A Comparative Study of Yucatec Maya Sign Languages

Home , Adamorobe Sign Language, As language, Bimodal bilingualism, French manual alphabet, French Sign Language, Mardin Sign Language

Josefina Safar A comparative study of Yucatec Maya Sign Languages

A comparative study of Yucatec Maya Sign Languages Josefina Safar

ISBN 978-91-7911-298-1

Department of Linguistics

Doctoral Thesis in Linguistics at Stockholm University, Sweden 2020

A comparative study of Yucatec Maya Sign Languages Josefina Safar Academic dissertation for the Degree of Doctor of Philosophy in Linguistics at Stockholm University to be publicly defended on Friday 30 October 2020 at 09.00 in hörsal 11, hus F, Universitetsvägen 10 F, digitally via conference (Zoom), public link at department https://www.ling.su.se/

Abstract In my dissertation, I focus on the documentation and comparison of indigenous sign languages in Yucatán, Mexico. I conducted fieldwork in four Yucatec Maya communities with a high incidence of deafness. Because deaf people born into these villages have never had access to an established sign language, they have developed their own local sign languages to communicate with each other and their hearing relatives. Yucatec Maya Sign Languages (YMSLs) are young languages that have emerged over the past decades. The sign languages in the four communities are historically unrelated, but their shared cultural background and the influence of co-speech gestures used by hearing speakers of Yucatec Maya lead to striking similarities in their lexicon and grammar. At the same time, YMSLs display a high degree of variation related to sociolinguistic factors, such as family membership, age, education or language acquisition from deaf adults. In my dissertation, I argue that we can use the phenomenon of variation in young, micro-community sign languages as a window to find out how linguistic conventions are established and which sociolinguistic variables are relevant for shaping sign language structures. My dissertation consists of four sub-studies. In Study I, I employ the framework of translanguaging to examine the semiotic resources used by deaf and hearing Yucatec Maya to interact with each other and with people from other communities. I demonstrate that the repertoire of Yucatec Maya conventional gestures, positive attitudes towards deafness and sign language, as well as shared cultural knowledge, facilitate communication between deaf and hearing people and lead to overlap between sign languages without any historical affiliation. This study constitutes the first application of the translanguaging theory to studies of sign language emergence. Study II investigates cardinal numbers in YMSLs from three villages. I found that some features of Yucatec Maya counting gestures are preserved but that distinct number paradigms evolved in the three YMSL communities. YMSL numerals exhibit variation as a result of linguistic and sociolinguistic factors. Study III explores how YMSL signers convey a linguistic distinction between objects and actions and discusses if these strategies more generally distinguish nouns from verbs. Two possible strategies of the noun-verb distinction were examined; both have an equivalent in hearing people’s gestures but have been integrated into YMSLs in different ways. In Study IV, I focus on a conventional gesture used by hearing Yucatec Maya to specify the height of upright – usually human – referents and analyse how it has been incorporated into YMSLs from four villages. Comparing the form, meaning and distribution of height-specifiers in Yucatec Maya gestures and YMSLs, I demonstrate paths of lexicalisation and grammaticalisation from gesture to sign. Apart from providing documentation of underdescribed, endangered languages, my dissertation makes several theoretical contributions. I demonstrate that language age is not the only variable relevant to the emergence of complex linguistic structures, but that other sociolinguistic factors, such as the distribution of deafness across families, networks of interaction and attitudes toward deafness play a role. Moreover, I present evidence that gestures can enter lexicon and grammar of a sign language simultaneously, challenging previous accounts of lexicalisation and grammaticalisation.

Keywords: Yucatec Maya Sign Language, Yucatec Maya, Mexico, Mesoamerica, shared sign language, village sign language, language emergence, language evolution, sociolinguistic variation, gesture-sign interface, grammaticalisation, lexicalisation, cardinal numbers, size-and-shape specifiers, translanguaging, noun-verb distinction.

Stockholm 2020 http://urn.kb.se/resolve?urn=urn:nbn:se:su:diva-185092

ISBN 978-91-7911-298-1 ISBN 978-91-7911-299-8

Department of Linguistics

Stockholm University, 106 91 Stockholm

A COMPARATIVE STUDY OF YUCATEC MAYA SIGN LANGUAGES

Josefina Safar

A comparative study of Yucatec Maya Sign Languages

Josefina Safar ©Josefina Safar, Stockholm University 2020

ISBN print 978-91-7911-298-1 ISBN PDF 978-91-7911-299-8

Cover photo: Josefina Safar

Printed in Sweden by Universitetsservice US-AB, Stockholm 2020

I opened my notebook and the lesson began. ‘Bhembpungungwa,’ Mowooma said. ‘Bhembpungungwa?’ I repeated. ‘Eh-heh,’ he replied. ‘Write it down.’ He watched carefully as I wrote, a contemplative smile transforming his face, miraculously, into an expression of innocent friendliness. When I finished he asked me to read it aloud. ‘Bhembpungungwa,’ I read. Mowooma smiled with satisfaction. ‘But what does it mean,’ I asked, ‘in French?’ ‘Never mind what it means in French!’ he replied impatiently. ‘You already know French!’

Louis Sarno, Song from the Forest. My Life Among the Ba-Benjellé Pygmies.

Abstract

In my dissertation, I focus on the documentation and comparison of indigenous sign languages in Yucatán, Mexico. I conducted fieldwork in four Yucatec Maya communities with a high incidence of deafness. Because deaf people born into these villages have never had access to an established sign language, they have developed their own local sign languages to communicate with each other and their hearing relatives. Yucatec Maya Sign Languages (YMSLs) are young languages that have emerged over the past decades. The sign languages in the four communities are historically unrelated, but their shared cultural background and the influence of co-speech gestures used by hearing speakers of Yucatec Maya lead to striking similarities in their lexicon and grammar. At the same time, YMSLs display a high degree of variation related to sociolinguistic factors, such as family membership, age, education or language acquisition from deaf adults. In my dissertation, I argue that we can use the phenomenon of variation in young, micro-community sign languages as a window to find out how linguistic conventions are established and which sociolinguistic variables are relevant for shaping sign language structures. My dissertation consists of four sub-studies. In Study I, I employ the framework of translanguaging to examine the semiotic resources used by deaf and hearing Yucatec Maya to interact with each other and with people from other communities. I demonstrate that the repertoire of Yucatec Maya conventional gestures, positive attitudes towards deafness and sign language, as well as shared cultural knowledge, facilitate communication between deaf and hearing people and lead to overlap between sign languages without any historical affiliation. This study constitutes the first application of the translanguaging theory to studies of sign language emergence. Study II investigates cardinal numbers in YMSLs from three villages. I found that some features of Yucatec Maya counting gestures are preserved but that distinct number paradigms evolved in the three YMSL communities. YMSL numerals exhibit variation as a result of linguistic and sociolinguistic factors. Study III explores how YMSL signers convey a linguistic distinction between objects and actions and discusses if these strategies more generally distinguish nouns from verbs. Two possible strategies of the noun-verb distinction were examined; both have an equivalent in hearing people’s gestures but have been integrated into YMSLs in different ways. In Study IV, I focus on a conventional gesture used by hearing Yucatec Maya to specify the height of upright – usually human – referents and analyse how it has been incorporated into YMSLs from four villages. Comparing the form, meaning and distribution of height- specifiers in Yucatec Maya gestures and YMSLs, I demonstrate paths of lexicalisation and grammaticalisation from gesture to sign. Apart from providing documentation of underdescribed, endangered languages, my dissertation makes several theoretical contributions. I demonstrate that language age is not the only variable relevant to the emergence of complex linguistic structures, but that other sociolinguistic factors, such as the distribution of deafness across families, networks of interaction and attitudes toward deafness play a role. Moreover, I present evidence that gestures can enter lexicon and grammar of a sign language simultaneously, challenging previous accounts of lexicalisation and grammaticalisation.

Keywords

Yucatec Maya Sign Language; Yucatec Maya; Mexico; Mesoamerica; shared sign language; village sign language; language emergence; language evolution; sociolinguistic variation; gesture-sign interface; grammaticalisation; lexicalisation; cardinal numbers; size-and-shape specifiers; translanguaging; noun-verb distinction

Table of Contents

List of papers ...... vii

List of figures ...... viii

List of tables ...... ix

List of sign language acronyms and abbreviations ...... x

Acknowledgements ...... xi

1 Introduction ...... 1

2 The emergence of sign languages in Yucatec Maya communities ...... 3 2.1 Introducing village sign languages and emerging sign languages ...... 3 2.2 The grey area: non-prototypical constellations of signing communities ...... 7 2.3 Yucatec Maya Sign Languages: The communities under study ...... 8 2.3.1 Previous studies on Yucatec Maya Sign Languages ...... 9 2.3.2 Deafness in the communities under study ...... 10 2.3.3 Deafness across generations and families ...... 12 2.3.4 Language use in the communities ...... 13 2.3.5 Cultural characteristics ...... 16 2.3.6 Attitudes towards deafness and sign language ...... 18 2.3.7 Language contact with Mexican Sign Language ...... 21 2.3.8 Kopchen: A “control group” for the study of Yucatec Maya gesture ...... 21

3 Motivations for the study ...... 22

4 Research questions ...... 29

5 Fieldwork methodology and data ...... 31 5.1 Fieldwork and data collection ...... 32 5.1.1 Fieldwork methodology ...... 32 5.1.2 Collaboration with the communities ...... 36 5.1.3 Looking back: Challenges and learning curves ...... 40 5.2 Datasets and coding ...... 45

6 The four studies: Summary and reflexion ...... 50 6.1 Translanguaging in Yucatec Maya signing communities ...... 50 6.2 Numeral variation in Yucatec Maya Sign Languages ...... 53 6.3 Strategies of the noun-verb distinction in Yucatec Maya Sign Languages ...... 56 6.4 “When you were that little…” – From Yucatec Maya height-specifier gestures to Yucatec Maya Sign Language person-classifier signs ...... 63

7 One language or several languages? The perpetual question ...... 68 7.1 Embedding the debate in previous literature on YMSL(s) ...... 68 7.2 YMSLs and their sociolinguistic networks ...... 70

7.3 What’s your sign for TORTILLA? Variation in YMSLs ...... 72 7.3.1 Variation between communities ...... 73 7.3.1.1 A continuum of lexical variation ...... 73

v 7.3.1.2 Variation in grammatical structures ...... 79 7.3.2 Variation within communities ...... 81 7.3.2.1 Family membership/interactional groups ...... 82 7.3.2.2 Age of the signers ...... 85 7.3.2.3 Hearing status ...... 87 7.3.2.4 Gender and religious affiliation ...... 88

8 The contribution of YMSLs to studies on sign language emergence and evolution ...... 90 8.1 From Yucatec Maya gestures to YMSL signs ...... 90 8.2 Paths of conventionalisation from gestures to signs ...... 93 8.2.1 Lexicalisation: Gestures become lexical signs ...... 94 8.2.2 Grammaticalisation: Gestures become grammatical markers ...... 96 8.2.3 The relation between lexicalisation and grammaticalisation ...... 100 8.2.4 Sociolinguistic factors that play a role in sign language emergence: Lessons from YMSLs ...... 102

9 Conclusions ...... 107

Appendix A: Coding protocols ...... 109 Coding protocol Study I ...... 109 Coding protocol Study II ...... 113 Coding protocol Study III ...... 117 Coding protocol Study IV ...... 122

Appendix B: List of handshape names ...... 126

Sammanfattning på svenska ...... 127

Bibliography ...... 132

vi List of papers

I. Safar, Josefina. 2019. Translanguaging in Yucatec Maya signing Communities. Applied Linguistics Review 10(1). 31–53.

II. Safar, Josefina, Olivier Le Guen, Geli Collí Collí & Merli Collí Hau. 2018. Numeral variation in Yucatec Maya Sign Languages. Sign Language Studies 18(4). 488–516.

III. Safar, Josefina & Rodrigo Petatillo Chan (in press). Strategies of noun-verb distinction in Yucatec Maya Sign Languages. In Olivier Le Guen, Josefina Safar & Marie Coppola (eds.), Emerging sign languages of the Americas. Berlin: De Gruyter Mouton.

IV. Safar, Josefina (forthcoming). “When you were that little…” – From Yucatec Maya height-specifier gestures to Yucatec Maya Sign Language person- classifier signs. Gesture 18(1).

vii List of figures

Figure 1 Location of the YMSL communities 9 Figure 2 Two deaf men chatting on the street in Chicán 15 Figure 3 Traditional Yucatec Maya solar 16 Figure 4 Yucatec Maya woman in Chicán making embroidery 17 Figure 5 A deaf mason in Chicán (right) and his hearing apprentice (left) 19 Figure 6 Location of Kopchen and the YMSL communities 22 Figure 7 Myself conversing with two deaf YMSL signers 34 Figure 8 Linguistics students Rodrigo Petatillo Chan and Rebecca Petatillo Balam from Kopchen (left of the table) as well as one deaf (Maria Yeseña Kinil) and one hearing YMSL signer (Rossy Kinil Canché) from Nohkop (right of the table) making transcriptions in ELAN 37 Figure 9 Two research assistants (Merli Collí Hau in the pink shirt and Geli Collí Collí in the blue shirt) assisting with an interview of a deaf signer in Chicán (woman in the white huipil) 38 Figure 10 Customised “Controlled Vocabulary” function in ELAN (left) and self-created sign language transcription symbols by Geli (right) 39 Figure 11 Stimulus materials “Give and Take” from the MPI (left) and adapted to the Yucatec Maya context (right) 41 Figure 12 Continuum of lexical variation in YMSLs 73

Figure 13 YMSL variants TORTILLA1 (left) and TORTILLA2 (right) 75

Figure 14 YMSL sign for SEPTEMBER in the interactional groups 1 and 2 in Chicán 84

Figure 15 YMSL sign for DECEMBER in Chicán 84

Figure 16 YMSL signs for FACEBOOK (left) and UPLOAD-TO-INTERNET (right) in Chicán 85

viii List of tables

Table 1 Overview of demographic data of the YMSL communities 11 Table 2 Overview of locations, levels of analysis and research questions treated in each of the four studies 30 Table 3 Data analysed for Study I 46 Table 4 Data analysed for Study II 46 Table 5 Data analysed for Study III/Sub-Study 1 47 Table 6 Data analysed for Study III/Sub-Study 2 48 Table 7 Data analysed for Study IV 48

ix List of sign language acronyms and abbreviations

Sign language acronyms

Acronym Stands for English translation (if applicable) ABSL Al-Sayyid Bedouin Sign Language AdaSL Adamorobe Sign Language ASL American Sign Language Auslan Australian Sign Language DGS Deutsche Gebärdensprache German Sign Language ISL Israeli Sign Language KK Kata Kolok LaSiBo Langue de Signes de Bouakako Bouakako Sign Language LSF Langue de Signes Française French Sign Language LSM Lengua de Señas Mexicana Mexican Sign Language NGT Nederlandse Gebarentaal Sign Language of the Netherlands NSL Nicaraguan Sign Language NZSL New Zealand Sign Language ÖGS Österreichische Gebärdensprache Austrian Sign Language SJQCSL San Juan Quiahije Chatino Sign Language STS Svenskt Teckenspråk Swedish Sign Language TİD Türk İşaret Dili Turkish Sign Language YMSL(s) Yucatec Maya Sign Language(s)

Other abbreviations

Abbreviation Stands for CODA Child of deaf adults L2 Second language MPI Max Planck Institute SASS Size-and-shape specifier

x Acknowledgements

It takes a village to raise a child, and writing this dissertation would not have been possible without the support and the company of numerous people around the world. I want to thank all of you for inspiring and encouraging me and for being a part of this journey. En primer lugar, quiero agradecer a los señantes sordos y oyentes de las Lenguas de Señas Maya Yucatecas y a sus familias por haber trabajado conmigo durante años. Gracias por darme el privilegio de estudiar su(s) lengua(s), por su generosidad, por su paciencia y por su confianza. Sin ustedes, no existiría esta tesis – les estoy muy agradecida por haber compartido su(s) idioma(s) y experiencias de vida conmigo. (En Chicán) A Adán, Lucía, Eulogia y Silveria; a Margarita, Guillermo, Pati, Zulema y Lisandro; a Audomaro y su esposa; a Santos, Silvia, Pedro, Merli y Greti; a José, Neidy, Geli y Cruz; a Ismael y su familia; a Adriana; a Simón; a Don Teodoro (¡qué descanse en paz!) y su familia; a Bernardino, Angélica, Dianela, Mirna, Kandi y Marisela; a Lupita, René, Brígida y su madre y también a Dolores y su esposa. (En Nohkop) A Yeseña, Nayeli, Mili, Rossy, Wicho y Jonny (¡qué descanse en paz!) y también a su abuela, sus esposos y sus hijos. (En Cepeda) A Abigail, Betty y su familia; a Mirna, Lucy y su familia; a Natanael y su familia; a José Luis, Fernando, María y Luis; a Abimael, Josué y su familia, así como a Don Lupercio. (En Trascorral) A todos los miembros sordos y oyentes de la familia. Mi agradecimiento especial va a las familias que compartieron conmigo sus casas y sus comidas: (En Chicán) A Berta, Cesar, Pedro, Ismael, Yoli, Temo, Alba, Gener, Rebeca, Juanita, Hector and Yameli – gracias por su inmensa hospitalidad y su divertidísimo sentido del humor. (En Cepeda) A Neyvy, David, Monserrat y Jair David – gracias por su curiosidad y por abrir las puertas de su casa a una persona completamente desconocida – y a Don Julián por sus canciones y historias. En Kopchen, quiero agradecer a Susana y sus familiares por darme la bienvenida en su familia, así como a Delio, Armando, Fausta y Antonia por sus expresivos y hermosos gestos.

xi Además, estoy agradecida a mis asistentes de investigación: Geli Collí Collí, Merli Collí Hau, Rossy Kinil Canché y Dianela Collí Chim. Gracias por su excelente trabajo, su inteligencia y sus buenas observaciones sobre las Lenguas de Señas Maya Yucatecas. ¡Trabajar con ustedes ha sido un verdadero placer! En México, tuve una red de personas que me apoyaron en numerosos aspectos prácticos relacionados con mi trabajo de campo: Gracias a Ernesto Escobedo Delgado (por ser la primera persona que me animó a ir a Chicán y por muchas conversaciones y consejos útiles), a Debora Castillo (por las aventuras durante la tormenta en su VW Vocho verde), a Rodrigo Petatillo Chan, a Rebeca Petatillo Balam y a Maribel Pacheco Cauich (por la compañía agradable durante el trabajo de campo), a Lorena Pool Balam y su familia (por los muchos conocimientos culturales y por hospedarme en Chemax), a John Lucy y Suzanne Gaskins (por las buenas converaciones), a Rosalie Hübl (por ser mi casa en México), a Tamara Burón (por abrirme su puerta en Mérida – me permitiste tener un descanso durante el trabajo de campo), a Geoff Brennan (for saving my life when I was feeling miserable and sick as a dog) y a Alejandra Alvarez (por tus hermosas fotos). My team of PhD supervisors, Johanna Mesch and Olivier Le Guen, both contributed their complementary expertise and skills in guiding me through the process. Thank you, Johanna, for your open mind and your readiness to engage in a project unrelated to your previous research and for always being available to discuss big and small questions. Every time, your positive and optimistic attitude made me leave your office thinking “it is going to be okay!”. Thank you, Olivier, for introducing me into the YMSL communities and for teaching me a lot about the Yucatec Maya and about fieldwork. Your natural attitude in the field and your collaborative approach to doing research have been an inspiration for me. Thank you for your generosity, your feedback, for hosting me in Mexico City, for heated debates and many good (and bad) jokes. I am especially grateful to my mentor Verena Krausneker, who first sparked my interest in sign language research and ever since continued to be a part of my academic journey. Thank you, Verena, for your idealism and your integrity and for being someone I look up to, both as an academic and as a person. In the process of finalising my dissertation, I have immensely benefitted from the feedback of my “mock” opponent Kate Mesh. Thank you, Kate, for your stimulating questions during my mock defense and for supporting me with numerous issues that went way beyond your mock opponent duties. Your ability to give feedback that is at xii one and the same time meticulous and constructive and your views on sign language emergence and gesture make you a most invaluable colleague. My (present and former) colleagues at Stockholm University provided a fruitful and comfortable environment to learn and do research. Above all, Calle Börstell and Pia Simper-Allen went with me through all the ups and downs. Thank you, Calle, for helping me with so many big and small things when I moved to Sweden, for listening, translating, creating REAL maps, and for empowering me by teaching me how to google properly. It is a pleasure to work with someone who is not only exceptionally intelligent but also has the rare gift of using their intelligence in the most generous and collaborative way. Thank you, Pia, for being my colleague and becoming my friend – this made it both necessary and possible for me to improve my Swedish Sign Language skills. I appreciate your independence, your reliability, your strong opinions about linguistics and everything else, and your true passion for sign language (research). My special thanks go to Sofia Gustafson Capková, director of PhD studies at our department, for the constant support during my somehow nomadic PhD life. It has been a blessing to be a part of the department’s sign language section – a warm thank you to Eira Balkstam, Josefin Bark, Joel Bäckström, Thomas Björkstrand, Nora Duggan, Sebastian Embacher, Anette Ekberg, Sílvia Gabarró-López, Åsa Gustafsson, Moa Gärdenfors, Ingela Holmström, Anders Hultén, Mats Jonsson, Nik Riemer Kankkonen (for the daily dose of Viennese humour in our office), Juli af Klintberg, Magnus Ryttervik, Krister Schönström and Josephine Willing, for being such kind and fun colleagues and for helping me to settle in. Even though we have not directly worked together, I always enjoyed seeing the faces of some former members of the sign language section: Brita Bergman (thank you for your critical eye and for always hitting the nail on the head), Lars Wallin, Anna-Lena Nilsson and Kristina Svartholm. Huge thanks to the STS interpreters who I worked with during numerous presentations and at my mock defense, in particular Thomas Pirander and Cecilia Homann. Of my other colleagues at SU I want to thank in particular: Henrik Bergqvist (for your interesting classes and for being my Mesoamericanist colleague at the department), Johannes Bjerva (for your cold-blooded word-cutting skills), Ahmed Cherif (for a very enjoyable Swedish Sign Language – English tandem), Francesca Di Garbo (for your friendship and your ability to talk about complicated stuff in a totally unpretentious way), Elisabeth Eir Cortes, Pernilla Hallonsten Halling (for your calm

xiii intelligence and your help with my Swedish summary), Henrik Liljegren (for setting a great example of how to do language documentation), Maria Koptjevskaja-Tamm (for your enthusiasm and for opening doors for me), Richard Kowalik (for being my fieldwork buddy when everyone else in the room was talking about EEG waves), Eva Lindström, Guillermo Montero Melis, Bruno Olsson, Franco Pauletto, Geraldine Quartararo (for sticking together during the alienating moments of being a foreigner in Sweden and your help with writing in Spanish), Johan Sjons (for being the most un- Swedish Swedish person I know), Signe Tonér (for your pragmatic way of thinking), Ghazaleh Vafaiean (you are such a cool woman!), Ljuba Veselinova, Bernhard Wälchli (for always thinking ten steps ahead but never showing it) and Hatice Zora (for being the kind, brilliant and humble person you are). A special award should be given to Linda Habermann, Nada Djokic and Cilla Nilsson who are the most efficient and good- spirited admin team ever – without your work, I (or anyone else at the department, for that matter) would not have written a single page. Thanks also to Jeffy Mwakalinga for helping with IT stuff, for fun badminton sessions and for chatting with me in Kiswahili. I cannot neglect to mention Marjatta Sikström for giving me a home in Stockholm that I always love returning to. I am grateful to the group that formed at the Emerging sign languages in the Americas colloquium in Mexico City and during the compilation of the resulting volume. Thank you for facilitating a forum to discuss my work with scholars doing research in contexts that are in so many ways comparable yet so diverse: Benjamin Braithwaite, Marie Coppola, Ivani Fusellier, Sara Goico, John Haviland, Laura Horton, Lina Hou, Olivier Le Guen, Emmanuella Martinod and Kate Mesh. I truly hope that our exchange will continue in the future. A big thank you to Olivier for setting up this collaboration and to Olivier and Marie for pushing the editing and publication process. During my PhD research I have benefitted from the exchange with academic colleagues around the globe, who motivated and supported me over the years through their work as well as through face-to-face and/or email correspondence. Thank you to this team of “invisible supervisors”: Tashi Bradford, Kang-Suk Byun, Richard Cokart, Onno Crasborn (for helping me to formulate my project for the first time and for keeping in touch), Connie de Vos, Mark Dingemanse, Nick Evans, Lindsay Ferrara, Carlo Geraci, Mara Green, Marianne Gullberg (for a great course in Porquerolles), Julie Hochgesang, Jana Hosemann, Annelies Kusters (for your thought-provoking work and precious feedback), Okan Kubus, Ryan Lepic, Ceil Lucas, Claudia Macht and xiv Konstantin Grin (for the many years of being friends and conference buddies), David McKee, Rachel McKee, Irit Meir, Maartje de Meulder, Carla Morris, Victoria Nyst (you were the one who made me realise that linguistic fieldwork is what I want to do and encouraged me to get started – I can’t thank you enough), Gerardo Ortega (for your positive spirit as well as a lot of shared interests and observations), Carol Padden, Nick Palfreyman, Sibaji Panda, Liona Paulus, Pamela Perniss, Roland Pfau, Lauren Reed, Keiko Sagara, Wendy Sandler, Katharina Schalber, Kristin Snoddon, Rose Stamp, Jean-Jacques Tano, Agnes Villwock, Jenny Webster (thanks for being the ideal co- author) and Ulrike Zeshan. I consider myself lucky for having been able to spend six months as a guest researcher at Victoria University in Wellington. This visit was made possible through a scholarship awarded by Kungliga Vitterhetsakademien in Sweden. In New Zealand, I was more than well received by the team at the School of Linguistics and Applied Language Studies, in particular the Deaf Studies Research Unit: Rachel and David McKee, Sara Pivac Alexander and Micky Vale. Hats off to your commitment and the important work you are doing for and with the sign language community in New Zealand. Thank you, Rachel, for helping me to settle in, showing me around, and for being a lovely person to work and spend time with. Thank you, David, for welcoming me at your house and letting me use your beautiful office. Also, I am most thankful to Miriam Meyerhoff (for your straightforward attitude and for your perspective on sign language variation), to Laurie Bauer (for good discussions about blurred boundaries), to Bernie Hambleton (for taking care of all the paperwork) and to John and Paula Hine (for hosting me and for being among the kindest people I have ever met). None of my academic work would have happened without the input of the various talented sign language teachers I have had over the years: Günter Roiss (for ÖGS), Pamela Sundhausen and Simon Kollien (for DGS), Stefan Goldschmidt (for ASL), Okan Kubus (for TİD), Mats Jonsson and Joel Bäckström (for STS) and Sara Pivac Alexander (for NZSL). By introducing me to your languages you have given me new perspectives on the world. For scholarships that enabled me to cover my fieldwork expenses and attend conferences I wish to thank the following foundations: Wallenberg Stiftelse, Gålö Stiftelse and Kungliga Vetenskapsakademien. Thanks to Angela Terrill from Punctilious Editing Sweden for her swift and efficient proofreading. Any remaining mistakes are my own. The team of

xv avhandlingsstöd at SU as well as the US-AB printing office were immensely helpful with the formalities around printing a dissertation during a pandemic. Grant Mc Williams from mallsupport kindly took care of some formatting nightmares. Otto and Ilse Weissinger gave me the possibility to finish my dissertation in the most peaceful and idyllic place one can imagine. During the different stages of my project, my family always had my back and shared my enthusiasm. Thank you to my mother Nina, for your love and your young spirit; Udo, for being a rock for all of us; my grandmother Margareta, for communicating in the best way you can; Maria, for becoming and staying a part of my family; my aunt Susanna, whose beautiful house became my writing retreat more than once; my brother Pauli, for your serenity and your big heart, and for being the little brother I can always rely on; Moni, for your down-to-earth character and your eye for aesthetics; my sister Flora, my soulmate, with whom one could discuss linguistics and philosophy already at age eight; my sister Lotti, for your empathy and emotional intelligence beyond compare and for shining so bright. Finally, a huge thank you to my partner Juri for helping me to find the space to write when the days were packed and there was no time to anuku. Thanks for lending me your ears, your hands, your shoulders – and your English skills. The biggest support comes from the smallest person: Thanks to my daughter Zula, for putting everything into perspective.

xvi 1 Introduction

Su lengua de ellos es muy propia, muy local, muy de acá. (‘Their language is their very own, very local, very much from here.’)

These are the words a man chooses to talk about a language used in his home village, describing it as being intimately connected to the place and the people in the community. But what characterises this particular system of communication and what factors set it apart from other languages? The man was born into one of the four Yucatec Maya villages in Mexico which are the topic of this dissertation. What is special about these communities is that they include an unusually high proportion of deaf people in relation to their overall population size. Because these deaf individuals never had access to a Deaf school or an already existing sign language, they created their own sign languages out of the necessity to communicate with each other in their daily lives. These Yucatec Maya Sign Languages are “their very own” in the sense that their lexicon and grammar are completely unrelated to Mexican Sign Language or any other sign language. They thus offer a window to observe the emergence of linguistic structures without any influence of a previously established system. The man quoted above is hearing, but he grew up with a deaf brother and a deaf sister. He thus learned the local sign language from an early age and is a fluent signer. His situation is not exceptional: In these four villages, most hearing people have one or several deaf relatives, neighbours or friends, and for this reason, the Yucatec Maya Sign Languages are shared by deaf and hearing community members alike. What is interesting is that people from the four communities I studied had never met each other in the past, but their sign languages look strikingly similar and are to a certain extent mutually intelligible. In this dissertation, I will explore why this is the case and demonstrate that these sign languages include both elements that are shared between several Yucatec Maya communities as well as features that are specific to the particular villages. I argue that these “very local” sign languages are inseparable from the socio-cultural context from which they emerge and that the system of co-speech gestures used by hearing Yucatec Maya has a major impact on sign language creation.

1 Another point that caught my attention is that even in these small villages, where people created a sign language together in close interaction, there are a lot of individual differences between signers concerning their vocabulary and grammar. This lexical and structural variation, however, is not totally random. It is a common phenomenon in spoken (see e.g. Labov 1972; Meyerhoff 2006) and signed languages (see e.g. Lucas et al. 2003) that language users have different ways of saying the same thing. Why people express themselves in a certain manner has to do with where they come from, when and how they grew up, which ethnic group they belong to, if and where they went to school, who they are talking to, etc. Throughout my thesis, I will investigate the factors which are relevant for shaping sociolinguistic variation in Yucatec Maya Sign Languages. My dissertation consists of a compilation of four papers. In this document, I will provide an introduction to the field of research as well as a conclusion about how the four studies are interlinked and what they contribute to the wider discipline of sign language and gesture studies. The summary of my thesis is structured as follows: In Section 2, I will introduce previous research on village/emerging sign languages, reflect on the terminology for classifying sign languages and give an overview of the sociolinguistic context of the Yucatec Maya signing communities. In Section 3, I will explain the motivations which drove me to pursue this project and why it is relevant to conduct research on this topic. Section 4 contains a summary of the overall research questions of my dissertation as well as the sub-questions treated in the four studies. Section 5 is dedicated to a rationale of my methods of data collection and analysis and a critical reflection thereof. A summary of the four sub-studies follows in Section 6. In Section 7, I will discuss the results concerning inter- and intracommunity variation and try to answer the question whether the subject of my study should be considered one or several languages. I will then, in Section 8, move on to review the contribution of my results to research on (sign) language emergence and evolution, focusing on paths of development from gestures to signs as well as the influence of various sociolinguistic factors on language creation. Finally, I will wrap up with a conclusion in Section 9.

2 2 The emergence of sign languages in Yucatec Maya communities

2.1 Introducing village sign languages and emerging sign languages

Since the beginning of the discipline in the 1960s, sign language research has for a long time focused on sign languages of larger urban Deaf1 communities, such as American Sign Language (ASL), French Sign Language (LSF) or Swedish Sign Language (STS). Compared to spoken languages, sign languages are generally young, although many national sign languages can be traced back at least a few centuries (MacBurney 2012). “National” or “institutionalised” sign languages developed and became established in institutional contexts such as Deaf schools, Deaf clubs etc. and are often legally recognised in their respective countries (De Meulder et al. 2019). Users of national sign languages are predominantly deaf people whose language is an expression of their shared culture, identity and history (see e.g. Padden & Humphries 1988). Other than deaf people’s family members, interpreters and other professionals working with sign language, relatively few hearing people sign.2 Such a situation imposes major communication barriers between deaf and hearing people and obstructs equal access for deaf people to various domains of everyday life. Apart from a few early studies (Kakumasu 1968; Washabaugh 1979; Kendon 1980) it is only in the past few decades that researchers have developed a growing interest in sign languages that arise in very different scenarios, namely in small rural communities with a high

1 In the literature on sign languages, it has become customary to use the distinction between the lower-case spelling deaf when referring to deafness as the biological condition of hearing loss and the upper-case Deaf when one wishes to characterise a person as a member of a cultural and linguistic minority (Padden & Humphries 1988). However, the d/D-distinction has become a subject to much criticism because the binary dichotomy between deaf people and Deaf people is an oversimplification of the various possible identities of people who cannot hear (see Kusters et al. 2017b). Moreover, it can be ethnocentric when applied outside of the context of Deaf communities in Europe or the US, where the distinction originated (see Woodward & Horejes 2016). Many scholars in the area of Deaf studies or sign language linguistics, but especially in the literature on shared/village sign languages, opt to give up the “capital-D” because the life experiences of deaf people in these environments often do not mirror the ones of Deaf communities in urban settings. Throughout my thesis, I will use the lowercase deaf, unless when referring to established concepts such as Deaf community, Deaf education or Deaf studies (see Kusters et al. 2017b: 15). 2 Although in some countries the number of hearing L2 learners of sign language, e.g. in schools or universities, is constantly on the rise and sometimes even exceeds the total number of deaf signers, for instance in the US, Sweden or the UK (see De Meulder 2018).

