DOCSLIB.ORG

Arxiv:1804.06024V1 [Cs.CL] 17 Apr 2018 [email protected] Tefis W Uhr Otiue Equally

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Arxiv:1804.06024V1 [Cs.CL] 17 Apr 2018 Ivanvladimir@Turing.Iimas.Unam.Mx Tefis W Uhr Otiue Equally Fortification of Neural Morphological Segmentation Models for Polysynthetic Minimal-Resource Languages Katharina Kann∗ Manuel Mager∗ Center for Information and Instituto de Investigaciones en Language Processing Matem´aticas Aplicadas y en Sistemas LMU Munich, Germany Universidad Nacional Aut´onoma de M´exico [email protected] [email protected] Ivan Meza-Ruiz Hinrich Schutze¨ Instituto de Investigaciones en Center for Information and Matem´aticas Aplicadas y en Sistemas Language Processing Universidad Nacional Aut´onoma de M´exico LMU Munich, Germany [email protected] [email protected] Abstract 1 Introduction Morphological segmentation for polysyn- Due to the advent of computing technologies thetic languages is challenging, because to indigenous communities all over the world, a word may consist of many individual natural language processing (NLP) applications morphemes and training data can be ex- for languages with limited computer-readable tremely scarce. Since neural sequence- textual data are getting increasingly important. to-sequence (seq2seq) models define the This contrasts with current research, which fo- state of the art for morphological seg- cuses strongly on approaches which require large mentation in high-resource settings and amounts of training data, e.g., deep neural net- for (mostly) European languages, we first works. Those are not trivially applicable to show that they also obtain competitive per- minimal-resource settings with less than 1, 000 formance for Mexican polysynthetic lan- available training examples. We aim at closing this guages in minimal-resource settings. We gap for morphological surface segmentation, the then propose two novel multi-task train- task of splitting a word into the surface forms of its ing approaches—one with, one without smallest meaning-bearing units, its morphemes. need for external unlabeled resources—, Recovering morphemes provides information and two corresponding data augmentation about unknown words and is thus especially im- arXiv:1804.06024v1 [cs.CL] 17 Apr 2018 methods, improving over the neural base- portant for polysynthetic languages with a high line for all languages. Finally, we explore morpheme-to-word ratio and a consequently large cross-lingual transfer as a third way to for- overall number of words. To illustrate how seg- tify our neural model and show that we can mentation helps understanding unknown multiple- train one single multi-lingual model for morpheme words, consider an example in this pa- related languages while maintaining com- per’s language of writing: even if the word uncon- parable or even improved performance, ditionally did not appear in a given training corpus, thus reducing the amount of parameters by its meaning could still be derived from a combina- close to 75%. We provide our morphologi- tion of its morphs un, condition, al and ly. cal segmentation datasets for Mexicanero, Due to its importance for down-stream tasks Nahuatl, Wixarika and Yorem Nokki for (Creutz et al., 2007; Dyer et al., 2008), segmenta- future research. tion has been tackled in many different ways, con- *The∗ first two authors contributed equally. sidering unsupervised (Creutz and Lagus, 2002), supervised (Ruokolainen et al., 2013) and semi- Mexicanero Nahuatl Wixarika Yorem N. supervised settings (Ruokolainen et al., 2014). frq. m. frq. m. frq. m. frq. m. Here, we add three new questions to this line of re- 136 ni 155 o 327 p+ 102 k search: (i) Are data-hungry neural network models 128 ki 99 ni 230 ne 88 m applicable to segmentation of polysynthetic lan- 114 ti 84 ti 173 p 87 ne guages in minimal-resource