showed only limited support for the agency of a substrate language. On the one hand, the addition of voiced stops and other phonemes to the phonological inventory can partly be attributed to the addition of lexical items from a substratal language, something that is also true for the Quechua of Cajamarca (Adelaar & Muysken, 2004: 403, Urban, 2021b). On the other hand, it is considered possible that the stress pattern of Chachapoyas Quechua, while still not fully understood, may be a retention from proto-Quechua rather than a substrate effect (Torero, 1964: 461–462; Valqui, 2018), and also the particulars of vowel reduction and elision that so strongly characterize the Chachapoyas Quechua lects cannot at present be explained through a Cholón substrate. In sum, independent lines of evidence coming mostly from toponymy and substrate vocabulary are consistent in pointing to a Cholón affiliation for the substratal

References

137

Chacha language of the Chachapoyas, but, of the three case studies considered in this chapter, this one remains the weakest. Tacitly assuming that linguistic and cultural distributions in geographical space should cluster together, Valqui Culqui (2004) posits that the Chacha language should once have been spoken in a wider area, prominently including a southern extension in the Marañón-Huallaga Interfluve (Bolívar province, La Libertad department). Although this rests on a problematic expectation to find matches between archeological and linguistic distributions, as pointed out in (Urban to appear), in this particular case, the interpretation of Chacha Sprachdenkmäler through Cholón actually do lead to a better match between the distribution of toponyms and, by proxy, languages and archeological traits.

References Adelaar, W. F. H. (2012). Languages of the Middle Andes in areal-typological perspective: Emphasis on Quechuan and Aymaran. In L. Campbell, & V. Grondona (Eds.), The indigenous languages of South America: A comprehensive guide (pp. 575–624). Walter de Gruyter. Adelaar, W. F. H., & Muysken, P. C. (2004). The languages of the Andes. Cambridge University Press. Alexander-Bakkerus, A. (2005). Eighteenth-century Cholón. LOT. Andrade Ciudad, L. (2010). Contactos y fronteras de lenguas en la Cajamarca prehispánica. In P. Kaulicke, R. Cerrón-Palomino, P. Heggarty, & D. Beresford-Jones (Eds.), Lenguas y sociedades en el antiguo Perú: Hacia un enfoque interdisciplinario (pp. 165–180). Special Issue of Boletín de Arqueología 14. Anonymous. (1865) [~1560]. Relación de la religión y ritos del Perú, hecha por los primeros religiosos Agustinos que allí pasaron para la conversión de los naturales. In J. F. Pacheco, D. Francisco de Cárdenas, & D. Luis Torres de Mendoza (Eds.), Colleción de documentos inéditos relativos al descubrimiento, conquista y colonización de las posesiones españolas en América y Occeanía, sacados, en su major parte, del Real Archivo de Indias (Vol. 3, pp. 5–58). Manuel B. de Quirós. Auer, P. (1993). Is a rhythm-based typology possible? A study of the role of prosody in phonological theory. KontRI Working Paper No. 21. Bandelier, A. F. A. (1907). The Indians and aboriginal ruins near Chachapoyas in Northern Peru (explorations made in 1893 under the patronage of the late Mr. Henry Villard, to the memory of whom this monograph is respectfully dedicated). Historical Records and Studies. Barbieri, C., Sandoval, J. R., Valqui, J., Shimelman, A., Ziemendorff, S., Schröder, R., Geppert, M., Roewer, L., Gray, R., Stoneking, M., Fujita, R., & Heggarty, P. (2017). Enclaves of genetic diversity resisted Inca impacts on population history. Scientific Reports, 7, 17411. https://doi. org/10.1038/s41598-017-17728-w Brüning, H. H. (2004). Mochica Wörterbuch: Diccionario mochica. Mochica – castellano / castellano – mochica (José Antonio Salas García, Ed.). Universidad de San Martín de Porres, Escuela Profesional de Turismo y Hotelería. Cerrón-Palomino, R. (1990). Reconsideracion del llamado ‘quechua costeño.’ Revista Andina, 8(2), 335–386. Cerrón-Palomino, R. (2003) [1987]. Lingüística quechua (2nd ed.). Centro de Estudios Regionales Andinos ‘Bartolomé de las Casas.’

