Notes on the grammar of Andakí

Notes on the grammar of Andakí Jelien Moens Matthias Pache language science press Studies in Amerindian Linguistics 2 Studies in Amerindian Linguistics Editors: Carolina González (Florida State University) Siri Tuttle (Navajo Technical University in New Mexico) Thiago Chacon (University of Brasilia) Heriberto Avelino (National Institute for Anthropology and History) In this series: 1. Dyck, Carrie, Frances Froman, Alfred Keye & Lottie Keye. A grammar and dictionary of Gayogo hó:nǫˀ (Cayuga). 2. Moens Jelien & Matthias Pache. Notes on the grammar of Andakí. ISSN (print): 2943-0569 ISSN (electronic): 2943-0658 Notes on the grammar of Andakí Jelien Moens Matthias Pache language science press Jelien Moens & Matthias Pache. 2025. Notes on the grammar of Andakí (Studies in Amerindian Linguistics 2). Berlin: Language Science Press. This title can be downloaded at: http://langsci-press.org/catalog/book/528 © 2025, Jelien Moens & Matthias Pache Published under the Creative Commons Attribution 4.0 Licence (CC BY 4.0): http://creativecommons.org/licenses/by/4.0/ ISBN: 978-3-96110-539-7 (Digital) 978-3-98554-159-1 (Hardcover) ISSN (print): 2943-0569 ISSN (electronic): 2943-0658 DOI: 10.5281/zenodo.16964545 Source code available from www.github.com/langsci/528 Errata: paperhive.org/documents/remote?type=langsci&id=528 Cover and concept of design: Ulrike Harbort Typesetting: Sebastian Nordhoff Proofreading: Christopher Straughn, Elliott Pearl, Frederic Blum, Gerald Delahunty, Hella Olbertz, Jeroen van de Weijer, Lachlan Mackenzie, Ludger Paschen, Mary Ann Walter, Nicoletta Romeo, Patricia Cabredo, Russell Barlow, Tihomir Rangelov Fonts: Libertinus, Arimo, DejaVu Sans Mono Typesetting software: XƎL A T EX Language Science Press Scharnweberstraße 10 10247 Berlin, Germany http://langsci-press.org [email protected] Storage and cataloguing done by FU Berlin Contents Preface v List of abbrevations vii 1 Introduction 1 2 Cultural background 3 3 Sources and previous works on Andakí 7 4 Approach and challenges 13 5 Phonology 21 5.1 Phonotactics, syllable and root structure . . . . . . . . . . . . . 22 5.2 Vowels and suprasegmentals . . . . . . . . . . . . . . . . . . . . 23 5.2.1 Suprasegmentals . . . . . . . . . . . . . . . . . . . . . . 24 5.2.2 Vowelnasality ....................... 24 5.2.3 /i/,/ĩ/............................. 24 5.2.4 /u/,/ũ/............................ 25 5.2.5 /a/,/ã/............................ 25 5.2.6 Sequences of identical vowels . . . . . . . . . . . . . . . 26 5.3 Consonants.............................. 27 5.3.1 /p/ .............................. 28 5.3.2 /b/ .............................. 28 5.3.3 /m/.............................. 29 5.3.4 /t/ .............................. 29 5.3.5 /s/ .............................. 30 5.3.6 /n/ .............................. 31 5.3.7 /r/ .............................. 31 5.3.8 /ʧ/ .............................. 32 5.3.9 /k/ .............................. 32 5.3.10 /kʷ/.............................. 33 5.3.11 /x/ .............................. 34 Contents 5.3.12 /h/ .............................. 34 6 Nouns 37 7 Numerals 41 8 Demonstratives: <ri-> ‘proximate’, <chi-> ‘distal’ 43 9 Personal pronouns: <ninga> ‘I’, <rica> ‘you (sg)’, <ricacua> ‘you (pl)’ 45 10 Interrogative pronouns 49 11 Case markers and postpositions 55 11.1 Oblique: <-xa> ~ <-ja> . . . . . . . . . . . . . . . . . . . . . . . 55 11.2 Genitive: <-na>, <-azu>, <-ahe> . . . . . . . . . . . . . . . . . . 58 11.3 Locative:<-ni>............................ 59 11.4 Allative:<-ra>............................ 60 11.5 Purpose:<-xare>~.......................... 62 12 Ownership: <inca> 65 13 Adjectives and adverbs 67 14 Verbs 71 14.1 Causative:<ra->........................... 74 14.2 Spatial notions: <bu-> ‘translocative’, <chi-> ‘inward’, <fiqui-> ‘upward’ ............................... 75 14.3 Aspectual notions: <ra-> ‘aorist’, <bi-> ‘rapid’ . . . . . . . . . . 76 14.4 Verbal person marking . . . . . . . . . . . . . . . . . . . . . . . 78 14.5 Mood................................. 82 14.5.1 Realis mood: <-hi> . . . . . . . . . . . . . . . . . . . . . 82 14.5.2 Imperative1:<-(a)ba>.................... 85 14.5.3 Imperative2:<-ra>~<-za>................. 88 14.5.4 Imperative3:<-ni> ..................... 91 14.5.5 Prohibitive: <ni-…-qua> . . . . . . . . . . . . . . . . . . 92 14.6 Modality ............................... 93 14.6.1 Desiderative: <-zea> . . . . . . . . . . . . . . . . . . . . 93 14.6.2 Non-desiderative: <ficoa> . . . . . . . . . . . . . . . . . 94 14.7 Suppletion: <yu> and <qui> ‘to come’ . . . . . . . . . . . . . . . 96 14.8 Copula:<qui> ............................ 98 ii Contents 15 Further grammatical morphemes and features 101 15.1 Plural:<-ya>............................. 101 15.2 Negation: <para>, <ra-…-nca> . . . . . . . . . . . . . . . . . . . 102 15.3 Interrogative: <ni> . . . . . . . . . . . . . . . . . . . . . . . . . 105 15.4 Hortative: <ynszi> . . . . . . . . . . . . . . . . . . . . . . . . . . 107 15.5 Augmentative, diminutive, and related notions . . . . . . . . . . 108 15.6 Tense................................. 108 16 Concluding summary 111 Acknowledgements 113 References 115 Index 121 Nameindex................................. 121 iii Preface The history of the Andakí is that of one of the most enigmatic indigenous peoples of South America. From the beginning, their origins were wrapped in mystery. They have been described as remnants of the Precolumbian people who built the wonderful stone monuments surrounding the village of San Agustín in the upper Magdalena highlands, but also as Post-San Agustín invaders from the eastern lowland who might have welcomed the first Spanish conquistadors upon their arrival in the region. A more realistic and well-informed account of their origins is provided by the historian Juan Friede in his book Los andakí 1538–1947: Historia de la aculturación de una tribu selvática (México D.F., 1953). He describes the Andakí as an invasive nation who penetrated the Alto Magdalena region from the eastern Andean foothills around 1600. For two consecutive centuries they sowed fear and terror among the local Spanish settlers and their indigenous associates. The latter had been ruthlessly exploited since their incorporation in the colonial encomienda system. This treatment eventually resulted in the near disappearance of the local population, some of whom escaped to the eastern foothills and lowland, where they may have joined the ranks of Andakí fighters. From 1600 onward attacks of the Andakí obtained the character of targeted revenge actions, including nighttime abductions and the forced disappearance of entire families. Punitive expeditions organised by the colonial authorities were helpless against the guerrilla tactics of the Andakí, who had learned to use effective weaponry and refined intelligence against the encomenderos and their subjects. Their attacks continued throughout the 18th century, whereafter they diminished with the withdrawal of the Andakí to jungle bases on the Caquetá and Fragua rivers. Eventually, the memory of their feats slipped into oblivion, and only place names such as Belén de los Andaquíes in the Caquetá department are reminiscent of their former existence. The language of the Andakí has been poorly documented so far and remains unclassified. There have been no recent reports of possible surviving speakers. Consequently, the Andakí language must be considered extinct, unless evidence of the contrary is brought forward. Jelien Moens and Matthias Pache have taken up the challenge of analysing and systematising the scarce language data of Andakí that have been preserved. The present book exhibits the 2 Cultural background Figure 2.1: Map of the approximate distribution area of Andakí, created by Arjan Mossel, based on Eberhard et al. (2025) 4 (A_008).1It is interesting to see that Andakí terms for metals and items made of metal do not seem to be borrowings from Spanish: <bininto> / <binintó> ‘iron’, “hierro” (M_476), <yrajaró> ‘steel’, “acero” (M_477), <benije> (A_009), <binixi> (M_480) ‘machete’, <mifi> ‘needle’, “aguja” (M_482) – the term for ‘needle’, for instance, is often a borrowing from Spanish or Portuguese in Lowland South American languages (e.g., Pache & Urban, forthc.). Andakí houses had a circular shape and were distributed around what would have been the residence of the cacique or the chief of the tribe (Vargas Motta 1962: 15). In their battles, the Andakí made use of spears, darts, clubs, and slings; their spears and clubs were made from the wood of chonta palms (Vargas Motta 1962: 16–17; see Friede 1953: 107–108 for a description of the Andakí inventory of arms). Like their neighbors, the Andakí had the reputation of being “exceedingly wild and warlike” (Albis 1860–1861: 59) and they regularly united with them to fight the non-indigenous invaders (Friede 1953: 91; Vargas Motta 1962: 36), resisting them for almost 250 years (Friede 1953: 11). According to Colombia’s 2018 National Population and Housing Census (CNPV 2018), there were still some 248 persons identifying themselves as Andakí at that time. Although there is no recent evidence that Andakí is still spoken – Glottolog (Hammarström et al. 2025) and Ethnologue (Eberhard et al. 2025) consider the language extinct (Glottolog) or endangered/dormant, that is, “no longer used as a first language by any remaining members of the ethnic community” (Ethnologue) – the possibility that there are still some Andakí groups or individuals speaking Andakí cannot totally be excluded. 1The original Spanish translations from 18thand 19th-century sources are provided wherever appropriate. For verbs, these are often cited in inflected rather than infinitive forms. However, this convention is not followed when verbs appear as inflected forms in the sources but only the root is relevant. We generally preserve the spelling found in the original sources, although in some cases it is adapted to modern Peninsular Spanish; the original spelling is always provided in the appendix. To ensure typographical consistency, individual Andakí forms in the running text are mostly cited with a lowercase initial, even though both Spanish and Andakí entries in the manuscripts consistently begin with an uppercase letter. Original capitalization is retained only in glossed examples. The use of voseo in the Spanish translations (e.g., vení instead of ven ‘come (sg)!’) has been retained. Pointed brackets indicate graphemic representations. 5 3 Sources and previous works on Andakí There are three primary sources on Andakí. First, there are two nearly identical vocabulary lists which are part of a larger collection of indigenous vocabularies sent to Madrid by the scholar José Celestino Mutis (1732–1808) by order of the archbishop-viceroy of Santafé de Bogotá, Don Antonio Caballero y Góngora. The word lists were a request to Carlos III by Catherine II of Russia who wanted to document all the languages of the world. These Andakí materials sent to Madrid in 1787 by Mutis never reached the empress (Ramos 1959: 382; Ortega Ricaurte 1978: 102; Adelaar & Muysken 2004: 54) and were eventually published by the Royal Library of Madrid in 1928 (collection Lenguas de América). Scans of the two copies of the Vocabulario andaqui–español, Ms. II/2911 and Ms. II/2912, are available online at https://rbdigital.realbiblioteca.es/s/realbiblioteca/item/15059 and https://rbdigital.realbiblioteca.es/s/realbiblioteca/item/15084. Ms. II/2911 has 13 numbered pages, while Ms. II/2912 has 18. These two manuscript lists are nearly identical: they contain mostly the same concepts, but regularly differ in spelling. They offer 772 different Andakí words and phrases, accompanied by their Spanish translations. For spelling differences between both lists, compare, for instance, ‘Howis your (sg) wife?’, “Cómo está tu mujer?” (M_288), transcribed as <Karupegua hszazini> in Ms. II/2911 and as <Karupegua ñszazini?> in Ms. II/2912. For further examples of different spelling conventions used in the two 18th-century lists, see (1–2). (1) a. <Ychuyzi kaquani?> ychuyzi what ka-qua 2-say ni? q ‘What do you (sg) say?’, “Qué dices?”1(M_008) 1In the second line, we use a slightly simplified transcription of Andakí, which does not reproduce diacritics. 3 Sources and previous works on Andakí b. <Ychuhizi caquani?> ychuhi-zi what ca-qua 2-say ni? q ‘What do you (sg) say?’, “Qué dices?” (M_751) (2) a. <Ychuyzi kazini?> ychuyzi what ka-zi 2-do ni? q ‘What do you (sg) do?’, “Qué haces?” (M_009) b. <Ychuhizi cazini?> ychuhizi what ca-zi 2-do ni? q ‘What do you (sg) do?’, “Qué haces?” (M_752) More information on spelling conventions in the different Andakí sources is provided in Section 5. In some cases, two consecutive entries in the 18th-century lists seem to relate to each other in terms of question and answer, as in the case of (3–4). (3) <Bizejazini?> / <Bizefazini> bize-jazi what-purp ni? q ‘What for?’, “Para qué?” (M_185) (4) <Choyaxazé> / <Choya xaré> choya-xaze eat-purp ‘In order to eat’, “Para comer” (M_186) Note the variation in the rendering of the same grammatical morpheme <fazi>, <jazi>, <xaré>, <xazé> ‘purposive’ in the two consecutive examples. Given that thereare morecorrections in Ms. II/2911 than in Ms. II/2912, it seems that the latter is of slightly later origin and a copy of Ms. II/2911. In single cases, either of both copies may contain a form that is not present in the other one: seven entries that are found in Ms. II/2911 have no counterpart in Ms. II/2912, whereas only two entries in Ms. II/2912 have no counterpart in Ms. II/2911. Therefore, it is important to consider both sources. Overall, the handwriting is quite different between the two manuscripts, and the documents may have been written by different hands. Some 400 of the 772 entries in these two 18th-century sources are at least short phrases. Gómez & Torres (2012–2013a,b) provide a transcription 8 of the anonymous 18th-century manuscripts, with additional comments on the sources and on some of the Spanish translations. A second, shorter vocabulary list was composed by the Presbyter Manuel Maria Albis in 1854. This word list was part of his notes of a trip to the Provincia de los Andaquíes (Albis 1860–1861: 53). According to the introduction by the editors, Albis’ notes were delivered by a certain Dr. Manuel Antonio Bueno (Albis 1860–1861: 53). These data were first published in Spanish in 1855 in a work with the title Los indios del Andaquí: memorias de un viajero (Vergara y Vergara & Delgado 1855), and translated into English and published in the Bulletin of the American Ethnological Society (1860–1861). In his work, Albis describes the habits and language not only of the Andakí, but also of the Ingano (speakers of a Quechua variety), Koreguaje (Tucanoan language family), Guaque (Cariban language family), and Macaguaje (Tucanoan language family), who also inhabit this area. The Andakí vocabulary section of Albis contains 159 words and four phrases. Since the present work is mainly concerned with the grammar of Andakí, it is largely based on the Mutis materials from the 18th century; data from Albis (1860–1861) will be discussed primarily in the phonology chapter. Works based on these three primary Andakí sources have been relatively few so far. The first one is a presentation of 21 Andakí lexical items from Albis’ lists and a comparison with their counterparts in Koreguaje, a Tucanoan language (Latham 1862: 481). Besides presenting data in a table, however, Latham does not further comment on Andakí, or on coinciding forms in Andakí and Koreguaje. Indeed, there is no evidence for a genealogical connection between Andakí and Tucanoan languages, although there may be some single loanwords such as Andakí <paga> / <pagá> (M_499), <pagà> (A_112) ‘manioc’, “yuca”, which is reminiscent of Proto-Tucanoan *poʔɡa ‘manioc flour’ (Waltz & Wheeler 1972: 144). Rivet (1924), likewise based on Albis’ data, is the first study to discuss Andakí grammar: the author identifies a few grammatical morphemes among which are <-guae> ‘diminutive’ and a prefix <ma->, <man->, <min->, attested, for instance in <min-gosoa> ‘arm’ (cf. the entry in A_100) and <ma-esegua> ‘tail’ (cf. the entry in A_097); also, Rivet (1924: 100–101) states that subject person is marked on the verb by prefixation in this language. A few grammatical morphemes are also mentioned by Friede (1952: 190), although he does not comment on their meaning. More recently, papers on Andakí phonology (Coronas Urzúa 1994) and lexicon (Coronas Urzúa 1995) have been published, based on the two 18th-century lists (Anonymous 1701–1799[a],[b]) and on Albis (1860–1861). Coronas Urzúa’s (1994) observations on Andakí phonology will be addressed in more detail in this book, in Section 5 on Andakí phonology. Apart from lexical roots, Coronas Urzúa (1994, 9 3 Sources and previous works on Andakí 1995) also mentions a small number of grammatical morphemes: for instance, the author argues that /-kʷa/ is a negation marker (Coronas Urzúa 1994: 85); Andakí /haa/ is interpreted as an accusative or dative marker (Coronas Urzúa 1995: 82), the suffix /-ra/ as a locative marker (Coronas Urzúa 1995: 91), /-sa/ as an imperative marker (Coronas Urzúa 1995: 92), and /-si/ as a suffix forming adjectives (Coronas Urzúa 1995: 94). Some further insights into Andakí grammar are provided by Adelaar & Muysken (2004: 140): <ca->, for instance, marks the 2nd-person subject, as shown in (5). (5) <Ninga camimi> ninga I ca-mimi 2-love ‘Do you love me?’ (M_232; adapted from Adelaar & Muysken 2004: 140) Modal categories, by contrast, are encoded by suffixes, as shown in (6); the same is true for nominalization and case (Adelaar & Muysken 2004: 140). (6) <Fsrrajanozá> fsrrajano-za lie.down-imp ‘Lie down!’ (M_366; adapted from Adelaar & Muysken 2004: 140) Finally, Jolkesky (2016: 525–526) also discusses a few aspects of Andakí grammar, for instance the 2nd-person singular ka-, or the 2nd-person plural pronoun rika-kʷa. Besides publications addressing Andakí itself, there have also been several attempts to connect this language to other languages or language families in the Americas. Some (e.g., Rivet 1924; Greenberg 1956, 1987; Swadesh 1962; Loukotka 1968) have considered Andakí to be a Chibchan or Páezan language. Rivet (1924), for instance, compares Andakí to Nasa Yuwe (also known as Páez), but also to Barbacoan languages and Chibchan languages, and considers them all to be genealogically related. The last decades, by contrast, have seen a rise of linguists proposing a different view, namely, of Andakí as a language isolate (Constenla Umaña 1981; Constenla Umaña 1991; Coronas Urzúa 1994; Coronas Urzúa 1995; Kaufman & Berlin 2007; Campbell 2012; Seifart & Hammarström 2018: Campbell 2024). Nevertheless, Adelaar and Muysken suggest that Andakí is “possibly related to Páez” (italics in original, Adelaar & Muysken 2004: 611), and the 2025 10 edition of Ethnologue classifies Andakí as Páezan (Eberhard et al. 2025). Similarities with Nasa Yuwe (Páez) do exist above all in the lexicon (e.g., Pache 2018a: 572–574; Pache 2024). Andakí has also been investigated in terms of language contact and areal typology. For instance, the Andakí vowel system is characteristic of Amazonian languages in that it distinguishes between nasalized and oral vowels (Coronas Urzúa 1994: 80). The weakly developed numeral system of Andakí has likewise been argued to indicate a possible Amazonian origin of this language (Adelaar & Muysken 2004: 140). In terms of loanwords, Jolkesky (2016) notes several lexical parallels with Chibchan languages, Nasa Yuwe, and Tinigua (Tinigua– Pamiguan). 11 4 Approach and challenges For this book, we analyzed the language materials provided in 18th-century lists (Anonymous 1701–1799[a],[b]), sent to Spain by Mutis, and the slightly more recent word lists published in Albis (1860–1861). The first author of this book, Jelien Moens, built a database combining the available Andakí materials. The reader can consult it here: https://sites.google.com/view/Andaki-database. The two anonymous 18th-century lists, Ms. II/2911 and Ms. II/2912, are both provided in the database. Their transcriptions by Gómez & Torres (2012–2013a,b) have also been added, in order to illustrate differences with our own transcriptions. A synthesis column illustrates the sometimes divergent information and transcriptions. Also, modern equivalents of the more archaic spelling of the Spanish translations were added wherever deemed helpful. Finally, every entry is assigned an ID ranging from M_001 to M_772 (M for Mutis). As to the Albis sources (Vergara y Vergara & Delgado 1855; Albis 1860–1861), we consulted Albis’ travel notes, published in Spanish in Los indios del Andaquí: memorias de un viajero in 1855, which is available online,1but is sometimes difficult to read. To solve this problem, and to stay as close to the original source as possible, the English translation of Albis (1860–1861) was consulted whenever needed. As with the Mutis lists, a modern-day Peninsular Spanish column has been added to facilitate the process; the resulting entries range from A_001 to A_163 (A for Albis). The Albis source only contains four short sentences, while the remaining 159 entries are lexical (predominantly nouns). In sum, the excel sheet in the database contains 935 rows of linguistic material to work with. In this book and in the data shown below, all data are presented together with their IDs. The reader can thus consult the corresponding original spellings by looking up the relevant ID in the database. As with other scarcely documented, extinct languages (see, e.g., Urban 2019), working with the Andakí materials turns out to be quite complex in several respects. For instance, there is a lot of orthographic and allophonic variation, discussed in Section 5 on phonology, and in Coronas Urzúa (1994). Also, it is not 1https://bdh-rd.bne.es/viewer.vm?id=0000123147 4 Approach and challenges with certainty whether a particular phenomenon is absent from Andakí itself, or merely absent from the available language data. This book therefore by no means aims to provide an exhaustive description of Andakí grammar: linguistic examples and claims presented do not cover all aspects of Andakí grammar and are representative of a tendency in the limited dataset only. 20 5 Phonology This chapter provides a sketch of Andakí phonology, based largely on a paper dedicated to this topic (Coronas Urzúa 1994). An overview of Andakí phonology must take into account the limited availability of data and hence of minimal pairs, and the fact that we rely on data that are some 180 to 230 years old. The graphemes and grapheme combinations used in these sources do not necessarily represent phonemes, and their use is often quite variable. Variation in the orthography may have its origin, for instance, in spelling inconsistencies; in other instances, it may reflect alternation or free variation of sounds. An example of this variation is given in (1), where the root ‘to eat’ is given as <chiya> and <choya> and the purposive marker as <-jazi>, <-jaze>, <-xaré>, and <-xazé>. (1) a. <Chiya jaze> chiya-jaze eat-purp ‘In order to eat’, “Para comer” (M_147) b. <Choyaxazé> / <Choya xaré> choya-xaze eat-purp ‘In order to eat’, “Para comer” (M_186) In a number of cases, forms provided in Albis (1860–1861) can be quite different from those found in the 18th-century lists, as shown in the case of <condefui> (A_039) in the 19th-century list and the related <szuntijohé> (M_656) in the 18thcentury list, both ‘tapir’, “danta”. Given that these differences in spelling are not always systematic, it is not easy to determine to what extent such differences between the 18th and 19th-century sources reflect dialectal differences, distinct ways to represent Andakí sounds or, in the case of the ‘tapir’ term, a reference to different species. Neither for the 18th-century nor for the 19th-century materials, is there explicit information on how Andakí sounds were pronounced and on how exactly the data were gathered. Finally, there are several homographs in the data, suggesting that Andakí may have had many homophonous forms, unless tone, which is not represented in the spelling, played a role. According to Coronas 5 Phonology Urzúa’s (1995: 87) analysis of Andakí phonology, /kʷa/, for instance, can have the meanings ‘to kill’, ‘to cook’, ‘to do’, and ‘to say’, among others. Another example of Andakí homophony is <minzará> ‘sand’, “arena” (M_695) versus <minzará> a tree, Genipa americana, “jagua” (M_696). 5.1 Phonotactics, syllable and root structure The Andakí basic syllable structure is CV, as in <rica> / <ricá> ‘you’, “tú” (M_013). A syllable structure V or VC is attested only in exceptional cases such as in <acai> ‘lemon’, “limón” (A_001) or in <ypchize> ‘what?’, “qué” (M_303; M_304). If a word starts with a vowel, it is mostly afollowed by <n>, as in <antazá> ‘up (sg)!, raise (sg) it!’, “alza” (M_380); <andagu> ‘fast’, “presto” (M_550) – compare <dacoze> ‘fast’, “presto” (M_549), which seems to contain a related root *dako – or <anduazo> ‘banana’, “plátano” (M_113) – compare <duazo> ‘banana’, “plátano” (M_054). There are different interpretations possible for this phenomenon, and we tentatively propose here that the ais epenthetic before a prenazalized stop in word-initial position (pace Coronas Urzúa 1994). If nasal consonants occur in the syllable coda in other environments, they indicate nasality of the preceding vowel, for instance, in <ringa> ‘I’, “yo” (M_011), /rĩka/ (Coronas Urzúa 1994: 74– 75). CVC syllables do not occur in Andakí. Consonant clusters in intervocalic position do not seem to exist either, except in single cases such as <ypchize> ‘what?’, “qué” (M_303; M_304). Consonant clusters in the onset of a syllable are exceedingly rare if they exist. We found a few potential examples which include a nasal consonant /n/ and a sibilant, as in <nszajini?> / <nszãjini> ‘where are they?’, “dónde están?” (M_176); elsewhere, we find the respective forms without initial <n>: <sxaxini> ‘where is she/he/it?’, “dónde está?” (M_067). Similar clusters involving <n> and grapheme combinations referring to sibilants are attested in <nszazincaha> / <nszâzincahá> ‘because I (m) was sick’, “porque estuve enfermo” (M_326) or <nszazincahe?> / <nszazincaque?> ‘are you (sg, (m) sick?’, “estás enfermo?” (M_321). The possibility cannot be excluded, however, that we are not dealing with consonant clusters in onset position in these cases, but that <n> in these examples refers to a syllabic nasal consonant, as proposed by Coronas Urzúa (1994: 95). In other cases, grapheme combinations – like <sz>, <fsrr>, or <jx>, as in <jxiizá> / <jxizá> ‘walk (sg)!’, “camina” (M_365) – seem to refer to single sounds such as sibilants (in the case of <sz> and <fsrr>) or to a laryngeal fricative (in the case of <jx>), not to consonant clusters (for discussion, see Section 5.2). Sequences of same and distinct vowels both occur. Sequences of same vowels are discussed below, in Section 5.2.6. Among the sequences of distinct vow22 5.2 Vowels and suprasegmentals els, <ea> or <ae> are among the most frequently attested; the sequence <oa> is less common in Andakí, <ao> does not exist at all in the available sources; an overview of the exact frequencies of particular combinations is provided in Coronas Urzúa (1994: 79–80). Some of these sequences, such as <oa>, may also be interpreted as a sequence of a glide plus a syllabic vowel. In single instances, sequences of up to three vowels are attested: compare <janszeaezá> ‘get brave (sg, m)!’, “hazte bravo” (M_412). Most Andakí roots have the structure CVCV, as, for instance, <chisi> ‘to urinate’ (cf. M_389). Monosyllabic roots are infrequent — for instance, <bi> ‘to carry’ (cf. M_378) or <ko> ‘to blow’ (cf. M_381) (cf. Coronas Urzúa 1995). Grammatical morphemes often have the structure CV, such as <-ni> ‘locative’ or <ra-> ‘causative’ (see Section 11 and Section 14); only a few polysyllabic grammatical elements have been identified here, such as non-desiderative <ficoa> (see Section 14.6.2). 5.2 Vowels and suprasegmentals The Andakí vowel inventory has been argued to contain six vowels: /a/, /i/, /u/, and their nasal counterparts /ã/, /ĩ/, /ũ/ (Coronas Urzúa 1994: 85); it is shown in Table 5.1. Table 5.1: Vowel phoneme inventory of Andakí as established by Coronas Urzúa (1994: 80) and translated into English Front Central Back High i, ĩ u, ũ Non-high a, ã There is little evidence for the existence of a high central vowel ɨ,ɨ, which is frequently found in Lowland South American languages (cf. Dixon & Aikhenvald 1999: 8); if there was one in Andakí, it would probably be represented by <y>, which also seems to substitute <i> or <u> in some instances, as in <symquarahá> / <sumquarahá> ‘smoke’,1“humo” (M_531) and <ynsci> / <ynszi> ‘let us go away!’, “vámonos” (M_313), <inszi> ‘hortative; let us go!’, “vamos” (M_129). Mid vowels from loanwords are represented as such in the orthography of the Andakí sources: <e> is found in <buytreni> ‘vulture’ (M_606, from Spanish buitre 1The boldfacing in Andakí examples discussed in this and other sections is not in the original, but has been added by the authors. 23 5 Phonology ‘vulture’), <e> and <o> occur in <ovejani> ‘sheep’ (M_097, from Spanish oveja ‘sheep’) where they may have been pronounced as mid vowels by Andakí speakers. By contrast, Spanish /e/ seems to have been adapted to <i> [i] in <parini> ‘priest’ (M_560), which might be a loan from Spanish (padre ‘priest’). 5.2.1 Suprasegmentals Little is known about Andakí suprasegmentals. There is no direct evidence for tone, although the high frequency of homographs in the Andakí sources suggests that lexical tone may have existed in this language. Acute or grave accent is nearly always found in word-final position and may indicate stress, for instance, in <Ypchize canahá quine?> / <Ypchizé canahá quiné?> ‘What did you (sg) bring?’, “Qué has traído?” (M_303); there are no indications for stress being a contrastive feature in Andakí (Coronas Urzúa 1994: 97). 5.2.2 Vowel nasality The sequence of a vowel and <n> or <m> has been interpreted as representing a nasal vowel (Coronas Urzúa 1994: 74–75); in a few instances, nasality has also been argued to be represented by a circumflex on the vowel or by a postposed <ñ> (Coronas Urzúa 1994: 74–75). Among the few sets suggesting a phonemic distinction between a nasal and an oral vowel in Andakí is <ricaxa> ‘to you (sg)’, “a vos, a ti” (M_017) versus <rincaxa> ‘to me’ (cf. M_105). The orthography used in the different sources is not always consistent, and sometimes, a vowel seems to have been nasal in one case and oral in another case. This is illustrated by <shuntahé> ‘deer’, “venado” (M_657) with a nasal ũversus <shotajihi> ‘deer’ with non-nasal o, as it occurs in <mixinchi shotajihi> / <mixinehi shotajihi> ‘jaguar’, “tigre” (M_659) – compare <mixenehi> / <mixinehi> ‘dog’, “perro” (M_065). 