settings? (ii) How can 105 u 81 k 169 ti 83 ka the performance of neural networks for surface 70 s 61 tl 167 ka 79 ta segmentation be improved if we have only unla- 44 mo 59 mo 98 u 54 po beled or no external data at hand? (iii) Is cross- 42 ka 55 s 97 ta 50 e’ lingual transfer for this task possible between re- 39 a 52 ki 95 a 36 ye lated languages? The last two questions are cru- 31 nich 48 i 92 pe 36 su cial: While for many languages it is difficult to 31 $i 43 tla 91 e 36 ri obtain the number of annotated examples used in 24 ta 39 ’ke 80 r 34 a earlier work on (semi-)supervised methods, a lim- 24 l 34 nech 74 wa 31 me ited amount might still be obtainable. 22 tahtanili 31 no 69 me 30 wa We experiment on four polysynthetic Mexican 21 no 27 ya 68 ni 30 re languages: Mexicanero, Nahuatl, Wixarika and 17 ya 27 tli 68 ke 27 na Yorem Nokki (details in §2). The datasets we use 17 t 24 x 66 eu 24 wi are, as far as we know, the first computer-readable 17 ke 23 tlanilia 58 ye 24 a datasets annotated for morphological segmenta- 17 ita 23 e 57 ri 23 te tion in those languages. 16 piya 21 tika 52 tsi 20 si Our experiments show that neural seq2seq mod- 15 an 21 n 52 te 16 ’wi els perform on par with or better than other strong Table 1: The most frequent morphs (m.) together baselines for our polysynthetic languages in a with their frequencies (frq.) in our datasets. minimal-resource setting. However, we further propose two novel multi-task approaches and two new data augmentation methods. Combining them 2 Polysynthetic Languages with our neural model yields up to 5.05% abso- lute accuracy or 3.40% F1 improvements over our Polysynthetic languages are morphologically rich strongest baseline. languages which are highly synthetic, i.e., sin- gle words can be composed of many individual Finally, following earlier work on cross- morphemes. In extreme cases, entire sentences lingual knowledge transfer for seq2seq tasks consist of only one single token, whereupon “ev- (Johnson et al., 2017; Kann et al., 2017), we in- ery argument of a predicate must be expressed vestigate training one single model for all lan- by morphology on the word that contains that as- guages, while sharing parameters. The resulting signer” (Baker, 2006). This property makes sur- model performs comparably to or better than the face segmentation of polysynthetic languages at individual models, but requires only roughly as the same time complex and particularly relevant many parameters as one single model. for further linguistic analysis. In this paper, we experiment on four polysyn- Contributions. To sum up, we make the follow- thetic languages of the Yuto-Aztecan family ing contributions: (i) we confirm the applicability (Baker, 1997), with the goal of improving the of neural seq2seq models to morphological seg- performance of neural seq2seq models. The lan- mentation of polysynthetic languages in minimal- guages will be described in the rest of this section. resource settings; (ii) we propose two novel multi-task training approaches and two novel data Mexicanero is a Western Peripheral Nahuatl augmentation methods for neural segmentation variant, spoken in the Mexican state of Durango models; (iii) we investigate the effectiveness of by approximately one thousand people. This di- cross-lingual transfer between related languages; alect is isolated from the rest of the other branches and (iv) we provide morphological segmentation and has a strong process of Spanish stem incorpo- datasets for Mexicanero, Nahuatl, Wixarika and ration, while also having borrowed some suffixes Yorem Nokki. from that language (Vanhove et al., 2012). It is common to see Spanish words mixed with Nahu- Mexicanero Nahuatl Wixarika Yorem N. atl agglutinations. In the following example we train 427 540 665 511 can see an intrasentencial mixing of Spanish (in dev 106 134 176 