3 incidence of (often hereditary) deafness.3 These sign languages, often labelled village sign languages (e.g. Zeshan 2011), indigenous sign languages (e.g. Nonaka 2007), micro- community sign languages (e.g. Nyst et al. 2012) or rural sign languages (e.g. de Vos 2011) develop independently from any institutional context (such as Deaf schools etc.), out of the necessity for deaf and hearing people to communicate with each other in everyday life. Village sign languages are unrelated to the national sign languages of their respective countries and usually not recognised by law. Examples have been identified all over the globe, for instance Adamorobe Sign Language in Ghana (Nyst 2007; Kusters 2015), Al-Sayyid Bedouin Sign Language in Israel (Sandler et al. 2005), Kata Kolok in Bali (de Vos 2012), Yolngu Sign Language in Australia (Bauer 2014), Ban Khor Sign Language in Thailand (Nonaka 2007), San Juan Quiahije Chatino Sign Language in Mexico (Hou 2016; Mesh 2017), Central Taurus Sign Language in Turkey (Ergin 2017) or Miyakubo Sign Language in Japan (Yano & Matsuoka 2018), and new cases keep being discovered. Although the abovementioned language communities are embedded in very diverse geographic and cultural settings, they share a number of characteristic sociolinguistic traits. In many of these small, face-to-face communities, not only deaf people but also hearing people use the local sign language and hearing L2 signers often outnumber deaf signers. For this reason, they have also been called “shared signing communities” (Kisch 2008) and their respective sign languages “shared sign languages” (Nyst 2012). Typically, there is no separate “Deaf community” as in the case of many urban/national sign languages, but deaf and hearing people are in close interaction and there is a low degree of differentiation between deaf and hearing community members in terms of occupation and education (Nonaka 2012). However, despite certain parallels, shared signing communities are by no means homogenous. Kusters (2010) urges not to draw sweeping generalisations concerning the relations between deaf and hearing people and sign language use in the communities. For instance, deaf people’s marriage rate or the extent to which deaf people participate in village activities vary strongly across different communities examined in the literature. Therefore, Kusters (2010) stresses the need for in-depth ethnographic research, including studies conducted by deaf scholars, in order to understand the diversity of shared signing communities. Apart from their sociolinguistic context being radically different from urban, institutionalised sign languages, village sign languages often exhibit unusual linguistic

3 What constitutes a “high incidence of deafness” can differ considerably, as can the overall community size (see Nyst 2012, or de Vos & Zeshan, 2012, for an overview).

4 features. Many of them are linguistic isolates and provide valuable additions to (sign) language typology (de Vos & Zeshan 2012; de Vos & Pfau 2015), which can give us an idea of the true diversity of the world’s languages. Certain structures that have been thought of as universal in the signed modality proved to be non-existent or fundamentally different in some village sign languages, for instance entity classifiers (Nyst 2007) or the use of space for directional verbs (de Vos 2012; Bauer 2014). Some village sign languages also exhibit unique numeral systems (Zeshan et al. 2013; Zeshan & Sagara 2016) or restricted lexicons for specific domains such as colour (Adone et al. 2012; Zeshan & Sagara 2016). Similar to spoken minority languages, many village sign languages are currently endangered and might disappear before they are properly documented (Nonaka 2004; de Vos & Zeshan 2012; Webster & Safar 2019). Some village sign languages are reported to have died out already, e.g. Martha’s Vineyard Sign Language in the US (Groce 1985) or Urubu-Ka’apor Sign Language in Brazil.4 Another closely related, but not synonymous label is “emerging sign languages” (e.g. Meir et al. 2010a; Pyers et al. 2010). These are sign languages that have arisen spontaneously within the past few generations, either inside or outside of institutions. A particularly well- known case is Nicaraguan Sign Language (NSL), which evolved when the first Deaf school in Nicaragua was established, and homesigners from all over the country came together for the first time (e.g. Kegl et al. 1999). They initially had no shared sign language but together came to develop an increasingly complex system over the course of a few consecutive cohorts who entered the school. Linguists and psychologists are interested in studying emerging sign languages for a number of reasons. Unlike in the case of spoken languages, where it is not possible to trace back a language to its very beginnings, emerging sign languages offer a window to observe the evolution of a language “from scratch”. As emerging sign languages develop without the influence of any other established linguistic system, they can give valuable cues about the innate human capacity to learn and create language (e.g. Kegl et al. 1999; Senghas & Coppola 2001; Brentari & Coppola 2012). Observing the evolution of emerging sign languages over time can show which linguistic structures arise early in a language and which form later and how they change over the course of generations or cohorts. These questions have been investigated for various linguistic features, such as phonology (Sandler et al. 2011), the use of

4 https://www.ethnologue.com/language/uks (accessed 28 September, 2020)

5 space (Senghas 2011), referential shift (Kocab et al. 2015) and deixis (Coppola & Senghas 2010). Other researchers emphasise that, as a matter of fact, emerging sign languages do not develop completely from scratch because deaf signers are always surrounded by hearing communities, who use (at least) one spoken language as well as a system of conventional co- speech gestures. They have looked at various ways how hearing people’s gestures can serve as “raw material” for emerging sign languages and become integrated into their linguistic systems (e.g. Nyst 2016a, for Adamorobe Sign Language; Mesh 2017, for San Juan Quiahije Chatino Sign Language; Le Guen 2012; Le Guen et al., in press-a, for Yucatec Maya Sign Language). Another angle for studying emerging sign languages is the question of how the sociolinguistic constellation of a signing community impacts the formation of emerging sign languages. For instance, Hou (2016) shows that San Juan Quiahije Chatino Sign Language emerged within the network of multiple families in co-residence and is shaped by both variation and overlap between the lexicons used in different families. One critical issue related to the term emerging sign language is the question of time- depth: Although several of the abovementioned village sign languages are young and emerged over the course of only a few generations, not all village sign languages are emerging. As de Vos and Nyst (2018) point out, Adamorobe Sign Language (AdaSL) and Kata Kolok (KK) are not young languages but of comparable age as (or possibly even older than) some urban sign languages. In my view, the label emerging sign language has little analytical value: Similar to the (today outdated) term “developing country” it only informs us that a process of change is under way, but not where it is developing from and what its destination is. The term emerging sign language puts the focus on a specific scenario of language evolution (which, as argued above, can also happen under very diverse circumstances) and does not tell us much about the particular language “under construction”. The question of when a language stops being “emerging”, and at which point its evolution can be considered a regular instance of language change is still unresolved (de Vos & Nyst 2018: 482). In the following section, I will critically discuss why labels such as “village sign language”, “homesign” or “emerging sign language” do not encompass the whole range of situations in which sign languages evolve – including Yucatec Maya Sign Languages (YMSLs). In Study I (see Section 6.1), I reflect in more depth on the question of which term is most suitable to describe Yucatec Maya Sign Languages and how the framework of translanguaging can be used to critically deconstruct existing typologies of sign languages. Throughout my dissertation, I will use the labels “village sign language”, “shared sign

6 language”, “micro-community sign language”, “indigenous sign language” and “emerging sign language”. These terms are related but not interchangeable as each of them covers a specific aspect of YMSLs’ language ecology/ecologies.

2.2 The grey area: non-prototypical constellations of signing communities

In order to trace the path of development of emerging sign languages some researchers (see Goldin-Meadow 2005; Brentari & Coppola 2012) have suggested a continuum with ‘homesign’ or ‘homesign systems’ on one end and ‘mature/fully-formed sign languages’ on the other end. According to their model, sign languages can be placed on this continuum based on criteria such as linguistic complexity, degree of conventionalisation, intergenerational transmission and network density between deaf signers. However, Nyst (2012) points out that a number of sign languages – for instance, shared sign languages in rural settings, but also sign languages of urban Deaf communities without any access to formal education – do not neatly fit into any of these categories. Nyst rejects the idea of such a continuum because it leaves some sign languages “trapped somewhere in the middle of a developmental cline” (Nyst 2012: 567). Instead, she suggests that all sign languages must be described in their own terms, taking into account their specific sociolinguistic ecology, including factors such as contact between signers, acquisition patterns, language attitudes, the role of hearing signers, the influence of conventional gestures, etc. (Nyst 2012; Nyst et al. 2012). Criteria such as linguistic complexity cannot be taken as a base for judging how developed a language is, as languages naturally exhibit varying levels of complexity in different linguistic domains. A growing number of publications have since reported scenarios that lie somewhere in the “grey area” (Nyst 2012: 568) between homesign and sign languages of larger (either rural or urban) deaf communities, e.g. communities with loose contact between individual deaf people or with limited deaf-to-deaf intergenerational transmission. A variety of new labels has been introduced to describe these non-prototypical situations, such as “communal home sign” (Zeshan 2011), “rural home sign” (Nyst et al. 2012), “shared home sign” (Horton, in press-a), “multi-generation homesign” (Neveu 2019), “family sign languages” (Nyst et al. 2012), “nucleated network sign languages” (Reed 2019), or “local sign” (Green 2014). The “grey area” of sign languages that do not fit into any box seems to be larger than initially thought – perhaps even larger than the prototypical categories themselves. When documenting a sign language, a thorough ethnographic description of its linguistic ecology and the social networks between signers is indispensable. Many more studies about rural sign languages that emerge in

7 different kinds of constellations are needed to be able to create a comparative framework that would eventually lead to a linguistic typology of emerging/ village/ rural/ shared/ indigenous/ micro-community sign languages. As I argue throughout my dissertation, the analysis of YMSLs adds another piece to the investigation of sign languages in the “grey area”. It proves that categories like “homesign”, “emerging sign language”, “shared sign language”, “village sign language”, etc. each capture certain characteristics of the sociolinguistic landscape in which YMSLs are used – but they are not exhaustive nor are they always mutually exclusive. For this reason, critically scrutinising sign language taxonomies should be part of the thorough description of a sign language. YMSLs span several villages with uneven numbers of deaf families and generations. Their linguistic structures fluctuate between variability, convergence and (re-)negotiation, which poses a challenge to current classifications of sign languages. I will return to the issues of terminology and taxonomy in the Sections 3 and 6.1.

2.3 Yucatec Maya Sign Languages: The communities under study

In the following section, I will give an overview of the Yucatec Maya Sign Language communities, their demographic composition, the use of spoken and signed languages in the communities, cultural characteristics and language attitudes. The information in this section is adapted from the sociolinguistic sketch of the YMSL communities in Safar and Le Guen (in press). Yucatec Maya Sign Languages (YMSLs) are indigenous sign languages used by deaf and hearing signers in Yucatec Maya communities with a high incidence of deafness in the peninsula of Yucatán, Mexico. Deaf people born into these rural communities did not have access to Deaf schools, other institutions or an urban Deaf community, where they could learn an already established sign language. Needing a visually accessible medium to communicate with each other and with their hearing family members, they developed their own local sign languages, which are completely unrelated to the national Mexican Sign Language (Lengua de Señas Mexicana; LSM). To this date, four communities with multiple deaf members have been identified in the state of Yucatán: Chicán, Nohkop, Trascorral and Cepeda Peraza (Cepeda for short). The map in Figure 1 shows the location of the YMSL communities on the peninsula of Yucatán.

Figure 1: Location of the YMSL communities5

The team of the YMSL documentation project6 found several other villages across the peninsula with one or two deaf individuals, but it is not unlikely that further larger, yet unknown signing communities exist in the region. The YMSL documentation project, of which I am a part, is a team of linguists and anthropologists as well as deaf and hearing signers from the YMSL communities, who work together on the documentation of Yucatec Maya Sign Languages. The YMSL project team and our methods of language documentation will be introduced in more detail in Section 5.1.2.

2.3.1 Previous studies on Yucatec Maya Sign Languages

The existence of a local sign language in Yucatán was first mentioned in the 1970s by the US anthropologist Hubert Smith, who filmed a documentary in Chicán about social structures among the Yucatec Maya. He stumbled upon the presence of a high proportion of deaf people in the village and realised that hearing community members were able to sign as well. This aroused the interest of linguists and anthropologists. Johnson (1991) noted that the sociolinguistic setup of Chicán is radically different from urban Deaf communities (e.g. in the

5 Thanks to Calle Börstell for creating this map. 6 http://ymslproject.org (accessed 28 September, 2020); see Section 5.1.3.

9 US) because due to the widespread use of sign language in the village “social and economic benefits are more readily accessible to deaf people and […] the formation of a strong ethnic group and the politicization of deafness are unnecessary” (Johnson 1991: 471). Shuman (1980) and Shuman and Cherry-Shuman (1981) provided a first description of the sign language’s lexicon and selected grammatical features, although their analysis presented several methodological flaws and remained, as the authors themselves acknowledge, “preliminary”. For instance, Shuman translated a written word list from English to Spanish to Yucatec Maya and collected data mainly from one hearing signer (father of three deaf children), who provided the translations into YMSL. An elicitation method based on spoken/written language and the lack of deaf informants raises concerns about the reliability of Shuman’s data (see Safar, forthcoming-a). In later years, a number of further studies were carried out in Chicán, looking at possible relations of the local sign language to other sign languages in Mesoamerica (Fox Tree 2009), aspects of communication and identity among deaf people (MacDougall 2012), the sociolinguistic profile of the community (Escobedo Delgado 2012) as well as selected linguistic features, such as time expressions (Le Guen 2012) and numbers (Zeshan et al. 2013). Olivier Le Guen and his project team later identified the existence of another Yucatec Maya signing community in Nohkop (Safar et al. 2018; Safar & Petatillo Chan, in press; Le Guen et al., in press-a; Safar, forthcoming-a; Safar, forthcoming-b). The studies of this dissertation are the first ones to investigate the sign languages of Cepeda Peraza and Trascorral. Authors of previous studies have alternatively referred to YMSL as Chicán Sign Language/Lengua de Señas Chicana (Escobedo Delgado 2012; Zeshan et al. 2013), Nohya Sign Language (Shuman 1980) or Meemul Tziij (Fox Tree 2009). I will provide an in-depth discussion of the choice of terminology and its theoretical implications in Section 7.

2.3.2 Deafness in the communities under study

The four communities under study differ from each other in their overall population size as well as their number and distribution of deaf people. Chicán is a village of 720 inhabitants, including 16 deaf people who are between 16 and 67 years old. The oldest deaf man in Chicán was in his early eighties when he passed away in January 2020. In Nohkop, there is a family of originally five siblings, four of whom are deaf and between 17 and 24 years old. In July 2020, the hearing boy, who was the youngest sibling of the family, tragically passed away due to an illness at the age of 16. Trascorral is home to a family of 13 siblings, six of

10 whom are deaf and between 9 and 27 years old. In Cepeda Peraza, there are ten deaf community members from different families who are between 28 and 47 years of age. The following table gives an overview of demographic data of the YMSL communities; additional information about the family distribution of deaf people and the generations of deaf people in each community is presented in Section 2.3.3.

Table 1: Overview of demographic data of the YMSL communities

Chicán Nohkop7 Trascorral Cepeda Peraza Number of 720 (Escobedo No exact figure ~300 ~700 inhabitants Delgado 2012) (around 30) Number of deaf 16 4 6 10 people Percentage of ~2.2% No exact figure ~2% ~1.4% deaf people Age of deaf 16-67 17-24 9-27 28-47 people in 2020 (approximately) Gender 8 female, 3 female, 2 female, 4 female, distribution of 8 male 1 male 4 male 6 male deaf people Family Multiple Siblings of one Siblings of one Multiple families distribution of families family (family family (family deaf people of 5 siblings) of 13 siblings) Generations of 3 1 1 1 deaf people Interactional 6 1 1 6 groups including deaf people

Geneticists have investigated the high incidence of deafness in Chicán and Cepeda Peraza and concluded that deafness in these villages is likely to be caused by a recessive gene. Many people in these two communities are kin-related to each other and most deaf people carry the same surname (Collí in Chicán, Chi or Ek’ in Cepeda). The origin of deafness in Nohkop and Trascorral is unknown, but it is likely to have other, non-genetic causes. However, we have not investigated this question in further detail and the biological causes of deafness are not part of this study. In Chicán, government authorities and NGOs carried out audiometry and

7 Nohkop is a pseudonym for a small neighbourhood in the outskirts of the town Chemax.

11 distributed hearing aids to deaf inhabitants. Deaf people, however, do not use them and rejected these interventions, explaining that they did not correspond to their needs (Dikyuva et al. 2012: 319; MacDougall 2012; Safar 2015).

2.3.3 Deafness across generations and families

All YMSLs are young languages, with a maximum depth of three generations in Chicán. The oldest signer in Chicán, Don Teodoro, a man who recently died in his eighties, is reported to be the first deaf person in the village. In extensive interviews, Don Teodoro as well as other elders in the village asserted that they could not recall the presence of any other deaf people in Chicán at the time Don Teodoro was born, neither had they heard of the existence of previous generations of deaf people in the village. Only the subsequent generation, whose members are today between 45 and 67 years old, constituted the first larger group of deaf signers. The deaf signers of the youngest generation are today between 16 and 32 years old (see Generations of deaf people; Table 1). As Kisch (2012) emphasises, it is important to acknowledge that the definition of “generations” in shared signing communities cannot be based on age-criteria alone, but additional factors, such as communicative networks, patterns of socialisation, educational background and exposure to other languages need to be taken into account. In order to achieve a more meaningful grouping of deaf people in Chicán, Le Guen (2012) divided them into seven “interactional groups”. These are formed by members of nuclear or extended families that share the same household. Most interactional groups include several hearing members (siblings, spouses, hearing Children of Deaf Adults (CODAs)), but two of them are made up entirely of deaf members. Most deaf people in Chicán have hearing parents, but two deaf siblings (16 and 20 years old in 2020) were born to deaf parents and grew up in a deaf-only household. As the oldest deaf man, who, together with his family, had comprised a separate interactional group, recently passed away, there are six interactional groups with deaf signers left in Chicán (see Interactional groups including deaf people; Table 1). In Nohkop and Trascorral, deaf YMSL signers belong to one family each and constitute the first generation of users of their sign language (see Family distribution of deaf people and Generations of deaf people; Table 1). The five siblings in Nohkop (three deaf girls, one deaf boy, one hearing boy) grew up with their grandmother after their mother’s death. The Nohkop signing community includes around 30 additional hearing signers in the extended family and

12 neighbourhood (cousins, peers, neighbours etc.). In recent years, the signing community of Nohkop dissolved, as the three deaf sisters got married. One of them stayed in her home village but moved to the house of her husband’s family. Two of the deaf sisters married two brothers and continue living together in the same house in their husbands’ village. The husbands are all hearing and learned YMSL in interaction with their wives. All three sisters had hearing children who are acquiring YMSL as their first language. In Cepeda Peraza, the ten deaf signers are distributed across six families. There are four pairs of deaf siblings and two deaf young men who are the only deaf members of their families. Currently, there are six interactional groups with deaf signers in Cepeda. Unlike in Chicán, none of them is composed entirely of deaf people, but all of the interactional groups include hearing family members, too (see Family distribution of deaf people and Interactional groups including deaf people; Table 1). Members of the four communities assert that they have not been in contact in the past. Their sign languages emerged independently from each other. Note that even though the villages are part of the same region, they are located at one to several hours drive from each other. In rural Yucatán, transport options are relatively limited, people travel little and contact between villages is scarce if people are not kin-related. In-depth interviews, including with the eldest community members, failed to provide evidence for any historical link between the sign languages from the four communities. The eldest signers in each community reported that when they were little, there were no older deaf people around from whom they could learn sign language but that they were the ones who created the sign languages in interaction with their (deaf and hearing) family members. More recently, within the framework of the YMSL documentation project, some of the deaf and hearing research assistants visited other YMSL communities than the one they were born in. These visits provided valuable opportunities to gauge the degree of mutual intelligibility and study phenomena such as linguistic adaptation and translanguaging. A detailed analysis of such an encounter is included in Study I.

2.3.4 Language use in the communities

The main spoken language in the communities under study is Yucatec Maya, which is part of the Mayan language family. Yucatec Maya is an official national language of Mexico and with around 800,000 speakers (INEGI 2010) it is one of the most widely spoken indigenous languages in the country. With the growing influence of Spanish, hearing community members are becoming increasingly bilingual in Yucatec Maya and Spanish (or trilingual, including

13 YMSL). Although Yucatec Maya remains the primary language of socialisation and interaction within the families, Spanish is the dominant language in public domains, most crucially education. Children and teenagers learn Spanish in school and can sometimes be observed speaking Spanish to each other. Ethnographic research during my fieldwork showed certain differences between the communities regarding their degree of bilingualism: In Cepeda Peraza, parents speak Spanish to their children more frequently than in Nohkop, Chicán or Trascorral, where they rely almost exclusively on Yucatec Maya. Yucatec Maya is a language that is used primarily in face-to-face settings and characterised by an oral tradition. Before the Spanish conquest, the Maya civilisation had its own script, which consisted of a set of glyphs, but it was never widespread and remained reserved for an elite (Hanks 2010). After the colonialisation, the Latin script was introduced, but the majority of the population has not been using it for writing Yucatec Maya. Even though language conservation activists have tried to establish a written form of Yucatec Maya, its use stays restricted to specific domains (e.g. universities, academic discourse, translation of official documents such as the Mexican constitution, etc.). Many Yucatec Maya, especially from the generation over 45, are non-literate, and if they can read and write, they use Spanish for written communication. YMSLs are not signed versions of spoken Yucatec Maya but have independent lexicons and grammars. However, spoken Yucatec Maya and YMSLs are in active contact and specifically Yucatec Maya co-speech gestures have a critical impact on the emergence of YMSLs, as I will demonstrate in Studies I, II, III and IV. For deaf people in the Yucatec Maya communities, YMSLs is the primary means of communication. During my fieldwork, I was able to witness the use of sign language in all domains of village life, such as family matters, domestic issues, work, farming, economy, celebrations, weddings, funerals, gossip, mythology, trade, discussing politics, etc. In Section 5.1.1, I describe my approach to the ethnographic study of sign language use in the YMSL communities.

Figure 2: Two deaf men chatting on the street in Chicán (Photo: Alejandra Álvarez)

Deaf children are socialised in YMSL and acquire sign language in interaction with deaf family members or peers. As most hearing community members have deaf relatives, friends or neighbours, the use of the sign language is not restricted to deaf people but shared between deaf and hearing signers. As in other shared signing communities, hearing L2 signers constitute the majority of signers. Hearing people’s level of signing proficiency varies according to their proximity in kinship and their frequency of interaction with deaf people. The most competent hearing signers are CODAs, younger siblings, cousins or peers of deaf signers. However, even hearing villagers who do not have immediate deaf family members are able to communicate with deaf people using YMSL. In cases where a person from outside visits a YMSL community (e.g. people who come to trade in the village), they rely heavily on the repertoire of Yucatec Maya conventional gestures when they communicate with a deaf person. Often, hearing relatives act as “interpreters”, but these interpreting settings are spontaneous, and planned continuous and formal sign language interpretation does not exist (Pacheco, forthcoming). In Study I (Section 6.1), I reflect on encounters of deaf signers with people from outside their community and on the role of shared context and language attitudes on mutual understanding.

15 2.3.5 Cultural characteristics

The communities under study share numerous cultural and sociolinguistic traits. Among the Yucatec Maya, the extended family typically lives together in several houses on a shared family compound called a solar.

Figure 3: Traditional Yucatec Maya solar

It is a traditionally patrilocal society, which means that when a couple gets married, the woman moves to the compound of her husband’s family. Children grow up in the network of the extended family, with not only the parents as their main caregivers but also surrounded by their grandparents, aunts, uncles and cousins. Today, we can witness a tendency towards neolocalisation: after getting married, a couple sets up a new, independent household. This leads to changes in traditional patterns of family networks and socialisation. The annual cycle in Yucatec Maya communities is structured by the maize harvest and a sequence of religious rituals and ceremonies – traditional Maya as well as Christian ones. As all over Mexico, both Catholic and Protestant religions have been adopted in Yucatec Maya communities. In Chicán, the population is divided into a Catholic and a Protestant part (see Escobedo Delgado 2012: 378). Cepeda Peraza is a predominantly Protestant village. The

16 signing community in Trascorral is Protestant, and the one in Nohkop Catholic, except for the husband of one deaf woman. Traditionally, Yucatec Maya villages are agricultural communities whose members live off subsistence farming. Yucatec Maya men practice a particular form of slash-and-burn agriculture to grow corn (the staple food in Yucatán and large parts of Mesoamerica) and other crops (e.g. pumpkins, beans) on their milpa (agricultural field), go hunting and keep bees. The women traditionally take care of the house, grind maize at the mill and prepare food, look after the children and domestic animals (chickens, turkeys, pigs). Additionally, men and women also produce and sell artefacts such as hammocks and embroidery.

Figure 4: Yucatec Maya woman in Chicán making embroidery (Photo: Olivier Le Guen)

In recent years, farming is becoming an increasingly precarious activity due to frequent droughts and other natural disasters and community members are often forced to buy corn at exorbitant prices. Hunting also proves to be more and more difficult as less game lives in the area. As a result, new forms of income have emerged, and many people commute or permanently move to nearby cities (commonly to Mérida from Chicán, Tek’ax from Cepeda, Cancún or Tulum from Trascorral) to seek employment. Commonly, men work in the construction business and women are employed as housekeepers and babysitters by wealthier

17 families. In Cepeda, several hearing men have migrated to the US over the last decades to make a living there and regularly send back money to their families. Up to now, no deaf community member has left for the US. Comparable to other shared signing communities (see e.g. Nonaka 2012), deaf and hearing people had similar professional opportunities within traditional Yucatec Maya community structures and deaf individuals did not experience a disadvantage in terms of work and economic benefits (Johnson 1991). Today, as a result of radical demographic and economic changes due to globalisation and climate change, deaf people are facing a different situation. Even though several deaf men and women took on jobs outside their home villages, a change in location imposes new challenges and barriers in terms of communication and equal access.

2.3.6 Attitudes towards deafness and sign language

To refer to deaf people, hearing speakers of Yucatec Maya use the words kóok (‘deaf’), toot (‘mute’) or, more commonly, the Spanish loan (sordo-)muudo. The latter term is considered pejorative by Deaf communities in Spanish-speaking countries but does not carry any negative connotations in the Yucatec Mayan context. A common paraphrase for ‘deaf’ in Yucatec Maya is mina’an ut’aan ‘(those who) do not have words/voice’. When referring to sign language, hearing people normally use the Spanish loan seenyas ‘signs’. In their reflections about manual-visual communication, deaf and hearing people do not seem to make a conceptual distinction between deaf people’s signs and hearing people’s gestures (see also Kusters & Sahasrabudhe 2018; Mesh & Hou 2018) – although deaf signers claim that the accessibility of information through co-speech gestures alone remains limited compared to

YMSLs (see Study I). Deaf people refer to themselves as HEAR-NEG/(SPEAK-NEG)8 ‘does not hear/(does not speak)’. The action of signing is referred to as SIGNS.9 The Yucatec Maya generally express positive attitudes towards deafness. As several authors have observed (Le Guen 2012; Johnson 1991; Escobedo Delgado 2012; MacDougall 2012), deafness is not considered problematic but “a normal phenomenon” (Escobedo Delgado 2012: 378) and deaf people are not discriminated against in the Yucatec Maya context. This

8 Sign language glosses are written in small caps. NEG stands for negation marker. 9 See the links http://ymslproject.org/VIDEOS_datos/se%C3%B1as.mp4 for the YMSL sign SEÑAS (‘signs’) and http://ymslproject.org/VIDEOS_Datos_CH/SOORDO_converted.mp4 for the YMSL sign SORDO (‘deaf’) (accessed 28 September, 2020).

18 ideology is not restricted to deafness but embedded in a broader, cultural understanding of difference and diversity. The Yucatec Maya consider that children are born just like God created them and they should be accepted the way they are. In Cepeda, where Protestantism is prevalent, I encountered similar positive attitudes as in the other communities but slightly different explanations, which are more influenced by Christian values of benevolence and tolerance. Community members acknowledge that deaf persons are able to understand everything and achieve the same things as hearing individuals – provided the fact that they can communicate in sign language with people in their environment. Deaf people fulfil similar social roles as hearing people: they marry deaf or hearing partners, have children, pursue jobs and perform tasks within the community. They are regarded as intelligent, competent members of society and are appreciated for their individual skills. A young deaf man in Cepeda, for instance, is regularly contacted by the municipality of Tek’ax to take visitors to some nearby caves because locals agree that he can guide the way better than anyone else in the village.

Figure 5: A deaf mason in Chicán (right) and his hearing apprentice (left) (Photo: Alejandra Álvarez)

19 As some authors note (Johnson 1991; Kusters 2010; Escobedo Delgado 2012), deaf people’s marriage rate in Chicán is somewhat lower than among the general population, but people are reluctant to provide an explanation for this situation. In Cepeda, only one deaf man is currently married, but several others have boy- or girlfriends. In Nohkop, three deaf young women are married to hearing spouses; their younger deaf brother lives with their grandmother. Positive evaluations of deafness are strongly interlaced with the perception of sign language as an effective, pleasant and effortless means of communication. Deaf and hearing community members state that YMSL is a fully-fledged communication system and it is possible to express everything in sign language, including complex topics. In Chicán and Cepeda, Jehovah’s witnesses have attempted to teach villagers LSM, but deaf people stated that they did not feel the necessity to learn another sign language, as their own signs were perfectly suited to their environment and sufficient to cover their communicative needs. The signer quoted in the introduction describes this as a sign language that is “very local, very much from here”. LSM, in contrast, is perceived as a “time-consuming way of communication” by deaf and hearing signers and strongly associated with the use of a manual alphabet (see Study I). A thorough discussion of language attitudes and their importance for communication practices in Yucatec Maya signing communities is included in Study I of my dissertation. Despite positive attitudes towards deafness and widespread sign language use, the most crucial difference between deaf and hearing Yucatec Maya concerns their opportunities in terms of education. In the communities under study, education in sign language is not available and classes are taught only in Spanish. Deaf pupils sometimes rely on their hearing classmates as tutors but, as a matter of fact, mostly sit through their school career and copy their peers’ homework without being able to actively participate. Apart from two young deaf people in Chicán and two in Cepeda, who finished secondary school, deaf YMSL signers received only basic schooling. Two of the younger deaf signers in Chicán attended a special education centre in Tek’ax once a week for a few years, but this school is designed for children with various types of disabilities and the teacher knows only some basic lexical LSM signs and the manual alphabet. Due to the barriers in the educational system deaf community members do not (or only to a minimal extent) learn to read Spanish, which becomes an increasingly important requirement in the labour market. While formal education and literacy is generally low among the older generation (45 and older) in rural Yucatán, the imbalance in professional opportunities is growing between deaf and hearing Yucatec Maya of the younger generation (see also Nonaka 2012, for Ban Khor Sign Language). So far, no deaf person from the YMSL

20 communities has been able to enter tertiary education or enrol in any professional training program.

2.3.7 Language contact with Mexican Sign Language

Deaf and hearing YMSL signers all agree that YMSL and Mexican Sign Language (Lengua de Señas Mexicana; LSM) are mutually unintelligible languages. To date, there is relatively little language contact between YMSLs and LSM. Some younger signers know the manual alphabet of LSM (YMSLs do not have a manual alphabet) and occasionally use LSM signs that are not, or only partially, encoded in the YMSL lexicon, such as weekdays or colours. Only two younger deaf signers in Chicán are bilingual in YMSL and LSM. One young deaf woman worked as a housekeeper and babysitter for a family in Mérida, who paid for her to take classes in LSM. The woman got in contact with the Mexican Deaf community and is now married to a deaf man from Mérida, who works in a bakery. A few years ago, her deaf brother also moved to Mérida and joined the same business as his brother-in-law. Thereafter, he returned to Chicán, opened his own bakery and married a deaf LSM signer with whom he has two children. The two deaf siblings continue using YMSL with their family in Chicán. In future studies, it would be interesting to investigate code-switching between LSM and YMSL among this interactional group of signers. So far, LSM has not had any observable impact on the structure of YMSLs. However, with ongoing societal changes, such as the increased mobility of the signers as well as the rising accessibility of communication technology and social media, networks with the national Mexican Deaf community are facilitated and we cannot predict how this will affect the younger generation’s signing in the near future.

2.3.8 Kopchen: A “control group” for the study of Yucatec Maya gesture

In order to draw comparisons between Yucatec Maya gestures and signs, I collected data not only from Chicán, Nohkop, Cepeda Peraza and Trascorral but also from another Yucatec Maya village without any deaf members. The motivation for this choice was to rule out the possibility that their knowledge of sign language prompts hearing people in the YMSL communities to “code-blend” (Emmorey et al. 2005), i.e. to produce signs from a sign language while they speak. The community where I collected gesture data, called Kopchen, has about 500 inhabitants and is in many ways representative of a typical Yucatec Maya village in terms of its community structure, the use of spoken Yucatec Maya in all domains of everyday life,

21 little use of Spanish and the presence of traditional ceremonies and rituals. Kopchen is located on the Yucatec peninsula – not in the state of Yucatán as Chicán, Nohkop, Cepeda and Trascorral are but in the neighbouring state Quintana Roo.

Figure 6: Location of Kopchen and the YMSL communities10

Travelling from Kopchen to the YMSL communities takes about half a day to one day on public transport. People from Kopchen (at least those who participated in my studies) do not have any contact with deaf people and do not know any sign language and we can thus rule out the influence of YMSL on their gesture production. For this reason, Kopchen represents an ideal source for the comparison of Yucatec Maya gestures with YMSL signs.

3 Motivations for the study

In the following section I will reflect on why the investigation of YMSLs is relevant and what contributions it makes to the disciplines of sign language and gesture studies.