138

6 Chachapoyas

Church, W. B., & von Hagen, A. (2008). Chachapoyas: Cultural development at an Andean Cloud Forest crossroads. In H. Silverman, & W. H. Isbell (Eds.), Handbook of South American archaeology (pp. 903–926). Springer. Cieza de León, P. (2005) [1553]. Crónica del Perú: El Señorío de los Incas (Franklin Pease (Ed.)). Biblioteca Ayacucho. D’Avalos y Figueroa, D. (1602). Primera parte de la miscelánea austral de Don Diego D’Avalos y Figueroa, en varios coloquios, interlocutores, delio, y cilena. Antonio Ricardo. Middendorf, E. W. (1895). Peru: Beobachtungen und Studien über das Land und seine Bewohner während eines 25jährigen Aufenthalts. Das Hochland von Peru (Vol. 3). Robert Oppenheim (Gustav Schmidt). Latorre, S. (2014). Type de langue et certains emprunts sur l’ordre de constituants en Cholón. Paper presented at the Sedyl Ph.D. seminar “Ordre des constituants et contact de langues,” June 2–6 2014, Villejuif. https://f-origin.hypotheses.org/wp-content/blogs.dir/1815/files/2014/05/Typede-langue-et-certains-emprunts-sur-lordre-des-constituants-en-cholon.pdf. Accessed 31 March 2023. Valqui, J., & Ziemendorff, M. (2016). Vestigios de una lengua originaria en el territorio de la cultura chachapoya. Letras, 87(125), 5–32. https://doi.org/10.30920/letras.87.125.1 de Mogrovejo, T. (2006) [1593–1605]. Libros de visitas de Santo Toribio Mogrovejo (1593–1605) (José Antonio Benito (Ed.)). Fondo Editorial de la Pontificia Universidad Católica del Perú. de la Mata, P. (2007) [1748]. Arte de la lengua cholona (Astrid-Alexander Bakkerus (Ed.)). Iberoamericana/Vervuert. Escribens Trisano, A. (1977). Fonología del quechua de Ferreñafe [Ph.D. dissertation]. Universidad Nacional Mayor de San Marcos. Espinoza Soriano, W. (1977). La poliginia señorial en el reino de Caxamarca, siglos xv y xvi. Revista del Museo Nacional, 63, 399–466. Heggarty, P., & Beresford-Jones, D. (2013). Andes: Linguistic history. In P. Bellwood (Ed.), The global prehistory of human migration (pp. 401–409). Wiley Blackwell. Hintz, D. M. (2006). Stress in South Conchucos Quechua: A phonetic and phonological study. International Journal of American Linguistics, 72(4), 477–521. https://doi.org/10.1086/513058. Landerman, P. N. (1991). Quechua dialects and their classification [Ph.D. dissertation]. University of California, Los Angeles. Martínez Compañón, B. J. (1985) [1782–1790]. Trujillo del Perú en el siglo xviii (Vol. 2). Ediciones Cultura Hispánica. Nuckolls, J. B. (1996). Sounds like life: Sound-symbolic grammar, performance, and cognition in Pastaza Quechua. Oxford University Press. Parker, G. J. (1969). Comparative Quechua phonology and gramar III: Proto_Quechua lexicon. University of Hawaii Working Papers in Linguistics, 1(4): 1–61. Parker, G. J., & Chávez, A. (1976). Diccionario quechua: Ancash-Huailas. Ministerio de Educación/ Instituto de Estudios Peruanos. Pärssinen, M., Siiriäinen, A., & Korpisaari, A. (2003). Fortifications related to the Inca expansion. In M. Pärssinen, & A. Korpisaari (Eds.), Western Amazonia–Amazônia occidental: Multidisciplinary studies on ancient expansionistic movements, fortifications and sedentary life (pp. 29–72). Renvall Institute for Area and Cultural Studies, University of Helsinki. de la Vega, G. (1829) [1609]. Primera parte de los comentarios reales, que tratan del origen de los Incas, reyes que fueron del Perú, de su idolatría, leyes y gobierno, en paz y en guerra de sus vidas y conquistas, y de todo lo que fue aquel imperio y su república antes que los españoles pasáran a él. Hijos de Doña Catalina Piñuela. Pinho De Valhery Jolkesky, M. (2016). Estudo arqueo-ecolinguístico das terras tropicais sulamericanas [Ph.D. dissertation]. Universidade de Brasília. Renard-Casevitz, F. M., Saignes, T., & Taylor, A. C. (1988). Al este de los Andes: Relaciones entre las sociedades amazónicas y andinas entre los siglos xv y xvii (Trans. Juan Carrera Colin and rev. Gonzalo Flores and Olinda Celestino). Abya-Yala.