5.2.3 /i/, /ĩ/ In the available Andakí data, <i> seems to appear in free variation with <e> when these letters refer to a syllabic vowel, as in <nanquizi> (M_451) versus <nanguize> / <nañguize> (M_094), both ‘meat’, “carne”, in <cantijuchi> (M_504) versus <cantejoche> / <cantejoché> (M_505), both ‘pineapple’, “piña” (Coronas Urzúa 1994: 73), or in <inca> (M_284) versus <enca> / <encá> (M_173), both ‘ownership’ (see Section 12). There is no minimal pair suggesting a phonemic distinction between iand e; the contrast between <guajije> ‘pot’, “olla” (A_056) versus <guajiji> ‘small pot’, “olleta” (A_057) rather reflects sound symbolism. In some 24 5.2 Vowels and suprasegmentals instances, the Andakí forms in the 18th century lists have <i> where they have <e> (sometimes <a>) in the 19th century list, as in <guahizi> / <guajixi> (M_110) versus <guajije> (A_056), both ‘pot’, “olla”, or in <bisinguay> / <bintiguay> (M_647) versus <mandeguae> (A_084), both ‘bat’, “murciélago”. Both in the 18thcentury and 19th-century lists, <i> is more frequent than <e> (Coronas Urzúa 1994: 74). The factors conditioning the choice of <i> or <e> are not yet understood; in vowel sequences, if preceded by <a>, <e> occurs more frequently than <i>, both in the 18thand 19th-century sources (Coronas Urzúa 1994: 79). Among the few minimal pairs opposing <i> to <a> is <nanquizi> ‘meat’, “carne” (M_451) and <nãnquazi> / <nanquazi> ‘beautiful’, “bonito” (M_724) (Coronas Urzúa 1994: 74), and <quinaji> ‘head’, “cabeza” (A_117) versus <quinaja> ‘hair of the head’, “cabello” (A_116). A near-minimal pair opposing Andakí /i/ to /u/ is <chincahi> ‘brother-in-law’, “cuñado” (M_573) versus <chuncahá> ‘all’, “todo” (M_263). Non-syllabic /i/ is represented by <y> in intervocalic position, as in <chiyazi> ‘food’, “comida” (M_041) and in <guatiye> ‘stone’, “piedra” (A_062), or in final position, as in <bisinguay> / <bintiguay> ‘bat’, “murciélago” (M_647). 5.2.4 /u/, /ũ/ In the available Andakí sources, <u> seems to appear in free variation with <o>, as in <bujeaba> (M_281) versus <boxeaba>, both ‘let us go away!’, “vámonos” (M_085). Possible factors that determine the choice of <u> or <o> in the available language materials remain unknown. In several cases, <u> in the 18th-century list corresponds to <o> in the 19th-century lists, as in <fuchigua> (M_580) versus <ochegua> (A_110), both ‘hook’, “anzuelo”. Indeed, the grapheme <u> occurs more frequently than <o> in the 18th-century sources, whereas <o> is slightly more frequent than <u> in the 19th-century list (Coronas Urzúa 1994: 73). There are some instances where /u/ is represented by <v>, as in <vntahe> / <untahé> ‘moon’, “luna” (M_464) or in <taraunhé> / <taravnhé> ‘chicken’, “gallina” (M_098). We found no minimal pair suggesting an opposition between u and oin this language. 5.2.5 /a/, /ã/ The third vowel to be discussed here is /a/ <a>. Among the (near-)minimal pairs opposing /a/ <a> to /i/ <e> are <sacahá> ‘hands’, “manos” (M_431) versus <szecahá> ‘foot, leg’, “pie, pierna” (M_428), <antazá> ‘up (sg)!, raise (sg) it!’, “alza” (M_380) versus <antezá> ‘shout (sg)!’, “grita” (M_375), and, possibly, <mansesai> ‘large turtledove’, “tórtola grande” (A_090) versus <mensesai> 25 5 Phonology ‘small turtledove’, “tórtola pequeña” (A_094).2A minimal pair opposing /a/ <a> to /u/ <o> is <kagá> ‘sweet potato’, “batata” (M_500) versus <cogo> ‘farm’, “casa, rancho” (M_512). Additionally, it has been observed that Andakí /a/ is realized as a back vowel in some cases, for instance, in <anduozo> ‘banana’, “plátano” (M_144) versus <anduazo> ‘bananas’, “plátanos” (M_227) (Coronas Urzúa 1994: 74). In other cases, /a/ may have been realized as a high front vowel, for instance, in the case of <naqui> (M_002) versus <naqua> (M_003), both ‘who?’, “quién”. The determining factors remain unknown. 5.2.6 Sequences of identical vowels Sequences of identical vowels are widely found in the available Andakí data (Coronas Urzúa 1994: 79). That we are dealing with vowel sequences, not with long vowels, is supported by the observation that in several of these cases, one of the vowels carries an acute accent and the other does not, as in <Kariszina aqueé?> ‘Do you use to drink?’, “Sabéis beber?” (M_152). Sequences of identical vowels occur much less frequently in the Albis (1860–1861) list than in the two Mutis lists (Anonymous 1701–1799[a],[b]), and they are split by <h> in most cases in the latter source, although there are exceptions such as <guuhe> / <guuhé> ‘other’, “otro” (M_016) (Coronas Urzúa 1994: 77–78). It is not yet clear what intervocalic <h> indicates if it occurs between two identical vowels, and whether it may refer, for instance, to voicelessness or aspiration.3 Sequences of identical vowels that are split by <h> must be carefully distinguished from cases where a suffix -hV is attached to a word or morpheme that ends in the same vowel as the suffix vowel. In Andakí <quihi> ‘I came’, “vine” (M_733), for instance, final <-hi> must not be confused with the realis mood marker <-hi> (see Section 14.5.1), because <hi> is also present in Andakí <quihiza> ‘come (sg)!’, “vení” (M_127), which contains an imperative mood suffix <-za>. Although in this particular example the matter is clear, this is not always the case. Most frequently, same-vowel sequences in one source do not correspond to same-vowel sequences in the other source – as in <sonjuhu> ‘netbag’, “mochila de red” (M_628) versus <suju> ‘netbag’, “mochila” (A_154). Only in a few instances does a vowel sequence of the 18th-century source correspond to a vowel sequence in the 19th-century source, as in <sacahá> (M_431) ‘hands’, “manos” 2This pair, however, like the pair <guajije> ‘pot’, “olla” (A_056) versus <guajiji> ‘small pot’, “olleta” (A_057), discussed above, may rather reflect a widespread pattern of sound symbolism, with the high front vowel iassociated with smallness (e.g., Jespersen 1933). 3An anonymous reviewer suggests that <h> may have referred to a glottal stop. 26 5.3 Consonants versus <sacaà> ‘hand’, “mano” (A_121) (Coronas Urzúa 1994: 78), or in <jihi> ‘spirit’, “demonio” (M_461) versus <gii> ‘spirit’, “diablo” (A_050). A minimal pair that seems to oppose <ihi> to <i> is <quihi> ‘to come’ (cf. M_410) versus <qui> copula ‘to be’ (cf. M_762). In most but not all cases, sequences of identical vowels occur in word-final position (Coronas Urzúa 1994: 78). There are hardly any minimal pairs opposing a sequence of identical vowels to a singleton vowel – an exception is the (near-)minimal pair <sacaà> ‘hand’, “mano” (A_121) versus <sacca> ‘reed’, “caña” (A_124). 5.3 Consonants The Andakí consonant inventory proposed by Coronas Urzúa (1994: 96) and adapted in the present work is shown in Table 5.2. Table 5.2: Consonant phoneme inventory of Andakí as adapted from Coronas Urzúa (1994: 96) Bilabial Alveolar Prepalatal Velar Labiovelar Laryngeal Voiceless stops p t [t, d, ⁿd] k [k, ɡ, ᵑk] kʷ [kʷ, ɡʷ, w] Voiced stop b [b, ɸ, β] (ɟ)4 Affricate ʧ Voiceless fricatives s [s, ʃ] x [x, h]5h [h, x, ɸ] Nasals m [m, b] n Trill r [ɾ, n, h] The historical phonology of Andakí still leaves many open questions, yet is beyond the scope of this book. In several instances, forms seem to be etymologically related, yet the sound changes leading to the distinct forms are difficult to trace; one of these cases is <nanszihisze> ‘two’, “dos” (M_625), which seems to contain an element <szihi>, related to <szifi> ‘eyes’, “ojos” (M_423) (see Section 7). Andakí <szifi> ‘eye’, in turn, may be related to <chipi> as attested in <chipina> ‘face’, “rostro” (M_425). As illustrated below, there is considerable variation in the graphemes used in the manuscript; variation seems to be particularly frequent before <i>. 4Proposed as represented by <y> in Coronas Urzúa (1994: 88); there are no minimal pairs suggesting that <y> refers to a phoneme that is distinct from /i/. 5This phoneme is tentatively proposed in the present contribution; it is not proposed by Coronas Urzúa (1994). 27 5 Phonology 5.3.1 /p/ Three bilabial consonant phonemes have been postulated for Andakí by Coronas Urzúa (1994): /p/, /b/, and /m/. The voiceless bilabial stop /p/ is represented by <p> in the orthography of the available sources and frequently occurs before /a/ or /ã/, less frequently before front vowels and hardly before back vowels. It does not occur in complementary distribution with other bilabial consonants but in the same environments (Coronas Urzúa 1994: 81). (Near-)minimal pairs suggesting a phonemic distinction between Andakí /p/ and /b/ or /m/ are <piszihi> ‘rooster’, “gallo” (M_595) versus <biszica> ‘man’, “hombre” (M_732) and <muizihi> / <miszihi> ‘man’, “hombre” (M_459).6Andakí /p/ <p> is opposed to /b/ <b> in augmentative <-pi> ~ <-pihi> versus diminutive <-bi> ~ <-bihi> (see Section 15.5; sound symbolism may play a role in this context). A minimal pair opposing /p/ <p> to /k/ <k> is <paga> / <pagá> ‘manioc’, “yuca” (M_499) versus <kagá> ‘sweet potato’, “batata” (M_500). 5.3.2 /b/ The phoneme /b/ has been argued to be represented by <b> and <v> (Coronas Urzúa 1994: 87–88). The grapheme <v>, probably representing a voiced bilabial fricative allophone of /b/ – mostly in 19th-century Andakí – occurs only before <e> or <i> – compare <vegaè> (A_163) and <bicahi> (M_630), both ‘bream’, “dorada”. Additionally, we suggest that there may have been a voiceless bilabial fricative variant of /b/ in 18thand 19th-century Andakí alike, represented by <f> – this is suggested by the occurrence of forms such as <biza> / <bizá> ‘what?’, “qué” (M_063), and its variant <fiza> (cf. M_035). We tentatively propose that <v> and <f> refer to bilabial fricatives rather than to labiodental fricatives. Another fricative variant of /b/ may have been represented by <h> or zero, as suggested, for instance, by the variants <-(a)ba> and <-(a)ha> of the imperative1marker (see Section 14.5.2) and by <biza> / <bizá> (M_063), alternating with <yza> ‘what?’ (cf. M_146). In some single instances, we found evidence for variation between <b> and <m>, as this is possibly the case in <biszica> (cf. M_732) and <muizihi> / <miszihi> (M_459), both ‘man’, “hombre”. In two further cases, there are some b:mcorrespondences between the data provided by Mutis (Anonymous 1701– 1799[a],[b]) and Albis (1860–1861): compare <majihi> ‘macao’, “guacamayo” (M_599) versus <bafé> ‘macao’, “guacamaya” (A_004) and <bisinguay> / <bintiguay> (M_647) versus <mandeguae> (A_084), both ‘bat’, “murciélago”. The ‘bat’ term 6It is impossible to know whether the underlying phoneme is /b/ or /m/ in the Andakí term for ‘man’ (cf. also Coronas Urzúa 1994: 94). 28 5.3 Consonants suggests that there may have been nasalization of bif it is followed by a nasal vowel, similar to what can incidentally be observed for r(Coronas Urzúa 1994: 87–88). 5.3.3 /m/ An Andakí phoneme /m/ has been argued to be represented by <m> (Coronas Urzúa 1994: 94–95). The documentation of the two forms <muizihi> / <miszihi> ‘man’ (M_459), suggests that wis inserted between mand iin some cases.7A possible variation between band min Andakí has been briefly discussed in the previous paragraph and by Coronas Urzúa (1994: 87–88). In other cases, /b/ <f> and /m/ <m> are distinct phonemes: a minimal pair identified in the available Andakí data, opposing /b/ to /m/ is <mifi> /mibi/ ‘needle’, “aguja” (M_482) versus <fifi> / <fifi> /bibi/ ‘to call’ (cf. M_357). 5.3.4 /t/ Only one oral alveolar stop is postulated for Andakí by Coronas Urzúa (1994: 82): /t/. The non-existence of /d/ in Coronas Urzúa’s (1994) consonant inventory of Andakí is a remarkable gap; yet, there are no minimal pairs suggesting an opposition of Andakí tand d. In most cases where we find <d>, it occurs in the 19thcenutry word lists after a nasal vowel and corresponds to a <t> in the 18th-century word lists: compare <bisinguay> / <bintiguay> (M_647) versus <mandeguae> (A_084), both ‘bat’, “murciélago”, <szuntijohé> (M_656) versus <condefui> (A_039), both ‘tapir’, “danta” or <shuntahé> (M_657) versus <sondai> both ‘deer’, “venado” (A_152). That is, /t/ was probably realized as [d] in 19th-century Andakí after nasal vowels. In a few cases, a similar variation is found within the 18thcentury materials, as in <antagoni> ‘quickly’, “aprisa” (M_555) versus <andagu> ‘quickly’, “presto” (M_550). That is, after a nasal vowel, and in a few cases, Andakí /t/ could facultatively be realized as [d] in the Andakí variety documented in the 18th century as well (Coronas Urzúa 1994: 82). Elsewhere, <d> appears only in word-initial position before <i>; in the case of <dingá> / <dinga> ‘I’ (M_012), it represents an allophone of /r/ (see Section 9); in 7A similar phenomenon is attested in Muisca, an extinct Chibchan language of Central Colombia, where a non-syllabic element wis inserted if a labial consonant is followed by ɨ, as in /mɨska/ [mʷɨska] ‘human being’ (Adelaar & Muysken 2004: 84). Whether the terms for ‘human being’ in Andakí and Muisca are related or not remains to be established. Borrowing of a term for ‘man’ or ‘human being’ is also attested in other cases in South America, for instance, in Quechua runa (Rosat Pontacti 2004: 949) versus Cholón nun ‘man, male person’ (AlexanderBakkerus 2005: 363, 525). 29 6 Nouns Most single-word entries in the Andakí language materials are nouns, that is, forms that belong to an open class of items. They may refer to concrete entities, function as the argument of a predicate and as the head of a noun phrase, and take specific derivational morphemes, case markers, or specific plural markers (see, e.g., Dixon 2010: 39, 50). Nouns are particularly prominent in Albis’ (1860–1861) list, which almost exclusively contains nouns. In the available Andakí language materials, nouns include basic vocabulary items, such as <ynquifi> ‘nose’, “nariz” (M_447), <sacaà> ‘hand’, “mano” (A_121), or <sacanaji> ‘knee’, “rodilla” (A_122). Others belong to cultural vocabulary, for instance, <batonafi> ‘plate’, “plato” (A_008) or <bujucahá> (M_479), <bojoca> (A_010), both ‘ax’, “hacha”. Many Andakí nouns contain a final element <hi>, <he>, or <e> (cf. Adelaar & Muysken 2004: 140), which we tentatively interpret as a stem formative. It occurs, for instance, in <mixinehi> ‘dog’, “perro” (M_099), <shuntahé> ‘deer’, “venado” (M_657), <batihi> ‘gourd’, “mate” (M_111), <muizihi> / <miszihi> ‘man, human being’, “hombre” (M_459), <kaquihi> ‘sun’, “sol” (M_463), <chiguae> ‘son’, “hijo” (A_032), but not in forms with female referents (see Table 6.1 below). This final element has the same shape as the realis mood morpheme <-hi> and its variants (see Section 14.5.1), but is probably not identical with it, as shown in (1), where we find <-hi> ‘realis mood’ co-occurring with the presumed stem formative <-he> in the same sentence (albeit in distinct syntactic units).1Given the uncertain status of this element, we prefer not to gloss it separately. (1) <Quahini mijinahé?