10 Thanks to Calle Börstell for creating this map.

22 ● Language documentation of underdescribed, endangered languages

Although first mentions of a Yucatec Maya Sign Language go back as far as the 1970s and a number of studies have been published about its sociolinguistic setting and selected linguistic features (see Section 2.3.1), we still know relatively little about YMSLs’ lexicon and grammar. The documentation and description of YMSLs is an urgent matter because, as many other spoken and signed minority languages, YMSLs are threatened by extinction (see also Nonaka 2004; de Vos & Zeshan 2012; Webster & Safar 2019; Webster & Safar 2020). In the UNESCO’s Atlas of the World’s Languages in Danger the sign languages of Chicán and Nohkop were classified as severely endangered (Webster & Safar 2019; Webster & Safar 2020). The main reason for this precarious situation is that no more deaf babies have been born in recent years, and although hearing children also acquire YMSL from deaf caregivers and peers, it is likely that hearing people will eventually stop using sign language once there are no more deaf interlocutors around (see Groce 1985, for Martha’s Vineyard Sign Language). Another threatening factor is the ongoing dispersion of the YMSL signing communities when deaf signers leave their home village because of their marriage or jobs. Due to people’s increased mobility and the availability of cell phones, Internet and social media, YMSL signers can get in contact with members of the Mexican Deaf community and with LSM more easily than before. Apart from contributing new findings to sign language linguistics and gesture studies, my dissertation is also an effort of language documentation: It provides a description of selected linguistic features of YMSLs and gives a profound overview of the sociolinguistic constellation of four communities. Two of the four sign languages included in my project (YMSLs of Trascorral and Cepeda Peraza) are being investigated for the first time in the studies of my dissertation. Additionally, the field research for this dissertation resulted in many hours of video- recorded signing of adults and children, including different text types in YMSLs (elicited data, interviews, narratives, conversations) as well as the documentation of Yucatec Maya traditions and cultural routines. The material will be archived in The Language Archive of the Max Planck Institute (MPI) for Psycholinguistics in Nijmegen, The Netherlands.

23 • The role of gesture for language emergence

Emerging sign languages have been described as being “natural laboratories for the study of some fundamental theoretical issues in language evolution, emergence and development” (Meir et al. 2010a) because they evolve spontaneously without the influence from any established signed or spoken language. Some scholars have even gone as far as to claim that deaf people in a “linguistic void” (Meir & Sandler 2020) can form a language “out of nothing”, referring to situations “when deaf people without any prior exposure to either signed or spoken language find themselves together and form a communicative community” (Meir et al. 2010a). Such constellations are nonexistent for spoken languages11 and can give valuable cues about the human capacity to create language. Nevertheless, sign languages do not emerge out of nowhere. The idea that deaf people construct “something out of nothing”12 is, of course, effective in order to sell the relevance of research into emerging sign languages to a broader public, but it is sensationalist and not entirely accurate. Deaf signers do not live in a vacuum – they are always surrounded by hearing communities, who use not only one or several spoken languages but also a set of conventional co-speech gestures. Some of these gestures are taken up by deaf people and incorporated as raw material into their signed languages. Several previous studies have looked at diachronic paths of development from gestures to institutionalised/national sign languages such as ASL, LSF or DGS (e.g. Janzen 2012; Wilcox 2005; Wilcox et al. 2010; Janzen & Shaffer 2002; Loon et al. 2014). They demonstrated that lexical as well as grammatical elements of signed languages often have their historic origin in conventional co-speech gestures of hearing people. Over the course of generations, these gestures have been integrated into the conventionalised sign language system and modified according to language-specific phonological, morphological and syntactic constraints (Janzen 2012: 824). However, such diachronic analyses are often based on fragmentary or anecdotal descriptions of the gestural roots of signs and usually focus on older sign languages with a longer history of intergenerational transmission. Young, shared sign languages with dense networks between deaf and hearing and a large proportion of hearing L2 signers constitute a prime opportunity to observe the impact of

11 Although some argue (e.g. Fischer 1978) that creole/pidgin languages are comparable in many respects, but in these cases, language users do have a previously established system they can draw from. 12 See Rabia Ergin’s TEDxTufts talk “Something out of nothing: A brand-new language)” on Central Taurus Sign Language. Online: https://www.youtube.com/watch?v=RO1cXxEQm-w (accessed 28 September, 2020).

24 gestural conventions on a sign language in its early stages of development. Nonetheless, studies that systematically compare form, function and meaning of gestures versus signs are still relatively scarce. Tano and Nyst (2018) find cross-linguistic differences in the extent to which conventional gestures are incorporated into different shared sign languages and in the linguistic functions they fulfil. Other authors (Mesh & Hou 2018; Mesh 2017; Coppola, in press) show evidence for changes in form and meaning once gestures become part of a signed linguistic system and undergo a path of lexicalisation or grammaticalisation. Coppola (in press) notes that the comparison of hearing people’s gestures with the lexicon and grammar of emerging sign languages gives us an opportunity to study the development from gesture to sign not only from a diachronic but also from a synchronic point of view, as emerging sign languages are still in active transformation. For YMSLs, similar processes of lexicalisation and grammaticalisation from gestures to signs have been investigated, for instance for time expressions such as ‘long time from now’ (Le Guen 2012), emblematic gestures (i.e. gestures with a stable form-meaning pairing) or holophrastic gestures (i.e. gestures that replace not only one word but an entire speech act) (Le Guen et al. in press-a). The studies of this dissertation contribute further findings about the following domains:

- cardinal numbers (Study II) - size-and-shape specifiers (Study III) - patterned iconicity, i.e. the recurring use of iconic handshapes in specific semantic contexts (Study III) - height-specifiers, i.e. gestures and signs that depict a referent’s height (Study IV)

These studies can help answer questions such as: How can the common gestural substrate explain similarities between unrelated sign languages in Yucatec Maya villages? In what way do YMSLs resemble other national or village sign languages in their conventionalisation of gestural resources? Are their paths of lexicalisation and grammaticalisation compatible with models suggested in previous literature? Are there any features specific to the Yucatec Maya context that play a role in this process? Where do the limits of the influence of gestures lie and which elements of YMSLs are sign language-specific?

25 ● The influence of the sociolinguistic and cultural context on language emergence and evolution

Sign languages can emerge in very distinct settings: in urban or rural environments, in schools or families, with predominantly deaf or predominantly hearing users, etc. (see Le Guen et al., in press-b). How do these variables shape the emergence of linguistic structures? In order to shed light on this question, scholars studying village/emerging sign languages have made cross-linguistic comparisons on different levels: village vs. national sign languages (e.g. Nyst 2018; Lutzenberger 2018), emerging sign languages in villages vs. in urban Deaf communities (e.g. Meir et al. 2010a; Tkachman & Sandler 2013), village sign languages from different countries and continents (e.g. Zeshan et al. 2013), older vs. younger village sign languages (e.g. Tano & Nyst 2018), or emerging sign languages versus homesign systems (e.g. Brentari & Goldin-Meadow 2017; Coppola, in press). The case of YMSLs offers the opportunity for yet another level of comparison that has been absent from previous work, as Nyst (2012: 568) remarks: “To evaluate more closely the effect of a high incidence of deafness on a community and its linguistic mosaic, it is necessary to compare communities that resemble each other to a great extent and differ only in one or two respects.” Braithwaite (in press) attempts such a comparison of different types of signing communities (national vs. micro-community sign languages, varying incidences of deafness, varying levels of contact with other sign languages) across the Caribbean region, which – despite sharing many cultural and historic traits – still remains a fairly large and diverse area. YMSLs allow us to narrow down the scale of comparison even further. These languages developed in the same geographic region, from a shared cultural background (Yucatec Maya community structure, cultural practices, ideologies and attitudes), in contact with the same spoken languages (Yucatec Maya and Spanish as well as conventional co-speech gestures). We are looking at historically unrelated sign languages that emerged in fairly similar sociological settings, i.e. in village communities without any access to Deaf schools or similar institutions. Yet, they differ in certain aspects of their sociolinguistic composition such as their overall community size, the number of generations of deaf signers, the age of deaf signers, the presence of one versus multiple families with deaf members, deaf signers’ levels of formal education, etc. Given this unique set of circumstances, what parallels and differences can we find between YMSLs from the four communities? The Yucatec Maya case can help us to disentangle the impact of different sociolinguistic factors, which are relevant for sign language emergence and evolution.

26 • The role of variation

Another issue that linguists studying young micro-community sign languages have been trying to tackle is their high degree of variation. As many of these languages are only a few generations old, they have not had much time to converge on solid lexical and grammatical conventions. Moreover, the sociolinguistic setting of shared sign languages promotes their large amount of variation (Meir et al. 2012; Meir & Sandler 2020; Mudd et al. 2020): Shared sign languages are used in small, close-knit, socially relatively homogenous communities with a critical amount of shared background knowledge between signers and can thus tolerate a higher degree of linguistic variation without it being an obstacle to communication (e.g. Meir et al. 2012; Meir & Sandler 2020). In the absence of any institutions such as Deaf schools, as well as grammars and/or teaching materials, there is also little pressure towards standardisation. National sign languages, in contrast, have larger communities of users, who are mostly unrelated to each other, geographically dispersed, socially stratified and share relatively loose social networks. These factors increase the need for establishing shared linguistic conventions. Previous studies have demonstrated that lexical and sublexical variation is abundant in young village sign languages such as Al-Sayyid Bedouin Sign Language (ABSL) in Israel (Israel 2009; Meir et al. 2012), Kata Kolok (KK) in Bali (Mudd et al. 2020) or Langue de Signes de Bouakako (LaSiBo) in Côte d’Ivoire (Tano 2016). Variation can be found across families, between individuals and even within individual signers (e.g. Meir et al. 2012). This makes linguistic description challenging, to say the least, and the issue of variation is relevant not only on the level of an individual language but also in cross-linguistic comparisons. It remains an open question how to deal with the high amount of variability e.g. in typological studies (Zeshan et al. 2013: 392). I understand variation to be an inherent property of young shared sign languages and claim that we can use this phenomenon as a window to find out how and when linguistic conventions emerge in a community and where variability persists. Meir and Sandler (2020) argue that languages are constantly being (re)shaped by a pull between variation and conventionalisation and that it is necessary to understand the interactions between the two forces in order to trace the path of evolution of any particular language. Taking variation seriously, we need to embrace the fact that “real data are messy” (Johnston et al. 2007) instead of selecting only the most convenient (i.e. typologically unusual, “exotic”) features for analysis. De Vos and Nyst (2012: 482) suggest that a combination of qualitative and quantitative analyses may be useful to capture the phenomenon of variation in rural sign languages in a

27 comprehensive way. Young village sign languages can be informative not only in respect to the amount but also the type of linguistic variation of a language in its early stages. Meir et al. (2012) claim that in young, micro-community sign languages, variation occurs mainly on an individual level and it is only at a later stage of language development that this variation comes to display sociolinguistic patterns and index social identities.13 Studying variation in YMSLs allows us to test this claim. Although YMSL signing communities are small, they are far from homogenous and a range of social variables such as family background, age, education and gender have the potential to shape distinct signing varieties. A qualitative and quantitative analysis of selected linguistic features, paired with thorough metadata about individual signers and their social networks, can shed light on the emergence of sociolinguistic variation on a micro-level.

• Deconstruction of sign language “typologies”

Lastly, YMSLs can inform our understanding of the vast diversity of existing sign languages and possible ways to categorise them. YMSLs constitute one more example of sign languages in the “grey area” (Nyst 2012: 568): Traditional classifications, such as “village sign language” or “homesign” cannot satisfyingly capture the “multi-layered network of different villages, families, generations and overlapping deaf and hearing spaces” (Safar 2019: 31) in YMSL communities. As the studies of this dissertation demonstrate, we encounter a lot of overlap between communities on the one hand as well as a high level of variability within communities on the other hand. Parameters like linguistic complexity, level of conventionalisation, intergenerational transmission or network density between signers cannot be clearly delineated. It is questionable where to place YMSLs on a hypothetical continuum from “homesign” to “mature sign languages” and whether such an immensely simplified continuum makes any sense at all. By looking more closely at interactions of deaf and hearing signers from different communities in a variety of constellations, the investigation of YMSLs allows us to critically deconstruct existing typologies of sign languages and varieties. In my dissertation, I attempt to find an approach that can capture sign languages in the “grey area” and their unique language ecologies in an authentic way. This issue will be discussed in detail in Study I.

13 Although, in a later publication, Meir and Sandler (2020) write that the distinctive “accent” of ABSL signers when they use Israeli Sign Language (ISL) signals their group identity as members of the Al-Sayyid community.

28 4 Research questions

In my dissertation, I aim to answer the following overarching questions:

• What variation in lexicon and grammar do we encounter in YMSLs from different communities? • What variation in lexicon and grammar do we encounter within the individual YMSL communities? • How are conventional gestures of Yucatec Maya speakers incorporated into YMSLs? Which changes do these gestures undergo once they become part of the signed linguistic system? • How do the different sociolinguistic constellations of the four communities shape their sign languages?

The dissertation consists of four sub-studies, each of which treats parts of the overall research questions. The four studies tackle the central issues of the dissertation from different angles by analysing distinct linguistic levels of YMSLs. The progression of the studies moves from more general to more specific: In regard to YMSLs, I proceed from discourse (Study I) to lexicon (Study II, III and IV) to selected grammatical features (Study III and IV). Looking at hearing people’s gestures, I advance from looking at the larger semiotic repertoire of hearing Yucatec Maya (Study I) to specific types of gestures (Study II and III) to the examination of one particular gestural form (Study IV). Table 2 gives an overview of the four studies, the villages that were included, the level of analysis I adopted and the research questions I focused on.

29 Table 2: Overview of locations, levels of analysis and research questions treated in each of the four studies

Study Title Villages Level of analysis Research questions I Translanguaging in Yucatec Chicán, Nohkop, Interaction, language • Which sociolinguistic and cultural factors facilitate translanguaging Maya signing communities Cepeda, attitudes practices in Yucatec Maya signing communities? Trascorral • How can a translanguaging approach redefine traditional sign language classifications? II Numeral variation in Yucatec Chicán, Nohkop, Lexicon (numbers) • What variation can be found in YMSL numeral signs and numeral Maya Sign Languages Cepeda strategies? • Which linguistic and sociolinguistic factors influence the choice of a certain variant or strategy in YMSLs? III Strategies of noun-verb Chicán, Nohkop, Lexicon (tools), • Do YMSLs use size-and-shape specifiers and patterned iconicity to distinction in Yucatec Maya Cepeda, grammar (noun-verb mark the noun-verb distinction? Sign Languages Trascorral; distinction) • If so, in what way do these strategies differ from their gestural Kopchen (gesture) precursors? • What patterns of variation can be found between villages and among individual signers? IV “When you were that Chicán, Nohkop, Lexicon (human • How are conventional “height-specifier” gestures used by hearing little…” – From Yucatec Cepeda, referents), Yucatec Maya incorporated into YMSLs? Maya height-specifier Trascorral; grammar (noun • What are the differences in form, meaning and function between gestures to YMSL person- Kopchen (gesture) classifiers) height-specifying gestures and signs? classifier signs

30 In the following section, I will describe the methodology I adopted in order to answer these research questions.

5 Fieldwork methodology and data

The aim of the following section is to give an overview of my methods of investigation and the data I collected. In recent years, general awareness about the necessity for reproducible research in the discipline of linguistics has been increasing (Berez-Kroeker et al. 2018). Within the research traditions of the humanities or social sciences, it is not possible to apply the same guidelines for scientific rigor as in natural sciences, where researchers conduct experiments under controlled lab conditions and strive to make it possible for other researchers to replicate the same procedure and obtain the same results. Due to the nature of linguistic field research – a human being who studies complex, uncontrolled and uncontrollable interactions of other human beings – overcoming the “observer’s paradox” is virtually impossible. Many variables involved are contingent on the specific situation under which the data are collected, and on the subject of the researcher him- or herself. Nevertheless, it is desirable to provide a maximum level of transparency about how our results and conclusions about linguistic phenomena are generated. In their position statement, Berez-Kroeker et al. (2018) call for transparency on different levels: - transparency about methods of data collection (e.g.: Who collected the data and under what conditions? Which tools and materials were used for data collection?) - transparency about methods of data analysis (e.g. Which tools were used for transcription and analysis? How was the data coded? etc.) - information about participants (Who provided the data for analysis?) - transparency about source data (e.g. What types of data were used for analysis? Where can the data be accessed?) Transparency means becoming visible, exposing not only the successful outcomes of one’s research but disclosing also possible shortcomings in one’s methodology. Ultimately, however, granting other people access to one’s tools for data collection and analysis as well as to the datasets themselves contributes to creating a more open and collaborative framework for doing research and to improving the quality of linguistic studies.

31 In line with these standards, I will attempt to make my studies reproducible for other researchers by providing the following: a detailed explanation of my methods of fieldwork, the conditions under which I collected data, and the types of data I collected (Section 5.1.1); an account of how I collaborated with members of the YMSL communities in the process of data collection and analysis (Section 5.1.2); a critical reflection of my fieldwork methods (Section 5.1.3); and a detailed description of the datasets used for each of the four studies of my dissertation (Section 5.2). The coding protocols for the four studies are attached in the appendix (see Appendix A). The original datasets, on which the four studies are based, as well as the coded data in spreadsheet form are made available via the platform Open Science Framework (see Section 5.2), provided that the participants had agreed for this data to be shared. Researchers can request access to the files after having registered at the website.

5.1 Fieldwork and data collection

As in most fieldwork-based language documentation projects, I used a combination of various methodological approaches, in order to capture different linguistic phenomena as well as the sociolinguistic context of YMSLs. Data collection in the field was carried out in collaboration with other researchers, and members of the YMSL communities acted as research assistants. In the following sections, I outline my fieldwork methodology (Section 5.1.1), describe my ways of cooperation with the YMSL communities (Section 5.1.2) and reflect in hindsight about some methodological decisions (Section 5.1.3).

5.1.1 Fieldwork methodology

The first fieldtrip to Yucatán I undertook was for my MA thesis in 2013, where I stayed for about four weeks in Yucatán (most of the time in Chicán, as well as some days in Mérida and Kopchen). I had the chance to establish first connections with the community members and other researchers, carry out ethnography and participant observation and assist Olivier Le Guen with elicitation tasks. Moreover, I did some research in an archive in Mérida in order to gain access to old newspaper reports about deafness in Yucatán. This first visit was followed by two periods of fieldwork for my PhD: In 2015, I spent a total of six weeks in Yucatán. Most of this time, I stayed in Chicán and Nohkop, with a visit of several days in Kopchen (Yucatec Maya village without any deaf members, to collect gesture data) and day trips to Trascorral

32 and Cepeda. In 2016/2017, I was in the field for a total of three months, most of which I spent in Chicán and Cepeda. Additionally, I visited Nohkop and other surrounding communities, where Nohkop signers had moved in recent years. In Chicán, Nohkop, Trascorral and Kopchen, I was able to benefit from contacts, structures and routines that Olivier Le Guen and his project team had established during their previous periods of fieldwork. In Cepeda Peraza, no other linguist had worked before and I had to set up these connections myself. I will reflect on this process in more detail in Section 5.1.3. My methodological approach can be described as an instance of “linguistic ethnography”, an interdisciplinary framework, which aims to provide an analysis of language embedded in its social context (e.g. Copland et al. 2015). During my fieldtrips, I lived in the Yucatec Maya villages and stayed in the homes of hearing families without any deaf members.14 Once I got accustomed to life in the communities, I established a daily routine, which followed a certain rhythm but turned out a little different every day, according to the opportunities that presented themselves. I visited deaf and hearing people in their homes and took part in their everyday activities, such as cooking, exchanging gossip while preparing tortillas, weaving hammocks or sharing meals. I accompanied people to their milpa or beehives, and attended social gatherings, celebrations and rituals. By participating in these daily practices, I was able to observe interactions between deaf and hearing community members and witness encounters with people from outside (e.g. visitors from other villages, government representatives, students, NGO workers, church members, other researchers, etc.). I always carried a small notebook with me, in which I jotted down the most important observations and key facts I came across during the day. Every evening, when relaxing in my hammock, I wrote more elaborate field notes. Another part of my evening routine was to transfer all the files from my camera’s memory card to an external hard-drive, make a backup to another hard-drive, save all metadata about the recordings (place and time of recording, types of data recorded, participants, comments, etc.) into an Excel spreadsheet, have a brief look at the recordings and note down open questions on which I wanted to follow up.

14 The reasons for the choice to live with hearing families were purely practical in the first place – the people I ended up staying with were the ones who had the space and resources to host me for an extended period of time. It proved, however, advantageous not to live with any of my main informants: In my experience, feelings of jealousy arose quickly between community members whenever they had the impression that I would “prefer” one family over another.

33 I am a hearing person, speak Spanish proficiently, but at the time I first arrived in Yucatán, I did not know any YMSL or spoken Yucatec Maya. Gradually picking up YMSL was both an aim and a side effect of my fieldwork and from the onset, I always strove to use YMSL in everyday interactions with deaf people. Whenever I ran into trouble, I relied on gestures, pantomime or other “semiotic resources” (see Study I) for communication but tried to avoid the use of any other sign language I know, for instance Swedish Sign Language or Austrian Sign Language. For in-depth interviews about more complex topics, such as language attitudes (see Study I), I benefited from the help of hearing research assistants (fluent users of YMSL, Yucatec Maya and Spanish) as “interpreters”. With hearing people, I communicated in Spanish and acquired basic knowledge of Yucatec Maya over the course of my fieldwork. Interviews with hearing people were carried out either in Spanish or in Yucatec Maya, in collaboration with other members of the YMSL project who are native speakers of Yucatec Maya.

Figure 7: Myself conversing with two deaf YMSL signers

The prolonged amount of time I spent in the villages and the ethnographic investigation of everyday life in the communities was crucial for me in order to understand the social context of YMSLs and to get a feel of how sign language is used, by whom, in what constellations etc.

34 Parallel to the ethnographic research, I video-recorded different types of linguistic data: elicited data using picture and video stimuli (see Section 5.1.3, for reflection), informal interviews as well as natural discourse (narratives, conversations between two or several signers). Many of the elicitation sessions were carried out together with one or two research assistants. During some of the recordings of spontaneous conversations I was present, during others I just installed the camera and let the signers converse on their own.15 In Section 5.2, I provide a detailed breakdown of the different types of data used in each of the four studies of my dissertation. Informed consent was obtained from all participants: with deaf participants, a deaf or hearing research assistant explained the research procedures in YMSL and asked the participants whether or not they agreed to their data being used for different purposes. The whole procedure was recorded on video. Hearing participants gave their verbal consent either in Spanish or Yucatec Maya, which was also video-recorded. I discuss the process of finding an adequate way for obtaining informed consent and other ethical considerations in detail in Safar (forthcoming-a). Over the years, I employed different types of picture and video stimuli for elicitation. For the lexical elicitation task, on which parts of the Studies II, III and IV are based, I designed my own set of photo stimuli, using mostly photos taken in Yucatán (by Olivier Le Guen or myself). For other tasks, I made use of elicitation materials designed by other researchers for different grammatical domains, for instance ‘giving and taking’ events (courtesy of the MPI for Psycholinguistics), ‘cutting and breaking’ events (Bohnemeyer et al. 2001), handling vs. instrument handshapes (Padden et al. 2013) or transitive vs. reflexive verbs (Börstell 2017). In Section 5.1.3, I will provide some reflections concerning the use of elicitation materials. The following section is dedicated to an introduction of the YMSL documentation project and our ways of collaboration with the communities.

15 Of course, even in my absence, people are aware of being filmed and video-recorded data can thus never be one-hundred percent “authentic” but, in my experience, participants tend to forget about the camera after a while, especially those who are already used to working with researchers. Much to my amusement, I have videos where signers start making fun of my hairstyle fifteen minutes into the recording and then blush when remembering the camera.

35 5.1.2 Collaboration with the communities

During my PhD research, I was part of the YMSL documentation project.16 The main aims of this project are the compilation of an online dictionary for Chicán and Nohkop,17 the description of selected grammatical structures and the collection of a corpus of signed texts. The project team has also produced a non-academic publication about YMSLs, a bilingual book in Yucatec Maya and Spanish with portraits of deaf signers in Chicán and Nohkop.18 The YMSL project is led by Olivier Le Guen and comprises a team of anthropologists and linguists (Olivier Le Guen, Lorena Pool Balam and myself) and undergraduate linguistics/cultural studies students from Yucatán (at the time of my fieldwork these were Rebeca Petatillo Balam and Rodrigo Petatillo Chan from Kopchen as well as Maribel Pacheco from Mérida – at the time of writing, further members have joined the project). Most project members are native or highly proficient speakers of Yucatec Maya. Additionally, the project team includes deaf and hearing community members from Chicán and Nohkop, who were trained as research assistants. My own background as a hearing, female, European academic inevitably shaped my access to the communities and my perception of the phenomena under study. The close collaboration with native signers as research assistants was of great support to the success of my fieldwork and the rapport I was able to build with my informants. The YMSL signers learned to assist with elicitation tasks (operating the camera, explaining the elicitation procedure, presenting stimuli on a tablet), transcribe using the annotation program ELAN19 and make translations into Yucatec Maya and Spanish. They were also involved in interpreting the process of informed consent as well as interviews with deaf signers. The research assistants were equipped with laptop computers and received weekly payment for their work.

16 http://ymslproject.org/ (accessed 28 September, 2020) 17 http://ymslproject.org/index.html#voc (accessed 28 September, 2020) 18 The book can be downloaded from the following link: https://www.gob.mx/cms/uploads/attachment/file/438175/inpi-el_habla-de-la-mano-lengua- de-senas-maya-yucateca-movil.pdf (accessed 28 September, 2020) 19 ELAN is a tool for the annotation of video and audio data, which was developed at the Max Planck Institute for Psycholinguistics in Nijmegen (see e.g. Wittenburg et al. 2006). The software can be downloaded free at https://archive.mpi.nl/tla/elan (accessed 28 September, 2020).

Figure 8: Linguistics students Rodrigo Petatillo Chan and Rebecca Petatillo Balam from Kopchen (left of the table) as well as one deaf (Maria Yeseña Kinil) and one hearing YMSL signer (Rossy Kinil Canché) from Nohkop (right of the table) making transcriptions in ELAN (Photo: Olivier Le Guen)

Finding community members who were willing and able to work with us for an extended period and dedicate their time and effort to doing research about their native sign language was not easy. Potential research assistants need to have essential computer skills, and ideally should not be occupied with other jobs or family obligations. Basically, these criteria only apply to young people. During my first field trip, I worked with two hearing CODAs from Chicán, who are trilingual in YMSL, Yucatec Maya and Spanish. Dianela Collí Chim has a deaf father and a deaf aunt and grew up using YMSL at home on a daily basis. Merli Collí Hau has a deaf father and several other deaf relatives (aunts, uncle, cousins) and was socialised in an active signing environment. In Nohkop, I was able to benefit from the support of Rossy Kinil Canché, who is hearing and grew up in close interaction with her four deaf cousins. She is fluent in YMSL, Yucatec Maya and Spanish. Unfortunately, Rossy quit her collaboration with the YMSL project when she got married and had a child. During my second field trip, I worked most intensively with two research assistants from Chicán: Merli Collí Hau and her cousin Geli Collí Collí, a deaf native signer of YMSL. Both of Geli’s parents as well as her

37 brother are deaf and she also has a deaf aunt, a deaf uncle and a deaf cousin. When I was in Chicán, I went to see Geli and Merli at their houses every day, checked their transcriptions and talked to them about their observations. This collaboration was one of the most enjoyable elements of my fieldwork, especially once they gained more confidence about their role as experts of their own language and started to contradict me when they disagreed with my interpretations. They acquired some basic linguistic terminology and developed very good intuitions about their own language.20

Figure 9: Two research assistants (Merli Collí Hau in the pink shirt and Geli Collí Collí in the blue shirt) assisting with an interview of a deaf signer in Chicán (woman in the white huipil)21

Even though Geli has only basic literacy in Spanish, she learned to make transcriptions and translations in ELAN, with the help of the use of the “Controlled Vocabulary” function

20 A particularly delightful moment occurred during a car journey in rural Yucatán with the YMSL project team. Merli, at the time 14 years old and attending basic education, was having a signed conversation with her cousin, when she suddenly interrupted herself and asked me whether the YMSL signs DIZZY and DREAM constituted a minimal pair, which indeed they do. 21 Traditional clothing worn by Yucatec Maya women

38 (where the person making the transcriptions in ELAN can select from a list of possible glosses in a drop-down menu) as well as her excellent visual memory and sign transcription symbols she invented.

Figure 10: Customised “Controlled Vocabulary” function in ELAN (left) and self-created sign language transcription symbols by Geli (right)

Geli and Merli were also involved in the analysis of the data we collected together, especially for Study II, and they are co-authors of the resulting paper. At the end of my second field trip, they visited a conference in the town Valladolid together with the rest of the YMSL project team, which was a good occasion for them to see what our academic work consists of and what happens with the data. In the following section, I will critically reflect on some methodological decisions, certain challenges I encountered during fieldwork, and experiences I gathered over the years.

39 5.1.3 Looking back: Challenges and learning curves

A language documentation project that takes place over the course of several years is a dynamic process where ideas, approaches, skills, techniques, relationships and – last but not least – the researcher him- or herself change and evolve over time. Looking back at the materials I collected during the two periods of fieldwork and at the four sub-studies of my dissertation, I see a progression on various levels, for instance: the quality of my recordings, transcriptions and fieldnotes, the systematicity in data coding and analysis, my ability to flexibly apply various methods of data collection, my signing skills in YMSL, and my ways of relating to the participants of my studies. It took some rounds of trial and error to find out how to best set up the camera in sparsely lit Maya huts, how to position myself and encourage participants’ responses during elicitation sessions, how to filter the most important information out of the vast amount of new impressions, or how to design annotation templates that my research assistants could use with maximum ease and efficiency. Now, that my project is close to completion, I am able to re-examine some of the obstacles I encountered and how I attempted to overcome them. One major challenge everybody documenting a village sign language eventually faces is the choice of adequate elicitation materials (Nyst 2015; de Vos 2016). Materials designed for studying sign languages of larger, urban Deaf communities and/or spoken minority languages can be useful but often need to be adapted to fit the specific requirements and the sociocultural context of the community/communities under study. Some materials used in other studies cannot be applied to non-literate participants with little formal education. For instance, certain tasks to elicit cardinal numbers used by Zeshan et al. (2013) could not be replicated with deaf Yucatec Maya participants, many of whom cannot read written numbers (see Study II).22 Particular stimulus sets that had been designed for other signed or spoken languages are not meaningful to people who live in a Yucatec Maya village. These issues make cross- linguistic comparisons with other studies difficult and impose more workload on the individual researcher who needs to compile his/her own materials. What proved to be the most successful strategy in my fieldwork was to use a combination of materials used in previous linguistic or psycholinguistic studies, such as some of the materials available on the L&C Field Manuals

22 We resolved this issue by using photos of Mexican Pesos (bills and coins in variable quantities and arrangements) as stimuli, as even the non-literate participants are used to handling money on a daily basis and are familiar with its value. This turned out to be a successful method of collecting data on YMSL cardinal numbers.

40 and Stimulus Materials Website of the MPI for Psycholinguistics23 and materials specifically designed for the Yucatec Maya context. For instance, in order to elicit ditransitive verbs for giving and taking events (Le Guen & Safar, in prep.), we used video stimuli elaborated by the MPI for Psycholinguistics in Nijmegen but reproduced the videos with Yucatec Maya actors and objects of daily use in a Yucatec Maya village.