References

139

Rivarola, J. L. (2007). Antroponimia chachapoya. In La Romania en interacción: Entre historia, contacto y política. In M. Schrader-Kriffki, & L. Morgenthaler García (Eds.), Ensayos en homenaje a Klaus Zimmermann (pp. 295–305). Iberoamericana/Vervuert. Salomon, F., & Grosboll, S. (1986). Names and peoples in Incaic Quito: Retrieving undocumented historic processes through anthroponymy and statistics. American Anthropologist, 88(2), 387– 399. https://doi.org/10.1525/aa.1986.88.2.02a00060 De Santo Tomás, D. (1560). Lexicon, o vocabulario de la lengua general del Perú. Francisco Fernandez de Cordoua. Schjellerup, I. R. (1997). Incas and Spaniards in the conquest of the Chachapoyas: Archaeological and ethnohistorical research in the North-eastern Andes of Peru. Göteborg University, Department of Archaeology. Seitz, G. (2017). Cultural discontinuity: The new social face of the Awajún. Amakella. Taylor, G. (1979). Diccionario normalizado y comparativo quechua: Chachapoyas-Lamas. L’Harmattan. Taylor, G. (1990). La lengua de los antiguos Chachapuyas. In R. Cerrón-Palomino, & G. Solís Fonseca (Eds.), Temas de lingüística amerindia: Primer Congreso Nacional de Investigaciones Lingüístico-Filológicas (pp. 121–139). Consejo Nacional de Ciencia y Tecnología/Programa de Educación Bilingüe de Puno. Taylor, G. (1996). La tradición oral quechua de Chachapoyas. Instituto Francés de Estudios Andinos/Asociación ATOQ. Taylor, G. (1999). Estudios de dialectología quechua (chachapoyas, ferreñafe, yauyos). Editorial Universitaria (EUNE), Universidad Nacional de Educación. Taylor, G. (2000). Estudios lingüísticos sobre Chachapoyas. Instituto Francés de Estudios Andinos/ Fondo Editorial de la Universidad Nacional Mayor de San Marcos. Thomason, S. G. (2009). How to establish substratum interference. In Y. Nagano (Ed.), Issues in Tibeto-Burman historical linguistics (pp. 319–329). National Museum of Ethnology. Torero, A. (1964). Los dialectos quechuas. Anales Científicos de la Universidad Agraria, 2(4), 446–478. Torero, A. (1989). Areas toponímicas e idiomas en la sierra norte peruana: Un trabajo de recuperación lingüística. Revista Andina, 7(1), 217–257. Torero, A. (1993). Lenguas del nororiente Peruano: La hoya de Jaén en el siglo xvi. Revista Andina, 11(2), 447–472. Torero, A. (2002). Idiomas de los Andes: Lingüística e historia. Instituto Francés de Estudios Andinos/Editorial Horizonte. Urban, M. (2019a). Lost languages of the Peruvian North Coast. Ibero-American Institute/Gebr. Mann. Urban, M. (2019b). Is there a Central Andean language area? A view from the “minor” languages. Journal of Language Contact, 12(2), 271–304. https://doi.org/10.1163/19552629-01202002 Urban, M. (2021a). Cholón and the linguistic prehistory of northern Peru: Triangulating toponymy, substrate lexis, and areal typology. Linguistic Discovery, 17(1). https://doi.org/10.1349/PS1. 1537-0852.A.513 Urban, M. (2021b). The foreign contribution to the lexicon of Northern Peruvian Quechua varieties: An exploratory study. Language Dynamics and Change, 11(1), 77–129. https://doi.org/10.1163/ 22105832-00902007 Urban, M. (To appear). Small and extinct languages of Northern Peru. In M. Urban (Ed.), The Oxford Guide to the Languages of the Central Andes. Oxford University Press. Valqui Culqui, J. (2003). Rastreo a la extinta lengua de los Chachapoyas. Lengua y Sociedad, 5, 62–71. Valqui Culqui, J. (2004). Reconstrucción de la lengua Chacha mediante un estudio toponímico en el distrito de La Jalca grande (Chachapoyas-Amazonas) [Licenciate thesis]. Universidad Nacional Mayor de San Marcos. Valqui, J. (2018). El acento en el quechua de Chachapoyas: Un sustrato de la lengua chacha o un vestigio protoquechua. Letras, 89(130), 79–99. https://doi.org/10.30920/letras.89.130.4