> qua-hi good-rea ni q mijinahe? dog ‘Is your (sg) dog good?’, “Es bueno tu perro?” (M_271) In other nouns, there seems to be an element <-na>, which may have referred to longish shape originally.2It occurs in forms such as <sana> ‘finger’, “dedo” 1An anonymous reviewer suspects that the nominal element <-he> and its allomorphs have structural significance and observes that in Tinigua a similar element -he functions as a verbal suffix, though its precise role remains unclear. 2An anonymous reviewer points out the similarity between the Andakí shape-related element -na and many Amazonian languages which also use -na as a classifier for elongated referents. 6 Nouns (A_127) (compare the slightly different form <safsrí> / <safsri> ‘finger’, “dedo” in the 18th-century materials, M_448), <chipina> ‘face’, “rostro” (M_425, possibly related to Andakí <szifi> ‘eyes’, “ojos”, M_423, and Nasa Yuwe dʲiʔp ‘face’, Gerdel 2023), <szixaná> ‘chest’, “pecho” (M_437). Note that the actual shape of the entity referred to by the root and the shape originally referred to by the classifying morpheme do not necessarily always coincide in a language (see, e.g., Pache 2016). A stem-formative originally referring to roundish entities may be <-fi>, as attested in <szifi> ‘eye’, “ojos” (M_423), or <chimbufi> ‘navel’, “ombligo” (M_446), and possibly <szitafi> ‘hill’, “cerro” (M_467) – a root related to <szita> occurs in <chitachini> ‘above’, “arriba” (M_569). Finally, <-xe> occurs in several nouns referring to liquids, for instance, <bacoxe> ‘chicha’ (M_032), <vnanszaxi>, <unanszaxi> ‘spirits’, “aguardiente” (M_563), or <jixe> ‘water’, “agua” (M_031). None of these stem formatives/classifying morphemes is glossed in this book, given that the meaning of the root is often unknown (for instance, the meaning of <baco> in <bacoxe> ‘chicha’). Gender marking does not occur in Andakí pronouns, and it is marginal in Andakí nouns. Certain female kinship terms or person designations take a derivational suffix <-gua> (or a variant of it), whereas their male counterparts take <-he> or <-che>, as shown in Table 6.1. Table 6.1: Derivational gender marking in Andakí nouns with male and female referents. Female Male <cachinigua> ‘granddaughter’, “nieta” (M_491) <cachinehe> / <cachineche> ‘grandson’, “nieto” (M_490) <szasejagua> ‘cousin (female)’, “prima” (M_493) <szasejahe> / <szasejahé> ‘cousin (male)’, “primo” (M_492) <chiguacoa> ‘girl’, “muchacha” (M_702) <chiguahe> ‘boy’, “muchacho” (M_701) A suffix <-ni>, <-ne>, or <-re> with an unknown meaning occurs in Spanish loanwords in Andakí. Examples are <buytre-ni> ‘vulture’ (M_606, from Spanish buitre ‘vulture’), <oveja-ni> ‘sheep’ (M_097, from Spanish oveja ‘sheep’), and, possibly, <pari-ni> ‘priest’ (M_560, possibly from Spanish padre ‘priest’). A variant of <-ni> might be <-re>, attested in <guacaré> ‘cow’ (M_096, from Spanish vaca ‘cow’). Andakí nouns of non-Spanish origin do not seem to carry this morpheme. If Andakí <-ni> (and related forms) is indeed a classifying morpheme, it 38 is remarkable in that its use does not seem to be linked to semantics but to the loanword status of the nominal root to which it is attached. It does not occur with borrowed verbal roots, as illustrated by (2). The root <reza> is borrowed from Spanish (rezar ‘to pray’). (2) <Inszi rezara> / <Inszirezara> inszi let.us.go reza-ra pray-all ‘Let us pray!’, “Vamos a rezar” (M_129) There are also several other, apparently derivational morphemes in Andakí nouns which have an unknown meaning – compare, for instance, <sacaà> ‘hand’, “mano” (A_121) versus <sacanifi> ‘forked stick’, “horqueta (palo bifurcado)” (A_123); the meaning of <nifi> remains unknown for the moment. Another case is <-jo>, illustrated in Table 6.2. Note the different position of <-jo> with respect to the presumed stem formative/classifying morphemes in Sets 1 and 2 (<-jo> plus <-hé>), Set 3 (<-fi> plus <-jo>), and Set 4 (<-jo> replaces <-he> / <-hé>). Table 6.2: The derivational suffix <-jo> Form 1 Form 2 1 <shuntahé> ‘deer’, “venado” (M_657) <szuntijohé> ‘tapir’, “danta” (M_656) 2 <szahé> ‘worm’, “gusano” (M_645) <szajohé> ‘dragonfly’, “chapul” (M_644) 3 <szifi> ‘eyes’, “ojos” (A_138) <sifijo> ‘eyebrows’, “cejas” (A_139) 4 <chunguahe> / <chunguahé> ‘ears’, “orejas” (M_424) <sunguajo> ‘ear’, “oreja” (A_155) Finally, in some cases, there are specific word families in Andakí, for instance, nouns derived from a root <chigua> by a range of unknown suffixes: <chiguae> ‘son, offspring’, “hijo” (A_032), <chiguaco> ‘child’, “niño” (A_028), <chiguana> ‘chest’ (A_029), <chiguagua> ‘cousin’, “primo” (A_030), <chiguagus> ‘grandchild’, “nieto” (A_031). In a few cases, the lefthand morpheme of a compound or derived form remains unknown, such as in <jorapahi> ‘caiman’, “caimán” (M_634), which is obviously related to <rapae> ‘caiman’, “caimán” (A_118),3or in the case of <szajihi> ‘lake’, 3Compare, in Chicham languages, the formally similar terms Aguaruna dapi ‘snake’, nampiʧ ‘worm’ (Wipio Deicat & Ritchie 2023), Huambisa napi ‘snake’, nampiʧ ‘worm’ (Jakway 2008: 66–67). 39 6 Nouns “laguna” (M_710), composed of <jixe> ‘water’, “agua” (M_031), and an unknown element <sza->. As first observed by Rivet (1924: 101), some Andakí nouns may carry a prefix <m(V)-> of unknown meaning, illustrated in Table 6.3. Table 6.3: Andakí nouns with a prefix <m(V)-> Andakí noun with m(V)-prefix Andakí noun without m(V)-prefix <manduguaso> ‘banana’, “plátano” (A_089) <duazo> (M_054) and <anduazo> ‘banana’, “plátano” (M_113) <mijina> ‘earth’, “tierra” (A_097) <jixena> ‘earth’, “tierra” (M_338) <micaffi> ‘roasted corn’, “maíz tostado” (M_539) <cahixi> ‘corn porridge’, “mazamorra”, (M_541), <finticahe> ‘empty corn cob’, “tusa de maíz” (M_ 540) Like most other derivational morphemes discussed above, the prefix m(V)- is not frequently attested in the available Andakí materials. It may have a variant (V)mor (V)n-: compare <mitae> (A_101) versus <vntahe> / <untahé> (M_464) ‘moon’, “luna”, <quifi> (A_115) versus <ynquifi> (M_447) ‘nose’, “nariz”, and <bihina> (M_693) versus <imbina> (A_066), ‘silver’, “plata” (cf. Coronas Urzúa 1994: 77). 40 7 Numerals Only a few numerals have been documented for Andakí. They are presented and discussed in the present chapter. The numerals are shown in Table 7.1. Table 7.1: Andakí numerals Andakí numeral English, Spanish <guhigo> ‘one’, “uno” (M_624) <nanszihisze> ‘two’, “dos” (M_625) <nanszihisze haniguhe> / <nanszihisze haniguhé> ‘three’, “tres” (M_626) <saquan cuacho aqua> / <saquan cuachoaqua> ‘five’, “cinco” (M_627) These numerals are all morphologically complex: <guhigo> ‘one’ contains a morpheme that is probably related to <guuhe> / <guuhé> ‘other’, “otro” (M_016) and an unknown ending <-go>. A similar case of colexification of ‘one’ and ‘other’ occurs in Totoro (Barbacoan) and Ika (Chibchan) (Rzymski et al. 2019), two languages that are likewise spoken in northwestern South America. The element <szihi> in the term for ‘two’, <nanszihisze>, may be compared to <szifi> ‘eye’, “ojos” (M_423). A similar case of ‘two’ derived from ‘eye’ is found in some Nadahup languages (cf. Epps 2006), likewise from northwestern South America. The numeral <nanszihisze haniguhe> ‘three’ is built on <nanszihisze> ‘two’ and an element <haniguhe>, which is probably analyzable as <hani-guhe> and may contain the root for ‘other’ or ‘one’, <guuhe>. Note the absence of a numeral ‘four’ in the available Andakí materials. The numeral ‘four’ can have a particular position in South American counting systems (for details, see, e.g., Epps 2006; Pache 2018b). Finally, <saquan cuacho aqua> or <saquan cuachoaqua> ‘five’ may contain a morpheme that is related to or identical with <sacahá> ‘hands’, “manos” (M_431), <sacaà> ‘hand’, “mano” (A_121). 8 Demonstratives: <ri-> ‘proximate’, <chi-> ‘distal’ Andakí demonstratives encode at least a two-way contrast. The demonstrative adverb <rini> (M_553) or <rimi> (M_261) is translated as ‘here’, “aquí”. The pronominal equivalent is <rihizi> ‘this’, “este” (M_014). Its use is illustrated in (1). (1) <Rajiza rihizi> ra-ji-za caus-go-imp rihizi this ‘Take (sg) this!’, “Lleva a este” (M_194) The distant demonstrative adverbs are <chinta> / <chintá> ‘there’, “allá” (M_177; M_552) and <chini> ‘there’, “allí” (M_554). Their pronominal equivalent is <chihi> or <chichi> / <chihi> ‘that’, “aquel” (M_015). This suggest that <-ni> in <rini> ‘here’ and <chini> ‘there’ has a locative meaning (see Section 11.3).1 There are no examples that would illustrate the use of Andakí demonstratives as pronominal modifiers. 1The demonstrative morphemes <ri-> ‘proximate’ and <chi-> ‘distal’ are reminiscent of Nasa Yuwe na ‘dem.prox’ and txã,txa ‘dem.dist’ (see Díaz Montenegro 2019: 276, 295, 764). For a further possible instance of Andakí rand icorresponding with Nasa Yuwe nand a, see see Footnote 6 (page 60). 9 Personal pronouns: <ninga> ‘I’, <rica> ‘you (sg)’, <ricacua> ‘you (pl)’ The data in the word lists contain Andakí 1stand 2nd-person pronouns, but no 3rd-person pronouns. The 1st-person singular pronoun has several variants, among which are <ninga> (M_010), <ningua> (M_239), <ringa> (M_011), or <dingá> / <dinga> (M_012). There are no examples of 1st-person plural pronouns. Accordingly, we do not know whether Andakí made a distinction between the 1st-person plural inclusive and exclusive. The 2nd-person singular personal pronoun is <rica> / <ricá> (M_013) and has a variant <recá> (M_050). The 2nd-person plural pronoun is derived from the singular form by the plural marker <-coa> ~ <-qua>, as shown in (1). (1) <Ninga yubi ricacoaxa fsrrane janora joazajuexa riguanto ajaze> / <Ninga yubi rica coaxa fsr-rane janora joazajuexa ringuanto ajaze>1/ ninga I yu-bi come-rea rica-coa-xa you-pronpl-obl fsrranejano-ra teach-all joazajue-xa God-obl riguanto believe ajaze purp ‘I come to teach you (pl) the law of God so that you may believe in him’, “Vengo a enseñaros la ley de Dios para que creáis en él” (M_135) The plural morpheme <-coa> ~ <-qua> occurs only with pronouns, not with nouns. The element <ri> as attested in <ringa> ‘I’, “yo” (M_011) and <rica> / <ricá> ‘you’, “tú” (M_013) is probably identical with the element <ri> that occurs in <rini> ‘here’, “aquí” (M_553) or <rihizi> ‘this’, “este” (M_014) (see Section 8). In several instances, the use of personal pronouns appears to be optional in Andakí. In (2) the 1st-person subject is not overtly expressed at all. (2) <Quihi> quihi come ‘I came’, “vine” (M_733) 1The hyphen is due to a line break in the manuscript. 10 Interrogative pronouns occurs in <rihizi> ‘this’ (see Section 8) and to the productive relativizing suffix <-zi> discussed in Section 13, which deals with Andakí expressions translated as adjectives and adverbs. Lastly, Andakí <qua> (M_003) is translated as ‘who?’; its variant <qui> is illustrated in (12). (12) <Naquini> na-qui cop-who ni q ‘Who?’, “Quién” (M_002) Interrogative constructions comprising Andakí <qua> or <qui> ‘who?’ are quite different from the interrogative constructions discussed above: Andakí <naqua> or <naqui>, as illustrated in (12), is morphologically complex; we tentatively interpret <na-> as a copular prefix. This construction occurs in cleft sentences like those shown in (13–14), where <na-qua> ‘cop-who’ is not followed by the interrogative marker <ni>. (13) <Naqua fifiquani?> na-qua cop-who fifi call qua who ni? q ‘Who calls you (sg)?’, “Quién te llama?” (M_003) (14) <Naqua naquiquani> / <Náqua náquiguani?> na-qua cop-who na-qui caus-come qua who ni? q ‘Who brought you (sg)?’, “Quién te trajó?” (M_202) In a number of cases, the interrogative suffix <ni> follows <na-qua> ‘cop-who’ and occurs in both clauses of the cleft sentence. This is illustrated in (15–17). (15) <Naqua niuquani?> na-qua cop-who ni q u cook qua who ni? q ‘Who cooks?’, “Quién cocina?” (M_004) (16) <Naqua nimaquani?> na-qua cop-who ni q ma die qua who ni? q ‘Who dies?’, “Quién muere?” (M_005) 52 (17) <Naqua niquaquani?> na-qua cop-who ni q qua kill qua who ni? q ‘Who kills?’, “Quién mata?” (M_006) It is difficult to know to what extent the use of <ni> expresses information structure and whether the focus of the question is different in (13–14) and (15–17), directed at the action in (13–14) and at the person in (15–17). Andakí <na-qua> and <na-qui> ‘cop-who’ is found only in subject position and never in object position in the available Andakí data. 53 11 Case markers and postpositions The present chapter is dedicated to Andakí case markers and postpositions. Although we transcribe all these grammatical morphemes with a hyphen, there is not always sufficient information that would allow us to clearly determine their status as suffixes, enclitics, or free forms. In Section 11.1, we suggest an interpretation of <-xa> ~ <-ja> as an oblique case marker. Section 11.2 will argue in favor of <-na>, <-uza>, and <-ahé> as genitive markers. In Section 11.3, we discuss locative <-ni>, in Section 11.4, allative <-ra>, and in Section 11.5, purposive <-xaré> ~ <-jazi>. The character of these elements as bound or unbound morphemes is not always easy to determine. 11.1 Oblique: <-xa> ~ <-ja> The oblique marker <-xa> ~ <-ja> can be attached to personal pronouns and nouns alike. Its use with personal pronouns is illustrated in (1–3). Its use might add emphasis, as suggested by the Spanish translations in (2) and (3). (1) <Ricaxa> rica-xa you-obl ‘To you (sg)’, “A vos, o a ti” (M_017) (2) <Ningaxa riguantoza> ninga-xa I-obl riguanto-za believe-imp ‘Believe (pl) me!/It is me you (pl) should believe’, “Creedme a mí” (M_348) (3) <Ricaxa fifihe> / <Ricaxa fifihé> rica-xa you-obl fifi-he call-rea ‘It is you (sg) she/he/it calls’, “A vos te llama” (M_193) There seems to be differential object marking depending on the animacy of the referent: <-xa> attaches only to animate patients. In (4–5), the animate patients <verayue> ‘rabbit’ and <nunszue> ‘fish’ are marked as direct objects. 11 Case markers and postpositions (4) <Nunquiza verayueja> / <Nunquizá veraýueja> nunqui-za look-imp verayue-ja rabbit-obl ‘Look (sg) at the rabbit!’, “Mira el conejo” (M_191) (5) <Nunquiza nunszueja> nunqui-za look-imp nunszue-ja fish-obl ‘Look (sg) at the fish!’, “Mira el pez” (M_192) Interrogative pronouns referring to inanimate entities (6) and nouns referring to inanimate entities such as food (7), drinks (8), and clothes (9), do not appear to receive the morpheme <-xa> ~ <-ja> in object position. (6) <Ychuize camine?> / <Ypchize caminé?> ypchize what ca-mi 2-want ne? q ‘What do you (sg) want?’, “Qué quieres?” (M_304) (7) <Chiya nenguihi=> / <Chiya nênguihi> chiya eat nenguihi fish ‘I want to eat fish’, “Quiero comer pescado” (M_082) (8) <Andagu biguaza bacoxe> / <Andagu biquaza bacoxé> andagu quickly bi-gua-za rap-make-imp bacoxe chicha ‘Make (sg) chicha quickly!’, “Haz presto chicha” (M_150) (9) <Quanca jehá sasza> / <Quanajeha sasza> quan-caje-ha 2.imp-wash-imp sasza clothes ‘Wash (sg) the clothes!’, “Lava la ropa” (M_210) Compare the direct object ‘fish’ in (5) and (7) above: in (5), the term for fish refers to a living entity and receives object marking, whereas in (7), it refers to fish in terms of food and the noun in object position does not carry any object marking. In some cases, however, <-xa> seems to be missing with animate direct objects, too; that is, the use of this direct object marker does not seem to be obligatory, at least with pronouns. This is illustrated in (10). 56 11.1 Oblique: <-xa> ~ <-ja> (10) <Ninga ca-mimi?