Figure 11: Stimulus materials “Give and Take” from the MPI (left) and adapted to the Yucatec Maya context (right)

In my experience, stimulus materials that depict a familiar environment usually prompted participants to produce livelier and more elaborate signed descriptions than materials produced in a context they do not recognise. But of course, not all informants react in the same way and I always found it worth trying out a task rather than discarding it from the start. Ultimately, doing fieldwork is very much a path of trial and error. In another paper (Safar, forthcoming-a), I describe in more detail the process of compiling stimulus materials for the lexical elicitation study. Elicitation tasks are helpful tools in order to target specific linguistic phenomena of interest. They are easy to use, especially in the beginning of a language documentation project where a researcher only has limited skills in the language he or she studies. However, elicited data in isolation is not sufficient to make general claims about linguistic structures; instead, the findings need to be verified in natural discourse. I used this approach of combining elicited with narrative/conversational data (see e.g. Chelliah 2001, for advantages of such a

23 See http://fieldmanuals.mpi.nl/ (accessed 28 September, 2020)

41 combination) in the Studies II, III, IV and it proved to be a fruitful method: First, I was able to discern certain patterns in elicited data, then I analysed data from narratives and conversations in order to verify whether these structures really occur in spontaneous signing and how they are used in context. With the progression of the project, a further step entered the workflow: First, identifying potentially interesting phenomena in natural discourse, second, exploring these phenomena through structured elicitation tasks, third, cross-checking the findings with data from natural discourse. A cycle between the examination of spontaneous language use and the analysis of controlled, elicited data is, in my view, the most successful approach to language documentation. An interesting learning process for me was to discover how individual participants respond to different methodological approaches. In the beginning, in order to maximise the comparability of the data, I wanted to use the same procedures with everyone and found that while they were very successful with some informants, they did not work well with others. All participants have their individual skills: some people enjoy being given tasks, some are brilliant storytellers, others are shy but sharp observers, some are at their most expressive when working in pairs with others, some come up with astonishing metalinguistic reflections, some are encyclopaedias of cultural knowledge, some are an excellent source of gossip, others tell funny jokes that can cheer one up on a day when work seems to be progressing painfully slowly. None of these resources should be underestimated in a language documentation project and all of them add a little piece to the puzzle of unravelling a hitherto undescribed language. With time I learned to be more flexible, spontaneous and creative and developed a better eye for opportunities that presented themselves. For that to occur it was necessary to become familiar with people, their languages and their lifestyle and to gain confidence about my role as a researcher in the communities. And still, as probably every fieldworker can confirm, the most relevant things always happen when the camera is off. While the focus was mostly on elicitation during my first period of PhD fieldwork, I collected mainly narrative and conversational data during the following field trip. These materials not only present a rich resource of naturalistic language data but also offer a glimpse into people’s biographies and worldviews. The challenge, however, is that narratives (monologues) and, even more so, conversations (in dyads or groups) are difficult and extremely time-consuming to transcribe and translate and without the support of my research assistants, who are native signers of YMSL, this would have been impossible. Over the course of the project, the research assistants became very proficient transcribing using the program ELAN but were restricted by their school curriculum and material constraints (availability of technical

42 equipment, space to work, etc.). As a result, I have many, many hours of raw, untranscribed YMSL data, which are nevertheless valuable for the purposes of language documentation. They will be archived and made accessible to other researchers in The Language Archive of the MPI for Psycholinguistics, Nijmegen. Another inevitable learning process for a field linguist is that academic aspirations and the realities of fieldwork do not always coincide with one another. For a comparative study, my initial goal was to collect similar amounts of the same types of data from a maximum number of signers from all four communities. However, in real life, this turned out to be impossible. For a number of practical reasons, I ended up spending most of the time in Chicán. First of all, the village is easily accessible by public transport and it includes a (comparatively) large number of signers from different age groups and families. Other linguists and anthropologists have worked there before me, which means that deaf and hearing signers are familiar with being filmed and interviewed. Because of my collaboration with Olivier Le Guen and the YMSL project team, I was lucky to benefit from structures that had been established during their previous research in Chicán. This meant that upon arrival in the village, I was able to start working with deaf and hearing informants straight away and relations of mutual trust between community members and the YMSL project team already existed. Over the years, I built on these initial arrangements, set up my own contacts and expanded my social networks in the communities. The YMSL project team had also arranged to find a convenient place to stay on the compound of a generous and helpful family, where I benefited from having a tiny bit of privacy and enough space to work. Last but not least, we dedicated a lot of time training three research assistants in Chicán to do transcriptions, translations and elicitation, which facilitated the process of data collection and analysis immensely. For these reasons, I was able to gather much more data from Chicán than from the other communities. In Nohkop, Le Guen and his team had established an initially very well-functioning setup of collaboration with the signing community, where some deaf and hearing signers were actively involved in collecting and transcribing data from their native sign language. Over time, however, signers’ lives changed and today, the signing community of Nohkop is largely dispersed. Four signers (three deaf sisters, one hearing cousin) that used to be in daily contact have now married and moved to their husbands’ homes. They do not get the chance to interact often as a group, and their husbands are not very much in favour of their wives continuing to work on the YMSL project. However, as two of the deaf sisters married two brothers, they now continue living together in one house in their husbands’ hometown and carry on using YMSL

43 with each other, their husbands and their children. Visiting some of the Nohkop signers in their new homes entails driving long distances to rather remote locations. The signers in Trascorral were welcoming and in favour of us doing research about their sign language, but they are dealing with numerous private problems in the family. For this reason, we preferred not to intrude too much and only visited them occasionally. The village of Trascorral is fairly remote, and I was not able to go there when I was in the field on my own and travelling only by public buses. For these reasons, my studies include only a small amount of linguistic data from Trascorral as well as relatively limited sociolinguistic information and metadata. Some of these data were included in Studies III and IV. However, it is fair to say that at this point, we have managed to capture only a snapshot of YMSL from Trascorral and its language ecology. In Cepeda, I was the first linguist to work in the community and had to create a setup for conducting fieldwork there – a process that proved to be equally challenging and rewarding. I had heard about the high incidence of deafness in Cepeda from community members in Chicán and from Ernesto Escobedo Delgado, a deaf Mexican linguist, who had done fieldwork in Chicán. In 2015, I visited Cepeda for one day together with a hearing man and his hearing son, with whom I lived in Chicán. During that first visit, I got to know a family that includes two deaf sisters and made some first video-recordings at their home (elicitation of lexicon and grammar as well as some conversational data). During my next field period in 2016, I undertook another day trip together with the same two men as well as two research assistants from Chicán. This encounter is described in detail in Study I. After that, I went back alone to stay for a few weeks in the village. I met a family who was able to host me, conducted a sociolinguistic survey of the community and started working with deaf signers. I found myself very well received by community members, who were surprised and intrigued that someone would be interested in studying their sign language. It took a while to convey to people the aims and the significance of my research and after overcoming an initial period of mutual shyness and getting to know each other, most deaf people were eager to share their language with me and were comfortable with being video-recorded. Some of the deaf young men, however, were happy to meet for a chat but remained reluctant about being filmed. They explained that another foreigner had come to work with deaf people in the village before (NGO work not related to sign language) and not everyone approved of her behaviour. This demonstrated to me that one’s attitudes and actions in a field location do not only have an impact on the success and the quality of one’s own work but also affect other people who wish

44 to do research there in the future. The process of setting up a fieldwork-based linguistic project in a rural indigenous community requires patience and cultural sensitivity. Staying for an extended period of time in a small village inevitably means becoming part of its social structures. This has its advantages and disadvantages: Gaining the trust of community members and achieving a deeper understanding of their social relations enables us to collect more and more meaningful data. At the same time, it also entails getting entangled in social networks, becoming a source of gossip, being watched. Having to make decisions about who to collaborate with and how to compensate informants and research assistants for their work can be delicate ethical issues (see also Safar, forthcoming-a). Doing fieldwork in a remote geographical and cultural setting is an extremely rewarding and enriching experience, but it is also challenging on many levels. Typically, field linguists do not share these experiences openly but often tend to idealise and romanticise their time in the field (see Safar 2018, for reflection). Personal accounts of linguistic fieldwork as a holistic (academic, intercultural, interpersonal) experience can be found in Sarvasy and Forker (2018).

5.2 Datasets and coding

In the following section I aim to give an overview of the types of data I collected and the proportions to which different text types were included in each of the four studies. In order to provide readers with a better understanding of how the data was coded for analysis, a coding protocol for each of the four studies is attached in the appendix (see Appendix A). In these protocols, I explain in detail which materials I selected for analysis, how I identified relevant sub-parts of the videos for transcription, what tiers I used for creating annotations in ELAN, what criteria I applied for inclusion or exclusion of a response in the analysis and who was involved in transcribing and translating the data. In Study I (Translanguaging in Yucatec Maya signing communities; see Section 6.1), I analysed conversational data from deaf and hearing YMSL signers recorded in Cepeda and Chicán, semi-structured interviews with deaf and hearing YMSL signers recorded in Chicán and Trascorral, as well as ethnographic data from participant observation and unrecorded informal interviews from all four communities. The latter type of data is difficult to quantify, as it is a result of approximately six months of fieldwork and many pages of scribbled fieldnotes. In Table 3 is a list of the data used for Study I.

45 Table 3: Data analysed for Study I

Type of data Length (hh:mm:ss) Conversations 03:48:29 Interviews 03:07:02 (video-recorded) Additional data: Participant observation, interviews (not video-recorded), field notes

The dataset for Study I can be accessed via the platform Open Science Framework (OSF) (Safar 2020a). Access restrictions are placed according to my agreements with the participants. Additionally, the coding protocol for Study I and an example for a transcribed and coded video section are available on OSF (Safar 2020a). In Study II (Numeral variation in Yucatec Maya Sign Languages; see Section 6.2), we included data from a picture elicitation task as well as from narratives and conversations recorded in three villages: Chicán, Nohkop and Cepeda. Additionally, our analysis was supported by ethnographic data from participant observation and unrecorded informal interviews. Table 4 gives an overview of the types of data used in Study II.

Table 4: Data analysed for Study II

Type of data Length (hh:mm:ss) Chicán Nohkop Cepeda TOTAL Elicitation 01:11:50 00:32:02 00:08:47 01:52:39 Narratives/conversations/ 01:08:06 00:19:59 00:20:14 01:48:19 interviews Additional data: Participant observation, interviews (not video-recorded), field notes

The datasets for Study II, as well as the coding protocol and the coded data in Excel format can be accessed via OSF (Safar 2020b). Study III (Strategies of noun-verb distinction in Yucatec Maya Sign Languages; see Section 6.3) consists of two parts (sub-Study 1 and sub-Study 2), both of which included different types of data. In sub-Study 1 concerning size-and-shape specifiers (SASSes), we employed data from a picture elicitation task conducted with YMSL signers from Chicán,

46 Nohkop and Cepeda. In order to test whether the SASS constructions that we observed in the elicitation task also occur in spontaneous signing, we looked at the use of SASSes in video- recorded conversations including YMSL signers from Chicán, Nohkop and Cepeda. For the comparison with size-and-shape gestures used by hearing speakers of Yucatec Maya, we ran the same elicitation task with gesturers from Kopchen. Moreover, it was necessary to examine if the silent gestures that Yucatec Maya speakers produced under the conditions of an experimental task are also used in spontaneous talk, i.e. along with spoken Yucatec Maya. To that end, we included video-recorded narratives and interviews with Yucatec Maya gesturers from Kopchen in our analysis. Table 5 shows a breakdown of the types of data used in Study III/sub-Study 1.

Table 5: Data analysed for Study III/sub-Study 1

Type of data Length (hh:mm:ss) Sub-Study 1 Chicán Nohkop Cepeda Kopchen TOTAL (YMSL) (YMSL) (YMSL) (gesture) Elicitation 02:44:03 01:01:31 00:13:26 01:13:04 05:12:04 Narratives/conversations 01:26:56 00:06:28 00:00:00 00:59:19 02:32:43 Additional data: Participant observation

For the second part of Study III, exclusively elicited data was used. The same elicitation task was run with YMSL signers from three villages (Chicán, Nohkop and Trascorral) and hearing gesturers from Kopchen. Additional data comes from observations in the field made by Olivier Le Guen (fluent speaker of Yucatec Maya) and Rodrigo Petatillo Chan (native speaker of Yucatec Maya). Table 6 provides an overview of the amount of data recorded in each village.

47 Table 6: Data analysed for Study III/sub-Study 2

Type of data Length (hh:mm:ss) Sub-Study 2 Chicán Nohkop Trascorral Kopchen TOTAL (YMSL) (YMSL) (YMSL) (gesture) Elicitation 01:40:30 00:51:33 00:13:05 01:57:23 04:42:31 Additional data: Participant observation

The datasets for both parts of Study III are available via OSF (Safar 2020c). Access restrictions are imposed according to our agreements with the participants of the study. The coding protocols for both sub-studies as well as the Excel spreadsheets with the coded data are also included in the OSF repository (Safar 2020c). In Study IV (“When you were that little…” – From Yucatec Maya height-specifier gestures to Yucatec Maya Sign Language person-classifier signs; see Section 6.4), I analysed data from YMSL signers from Chicán, Nohkop, Trascorral and Cepeda as well as hearing non- signing speakers of Yucatec Maya from Kopchen and Chemax. The YMSL dataset consisted of elicited data from a picture elicitation task, video-recorded narratives and conversations. The gesture data analysed came from the same elicitation task (with Yuactec Maya speakers in Kopchen) as well as from video-recorded narratives, conversations and interviews (recorded in Kopchen and Chemax). Additionally, I included observations I made in the field outside of filming contexts. The types of data analysed in Study IV are listed in Table 7.

Table 7: Data analysed for Study IV

Type of data Length (hh:mm:ss) Chicán Nohkop Trascorral Cepeda Kopchen/ TOTAL Chemax (YMSL) (YMSL) (YMSL) (YMSL) (gesture) Elicitation 01:31:03 00:17:18 00:08:47 00:05:33 00:16:50 02:19:31

Narratives/ 02:46:53 00:20:35 00:00:00 00:55:35 01:13:23 05:16:26

conversations Additional data: Participant observation, interviews (not video-recorded)

48 The datasets for Study IV are accessible via OSF (Safar 2020d). Access restrictions are imposed according to my agreements with the participants. Additionally, the coding protocol for Study IV and the coded data in Excel spreadsheet form can be found on OSF (Safar 2020d). After having introduced the datasets, a number of critical remarks must be made regarding the different types of data used in the four studies. First, even though Studies II, III and IV included both elicited and spontaneous signing, the frequency with which a specific phenomenon occurs is naturally higher in elicited data than in narratives or conversations. Finding clear examples for the use of a specific form in natural language data is more difficult than under the controlled conditions of an elicitation task aimed at targeting a certain phenomenon. However, the domains of interest for Studies II, III and IV (cardinal numbers, SASSes, height-specifiers) all occur with high frequency in YMSLs and it was easy to identify instances of their use in spontaneous discourse. The findings from narrative/conversational data were supported by observations that myself, my co-authors and my research assistants made in the field. Second, in order to draw parallels between Yucatec Maya gestures and YMSL signs, we conducted the same elicitation tasks as in YMSLs with hearing speakers of Yucatec Maya, and asked them to describe the picture stimuli without speaking. These silent gestures that hearing people produced under the controlled conditions of an elicitation task are not equivalent to gestures they spontaneously produce in natural settings along with speech. To be able to claim that gestures used by hearing people are incorporated into YMSLs it is thus necessary to work with data gathered from spontaneous discourse. As shown in the tables above, this important step was made for Study III/sub-Study 1 (see Table 5) and for Study IV (see Table 7). However, no narrative/conversational data from Yucatec Maya gesturers were included in Study III/sub-Study 2. This lack was counterbalanced by the fact that Olivier Le Guen (fluent speaker of Yucatec Maya who lives in Mexico and has been doing fieldwork in Yucatán for 17 years) and Rodrigo Petatillo Chan (native speaker of Yucatec Maya who grew up in Kopchen) were able to contribute their observations about the use of handling/instrument handshapes by Yucatec Maya gesturers to the study. Critical reflections about the methodology used in Study III are included in Section 6.3. Third, the different types of data were not given the same weight in all four studies of my dissertation. In Study II and Study III, elicited data served as the point of departure, and narrative/conversational data helped to subsequently verify the results from the elicitation tasks. As I increasingly became aware of the advantages of using natural language data from the onset (see my critical comments in Section 5.1.3), I chose a different approach for Study

49 IV. I started off with data from both elicitation tasks and narratives/conversations and coded them in the same level of detail. This procedure generated results about the broad spectrum of possible uses of height-specifiers in YMSL and Yucatec Maya gesture, some of which I might have overlooked if I had initially focused on elicited data. Fourth, the reader can note that the amount of data used as well as the proportions of elicited versus narrative/conversational data are not balanced for the different communities included in Studies II, III and IV. This is a result of the uneven size of the signing communities on the one hand, and a shift in my approach to data collection on the other hand. While I focused mainly on a lexical elicitation task during my first fieldtrip in 2015, I recorded a much broader variety of data during my second period of fieldwork in 2016/2017. For this reason, I conducted more elicitation sessions in Chicán and Nohkop than in Cepeda Peraza, where I started working more extensively only during my second fieldtrip. As I spent the least amount of time in the community of Trascorral, I was able to record relatively little data from signers in this village. The full YMSL corpus that I collected during my PhD project will be archived and made available after the completion of my PhD via the website of The Language Archive of the Max Planck Institute (MPI) for Psycholinguistics, Nijmegen. It can be found by browsing the archive for “Yucatec Maya Sign Languages” or accessed via a permanent link.24 Due to constraints in time and resources, only a small part of the corpus has been annotated, translated and analysed to date.

6 The four studies: Summary and reflexion

In the following section, I will provide an overview of the four sub-studies that make up my dissertation.

6.1 Translanguaging in Yucatec Maya signing communities

In Study I, I looked at interactions between deaf and hearing people in Yucatec Maya signing communities in different constellations: interactions between deaf and hearing people from the same village, deaf people in interaction with hearing people from other villages in Yucatán and interactions between deaf people from different communities. For this analysis, I used the theoretical framework of translanguaging, which encompasses the “deployment of a

24 https://hdl.handle.net/1839/8da05fe4-c881-4d83-aaf9-e09b2a65530a (accessed 28 September, 2020)

50 speaker’s full linguistic repertoire without regard for watchful adherence to the socially and politically defined boundaries of named (and usually national and state) languages” (Otheguy et al. 2015: 281). Deaf and hearing people in Yucatec Maya communities use a broad array of communicative resources to interact with each other: they sign YMSL with various degrees of fluency, speak Yucatec Maya and/or Spanish, use gestures, draw on paper, in the air or on the ground, point and incorporate objects of their physical surroundings. Kusters et al. (2017a) extend the notion of “linguistic repertoire”, arguing that translanguaging practices are inherently multimodal and instead introduce the term “semiotic repertoire”. When signers and speakers communicate with others, they flexibly exploit their semiotic repertoires, selecting elements that suit the specific dynamics of the interaction as well as the communicative needs and resources of their interlocutor(s). In my study, I used a combination of ethnographic data gathered through participant observation, informal and semi-structured interviews as well as linguistic data from elicitation tasks, narratives and natural conversations in YMSLs (see Table 3). I investigated the semiotic resources which Yucatec Maya signers integrate into their interactions; which cultural and sociolinguistic features of Yucatec Maya signing communities facilitate translanguaging practices, and where their boundaries lie. Our “translanguaging instinct” (García & Wei 2014: 32) is a general human capacity, but in the Yucatec Maya context it is scaffolded by certain sociocultural characteristics. At the base of translanguaging encounters are cultural positive attitudes towards deafness and sign language: hearing Yucatec Maya consider deaf people as intelligent, competent interlocutors and sign language as an adequate, fully fledged channel of communication (see Sections 2.3 and 8.3). Their shared gestural and cultural repertoires facilitate communication between deaf and hearing people and lay a foundation for the emergence of shared sign languages when deaf people are born into a Yucatec Maya community. YMSL signers draw from the rich inventory of conventional co-speech gestures used by speakers of Yucatec Maya and incorporate them into their sign languages (see Sections 8.1 and 8.2), which leads to similarities in lexicon and grammar between historically unrelated sign languages. Also, collective cultural knowledge about Yucatec Maya customs, routines, beliefs, etc. helps to facilitate understanding even between strangers without a shared language. Importantly, these parallels go beyond similarities that are a result of modality-specific properties, i.e. due to iconicity (Meier et al. 2002; Occhino et al. 2017). There are, however, limits to translanguaging practices as not everybody has access to the same semiotic resources (De Meulder et al. 2019). The analysis of translanguaging in Yucatec Maya signing communities mirrors observations made, for

51 instance, by Kusters (2017): The more the repertoires of different communication participants overlap, the more easily meaning can be negotiated. Both interviews with deaf and hearing signers and ethnographic observations of their interactions demonstrate that knowledge of YMSL is needed in order to discuss more complex topics. At the core of my study is a detailed analysis of an encounter between deaf and hearing YMSL signers from Chicán (research assistants in the YMSL project) and a deaf YMSL signer from Cepeda. The participants of this conversation are “both going between different linguistic structures and systems and going beyond them” (Wei 2011: 1222). My analysis reveals linguistic similarities as well as differences between the sign languages of Chicán and Cepeda and explores which semiotic resources signers employ in order to bridge communicative gaps. “Translanguaging spaces” (Wei 2011) are multi-layered and participants are located at variable distances to each other, according to the amount of overlap between their semiotic repertoires. Adopting the framework of translanguaging allows us to look at interactions between and beyond different languages and modalities as creative processes that cannot and should not be squeezed into predetermined categories. From this angle, sign classifications on a continuum from “home sign” to “mature/fully-fledged sign languages” (Section 2.2) become obsolete as the focus lies on language as a fluid, contingent practice instead of a homogenous entity with clear borders. At the same time, the question “Is YMSL one language or several languages?” (see Section 7) becomes less pressing. By breaking up boundaries between different languages, varieties and modalities, Yucatec Maya signing communities are portrayed as a “multi-layered network of different villages, families, generations and overlapping deaf and hearing spaces, where translanguaging takes place” (Safar 2019: 31). Tying in with this study, there are numerous avenues that could be pursued further in future research. Given that translanguaging is “the norm rather than the exception” in Yucatec Maya signing communities, there is a huge variety of communicative arrangements one could look at in more detail: YMSL signers from different families, communication between women and men, between adults and children, in groups with predominantly hearing or predominantly deaf participants, etc. While Study I of my dissertation gives a broad overview of the kind of “translanguaging spaces” that exist in YMSL communities, such a micro-level approach could give a better understanding of the factors that are at play in individual constellations. Another question that deserves further analysis concerns the language ideologies surrounding the concepts of “gesture” and “sign” among deaf and hearing Yucatec Maya (see Kusters & Sahasrabudhe 2018). My analysis of translanguaging practices but also the gesture- sign comparisons in Studies II, III and IV, show that there is substantial overlap between

52 Yucatec Maya gestures and YMSL signs. No matter whether we think of the difference between gesture and sign as categorical (e.g. Emmorey 1999) or gradual (e.g. Kendon 2008), our understanding of “gesture” and “sign” is based on academic ideologies. It would be informative to find out how signers and speakers in Yucatec Maya communities conceptualise the notions of “gesture” and “sign” within their linguistic practices. A further point that could be explored in more depth concerns “sensory asymmetries” (De Meulder et al. 2019) between deaf and hearing people and how they affect translanguaging practices in Yucatec Maya villages. This could help us determine whether it is true that in these communities “everyone has equal access to social expression and understanding” (MacDougall 2012: 202) or “to what extent deaf people [feel] disadvantaged through lacking auditory access to spoken languages and the cultural activities they engendered” (Kusters 2010: 7).

6.2 Numeral variation in Yucatec Maya Sign Languages25

The topic of Study II is cardinal numbers in YMSLs and their linguistic and sociolinguistic variation, as well as their relation to number gestures used by hearing Yucatec Maya. The study includes data collected from elicitation, natural conversations and interviews from a total of 27 YMSL signers in three communities (17 from Chicán, 6 from Nohkop, 4 from Cepeda) as well as hearing non-signing speakers of Yucatec Maya (see Table 4). Based on fieldwork observations of Yucatec Maya number gestures, we found that several features of these gestures are preserved in YMSLs, namely the use of number-for- number iconicity (Taub 2001) and an opposition of cardinal numbers (starting with the index for ONE) and ordinal numbers (starting with the pinky for ONE). Moreover, certain lexical numerals (FIFTY(half)) and interrogatives related to quantities (HOW-MUCH/HOW-MANY) have their origins in number gestures. Numeral signs in YMSLs, however, go beyond their common gestural and iconic roots. We found that distinct number paradigms have evolved in three historically unrelated YMSL communities and that these signed numerals are subject to systematic variation as a result of linguistic and sociolinguistic factors.

25 The published article in Sign Language Studies (SLS) includes a link to video versions of the figures provided in the paper. Unfortunately, at the time of writing, that link to the online archive is no longer working. The supplementary files are now accessible on https://osf.io/rj7k9/ (accessed 28 September, 2020). A description of each video file is provided in the Appendix “Supplementary video materials” of the article in SLS.

53 Previous studies on other sign languages showed that the domain of cardinal numbers is characterised by a high degree of variation, related to age, region, education and family background of the signers (e.g. McKee et al. 2011, for NZSL; Stamp et al. 2014, for BSL). Moreover, numerals are often affected by rapid language change as a result of transformations in Deaf education and/or contact with other sign languages or varieties. The study of YMSL numerals offers a window to observe variation in young, micro-community sign languages that are not exposed to any normative forces from institutions, e.g. Deaf schools, dictionaries, etc. Sign languages of the world make use of a range of strategies to express cardinal numbers, for instance lexical numerals, digital numerals, additive, multiplicative or subtractive strategies, simultaneous morphology or spatial modification (Sagara & Zeshan 2016; Zeshan et al. 2013). Zeshan et al. (2013) identified the existence of a typologically unusual system in YMSL of Chicán that includes additive numerals with 20 and 50 as numeral bases, such as in

FIFTY^TWENTY^TEN for ‘eighty’.26 This strategy was previously unattested in urban sign languages and it is neither related to the number systems of the surrounding spoken languages Yucatec Maya and Spanish, nor to Mexican Sign Language (LSM). This is an example how village sign languages can add new insights to cross-modal and intra-modal typology (Zeshan et al. 2013; Sagara & Zeshan 2016). Taking a closer look, however, we find that the numeral system used in Chicán is far from homogenous but is affected by a substantial amount of variation. Analysing data from deaf and hearing signers from different families and age groups, we encountered a number of strategies absent from previous descriptions of Chicán’s number signs, namely digital numerals, simultaneous morphology along with non-manual intensifiers, modification in place of articulation and size/manner of movement, context-dependent numerals as well as writing/tracing strategies. Furthermore, we found evidence for the existence of familylects in the number domain, i.e. signing varieties that are linked to an extended family of signers (see Section 7.3.2.1). While most signers in Chicán use additive numbers, three deaf siblings under 30 years of age employ digital numbers instead. Moreover, the lexical numeral for

HUNDRED(deer) (described by Zeshan et al. 2013) is restricted to two closely related interactional groups; signers from other families in Chicán do not use it and even formulate strong attitudes against it.

26 In Study II, we adapted the annotation conventions applied by Zeshan et al. (2013), using the ^ symbol for the sequential articulation of numbers such as in FIFTY^TWENTY^TEN.

54 Apart from these sociolinguistic factors, variation in YMSL number signs is also caused by linguistic constraints, such as the use of distinct number variants (with varying handshape and/or palm orientation) for THREE and TEN in compounds versus in isolation or the phonological assimilation of FIFTY(half) to the handshape of the previous sign. Our study provides the first description of the numeral systems of the sign languages in Nohkop and Cepeda Peraza. In these communities, signers commonly use additive strategies for numbers up to 19 and employ digital numerals, writing/tracing strategies and/or context- dependent numerals for higher cardinal numbers. The comparison of numerals in Chicán, Nohkop and Cepeda shows that some strategies are employed across all three communities, while others are specific to YMSL of Chicán, and not used nor understood by signers from other communities (see also Study I). The combination of various strategies for number expression is also permitted and indeed common. One important finding of our study is that signed numeral systems are shaped by the level of literacy of the signers as well as by the point in time and the material culture in which a sign language emerges. Participants from Nohkop and Cepeda are all under 35 and are first- generation users of their sign languages. The increased accessibility of written numbers among the younger generation through formal schooling and access to smartphones and (social) media makes the adaptation of writing-based number strategies, i.e. digital or writing/tracing strategies, a likely choice. Deaf signers from Chicán belonging to the age group over forty received little or no formal education and had much less access to writing. They came to develop a system completely independent from written numbers, which they passed on to the younger generation.27

27 Spaepen et al. (2013) argue that homesigners in Nicaragua, who never received input in any conventional sign language, developed representations of numbers that are derived from number gestures used by their hearing interlocutors but that these signed numbers do not seem to function as single units in their short-term memory. Instead, they behave like indexes for individual members of a set, i.e. four fingers are perceived and memorised as one+one+one+one instead of a set of four. Even though we have not performed a recall task similar to Spaepen et al. (2013), I strongly doubt that this is also the case for YMSL signers. The results from our elicitation task show that YMSL signers are able to recognise and represent much higher numerals than the ones tested in Spaepen et al.’s study (which investigated number gestures up to nine). In Yucatec Maya communities, money constitutes the main domain to talk about exact quantities of higher cardinal numbers. We found, however, plenty of evidence from YMSL conversational data that the numeral signs described in Study II are also present in other domains, such as age, distance, weight, etc. I thus feel confident to say that YMSL numerals do indeed represent cardinal values and that, in this respect, YMSLs differ from the Nicaraguan homesign systems investigated by Spaepen et al. (2013).

55 Due to their small community size, it is not possible to conduct large statistical analyses based on a socially balanced sample for sign languages such as YMSLs, but the careful qualitative analysis demonstrates the emergence of sociolinguistic variation on a micro-level. Although a high level of individual variation persists, we are able to observe incipient patterns of sociolinguistic variation even during an early stage of language evolution. Linguistic variation indexes signers’ social identities and goes hand in hand with their language attitudes regarding the choice of a variant over another. Our study also raises methodological issues related to the question of how much variation to include in language description and typology. We argue that it is not sufficient to cherry-pick only the most “exotic”, typologically unique features but that language variation on the micro-level needs to be accounted for in order to get an accurate picture of a young, shared sign language. An avenue to pursue in further research would be to include more data from narratives and natural conversations in order to identify further possible sources of linguistic variation in the number domain and to analyse how numerals interact with other linguistic structures. Also, it would be informative to expand the data set for all participants with data from other lexical domains that typically exhibit a high degree of variation, such as colours or weekdays, and analyse whether we find similar sociolinguistic patterns of variation. A study of colour terms could also inform us how YMSLs fit into a universal colour hierarchy (Berlin & Kay 1969; see also Section 7.3.1.1) and the analysis of calendrical terms (weekdays, months, etc.) could reveal the influence of Yucatec Maya cultural routines on the YMSL lexicon as well as the impact of language contact with LSM.

6.3 Strategies of the noun-verb distinction in Yucatec Maya Sign Languages

In Study III, we deal with the question of how YMSL signers convey a distinction between objects and actions and discuss whether these linguistic strategies more generally discern nouns from verbs. We look at the influence of iconic co-speech gestures of hearing Yucatec Maya on the emergence of YMSLs, examine general cognitive processes that shape the linguistic representation of objects and related actions and describe paths of grammaticalisation from gesture to sign. Moreover, we take individual and sociolinguistic variation in YMSLs into consideration. The study of strategies of the noun-verb distinction in YMSLs and their comparison with previous findings on other urban and village sign languages

56 provides new insights into the emergence and conventionalisation of part-of-speech categories in young, micro-community sign languages. In many sign languages, nouns and semantically related verbs can look identical or very similar, for instance the ASL signs for the object ‘an iron’ and the activity ‘to iron’ (see Supalla & Newport 1978). These signs are formationally related as they share the same hand configuration and place of articulation. However, previous studies showed that sign languages make use of a range of different strategies to distinguish nouns from verbs, such as the frequency/number of repetitions of a sign, its duration, size, manner of movement or the presence or absence of mouthing (e.g. Kimmelman 2009; Johnston 2001; Hunger 2006; see Tkachman & Sandler 2013 for a detailed literature review). Certain tendencies of distinguishing objects and actions can also be observed in gestures of hearing non-signers. Some authors have suggested that deaf signers and hearing gesturers have shared cognitive biases concerning the distinction of objects and actions, which leads to similar iconic representations of objects versus actions in sign and gesture (e.g. Padden et al. 2015; Ortega & Özyürek 2016). These abovementioned studies are based on silent gestures that hearing people produced during elicitation tasks and not on gestures that hearing people use spontaneously along with speech. If we aim to find out whether strategies of the noun-verb distinction in sign languages have their roots in hearing people’s gestures, we need to analyse whether and how these gestures occur in natural discourse, along with spoken language. Study III takes a first step in this direction. Study III is divided into two parts, in which we investigate two strategies of the noun- verb distinction identified in previous literature. Sub-Study 1 analyses the use of size-and-shape specifiers (SASSes) to mark nouns and sub-Study 2 focuses on the distribution of iconic handshapes – also called “patterned iconicity” (Padden et al. 2013, Padden et al. 2015) – for objects and actions in the domain of tools. The use of SASSes as nominal markers, which is the topic of sub-Study 1, has been described by Tkachman and Sandler (2013). They encountered this strategy among signers of two young, unrelated sign languages, namely Israeli Sign Language (ISL) and Al-Sayyid Bedouin Sign Language (ABSL), although the two languages employ it to a different extent and exhibit distinct patterns of variation. The verb is expressed as an unmarked form, whereas for the related noun, a classifier depicting salient visual characteristics of the referent (such as

‘round’, ‘wide’, ‘tall’, ‘thin’, etc.) is attached. An example from YMSLs is the sign PUT-ON-

57 LIPSTICK, which refers to the action ‘to put on lipstick’, while the sign PUT-ON- 28 LIPSTICK+SASSSMALL-OBJECT refers to the corresponding object ‘a lipstick’. For the YMSL study about SASSes, we interviewed 22 deaf and hearing YMSL signers from Chicán (n=15), Nohkop (n=6) and Cepeda Peraza (n=1). Similarly to Tkachman and Sandler’s (2013) study, we conducted a picture elicitation task, in which participants saw two sets of stimuli: one depicting objects (tools, household items, containers and other objects of common use) and the other one depicting the same objects in use by a human actor. For each participant, 22 object-action pairs were analysed. We examined whether or not participants employed a SASS, what type of SASS they used and in what position the SASS occurred in relation to the base sign. The elicited data was paralleled with data from four Yucatec Maya hearing non-signers from the village Kopchen, who were asked to describe the same stimuli in silent gestures without speaking. Additionally, we looked at the use of size-and-shape gestures in narratives and interviews in spoken Yucatec Maya (see Table 5). Overall, SASSes in YMSLs exhibited little inter-signer variation in form and displayed strong similarities to size-and-shape depictions that hearing speakers of Yucatec Maya produced in the elicitation task as well as in spontaneous talk. It is remarkable that in the elicitation task, the hearing non-signers often used similar constructions as YMSL signers, where they represented an object by combining a size-and-shape gesture and a gesture imitating a motoric activity. However, while size-and-shape gestures often depict geometric properties of a referent in a detailed way, we found SASSes in YMSLs to be more reduced and schematic. As for frequency, 46% of the YMSL signs for objects analysed in our study included a SASS and 54% did not. SASSes occurred exclusively in conjunction with objects and never with a sign denoting an action. We did not encounter any significant difference between signers from different villages regarding their frequency of SASS use. It is noteworthy that YMSL signers used SASSes to distinguish objects from actions much more frequently than signers of ISL (24%) or ABSL (15%). Apart from the elicitation task, an analysis of conversational data showed that the addition of a SASS to a YMSL sign denoting an action changes its syntactic position and its co-occurrence with other parts-of-speech. These findings indicate that SASSes indeed fulfil a broader grammatical function of the noun-verb distinction. However, the use of SASSes as nominal markers is not obligatory in YMSLs and we identified a number of

28 In the Studies III and IV, I use the + symbol for glossing constructions that involve a SASS

(e.g. PUT-ON-LIPSTICK+SASSSMALL OBJECT) or a human-height specifier (e.g. MALE+SPEC-H:HUM), following the conventions used in similar studies, e.g. by Tkachman and Sandler (2013) or Tkachman and Meir (2018).