127 uppercases) and Mexicanero: test 355 449 553 425 total 888 1123 1394 1063 u|ni|ye MALO – I was sick Table 2: Number of examples in the final data splits for all languages. Nahuatl is a large subgroup of the Yuto- Aztecan language family, and, including all of its Yorem Nokki is part of Taracachita subgroup of variants, the most spoken native language in Mex- the Yuto-Aztecan language family. Its Southern ico. Its almost two million native speakers live dialect is spoken by close to forty thousand people mainly in Puebla, Guerrero, Hidalgo, Veracruz, in the Mexican states of Sinaloa and Sonora, while and San Luis Potosi, but also in Oaxaca, Durango, its Northern dialect has about twenty thousand Modelos, Mexico City, Tlaxcala, Michoacan, Na- speakers. In this work, we consider the South- yarit and the State of Mexico. Three dialectical ern dialect. The nominal morphology of Yorem groups are known: Central Nahuatl, Occidental Nokki is rather simple, but, like in the other Yuto- Nahuatl and Oriental Nahuatl. The data collected Aztecan languages, the verb is highly complex. Its for this work belongs to the Oriental branch spo- alphabet consists of 28 characters and contains 8 ken by 70 thousand people in Northern Puebla. different vowels. An example verb is: Like all languages of the Yuto-Aztecan family, ko’kore|ye|ne – I was sick Nahuatl is agglutinative and one word can consist of a combination of many different morphemes. 3 Morphological Segmentation Datasets Usually, the verb functions as the stem and gets extended by morphemes specifying, e.g., subject, To create our datasets, we make use of both seg- patient, object or indirect object. The most com- mentable (i.e., consisting of multiple morphemes) mon syntax sequence for Nahuatl is SOV. An ex- and non-segmentable (i.e., consisting of one single ample word is: morpheme) words described in books of the col- lection Archive of Indigenous Languages in Mexi- o|ne|mo|kokowa|ya – I was sick canero (Canger, 2001), Nahuatl (Lastra de Su´arez, 1980), Wixarika (G´omez and L´opez, 1999), and Wixarika is a language spoken in the states of Yorem Nokki (Freeze, 1989).
Load more

Recommended publications
  • (Huichol) of Tateikita, Jalisco, Mexico
    ETHNO-NATIONALIST POLITICS AND CULTURAL PRESERVATION: EDUCATION AND BORDERED IDENTITIES AMONG THE WIXARITARI (HUICHOL) OF TATEIKITA, JALISCO, MEXICO By BRAD MORRIS BIGLOW A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2001 Copyright 2001 by Brad Morris Biglow Dedicated to the Wixaritari of Tateikita and the Centro Educativo Tatutsi Maxa Kwaxi (CETMK): For teaching me the true meaning of what it is to follow in the footsteps of Tatutsi, and for allowing this teiwari to experience what you call tame tep+xeinuiwari. My heart will forever remain with you. ACKNOWLEDGMENTS I would like to thank my committee members–Dr. John Moore for being ever- supportive of my work with native peoples; Dr. Allan Burns for instilling in me the interest and drive to engage in Latin American anthropology, and helping me to discover the Huichol; Dr. Gerald Murray for our shared interests in language, culture, and education; Dr. Paul Magnarella for guidance and support in human rights activism, law, and intellectual property; and Dr. Robert Sherman for our mutual love of educational philosophy. Without you, this dissertation would be a mere dream. My life in the Sierra has been filled with countless names and memories. I would like to thank all of my “friends and family” at the CETMK, especially Carlos and Ciela, Marina and Ángel, Agustín, Pablo, Feliciano, Everardo, Amalia, Rodolfo, and Armando, for opening your families and lives to me. In addition, I thank my former students, including los chavos (Benjamín, Salvador, Miguel, and Catarino), las chicas (Sofía, Miguelina, Viviana, and Angélica), and los músicos (Guadalupe and Magdaleno).