140

6 Chachapoyas

Zariquiey Biondi, R. (2004). Fonología del Quichua del Napo: Una aproximación a su sincronía y a su historia. Boletín del Instituto Riva-Agüero, 31, 291–320. Zevallos Quiñones, J. (1966). Onomástica prehispánica de Chachapoyas. Lenguaje y Ciencias, 20, 27–41.

Chapter 7

Synopsis and Conclusion

Abstract This concluding chapter provides a comparative synopsis of the results obtained in the three case studies and a summary evaluation of the results. It compares the three case studies according to several parameters of variation, showing that in all three study regions, there is evidence for non-Quechuan toponymy and also structural changes in the local Quechua lects, in particular in the phonology. However, the results differ with regard to the confidence with which we can attribute these to contact-influence of the substrate kind, and also the degree to which the lexicon has been enriched by an unrelated language. Most importantly, the confidence with which one can identify a particular language or language family with which the replaced language may have been affiliated varies from strong in the case of Southern Peru, intermediate in the Southern Ecuadorian highlands, to weakest in the case of the Chachapoyas region. Keywords Linguistic stratigraphy · Sprachdenkmäler · Language shift · Toponyms · Substrate · Contact-induced language change

The preceding three chapters have evaluated the evidence for Sprachdenkmäler of substratal languages in three regions of the Central Andes: Southern Peru, the southern Ecuadorian highlands, and the Chachapoyas cloud forests of Northern Peru. The characteristics of the local toponymy, structural peculiarities of the Quechuan lects spoken in these regions, and vocabulary items that qualify as possible remnants from the preceding non-Quechuan languages were taken into account. We have seen some commonalities between the regions, but also some differences. In Southern Peru and the southern Ecuadorian highlands, early colonial sources mention a non-Quechuan language still spoken alongside a form of Quechuan— Aymara in Southern Peru and the undocumented Cañari language in southern Ecuador. However, also for the Chachapoyas region we have seen strong evidence for non-Quechuan toponymy. Only in the Southern Peruvian case study this is unequivocally explainable through a known language. In the southern Ecuadorian Highlands, the distribution of characteristic toponyms that sometimes showcase extensions to the Ecuadorian highlands as a whole, sometimes to the Ecuadorian-Colombian border area specifically, suggest a Barbacoan substratal layer for the Ecuadorian highlands as © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Urban, Linguistic Stratigraphy, https://doi.org/10.1007/978-3-031-42102-0_7