58 alternative strategies to distinguish nouns from verbs. Interestingly, we found that certain referents were marked significantly more often with a SASS than others. This corresponds to what Tkachman and Sandler (2013) found for ISL and ABSL, but so far, we can only hypothesise about which factors influence this variation. Moreover, we encountered significant variation in the use of SASSes between individual signers. Signers from two interactional groups in Chicán, younger signers as well as hearing bilinguals used more SASSes than other participants. We conclude that SASSes as nominal markers are part of the familylect of one extended family in Chicán and that language transmission from deaf adults and/or bilingualism are variables that possibly influence the use of SASSes for the noun-verb distinction. As for the position of the SASS in relation to the base sign, we found that in YMSL of Chicán – similarly to ISL and ABSL29 – the SASS has a strong tendency to appear in a final position

(such as in WASH-CLOTHES+SASSWIDE-OBJECT‘washing board’) but that in Nohkop, its position is more flexible. To sum up, our study demonstrates that SASS signs used by YMSL signers strongly resemble size-and-shape gestures that hearing speakers of Yucatec Maya produce but that in YMSLs, SASSes obtain a new grammatical function, namely that of the noun-verb distinction. Based on our observations of size-and-shape gestures in spontaneous language use in Yucatec Maya, I suggest that deaf YMSL signers pick up conventional size-and-shape gestures that they see being used by hearing people and incorporate them into their sign languages in order to convey a linguistic distinction between part-of-speech categories. The gestures are reduced in form and in YMSLs, SASSes serve the purpose of a broader classification of a referent rather than an iconic description. Once they become part of YMSLs, SASSes obey linguistic constraints (such as the position of the SASS in relation to the base sign, constituent order, co- occurrence with other parts-of-speech) and are subject to sociolinguistic variation, influenced by family membership, age or hearing status of the signer. Sub-Study 2 builds on previous work by Padden et al. (2013) who described two main representation techniques to depict the use of tools and coined the term “patterned iconicity”. The first one is called a handling strategy, in which the signer’s hand represents a hand manipulating an object. For instance, a fist depicts a hand holding an (invisible) toothbrush. The other alternative is an instrument strategy, where the signer’s hand represents the object

29 Note that Tkachman and Sandler (2013: 269) observed in their study that the SASS occurs in a final position in both ISL and ABSL, while Tkachman and Meir (2018) found that ABSL had a preference for a SASS-final order but that in ISL, there was no significant difference between SASS-final and SASS-initial constructions.

59 itself, e.g. an extended index finger for a toothbrush. Both types of handshapes can be combined with a movement that imitates the action of brushing one’s teeth. Padden et al. (2013) found that signers and hearing gesturers use both strategies, but while gesturers generally prefer handling strategies, there are cross-linguistic differences between sign languages in their choice of iconic patterns. In a follow-up study, Padden et al. (2015) compared gesturers’ and signers’ use of iconic handshapes for objects and actions carried out with these objects. They concluded that hearing non-signers have a general preference for handling depictions, but ASL signers employ the handling/instrument opposition to convey a distinction between grammatical categories, namely handling for verbs and instrument for nouns. Sub-Study 2 of this paper, “Patterned iconicity in YMSLs”, is further divided into two sub-studies. Sub-Study 2a compares the use of handling versus instrument handshapes for the depiction of objects (tools) among YMSL signers from three communities (12 from Chicán, 7 from Nohkop, 2 from Trascorral) with hearing non-signing gesturers (11 participants from Kopchen). Signers as well as gesturers were shown the same set of photo stimuli used by Padden et al. (2013). It consists of photos of 24 familiar hand-held tools such as a handsaw, a toothbrush, a knife, a screwdriver, etc. The comparison with Padden et al.’s (2013) findings revealed some interesting differences between the sign languages examined in their study and YMSLs. First, the split between Yucatec Maya gesturers and YMSL signers is not as drastic as in the previous study, and second, the preference for one strategy over the other is not as pronounced in any of the participant groups as it is in Padden et al.’s study. While Yucatec Maya non-signers used handling and instrument handshapes to equal degrees in their silent gestures, YMSL signers had a slight preference for instrument overall, but this finding proved to be significant only for the group of Chicán signers. This could possibly be a result of language age and intergenerational transmission, due to the presence of multiple generations of signers in Chicán. In Nohkop, we detected a slight difference between hearing signers (preference for handling) and deaf signers (preference for instrument), but it proved not to be statistically significant. More striking than comparisons between participant groups, however, is the link between the choice of a particular strategy and a particular type of object. Certain objects are more likely to be represented by a handling strategy, others by an instrument strategy, while some objects can trigger either handling or instrument. This tendency is irrespective of whether the participant is hearing or deaf or which village she or he comes from. It seems that we are witnessing a more general cognitive phenomenon, but the correlation between a specific strategy and a specific object is more solid among YMSL signers than among hearing non-

60 signers. Some of the patterns of variation we found overlap with Padden et al.’s (2013) findings, while others seem to be specific to YMSLs.30 It remains an open question to what extent properties of the referent influence the choice of iconic strategy and what role conventional emblematic gestures play in this process. We suggest that iconic patterning in Yucatec Maya gestures and YMSL signs related to tool use is shaped by an interplay of characteristics of the referent and cultural conventions. In sub-Study 2b, we then move on to compare the choice of iconic strategy for objects and related actions. We included data from 30 participants (11 from Kopchen, 12 from Chicán, 7 from Nohkop), who saw two sets of photo stimuli depicting objects and corresponding actions carried out with these objects (comparable to the stimuli used in sub-Study 1). These materials were adapted from Padden et al. (2013), but they depicted Yucatec Maya actors and culturally typical objects. A subset of 12 object-action pairs from each participant was selected for our analysis. Importantly, no significant difference in choice of iconic strategy for depicting objects versus actions could be detected in YMSLs or Yucatec Maya silent gesture. The distribution of iconic handshapes was very similar across all groups: Signers from both communities as well as hearing non-signers showed a clear preference for instrument over handling for both objects and actions for the subset of 12 stimuli. As in sub-Study 2a, this tendency was most pronounced among signers from Chicán. Our findings show that in YMSLs and in Yucatec Maya silent gesture, the type of object is a more reliable predictor for the choice of a handling versus an instrument strategy than the presence or absence of human agency. The different frequency distribution of instrument versus handling in the results from Padden et al.’s stimulus set (sub-Study 2a) and the Yucatec Maya stimulus set (sub-Study 2b) can be explained by a bias towards objects that are generally represented by instrument handshapes in the stimulus set for sub-Study 2b. This underlines the importance of carefully selected and pilot-tested stimuli for elicitation tasks. Sub-Studies 2a and 2b demonstrate that patterns of iconicity exist in both gesture and sign. While there is little individual variation in the choice of instrument versus handling handshapes between participants there is a lot of variation related to the type of referent. So far, there is no evidence that YMSL signers use the handling/instrument opposition for the grammatical purpose of the noun-verb distinction. We can only speculate why YMSLs differ from ASL in this respect. There might be cross-linguistic differences in regard to the devices

30 Padden et al. (2013) also note that certain objects are more likely to elicit instrument forms than others, but their explanations for the variation in patterned iconicity related to the type of object remain rather superficial.

61 that signers recruit for the noun-verb distinction. YMSLs being young micro-community sign languages and ASL being an older, macro-community sign language, the degree of conventionalisation of strategies of the noun-verb distinction might also be a result of language age. Finally, we cannot exclude the possibility that our divergent findings are due to methodological reasons (choice of stimuli, picture vs. video condition, randomised vs. fixed order of stimuli, etc.). To conclude, Study III of my dissertation shows that two different kinds of iconic gestures have been adapted into YMSLs in different ways and they have been conventionalised to a different extent. SASSes as nominal markers in YMSLs constitute a case of grammaticalisation from co-speech gestures (see Section 8). For iconic handshapes, there seems to be an incipient preference for instrument over handling in YMSLs, but this tendency was conventionalised to a different extent in the three villages included in this study. The sociolinguistic makeup of the communities (number of generations of deaf signers, age of the signers, distribution of deafness across one vs. multiple families) is likely to play a crucial role here. It is interesting to look at the type of variation encountered in the two studies. Individual variation between participants was high in sub-Study 1 and low in sub-Study 2. This indicates that SASSes as nominal markers are an already conventionalised linguistic structure that shows patterned sociolinguistic variation while in sub-Study 2, more general, cognitive and cultural factors seem to be at play. The data from non-signing participants show that hearing speakers of Yucatec Maya are already remarkably systematic and consistent in their gesture production, which leads to striking similarities between unrelated YMSL communities in both studies. Our findings demonstrate that even in young, micro-community sign languages, the seeds for multiple strategies of the noun-verb distinction are present. They are often rooted in hearing people’s co-speech gestures and different factors (community size, generational depth, number of deaf signers, networks of interaction) spur their conventionalisation into the signed linguistic system. Looking back, the most important conceptual problem I encountered in Study III is the attempt to draw a generalisation from a) an object-action distinction to a noun-verb distinction and b) from the semantic category of tools to the broader sign language lexicon. This issue is not unique to my study but also applies to other previous studies on these topics (e.g. Tkachman

62 & Sandler 2013; Padden et al. 2015).31 A part of the dilemma stems from an analysis of mainly elicited data instead of a corpus-based study. Although we did cross-check our results with data from natural conversations in YMSLs, the findings of this study can only be generalised to a limited extent. In retrospect, if I were to reproduce this study, I would start off with more conversational data in order to investigate YMSL nouns and verbs in context (see Section 5.2). In order to determine how Yucatec Maya gesturers’ depictions of size and shape become increasingly schematised and reduced in YMSLs, future research could provide a more fine-grained comparison of gesturers’ and signers’ depictions of objects. This could be done by looking at parameters like handshape complexity and finger selection (see e.g. Brentari et al. 2012) or by comparing in detail the types of iconic devices used by gesturers and signers, such as shape-for-shape or distance-for-size depiction (Nyst 2016b). Another issue of interest in relation to Study IV (see the following Section 6.4) would be to look at eye gaze behaviour in co-occurrence with SASSes as nominal markers in order to see whether it resembles the patterns found with height-specifiers as person-classifiers.

6.4 “When you were that little…” – From Yucatec Maya height-specifier gestures to Yucatec Maya Sign Language person-classifier signs

While in Study III, I explored more generally the relation between iconic gestures of hearing Yucatec Maya and YMSL signs, I decided to focus on one specific conventional gesture in Study IV and looked at the different ways how this form was incorporated into YMSLs. Such a micro-analysis offers an opportunity to investigate in more depth the differences in form, meaning and distribution between gestures and signs and explore possible paths of lexicalisation and grammaticalisation. To date, systematic comparisons of gestures and signs within the same language community are scarce (but see Mesh & Hou 2018; Mesh 2017; Tano & Nyst 2018; Nyst 2016a; Fenlon et al. 2019). Shared signing communities where deaf and hearing people are in close interaction and co-create a sign language offer an excellent opportunity to study the gesture-sign interface in detail and provide new insights into language emergence and evolution. Study IV contributes to the existing body of literature by providing a detailed analysis of “human-height specifiers” in Yucatec Maya and YMSLs. Previous work (Foster 1948; Meo- Zilio & Mejía 1980; Zavala 2000; Fox Tree 2009; Le Guen et al., in press-a) has shown that

31 I wish to thank Irit Meir for pointing this out in her review of our paper and for being upright enough to not exempt her own work from this criticism.

63 speakers of many Mesoamerican languages, including Yucatec Maya, use a set of conventional gestures that ascribe a referent to a particular semantic class. They vary handshape and/or orientation of the hand according to the type of entity they are talking about, for instance a human being, a four-legged animal, a bird, a plant, etc. The size of the inventory of such “gestural classifiers” and their specific forms vary cross-linguistically, but it seems to be a widespread areal phenomenon in Mesoamerica. Many of these classificatory gestures iconically depict the height of the respective referent; I thus use the term “height-specifiers”. As these height-specifier gestures occur with high frequency in spontaneous talk and their form is consistent across gesturers, they represent ideal candidates for being a source for lexical and grammatical items in emerging sign languages. Indeed, it has been shown that shared sign languages that develop in the geographic region of Mesoamerica often incorporate such conventional classificatory gestures into their linguistic systems (see e.g. Fox Tree 2009; Haviland 2013; Hou 2018; Horton, in press-a). In Study IV, I focus on a conventional gesture used by speakers of Yucatec Maya, performed by a flat B- or 5-handshape,32 with the palm facing downwards. It depicts the height of an upright, typically human referent, from the top of the head to the ground. The Yucatec Maya gesture can also refer to non-human upright entities, such as plants or other tall objects. We observed the same manual form in YMSLs from all four communities. The aim of this study was to find out which features of the original Yucatec Maya gesture are preserved in YMSLs, if YMSL signers attribute new meanings and functions to the gesture and how the height-specifier is integrated into the morpho-syntactic structure of the sign languages. Furthermore, I looked at variation in the use of the human height-specifier between YMSL signers from different communities and between individuals from the same community. To this end, I collected data from two groups of participants: a total of 34 deaf and hearing YMSL signers from Chicán (n=21), Nohkop (n=5), Trascorral (n=3) and Cepeda (n=5) as well as eight hearing gesturers from the village Kopchen and the small town Chemax. The participants in the gesture group do not know any sign language. For my analysis, I combined different types of data (see Table 7): I conducted a picture elicitation task, using photo stimuli showing people of various ages, gender and professions. YMSL signers were instructed to describe the stimuli in sign language and hearing gesturers were asked to explain what they see without speaking. The elicited data was complemented with short narratives and spontaneous

32 See Appendix B for a description of handshape names used in my dissertation.

64 conversations (in dyads and groups) in YMSLs as well as narratives, conversations and short semi-structured interviews in spoken Yucatec Maya (for the gesture group). The study revealed interesting parallels as well as differences between height-specifier gestures and signs. In spoken Yucatec Maya, the gesture is typically accompanied and temporally aligned with the verbal deictics beya ‘like that’ or buka’aj/beytak ‘of that size’. These direct the interlocutor’s attention to the fact that relevant information is being conveyed by the manual channel. The gesture adds semantic content to the spoken channel by providing a schematic depiction of the height of the referent in question – information that often remains vague or unsaid in speech alone. When the height-specifier stands for a human referent, usually a child, it often represents not only his/her physical measurements but also metaphorically stands for his/her age. In this way, the gesture can function as a time anchor to specify when a certain event occurred, in an utterance such as “When you were that little, we moved to another house.” Typically, the gesture is accompanied by the speaker’s eye gaze to his or her hand.

What about YMSLs? The human-height specifier, glossed SPECIFIER-HEIGHT:HUMAN, or SPEC-H:HUM for short, is a high-frequency sign in YMSLs from different communities. YMSL signers preserved some of its original features but mapped new meanings and functions onto the co-speech gesture. When employed as a predicate in YMSLs, form and function of

SPEC-H:HUM are identical to its use in gesture. It iconically depicts the height – and metaphorically, the age – of a specific referent and the signer’s eye gaze is directed at his or her hand. Furthermore, SPEC-H:HUM was lexicalised as a noun in YMSLs from all four communities. In its lexical use, SPEC-H:HUM exclusively refers to human referents and no longer to upright entities more generally. The most common meanings are CHILD,33

DAUGHTER/SON or YOUNGER-SIBLING (in all four communities), but in Nohkop and Cepeda,

SPEC-H:HUM can also refer to a member of an older generation, namely FATHER or (GRAND-

)MOTHER. When used as a noun, the YMSL sign still shares the same manual form with the Yucatec Maya gesture, but its height of articulation is generic and does not necessarily reflect the referent’s actual height. The eye gaze is detached from the hand and instead directed at the interlocutor or a neutral location. SPEC-H:HUM also exists as a plural form, expressed through

33 Note that many other sign languages of the world use the same form for CHILD as in YMSLs, for instance Sign Language of the Netherlands (NGT) (Nyst 2018). Interestingly, Nyst (2018) shows that Adamorobe Sign Language (AdaSL) uses a completely different sign for CHILD, in which the lower arm and fist embody a person. She demonstrates that the lexicon of a sign language is shaped by a general preference for a certain type of iconic depiction, i.e. embodied shape-depictions in the case of AdaSL. These language-specific preferences are themselves influenced by gestural patterns of the surrounding hearing communities.

65 rapid repeated articulation on a horizontal axis, meaning CHILDREN. When articulated repeatedly on a vertical axis, the human-height specifier becomes a verb for GROW. A further possible function of SPEC-H:HUM in YMSLs is the one of a spatial referential device: SPEC-

H:HUM can be used to locate discourse referents in the signing space and in this case, it interacts with other grammatical structures such as pointing and/or directional verbs.

The most notable example of how SPEC-H:HUM has been adapted and integrated into the linguistic structure of YMSLs is that it can function as a person-classifier, by attaching it to a noun indicating gender (FEMALE or MALE) or age (OLD-PERSON or INFANT) or a combination of both gender and age. In this case, SPEC-H:HUM no longer refers to a concrete person but serves more generally to ascribe the referent to the semantic class of human beings. The sign is articulated at neutral height and its descriptive, height-depicting function is lost. During the articulation of the person-classifier, the signer’s eye gaze is directed at the interlocutor or a neutral location instead of the hand. In these constructions, SPEC-H:HUM typically stands in a final position, following the age or gender sign, such as FEMALE+SPEC-H:HUM. Among younger signers in Chicán, SPEC-H:HUM also occurs as an agentive marker in compounds denoting professions, such as FEMALE+SPEC-H:HUM+WRITE ‘teacher’. Finally, I documented the use of

SPEC-H:HUM as a part of name signs in Chicán, such as JOSÉ+FEMALE+SPEC-H:HUM ‘Geli (lit.: José’s daughter)’. My findings demonstrate that the integration of conventional gestures into sign languages is not a simple and linear process. Height-specifiers have become incorporated into YMSLs in various ways and there are parallels as well as differences between the four YMSL communities. YMSL height-specifier signs have taken on a range of new functions, some of them close to the corresponding Yucatec Maya gesture, some further removed. Study IV makes two main contributions. First, it emphasises that the comparison of co- speech gestures and signs in terms of their form, meaning and function can be a complex endeavour. The analysis of height-specifiers shows that, even though gestures and signs are often very similar in form and meaning, viewing the transfer from gestures to signs as “wholesale borrowing” (i.e. a one-to-one correspondence between a gesture and a sign) is not accurate (Mesh & Hou 2018; see also Mesh 2017). Gesturers and signers in shared signing communities have the same cognitive disposition and a common cultural foundation. Deaf signers can visually access the conventional co-speech gestures used by hearing people and readily incorporate these forms as “raw material” into their sign languages. There is, however, a crucial difference: hearing people use gestures along with speech as part of “composite utterances” (Enfield 2009), where the visual and the auditory modality complement each other.

66 The visual-manual channel being the sole mode of expression that is fully accessible to them, deaf signers often alter or extend the meaning of hearing people’s gestures (Mesh & Hou 2018; Mesh 2017; Coppola, in press). In the case of the height-specifiers analysed in this study, Yucatec Maya gestures and signs share a core meaning, namely ‘the height of an upright entity’. In YMSLs, additional meanings emerged, while the class of the entity becomes more constrained (from upright entities in general in gesture, to exclusively human referents in sign). Also, we encounter subtle differences in form (place of articulation, eye gaze) between gestures and signs. This underlines the fact that in gesture-sign comparisons it is crucial to consider not only the most salient, manual components but to look at multiple articulatory features of gestures and signs in context. The second important outcome of this study is that it prompts us to revise suggested paths of lexicalisation and grammaticalisation from gestures to signs (e.g. Wilcox 2005; Wilcox et al. 2010; Janzen 2012). The analysis of height-specifiers in YMSLs showed that they entered both lexicon and grammar simultaneously, which points to a dynamic view of lexicalisation and grammaticalisation as intertwined processes, rather than a linear view in which lexicalisation precedes grammaticalisation (see Section 8.2.3, for discussion). One specific gestural form became multifunctional in signed languages and coexists as a lexical as well as a grammatical device.34 As a result of lexicalisation, SPEC-H:HUM acquires multiple meanings in YMSLs (CHILD, DAUGHTER/SON, YOUNGER-SIBLING, FATHER, (GRAND-)MOTHER,

CHILDREN, GROW). Undergoing a path of grammaticalisation, SPEC-H:HUM became a person- classifier in YMSLs. A process of “semantic generalisation” (Bybee et al. 1994) took place from specific human referents to human beings as a class. SPEC-H:HUM was integrated into the morphosyntactic structure of YMSLs and interacts with other linguistic features, such as eye gaze and other non-manual features, organisation of the signing space, sign/constituent order, or movement (e.g. in the verb GROW). The syntactic structure of the person-classifier standing in a final position (following an age/gender sign) parallels observations on SASSes as nominal markers, e.g. CHOP+SASSLONG-OBJECT ‘machete’ (Study III) and constructions involving a height-

34 Janzen (2012) and Janzen & Shaffer (2002) describe similar instances of multifunctionality in the grammaticalisation path of the ASL future marker FUTURE. In their analyses, they reconstruct these changes as being the result of a diachronic process where distinct functions of a form arise in a particular order (gesture < lexical item < grammatical item). In YMSLs, the SPEC-H:HUM is polyfunctional already among the first generation of signers and we can only speculate which function came first and whether or not some functions will disappear over time. For these reasons, I suggest that a dynamic view of lexicalisation and grammaticalisation is much more adequate to capture the synchronic process of conventionalisation of co-speech gestures into shared sign languages. See Section 8.2.3 for more details.

67 specifier for four-legged animals, e.g. HORSE+SPEC-H:4LEG ‘horse’ (see Section 8.2.2, for a more detailed comparison of height-specifiers and SASSes). Both the SASS as well as the height-specifier for four-legged animals also tend to occur in a final position, which indicates that we might be observing the emergence of a consistent constituent order for YMSL constructions involving a size-and-shape/height-specifier (see Tkachman & Meir 2018, on the conventionalisation of SASS-constructions in ISL and ABSL). Further research on YMSLs’ morpho-syntax is needed to test this hypothesis. In possible follow-up studies, I would like to adopt similar micro-analyses of specific forms used by gesturers and signers to other linguistic domains, for instance manual and non- manual negation markers (see Mesh & Hou 2018) or the functions of PALM-UP in Yucatec Maya gesture and sign (see Cooperrider et al. 2018).

7 One language or several languages? The perpetual question

In the following section, I will synthesise the results of the four sub-studies presented above. These will be combined with further, yet unpublished findings from my PhD research, with observations I made during my fieldwork and with other literature on YMSLs. In the light of these results, I will reflect on whether the subject of my study is one single language with regional varieties in different communities or if rather we are dealing with different sign languages in each community.

7.1 Embedding the debate in previous literature on YMSL(s)

The question “One language or several languages?” turned out to be a central point in my dissertation, but even after seven years of working with YMSL(s?), it remains extremely difficult to answer. Ultimately, the conclusion depends on our definition of “language” and “dialect” and as we know, the boundaries between the two categories are blurred and ideologically loaded (e.g. Makoni & Pennycook 2006). In the literature on sign languages in rural Yucatán, the debate has been ongoing for a while. Some authors (e.g. Le Guen 2012; Le Guen et al., in press-a; MacDougall 2012; Johnson 1991) speak of one common (Yucatec) Maya Sign Language. Le Guen et al. (in press-a) consider that the forms of signing in different communities35 are varieties of one unified sign

35 Le Guen et al.’s analysis includes a comparison of the sign languages from Chicán and Nohkop.

68 language across the peninsula of Yucatán and argues that Yucatec Mayan gesture serves as basis that gives rise to the emergence of lexically and grammatically similar signing varieties in locations without any historical link. Fox Tree (2009) even proposes the existence of a pre-Hispanic, pan-Mayan sign language complex he calls Meemul Tziij – “maybe the oldest documentable sign language in the world” (Fox Tree 2009: 355) – across Mesoamerica, resulting in historically related sign languages in different parts of Mexico and Guatemala. There is, however, no solid evidence from historical documents or from people’s memories that could sustain this hypothesis. Fox Tree’s claims about a connection between old Mayan iconography and handshapes in contemporary Mesoamerican sign languages remain speculative. Other scholars (Escobedo Delgado 2012; Zeshan et al. 2013) insist that the sign language they studied is unique to the village Chicán and should be labelled as such (Chican Sign Language/Lengua de Señas Chicana). Shuman (1980) and Shuman and Cherry-Shuman (1981) also limited their analysis to the village of Chicán but assigned to it the pseudonym “Nohya” (at the time, the name of the village was not yet widely known in the scientific community). The four studies of my dissertation as well as other publications (Le Guen et al., in press-a; Le Guen 2012) provide ample evidence that YMSLs from different communities exhibit an important degree of overlap in their lexicon and grammar. This fact can partly be explained by their shared sociocultural background and their common gestural precursors. However, to date, we still know very little about the lexicon and grammar of the sign languages in question and systematic comparisons between communities are recent. Scholarly findings about YMSLs mirror community members’ observations, who describe YMSLs from unrelated communities as “similar but different” (see Study I). So, once more: are we dealing with one sign language or several sign languages? In some cases, when we feel stuck in answering a question, it is because previous models and terminology do not really capture the new situation we are looking at. As I argue in Study I, traditional definitions of “language” or “language variety” are not suitable to grasp the complex sociolinguistic constellation of Yucatec Maya signing communities and its network of different villages, families and generations. The circumstance that languages have no historical affiliation but resemble each other due to a common cultural, cognitive and gestural substrate does not exist in the same way for spoken languages. For sign languages, similar situations have been mentioned, e.g. by Nyst (2013), Nyst (2018), Tano and Nyst (2018) or Reed and Rumsey (2020). For instance, Nyst (2018) suggests that certain similarities in the strategies of iconic

69 depiction of various unrelated sign languages in West Africa are an areal phenomenon and result from a common gestural foundation. However, to my knowledge, no in-depth comparative studies between gestures and signs across a larger geographic region have been conducted to date. Given that the sign languages in Yucatán are historically unrelated, I have opted to use the plural form “sign languages” to account for their independent roots – but in conjunction with the unifying label “Yucatec Maya” in order to emphasise their common cultural and ethnic background.36 In order to make more well-informed claims about similarities and variation in YMSLs, we must go deeper and approach the question from both a sociolinguistic and a linguistic perspective.

7.2 YMSLs and their sociolinguistic networks

YMSLs do not conform to the prototypical case of an isolated and homogenous rural sign language with clearly defined geographic boundaries – consider, for instance, Providence Island Sign Language that emerged due to genetic deafness on a Caribbean Island (Washabaugh et al. 1978; Braithwaite, in press), Algerian Jewish Sign Language, which developed in a Jewish enclave in Algeria (Lanesman & Meir 2012) or Mardin Sign Language, used by a single extended family with several deaf members in a town in Turkey (Dikyuva 2012).37 In Yucatán, we are dealing with a more complex constellation including several villages with dissimilar numbers of generations of deaf people and loose ties between multiple families. Traditionally, people in rural Yucatán have few connections to other villages, unless they have family members who live there.38 Nevertheless, it is questionable whether designating a label to the sign language of one specific village, i.e. Chicán Sign Language (Escobedo Delgado 2012; Zeshan et al. 2013), is adequate in the Yucatec Maya case, as it evokes the image of a more or less homogenous, coherent linguistic situation in that

36 Lynn Hou and Kate Mesh, who work in a community in rural Oaxaca, Mexico (Hou 2016; Mesh 2017; Mesh & Hou 2018) found a similar solution and adopted the label “San Juan Quiahije Chatino Sign Language (SJQCSL)” for the sign language used in the municipality of San Juan Quiahije, which is home to the indigenous ethnic group of Chatinos. The admittedly tongue-twisting label SJQCSL leaves room for the possibility that other sign languages exist in other Chatino communities in the wider region. 37 Although – as I discuss in Section 2.2 – an in-depth ethnographic investigation is indispensable in order to determine how “prototypical” a constellation is. 38 Marsaja (2008) mentions a similar situation in Bali, where social life is organised around ritual activities in the communities and people have a strong sense of village identity. Thanks to Connie de Vos for the comment.

70 community. Even though Chicán is a tiny village and it takes about ten minutes to walk from one end to the other, some families with deaf members who live on opposite sides of the village actually have very little contact (due to private, political or religious differences). As I argue in Studies I, II and III, this situation results in substantial variation between the signing varieties of different extended families (“familylects”) in Chicán. Some hearing relatives of deaf people report that they do not understand deaf people from other families because “they sign differently” (see Study I). Whether this is a result of actual divergence in lexicon and grammar or rather a manifestation of language ideologies in play is an open question, but it gives us a clue about the sociolinguistic landscape of Chicán. Although I had much less time to study interactions between deaf signers from different families in Cepeda, I can state that the overall setting seems to be very similar to Chicán: some deaf signers are in regular exchange with deaf members of other families, while others only meet sporadically or by coincidence. Understanding culturally typical networks of interaction among the Yucatec Maya is crucial here. People interact primarily with members of their (extended) families rather than with peers outside of their families (Le Guen 2012; Safar & Le Guen, in press; see also Horton (in press-b) about signers in Nebaj, Guatemala, or Hou 2016, and Mesh 2017, for SJQCSL). Sometimes even neighbours living on adjacent compounds in Yucatec Maya villages communicate little with each other. In this regard, deaf people constitute no exception. Deaf signers from different families in Chicán or Cepeda sometimes meet due to a social event in the community or jointly participate in government-related activities aimed at the deaf population, but they generally do not socialise based on their shared experience of deafness. In Chicán, a group of deaf men regularly meet for a chat on early evenings (sometimes with hearing signers present, sometimes without), which shows that “deaf spaces” (Kusters 2015) do exist – but they conform to general patterns of social networks customary among the Yucatec Maya. Deaf-deaf interactions largely coincide with kinship relations and affinities between members of (extended) families. Hou (2016) analyses SJQCSL as “a constellation of family sign languages […] [of] multiple extended families in co-residence”. Different families create distinct signing varieties that overlap due to the presence of shared conventional gestures used by hearing Chatinos. Despite the many similarities between the Yucatec Maya and the Chatino situation, it is doubtful whether this is an adequate way to describe YMSLs. Although YMSL of Chicán (and presumably also of Cepeda) is, as I argue, characterised by the existence of different familylects (see Section 7.3.2.1), it exhibits, at the same time, many lexical and grammatical structures that are shared across the whole community. Some of these structures bear no obvious relation to

71 conventional gestures. Further linguistic analyses are needed in order to establish the degree of unified conventions in Chicán and to gauge how much linguistic accommodation is occurring between signers from different families.39 It is interesting, however, that in the case of YMSLs, the shared cultural and gestural elements, which Hou (2016) mentions for SJQCSL, extend beyond community borders. In some domains, we find more linguistic parallels between the signing varieties of some interactional groups in Chicán and signers from Nohkop than between interactional groups of the same village. Although historically unrelated, there are some instances of convergence between the communities that transcend pure chance and modality- specific effects (i.e. the use of iconicity) and these are linked to Yucatec Maya culture (see Study I). Next, I will move on to present the linguistic evidence I have gathered about variation between YMSLs from different communities (Section 7.3.1) as well as variation within these communities (Section 7.3.2).

7.3 What’s your sign for TORTILLA? Variation in YMSLs

Telling me about a visit from a group of deaf people from Cepeda (who were invited by the Mexican linguist Escobedo Delgado), a hearing signer in Chicán commented:

“I was impressed. Of course, some things are similar – they are yucatecos like us. But even though they don’t live far, their signs are not the same!”

In this signer’s perception, the linguistic structures of the sign languages from the two villages are distinct, but the regional and cultural affiliations of the signers form a link between them (see Study I). At this stage, as we are only scratching the surface in inter-community comparisons of YMSLs, it is premature to make a general statement about the extent of lexical and grammatical similarities. Also, we are just beginning to tease apart which instances of linguistic variation within the communities are idiosyncratic and which display systematic sociolinguistic patterns. The examination of specific lexical and grammatical domains that I carried out in the four studies of my dissertation allows us to shed more light on these questions.

39 Members of the YMSL project are currently investigating strategies of linguistic adaptation between YMSL signers in more detail.

72 7.3.1 Variation between communities

Let us first examine parallels and differences between YMSLs from different communities.

7.3.1.1 A continuum of lexical variation

During my first period of fieldwork in 2015, I conducted a lexical elicitation task, in which I showed the participants a series of photo stimuli that target different semantic fields: animals, human referents, plants and food, buildings, weather and substances, activities, objects of common use, colours, numbers, vehicles, clothing, opposites, Yucatec Maya culture/ceremonies as well as emotions. Some of these stimuli yielded more compelling results than others (see Safar, forthcoming-a, for an in-depth discussion of my methodology) and I selected a few semantic domains for a detailed variation analysis. The overall findings from the task suggest that there is a continuum of variation in different domains of the YMSL lexicon.