    [Show full text]
  • The Reorganization of the Huichol Ceremonial Precinct (Tukipa) of Guadalupe Ocotán, Nayarit, México Translation of the Spanish by Eduardo Williams
    FAMSI © 2007: Víctor Manuel Téllez Lozano The Reorganization of the Huichol Ceremonial Precinct (Tukipa) of Guadalupe Ocotán, Nayarit, México Translation of the Spanish by Eduardo Williams Research Year : 2005 Culture : Huichol Chronology : Modern Location : Nayarit, México Site : Guadalupe Ocotán Table of Contents Abstract Resumen Linguistic Note Introduction Architectural Influences The Tukipa District of Xatsitsarie The Revolutionary Period and the Reorganization of the Community The Fragmentation of the Community The Tukipa Precinct of Xatsitsarie Conclusions Acknowledgements Appendix: Ceremonial precincts derived from Xatsitsarie’s Tuki List of Figures Sources Cited Abstract This report summarizes the results of research undertaken in Guadalupe Ocotán, a dependency and agrarian community located in the municipality of La Yesca, Nayarit. This study explores in greater depth the political and ceremonial relations that existed between the ceremonial district of Xatsitsarie and San Andrés Cohamiata , one of three Wixaritari (Huichol) communities in the area of the Chapalagana River, in the northern area of the state of Jalisco ( Figure 1 , shown below). Moreover, it analyzes how the destruction of the Temple ( Tuki ) of Guadalupe Ocotán, together with the modification of the community's territory, determined the collapse of these ceremonial links in the second half of the 20th century. The ceremonial reorganization of this district is analyzed using a diachronic perspective, in which the ethnographic record, which begins with Lumholtz' work in the late 19th century, is contrasted with reports by missionaries and oral history. Similarly, on the basis of ethnographic data and information provided by archaeological studies, this study offers a reinterpretation of certain ethnohistorical sources related to the antecedents of these ceremonial centers.
    [Show full text]
  • Final MA Thesis
    UNIVERSITY OF CALIFORNIA Los Angeles Teachings From Within: Urban Wixárika Women Re-making Motherland A thesis submitted in partial satisfaction of the requirements for the degree Master of Arts in Culture and Performance by Cyndy Margarita Garcia-Weyandt 2015 © Copyright by Cyndy Margarita Garcia-Weyandt 2015 ABSTRACT OF THE THESIS Teachings From Within: Urban Wixárika Women Re-making Motherland by Cyndy Margarita Garcia-Weyandt Master of Arts in Culture and Performance University of California, Los Angeles, 2015 Professor David D. Shorter, Chair The Indigenous Wixárika community, from Western México, commonly known as the “Huichol,” continues to practice ancestral traditions despite the numerous families who have left their motherland to re-establish communities in urban areas in the state of Nayarit, México. As a direct result of these moves, Wixárika families have adopted mestizo ways of living, in regards to education and modes of production. This thesis explores how they implement forms of resistance in urban communities, allowing them to continue practicing Wixárika rituals and ceremonies. Drawing from my ethnographic fieldwork, I illustrate how Wixárika women transform their urban space to cultural knowledge. I include a discussion on how women use physical labor to transmit cultural identity in urban areas in practices that include preparation of corn-based meals and art production such as beading or weaving. In addition, the use of poetics in traditional Wixárika names provides the metaphors to link children and youth to cultural identity. This thesis serves as a case study to illustrate how Indigenous women empower children by providing the skills to continue the practice community-based knowledge.
    [Show full text]
  • Schaefer Revised
    Portraits TOC HUICHOL: BECOMING A GODMOTHER Stacy B. Schaefer MY INTRODUCTION TO SPROUTING CORN Nauxé, á ítsari, éna pu ta, éna pu ta. Look, your loom, put this here, put this here. Mükü pu ta, éna pu ta. Aixru pu áne. Put that there, put this here. That’s good. Pe ti kwa’á? Teté kwaní. Are you hungry? It’s time to eat. hese were the first words spoken to me by a little four-year-old Huichol girl named Sprouting Corn, who was later to become Tmy godchild. Together, as godmother and goddaughter, we would develop an extraordinary relationship that would transcend the traditional role of anthropologist and young native informant. It was midday and we were sitting in the dirt patio of Sprouting Corn’s family rancho in the Huichol Indian Sierra com- munity of San Andres Cohamiata, located on the pine-covered mesa of the Sierra Madre Occidental Mountains in the state of Jalisco. I was just beginning my long-term fieldwork researching backstrap loom weaving and the role of women among the Huichol Indians of Mexico. Sprouting Corn’s mother, Swift Rabbit, a master weaver, had agreed to teach me the ancient art of Huichol weaving. Although I was fluent in Spanish, I knew very little of the Huichol language—a language related to Nahuatl (spoken by Aztecs) and Hopi—which derives from the Uto-Aztecan language family. Thousands of years of separation have made it difficult for the speakers of these related languages to understand one another. So when this little girl, dressed in a bright print gathered skirt and shocking pink blouse with ribbons sewn on, with strands of blue, orange, and yellow beads circling her neck, started chattering to me as if I understood completely what she was saying, I nodded in confusion.