141

142

7 Synopsis and Conclusion

a whole, though in this case, the Achilles Heel is that relevant toponymic endings still need to be etymologized through extant Barbacoan languages. In the Chachapoyas, the patterning of toponymic endings is generally similar in that some are restricted to the Chachapoyas narrowly, but others have a much wider extension in northern Peru as a whole. Three of these endings can be explained through a language that was still spoken in historical times on the eastern slopes of the Northern Peruvian Andes, Cholón. In this case, an issue that is also noted for southern Ecuador is even more virulent and problematic: the relevant toponymic endings are surprisingly thin on the ground in the area where Cholón is known to have been spoken in historical times. In all three cases, we have good evidence for changes in the phonological system of the local Quechuan lects that can be explained, at least partially, through the adoption of non-Quechuan vocabulary from a pre-existing non-Quechuan language. In the Southern Peruvian case study, newly introduced phonemes have spread to native vocabulary, something that does not seem to be the case in Cañar Quichua (though the vagaries of the nonphonemic orthography of the consulted Cañar Quichua dictionary leave open the possibility of some minor occurrences of newly introduced phonemes in native Quechuan material) or Chachapoyas Quechua. An important observation is that in both cases, the phonology and phonotactics of the non-Quechuan toponymy and the structural changes by which the phonological systems have become enriched are consistent, and thus point to a common origin in the same or similar substratal language(s) (in Southern Peru, the same may well be true, though this is obscured by the deletion of the nonmodal features of glottalization and aspiration as toponyms become adopted to Spanish). In all three Quechuan lects, one can find some amount of vocabulary that does not derive from proto-Quechua, but could have been acquired from a local source. In some cases, the phonology and phonotactics of these items reinforce this idea. In the Southern Peruvian case, it is doubtful given the semantics of relevant items and the continuing contact situation between Quechuan and Aymara whether Aymara loanwords in Cuzco Quechua proper were transferred during language shift. Semantic domains in which items without clear Quechuan sources cluster are more suggestive of this in southern Ecuador and the Chachapoyas, though also here, an alternative scenario according to which borrowing occurred in a situation of contact that preceded shift cannot be ruled out. Source forms in relatives of the suspected substratal languages—the Barbacoan and Cholón languages respectively—proved difficult to find, especially for the Chachapoyas case. This may have something to do with the limited amount of lexical documentation of Barbacoan languages and Cholón. All things considered, one can say that, for all three areas, it is very likely that Quechuan lects were influenced by a local non-Quechuan language, but that the identity of that language can only be firmly asserted for Southern Peru, where, in spite of the problematic aspects of the situation discussed in Sect. 4.3, a former presence of an Aymara language seems strongly indicated. The Barbacoan languages may have extended to the Cañar -Azuay region in the southern Ecuadorian highlands, but the evidence for this is weaker, and still weaker is the evidence for an affiliation of Chacha, the substratal language of the Chachapoyas, with Cholón. At the same

7 Synopsis and Conclusion

143

Table 7.1 Qualitative synopsis of evidence for substratal languages in the three regions of the Andes Southern Peru (Chap. 4)

Southern Ecuadorian Highlands (Chap. 5)

Chachapoyas (Chap. 6)

Non-Quechuan languages mentioned in colonial sources

Yes

Yes

No

Non-Quechuan toponymy

Yes

Yes

Yes

Structural changes

Yes

Yes

Yes

Influx of vocabulary

Some, but not only

Some

Little

Including relevant domains? Yes, but limited

Yes

Yes

Data situation

Intermediate

Poor

Overall strength of evidence Good

Intermediate

Weak

Extralinguistic evidence for language shift?