> ninga I ca-mimi? 2-like ‘Do you love me?’, “Me queréis?” (M_232) It seems that <-xa> ~ <-ja> also has other functions than marking the direct object: with the verb <caya> ‘to sit; to live’ and its variants, for instance, it appears to express a comitative meaning (‘with’), as in (11–12). (11) <Cafianzuhi ningaxa ca-yasesza?>1/ <Cafianzuhi ningaxá cayasesza?> ca-fianzu-hi 2-like-rea ninga-xa I-obl caya-sesza? live-des ‘Do you (sg) want to live with me?’, “Quieres vivir conmigo?” (M_353) (12) <Nincaxa coaya> ninca-xa I-obl coaya live.imp ‘Live (sg) with me!’, “Vive conmigo”2(M_223) The morpheme <-xa> is translated as ‘about’, “de” in (13), and may also have been used to convey meanings such as conversation topic. (13) <Nszazi raguayani xin-gaxá?>3/ <Nszâzi raguayani xingaxá?> nszazi how/where ra-gua-ya aor-say-pl ni q xinga-xa? I-obl ‘What do they say about me?’, “Qué dicen de mí?” (M_758) Usually, terms that refer to feelings seem to take an unmarked subject; note that there are only examples of such constructions with a 1st-person subject. Two cases are illustrated in (14–15). (14) <Ninga firajichi> ninga I firajichi hungry ‘I am hungry’, “Yo tengo hambre” (M_029) 1Hyphenation is due to a line break in the manuscript. 2(12) is tentatively interpreted as an imperative construction, given that we also find the single form of <coaya> / <coayá> ‘sit down (sg)!’, “siéntate” (M_088). Yet, neither the Spanish translation nor the Andakí forms allow it to clearly distinguish this construction from a construction involving a 3rd-person subject and indicative mood. 3The hyphen is due to a line break in the manuscript. 57 11 Case markers and postpositions (15) <Ninga finajuche> ninga I finajuche thirsty ‘I am thirsty’, “Yo tengo sed”4(M_030) There is also one instance in the data, however, where the experiencer, <ninga> ‘I’ once more carries the oblique marker <-xa> ~ <-ja> in such a construction (16). (16) <Najanszichi ningaxa> najanszichi angry ninga-xa I-obl ‘I (m) am furious’, “Yo estoy rabioso” (M_333) The use of oblique <-xa> in (16) remains unexplained; as a difference from (14) and (15), the pronoun referring to the experiencer follows the verb in (16), whereas it precedes it in (14) and (15); the use of <-xa> in (16) might be related to this different word order. 11.2 Genitive: <-na>, <-azu>, <-ahe> Used with pronouns, the Andakí morpheme <-na> functions as a genitive or possessive marker. This is shown in (17). (17) <Nszajini jixena ricaná?> nszaji how/where ni q jixena land rica-na? you-pronposs ‘What is your (sg) land?’, “Cuál es tu tierra?” (M_339) In (18), the 1st-person pronoun is followed both by the genitive marker <-na> and the allative morpheme <-ra>. (18) <Ynszi jixena ninganara> ynszi let.us.go jixena land ninga-na-ra I-pronposs-all ‘Let us go to my land!’, “Vamos a mi tierra” (M_338) 4The terms for ‘thirsty’ and ‘hungry’ are almost certainly polymorphemic. For instance, the element <juche> in <finajuche> ‘thirsty’ may have a counterpart <xuchi> in <rúxuchi-hi> / <ruxuhi-hi> ‘I am cold’, “tengo frío” (M_140; hyphenation added by the authors). The ending <-che> or <-chi> recurs in forms referring to emotions and to hunger, thirst, cold and similar feelings and is not further analyzed here. 58 11.3 Locative: <-ni> The genitive marker <-na> seems to occur in possessive pronouns only. To what extent it is related to the locative nominalizer <-na> (see the introduction of Section 14) remains to be established. We found two examples in the data with a regular noun referring to the possessor. Here as well, the form referring to the possessor follows the noun referring to the possessed entity — yet, the possessor is marked by <-azu> in (19). (19) <Guasu nosehazu> guasu egg noseh-azu louse-gen ‘Nits’, “Liendres” (M_689) The term for ‘chicken egg’, “huevo de gallina” is given elsewhere as <huasho> (M_674), obviously the same term as <guasu> in (19), the term for ‘lice’, “piojos” is <nozihi> (M_688). In (20), no genitive marker can be securely identified. The separate entry for ‘cow’ is <guacaré> (M_096), from Spanish “vaca” plus an unknown, possibly lexicalized suffix <-re>; glossing is tentative in (20). (20) <Chiguahé guacarahe> / <Chiguahé guacarahé> chiguahe offspring guaca-r-ahe cow-ls-gen ‘Calf’, “Becerrito” (M_655) We tentatively interpret the final element <-ahe> in <guacarahé> ‘of the cow’ as a genitive marker. In all these possessive constructions, the noun referring to the possessed entity precedes the noun or pronoun that refers to the possessor. 11.3 Locative: <-ni> In Section 8, it was suggested that <-ni> in <rini> ‘here’, “aquí” (M_553) and <chini> ‘there’, “allí” (M_554) has a locative meaning. The use of <rini> and <chini> is illustrated in (21–22). To what extent the use of <-ni> ‘locative’ is restricted to these demonstratives or was productive in 18thand 19th-century Andakí is impossible to determine for lack of data. Note the different position of <rini> and <chini> in the two examples. (21) <Coayaza rini> / <Coyazarini> coaya-za sit-imp ri-ni prox-loc ‘Sit (sg) here!’, “Siéntate aquí” (M_189) 59 11 Case markers and postpositions (22) <Chini caxini> chi-ni dist-loc ca-xi 2-go ni q ‘Which way are you (sg) going?’, “Por dónde vas?” (M_092) In (23), the adverb <chinta> / <chintá> ‘there’, “allá” (M_177; M_552) seems to take a variant <-ne> of this locative marker. (23) <Chinta nechayaya> / <Chinta nhayaya> chinta-ne that-loc haya-ya live-pl ‘There they live’, “Allá viven”5(M_331) 11.4 Allative: <-ra> A frequently attested Andakí grammatical morpheme is <-ra> ‘allative’. It is formally similar to Nasa Yuwe allative -na (see Díaz Montenegro 2019: 496).6This morpheme marks the complement of the verb ‘to go’ and may be attached to nouns and verbs alike, as shown in (24–26) , which suggests a clitic status of this morpheme. (24) <Jibi yahara> / <Jibi yahàra> ji-bi go-rea yaha-ra river-all ‘I am going to the river’, “Me voy al río” (M_292) (25) <Jihiza chiyara> / <Jihiza chiyará> jihi-za go-imp chiya-ra eat-all ‘Go (sg) eat!’, “Anda come” (M_317) (26) <Jibi chihizerá> / <Jibi chihizera> ji-bi go-rea chihize-ra urinate-all ‘I am going to urinate’, “Voy a orinar” (M_753) 5We have no explanation for the form <haya> as a variant of <caya> ‘to live’. In entry (M_764), we find the variant <aya> ‘to live’ (3, Section 15.1), and in entry (M_754), we find <zaya> ‘to remain’. 6For a correspondence of Andakí <r>: Nasa Yuwe /n/, compare also Andakí <ri-> ‘proximate demonstrative’ and Nasa Yuwe na ‘proximate demonstrative’; see Section 8. 60 11.4 Allative: <-ra> Since grammaticalization from allative to complementizer is quite widespread (Heine & Kuteva 2002: 37), Andakí <-ra> is discussed here, in the chapter dealing with case markers and postpositions, even if it also occurs with verbs, as in (25– 26). Besides to the complements of <ji> ~ <jihi> ‘to go’, as in (24–26), <-ra> is also attached to the complements of other motion verbs such as <buji> ‘to go away’ and <raje> ‘to bring’, both containing the root <ji> ‘to go’. These cases are shown in (27–28). (27) <Ninga bujibi cogora> ninga I bu-ji-bi transl-go-rea cogo-ra house-all ‘I want to go home’, “Yo quiero irme a casa” (M_291) (28) <Quarajea cogora> qua-ra-je-a 2.imp-caus-go-imp cogo-ra house-all ‘Bring (sg) me to your (sg) house!’, “Llévame a tu casa” (M_771) Allative <-ra> also follows the complement of <ynszi> ‘let us go’, as shown in (29–30). (29) <Inszijicora> inszi let.us.go jico-ra church-all ‘Let us go to the church!’, “Vamos a la iglesia” (M_128) (30) <Inszi rezara> / <Inszirezara> inszi let.us.go reza-ra pray-all ‘Let us pray!’, “Vamos a rezar” (M_129) Allative <-ra> might also be present in forms like <junquera> ‘upwards’, “arriba” (M_568) and <canara> ‘downwards’, “abajo” (M_570). In the previous examples illustrating the uses of allative <-ra>, the verb always precedes the allative complement. This order is not fixed, however, as shown in (31). (31) <Fichanara rajihi> / <Fichanará rajichi> ficha-na-ra brush-loc-all ra-ji-hi aor-go-rea ‘She/he/it left for the field’, “Se fue a la chagra” (M_161) 61 13 Adjectives and adverbs (4) <Quazira chiyaya> qua-zi good-rel ra-chiya-ya aor-eat-pl ‘They have eaten everything’, “Todo han comido” (M_168) By contrast, in (5–7), Andakí <qua> or <quan> ‘good’ is used predicatively, as an argument in copula complement function, and does not carry <-zi> ‘relative’. (5) <Quancaquehé> quan good ca-que-he 2-cop–rea ‘Are you (sg, m) good?’, “Estás bueno?” (M_258; M_267) (6) <Quahini mijinahé?> qua-hi good-rea ni q mijinahe? dog ‘Is your (sg) dog good?’, “Es bueno tu perro?” (M_271) (7) <Quahi> qua-hi good-rea ‘It is good’, “Bueno es” (M_272) Attributively used adjectives and relative clauses are not collapsed in all instances in Andakí, however. The morpheme <qua> ‘good’, shown in (3–7) above, also appears to be present in <nanquazi> ‘beautiful’, “bonito” (M_724); yet, the form with final <-zi> appears only when listed as a separate entry in the 18thcentury Andakí materials. As illustrated below, unlike <qua> ‘good’, <nan-qua> ‘cop-good’ and its variant <nin-qua> appear without final <-zi> not only in those examples where they have a predicative function (8), but also where they have an attributive function (9). Andakí <na-> ~ <nan-> ~ <nin-> is tentatively interpreted as a copular prefix here. It is also found in certain Andakí interrogative cleft-constructions (see Section 10). First, an example of <nanqua> in predicative function is shown in (8). (8) <Nanqua gaqui> nan-qua cop-good ga-qui 2-cop ‘You (sg, fem) are beautiful’, “Eres bonita” (M_122) Second, (9), likewise of a cleft construction, illustrates <nin-qua> ‘cop-good’ in attributive function. 68 (9) <Ninquaca guabi>2 nin-qua cop-good ca-gua-bi 2-say-rea ‘You (sg) say very well’, “Muy bien dice” (M_209) Andakí <-cu>, by contrast, attested in the isolated entry <yazicu> ‘ugly’, “feo” (M_650) or <yaseco> ‘bad’, “malo” (A_162), seems to have a distribution that can be compared to that of Andakí <-zi> ‘relativizer’ in <nanquazi> ‘beautiful’. As an isolated entry in the word lists, the form in question appears with final <-gu> ~ <-co>, and it occurs without final <-gu> ~ <-co> both in predicative and in attributive function, as shown in (10) and (11), respectively. (10) <Yazin Kaqui> yazin bad ka-qui 2-cop ‘You (sg, fem) are ugly’, “Eres fea” (M_133) (11) <Andagu quagua yazi fi-rajichi>3/ <Andagu quagua yazi firajichi> andagu quickly qua-gua 2.imp-cook yazi bad firajichi hungry ‘Cook (sg) quickly, I am hungry’, “Cocina presto, que tengo hambre” (M_151) Thus, it seems that <yazicu> / <yazeco> ‘ugly’, ‘bad’ and <nanquazi> ‘beautiful’ only carry the relativizing suffixes <-cu> and <-zi> if they are elicited in isolation, and occur without them in predicative and in attributive function in the available Andakí examples. A few Andakí adverbs do not seem to occur with a final suffix <-zi> or <- cu>. This is the case for <juatuxa> ‘straight’, “derecho” (M_670). In attributive function, this modifier appears without any further morphology, as shown in (12). There is no corresponding predicative construction in the available Andakí data. (12) <Cohagea juatuxa> / <Cohajea juatuxá> coha-ge-a 2.imp-go-imp juatuxa straight ‘Walk (sg) straight!’, “Anda derecho” (M_295) 2In Ms. II/2911, the original form is <ninquaqua guabi>, final <qua> is crossed out and <ca> is placed above it. 3Hyphenation is due to a line break in the manuscript. 69 13 Adjectives and adverbs To what extent the uses of the copula prefix and of the relativizing suffixes are also governed by semantic and pragmatic principles is difficult to glean from the available materials. In the neighboring Nasa Yuwe language, there is a suffix -sa, possibly a counterpart of Andakí <-zi> ‘relative’, which derives headless relative clauses and can be attached to color terms or verbs in this language (e.g., Jung 2008: 144–145; Díaz Montenegro 2019: 261). The use of Nasa Yuwe -sa, for instance with a color term, has been argued to add emphasis, compared with the use of the color term without -sa (Jung 2008: 148). 70 14 Verbs In this chapter, we will discuss the grammar of the Andakí verb and particular phenomena of the verbal domain. First, we will discuss those features that are expressed by prefixes and proclitics: causative marking (Section 14.1), the marking of spatial notions (Section 14.2), aspect/aspectual notions (Section 14.3), and verbal person marking (Section 14.4). This is followed by a discussion of those bound and unbound grammatical elements that follow the verbal root and mark mood (Section 14.5) and modality (Section 14.6). This chapter also comprises a discussion of verbal root suppletion (Section 14.7) and the copula in Andakí (Section 14.8). The Andakí word lists – in particular those from the 18th century – contain several verbs, that is, forms that typically occur as heads of a predicate, refer to actions, and take specific morphology such as markers of tense, aspect, and modality (see, e.g., Dixon 2010: 39, 52–53). In the anonymous 18th-century Andakí materials, individually listed verbs are easily identifiable in that they carry <-za>, which marks imperative mood. In a few cases, the verbal root and its nominal counterpart are nearly identical and seem to be etymologically related. This is the case with <shungua> ‘to hear’ (cf. M_731) and the corresponding noun <chunguahe> / <chunguahé> ‘ears’, “orejas” (M_424), which carries a stem formative <-he> ~ <-hé> (see Section 6). In other cases, the verb is fundamentally distinct from and probably etymologically unrelated to the corresponding noun, as is the case with <zega> ‘urine’, “orines” (M_021) and <chisi> ‘to urinate’ (cf. M_389). Overall, verbs are less frequently found in the available Andakí lists than nouns. The verb tends to occur at the end of an Andakí utterance in transitive and intransitive constructions alike. All intransitive and most transitive constructions have S or A expressed by a pronoun in the available Andakí materials, not by a noun. In intransitive constructions, in which S is expressed by a pronoun, SV is the most frequently attested constituent order, shown in (1), although VS also occurs, shown in (2). 14 Verbs (1) <Ninga buxibi> ninga I bu-xi-bi transl-go-rea ‘I go away’, “Yo me voy” (M_320) (2) <Yubi ninga> yu-bi come-rea ninga I ‘I come’, “Yo vengo” (M_256) There are only a few transitive constructions. A construction with AOV constituent order in which the transitive subject (A) is expressed by a personal pronoun is shown in (3).1 (3) <Ninga chatize guabi> ninga I chatize truth gua-bi say-rea ‘I tell the truth’, “Yo digo verdad” (M_349) In (4), a question in which the agent is expressed by a regular noun, the object follows the verbal predicate, and the constituent order is AVO. (4) <Cabiyara chiyaya quique?> / <Cabiyara chiyayá quique?> cabi-ya parrot-pl ra-chiya-ya aor-eat-pl quique? corn ‘Have the parrots eaten the corn?’, “Han comido el maíz los loros?” (M_166) Given the very limited data available and given the unusual constituent order in the Spanish translation, it is difficult to state how representative (4) is for Andakí constituent order in transitive clauses. In imperative constructions such as those shown in (5–8), the verbal predicate is likewise followed by the object and the order is VO. (5) <Quananqueha Xifi=> / <Quananqueha Xifi> qua-nan-que-ha 2.imp-caus-come-imp xifi candle ‘Bring (sg) a candle!’, “Trae candela” (M_075) 1An anonymous reviewer suggests that, alternatively, <chatize> in (3) functions as an adverbial modifier, not as an object. 