Figure 12: Continuum of lexical variation in YMSLs

On one end of the continuum, we find that lexical domains that are typically represented by highly iconic signs (e.g. derived from the depiction of actions) and/or based on Yucatec Maya conventional gestures exhibit little variation between and within the communities. On the other end, we see that domains that are more abstract and lack a gestural representation, e.g. cardinal numbers higher than 10, colours or weekdays, show a high degree of inter- and

73 intra-community variation.40 Generally, these results are not surprising. Even historically unrelated sign languages from different geographic regions in the world proved to show substantial overlap in their lexicon due to modality-specific effects, i.e. the use of iconicity (see e.g. Currie et al. 2002; Occhino et al. 2017). Iconicity in language is generally understood to be a perceived correspondence between a linguistic form and its meaning and being visual- spatial languages, sign languages have the potential to make use of a vast array of iconic mappings on various linguistic levels (e.g. Taub 2001; Perniss et al. 2010). For this reason, sign language researchers have argued that for lexico-statistical comparisons of sign languages the original Swadesh list (Swadesh 1955), which is traditionally used for this purpose, needs to be adapted in order to remove highly iconic signs. Otherwise, the comparative analysis would yield “false cognates” between unrelated languages (Woodward 1993). Especially signs that are directly derived from the depiction of a motoric activity, e.g. the domain of hand-held tools and actions carried out with these tools, result in highly iconic representations and exhibit very little regional and individual variation (see Study III for YMSLs; see also Padden et al. 2013, Padden et al. 2015, Cormier et al. 2012). Examples for such signs are WASH-CLOTHES, BRUSH-TEETH, CHOP-WITH-MACHETE, WRITE etc. There are, however, differences in which iconic strategy signers choose for action-based depictions, i.e. handling versus instrument handshapes. I have shown in Study III that in YMSLs, the preference of a certain iconic strategy is mainly connected to the type of object depicted and that there are strong similarities to Yucatec Maya gestures in this respect. We see, however (see sub-Study 2a of Study III), that YMSL signers have a slight overall preference for instrument over handling handshapes for the category of hand-held tools. This preference for an instrument strategy is significant among signers in Chicán but not in the other communities. These results suggest that certain sociolinguistic features of YMSL in Chicán, such as the presence of multiple generations of deaf signers and/or the larger community size, may spur the conventionalisation of patterned iconicity in the lexicon. Another group of signs that exhibit strong uniformity between communities is signs that are derived from conventional emblematic gestures used among hearing Yucatec Maya.

40 One could hypothesise that the more iconic gestures (e.g. for activities or objects) are used with more frequency by speakers of Yucatec Maya and are thus incorporated more often and more uniformly into YMSLs. However, Le Guen (2012) shows for the domain of time that gestures for “abstract” concepts are also incorporated into a sign language’s lexicon and grammar. The focus on semantic domains that can be visually represented using picture stimuli is a bias that results from the design of my study; a corpus analysis could yield a more fine- grained representation of the continuum of lexical variation in YMSLs.

74 Emblematic gestures, also called “quotable gestures” or “emblems” (Kendon 1992), exhibit a stable connection between form and meaning and are used consistently across speakers, such as the widely known “thumbs-up” gesture. For these reasons, Le Guen et al. (in press-a) claim that quotable gestures constitute a form of “proto-signs” (see also Kendon 1988). Emblems that are used either locally in Yucatán, more generally across Mexico or internationally are prime candidates to become lexicalised into YMSLs (Le Guen et al., in press-a) and the resulting lexical signs are usually identical or very similar across communities. Examples are the signs for MONEY, FINISH, RAIN, YES, COME-HERE, WAIT, DON’T-KNOW, etc. (Le Guen et al., in press- a). As I have shown in Study I, some parallels between the lexicons of different YMSL communities transcend the effect of iconicity and gesture. The reason for these similarities lies in the shared repertoire of cultural knowledge, lifestyle, rituals and beliefs across Yucatán. A good example for such a culture-specific lexical item is the sign for TORTILLA, which is the staple food in Yucatán (as well as in other regions of Mexico and Mesoamerica). In the YMSL data, I found two main variants for TORTILLA, which look fairly similar to each other but are derived from different iconic “bases” (Kendon 1980). One variant (TORTILLA1) depicts the action of passing the tortilla quickly from one hand to the other, in order to let it cool off after it is served fresh from the fire. The other variant (TORTILLA2) is based on the procedure of moulding the dough into a flat, round shape.

Figure 13: YMSL variants TORTILLA1 (left) and TORTILLA2 (right)

I encountered both variants across different communities, although TORTILLA1 seems to be overall more frequent than TORTILLA2. The sign TORTILLA can also be combined with a SASS for a round shape (see Study III/sub-Study 1). Both variants are easily recognisable even

75 to non-signing hearing people as the process of preparing and handling tortillas is so deeply entrenched in the daily life in a Yucatec Maya village. Similarly, the sign for TAMALES also refers to the customary preparation of a typical Yucatec dish, for which corn mass is stuffed with meat and wrapped into banana leaves. Other signs based on certain culture-specific concepts refer for instance, to certain properties of plants that are typical to the region (e.g. CHAYA),41 traditional ways of capturing animals (e.g. GOPHER),42 or rituals performed at Yucatec Maya celebrations (e.g. JÉETS’-

MÉEK’).43 The resulting signs are iconic, but importantly, their iconicity is only transparent to “insiders” who are familiar with these cultural elements. While an iconic relation between form and referent gives way to similar signs in unrelated sign languages, the interpretation of iconic signs lies “in the eye of the beholder” (Occhino et al. 2017). The shared cultural context can yield even more concrete similarities between YMSLs from different communities and facilitate understanding beyond community borders. In this way, translanguaging between members of YMSL communities is different from “cross-signing” (Zeshan 2015), i.e. the ad- hoc communication in encounters between deaf signers with no shared sign language: YMSL signers from different communities have no common sign language, but they do share the same cultural and gestural background (see Study I). On the other end of the spectrum are domains that are not gesturally encoded and do not lend themselves to iconic representations as easily as others. These parts of the lexicon, for instance cardinal numbers higher than ten, colours or calendrical terms, give rise to distinct variants in different communities. In sign languages, the aforementioned lexical domains typically exhibit a high degree of sociolinguistic variation, connected to the regional origin, age, education and/or family background (hearing vs. deaf parents) of the signer (see e.g. Stamp et al. 2014 for BSL; McKee et al. 2011 for NZSL). In Study II, we investigated numeral variation in YMSLs. Cardinal numbers constitute a prime domain for cross-linguistic comparisons as well as typological studies. Supposedly, all languages of the world have some way to express quantification but can do so in very different ways (see Sagara & Zeshan 2016). In the case of YMSLs, signers from different communities use not only different lexical variants for numbers but have developed totally distinct strategies to construct numbers higher than ten: additive numbers with non-decimal bases in Chicán – a typologically unusual system –, simultaneous morphology (using movement inflection and

41 An edible plant with leaves that can cause an itchy rash 42 A small rodent, traditionally captured in a specific type of trap 43 A Yucatec Maya rite of passage performed when a baby is a few months old

76 non-manual markers) in Chicán, digital numbers in Chicán, Nohkop and Cepeda, tracing strategies in Nohkop and Cepeda as well as context-dependent numerals in all communities. These results reflect a) the independent paths of language evolution in YMSLs from different communities, which go beyond their common gestural precursors and iconic roots, and b) the different sociolinguistic contexts and points in time in which the sign languages emerged (see Study II). So far, no dedicated study has been carried out on YMSL colour signs, but the results from the lexical elicitation task show that this is another highly variable domain. The YMSLs from Chicán and Nohkop both have three fully lexicalised basic colour terms: BLACK, WHITE and RED.44 The lexical encoding of these colours conforms to the colour hierarchy proposed by

Berlin and Kay (1969). The sign for BLACK is identical in Chicán and Nohkop and performed by repeatedly touching one’s hair (which is, by default, black among the Yucatec Maya) with the index finger. The lexemes for RED have a different form in both communities: The sign from Chicán is articulated by rubbing one’s lips with the index finger and is derived from the colour of the lips. Signers from Nohkop use a variant that consists in quick repeated bending of the index and they explain that this sign refers to blood. The signs for WHITE are completely distinct: the Chicán variant is performed by brushing over one’s torso with the index finger, the Nohkop variant by rubbing the tips of thumb, index and middle finger against each other (the sign can be one-handed or two-handed). Chicán signers explain the etymology of their sign by the white colour of the huipil, the traditional clothing worn by Yucatec women. Nohkop signers cannot provide any etymological explanation for their variant of WHITE (see Le Guen et al., in press-a). In Chicán, all other colours are expressed ad-hoc, by touching of, or pointing to objects in the surroundings or describing an object that typically displays the colour in question. Note that younger signers from Chicán sometimes use LSM loans for colour signs.

In Nohkop, signers have developed a generic lexical sign for COLOUR, which is homonymous with the sign for RED, and often accompanied by pursed lips.45 This sign can be used in compounds with nouns for objects with a salient colour, e.g. COLOUR^BANANA for ‘yellow’ or

44 See http://ymslproject.org/vocabulario_ch.html#chican and http://ymslproject.org/Vocabulario_NK.html#nohkop for the signs for NEGRO (‘black’), BLANCO (‘white’) and ROJO (‘red’) in Chicán and Nohkop (accessed 28 September, 2020). 45 Adone et al. (2012) and Nyst (2007) also describe the use of a generic colour sign along with mouthings to express a variety of colours in Konchri Sain and Adamorobe Sign Language.

77 COLOUR^LIME for ‘green’.46 An additional strategy for colour expression in Nohkop is to combine the generic lexeme for COLOUR with a number sign that corresponds to the grade of high school in which the uniform in a specific colour (pink, purple and blue) is worn. From the data we have seen so far, it seems that pointing to or touching objects is the main strategy of colour expression in Trascorral and Cepeda. The comparison of colour terms in Chicán and Nohkop shows that they display certain parallels rooted in common sensory experiences of the signers (i.e. ‘hair is black’), but the domain of colours leaves ample space to choose from a variety of strategies for their encoding (see also Zeshan & Sagara 2016, for a typological survey of colour terms in sign languages). The typical strategy to express days of the week in both communities is to refer to an event that usually takes place on that particular day. However, the specific happenings chosen by signers to metonymically represent the corresponding day of the week differ and result in completely distinct and unrelated regional variants. For instance, the sign for SUNDAY in Chicán is derived from the sign for HUNT, Sunday being the day where Yucatec Maya men traditionally go hunting in groups. In Nohkop, the lexeme for SUNDAY refers to a cartoon broadcasted on Sundays on Mexican television (see Le Guen 2012: 243).47 The domain of calendrical terms offers the opportunity for typological comparisons and is shaped by an interplay of constraints of the visual-gestural modality, i.e. iconicity, sociolinguistic traits of a language community and its cultural characteristics. Another particularly interesting area of investigation is the domain of human referents. These signs occur with high frequency and are located somewhere in the middle on the continuum of variation. There is a certain amount of overlap between different communities as signs denoting human referents are often based on an iconic depiction of a person’s physical appearance, e.g. salient female or male bodyparts, hairstyle, typical accessories, etc. (see Reed & Rumsey 2020, for sign languages in Papua New Guinea; Iseli 2018, for homesign in Vanuatu). Furthermore, there is a cultural component that gives way to parallels across Yucatec Maya signing communities. For instance, signers from Chicán and Cepeda use similar variants

(differing only slightly in their place of articulation) for FEMALE that are derived from the typical hairstyle of Yucatec Maya women, who wear their hair in a bun. Signers from Chicán

46 As other sign language researchers, I use the ^ symbol to mark compounds that consist of two of more constituents, see e.g. the glossing conventions used in the Swedish Sign Language corpus (Mesch & Wallin 2015). 47 See http://ymslproject.org/vocabulario_ch.html#chican and http://ymslproject.org/Vocabulario_NK.html#nohkop for the signs for DOMINGO (‘Sunday’) in Chicán and Nohkop (accessed 28 September, 2020).

78 and Nohkop have a variant for FEMALE that is identical in both villages and depicts the earrings that Yucatec Maya girls wear from a young age. However, the iconic representation of a referent can result in very distinct realisations (Taub 2001), or, in other words, signs can be iconic in different ways when signers select different visual aspects that come to symbolise the referent as a whole. Reed and Rumsey (2020), following Kendon (1980), describe this type of variation as a result of choosing different iconic “bases” for the depiction of human referents.

In the dataset from the YMSL lexical elicitation task, I encountered eleven variants for FEMALE and nine variants for MALE in the four communities. Nouns for human referents are typically age-neutral in YMSLs and can be translated as BOY/MAN or GIRL/WOMAN, depending on the context, hence the glosses MALE and FEMALE. The domain of human referents shows that some lexical signs can exhibit variation and overlap at the same time, when the same underlying iconic image is represented in different ways (see e.g. Perniss et al. 2010). The findings from YMSLs contest Mudd et al.’s (2020) hypothesis that if a concept is discussed with high frequency, it will exhibit less variation across signers.48 Summing up, regional variation in the YMSL lexicon is pervasive and I explored different facets of this phenomenon in the four studies. In Study I, I argued that common cultural and gestural roots lead to similarities in the lexicon of unrelated YMSL communities. Lexical differences provoke strategies of linguistic negotiation and translanguaging between signers from different villages. In Study II, we showed that the lexical domain of cardinal numbers is characterised not only by distinct variants in each community but also more generally by different linguistic and cognitive strategies to construct higher cardinal numbers. Finally, Study III and Study IV demonstrated that irrespective of the presence of several lexical variants, YMSL signers from different communities use identical constructions (involving SASSes and height-specifiers) to encode their lexicon for tools and human referents.

7.3.1.2 Variation in grammatical structures

The studies of my dissertation demonstrate that similarities between YMSLs from different villages go beyond the lexicon and that there are striking parallels also in the grammatical structures of unrelated sign languages in Yucatán. This concerns for instance

48 Findings from other Deaf-community and shared sign languages also proved that semantic domains that occur with considerable frequency in a language, such as numbers, colours or kinship, commonly display a fair degree of variation (see e.g. Zeshan & Sagara 2016; Stamp et al. 2014; Meir & Sandler 2020).

79 SASSes for the noun-verb distinction (Study III) or the use of height-specifiers as person- classifiers (Study IV). Other examples for similarities in grammar and discourse structures between YMSL communities include strategies of time expression (Le Guen 2012), the organisation of the signing space (Le Guen & Safar, in prep.) or the use of discourse markers, such as FINISH or PALM-UP (Le Guen et al., in press-a). Differences in the use of certain grammatical features were observed between signers from Chicán and signers from the other communities. In Study III, we found that constructions involving SASSes as nominal markers occurred in three villages: Chicán, Nohkop and Cepeda (Trascorral was not included in the SASS study). There was no significant difference between communities in the overall frequency with which SASSes were used, but the syntactic position of the SASS was more robust in Chicán. In Chicán, the SASS occurred in a final position (following the base noun) in the majority of the cases, while in Nohkop, the structure seems to be more flexible. Cepeda was not included in the analysis of the syntactic distribution of SASSes because of the paucity of data from this community. Concerning the frequency of SASSes as nominal markers we discovered interesting patterns of variation within the communities and found that, in this respect, signers from two interactional groups in Chicán behave more similarly to some Nohkop signers than to signers from other interactional groups in their own village. This suggests that variables other than region influence the use of SASSes as nominal markers in YMSLs. I will return to this point in Section 7.3.2. In Study IV, I showed that in my dataset, signers from Chicán used the human height- specifier as a noun-classifier more frequently than those from Nohkop or Cepeda (although a detailed quantitative analysis has not yet been conducted). Possibly, the fact that YMSL in Chicán has already been subject to intergenerational transmission from deaf adults to deaf children contributes to more conventionalisation of grammatical structures (see e.g. Senghas 2003, for Nicaraguan Sign Language). I discuss this issue in more detail in Section 7.3.2.2. However, with all inter-community comparisons we must be cautious and keep in mind that we simply have more data from Chicán than from the other communities. This analytical bias may be responsible for the fact that we observed certain structures more frequently in Chicán than in the other villages. Also, it is necessary to have a detailed look at variation not only between but within the communities. YMSLs are not homogenous in each village, especially not in Chicán, and the regional origin of the signers is strongly intertwined with other social variables, such as age or family membership. I will discuss the issue of intra- community variation in the following section.

80 7.3.2 Variation within communities

“The sign FATHER! Ismael signs it like this [showing a beard]: The one with a beard, who is already older. These ones [signers from another family] not! [shows their sign, touching his hair]: The one who is already old and has white hair. This sign, they show it in different ways. – Sign language is the same thing as if someone speaks normally, orally. Because we can say, for instance ‘father’ and someone else could say ‘my old man’, ‘my boss’! That’s the same thing! This is how they [the deaf] communicate! They also have their different ways. But yes, they understand each other! For example, if I talk to a guy and he says: “My boss told me…” – I already understand that he is talking about his father, he is not referring to his superior at work! And like this, they also understand each other. The problem comes up when there is a third person, who does not know sign language: they won’t understand! If they only know one sign, even if another person is talking about the exact same thing, they won’t understand.”49

In this quote from an interview, a hearing man from Chicán, who grew up with a deaf brother and a deaf sister, explains that individual deaf people from his home village sign differently. Without using any linguistic terminology, he touches upon several crucial issues in the study of linguistic variation: First, just as spoken languages, sign languages have different ways of saying the same thing. Second, even though synonymous signs take on different forms, they are of equivalent communicative value. Third, a certain level of competence in the respective language is required in order to successfully decode other people’s variants: Deaf people who are proficient in a specific sign language have no problems in communicating with other signers of that language despite the existing linguistic variation, but people who have no or little command of the sign language often have trouble deciphering different variants. Studies on various national sign languages (e.g. Bayley et al. 2002; Lucas, Bayley & Valli 2001; Lucas et al. 2003 for ASL; McKee et al. 2011; McKee et al. 2007, for NZSL; Schembri et al. 2013; Stamp et al. 2014 for BSL; Johnston & Schembri 2007, for Australian Sign Language (Auslan); LeMaster & Dwyer 1991, for Irish Sign Language) and emerging/village sign languages (e.g. Israel 2009 for ABSL; Washabaugh 1986, for Providence Island Sign Language; Hou 2016 for SJQCSL; Tano 2016, for LaSiBo) demonstrated that, just as in spoken languages, linguistic variation is a common phenomenon also in the signed modality. For instance, in a well-known study entitled “What’s your sign for PIZZA?”, Lucas et

49 In the original quote in Spanish, the man uses the word jefe (here translated as ‘boss’), an informal term for ‘father’ common in Mexico. All translations from Spanish in this section are my own.

81 al. (2003) showed that ASL exhibits patterns of sociolinguistic variation, depending on the signers’ region of origin, age, gender and/or ethnicity. YMSLs constitute no exception in this respect. Variation within YMSL communities is connected to different social variables (see e.g. Labov 1972; Meyerhoff 2006). In the following sections, I will discuss the role of family membership, age, hearing status, gender and religious affiliation for sociolinguistic variation.

7.3.2.1 Family membership/interactional groups

Meir et al. (2012) argue that the family a signer belongs to is the primary source of variation in Al-Sayyid Bedouin Sign Language (ABSL). Sandler et al. (2011) used the term “familylect” (see Søndergaard 1991, for spoken languages) to refer to signing varieties that originate in different (extended) families of ABSL signers. They observe that despite the abundance of lexical and sublexical variation across the community there is a strong degree of lexical uniformity within households that include several signers. Similarly, Hou (2016) mentions the existence of different “family-based lexicons” in SJQCSL, which overlap due to effects of iconicity and the presence of conventional gestures. For YMSL in Chicán, Shuman (1980) already noted lexical variation between signers from different families, but his account remains anecdotal and he does not further analyse the phenomenon. Le Guen (2012: 216) categorises signers in Chicán into different “interactional groups”. These are extended families that live on the same compound and interact on a daily basis (see Section 2.3.3). Studies II, III and IV provide evidence for distinct signing varieties used by different interactional groups in Chicán and show that these “familylects” affect not only the lexicon but also the grammar of YMSLs. In Study II, we showed that the interactional groups in Chicán differ in their use of additive versus digital numerals. Additive numerals are widely used across most deaf and hearing signers in the community and are passed on from older to younger generations. An exception are the signers from interactional group 3: These three deaf siblings, who are all under 30 years old and have little contact to other deaf signers in the village, favour digital numerals over additive ones. In the Studies II, III and IV, the familylect of one extended family in Chicán proved to be especially conspicuous. This particular family is made up of the interactional groups 1 and 2, whose family compounds are located just across the road from each other. The family includes seven deaf and several highly proficient hearing signers of different ages (siblings,

82 spouses, children, grandchildren, nieces, nephews and cousins of deaf people) and has been the main sphere of intergenerational transmission of YMSL in Chicán. The findings from Studies II, III and IV highlighted that the signing variety of the members of this extended family exhibits certain characteristics that set them apart from other Chicán signers. An instance of family-related peculiarities in the number domain concerns the lexical sign for HUNDRED(deer). Although Zeshan et al. (2013) list this lexeme as the only translation for ‘100’, our analysis of data from a variety of signers in Chicán showed that the sign is used only by the members of the interactional groups 1 and 2. Deaf signers from interactional group 5 stated not only that they did not understand the semantics of sign, but they rejected its use because it originated from a family they are not particularly affectionate with. It is noteworthy that even in a micro-community sign language in its early stages of development, signers already have clear attitudes about certain signed variants being more correct, beautiful or authentic than others (see also Haviland 2016, for Zinacantán family homesign). Study III goes beyond the lexicon and demonstrates that family-related variation also affects grammatical features of YMSLs. Again, the interactional groups 1 and 2 in Chicán stood out in this study. We showed that signers from these two closely related interactional groups used SASSes as nominal markers significantly more often than other signers. Thus, we concluded that SASSes as nominal markers form a part of their “familylect”. Although I have not yet conducted a quantitative analysis in this respect, the results of Study IV similarly suggested that younger signers from the interactional groups 1 and 2 used height-specifiers as noun-classifiers more often than signers from other families. Due to the concentration of many deaf signers and the availability of numerous hearing signers who willingly act as interpreters for their deaf relatives, the members of this extended family have been the preferred informants for many linguists who have worked in Chicán. While this is a convenient methodological choice it creates a bias when one aims to represent the sign language of the entire community. Possibly even as a result of having worked extensively with linguists and being asked repeatedly to provide YMSL translations for different lexical domains, signers from this family created not only distinctive number signs

(see above for HUNDRED(deer)) but also lexemes for weekdays and months, which are unknown to other signers in the village. For instance, the sign for SEPTEMBER refers to the Mexican Independence Day on the 16th of September. The first part of the compound sign represents the salute to the Mexican flag, the second part depicts the march performed on Independence Day.

Figure 14: YMSL sign for SEPTEMBER in the interactional groups 1 and 2 in Chicán

An exception seems to be the sign for DECEMBER, which is homonymous with the sign for CHRISTMAS and derived from the depiction of the Divine Child. This sign is used across all interactional groups in Chicán.

Figure 15: YMSL sign for DECEMBER in Chicán

These observations underline the importance of including signers from different families in the documentation of young village sign languages with a high degree of variation. The sign languages of Nohkop and Trascorral are only used within a single family, thus, the notion of family-related variation does not apply to them. As the sociolinguistic setting of Cepeda resembles the one in Chicán regarding the distribution of deaf signers across multiple families, we can assume that YMSL of Cepeda also exhibits familylects. Given that the investigation of YMSL in Cepeda is still in its early stages, the comparison of signing varieties from different families remains a topic for future analyses.

84 7.3.2.2 Age of the signers

“They [the deaf] adapt to modern times. It’s not because they learn other signs. But rather, there are new things, new objects… So they look for a way to explain and make known the object they want to show. Things that sometimes, in the case of Santos, Ismael, Adán, those who are already older… for instance Don Teodoro [oldest deaf person in Chicán, now deceased; J.S.], do you know him? He doesn’t even know what a mobile phone is! If you ask him, he doesn’t even have a sign for ‘mobile’ because he has not experienced it, has not seen it. It is new! But if you ask Cruz [youngest deaf person in Chicán; J.S.] he will show you quickly [the signs for] Internet and all that!”

The same hearing signer that I quoted above comments on family-related differences between signers in Chicán (see Section 7.3.2) also observed some lexical innovations for newly emerging terms. These include for instance concepts related to communication technology. As in other (spoken and signed) languages, such new terms can be expressed through loans from other sign languages (see e.g. McKee & McKee 2020, for ASL borrowings in NZSL). For instance, the sign for FACEBOOK in Chicán is borrowed from LSM. Other lexemes are newly coined by YMSL signers, e.g. the sign for UPLOAD-TO-INTERNET in YMSL of Chicán.

FACEBOOK is an initialised sign, with an F-handshape touching the cheek. As YMSL does not make use of a manual alphabet, this strategy of coining new lexemes is not available in the language.50 The sign for UPLOAD-TO-INTERNET in Chicán is a combination of a sign that shows the action of typing on a keyboard and an indicating sign that describes an arc movement away and upwards from the signer’s body.

Figure 16: YMSL signs for FACEBOOK (left) and UPLOAD-TO-INTERNET (right) in Chicán

50 Once adopted into YMSL, the LSM loan can thus no longer be considered an initialised sign – the F-handshape is retained in YMSL, but the form itself is meaningless since it is not part of a manual alphabet.

Some instances of age-related variation in YMSLs are influenced by the distinct life experiences and access to material commodities that the signers grew up with and within which they created their sign languages. In Study II, we showed that the first generation of signers in Chicán, who are today over forty years old, received little to no formal schooling and came to develop an additive numeral system completely independent from written numbers. They passed this system on to the younger generation of signers in Chicán. Signers from Nohkop and Cepeda, who are overall younger than Chicán signers, had more exposure to formal education, as well as more access to written language via information and communication technology (television, smartphones, social media). These signers created number systems that are based on writing, i.e. digital and/or writing/tracing strategies. This also applies to three younger signers from interactional group 3 in Chicán, who do not have much contact with other signers in the community and thus did not acquire their additive number system. McKee et al. (2011) discuss examples of age-related variation in NZSL numbers as a result of changes in Deaf education policies, i.e. the introduction of Australasian Signed English in schools. Apart from differences that stem from the sociolinguistic context that signers live in, we also encounter differences between older and younger signers that can be attributed to the evolution of the language itself. Study III demonstrated that the closely related interactional groups 1 and 2 in Chicán used SASSes as nominal markers more often than other signers (see Section 7.3.2.1), but within these interactional groups, it was the younger deaf and hearing signers (under 30) who used this structure most frequently and most consistently. As these signers grew up with rich language input from deaf parents or other deaf caregivers, I hypothesise that we are witnessing an effect of language transmission. Younger signers learnt YMSL from adult deaf language models and restructured their linguistic input. Features that were used with more variability by the older generation became more consistent and conventionalised among the younger generation (see e.g. Senghas 2003, for NSL). Some findings from Study IV seem to strengthen this hypothesis. A detailed quantitative analysis on this issue remains to be conducted, but preliminary results show that in Chicán, the younger signers with deaf parents or other close deaf relatives (i.e. the same group as in Study III) use human height-specifiers as noun- classifiers more often than other signers from their village or from other communities. Also, the tendency to locate the height-specifier in a fixed syntactic position (following a gender/age noun) is most solid among this group of signers. Moreover, the younger signers from Chicán use the human height-specifier as an agentive marker in compounds for professions. This

86 function of the height-specifier was not documented among older signers and seems to be an innovation brought into YMSL by the younger age group. It is remarkable that the same individuals, i.e. signers under 30 from two main interactional groups in Chicán, stand out in different studies on a range of linguistic phenomena: SASSes as nominal markers (Study III), human height-specifiers (Study IV) as well as the use of the signing space for directional verbs (Le Guen & Safar, in prep.). Further studies need to be conducted in order to confirm whether language transmission from deaf adults is indeed the variable that distinguishes this group from other signers or if other sociolinguistic factors play a role herein.

7.3.2.3 Hearing status

Because YMSLs are shared sign languages that are co-created by deaf people and their hearing interaction partners, it is crucial to include not only deaf but also hearing signers in our analyses in order to obtain a comprehensive picture of the language (see also de Vos 2016; Nyst, 2015). As in other shared sign languages, hearing L2-signers form the majority of signers in the YMSL communities. In general, we discovered little evidence for linguistic differences between deaf and hearing signers but rather found that hearing people’s language use mirrors the patterns of their deaf relatives and/or other frequent interaction partners. However, we have not yet conducted any dedicated studies in this respect. Mesh and Hou (2018) point out differences in the use of negation signs between deaf and hearing signers of SJQCSL.51 It is not unlikely that an in-depth comparison of deaf and hearing signer’s productions could reveal variation between these two groups also in YMSLs. The fact that hearing YMSL signers are also speakers of Yucatec Maya and (at least to some extent) Spanish leads to intense language contact and is likely to cause linguistic transfers from spoken to signed language. Indeed, Le Guen et al. (in press-a) and Le Guen (2012) show evidence for linguistic calques from spoken Yucatec Maya into YMSLs. This is not an uncommon phenomenon and de Vos (2016: 212) also mentions the existence of calques from spoken Balinese in Kata Kolok. In the studies for my dissertation, the only instance of variation between deaf and hearing YMSL signers I encountered was in Study III. We found that hearing bilingual signers in Chicán as well as in Nohkop used SASSes for the noun-verb distinction with high frequency compared to deaf signers. For these participants, YMSL is not the only language they are

51 The authors admit, however, that these differences could as well be a result of having analysed more data from deaf than from hearing signers.

87 proficient in, but they also speak languages that employ overt morphological and/or lexical markers of the noun-verb distinction. It is plausible to assume that they carry over this distinction to YMSL. Another explanation could be that the hearing participant group, maybe because of their better access to formal education, detected the purpose of the task more clearly than the deaf participants. One note of caution must be made at this point. In Study III, some of the signers who used a particularly high proportion of SASSes are hearing, under 30, attended secondary school and belong to interactional group 1 and 2 in Chicán, thus grew up with deaf adult language models. This underlines that the social variables age, family membership, language acquisition, education and hearing status are closely entangled. Stamp et al. (2014) or McKee et al. (2011) point out such interactions between multiple social variables in their analysis of sociolinguistic variation in BSL and NZSL. They tackled this problem by running a multivariate analysis using the statistical program Varbrul in order to tease apart the influence of individual factors on linguistic variation. Due to the small community size and the impossibility of creating a socially balanced sample of participants, it is difficult to conduct similar statistical analyses to test for multi-variable effects in YMSLs. All claims concerning sociolinguistic variation in YMSLs must be regarded as part of a complex whole and need to be confirmed through in-depth qualitative analyses of data from a variety of participants.

7.3.2.4 Gender and religious affiliation

Concerning the effect of gender and religion on variation in YMSLs, we only have anecdotal evidence to date. A study by LeMaster and Dwyer (1991) proved that in Irish Sign Language, gender had an effect on the signers’ lexicons, as deaf men and women used to attend separate schools where they learnt distinct signing varieties. After these signers left school, men and women of course started to increasingly mingle with each other and their lexicons started to converge, but parts of the gender-based varieties still persist in contemporary Irish Sign Language (Leeson & Grehan 2004).52 As for religion, there are no in-depth studies on this topic, but Schembri and Johnston (2012) mention that British or Australian signers who attended Catholic schools, where Irish

52 Even though this gender-related variation in Irish Sign Language is not directly an effect of gender but rather indirectly a result of the separation in education, the studies quoted above are to this date the most well-known accounts of the impact that gender has on sign language variation.

88 Sign Language was the main language of instruction, acquired some Irish Sign Language vocabulary as well as the manual alphabet. A further example is language use by the British Jewish Deaf community (Jewish Deaf Association 2003). Jewish BSL signers were often educated in separate Jewish Deaf schools and came to develop specific lexical signs that represent key concepts of their religion.

As for YMSLs, we observed that male signers in Nohkop use a lexical variant for MALE, in which the extended thumb depicts the penis (see Study IV). Female signers state that they find this sign embarrassing and prefer to use other variants of MALE that refer to a man’s hairstyle or moustache. A further gender-related difference is that female signers from Nohkop seem to employ a greater variety of colour terms than their male counterparts. Le Guen et al. (in press-a) speculate that this is caused by the fact that these girls often order make-up from a catalogue and thus need to convey finer distinctions of colours. As for religious affiliation, some deaf people from Cepeda – a predominantly Protestant village – explained that they do not use the variant for FEMALE, which refers to the earrings that Yucatec Maya girls typically wear from a young age. While this sign is common in Chicán and Nohkop, Protestant signers from Cepeda clarified that their religion does not allow them to use jewellery. Although these observations on variation related to gender and religion are preliminary and need to be complemented by dedicated studies, it is striking that even micro-communities of signers like in YMSLs display patterns of language use connected to a range of social variables.

To conclude this section, I demonstrated that in YMSLs, we are witnessing the emergence of sociolinguistic variation on a micro-scale. This goes against Meir et al.’s (2012) claim that in young village sign languages, variation occurs mainly on an individual or family level and does not yet display systematic sociolinguistic patterns, because solid linguistic conventions are still not established. I argued, however, that through careful (qualitative and quantitative) analysis of data from a variety of signers from different regions, age groups, gender and interactional groups, the seeds of sociolinguistic variation are discernible already in young micro-community sign languages. In order to provide a representative picture of a shared sign language, it is essential to collect data from a broad spectrum of deaf and hearing signers and to take variation into account. Paying close attention to variation and the factors that are responsible for it can help us to understand processes of lexical and grammatical conventionalisation in shared sign languages.

89 8 The contribution of YMSLs to studies on sign language emergence and evolution

In the following section, I will outline the main contributions of my dissertation to the broader area of research on (sign) language emergence and (sign) language evolution. In Section 8.1, I will reflect more generally on the role of Yucatec Maya gestures for the emergence of YMSLs, while in Section 8.2, I will look in more detail at paths of lexicalisation (Section 8.2.1) and grammaticalisation (Section 8.2.2) from gesture to sign and at the relationship between the two phenomena (Section 8.2.3). In Section 8.3, I will draw some conclusions on what the case of YMSLs can teach us about the relevance of various sociolinguistic factors for (sign) language emergence.