No

Yes

Very good Yes

time, these are currently the best hypotheses, and it is left for future research to unearth further evidence for these affiliations or discard them. Table 7.1 summarizes the evidence form the different types of Sprachdenkmäler for the three regions in a comparative manner. The Ecuadorian and Chachapoyas cases have also raised the question of how to deal with the possibility of a shared substrate in Ecuadorian Quichua as a whole and the northern Peruvian Quechuan lects at large. In the former case, the relevant evidence concerned some structural properties that are suggestive of Barbacoan influence in Ecuadorian Quichua as a whole, but also the lexicon, more specifically tree names in -chi that suggest a Barbacoan origin. Pilchi, for instance, is a tree name also attested in the lowland varieties of the oriente, where, for all we know, Barbacoan languages were never spoken. Such observations are suggestive of a relatively early wave of Barbacoan influence on an early form of Ecuadorian Quichua, possibly overlain by more recent local layers of influence. The Sprachdenkmäler found in lects like that of Cañar-Azuay and Imbabura may partly belong to the former, partly to the latter layer. As far as Northern Peru is concerned, we have seen how vowel elision in Chachapoyas Quechua is to some extent also reflected in Lambayeque Quechua. While this is by no means indicative of a shared substrate, it may more plausibly reflect substrate languages of the type that are common in Northern Peru which feature an increased number of monosyllabic roots and heavier syllables vis-à-vis Quechua and Aymara (Urban, 2019a, 2019b). In sum, while the evidence is nowhere entirely unequivocal, the discussed data strongly support a former presence of Aymara (or closely related lects) in much of the regions where Quechua IIC lects are spoken today. The data suggest, less strongly, that Barbacoan languages may have extended to the Southern Ecuadorian Highlands, and, still less strongly, that one or more languages related to the historically attested Cholón language were spoken not only in the Chachapoyas region, but also

144

7 Synopsis and Conclusion

throughout much of the northern Peruvian highlands. One final word of caution is in order. In all three cases, the analyses presented build on hypotheses already formulated in earlier work. Accordingly, there was an inherent bias towards Aymara in the Southern Peruvian case study in Chap. 4, an inherent bias towards Barbacoan in the southern Ecuadorian case study in Chap. 5, and an inherent bias towards Cholón in the Chachapoyas case study in Chap. 6. One might thus ask if similar results could not be obtained for any suggested affiliation. As far as the substrate lexis of Cañar Quichua is concerned, a test search for similarities between the relevant portion of the lexicon of Cañar Quichua with the unrelated Páez language, spoken in roughly the same region as (Northern) Barbacoan languages, yielded results that were judged impressionistically to be inferior to those obtained for Barbacoan (Urban, 2018). This is a good sign, but of course the results cannot be generalized. It is accordingly possible that, on the basis of comparisons with other languages, better hypotheses as to the identity of the substrate languages that underlie the most recent stratigraphic layer of Central Andean language history could be formulated.

References Urban, M. (2018). The lexical legacy of substrate languages: A test case from the Southern Ecuadorian highlands. Transactions of the Philological Society, 116(3), 435–459. https://doi.org/10. 1111/1467-968X.12129 Urban, M. (2019a). Lost languages of the Peruvian North Coast. Ibero-American Institute/Gebr. Mann. Urban, M. (2019b). Is there a Central Andean language area? A view from the “minor” languages. Journal of Language Contact, 12(2), 271–304. https://doi.org/10.1163/19552629-01202002

Index

A Andean Spanish, 11 Argentina, 7, 14, 15, 19, 71 Atacameño. See Kunza Aymaran Aymara, 10, 15, 17, 30–33, 45–67, 78, 79, 99, 107, 108, 116, 120, 141–144 Jaqaru, 10, 53, 55, 59–64, 67