    [Show full text]
  • [.35 **Natural Language Processing Class Here Computational Linguistics See Manual at 006.35 Vs
    006 006 006 DeweyiDecimaliClassification006 006 [.35 **Natural language processing Class here computational linguistics See Manual at 006.35 vs. 410.285 *Use notation 019 from Table 1 as modified at 004.019 400 DeweyiDecimaliClassification 400 400 DeweyiDecimali400Classification Language 400 [400 [400 *‡Language Class here interdisciplinary works on language and literature For literature, see 800; for rhetoric, see 808. For the language of a specific discipline or subject, see the discipline or subject, plus notation 014 from Table 1, e.g., language of science 501.4 (Option A: To give local emphasis or a shorter number to a specific language, class in 410, where full instructions appear (Option B: To give local emphasis or a shorter number to a specific language, place before 420 through use of a letter or other symbol. Full instructions appear under 420–490) 400 DeweyiDecimali400Classification Language 400 SUMMARY [401–409 Standard subdivisions and bilingualism [410 Linguistics [420 English and Old English (Anglo-Saxon) [430 German and related languages [440 French and related Romance languages [450 Italian, Dalmatian, Romanian, Rhaetian, Sardinian, Corsican [460 Spanish, Portuguese, Galician [470 Latin and related Italic languages [480 Classical Greek and related Hellenic languages [490 Other languages 401 DeweyiDecimali401Classification Language 401 [401 *‡Philosophy and theory See Manual at 401 vs. 121.68, 149.94, 410.1 401 DeweyiDecimali401Classification Language 401 [.3 *‡International languages Class here universal languages; general
    [Show full text]
  • Arxiv:1806.04291V1 [Cs.CL] 12 Jun 2018 Hnwrigo Hsfil.Sneidgnu Agae R Di Are Languages Indigenous Since P We field
    Challenges of language technologies for the indigenous languages of the Americas Manuel Mager Ximena Gutierrez-Vasques Instituto de Investigaciones en Matem´aticas GIL IINGEN Aplicadas y en Sistemas Universidad Nacional Universidad Nacional Aut´onoma de M´exico Aut´onoma de M´exico [email protected] [email protected] Gerardo Sierra Ivan Meza GIL IINGEN Instituto de Investigaciones en Matem´aticas Universidad Nacional Aplicadas y en Sistemas Aut´onoma de M´exico Universidad Nacional Aut´onoma de M´exico [email protected] [email protected] Abstract Indigenous languages of the American continent are highly diverse. However, they have received little attention from the technological perspective. In this paper, we review the research, the dig- ital resources and the available NLP systems that focus on these languages. We present the main challenges and research questions that arise when distant languages and low-resource scenarios are faced. We would like to encourage NLP research in linguistically rich and diverse areas like the Americas. Title and Abstract in Nahuatl Masehualtlahtoltecnologias ipan Americatlalli In nepapan Americatlalli imacehualtlahtol, inin tlahtolli ahmo quinpiah miac tlahtoltecnolog´ıas (“tecnolog´ıas del lenguaje”). Ipan inin amatl, tictemoah nochin macehualtlahtoltin intequiuh, nochin recursos digitales ihuan nochin tlahtoltecnolog´ıas in ye mochiuhqueh. Cequintin problemas monextiah ihcuac tlahtolli quinpiah tepitzin recursos kenin amoxtli, niman, ohuic quinchihuaz tecnolog´ıa ihuan ohuic quinchihuaz macehualtlahtolmatiliztli. Cenca importante in ocachi ticchihuilizqueh tlahtoltecnolog´ıas macehualtlahtolli, niman tipalehuilizqueh ahmo mopolozqueh inin tlahtolli. 1 Introduction The American continent is linguistically diverse, it comprises many indigenous languages that are nowa- days spoken from North to South America.