8.1 From Yucatec Maya gestures to YMSL signs

Investigating the emergence and evolution of sign languages as well as cognitive aspects of language development, a number of studies have looked at the relation between hearing people’s gestures and signed languages from different parts of the world (e.g. Pfau & Steinbach 2011; Padden et al. 2013; Ortega et al. 2019; Johnston 2018; Cooperrider et al. 2018; Senghas et al. 2004 – to name only a few). Shared sign languages with close contact between deaf and hearing people present a particularly promising opportunity to study the development from gestures into signs (e.g. Tano & Nyst 2018; Mesh & Hou 2018; Mesh 2017) and the analysis of YMSLs ties in with this line of research. Such a comparative approach is crucial for the thorough description of shared and/or emerging sign languages: In order to understand whether and why a certain phenomenon in a sign language is extraordinary we need to investigate whether it has an equivalent in the gestural behaviour of hearing people (Mesh 2017: 36). Speakers of Yucatec Maya make extensive use of multimodal communication and have a rich inventory of conventional gestures that complement their spoken utterances. Le Guen et al. (in press-a) provide numerous examples of how gestures of hearing Yucatec Maya have become incorporated into YMSLs, via a process of lexicalisation, grammaticalisation or as direct “transfers” (i.e. without modification in form or function). They argue that certain dimensions of gestures53 (e.g. lexical quotable, holophrastic, pointing gestures, concept gestures etc.) tend to be systematically transformed into certain parts of speech in YMSLs.

53 Le Guen et al. (in press-a) distinguish the following categories: lexical quotable gestures, holophrastic quotable gestures, manual classifiers, concept gestures, iconic gestures and

90 In a similar vein, the four studies of my dissertation show, from different angles, how gestural practices of hearing speakers of Yucatec Maya scaffold the emergence of YMSLs. At the same time, I also demonstrate important differences between Yucatec Maya gestures and signs, showing that deaf signers restructure the gestural substrate. Once gestures are incorporated as part of a signed linguistic system, their form, meaning and distribution are transformed and they interact with other linguistic structures. Mesh (2017) describes this process as “adapting while adopting”: Deaf signers do not take up the gestural system exactly as it is but alter it and introduce new language-specific innovations. At this point, it seems appropriate to say a few words about the conceptual distinction between “gesture” and “sign”. As both gestures and signs are expressed via the visual-manual modality it is sometimes difficult to separate the gestural use of a form from the signed one, which led Emmorey (1999) to ask whether signers gesture at all. In the literature on gesture and sign language studies the relationship between gestures of hearing people and sign languages used by deaf people has been a topic of much debate throughout the history of the disciplines (see Müller 2018, for an overview). While some scholars claim that gestures and signs exhibit many common characteristics and that the boundary between the two categories is dynamic rather than clear-cut (e.g. Kendon 2008; McNeill 2000), others counter that there is a “cataclysmic break” once the visual-manual modality becomes the sole available channel of communication (e.g. Goldin-Meadow & Brentari 2017). Müller (2018) argues that these fundamentally diverging theoretical positions are based on the authors’ totally different notions of ‘gesture’. She underlines that in order to understand similarities and differences between gestures and signs it is not enough to draw from gestures produced under experimental conditions, but that one needs to look at multimodal language use “in the wild”. In their analysis of negation in SJQCSL, Mesh and Hou (2018) opted to abandon an overt terminological distinction of gestural versus signed manifestations of negation and speak of conventional gestures even once these forms have been incorporated into the emerging sign language. Their decision seems justified given that the strict separation of gestures and signs metonymic pointing. They describe these categories as being different types of gestures. This is somehow different from McNeill (1992) or Kendon (2004), who treat these categories as different dimensions that can coexist in one and the same gesture. This also applies to Le Guen et al.’s (in press-a) classification: In their definition, the “types” of gestures are not mutually exclusive – for instance, a concept gesture such as ‘looko (crazy)’ can at the same time be holophrastic and quotable, or a lexical quotable gesture such as ‘tak’in (money)’ can be iconic as well. In another publication, Le Guen (2018) refers to the categories “iconic”, “deictic”, “metaphoric” etc. as dimensions and emphasises that more than one property can be present in a single gesture.

91 in many cases does not mirror language user’s perceptions of their visual communication (Kusters & Sahasrabudhe 2018; see also Study I of my dissertation). However, Mesh and Hou’s terminological wariness somehow obscures the fact that their very analysis reveals important differences in meaning and syntactic distribution of the conventional negation gestures among hearing non-signing speakers of Chatino and signers of SJQCSL. I understand gestures and signs to share many overlapping formational and functional properties and to exist on a continuum (Kendon 2008; Müller 2018). For the sake of clarity, I will make a functional distinction and refer to the visual communicative behaviour of hearing, non-signing speakers that goes along with their spoken utterances as “gesture” and the one of deaf and hearing signers, which is restricted to the visual channel, as “sign language”. In their study, Le Guen et al. (in press-a) aim to provide a broad overview of the relation between Yucatec Maya gestures and signs. In my dissertation, I have chosen depth over breadth, and compared specific gestures and signs in more detail. Some of my findings match Le Guen et al.’s results, while others suggest that some of their claims about grammaticalisation and lexicalisation need to be re-examined. In terms of interaction, I showed in Study I that the semiotic repertoires of hearing and deaf Yucatec Maya overlap due to their shared cultural background and gestural substrate. These factors facilitate the communication between deaf and hearing members of one community as well as between signers from different villages. However, I also argued that semiotic resources are not equally accessible to everybody: Despite the fact that conventional gestures and cultural knowledge can bridge many gaps in communication there are linguistic elements in YMSLs that are not transparent to non-signers, and sign language competence is necessary in order to converse with deaf people about more complex topics. In Study II, we found that certain features of Yucatec Maya gestures for quantification are preserved in YMSL numerals, e.g. the opposition of starting to count from the index for cardinal and from the pinky for ordinal numbers, or the lexical sign for FIFTY(half) in Chicán. However, distinct number paradigms have evolved in YMSLs from different communities, which go beyond their gestural precursors. Most notably, signers from Chicán have developed an additive number system totally independent from gestures or written numbers. Once number signs become part of sign language structure they are subject to specific linguistic constraints, as we demonstrated for instance in the case of handshape assimilation of FIFTY(half), the possibility for spatial/non-manual modification of FIFTY(half), or the existence of distinct number variants in isolation and in compounds. These observations match findings by Coppola et al. (2013) on homesign in Nicaragua: They show that homesigners use devices for number

92 expression that resemble the number gestures used by their hearing communication partners. However, the homesigners’ systems display linguistic patterns not found among hearing gesturers, for instance the incorporation of number handshapes into other signs. These results demonstrate that the input for a signed numeral system can come from hearing people’s number gestures but that deaf signers reorganise this input into a linguistic system.

8.2 Paths of conventionalisation from gestures to signs

In the following sections, I will look in more detail at paths of conventionalisation from gestures to signs. I will discuss the phenomena of lexicalisation (in Section 8.2.1) and grammaticalisation (in Section 8.2.2), relating the findings from YMSLs to the wider body of literature for signed and spoken languages. In Section 8.2.3, I will reflect about the relationship between lexicalisation and grammaticalisation as separate but intertwined processes. Lexicalisation and grammaticalisation, i.e. the creation of new lexical and grammatical elements, have been described as universal phenomena across languages, although there are typological differences in how these processes unfold (see Narrog & Heine 2018, for a typology of grammaticalisation). Several studies (e.g. Janzen & Shaffer 2002; Janzen 2012; Pfau & Steinbach 2013; Wilcox 2005; Wilcox et al. 2010; Zeshan 2003) proved that lexicalisation and grammaticalisation are not limited to spoken languages but equally occur in signed languages. While some facets of these developments seem to be linguistic universals, other aspects are specific to the signed modality (see e.g. Pfau & Steinbach 2011). One modality-specific feature is that sign languages commonly recruit manual or non-manual gestures as a source for lexical and grammatical structures. The analysis of lexicalisation and grammaticalisation in sign languages comes with particular difficulties: Diachronic investigations of signed languages are methodologically challenging, as sign languages are generally young and historical sources are scarce (Janzen 2012). This issue concerns in particular emerging sign languages, which are only a few generations old and in active and ongoing contact with spoken languages and their accompanying gestures. At the same time, emerging/shared sign languages offer a brilliant opportunity to study lexicalisation and grammaticalisation “in real time” from a synchronic point of view (see also Coppola, in press).

93 8.2.1 Lexicalisation: Gestures become lexical signs

Brinton and Traugott (2005: 96) define the process of lexicalisation as follows: “Lexicalization is the change whereby in certain linguistic contexts speakers use a syntactic construction or word formation as a new contentful form with formal and semantic properties that are not completely derivable or predictable from the constituents of the construction or the word formation pattern.” In the case of spoken languages, lexicalisation can be the result of various word formation processes, such as derivation, compounding or conversion (Brinton & Traugott 2005). For sign languages, it has been proven that gestures often act as a source for building new lexical items in sign languages (e.g. Zeshan 2000; Janzen 2012; Wilcox 2005;

Wilcox et al., 2010; Cormier et al. 2012). For instance, the signs for THANK-YOU in Indo-

Pakistani Sign Language (Zeshan 2000) or STINGY in NSL (Coppola, in press) are based on conventional gestures used in India and Nicaragua. Le Guen et al. (in press-a) demonstrate this process in YMSL for various dimensions of gestures (see footnote 53), with a specific focus on quotable gestures (emblems). These gestures have a stable and quotable meaning, such as tak’in ‘money’. The authors argue that due to their consistent form-meaning mapping across speakers, quotable gestures can easily be incorporated into the emerging sign language lexicon without much semantic modification. Given that emblems already have language-like properties, there is no change in form and meaning when they become part of the YMSL lexicon, but they only undergo syntactic integration. It is, however, necessary to closely compare (manual and non-manual) form, meaning and distribution of gestures and signs, as the differences can be subtle and lead to hasty conclusions regarding the gesture-sign interface. For instance, Mesh and Hou (2018) find that even in the case of quotable negation gestures there are subtle shifts in form and meaning between Chatino gesturers’ and SJQCSL signers’ productions.54 Similarly, Coppola (in press) shows that conventional gestures used by hearing Nicaraguans undergo a number of semantic and phonological changes when they are integrated into NSL. We have yet to find out whether there are specific aspects that set apart YMSLs in regard to the lexicalisation of co-speech gestures and to what extent they pattern with findings about other established and emerging sign languages.

54 Although we have to bear in mind that in the case of negation we are dealing with a more complex phenomenon than some of the lexical domains Le Guen et al. (in press-a) are discussing. Possibly the relationship between emblematic gestures and lexical signs is more straightforward than with grammatical features, but we should not take this for granted. The same thorough analytical approach to gesture-sign comparisons is required, no matter what type of gestures we are looking at.

94 In Study II, we demonstrated that YMSL number lexemes from one to ten are identical to the counting gestures used by hearing Yucatec Maya. In contrast to number gestures though, the YMSL number signs are affected by language-specific phonological constraints: handshape and/or palm orientation vary when a number is used in isolation or as part of a compound.

Another instance of lexicalisation from gestures is the sign for FIFTY(half) in Chicán, which is derived from a widely-used gesture in Mexico meaning ‘half’. The meaning of the resulting number sign in YMSL of Chicán is more restricted, referring only to ‘fifty’, i.e. ‘half of hundred’.55 This kind of semantic specialisation is typical for lexicalisation processes (see e.g. Himmelmann 2004). Another example for semantic and articulatory changes from Yucatec

Maya number gestures to YMSL signs is the interrogative sign for HOW-MUCH/HOW-MANY, expressed by a fast and continuous articulation of one through five, starting from the pinky. The manual sign is accompanied by non-manual question markers, i.e. a backward head-tilt and nose wrinkle.56 This form is in itself not a conventional gesture, but rather is the YMSL sign HOW-MUCH/HOW-MANY a semantic reanalysis of the combination of a Yucatec Maya counting sequence with a question (lit.: ‘one? two? three? four? five?’). Study IV contributes further examples of lexicalisation from conventional Yucatec Maya gestures. I showed that the height-specifier gesture used for upright referents in spoken Yucatec Maya becomes a lexeme for human referents in YMSLs, which can have several possible meanings, depending on the context: CHILD, YOUNGER-SIBLING, DAUGHTER/SON (in all four communities), (GRAND-)MOTHER or FATHER in Nohkop and Cepeda. Here, we witness a shift in meaning from upright referents more generally in gesture, to exclusively human beings in the sign SPEC-H:HUM. Also, we see that one gestural form can become polysemous in the YMSL lexicon, referring to a variety of referents that all share the same core meaning ‘human being’. At the same time, there is a change in form, as the location shifts from an iconic (depicting the referent’s height) to a neutral location in front of the signer’s torso. Unlike during the articulation of the gesture, the signer’s eye gaze is detached from the hand and instead directed to the interlocutor or neutral space. These differences demonstrate that the lexicalisation of the height-specifier gesture for upright referents into YMSLs is not a simple transfer (i.e. without any changes in form and meaning) but involves semantic and formational

55 Some signers in Chicán also use a related variant for FIVE-HUNDRED, i.e. ‘half of thousand’. Another related variant exists in all three communities for ‘half an hour’, as in TWO^HALF (‘half past two’). 56 These non-manual question markers in YMSL are themselves grammaticalised from non- manual interrogative gestures used by hearing speakers of Yucatec Maya.

95 adaptations. The semantic restriction of SPEC-H:HUM to the category of human referents is also reflected in the lexicalisation of the height-specifier into the verb GROW, which has distinct parameters of orientation depending on whether it refers to the growth of a human being or a four-legged animal. Following Brinton and Traugott’s (2005) definition quoted above, the examples I presented (YMSL numerals and SPEC-H:HUM) constitute new contentful forms whose formal and semantic properties are not completely predictable from the underlying gestures, demonstrating that paths of lexicalisation in YMSLs display similar characteristics as in other spoken and signed languages.

8.2.2 Grammaticalisation: Gestures become grammatical markers

The process by which novel grammatical structures arise in a language is called grammaticalisation, and Hopper and Traugott (2003: 99) define it as follows: “Grammaticalization is the change whereby in certain linguistic contexts speakers use parts of a construction with a grammatical function. Over time the resulting grammatical item may become more grammatical by acquiring more grammatical functions and expanding its host- classes.” Again, there are parallels and differences between language modalities in regard to grammaticalisation: Just as spoken languages do, sign languages create grammatical elements out of items that before, were not grammatical in nature, or grammatical elements start to fulfil new grammatical functions. As it also is the case for lexicalisation, gestures can form a base for grammaticalisation in sign languages (e.g. Janzen & Shaffer 2002; Janzen 2012; Zeshan 2003; Wilcox 2005; Wilcox et al. 2010; Pfau 2015; Pfau & Steinbach 2011). These previous accounts have generally considered the incorporation of gestures into sign languages as a transformation of “prelinguistic” or “extralinguistic” material into linguistic elements (e.g. Janzen & Shaffer 2002; Wilcox et al. 2010). Le Guen et al. (in press-a) or Mesh (2017) question this point of view, arguing that co-speech gestures already possess certain language-like characteristics and are further modified when they become part of the sign language. My analysis of height-specifiers (Study IV) illustrates Le Guen et al.’s and Mesh’s point: These manual forms have a classificatory function already in gesture but become more conventionalised and constrained in YMSLs. Wilcox (2005) suggests two main “routes” for how gestural items can be grammaticalised into sign languages with or without the intervening stage of a lexeme. The first route is that a gesture enters a sign language as a lexical unit and develops further into a

96 grammatical morpheme. An example is the old LSF sign PARTIR ‘to go/leave’. It originated in a French quotable gesture meaning ‘let’s go’ and over time evolved into a future marker used in contemporary LSF and ASL (Janzen 2012). The second route is that a gesture bypasses the lexical stage, and instead becomes part of the prosody/intonation of a sign language, then further develops into a grammatical morpheme. Elements that become grammaticalised along this path often include non-manual gestures of the face or head as well as manner of movement of manual gestures/signs. Well-known examples for the second route are the grammaticalisation of the headshake gesture into a negation marker in sign languages (Pfau & Steinbach 2011)57 or questioning gestures that evolved into yes/no-question markers and further into topic markers in ASL (Janzen 2012). Unfortunately, many studies on the grammaticalisation of gestures into sign languages, such as Janzen and Shaffer (2002), Janzen (2012), Wilcox (2005) or Wilcox et al. (2010) are only based on historical accounts of gestures or old depictions in books, posters etc., or they rely on commonplace assumptions about gestures instead of attempting an analysis of the actual use of gestures in spoken discourse. If we want to understand how gestures become part of a sign language’s grammar and which transformations their form and meaning undergo it is crucial to carefully investigate these gestures in context, i.e. in relation to speech. Study IV attempts to contribute to filling this gap. The process of grammaticalisation is characterised by phonetic reduction and semantic generalisation (Bybee et al. 1994: 6). Previous literature (e.g. Zeshan 2003; Wilcox et al. 2010) demonstrated that these criteria equally apply to signed languages. When a form becomes (more) grammatical, one or several phonological parameters of a sign are reduced while its meaning changes from more concrete to more abstract. Wilcox et al. (2010) argue that the same general cognitive process, namely metonymic extension, functions as an underlying principle for grammaticalisation in spoken and signed languages. Most commonly, studies on lexicalisation and grammaticalisation take on a diachronic perspective and look at historic language data in order to establish how the use of a specific form has changed over time. However, grammaticalisation can happen within a relatively short time frame for instance in pidgin or creole languages (McWhorter 2018). The situation is similar for emerging sign languages: Previous research, for instance on NSL, demonstrated that signers rapidly created new grammatical structures over a very short period of time (see e.g.

57 Although Johnston (2018) argues that the headshake in Auslan is not grammatical but still gestural in nature.

97 Senghas & Coppola 2001). In young (sign) languages that have not yet developed rigid conventions for a range of linguistic domains, there is not only more “space” (McWhorter 2018) but also more need for new grammatical items. As argued in Section 8.2, conducting diachronic studies for sign languages is challenging, especially in the case of young, shared sign languages. One the one hand, historic data for these languages is obviously limited and on the other hand, they display a lot of variation (see Section 7), as their linguistic structures are still being conventionalised. In shared signing communities, deaf and hearing people are in close interaction and there is ongoing language contact between their shared sign language and the surrounding spoken language and its gestural system. Therefore, they represent an excellent opportunity for synchronic studies of grammaticalisation from gesture to sign, i.e. the investigation of forms that coexist in gesture and sign. I adopted this approach in Studies III and IV of my dissertation. In Study III, we showed that iconic gestures used by hearing non-signing speakers of Yucatec Maya to depict the size and/or shape of objects (e.g. wide, round, tall, tiny, etc.) became nominal markers in YMSLs. When a SASS is attached to a YMSL sign that, in isolation, refers to an action, such as WASH-CLOTHES ‘to wash clothes’, the verb is transformed into a semantically related noun, such as WASH-CLOTHES+SASSWIDE-OBJECT ‘washing board’. The form changes from a more freely elaborated articulation in gesture, where speakers of Yucatec Maya tend to faithfully represent visual properties of an object in detail, to a shorter, more quickly articulated expression in YMSLs, where size and manner of movement are reduced (‘phonetic reduction’). At the same time, the meaning of the gesture changes from the more concrete, i.e. a description of a specific object’s size and shape, to the more abstract, i.e. the categorisation of a referent as an object with certain visual-geometric properties (‘semantic generalisation’). The results of Study IV are in many ways comparable to Study III. My data showed that Yucatec Maya height-specifier gestures for upright referents became person-classifiers in

YMSLs, which can be attached to a sign denoting gender or/and age, for instance MALE+SPEC-

H:HUM ‘man’. In the process of grammaticalisation, the place of articulation moves from an iconic location depicting the height of a referent to a neutral location, usually at head-height.58

Another change in form concerns non-manual features of the sign SPEC-H:HUM: During the articulation of the height-specifier gesture, the speaker’s eye gaze is directed towards the hand,

58 See also Janzen (2012: 830), for a change in place of articulation of the grammaticalised future marker WILL in ASL and LSF compared to the lexeme DEPART.

98 but in the case of the person-classifier, the eye gaze rests at the interlocutor or a neutral location.

Semantically, SPEC-H:HUM no longer depicts the height of a concrete referent but refers more generically to human beings as members of a class. Again, this is an instance of semantic generalisation. The grammaticalisation of YMSL person-classifiers from Yucatec Maya height-specifier gestures shares certain parallels with the noun PERSON that became grammaticalised in many sign languages, e.g. as an agreement auxiliary, a case marker or a pronoun (e.g. Pfau & Steinbach 2013; see Börstell 2017, for a detailed literature review).

However, unlike in YMSLs, the sign PERSON as the source of grammaticalisation in other sign languages, e.g. ASL or DGS, refers to the outline of a human referent; it is not derived from a measurement gesture and not descriptive in terms of height.

In terms of the structure of the construction, SASSes and SPEC-H:HUM behave very similarly: The SASS as well as the SPEC-H:HUM have a strong tendency to stand in a final position, following the action sign or the gender/age noun. This syntactic structure is more solid for constructions involving SPEC-H:HUM than for constructions involving a SASS. In the case of SASSes as nominal markers, we also detected some regional differences regarding their syntactic distribution. Signers from Chicán use the SASS in a final position more consistently, while in Nohkop, there is more variation. In Study IV, I discussed differences between SASSes and height-specifiers in sign languages and defined YMSL height-specifiers as forming a category between SASSes and entity classifiers. An important difference between SASSes and height-specifiers is that the latter are more restricted to a particular class of entity, with specific contrastive handshapes assigned to specific semantic categories of referents (human beings, four-legged animals, etc.). SASS handshapes, in contrast, can refer to a variety of referents that share certain visual characteristics. For instance, SASSROUND-OBJECT in YMSLs can apply to a tortilla, a calabash or a football – the exact translation results from the juxtaposition with a verb (MAKE-TORTILLA,

DRINK, PLAY-FOOTBALL) and/or can be derived from the context. In this sense, SASSes provide a broader categorisation of a referent than height-specifiers do and unlike height-specifiers, they allow for handshape variation (for instance, two index-handshapes or two 5-handshapes can be used for a SASSWIDE-OBJECT in YMSLs). However, we could argue that the difference between SASSes and height-specifiers might be more gradual than categorical, the form of the human height-specifier being more conventionalised than others because we “so frequently observe and discuss humans interacting” (Janzen 2012: 823). In that sense, the YMSL SPEC-

H:HUM prototypically came to represent humans rather than other upright entities (as it does in Yucatec Maya gesture).

99 Another interesting finding from Study IV concerns the height-specifier for four-legged animals SPEC-H:4LEG. Constructions involving SPEC-H:4LEG exhibit the same structure as

SASSes or SPEC-H:HUM: the specifier follows a sign that identifies the type of animal, such as

HORSE+SPEC-H:4LEG ‘horse’. This construction is frequent among signers from Chicán and Nohkop, but from preliminary observations, it does not seem to be grammaticalised to the same extent as SPEC-H:HUM. Frequency of use is said to be a driving force behind grammaticalisation

(see Bybee 2011), which might account for the fact that SPEC-H:HUM is a more conventionalised classifier form in YMSLs than SPEC-H:4LEG. Another difference in the results from Study III and Study IV is the type of sign that the specifier combines with and which function it serves. I showed that in YMSLs, a SASS can be attached to a verb and transform this verb into a noun. The SPEC-H:HUM, in contrast, is attached to a noun denoting gender and/or age and does not change the part-of-speech category of the base sign. In other contexts, SASSes can also be used in a similar way in YMSLs, for instance in constructions such as EGG+SASSSMALL-OBJECT ‘egg’ (see Meir et al. 2010b, for similar examples in ABSL). In this function, the SASS does not mark a noun-verb distinction but provides further information about visual properties of the referent. I also found that neither the addition of a SASS nor of SPEC-H:HUM is obligatory in YMSLs. Study III and Study IV contribute case studies about the grammaticalisation of classificatory gestures into sign language classifiers (see Pfau & Steinbach 2011, for DGS). To summarise, it seems that SASSes and height-specifiers in YMSLs have a lot in common. More research about the use and distribution of these forms in YMSLs as they appear in natural discourse is needed in order to understand if we are observing the emergence of some sort of classifier system in YMSLs.

8.2.3 The relation between lexicalisation and grammaticalisation

In this section, I will try to consolidate my findings on lexicalisation and grammaticalisation in YMSLs. Models like the one proposed in Wilcox’ (2005) “Route 1” of grammaticalisation are based on a linear view of the process by which lexicalisation precedes grammaticalisation. My findings from Study IV indicate that a dynamic view of lexicalisation and grammaticalisation as intertwined processes may be more adequate to capture the integration of co-speech gestures into a shared sign language. Wilcox’ model is reduced to two prototypical paths of change from gestures to sign, but the findings from YMSLs (and see also Dachkovsky et al. 2019 for ABSL) suggest that there might be more than two possible routes.

100 While in Wilcox’ model, the stages gesture > lexicon > grammar constitute distinct steps of language change, I argue that these phases can occur in parallel and are not always separable. Le Guen et al. (in press-a) consider the incorporation of height-specifiers into YMSLs an instance of grammaticalisation of conventional “manual classifiers” used in spoken Yucatec Maya. However, my findings show that this claim needs to be revised at least in the case of the human height-specifier, which became incorporated into YMSLs as both a lexical and a grammatical item. This is not the only known example for a Yucatec Maya gesture that was lexicalised as well as grammaticalised. Le Guen et al. (in press-a) mention, for instance, that the concept gesture úuch ‘a long time from now’ can be used as a lexical sign LONG-

TIME(FROM-NOW) but also as a temporal marker to express distant past or future (see also Le Guen 2012). Similarly, Tano and Nyst (2018) point out that body-part size-and-shape gestures common in West Africa became part of the lexicon and the grammar of Adamorobe Sign Language. Dachkovsky et al. (2019) show that a gesture that indicates the passing of time entered ISL as both a lexeme and a grammatical marker.59

In the case of height-specifiers, it would of course be tempting to propose that SPEC-

H:HUM first entered the lexicon of YMSLs as a noun for CHILD, YOUNGER-SIBLING,

DAUGHTER/SON, FATHER or (GRAND-)MOTHER and then further evolved into a person-classifier. In this way, its course of development would conform to Wilcox’ (2005) “Route 1” of grammaticalisation. What would be indicative of this path is the gradual change in the role of iconic (height-depicting) location and eye gaze at the hand: With the gesture and YMSL predicate, these components are obligatory, with the lexical nouns, they are optional and with the person-classifier, they are absent. We could thus talk of a gradual shift of these features along the grammaticalisation path. Ultimately this is, however, an empirical question: How can we tell what came first? In young languages, the pace of linguistic conventionalisation is rapid (see e.g. Senghas 2003; Coppola, in press, on NSL; McWhorter 2018, on creole languages) and both lexical and grammatical functions of the same form are used simultaneously, already among the first generation of deaf YMSL signers. Hopper (1991: 22) speaks of “layering” in grammaticalisation when several forms co-exist in a language with distinct meanings and differences in their lexical/grammatical status. While lexicalisation and grammaticalisation are often conceptualised as separate, opposed processes, some scholars (e.g. Brinton & Traugott 2005; Lightfoot 2011;

59 Unfortunately, their analysis lacks any description of the gestural equivalent of the ISL sign TIME-PASS. A more systematic comparison would be informative and could possibly reveal parallels to the grammaticalisation of TIME-PASS in YMSL (Le Guen 2012).

101 Himmelmann 2004) urge for a more integrated view of the two phenomena as interrelated. Brinton and Traugott (2005) argue that both lexicalisation and grammaticalisation are subtypes of language change that result in “adoption [of a form] into the inventory” of a language. This inventory is characterised by the coexistence of various forms with different degrees of referentiality, productivity, obligatoriness, etc. (Brinton & Traugott 2005: 92). My analysis of height-specifiers in YMSLs demonstrates that lexicalisation and grammaticalisation are processes that can occur in parallel, while the gestural source of SPEC-

H:HUM continues to be used by hearing people along with spoken Yucatec Maya. As expected from the literature on grammaticalisation in general (see e.g. Hopper 1991) and on the grammaticalisation of co-speech gestures into signs (e.g. Wilcox 2005; Janzen 2012), a gestural form becomes polyfunctional in sign language. Some functions (or, in Hopper’s (1991) terms, layers) of SPEC-H:HUM in YMSLs are close to their gestural precursor, others are further removed. The case of YMSLs shows that developments from gesture to sign do not always happen in linear and predictable ways. The approach to lexicalisation and grammaticalisation as dynamic synchronic processes could be fruitful to study the incorporation of conventional gestures into shared sign languages more generally.

8.2.4 Sociolinguistic factors that play a role in sign language emergence: Lessons from YMSLs

The comparison of shared sign languages in four Yucatec Maya communities that are located in the same geographic region and share similar cultural traits, yet differ in certain sociolinguistic features offers important insights on how language emerges in interaction and which social and cultural factors are influential herein.

● Cultural attitudes toward deafness and sign language

Throughout my dissertation, I have argued that rural Yucatec Maya communities constitute a fruitful environment for sign languages to develop and flourish. Deaf people who are born into the context of a Yucatec Maya village have a communicative advantage compared to those who grow up in cultural environments with a prevailing oralist ideology that are less favourable for sign language creation (see Le Guen et al., in press-a). The Yucatec Maya have a rich inventory of conventional co-speech gestures and readily turn to visual communication

102 when confronted with a person who cannot hear. Deaf people are regarded as equally competent and intelligent members of society as hearing people. Both deaf and hearing people state that they find it easy to communicate with other people in their village as they grew up together and are used to each other (see Study I). This familiarity and experiential knowledge of sign language shapes the relationships between deaf and hearing people in the communities (see also Kisch 2008). Deaf and hearing community members also emphasise that in order to be able to unfold their potential, deaf people need communication in a modality which is fully accessible to them, i.e. sign language (see Study I). The Yucatec Maya consider sign language to be a fully functional and valuable tool of communication, which has the potential to cover all possible topics of conversation, including complex subjects, for instance medical terms. Deaf as well as hearing people recognise that sign language is something you need to learn in interaction with deaf signers. People from outside the community do not “automatically” know sign language although Yucatec Maya conventional gestures form a foundation, which facilitates communication between deaf and hearing people. In Study I, I concluded that these positive attitudes towards deafness and visual communication are the sine qua non for hearing community members to engage in regular and extended interactions with their deaf relatives and peers and are thus fundamental for the emergence of shared sign languages. In a similar vein, Green (2014) concluded that in many cases, successful communication is more dependent on the participants’ willingness to enter a dialogue than on the presence of a shared language. Even though such positive, non- discriminatory attitudes have also been described for village/shared sign languages in other parts of the world, e.g. by Nonaka (2012) on Ban Khor Sign Language in Thailand, this is by no way a general characteristic of all village/shared sign languages (see Kusters 2010, for a critical discussion). For instance, Haviland (in press) describes that in Zinacantán – a community in rural Chiapas, Mexico – hearing people frequently categorise deaf people as uma’ (‘dumb’) or chich (‘foolish’). Even though the Zinacantán signing community is located in the same country as the Yucatec Maya signing communities and forms part of another Mayan cultural and linguistic group, deafness is considered a “severe disability” among the Tzotzil Maya in Chiapas, whereas among the Yucatec Maya in Yucatán, it is not. Such a difference in attitude obviously not only determines deaf people’s opportunities in terms of communication access but also impacts their prospects for obtaining a job or starting a family. The wide range of attitudes even within the restricted geographic area of Mesoamerica underlines the need for a fine-grained ethnographic investigation of the respective sociocultural environment.

103 ● Family structures

Another crucial factor for the emergence of YMSLs is the characteristic family structures in rural Yucatán. The patterns of socialisation and interaction in Yucatec Maya villages are very much centred around the extended family (see also Hou 2016 for SJQCSL), whose members typically live on a shared family compound (solar). Children in Yucatec Maya villages are brought up within the broader network of the extended family, with not only their parents as the main caregivers but also their grandparents, aunts, uncles, cousins, etc. This creates a situation where a deaf person born into a Yucatec Maya community has access to many more close interaction partners than just his or her parents and siblings. It also means that if there is a deaf person in the extended family, hearing people grow up signing with them and acquire YMSL from an early age. Usually the immediate family members, especially those who are younger or of a similar age group as the deaf person, e.g. younger siblings or cousins, are the most proficient hearing YMSL signers. The extended family being the main environment of social interaction also entails that deaf people who are not kin-related usually do not have much contact with each other. Deaf people are much more likely to interact with hearing people from their own family than with deaf people from another family. This is a crucial factor for the emergence of shared sign languages in the Yucatec Maya context. Typical cultural networks of interaction shape patterns of sociolinguistic variation and give rise to “familylects” (see Section 7.3.2.1) that can be quite distinct within the same community. Today, under the influence of schooling, social networks of interaction are changing and children are more likely to be in regular contact with peers who are not members of their family. This situation opens up new possibilities for peer-to-peer language transmission of YMSLs.

● Language age and intergenerational transmission

Language age and the transmission of a sign language from one generation (or cohort) of deaf signers to the next are often said to be the main factors for the creation of new and increasingly complex linguistic structures (e.g. Senghas & Coppola 2001, for NSL; Sandler et al. 2005, for ABSL). The studies of my dissertation show that language age is an important but not the only relevant feature. A range of grammatical structures is present already among the

104 first generation of YMSL signers, e.g. SASSes for the noun-verb distinction (Study III) or height-specifiers as person-classifiers (Study IV), indicating that languages are complex from an early stage. The YMSL data confirm that intergenerational transmission is indeed a factor giving rise to new linguistic structures among the younger generation and supporting the conventionalisation of features that were used variably among the older generation. Various differences were detected between signers from Chicán (a community with three generations of deaf signers) and Nohkop, Trascorral and Cepeda (communities with one generation of deaf signers), concerning for instance the consistency of iconic patterns (Study III) or the syntactic position of SASSes (Study III) and person-classifiers (Study IV). Even more compellingly, my studies revealed important linguistic differences between younger signers in Chicán who grew up with deaf parents or other deaf adult language models and other groups of signers. Younger signers who learnt sign language primarily from deaf adults stand out in their use of SASSes for the noun-verb distinction (Study III) or of person-classifiers (Study IV). Another study (Le Guen & Safar, in prep.) shows that the same young signers also differ from others in their grammatical use of the signing space for directional verbs. However, as I have argued in Section 7.3.2, the interaction of several social variables plays a role in shaping sociolinguistic variation in YMSLs. Further studies are needed in order to disentangle the effects of intergenerational transmission on the emergence of new linguistic structures in YMSLs.