B Bagua, 18 Barbacoan, 9 Awa Pit, 9, 19, 73, 75, 76, 89–94, 96–98, 102, 103, 105, 107, 109 Cha’palaa, 9, 19, 73, 76, 97, 102, 105, 107, 108 Guambiano, 9, 19, 73, 96, 98, 102, 103, 105, 107 Nam Trik, 9, 19, 73, 98, 102, 103, 108 Totoró, 9, 19, 73, 96, 98, 102, 103, 105, 107 Tsafiqui, 9, 19, 73, 76, 97, 99, 102, 105, 107, 108 Bilingualism, 29, 38, 65, 120 Bolivia, 3, 7 Chuquisaca, 47 Cochabamba, 47 Lake Popoó, 17 Lake Titicaca, 10, 17 Oruro, 47

C Cañari, 18, 19, 77, 78, 80–83, 92, 94, 96, 97, 99, 102, 108, 109, 119, 121, 126, 128, 134, 135, 141 Cara, 18, 27, 76, 77, 82, 84, 89, 95, 99 Chacha, 18, 115, 116, 120, 123, 125, 127, 128, 133, 135, 137, 142 Chicham, 18, 77, 87 Chile, 10 Atacama desert, 15 Chirino, 18 Chocoan, 9, 73 Epena Pedee, 9 Cholón, 18, 31, 33, 115, 116, 118, 122–126, 131–137, 142, 143 Classification, 5, 7, 11, 12, 77 Cofán, 9 Cognate, 2, 39, 41, 60, 101, 102, 128 Colombia, 7, 9, 12, 71, 73, 80, 82, 85–88, 91, 92, 94, 97, 108 Cauca, 88 Chocó, 88 Huila, 88 Nariño, 73, 87, 88, 91, 93, 94, 97 Tolima, 87, 88 Valle del Cauca, 88 Copallín, 18, 122 Culli, 17, 28, 29, 31, 32, 34, 40, 122, 125 Culture history, 5–7

D Diaguita, 15

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. Urban, Linguistic Stratigraphy, https://doi.org/10.1007/978-3-031-42102-0

145

146 E Ecuador, 3, 143 Azuay, 71, 73, 77–100, 103, 108, 109, 142 Cañar, 71, 73, 77–98, 100, 103, 108, 109, 142 Carchi, 74, 76, 85, 91, 93, 97 Chimborazo, 72, 73, 77, 78, 82, 85–88, 91, 92 Cotopaxi, 73, 76, 85–88, 93, 95 El Oro, 77, 78, 80, 82, 86–88, 91 Guayas, 78, 87, 88, 94 Huila, 86, 92 Imbabura, 9, 27, 72, 73, 76, 85, 87, 88, 93, 98, 99, 143 Loja, 9, 18, 73, 77, 78, 80, 82, 84–88, 91, 94, 96, 107 Los Ríos, 88 Manabí, 73, 88 Morona Santiago, 85–88 Napo, 9, 72, 85, 87, 88, 116 oriente, 9, 95, 103, 109, 143 Pastaza, 9, 72, 88, 135 Pichincha, 73, 76, 85–88, 96 Santa Elena, 85, 86, 88 Sucumbíos, 85–88, 96 Tungurahua, 73, 76, 77, 85–88, 93 Zamora Chinchipe, 87 Etymology, 30, 31, 38, 47, 59, 78, 80, 94, 97, 101, 102, 115, 120, 124 Extralinguistic evidence, 40

G Genetics, 66 Glottochronology, 13

H Hibito, 18 Humahuaca, 15

I Inca, 6, 7, 13, 20, 31, 37, 45, 47, 53, 58, 72, 78, 80, 97, 99, 115, 118–120, 126 Interference through shift. See substrate influence

J Jaén, 77 Jivaroan. See Chicham

Index K Kakán, 15, 20 Kunza, 15

L Language family, 3, 5–7, 11–13, 18, 45, 53, 73, 94, 98, 109, 141 Late Horizon, 6, 7, 13