    [Show full text]
  • Sandawe : SOV : Khoisan 3 : Qxu^
    SI1. Word Order in the World's Languages 1 : Hadza : SVO : Khoisan 2 : Sandawe : SOV : Khoisan 3 : Qxu^ : SVO : Khoisan: Northern 4 : ǂAu.//e^ : SVO : Khoisan: Northern 5 : Maligo : SVO : Khoisan: Northern 6 : Ekoka : SVO : Khoisan: Northern 7 : Nama : SOV : Khoisan: Central: Nama 8 : Dama : SOV : Khoisan: Central: Nama 9 : !Ora : SOV : Khoisan: Central: Nama 10 : Xiri : SOV : Khoisan: Central: Nama 11 : Hai.n//um : SOV : Khoisan: Central: Hai.n//um 12 : Kwadi : SOV : Khoisan: Central: Kwadi 13 : G//abake : SOV : Khoisan: Central: Tshu-Khwe: Northeast 14 : Kwee : SOV : Khoisan: Central: Tshu-Khwe: Northeast 15 : Ganade : SOV : Khoisan: Central: Tshu-Khwe: North-Central 16 : Shua : SOV : Khoisan: Central: Tshu-Khwe: North-Central 17 : Danisin : SOV : Khoisan: Central: Tshu-Khwe: North-Central 18 : Deti : SOV : Khoisan: Central: Tshu-Khwe: Central 19 : Kxoe : SOV : Khoisan: Central: Tshu-Khwe: Central 20 : Buka : SOV : Khoisan: Central: Tshu-Khwe: Northwest 21 : Handa : SOV : Khoisan: Central: Tshu-Khwe: Northwest 22 : Xu^ : SOV : Khoisan: Central: Tshu-Khwe: Northwest 23 : G//ana : SOV : Khoisan: Central: Tshu-Khwe: Northwest 24 : G/wi : SOV : Khoisan: Central: Tshu-Khwe: Southwest 25 : Naron : SOV : Khoisan: Central: Tshu-Khwe: Southwest 26 : G//ani : SOV : Khoisan: Central: Tshu-Khwe: Southwest 27 : N/hain.tse : SOV : Khoisan: Central: Tshu-Khwe: Southwest 28 : !O^ : SVO : Khoisan: Southern: Ta'a 29 : ǂHu^ : SVO : Khoisan: Southern: Ta'a 30 : N/amani : SVO : Khoisan: Southern: Ta'a 31 : //Ng.
    [Show full text]
  • Jnls-Mnemonics
    SXjra pp: -- Techset Composition Ltd, Salisbury, U.K. // J. Lat. Amer. Stud. , – © Cambridge University Press doi:./SX 1 2 ‘ ’ 3 Civilising the Savage : State-Building, 4 Education and Huichol Autonomy in 5 6 Revolutionary Mexico, – 7 8 9 Q1 NATHANIEL MORRIS* 10 11 12 13 Abstract. Attempts to use schools to assimilate the Huichols into the Revolutionary nation-state prompted the development of divergent partnerships and conflicts in 14 their patria chica, involving rival Huichol communities and factions, local mestizos, 15 government officials and Cristero rebels. The provocations of teachers, the cupidity 16 of mestizo caciques, rebel violence and Huichol commitment to preserving communal 17 autonomy undermined alliances between Huichol leaders and federal officials, and led 18 to the ultimate failure of the government’s project. If anything, the short-lived Revolutionary education programme equipped a new generation of Huichol leaders 19 with the tools to better resist external assimilatory pressures into the s and 20 beyond. 21 22 Keywords: Huichols, Mexican Revolution, state-building, education, Cristero rebel- 23 lion, assimilation 24 25 Between the accession of Álvaro Obregón to Mexico’s presidency in , and 26 the end of Lázaro Cárdenas’ presidency in , the federal government 27 sought to politically, culturally and economically ‘incorporate’ the country’s 28 Indian peoples into the nation-state, predominantly via the efforts of the maes- 29 tros rurales (rural schoolteachers) of the Secretaría de Educación Pública 30 (Secretariat of Public Education, SEP). The case of the Huichols of northern 31 Jalisco (see Map ) – described in the early s by government officials as ‘an 32 almost savage tribe’, whose members ‘go around naked and subsist on 33 34 Nathaniel Morris, Research Fellow, University of Warwick.