72 (6) a. <Quananqueha Xizi. => / <Quananqueha Xizi> qua-nan-que-ha 2.imp-caus-come-imp xizi firewood ‘Bring (sg) firewood!’, “Trae leña”2(M_076) b. <Quananquea jizi> / <Guananquea fizi> qua-nan-que-a 2.imp-caus-come-imp jizi firewood ‘Bring (sg) firewood!’, “Trae leña” (M_142) (7) <Quananqueha Jexe. => / <Quananqueha Jexé> qua-nan-que-ha 2.imp-caus-come-imp jexe water ‘Bring (sg) water!’, “Trae agua” (M_074) (8) <Andagu biguaza bacoxe> / <Andagu biquaza bacoxé> andagu quickly bi-gua-za rap-make-imp bacoxe chicha ‘Make (sg) chicha quickly!’, “Haz presto chicha” (M_150) Cases where the object precedes the verbal predicate in imperative constructions (OV order) seem to occur somewhat less frequently; two such instances are illustrated in (9–10). The translation of (9) is the same as that of (7) above. (9) <Jixe nanqueza> jixe water nan-que-za caus-come-imp ‘Bring (sg) water!’, “Trae agua” (M_031) (10) <Bacoxe quananqueha riszijaxi.=> / <Bacoxe quananquehá riszijaxi> bacoxe chicha qua-nan-que-ha 2.imp-caus-come-imp riszi-jaxi drink-purp ‘Bring (sg) chicha in order to drink’, “Trae chicha para beber” (M_078) Constituent order in the Andakí imperative constructions (5–10) may be governed by information structure. As to the arrangement and ordering of grammatical morphemes in the verb or verb phrase, valency changes, spatial and aspectual notions, and mostly also verbal person (2nd-person subject) are marked by prefixes or proclitics, whereas 2It remains to be established whether or not the terms for candle <xifi> and firewood <xizi> are etymologically related. 73 14 Verbs mood and modality are marked by bound or unbound grammatical elements that follow the verb. Given the incomplete nature of the available Andakí materials, several domains of the grammar of the verb must remain open, because evidence for them is found only in single examples. Among these topics are subordination and nominalization. Besides the relativizer <-ze> discussed above, in Section 13, a possible subordinating ending in Andakí is <-chaza>. Its use is illustrated in (11). (11) <Firajiquichaza cobaquea rincaxahá> firaji hungry qui-chaza cop-sub coba-que-a 2.imp-come-imp rinca-xaha I-obl ‘When hungry, come (sg) to me!’, “En teniendo hambre ven a mí” (M_762) A suffix <-na> seems to function as a locative nominalizer, as shown in (12), where it derives the term for ‘field’ <fichana> (M_161) from the verb ‘to brush’ <ficha> (cf. M_162). (12) <Fichanara rajihi> / <Fichanará rajichi> ficha-na-ra brush-loc-all ra-ji-hi aor-go-rea ‘She/he/it went to the field’, “Se fue a la chagra” (M_161) Besides suffixation and prefixation, reduplication may be a marginal morphological device in Andakí; the only example of this is the case of <mi> ‘to want’ (cf. M_304) and <mimi> ‘to love’ (cf. M_232), where reduplication seems to indicate intensity. In what follows, we give an overview of those grammatical morphemes in the verbal domain for which we could find more substantial evidence. 14.1 Causative: <ra-> There is little information on verbal derivational morphology in Andakí. A derivational prefix which is attested with a range of verbs is the causative marker <ra-> (M_194), <za-> (M_757), <naha-> (M_303), <na-> [na-] (M_202), or <nan-> [nã-] (M_289). For instance, it derives the verb ‘to bring, to carry’ from the verb ‘to go’, as illustrated in (13–14). 74 14.2 Spatial notions: <bu-> ‘translocative’, <chi-> ‘inward’, <fiqui-> ‘upward’ (13) <Rica Kazaxihi>3 rica you ka-za-xihi 2-caus-go ‘You (sg) brought’, “Tú llevaste” (M_757)4 (14) <Rajiza rihizi> ra-ji-za caus-go-imp rihizi this ‘Take (sg) this!’, “Lleva a este” (M_194) Causative <ra-> has the variant <na-> and <naha->, illustrated in (15) and (16), respectively. (15) <Naqua naquiquani?> / <Náqua náquiguani?> na-qua cop-who na-qui caus-come qua who ni? q ‘Who brought you (sg)?’, “Quién te trajó?” (M_202) (16) <Ypchize canahá quine?> / <Ypchizé canahá quiné?> ypchize what ca-naha-qui 2-caus-come ne? q ‘What did you (sg) bring?’, “Qué has traído?” (M_303) In other cases, changes in verbal valency are lexically encoded in Andakí, for instance, in <qua> ‘to kill’ (cf. M_006) versus <ma> ‘to die’ (cf. M_005). 14.2 Spatial notions: <bu-> ‘translocative’, <chi-> ‘inward’, <fiqui-> ‘upward’ Andakí has some grammatical morphemes expressing spatial notions or associated motion on verbs. The prefix <bu->, for instance, may encode a translocative meaning, as suggested by the pairs <ji> ‘to go’ (cf. M_039, M_753) versus <buji> ‘to go away’ (cf. M_190). The prefix <bo-> or <bu-> may also express associated motion with non-motion verbs, as suggested by <daza> ‘sleep! (sg)’, “dormí” (M_356) versus <bondaza> ‘(go and) sleep! (sg)’, “dormí” (M_355) and <bunta> ‘go and sleep’ illustrated in (17).5 3It is not clear whether or not final <-hi> belongs to the root or is a realis mood suffix. 4This example is attested in Ms. II/2911 only. 5As pointed out by an anonymous reviewer, associated motion is a common category among languages of western South America; it occurs, for instance, in Mosetén (isolate), Pano-Tacanan and Quechuan (e.g., Ross 2021). 75 14 Verbs (17) <Quabuntahá> qua-bu-nta-ha 2.imp-am-sleep-imp ‘Go and sleep (sg)!’, “Anda duerme” (M_318) Inward motion is probably expressed by the prefix <chi->, illustrated in (18). (18) <Guachijiza andagu> gua-chi-ji-za 2.imp-iw-go-imp andagu quickly ‘Enter (sg) quickly!’, “Entra presto” (M_141) A prefix <fiqui> or <fiji>, shown in (19–20), is tentatively glossed as ‘upward motion’. (19) <Fiquijizá> fiqui-ji-za uw-go-imp ‘Stand up (sg)!’, “Levántate” (M_372) (20) <Quafixi Jea> / <Quafixijea> qua-fixi-je-a 2.imp-uw-go-imp ‘Stand up (sg)!’, “Levántate” (M_089) 14.3 Aspectual notions: <ra-> ‘aorist’, <bi-> ‘rapid’ We found no specific markers for imperfective aspect and related notions in the available Andakí data. There are two examples with 3rd-person subjects translated with a progressive meaning. These examples do not bear any obvious aspectual morphology, as illustrated in (21–22). (21) <Fichahe> / <Fichahé> ficha-he brush-rea ‘She/he/it is brushing’, “Está rozando” (M_162) (22) <Zancahaya> zancaha-ya talk-pl ‘They are talking’, “Están hablando” (M_125) 76 14.3 Aspectual notions: <ra-> ‘aorist’, <bi-> ‘rapid’ We tentatively suggest that the aorist (past tense with perfective aspect) is marked by a prefix <ra->. We found instances of this prefix only in clauses with a 3rd-person subject, as illustrated by (23–25); that is, this prefix may have been used exclusively with non-speech act participants. (23) <Canchihi raquahi> canchihi lightning ra-qua-hi aor-kill-rea ‘The lightning killed him/it’, “Lo mató el rayo” (M_220) (24) <Ramahi> ra-ma-hi aor-die-rea ‘She/he/it died’, “Murió” (M_322) (25) <Naqua Raguaquani?> / <Naqua raguaquani?> na-qua cop-who ra-gua aor-say qua who ni? q ‘Who told you (sg)?’, “Quién te dijo?” (M_730) In a few instances, <ra-> occurs in examples that are translated with the Spanish perfect, as in (26). (26) <Cabiyara chiyaya quique?> / <Cabiyara chiyayá quique?> cabi-ya parrot-pl ra-chiya-ya aor-eat-pl quique? corn ‘Have the parrots eaten the corn?’, “Han comido el maíz los loros?” (M_166) There is also one sentence containing <ra-> which is translated, however, with the Spanish present tense (27). (27) <Nszazi raguayani xin-gaxá?>6/ <Nszâzi raguayani xingaxá?> nszazi how/where ra-gua-ya aor-say-pl ni q xinga-xa? I-obl ‘What do they say about me?’, “Qué dicen de mí?” (M_758) The aorist-marking prefix <ra-> seems to have a variant <na->, illustrated in (28). 6The hyphen is due to a line break in the manuscript. 77 14 Verbs To what extent this distribution of <-bi> – in 1st-person subject statements and 2nd-person subject questions – can be linked to egophoric person marking, a relatively widespread feature in several languages of northwestern South America (e.g., Knuchel 2015; San Roque et al. 2018), is not easy to establish, given the lack of data and the variability of transcriptions. Also, in a few cases, <-hi> occurs with a 2nd-person subject in an interrogative clause (57), and in a few instances, <-bi> occurs with a 3rd-person subject (58). These cases, however, are exceptions in the available Andakí materials. (57) <Cafianzuhi ningaxa ca-yasesza?>10 / <Cafianzuhi ningaxá cayasesza?> ca-fianzu-hi 2-like-rea ninga-xa I-obl caya-sesza? live-des ‘Do you (sg) want to live with me?’, “Quieres vivir conmigo?” (M_353) (58) <Raxinachibi> / <Raxinachihic> ra-xi caus-go nachi-bi night-rea ‘Night fell’, “Anocheció” (M_124)11 Realis mood <-hi> seems to have the variants <-ha>, <-fa>, and <-a>. The use of <-ha> and <-fa> with nouns in predicative function is illustrated in (59–60). (59) <Bacoxe hane?> bacoxe-ha chicha-rea ne q ‘Is there chicha?’, “Hay chicha?” (M_046) (60) <Noszuefane?> / <Noszuefané?> noszue-fa fish-rea ne q ‘Is there fish?’, “Hay pescado?” (M_324) In (61–65), we only tentatively interpret final <ha> as a realis mood marker. In (61–62), we present some examples containing the negation marker <para> and related forms. (61) <Ninga pajaha> / <Ninga pajahá> ninga I paja-ha neg-rea ‘I do not have’, “Yo no tengo” (M_228) 10Hyphenation is due to a line break in the manuscript. 11This interpretation of <raxi> in (58) as ‘to cause to go, to bring’ is preliminary. 84 14.5 Mood (62) a. <Psajaha> / <Pajaá> psaja-ha neg-rea ‘I do not have; there is not’, “Yo no tengo; no hay” (M_229) b. <Pacahá> paca-ha neg-rea ‘I do not have; there is not’, “No tengo; no hay” (M_231) The negation marker may also occur without the realis mood marker, as illustrated by <pagá> ‘I do not have; there is not’, “no tengo; no hay” (M_230). (63–65) show further instances of the presumed realis mood marker <-ha> ~ <-há>. (63) <Kachiyaha> / <Kachiyahá> ka-chiya-ha 2-eat-rea ‘Did you (sg) already eat?’, “Ya comiste?” (M_734) (64) <Raguayahá> ra-gua-ya-ha aor-say-pl-rea ‘They said’, “Dijeron” (M_729) (65) <Nanquaha> / <Nanquahá> nan-qua-ha cop-good-rea ‘He/it is good’, “Bueno está” (M_725) 14.5.2 Imperative1: <-(a)ba> There are different markers of the imperative mood in Andakí which, for practical purposes, will be called imperative1, imperative2, and imperative3here. The pragmatics of these different markers cannot be determined due to the nature of the available Andakí data. In the case of imperative1we find different forms for the 1st and 2nd person, discussed in Section 14.5.2.1 and Section 14.5.2.2, respectively. 14.5.2.1 First person imperative1: <Ø-…-(a)ba> What we tentatively label imperative1in this work is marked by a morpheme <-ba>. Whereas in imperative mood, the 2nd person is marked by a dedicated prefix, the 1st person (hortative) is not. This is illustrated in (66). 85 14 Verbs (66) <Chiyaba gunfigo> chiya-ba eat-imp gunfigo one ‘Let us eat together!’, “Comamos juntos” (M_293) As in the indicative mood, a zero morpheme (Ø-chiya-ba ‘let us eat!’) may be postulated as a 1st-person subject marker. Imperative <-ba> has a variant represented by <-aba>, which seems to occur only with the verb <buxi> (and variants) ‘to go away’, as illustrated in (67–68). (67) <Nahachi boxeaba => / <Nahachi boxeaba> nahachi night bo-xe-aba transl-go-imp ‘Let us go away, it is late!’, “Vámonos, que es tarde” (M_087) (68) a. <Boxeaba => / <Boxeaba> bo-xe-aba transl-go-imp ‘Let us go away!’, “Vámonos” (M_085) b. <Bujeaba> bu-je-aba transl-go-imp ‘Let us go away!’, “Vámonos” (M_281) Another variant of the imperative suffix is represented by <-aha>; its use with a 1st-person singular subject being illustrated in (69). (69) <Quananquea chiguaca nunqueaha> / Quananquea chiguaca nunqueahá> qua-nan-que-a 2.imp-caus-come-imp chiguaca children nunque-aha look-imp ‘Bring me your (sg) children, I want to see them’, “Tráeme a tus hijos, que quiero verlos” (M_289) 14.5.2.2 Second person imperative1: <qua-…-ba> With the imperative marked by <-ba> or a variant of it, 2nd-person subjects are marked by a prefix <qua->, <coba->, <coha->, or <coa->. Some examples are shown in (70–75). 86 14.5 Mood (70) <Firajiquichaza cobaquea rincaxahá> firaji hungry qui-chaza cop-sub coba-que-a 2.imp-come-imp rinca-xaha I-obl ‘When hungry, come (sg) to me!’, “En teniendo hambre ven a mí” (M_762) (71) <Cohagea andagu> coha-ge-a 2.imp-go-imp andagu quickly ‘Go (sg) quickly!’, “Anda presto” (M_296) (72) <Coagua anduazo> coa-gua 2.imp-cook anduazo banana ‘Cook (sg) bananas!’, “Cocina plátanos” (M_227) (73) <Quanca jehá sasza> / <Quanajeha sasza> quan-caje-ha 2.imp-wash-imp sasza clothes ‘Wash (sg) the clothes!’, “Lava la ropa” (M_210) (74) <Quarichá> qua-ri-cha 2.imp-drink-imp ‘Drink (sg)!’, “Bebe”12 (M_047) (75) <Quaxeba => / <Quaxeba> qua-xe-ba 2.imp-go-imp ‘Go (sg)!’, “Anda” (M_083) The following case (76) is the only example where a personal pronoun occurs in an imperative construction of the kind discussed here. (76) <Quaxiha Rica> / <Quaxihá Ricá> qua-xi-ha 2.imp-go-imp rica you ‘You (sg) go!’, “Anda vos” (M_051) In this kind of construction, the personal pronoun might be used for emphasis (see Section 9). Similar constructions without the personal pronoun are shown in (75) above and in (77). 12That <ch> can refer to an allophone of the phoneme referred to by <h> is in line with pairs such as <cachinehe> / <cachineche> ‘grandson’, ”nieto” (M_490). 87 14 Verbs (77) <Quexiha => / <Quexiha> que-xi-ha 2.imp-go-imp ‘Go (sg)!’, “Anda” (M_039) 14.5.3 Imperative2: <-ra> ~ <-za> In single entries, verbs are presented with a suffix <-ra> or <-za> in the 18thcentury lists; these forms are translated as imperatives in Spanish. This is illustrated in (78–80). (78) <Chiyazá> chiya-za eat-imp ‘Eat (sg)!’, “Come” (M_354) (79) <Finszuzá> finszu-za clean-imp ‘Clean (sg)!’, “Limpia” (M_370) (80) <Cayaza> caya-za sit-imp ‘Sit down (sg)!’, “Siéntate” (M_376) Andakí <-za> is glossed here as an imperative marker. It may also occur with a 2nd-person plural subject, as illustrated in (81). (81) <Chuhuaza> / <Chuhuazá> chuhua-za hear-imp ‘Hear (pl)!’, “Oíd” (M_362) Andakí <-za> is also found with a 1st-person subject in (82). Note that the subject person is expressed by a pronoun in this case. 88 14.5 Mood (82) <Ninga nunquiza chin-queca riguacu>13 / <Ninga nunquiza chinqueca riguacu> ninga I nunqui-za look-imp chinqueca children riguacu you.pl ‘I have to look upon you (pl) as children’, “Yo os he de mirar como a hijos” (M_241) A possible use of <-za> with a 3rd-person subject is shown in (83), which, however, may also address a 2nd-person subject: given that in polite commands Spanish uses the 3rd-person subjunctive, this form is ambiguous. (83) <Quihizá> quihi-za come-imp ‘Let him/her/it come; may she/he/it come; come (sg)!’, “Que venga” (M_410) It is unclear to what extent the different imperatives, called imperative1and imperative2here for practical purposes, differ in illocutionary force and politeness. Judging from the Spanish translations provided in the available language materials, the imperative constructions in the present section (imperative2) seem to have the same meaning as those with <-(a)ba> and variants (imperative1), illustrated in the previous section. This is also suggested by the translations shown in (84–87), illustrating the two different imperative markers with the verb ‘to come’ (84–85) and with the verb ‘to bring’ (86–87). (84) <Quaqiha => / <Quaquiha> qua-qui-ha 2.imp-come-imp ‘Come (sg)!’ “Vení” (M_038) (85) <Quihizá> quihi-za come-imp ‘Come (sg)!’, “Ven” (M_419) 13The hyphen is due to a line break in the manuscript. The form <riguacu> ‘you.pl’ is unattested elsewhere in the available Andakí data. 89 14 Verbs (86) <Quananqueha Jexe. => / <Quananqueha Jexé> qua-nan-que-ha 2.imp-caus-come-imp jexe water ‘Bring (sg) water!’, “Trae agua” (M_074) (87) <Jixe nanqueza> jixe water nan-que-za caus-come-imp ‘Bring (sg) water!’, “Trae agua” (M_031) Note that the object (‘water’) follows the imperative verb in (86) and precedes it in (87). In some instances, there is a variant <-ra> of the imperative suffix, illustrated in (88–89); once more, it is not entirely clear whether the utterances address a 2nd or a 3rd person. (88) <Quijra> quij-ra come-imp ‘Let him/her/it come; may she/he/it come; come (sg)!’, “Que venga” (M_160) (89) <Andagu quira> / <Andaguquira> andagu quickly qui-ra come-imp ‘Let him/her/it come; may she/he/it come quickly; come quickly (sg)!’, “Que venga presto” (M_155) In one instance, we found <-ca> as a variant of imperative <-za>; this is illustrated in (90). (90) <Raxicá> / <Racizá> ra-xi-ca caus-go-imp ‘Carry (sg)!’, “Lleva” (M_379) The graphemic variation in the context of imperative2needs further investigation; since the underlying rules are not yet understood. Despite its more frequent occurrence as <-za>, we suggest in Section 5.3 that the underlying, phonemic form of the Andakí imperative1suffix is /-ra/. 90 14.5 Mood 14.5.4 Imperative3: <-ni> The word lists also provide us with imperative constructions that seem to contain a suffix or enclitic <-ni>, such as <sani> ‘wait (sg)!’, “espera” (M_066), <szuzini> ‘stay (pl) here!’, “estaos quedo” (M_236), <fsatani> / <fsâtani> ‘shut up, do not talk (sg)!’, “calla, no hables” (M_238). We have not been able to identify the verbal roots in question in other entries, which makes it difficult to interpret <-ni> as an imperative marker, except in the cases of (91–92); Andakí <qua> ‘to make’ is also attested elsewhere, for instance, in (8) and (30) above. (91) <Quani nanqua cogo> qua-ni make-imp nan-qua cop-good cogo house ‘Make (pl) a good ranch; they should make a good ranch!’, “Hagan buen rancho” (M_215) (92) <Cogo quani naqua> cogo house qua-ni make-imp na-qua cop-good ‘You have to make me a house!’, “Me habéis de hacer casa” (M_249) Note the different word order and translations in (91) and (92).14 The element <-ni> occurs not only with verbs, but also with adverbials, as illustrated in (93).15 (93) <Quixarani> quixara-ni far-imp ‘Let us go far away!’, “Vamos lejos” (M_736) The adverbial expression ‘far’, “lejos” is documented elsewhere as <quejara> (M_557). Finally, a suffix <-no> may express imperative mood in <yquano> / <yguano> ‘listen (sg)!’, “oiga” (M_264). The root <yqua> / <ygua> may be related to <shungua> ‘to hear, understand’ (cf. M_731), but it is unclear how <-no> relates to <-ni>. 14An anonymous reviewer observes that the translation does not reflect all elements of the original text in (92). As mentioned in Section 4 above, this happens in several cases in the available Andakí data. 15An anonymous reviewer observes that (93) looks like a rare case of verb root ellipsis. 91 14 Verbs 14.5.5 Prohibitive: <ni-…-qua> The prohibitive is marked by a circumfix <ni-…-qua>. For illustrative purposes, we juxtapose an imperative and a prohibitive form in (94–95). The imperative form is shown in (94), the prohibitive form in (95). (94) <Quaquaha> / <Quaquahá> qua-qua-ha 2.imp-fight-imp ‘Go fight (sg)!’, “Anda pelea” (M_300) (95) <Niquaquaha> / <Niquaquahá> ni-qua-quaha proh-fight-proh ‘Do not fight!’, “No peleéis” (M_301) The prohibitive suffix has several variants, among which are also <-qua> and <-coha>, illustrated in (96) and (97), respectively.16 (96) <Ninaquá> (M_197), <Ninaqua> (M_237) ni-na-qua proh-cry-proh ‘Do not cry (sg)!’, “No llores” (M_197; M_237)17 (97) <Nibugicoha> / <Nibujicohá> ni-bu-ji-coha proh-transl-go-proh ‘Do not go away (sg)!’, “No te vayas” (M_314) A remarkable phenomenon in Andakí is that the verb ‘to give’, illustrated in (98–101), looks like a prohibitive construction; the respective parts are highlighted in bold. However, the Spanish translations suggest otherwise and it has not been possible to analyze the constructions in question in more detail. The remarkable combination of letters <fsrr> probably refers to a sibilant (Coronas Urzúa 1994: 90–91; see also Section 5.3.5). 16An anonymous reviewer suggests that the suffix <-qua> and its variants may also be analyzed as the 2nd-person imperative marker. The fact that all provided examples involve 2nd-person referents suggests this as a possible interpretation. 17In another context, the root ‘to cry’ appears in a slightly different form: <naha> – compare naha-za ‘cry!’, ”llora” (M_384). 92 14.6 Modality (98) <Nifsrraquahá> nifsrraquaha give ‘You will give’, “Daréis” (M_726) (99) <Nifsrranquazá> nifsrranqua-za give-imp ‘Give (sg) me!’, “Dame” (M_415) (100) <Bacoza nifsrranquaza> bacoza mazato nifsrranqua-za give-imp ‘Give (sg) me mazato!’, “Dame mazato” (M_309) (101) <Bacoxe nifsrran => / <Bacoxe nifssrânquaza> bacoxe chicha nifssranqua-za give-imp ‘Give (sg) me chicha!’, “Dame chicha” (M_032) The two forms shown in (101), <nifsrran> and <nifssrânquaza>, each attested in a different version of Mutis’ Andakí materials, suggest that final <qua> in <nifsrranqua> is indeed a separate morpheme.18 14.6 Modality There are several markers of modality in Andakí, that is, morphemes that express inner states and attitudes of a speaker and indicate irrealis/counterfactual notions of will, obligation, necessity, or ability. This section will discuss the expression of desiderative (Section 14.6.1) and non-desiderative(Section 14.6.2) modality in Andakí. 14.6.1 Desiderative: <-zea> The morpheme <-zea> and its variants <-cea> and <-seza> encode a desiderative or volitive meaning. Their use is illustrated in (102–103). 18The formal similarity of <nifsrran> / <nifsrraqua> ‘to give’ with the construction <nifsrracaquahá> ‘do not stumble!’, “no tropieces” (M_750) is remarkable. 93 14 Verbs (129) <Guaca juncaquehe?> / <Guaca juncaquehé> guacaju canoe nca prop que-he? cop-rea ‘Do you have canoes?’, “Tenéis canoas?”27 (M_246) (130) <Chanca encaquehé?> / <Chanca encaquehé?> chanca husband enca prop que-he? cop-rea ‘Do you have a husband?’, “Tenéis marido?” (M_172) Note that we only have access to examples with 2nd-person subjects in the context of the Andakí copula, which limits the possible generalizations that can be made. 27In Ms. II/2911, the Spanish translation says: ”Tenéis cañas?”. 100 15 Further grammatical morphemes and features In this chapter, we will consider grammatical elements that cannot clearly be linked to the parts of speech discussed so far. They either occur with elements pertaining to different word classes, such as the plural clitic <-ya>, which can be attached to verbs and nouns, but not to pronouns (Section 15.1), or they are unbound forms such as the negation marker <para> (Section 15.2), the interrogative marker <ni> (Section 15.3), and the hortative expression <ynsci> (Section 15.4). Augmentative, diminutive, and related notions are discussed in Section 15.5. The absence of dedicated tense markers in Andakí is discussed in Section 15.6. 15.1 Plural: <-ya> Plural number is marked by a clitic <-ya>. This morpheme can be attached to verbs and nouns alike, yet not to pronouns. In verbs, there is no evidence that Andakí <-ya> also pluralizes 1stand 2nd-person subjects, and it seems that plural number is overtly marked by <-ya> only with a 3rd-person subject. The morpheme <-ya> marks plural number of human referents, as in (1). (1) <Chinta nehayaya> / <Chinta nchayaya> chinta-ne there-loc haya-ya live-pl ‘There they live’, “Allá viven” (M_331) Non-human, animate plural subjects also receive <-ya>. In (2), plural number is marked on both the subject and the verb. (2) <Cabiyara chiyaya quique?> / <Cabiyara chiyayá quique?> cabi-ya parrot-pl ra-chiya-ya aor-eat-pl quique? corn ‘Have the parrots eaten the corn?’, “Han comido el maíz los loros?” (M_166) 15 Further grammatical morphemes and features The morpheme <-ya> also attaches to <juanti> ‘big, a lot’, which in (3) has a human referent. (3) <Juantiya neayaya?> / <Juantiya neayayá?> juanti-ya big/a.lot-pl ne q aya-ya? live-pl ‘Are there many people in your land?’, “Hay mucha gente en tu tierra?” (M_764) It is difficult to determine whether the marking of plurality with <-ya> depends on the degree of animacy or of agency of the 3rd-person plural referent. Inanimate referents are rarely agents of an action expressed by a lexical verb, which is in line with the observation that there are no examples, in the available Andakí data, of pluralized nouns referring to inanimate entities. 15.2 Negation: <para>, <ra-…-nca> There are several ways to express negation in Andakí (see also Section 14.5.5 on the prohibitive and Section 14.6.2 on the non-desiderative). The most widespread negative morpheme used with both nouns and verbs is a preposed element <para> that is always represented as an unbound form in the manuscripts. In several cases, it seems to have grammaticalized and to have given rise to a negation-marking prefix or proclitic which only seems to occur with verbs, however. The unbound negation marker <para> has several variants, among which are <paxa> / <paxá> (M_143) and <paga> / <pagá> (M_310). Its use with nouns in negative nominal predicates is illustrated in (4) and (5), respectively. (4) <Paxa jizi> / <Paxá fizi> paxa neg jizi firewood ‘There is no firewood’, “No hay leña” (M_143) (5) <Sisza pagaha> sisza name paga-ha neg-rea ‘I have no name’, “No tengo nombre” (M_307) 102 15.2 Negation: <para>, <ra-…-nca> Although in most cases, <para> or one of its variants precedes the negated entity, as in (4), in a few instances it follows it, as illustrated in (5). The negation marker <para> and variants may also occur alone, and constitute a single utterance, as shown in (6). (6) <paja>; <paga> / <pajá>; <pagá> paja neg ‘I do not have; there is not’, “No tengo; no hay” (M_149) In verbal negation, Andakí <para> and variants occur together with other negative morphemes, as shown in (7–8). (7) <Pará nanqueanca> / <Pará nanquetanca>1 para neg nan-que-anca caus-go-neg ‘I bring nothing’, “Nada traigo” (M_737) (8) <Para riszifigua> para neg riszi drink figua ndes ‘I do not want to drink’, “No quiero beber” (M_049) In most cases, however, a truncated form, <ra->, occurs as a negative prefix or proclitic. Its use is shown in (9–10). (9) <Ragua chanca> ra-guacha-nca neg-find-neg ‘I have not found anything’, “Nada he topado” (M_165) (10) <Ninga rahasza ficoha> ninga I ra-hasza neg-leave ficoha ndes ‘I do not want to leave you (sg)’, “Yo no quiero dejarte” (M_251) As in the case of <para> and its variants, negation of a verb with <ra-> entails the use of a negation marker <ficoa> or <-anca>. The use of the negative suffix <-anca> or <-nca> is shown, for instance, in (7) and (9) above, and in (11) below. 1We have no explanation for <t> in this form; it may have an antihiatic function here or be an error. 103 15 Further grammatical morphemes and features (11) <Ninga januzira guanca> / <Ninga janûzira guanca> ninga I janu-zi big/a.lot-rel ra-gua-nca neg-say-neg ‘I do not ask you (sg) for much’, “Yo no te pido harto” (M_181) The use of <-anca> and variants appears to be mutually exclusive with the use of non-desiderative <ficoa> and related forms, the use of which is illustrated, for instance, in (12). (12) <Ramaficora> ra-ma neg-die ficora ndes ‘Do not die’, “No muráis” (M_323)2 The prefix <ra-> has a variant <za->, illustrated in (13); note also the variant <-yanca> of the negative suffix, with <y> introduced possibly for antihiatic reasons, like <t> in (7) above. (13) <Buxibi firajichiza chiya yanca> / <Buxibi firajichiza chiyayanca> bu-xi-bi transl-go-rea firajichi hungry za-chiya-yanca neg-eat-neg ‘I am going to eat, I am hungry’, “Voy a comer, que tengo hambre” (M_028) In (14), <ca-> is probably a negation marker and a variant of <ra->. (14) <Cachiyanca> / <Cachiyancá> ca-chiya-nca neg-eat-neg ‘I have not eaten’, “No he comido” (M_045) The interaction between negative <ra-> and the 2nd-person subject marker <ca-> are not yet fully understood; <ca-> and <ra-> do not seem to co-occur on the same word as separate prefixes, as shown in (15–16). (15) <Rafianzanca?> ra-fianz-anca? neg-like-neg ‘Why do you (sg) not want to?’, “Por qué no quieres?” (M_260)3 2Attested in Ms. II/2912 only. This form follows the form <ramahi> ‘she/he/it died’, “murió”. 3Note that there is no interrogative pronoun ‘why?’ in the Andakí example. 104 15.3 Interrogative: <ni> (16) <Yhiza raqueanquini?> yhiza why ra-que-anqui neg-come-neg ni? q ‘Why do you (sg) not come?’, “Por qué no vienes?” (M_325) In the last example, <-anqui> is interpreted as a variant of <-anca>; final <i> remains unexplained but might reflect assimilation to the high front vowel in the following interrogative morpheme <ni>. Finally, if it is attached to a noun, the negation prefix seems to occur without any further negation suffix, as in (17). (17) <Rashunguahé> ra-shunguahe neg-ear ‘Stupid (m)’, “Tonto” (M_649) 15.3 Interrogative: <ni> Questions in Andakí often contain an unbound interrogative morpheme <ni>. It occurs both in polar questions and in content questions. Its use in polar questions is illustrated in (18). (18) <Quahini mijinahé?> qua-hi good-rea ni q mijinahe? dog ‘Is your (sg) dog good?’, “Es bueno tu perro?” (M_271) It is also used in content questions, as shown in (19). (19) <Fizini?> / <Fizini.=> fizi what ni? q ‘What is she/he/it?’, “Qué es?” (M_072) In (18–19), the interrogative marker follows the predicate. If the construction contains only a verb, <ni> seems to occur only in initial position, as shown in (20a–20b). 105 15 Further grammatical morphemes and features (20) a. <Niyuhe?> ni q yu-he? come-rea ‘Does she/he/it come?; Has she/he/it come?’, “Viene?; Ha venido?” (M_156) b. <Niyuhe?> / <Niyuche?> ni q yu-he? come-rea ‘Does she/he/it come?; Has she/he/it come?’, “Viene?; Ha venido?” (M_024) In constructions that contain both a question word and a verb, <ni> seems to occur only in final position, as shown in (21–23). (21) <Ychuize camine?> / <Ypchize caminé?> ychuize what ca-mi 2-want ne? q ‘What do you (sg) want?’, “Qué quieres?” (M_304) (22) <Ychuyzi Kazini?> ychuyzi what ka-zi 2-do ni? q ‘What do you (sg) do?’, “Qué haces?” (M_009) (23) <Sazi ca-yu-ni?> sazi how/where ca-yu 2-come ni? q ‘How have you (sg) come?’, “Cómo has venido?” (M_308) In a number of questions, however, we find no use of <ni> at all, as illustrated in (24–25). (24) <Rafianzanca?> ra-fianz-anca? neg-like-neg ‘Why do you (sg) not want to?’, “Por qué no quieres?” (M_260) (25) <Cafsrriaxi?> ca-fsrriaxi? 2-forget ‘Did you (sg) forget?’, “Te olvidaste?” (M_117) 106 15.4 Hortative: <ynszi> Note that in (24), which is a content question, there is no interrogative pronoun. 