M Marañón, 134 Mochica, 17, 30, 31, 33, 77, 134 Morphology, 39

P Panzaleo, 18, 76, 77, 82, 89, 91, 95, 96 Pasto, 9, 18, 74, 76, 77, 84, 89–92, 96–98, 109 Patagón, 18 Peru, 7 Amazonas, 118, 121 Ancash, 9, 33, 121, 122, 125 Apurímac, 46, 49–51, 107 Arequipa, 47, 50, 65 Ayacucho, 12, 31, 46, 47, 49, 50, 52–54, 56, 57, 64–66, 98 Cajamarca, 122 Chachapoyas, 3, 115, 116, 118–121, 123–125, 128, 134–136, 141–143 Chachapoyas region, 9, 34, 36, 37, 81, 95, 97, 108, 116, 118, 121–125, 128, 143 Cuzco, 7, 9, 12, 13, 31, 37, 45–47, 49–54, 57, 58, 60–66, 100 Huancavelica, 11, 12, 49 Huánuco, 118 Jaén, 18, 116, 122 Junín, 9 La Libertad, 118, 119 Lambayeque, 122 Lima, 10, 12, 33, 45, 122, 124 Loreto, 31 Moquega, 50 Pacaraos, 12 Piura, 122 Puno, 50, 53, 65 San Martín, 117 Yauyos, 12 Phonology, 3, 12, 28, 33, 36, 37, 39, 71, 80, 96, 127, 136, 141, 142 Phonotactics, 34, 37, 39, 41, 59, 94

Index Puquina, 7, 10, 15, 32 Puruhuay, 18

Q Quechuan Ancash, 78, 129 Ayacucho, 49–52, 54, 56–58, 60, 62, 102, 107 Cajamarca, 9, 12, 14, 60, 118, 136 Cañar Quichua, 82, 91, 100, 144 Chachapoyas, 9, 10, 12, 29, 30, 40, 60, 72, 115–118, 120, 121, 123, 127–137, 142–144 continuous zone, 9–12, 14, 17 Cuzco, 45, 49–54, 58, 60, 61, 65, 72, 78, 101, 102, 107, 108, 142 Cuzco-Bolivian, 50, 52–59, 65 Ecuadorian Quichua, 9, 14, 27, 38, 41, 53, 60, 72, 77, 98, 99, 101, 103–106, 109, 143 Huallaga, 60 Huanca, 60 Imbabura, 99 Inga, 9, 60, 102, 107, 116 Lambayeque, 9, 12, 60, 118, 129, 130, 133, 136, 143 lengua general, 14, 83 Pacaraos, 60 Pacaraos Quechua, 12 Puno, 53–55, 58 Quechua I, 11, 12, 54, 60, 102, 134 Quechua II, 11–14, 54 Quechua IIA, 12 Quechua IIB, 12, 60, 72, 100, 102, 117 Quechua IIC, 12, 48, 49, 52, 54, 56, 58–60, 66, 67, 71, 102, 143 San Martín, 9, 60, 72, 102, 117, 118, 127, 129, 136

147 Santiago del Estero, 15, 19, 71 Santiago del Estero Quichua, 10 Tarma, 31, 60, 61 Tarma Quechua, 10 Yauyos Quechua, 11, 12, 60 Quingnam, 17, 133

S Sácata, 18 Sechura, 17 Semantics, 29, 30, 39, 55, 103, 122, 123, 126, 135 Sprachdenkmal, 1–3, 25, 27, 35, 38, 40, 41, 46, 66, 71, 81, 82, 96, 126, 137, 141, 143 Substrate influence, 19, 25, 27, 36, 38, 41, 51, 56, 58, 65, 127, 133

T Tabancale, 18 Tallán, 17 Third Lima Council, 45, 57 Tree model, 12

U Uru-Chipaya Chipaya, 10, 17 Uru, 10, 17

W Wanderwörter, 39

X Xoroca, 18