    [Show full text]
  • Arxiv:2104.08726V1 [Cs.CL] 18 Apr 2021
    AmericasNLI: Evaluating Zero-shot Natural Language Understanding of Pretrained Multilingual Models in Truly Low-resource Languages Abteen Ebrahimi} Manuel Mager♠ Arturo Oncevay~ Vishrav Chaudharyr Luis Chiruzzo4 Angela Fanr John OrtegaΩ Ricardo Ramosη Annette Rios Ivan Vladimir] Gustavo A. Gimenez-Lugo´ | Elisabeth Mager] Graham Neubig./ Alexis Palmer} Rolando A. Coto Solanof Ngoc Thang Vu♠ Katharina Kann} ./Carnegie Mellon University fDartmouth College rFacebook AI Research ΩNew York University 4Universidad de la Republica,´ Uruguay ηUniversidad Tecnologica´ de Tlaxcala ]Universidad Nacional Autonoma´ de Mexico´ |Universidade Tecnologica´ Federal do Parana´ }University of Colorado Boulder ~University of Edinburgh ♠University of Stuttgart University of Zurich Abstract Language ISO Family Dev Test Aymara aym Aymaran 743 750 Pretrained multilingual models are able to per- Ashaninka´ cni Arawak 658 750 form cross-lingual transfer in a zero-shot set- Bribri bzd Chibchan 743 750 ting, even for languages unseen during pre- Guarani gn Tupi-Guarani 743 750 training. However, prior work evaluating per- Nahuatl´ nah Uto-Aztecan 376 738 formance on unseen languages has largely Otom´ı oto Oto-Manguean 222 748 Quechua quy Quechuan 743 750 been limited to low-level, syntactic tasks, and Raramuri´ tar Uto-Aztecan 743 750 it remains unclear if zero-shot learning of Shipibo-Konibo shp Panoan 743 750 high-level, semantic tasks is possible for un- Wixarika hch Uto-Aztecan 743 750 seen languages. To explore this question, we present AmericasNLI, an extension of XNLI Table 1: The languages in AmericasNLI, along with (Conneau et al., 2018) to 10 indigenous lan- their ISO codes, language families, and dataset sizes. guages of the Americas.
    [Show full text]
  • The Living Camera in the Ritual Landscape: the Teachers of the Tatuutsi Maxakwaxi School, the Wixárika Ancestors, and the Teiwari Negotiate Videography
    Journal of Ethnology and Folkloristics 11 (1): 39–64 DOI: 10.1515/jef-2017-0004 THE LIVING CAMERA IN THE RITUAL LANDSCAPE: THE TEACHERS OF THE TATUUTSI MAXAKWAXI SCHOOL, THE WIXÁRIKA ANCESTORS, AND THE TEIWARI NEGOTIATE VIDEOGRAPHY LEA KANTONEN Professor of Artistic Research Academy of Fine Arts, University of the Arts Helsinki Elimäenkatu 25 A, Helsinki P.O. Box 10, 00097 Uniarts, Finland Postdoctoral Researcher Foundation for Cultural Policy Research Pitkänsillanranta 3B 00530 Helsinki, Finland e-mail: [email protected] PEKKA KANTONEN Doctoral Student Academy of Fine Arts, University of the Arts Helsinki Elimäenkatu 25 A, Helsinki P.O. Box 10, 00097 Uniarts, Finland e-mail: [email protected] ABSTRACT In this article, we outline the meanings modern Wixárika institutions, such as the school and the museum, may receive as parts of ritual landscape and how the community-based videos shot in the context of these institutions may increase our understanding of ritual landscapes in general. We discuss how ritual landscape can be researched using community-based documentary video art in a way that takes the ontological conceptions of the human and non-human relations of the community seriously. In this case, we understand community-based video art as artistic research in which the work is produced with the community for the com- munity. The making of art, discussed in this article, is a bodily activity as it includes walking with a camera in the Wixárika ritual landscape, interviewing people for the camera, and documenting the work and rituals of the pupils, teachers, and the mara’akate (shaman-priests) planning the community-based museum.