15.4 Hortative: <ynszi> An Andakí hortative expression is <ynsci> / <ynszi> (M_313), <inci> (A_067), <inszi> (M_129). These forms are translated as ‘let us go!’, “vamos” and ‘let us go away’, “vámonos”. They often occur without further morphemes, as in (26a–26b). (26) a. <Ynszi> / <Ynsci> ynszi let.us.go ‘Let us go away!’, “Vámonos” (M_313) b. <Inci> inci let.us.go ‘Let us go!’, “Vamos” (A_067) In a few cases, the imperative marker <-za> is attached to <ynszi> ‘let us go!’, as shown in (27). (27) <Ynsziza> ynszi-za let.us.go-imp ‘Let us go!’, “Vamos” (M_022) The most frequent construction in the data is <ynszi> ‘let us go!’ + verb/noun + <-ra> ‘allative’. The form glossed as ‘let us go!’ always occupies the first position in the sentence. In (28), the complement is a verb, while in (29), the complement is a noun. (28) <Ynsci fichara> / <Ynsci fichará> ynsci let.us.go ficha-ra brush-all ‘Let us go brush!’, “Vamos a rozar” (M_328) (29) <Ynszi zota jera> / <Ynszi zotajera> ynszi let.us.go zotaje-ra mountain-all ‘Let us go to the forest!’, “Vamos al monte” (M_253) 107 15 Further grammatical morphemes and features <Ynszi> (and variants) and the imperative with -(a)ba may also occur in a single (complex) sentence, as shown in (30), where they are found in two separate syntactic units. (30) <Ynszi yahara nâjubahá> ynszi let.us.go yaha-ra river-all naju-baha wash-hort ‘Let us go to the river to wash ourselves!’ / ‘Let us go to the river, let us wash!’, “Vamos al río a lavarnos” (M_254) 15.5 Augmentative, diminutive, and related notions Among the grammatical morphemes that may tentatively be identified in Andakí adjectives and nouns are those expressing augmentative, diminutive, and related notions. A suffix <-pi>, <-pihi>, or <-nipihi> and variants may have an augmentative meaning, and occur in forms such as <jushuampihi> / <jushuanipihi> ‘big’, “gordo” (M_700), <cupihi> ‘major’, “mayor” (M_698), <acozinipihi> ‘wicked’, “bellaco” (M_582), <quasinimpihi> ‘good’, “bueno” (M_583),4but also in nouns such as <chupihi> ‘grandfather’, “abuelo” (M_488). The diminutive counterpart <-bi> or <-bihi> may be attested in forms such as <chibihi> ‘minor’, “menor” (M_699)5and in nouns such as <cabihi> ‘parrot’, “loro” (M_600) and <rimbihi> ‘turtledove’, “tortola” (M_607). 15.6 Tense There is no evidence for either verbal tense or nominal tense marking in the available Andakí data, and it is difficult to know how notions of nominal or verbal tense were indicated at all. For instance, (31) is translated with past tense (indefinido) in Spanish, although there is no dedicated past tense marker; the only identifiable grammatical morphemes in (31) are <ca-> ‘2nd-person subject’ and <-he> ‘realis mood’. (31) <Caxihe?> / <Caxihé?> ca-xi-he? 2-go-rea ‘Did you go?’, “Fuisteis?” (M_040) 4This form contains an element <qua> or <quasi> ‘good’, discussed further above. 5Attested in Ms. II/2912 only. 108 15.6 Tense A form which is unmarked for tense, however, does not necessarily imply a past action or event, as shown in (32). (32) <Ynszaxa caxini> ynszaxa where ca-xi 2-go ni q ‘Where do you (sg) go?’, “Dónde vas” (M_059) In some cases, time reference may be expressed by adverbials. For instance, (33), which contains the term for ‘tomorrow’, is translated with future tense in Spanish. (33) <Ninga gozeha> / <Ninga gohezá> ninga I gozeha tomorrow ‘I will go’, “Yo iré” (M_276)6 Note that it is the verb ‘to go’ which is omitted in (33), similar to what happens in (93) in Section 14 above, repeated here as (34). (34) <Quixarani> quixara-ni far-imp ‘Let us go far away!’, ”Vamos lejos” (M_736) We suggest that verb ellipsis is pragmatically licensed in (33) and (34), and that the missing element could originally be inferred from the context. Verb root ellipsis is an infrequent phenomenon in the languages of the world. In South America, it has been observed to occur in Kwaza, a language isolate of Brazil (see Comrie & Zamponi 2019). 6Compare the element <gozeha> ‘tomorrow’, “mañana” in (33) with the single entry <nagosexa> ‘tomorrow’, “mañana” (M_472). An unidentified prefix is also attested in the single entry <canajisexa> ‘yesterday’, “ayer” (M_470), as opposed to <najize> ‘yesterday’, “ayer” (M_108). 109 References Constenla Umaña, Adolfo. 1981. Comparative Chibchan phonology. Philadelphia: University of Pennsylvania. (Doctoral dissertation). Constenla Umaña, Adolfo. 1991. Las lenguas del Área Intermedia: Introducción a su estudio areal. San José: Editorial de la Universidad de Costa Rica. Coronas Urzúa, Gabriela. 1994. Análisis fonológico de la lengua andaquí. Filología y Lingüística 20(1). 69–98. DOI: 10.15517/rfl.v20i1.20230. Coronas Urzúa, Gabriela. 1995. El léxico de la lengua andaquí. Filología y Lingüística 21(1). 79–113. DOI: 10.15517/rfl.v21i1.20265. Díaz Montenegro, Esteban. 2019. El habla nasa (páez) de Munchique: Nuevos acercamientos a su sociolingüística, fonología y sintaxis. Lyon: Université Lumière Lyon 2. (Doctoral dissertation). Dixon, R. M. W. 2010. Basic linguistic theory. Vol. 2: Grammatical topics. Oxford: Oxford University Press. Dixon, R. M. W. & Alexandra Y. Aikhenvald. 1999. Introduction. In R. M. W. Dixon & Alexandra Y. Aikhenvald (eds.), The Amazonian languages, 1–21. Cambridge: Cambridge University Press. Eberhard, David M., Gary F. Simons & Charles D. Fennig (eds.). 2025. Ethnologue: Languages of the world. http://www.ethnologue.com (12 August, 2025). Epps, Patience. 2006. Growing a numeral system: The historical development of numerals in an Amazonian language family. Diachronica 23(2). 259–288. DOI: 10.1075/dia.23.2.03epp. Fernández de Piedrahíta, Lucas. 1881. Historia general de las conquistas del Nuevo Reino de Granada. Bogotá: Imprenta de Medardo Riva. Friede, Juan. 1948a. Historia de los indios andakí del Valle del Suaza. Revista Universidad Nacional de Colombia 13. 109–158. Friede, Juan. 1948b. La investigación histórica y la lingüística americana. In Actes du XXVIIIe Congrès International des Américanistes, 115–124. Paris: Musée de l’Homme. Friede, Juan. 1952. Consideraciones sobre la lengua andaki. Revista Antropología y Etnología 6. 187–222. Friede, Juan. 1953. Los andakí 1538–1947: Historia de la aculturación de una tribu selvática. Mexico City: Fondo de Cultura Económica. Gerdel, Florence L. 2023. Páez dictionary. In Mary Ritchie Key & Bernard Comrie (eds.), The intercontinental dictionary series. Leipzig: Max Planck Institute for Evolutionary Anthropology. https://ids.clld.org/contributions/244 (19 November, 2023). Gil, David. 2013. Genitives, adjectives and relative clauses. In Matthew S. Dryer & Martin Haspelmath (eds.), The world atlas of language structures online 116 References (v2020.4). Zenodo. DOI: 10.5281/zenodo.13950591. http://wals.info/chapter/60 (19 November, 2023). Gómez, Diego F. & Jennifer Torres. 2012–2013a. Transcripción: Vocabulario de la lengua andaquí; anónimo, manuscrito ii/2911 de la Biblioteca del Palacio Real de Madrid. Bogotá: Instituto Colombiano de Antropología e Historia. http : //coleccionmutis.cubun.org/Manuscrito_2911_BPRM_todo (19 November, 2023). Gómez, Diego F. & Jennifer Torres. 2012–2013b. Transcripción: Vocabulario de la lengua andaquí; anónimo, manuscrito ii/2912 de la Biblioteca del Palacio Real de Madrid (Colección Mutis). Bogotá: Instituto Colombiano de Antropología e Historia. http://coleccionmutis.cubun.org/Manuscrito_2912_BPRM_todo (19 November, 2023). Greenberg, Joseph H. 1956. The general classification of Central and South American languages. In Anthony Wallace (ed.), Men and cultures: Selected papers of the 5th International Congress of Anthropological and Ethnological Sciences, 791– 794. Philadelphia: University of Pennsylvania Press. Greenberg, Joseph H. 1987. Language in the Americas. Stanford: Stanford University Press. Hammarström, Harald, Robert Forkel, Martin Haspelmath & Sebastian Bank. 2025. Glottolog (v5.2). Leipzig: Max Planck Institute for Evolutionary Anthropology. DOI: 10.5281/zenodo.15525265. http://glottolog.org (12 August, 2025). Heine, Bernd & Tania Kuteva. 2002. World lexicon of grammaticalization. Cambridge: Cambridge University Press. DOI: 10.1017/CBO9780511613463. Jakway, Martha. 2008. Vocabulario huambisa (Serie Lingüística Peruana 24). Lima: Summer Institute of Linguistics. http://repositorio.cultura.gob.pe/handle/ CULTURA/472 (19 November, 2023). Jespersen, Otto. 1933. Symbolic value of the vowel i. In Linguistica: Selected papers in English, French and German by Otto Jespersen, 283–303. Copenhagen: Levin & Munksgaard. Jolkesky, Marcelo Pinho de Valhery. 2016. Estudo arqueo-ecolinguístico das terras tropicais sul-americanas. Brasília: Universidade de Brasília. (Doctoral dissertation). DOI: 10.26512/2016.02.T.21671. Jung, Ingrid. 2008. Gramática del páez o nasa yuwe: Descripción de una lengua indígena de Colombia (Languages of the World/Materials 469). Munich: LINCOM Europa. Kaufman, Terrence & Brent Berlin. 2007. The native languages of South America. In Christopher Moseley & R. E. Asher (eds.), Atlas of the world’s languages, 2nd edn., 61–93. London: Routledge. 117 References Knuchel, Dominique. 2015. A comparative study of egophoric marking: Investigating its relation to person and epistemic marking in three language families. Stockholm: Stockholms Universitet. (MA thesis). Latham, Robert Gordon. 1862. Elements of comparative philology. London: Walton & Maberly. Loukotka, Čestmír. 1968. Classification of South American Indian languages. Johannes Wilbert (ed.) (Reference Series 7). Los Angeles: UCLA Latin American Center. Margery Peña, Enrique. 1989. Diccionario cabécar–español, español–cabécar. San José: Editorial de la Universidad de Costa Rica. Ortega Ricaurte, Carmen. 1978. Los estudios sobre lenguas indígenas de Colombia. Bogotá: Instituto Caro y Cuervo. Pache, Matthias. 2016. The grammaticalization of plant-part terms in Chibchan languages. International Journal of American Linguistics 82(4). 425–452. DOI: 10.1086/688391. Pache, Matthias. 2018a. Contributions to Chibchan historical linguistics. Leiden: Universiteit Leiden. (Doctoral dissertation). Pache, Matthias. 2018b. Lengua X: An Andean puzzle. International Journal of American Linguistics 84(2). 265–285. DOI: 10.1086/696199. Pache, Matthias. 2024. Tracing sound change in Nasa Yuwe: Evidence from Andaqui and Misumalpan languages. Linguas Indígenas Americanas 24, e024003. 1–44. DOI: 10.20396/liames.v24i00.8673031. Ramos, Demetrio. 1959. El país andaki, tipo de tierra de frontera, visto por López Ruiz en su busca de canela y quina. Revista de Indias 19. 373–419. Rivet, Paul. 1924. La langue andakí. Journal de la Société des Américanistes 16. 99– 110. DOI: 10.3406/jsa.1924.3760. Rosat Pontacti, Adalberto A. 2004. Diccionario enciclopédico quechua–castellano del mundo andino. Cochabamba: Verbo Divino. Ross, Daniel. 2021. A cross-linguistic survey of associated motion and directionals. In Antoine Guillaume & Harold Koch (eds.), Associated motion (Empirical Approaches to Language Typology 64), 31–86. Berlin: De Gruyter Mouton. DOI: 10.1515/9783110692099-002. Rzymski, Christoph, Tiago Tresoldi, Robert Forkel, Heini Arjava, Ingrid Blanco, Timotheus A. Bodt, Sophie Chang, Volker Gast, Russell D. Gray, Simon J. Greenhill, Abbie Hantgan, Carolin Hundt, Nataliia Hübler, Gereon A. Kaiping, Ezequiel Koile, Maria Koptjevskaja-Tamm, Yunfan Lai, Johann-Mattis List, Sergei Monakhov, Natalia Morozova, Steve Pepper, Kristina Pianykh, Mariann Proos, Sallona Ramesh, Nathanael E. Schweikhard, Briana Van Epps & Mei-Shin Wu. 2019. The database of cross-linguistic colexifications, reproducible 118 References analysis of cross-linguistic polysemies. Data set. DOI: 10.1038/s41597-019-0341-x. https://clics.clld.org/ (19 November, 2023). San Roque, Lila, Simeon Floyd & Elisabeth Norcliffe. 2018. Egophoricity: An introduction. In Simeon Floyd, Elisabeth Norcliffe & Lila San Roque (eds.), Egophoricity (Typological Studies in Language 118), 1–78. Amsterdam: John Benjamins. DOI: 10.1075/tsl.118. Seifart, Frank & Harald Hammarström. 2018. Language isolates in South America. In Lyle Campbell (ed.), Language isolates, 260–286. London: Routledge. DOI: 10.4324/9781315750026-10. Swadesh, Morris. 1962. Afinidades de las lenguas amerindias. In Akten des 34. Amerikanistenkongresses, Wien, 18.–25. Juli 1960, 729–738. Horn: Ferdinand Berger. Urban, Matthias. 2019. Notes on Yurumanguí grammar and lexicon. Línguas Indígenas Americanas 19, e019015. 1–25. DOI: 10.20396/liames.v19i0.8656013. Vargas Motta, Gilberto. 1962. Monografía histórica de Acevedo: Apuntes sobre los Andakí alrededor de la historia de San Francisco Javier de la Ceja de los Andakí (Biblioteca de Autores Huilenses 6). Neiva: Imprenta Departamental. Vergara y Vergara, José María & Evaristo Delgado. 1855. Los indios del andaquí: Memorias de un viajero. Popayán: La Matricaria. Waltz, Nathan E. & Alva Wheeler. 1972. Proto Tucanoan. In Esther Matteson, Alva Wheeler, Frances L. Jackson, Nathan E. Waltz & Diana R. Christian (eds.), Comparative studies in Amerindian languages (Janua Linguarum / Series Practica 127), 119–149. Paris: Mouton. Wipio Deicat, Gerardo & Mary Ritchie. 2023. Aguaruna. In Mary Ritchie Key & Bernard Comrie (eds.), The intercontinental dictionary series. Leipzig: Max Planck Institute for Evolutionary Anthropology. https : / / ids . clld . org / contributions/258 (19 November, 2023). 119 Name index Adelaar, Willem F. H., 7, 10, 11, 29, 37, 65 Aikhenvald, Alexandra Y., 23, 82 Albis, Manuel María, 3, 5, 9, 13, 21, 26, 28, 31, 34, 35, 37 Alexander-Bakkerus, Astrid, 29 Berlin, Brent, 10 Buenaventura Amézquita, Juan Guillermo, 3 Campbell, Lyle, 10 Comrie, Bernard, 109 Constenla Umaña, Adolfo, 10 Coronas Urzúa, Gabriela, 9–11, 13, 21–35, 40, 92 Delgado, Evaristo, 9, 13 Díaz Montenegro, Esteban, 43, 50, 60, 70, 97 Dixon, R. M. W., 23, 37, 71 Eberhard, David M., 4, 5, 11 Epps, Patience, 41 Fernández de Piedrahíta, Lucas, 3 Friede, Juan, 3, 5, 9 Gerdel, Florence L., 38 Gil, David, 67 Gómez, Diego F., 8, 13 Greenberg, Joseph H., 10 Hammarström, Harald, 5, 10 Heine, Bernd, 61, 62, 79 Jakway, Martha, 40 Jespersen, Otto, 26 Jolkesky, Marcelo Pinho de Valhery, 10, 11 Jung, Ingrid, 70 Kaufman, Terrence, 10 Knuchel, Dominique, 84 Kuteva, Tania, 61, 62, 79 Latham, Robert Gordon, 9 Loukotka, Čestmír, 10 Margery Peña, Enrique, 35 Muysken, Pieter C., 7, 10, 11, 29, 37 Ortega Ricaurte, Carmen, 7 Pache, Matthias, 11, 33, 38, 41, 97 Ramos, Demetrio, 7 Ritchie, Mary, 40 Rivet, Paul, 9, 10, 40 Rosat Pontacti, Adalberto A., 29 Ross, Daniel, 75 Rzymski, Christoph, 41, 51 San Roque, Lila, 84 Seifart, Frank, 10 Swadesh, Morris, 10 Torres, Jennifer, 8, 13 Name index Urban, Matthias, 13 Vargas Motta, Gilberto, 3, 5 Vergara y Vergara, José María, 9, 13 Waltz, Nathan E., 9 Wheeler, Alva, 9 Wipio Deicat, Gerardo, 40 Zamponi, Raoul, 109 122 Notes on the grammar of Andakí This book presents the first grammatical description of Andakí, an extinct language and presumed isolate once spoken in southern Colombia. Written in an accessible style, this book is valuable to both linguists and scholars of South American indigenous cultures. Although Andakí was documented to some extent in the late 18th and early 19th centuries, its grammar has never been systematically analyzed until now. Drawing on sometimes fragmentary data, all made available in an online database, the book offers the most complete grammatical description of Andakí to date and highlights connections to neighboring languages where relevant. This study provides essential groundwork for future comparative research and contributes to preserving the linguistic heritage of one of the world’s most linguistically diverse regions.