    [Show full text]
  • Linguistic Theory, Collaborative Language Documentation, and the Production of Pedagogical Materials
    Introduction: Collaborative approaches to the challenges of language documentation and conservation edited by Wilson de Lima Silva Katherine J. Riestenberg Language Documentation & Conservation Special Publication No. 20 PUBLISHED AS A SPECIAL PUBLICATION OF LANGUAGE DOCUMENTATION & CONSERVATION LANGUAGE DOCUMENTATION & CONSERVATION Department of Linguistics, UHM Moore Hall 569 1890 East-West Road Honolulu, Hawai'i 96822 USA UNIVERSITY OF HAWAI'I PRESS 2840 Kolowalu Street Honolulu Hawai'i 96822 1888 USA © All texts and images are copyright to the respective authors, 2020 All chapters are licensed under Creative Commons Licenses Attribution-Non-Commercial 4.0 International Cover designed by Katherine J. Riestenberg Library of Congress Cataloging in Publication data ISBN-13: 978-0-9973295-8-2 http://hdl.handle.net/24939 ii Contents Contributors iv 1. Introduction: Collaborative approaches to the challenges of language 1 documentation and conservation Wilson de Lima Silva and Katherine J. Riestenberg 2. Integrating collaboration into the classroom: Connecting community 6 service learning to language documentation training Kathryn Carreau, Melissa Dane, Kat Klassen, Joanne Mitchell, and Christopher Cox 3. Indigenous universities and language reclamation: Lessons in balancing 20 Linguistics, L2 teaching, and language frameworks from Blue Quills University Josh Holden 4. “Data is Nice:” Theoretical and pedagogical implications of an Eastern 38 Cherokee corpus Benjamin Frey 5. The Kawaiwete pedagogical grammar: Linguistic theory, collaborative 54 language documentation, and the production of pedagogical materials Suzi Lima 6. Supporting rich and meaningful interaction in language teaching for 73 revitalization: Lessons from Macuiltianguis Zapotec Katherine J. Riestenberg 7. The Online Terminology Forum for East Cree and Innu: A collaborative 89 approach to multi-format terminology development Laurel Anne Hasler, Marie Odile Junker, Marguerite MacKenzie, Mimie Neacappo, and Delasie Torkornoo 8.
    [Show full text]
  • Word Accent Robindra Nath Banerji [email protected] Haverford College Department of Linguistics
    Huichol (Wixárika) Word Accent Robindra Nath Banerji [email protected] Haverford College Department of Linguistics Abstract Huichol: Context and Sociolinguistic Background Evidence For Word Accent In Huichol The Huichol language (ISO code hch), also known by its autonym, Wixárika, is a language in the Uto-Aztecan In (3), the two words with segmental material /tame/ have different meanings based on the placement of The analysis of systems of tone, stress, and intonation in the world’s languages is a center of debate in family spoken by approximately 45,000 people (INEGI 2010). Huichol is spoken primarily in the Sierra Madre the accent—the phenomenon Iturrioz, Ramírez, and Pacheco (1999) called “lexical stress accent.” This modern phonology. While prototypical cases of stress languages such as English and tone languages such Occidental in the western Mexican states of Jalisco and Nayarit, as well as in parts of Durango and Zacatecas example provides evidence for the presence of differing accent in underlying forms of these two words, as they as Yorúbà are well understood, languages on the peripheries of these prototypes, especially those with (see Figure 1). were produced in isolation of any other linguistic material. It suggests that, at least in a certain subsection of interactions between tone and stress systems, often pose challenges to researchers and theorists. The the lexicon, Huichol uses accent to distinguish between different words with the same segmental material. interaction between stress and tone in the Huichol (Uto-Aztecan) language of western Mexico, can be While most speakers of Huichol live in the Huichol homeland, in modern times many have migrated to urban understood using van der Hulst’s (2014) concept of accent, a lexically underlying mark of prominence.
    [Show full text]