● “Critical mass” of deaf people?

One central issue in studies of sign language emergence is how many deaf people it takes to create a sign language (see e.g. Brentari et al. 2012: 17). There is no definite answer to this question, but the consensus seems to be that one is not enough and that “a linguistic community may be crucial for individual users to experience the system not only as conveyers of information but also as receivers of information” (Brentari & Goldin-Meadow 2017: 365). From this perspective, the communication system of an isolated deaf person without access to a linguistic community would rather be qualified as an instance of homesign (see e.g. Goldin- Meadow 2003).60 Homesign systems, i.e. the signing of isolated deaf individuals who have not

60 Some researchers, however, e.g. Fusellier-Souza (2006) or Martinod et al. (in press) call the signing of each isolated deaf individual in their studies a “sign language” and use this label and the term “homesign” interchangeably. Also Hou (2016) and Mesh (2017) opt to use the term “language” irrespective of the level of maturity or complexity of a system.

105 been exposed to an established sign language, are generally considered to “contain many, but not all of the properties of natural language” (Brentari & Goldin-Meadow 2017: 364; Goldin- Meadow 2003). Compared to (in Brentari & Goldin-Meadow’s terms) “fully-blown sign languages” homesign systems have a lower level of conventionalisation and linguistic complexity, in terms of their phonology, lexicon, morphology and syntax (see e.g. Goldin- Meadow 2003; 2005; Brentari & Coppola 2012; Brentari & Goldin-Meadow 2017). However, some researchers, e.g. Nyst et al. (2012) or Hou (2016), criticise that the term homesign originated in a specific sociolinguistic context, in which deaf signers do not have access to visual communication and are socialised to speak, as it is often the case, for instance, in the US or Europe. This definition may not always apply, for instance, to deaf individuals who grow up in rural settings without any access to deaf education and without a prevailing oralist ideology. Comparative studies including a variety of sociolinguistic settings in different geographic and cultural contexts are needed in order to achieve a more fine-grained classification of homesign systems (see Nyst et al. 2012). The delimitation “homesign” versus “sign language” is arbitrary, as natural languages can display diverse degrees of complexity on various linguistic levels. There are no reliable linguistic criteria to delineate the two types of systems from each other and to determine when a language has achieved the status of being “fully-blown”. Horton (in press-a) argues that the umbrella term “homesign” can refer to a variety of sociolinguistic situations and that the communicative experiences of child homesigners significantly shape the signed system they develop. The density of networks that a deaf child participates in is crucial for the emergence of linguistic structures and the presence of multiple deaf signers who are in regular contact spurs the solidification of linguistic conventions. Similarly, Coppola (in press) argues that the “richly-connected networks”, in which a shared community sign language is the medium of all interactions, accelerate the conventionalisation of linguistic structures compared to “sparsely-connected networks”, in which hearing people use sign language in interactions with an isolated deaf person but spoken language with each other. Richie et al. (in press) corroborate these findings with results from computational modelling, but they emphasise that such experiments need to be complemented with empirical data from observational fieldwork, because simulation models can mirror the complex reality of interactional networks only to a certain extent. How big the “critical mass” of deaf people in regular contact needs to be for a sign language to emerge remains a matter of debate and Nyst (2012: 568) points out that the threshold from homesign to a shared sign language may be sliding. In any case, the mere number of deaf people is not the only factor that determines the evolution of a signed linguistic system. Probably more relevant than

106 community size are dense networks between deaf signers as well as positive attitudes toward sign language, which open up opportunities for deaf people to engage in interactions with a variety of interlocutors. Importantly, in the case of shared sign languages it is not enough to focus only on the number of deaf signers, as hearing people also play a crucial role in sign language creation. The case of YMSL in Nohkop shows that even a group of only four deaf siblings (who are very close to each other in age) and their hearing interlocutors were able to create a complex signed system with stable lexical and grammatical conventions.

9 Conclusions

My dissertation makes relevant contributions to a number of disciplines and areas of research: to sign language linguistics, by providing a description of largely undocumented indigenous minority sign languages; to studies of language evolution, by investigating how shared sign languages emerge and how they change over time; to Deaf studies, by giving an account of how deaf people socialise and communicate in Yucatec Maya villages; to gesture studies, by analysing the interface between Yucatec Maya multimodal communication and Yucatec Maya Sign Languages; to sociolinguistics, by showing how the social constellation of Yucatec Maya signing communities has an impact on the structure of sign languages that emerge in that context; to linguistic anthropology, by pointing out how Yucatec Maya cultural elements shape interactions between deaf and hearing people; and to linguistic typology, by exploring how specific lexical and grammatical structures of YMSLs compare to other rural or urban signed languages. After having worked on my dissertation project for four years, what main conclusions can I draw? When we engage closely with a topic over an extended period of time, we have the impression of only ever scratching the surface and for every answer we find, many new questions arise. I will try to sum up the essence of my thesis in a few points:

● YMSLs are similar yet unrelated sign languages. They overlap in their lexicon and grammar due to the incorporation of Yucatec Maya gestural elements and because of shared cultural features.

● The high degree of intra-community linguistic variation reflects the sociolinguistic landscape of the communities.

107 ● Different lexical domains can be placed on a continuum of variation.

● Many linguistic structures are already present among the first generation of YMSLs signers, but the languages change over time. New lexical and grammatical conventions arise when the sign language is passed on to the next generation.

● Co-speech gestures used along with spoken Yucatec Maya serve as a substrate for YMSLs and undergo different paths of conventionalisation when they are integrated into the sign languages.

● The development from gestures to signs is multi-layered and does not always happen in linear and predictable ways.

The documentation of indigenous micro-community sign languages like YMSLs is an urgent issue because they are endangered languages that exhibit typologically unique features, such as Chicán’s unusual numeral system or the use of the signing space – and possibly many more, which have not yet been uncovered. Being at the same time emerging and endangered, there is a chance that YMSLs will stop being used before we are able to get a proper grasp of their linguistic structures and how they evolve over the years. I hope that the rapidly expanding field of research on village/ shared/ indigenous/ emerging/ micro-community/ “grey-area” sign languages will soon include more comparative studies such as the present one, both within a bounded geographical region as well as across different countries and continents. This will allow us to comprehend the true diversity of shared sign languages, and define the common linguistic and sociolinguistic traits that characterise them as a group.

108 Appendix A: Coding protocols

Coding protocol Study I Translanguaging in Yucatec Maya signing communities

1. Selection of video materials

The analysis in this study is based on:

a. Video-recorded conversations between deaf and hearing participants from Chicán and Cepeda Peraza and between deaf and hearing participants from Chicán and Nohkop (03:48:29; 5 YMSL signers from Chicán, 1 YMSL signer from Nohkop and 1 YMSL signer from Cepeda Peraza)

b. Video-recorded semi-structured interviews (03:07:02; 7 YMSL signers from Chicán and 1 YMSL signer from Trascorral)

c. Unrecorded informal interviews, fieldwork observations and field notes by Josefina Safar

The data used for this study was collected by Josefina Safar and Olivier Le Guen.

2. Identification of signers’/speakers’ semiotic resources in conversations

The video-recorded conversations were annotated using the program ELAN.61 I did not transcribe the total amount of data but selected stretches of video where the participants combined a variety of semiotic resources, either sequentially or simultaneously, or both. In the selected video sections, I identified the semiotic resources used by signers and speakers and made annotations on the following tiers:

Table 1: ELAN tiers used for the annotation of semiotic resources

Tier name Tier description Example Translation English Translation of the utterance into How old is your father? English YMSL_DH62 Participant uses his/her variety of BEANS YMSL; Articulation on the dominant hand YMSL_NonDH Participant uses his/her variety of TWENTY YMSL; Articulation on the non- dominant hand Spoken Yucatec Maya Participant speaks Yucatec Maya Bix?

61 ELAN can be downloaded free via the following URL: https://archive.mpi.nl/tla/elan. 62 I created individual tiers for different participants and individual letters were assigned to the participants. For instance, the tier YMSL_DH_a was used to make annotations of articulations on the dominant hand of participant a.

109 Spoken Spanish Participant speaks Spanish Un poco, casi no.

Translating Participants makes a translation from PRO2 AIRPLANE PRO2 LONG- one language to another TIME-FROM-NOW QU-MARK/ ¿Cuando viniste? (‘When did you arrive?’) Emblem Participant uses an emblematic money gesture Enactment Participant performs an enactment of put on make-up an activity Pointing Participant points to a visible or pointing to a plant on the invisible location or object compound Object handling Participant manipulates a real-world handling a purse object Fingerspelling Participant makes use of the LSM P-I-Ñ-A (‘pineapple’) manual alphabet Mouthing Participant silently articulates words [agua] (‘water’) or parts of words from spoken Spanish or Yucatec Maya Writing Participant writes on paper, ground or writing a number with a pencil palm on the palm Drawing Participant draws on paper, ground or drawing an animal with a pencil palm into a notebook Other Other semiotic resources scratching head NonMan Non-manual features head nod Eye gaze Direction of participant’s eye gaze at phone

3. Describing communicative strategies

Based on the annotations on the tiers described in Table 1, I was able to describe relevant characteristics of the conversation, for instance

- Chaining, i.e. the sequential use of various semiotic resources (Humphries & MacDougall 1999) - Simultaneous use of various semiotic resources - Turn-taking - Pauses - Repairs/corrections - etc.

4. Interviews with YMSL signers

I conducted semi-structured interviews with deaf YMSL signers (in YMSL, with hearing research assistants as interpreters) and hearing YMSL signers (in Spanish). Examples for interview questions were:

• I have noticed that people from different families in the village sometimes use different signs. Can you understand the signs from other families?

• What are the differences between the signs of young people and those of older people?

110

• (to deaf signers) When hearing people speak Yucatec Maya with each other, they also move their hands. How is this different from the signs that deaf people use? Can you understand what they say?

• (to deaf signers) When you meet someone on the street who does not know sign language, for instance a person from Mérida63 or Kinil64 – how will you communicate with him or her?

• (to hearing signers) How/when did you learn sign language?

• Have you ever met a deaf person from another village in Yucatán? Could you understand his/her signs? What are the differences from the signs used in this village?

• What are the differences between the sign language used in the village and the sign language you see on TV?

• Can you talk about everything in sign language?

In the transcribed interviews (or sections of interviews), I identified relevant themes, for instance:

- Variation between families - Variation between villages - Age-related variation - YMSL vs. Mexican Sign Language (LSM) - YMSL vs. co-speech gestures Etc.

These themes were further divided into sub-themes, for instance:

- “Signs from other villages are the same as ours” - “Signs from other villages are different from ours” - “Mexican Sign Language is complicated” - “Sign language and gesture are the same” - “Sign language is more exact than gestures” - “Young people use new signs” Etc.

5. Field notes

Additionally, I identified and coded relevant themes in the field notes based on my observations and informal interviews. These included for instance

- Interventions by the government/NGOs/church - Education opportunities for deaf people - Employment opportunities for deaf people

63 Capital of Yucatán 64 Neighbouring village of Chicán

111 - Relationships between deaf and hearing people - Acquisition of YMSL Etc.

These themes were further divided into sub-themes, for instance

- “Deaf people do not need hearing aids” - “Learning LSM is not necessary” - “Deaf and hearing people are used to each other” - “Working on the milpa65 is more difficult than it used to be” - “Hearing people grow up with YMSL” Etc.

References

Humphries, Tom & Francine MacDougall 1999. “Chaining” and other Links. Making connections between American Sign Language and English in two types of school settings. Visual Anthropology Review 15. 84–94.

65 Field used to grow corn and other crops for subsistence agriculture

112 Coding protocol Study II Numeral variation in Yucatec Maya Sign Languages

1. Selection of video materials for analysis

The analysis in this study is based on:

a. Data from an elicitation task using picture stimuli (01:52:39; 27 YMSL signers from Chicán, Nohkop and Cepeda Peraza)

b. Narratives, conversations and interviews (01:48:19; 12 YMSL signers from Chicán, Nohkop and Cepeda Peraza)

c. Fieldwork observations by Josefina Safar, Olivier Le Guen, Geli Collí Collí and Merli Collí Hau; Field notes by Josefina Safar

The video data used for this study was collected by Josefina Safar.

2. Annotation of participants’ responses to photo stimuli

The data for this study was annotated using the program ELAN.66 For the data from the elicitation task, annotations were created in the following way:

a. Selection of video sections for annotation

Signed utterances were annotated that overlap with stretches of video during which I (Josefina Safar) showed the stimulus pictures to the participants on an iPad.

b. Glossing conventions

We used the following symbols to annotate YMSL number signs, according to the conventions used by Zeshan et al. (2013):

- The ^ symbol was used to annotate sequential combinations of numbers, e.g. TWENTY^TEN for ‘30’.

- The # symbol was used to annotate simultaneous articulations of numbers on the dominant and the non-dominant hand, e.g. FIVE#TWO for ‘7’.

Moreover, it was necessary to distinguish different number variants in YMSLs. If more than one variant of a number sign occurred in the dataset, individual labels were assigned to identify the variants, listed in parentheses after the number gloss. For instance, ZERO signed with an S-handshape was annotated using the label ZERO(S) as opposed to ZERO signed with an O-handshape, which was glossed as ZERO(O).

66 ELAN can be downloaded free via the following URL: https://archive.mpi.nl/tla/elan.

113 c. Annotation tiers

Annotations were created on the following tiers. Note that not every sign has an entry on all tiers – on some tiers, annotations were only created if that specific parameter was relevant to distinguish different number variants.

Table 1: ELAN tiers used for the annotation of YMSL number signs

Tier name Tier description Example Stimulus Target number depicted on the photo 60 stimulus Gloss_DH Articulation on the dominant hand ZERO(O) Gloss_NonDH Articulation on the non-dominant FIVE hand Gloss Number English gloss of the number sign SIXTY Number composition Compositional representation of the FIVE#ONE^ZERO(O) number sign, based on the sequential and/or simultaneous combination of number signs on the dominant and/or the non-dominant hand Orientation Orientation of the palm during the palm-in articulation of the number sign (if palm-out relevant) palm-down

NonMan Non-manual markers that co-occur pursed lips with the articulation of manual signs squint (if relevant) brow raise blink Movement Movement parameter of the sign (if tense forward relevant) wrist rotation Point of contact Location on the body where contact wrist with the active manual articulator is thumb made (if relevant) Thigh palm Handshape Handshape of the dominant or non- V dominant hand (if relevant) Mouthing Silent articulation of words or parts of [pe] with the manual sign PESO words from spoken Spanish or Yucatec Maya (if relevant) Repetition Number of repetitions of a number ++ for two repetitions sign (if relevant) Comments Additional comments made by the Participant produced a number annotator sign that deviates from the number in the stimulus

3. Creation of annotations by researcher and research assistants, cross-checking with participants

114 Most transcriptions were made during my fieldwork, in collaboration with two research assistants (Geli Collí Collí, Merli Collí Hau), who are native YMSL signers from Chicán. The annotation process was as follows: The two research assistants identified the sections where participants responded to a stimulus item, created segmentations of the video files (where all segments of the video, in which the participant produced a sign, were marked), and subsequently created annotations in Spanish for all these segments on the following tiers: Gloss_DH, Gloss_NonDH, Mouthing, NonMan.67 They used the “Controlled Vocabulary (CV)” function in ELAN, where annotations can be made by selecting from a drop-down list, but they were also able to enter annotations manually in case the correct gloss was not included in the CV. I imported Merli’s and Geli’s annotations into the template with the tiers listed in the table above, translated the Spanish annotations into English and coded the annotated segments on the remaining tiers. Afterwards, the research assistants and I went through the annotations together and discussed unclear cases. In cases where the three of us were not sure about the correct translation of a specific number sign, we consulted the respective participants. This usually helped to resolve the issue. In the case of the sign HUNDRED(deer), the incongruent explanations by different community members led us to identify this variant as part of the familylect of one extended family.

4. Criteria for inclusion or exclusion of a number sign for analysis

One response for each stimulus item was included in the analysis. If a participant produced more than one response to the stimulus, only the first response was included (unless the participant corrected him- or herself and thus provided a different response than the first one). When a signer responded with a number sign that did not correspond to the stimulus item, it was excluded from the analysis. In cases, where a participant provided a visual description of the arrangement shown in the photo stimulus (e.g. TWO BILL FIVE-HUNDRED instead of the total amount ‘1000’), these were also excluded.

5. Categorisation of number signs

The number signs were assigned to the following categories, based on the strategies described by Zeshan et al. (2013) and Sagara and Zeshan (2016).

Table 2: Number strategies in YMSLs

Strategy Definition Example Lexical numerals Number is expressed by a TWENTY in YMSL of Chicán separate lexeme Additive numerals Number is expressed by FIFTY^TWENTY for ‘70’ in YMSL addition of two or more numeral of Chicán signs

67 The research assistants worked with a template in Spanish they had been using previously for other transcriptions. It includes the tiers man-der (mano derecha, ‘right hand’), man-izq (mano izquierda, ‘left hand’), boca (‘mouth’) and cara (‘face’). Since all participants of the study are right-handed, man-der and man-izq are equivalent to the tiers Gloss_DH and Gloss_NonDH used by myself. For the purpose of this study, the tiers boca and cara correspond to the tiers Mouthing and NonMan.

115 Digital numerals Number is expressed by signing THREE^ZERO for ‘30’ in YMSL of a sequence of individual digits Nohkop as they appear in written numbers Context-dependent numerals Number handshape can refer to 3-handshape for ‘30’ in YMSL of multiples of 1, 10, 100 or 1000, Cepeda depending on the context Simultaneous morphology Number handshape is Index (1-handshape) with tense simultaneously combined with a forward movement and non- specific movement to express manual marker for ‘100’ in YMSL multiples of 100 or 1000 of Chicán Writing Number is expressed by writing it with a stick on the ground Tracing Number is expressed by tracing its outline with the finger in the air, on the palm, the forearm or a wall

6. Identification of number signs in narratives, conversations and interviews

In addition to the data from the elicitation task, I analysed the use of YMSL number signs in narratives, conversations and interviews. I used 9 video files that had already been transcribed (or partly transcribed) by the research assistants or myself. These videos had originally been annotated using the tiers Gloss_DH and Gloss_NonDH and Translation (Spanish or English). I identified instances of number signs in the annotations and coded the relevant sections for the specific features relevant to this study (see Table 1). Then I assigned the number signs to the categories listed in Table 2 and checked whether any of the sign’s parameters (movement, handshape, location, orientation, non-manuals) differed from the number signs identified in the elicited data.

7. Comparison with number gestures of hearing Yucatec Maya

The account of number gestures used by hearing non-signing speakers of Yucatec Maya and Mexican Spanish is based on observations by the second author, Olivier Le Guen, during 17 years of fieldwork in Yucatán and nine years of living in Mexico City.

References

Sagara, Keiko & Ulrike Zeshan. 2016. Semantic fields in sign languages – A comparative typological study. In Ulrike Zeshan & Keiko Sagara (eds.), Semantic fields in sign languages. Colour, kinship and quantification, 3–37. Berlin: De Gruyter Mouton.

Zeshan, Ulrike, Cesar Ernesto Escobedo Delgado, Hasan Dikyuva, Sibaji Panda & Connie de Vos. 2013. Cardinal numerals in rural sign languages: Approaching cross-modal typology. Linguistic Typology 17(3). 357–396.

116 Coding protocol Study III Strategies of noun-verb distinction in Yucatec Maya Sign Languages

Sub-Study 1: SASSes as nominal markers

1. Selection of video materials for analysis

The analysis in this study is based on:

a. Data from an elicitation task using picture stimuli (03:59:00; 22 YMSL signers from Chicán, Nohkop and Cepeda Peraza; 01:13:04; 4 hearing gesturers from Kopchen)

b. Narratives, conversations and interviews (01:33:24; 6 YMSL signers from Chicán and Nohkop 00:59:19; 4 hearing gesturers from Kopchen, 1 hearing gesturer from Cepeda Peraza)

c. Fieldwork observations by Josefina Safar, Rodrigo Petatillo Chan and Olivier Le Guen; Field notes by Josefina Safar

The video data used for this study was collected by Josefina Safar and Olivier Le Guen.

2. Annotation of participants’ responses to photo stimuli

The data for this study was annotated using the program ELAN.68 For the data from the YMSL elicitation task, annotations were created in the following way:

a. Selection of video sections for annotation

Signed utterances were annotated that overlap with stretches of video during which I (Josefina Safar) showed the stimulus pictures to the participants on an iPad.

b. Glossing conventions

To gloss constructions involving a Size-and-Shape specifier (SASS), I used the + symbol (e.g. PUT-ON-LIPSTICK+SASSSMALL OBJECT), following the conventions used in similar studies, e.g. by Tkachman and Sandler (2013) or Tkachman and Meir (2018).

In order to distinguish different lexical variants in YMSLs, individual labels were assigned to identify each variant, listed in parentheses after the sign gloss. For instance, DRINK signed with an C-handshape was annotated using the label DRINK(C) as opposed to DRINK signed with an Y-handshape, which was glossed as DRINK(Y).

c. Annotation tiers

Annotations were created on the following tiers.

68 ELAN can be downloaded free via the following URL: https://archive.mpi.nl/tla/elan.

117 Table 1: ELAN tiers used for the annotation of SASSes in YMSL

Tier name Tier description Example Stimulus Concept depicted on the photo Knife stimulus Drink out of glass Gloss_DH Articulation on the dominant hand CUT Gloss_NonDH Articulation on the non-dominant DRINK(C) hand SASS Presence vs. absence of a SASS SASS no SASS SASS type Class of object that the SASS refers cylindrical object to, static vs. tracing type of SASS cylindrical object (tracing) flat object flat object (tracing) tall object tall object (tracing) wide object wide object (tracing) round object small object square object thin object thin object (tracing) SASS handshape Handshape of the SASS on the C dominant and/or non-dominant hand SASS position Position of the SASS in relation to the initial base sign medial final repeated SASS only Comments Additional comments made by the Participant used a real-world annotator object

d. Creation of annotations by researcher, cross-checking with research assistants

Annotations were made by myself (Josefina Safar), in collaboration with three research assistants from Chicán (Geli Collí Collí, Merli Collí Hau, Dianela Collí Chim) and one from Nohkop (Rossy Kinil Canché). The annotation process was as follows: I identified the sections where participants responded to a stimulus item, created segmentations of the video files (where all segments of the video, in which the participant produced a sign, were marked), and subsequently created annotations in English for all these segments on the tiers listed in Table 1. I used the “Controlled Vocabulary (CV)” function in ELAN, where annotations can be made by selecting from a drop-down list, but it is also possible to enter annotations manually in case the correct gloss is not included in the CV. The lexical elicitation task for objects and activities included more items than the ones included in this study. Only the items selected for the purposes of this study (22 objects, 22 activities) were annotated and coded on the tiers described in Table 1. In cases where I was unsure about the correct translation of a sign, I consulted the research assistants, who are native signers of YMSL. This helped to resolve ambiguities.

118 3. Criteria for inclusion or exclusion of a SASS for analysis

One SASS response per participant per stimulus item was included in the analysis. If a participant produced more than one SASS in response to the stimulus, only the first SASS was included. In cases, where a participant did not provide any response to a stimulus or used other strategies than YMSL signs (e.g. pointing to or handling a real-world object), these were excluded from the analysis.

4. Identification of SASSes in conversations

In addition to the data from the elicitation task, I analysed the use of SASSes in YMSL conversations. I used 3 video files that had already been transcribed (or partly transcribed) by the research assistants or myself. These videos had originally been annotated using the tiers Gloss_DH and Gloss_NonDH and Translation (Spanish or English). I identified instances of SASSes in the annotations and coded the relevant utterances for the specific features relevant to this study (see Table 1). Moreover, I determined the syntactic position of the SASS in the clause. I focused on comparing instances of verbs in isolation and the same verbs in combination with a SASS (such as DRINK vs. DRINK+SASSTALL-OBJECT) in order to identify differences in their syntactic position and the co-occurrence with other parts-of-speech.

5. Comparison with Yucatec Maya size-and-shape gestures

As it was not the primary focus of the study, the gestures of hearing Yucatec Maya participants were not transcribed and coded at the same level of detail as the YMSL data.

a. Identification of Yucatec Maya size-and-shape gestures in elicited data

In Yucatec Maya gesturers’ responses to the picture elicitation task I identified instances of size-and-shape gestures and checked whether the types of size-and-shape gestures and the hand configuration of size-and-shape gestures corresponded to the SASSes used by YMSL signers.

b. Identification of Yucatec Maya size-and-shape gestures in natural discourse

In order to find out whether the silent gestures produced by hearing speakers of Yucatec Maya under the conditions of an elicitation task are also used in spontaneous speech, i.e. along with spoken Yucatec Maya, I looked at 6 videos of narratives and interviews (four recorded in Kopchen, two recorded in Cepeda Peraza). I identified instances of size-and-shape gestures and checked whether the types of size-and-shape gestures and the hand configuration of size- and-shape gestures corresponded to the SASSes used by YMSL signers.

References

Tkachman, Oksana & Wendy Sandler. 2013. The noun-verb distinction in two young sign languages. Gesture 13(3). 253–286.

Tkachman, Oksana & Irit Meir. 2018. Novel compounding and the emergence of structure in two young sign languages. Glossa: A Journal of General Linguistics 3(1). 136. http://doi.org/10.5334/gjgl.632.

119 Sub-Study 2: Patterned iconicity in YMSLs and Yucatec Maya gesture

1. Selection of video materials for analysis

The analysis in this study is based on:

a. Data from an elicitation task using picture stimuli (02:45:08; 21 YMSL signers from Chicán, Nohkop and Trascorral 01:57:23; 11 hearing gesturers from Kopchen)

b. Fieldwork observations by Olivier Le Guen and Rodrigo Petatillo Chan

The video data was collected by Olivier Le Guen.

2. Annotation of participants’ responses to photo stimuli

The data for this study was annotated using the program ELAN.69 For the data from the YMSL elicitation task, annotations were created in the following way:

a. Selection of video sections for annotation

Signed utterances were annotated that overlap with stretches of video during which Olivier Le Guen showed the stimulus pictures to the participants on an iPad.

b. Annotation tiers

Annotations were created on the following tiers.

Table 1: ELAN tiers used for the annotation of instrument/handling handshapes in YMSL and Yucatec Maya gesture

Tier name Tier description Example Stimulus Nb Number of stimulus and concept 01-Brush teeth depicted on the stimulus picture Mano derecha Articulation on the right hand BRUSH-TEETH Mano izquierda Articulation on the left hand BROOM Instrument Participant produced a handshape BRUSH-TEETH(instrument) where the participant’s hand represents an object Handling Participant produced a handshape BRUSH-TEETH(handling) where the participant’s hand represents a hand manipulating an object Other Participant produced a response other Touching clothes, than instrument or handling pointing at broom

69 ELAN can be downloaded free via the following URL: https://archive.mpi.nl/tla/elan.

120 3. Creation of annotations by researchers

Annotations were made by Rodrigo Petatillo Chan, in collaboration with Olivier Le Guen. The annotation process was as follows: Rodrigo Petatillo Chan identified the sections where participants responded to a stimulus item, created segmentations of the video files (where all segments of the video, in which the participant produced a sign or gesture, were marked), and subsequently created annotations in English for all these segments on the tiers listed in Table 1. The elicitation task included more items than the ones included in this study. In the pilot study, some items proved to be not culturally adequate and were thus excluded from the analysis. Only the items selected for the purposes of this study (24 items for sub-Study 2a, 24 items for sub-Study 2b) were annotated and coded on the tiers described in Table 1. Rodrigo Petatillo Chan’s codings were cross-checked with Olivier Le Guen. Possible doubts were discussed and decided in agreement between the two coders.

4. Criteria for inclusion or exclusion of a response for analysis

One response per participant per stimulus item was included in the analysis. If a participant produced, for instance, first a handling and then an instrument handshape in response to the stimulus, only the first response was included.

121 Coding protocol Study IV “When you were that little…” – From Yucatec Maya height-specifier gestures to Yucatec Maya Sign Language person-classifier signs

1. Selection of video materials for analysis

The analysis in this study is based on:

a. Data from an elicitation task using picture stimuli (02:02:41; 28 YMSL signers70 from Chicán, Nohkop, Trascorral and Cepeda Peraza; 00:16:50; 3 hearing gesturers from Kopchen)

b. Narratives, conversations and interviews (04:03:03; 19 YMSL signers from Chicán, Nohkop and Cepeda Peraza 01:13:23; 8 hearing gesturers from Kopchen and Chemax)

c. Fieldwork observations by Josefina Safar and Olivier Le Guen; Field notes by Josefina Safar

The video data used for this study was collected by Josefina Safar and Olivier Le Guen.

2. Annotation of participants’ responses to photo stimuli

The data for this study was annotated using the program ELAN.71 For the data from the YMSL elicitation task, annotations were created in the following way:

a. Selection of video sections for annotation

Signed utterances were annotated that overlap with stretches of video during which I (Josefina Safar) showed the stimulus pictures to the participants on an iPad.

b. Glossing conventions

To gloss constructions involving a height-specifier (glossed as SPEC-H:HUM), I used the + symbol (e.g. FEMALE+SPEC-H:HUM), following the conventions used in similar studies, e.g. by Tkachman and Meir (2018).

In order to distinguish different lexical variants in YMSLs, individual labels were assigned to identify each variant in the dataset, listed in parentheses after the sign gloss. For instance, FEMALE, articulated by pinching one’s earlobe with thumb and index finger was annotated using the label FEMALE(ear) as opposed to FEMALE articulated by touching the side of one’s head with a B- or 5-handshape, which was glossed as FEMALE(head).

70 A total of 34 YMSL signers participated in the study. For 14 participants, I analysed elicited data only; for 15 participants, I analysed both elicited and narrative/conversational data; for 4 participants, I analysed narrative/conversational data only. 71 ELAN can be downloaded free via the following URL: https://archive.mpi.nl/tla/elan.

122 c. Annotation tiers

Annotations were created on the following tiers.

Table 1: ELAN tiers used for the annotation of human referents in YMSL

Tier name Tier description Example Stimulus Concept depicted on the photo Baby stimulus Old woman Gloss_DH Articulation on the dominant hand FEMALE(head) Gloss_NonDH Articulation on the non-dominant A-LOT hand HHS_Height72 Height of articulation of SPEC-H:HUM low (waist-high or below) mid (between waist-high and shoulder-high) high (above shoulder-high) neutral (in neutral signing space)73 moving up (hand moves from a lower to a higher position) HHS_Position Position of SPEC-H:HUM in relation to initial the base sign medial final HHS only Eyegaze Direction of the signer’s eyegaze hand during manual articulation of SPEC- interlocutor H:HUM stimulus neutral location Comments Additional comments made by the Participant used an LSM sign annotator

3. Creation of annotations by researcher and research assistants, cross-checking with participants

Annotations were made by myself (Josefina Safar), in collaboration with three research assistants from Chicán (Geli Collí Collí, Merli Collí Hau, Dianela Collí Chim). The annotation process was as follows: I identified the sections where participants responded to a stimulus item, created segmentations of the video files (where all segments of the video, in which the participant produced a sign, were marked), and subsequently created annotations in English for all these segments on the tiers listed in Table 1. I used the “Controlled Vocabulary (CV)” function in ELAN, where annotations can be made by selecting from a drop-down list, but it is also possible to enter annotations manually in case the correct gloss is not included in the CV.

72 HHS stands for ‘human height-specifier’. 73 Coding the height of SPEC-H:HUM was not always straightforward and, in particular, it was not always possible to clearly distinguish instances of SPEC-H:HUM articulated at mid-height or at neutral height. The height was coded as neutral if it did not display any correspondence to the referent’s measurements.

123 In cases where I was unsure about the correct translation of a sign, I consulted the research assistants, who are native signers of YMSL and/or the respective participants. This helped to resolve ambiguities.

4. Criteria for inclusion or exclusion of a height-specifier for analysis

As this study was not primarily concerned with the frequency of use or statistical distribution of height-specifiers across YMSL signers but rather sought to describe the range of linguistic functions of SPEC-H:HUM in YMSLs, all instances of SPEC-H:HUM were included in my analysis. If a participant produced more than one instance of SPEC-H:HUM in response to one stimulus, all occurrences of SPEC-H:HUM were taken into account.

5. Identification of height-specifiers in narratives and conversations

In addition to the data from the elicitation task, I analysed the use of height-specifiers in YMSLs in narratives and conversations. I used 20 video files; a part of them had already been transcribed (or partly transcribed) by the research assistants, another part was transcribed and coded by myself specifically for this study. The videos previously transcribed by the research assistants had been annotated using a template with the following tiers: man-der (mano derecha, ‘right hand’), man-izq (mano izquierda, ‘left hand’), cara (‘face’) and traducción español (Spanish translation).74 I identified instances of height-specifiers in the annotations and coded the relevant sections for the specific features relevant to this study (see Table 1). The videos I annotated specifically for the purposes of this study were not transcribed entirely. Instead, only the utterances that included a height-specifier were annotated sign by sign, using the tiers listed in Table 1.

6. Comparison with Yucatec Maya height-specifier gestures

a. Identification of Yucatec Maya height-specifier gestures in elicited data

In Yucatec Maya gesturers’ responses to the picture elicitation task I identified instances of height-specifier gestures and checked whether their hand configuration corresponded to the height-specifiers used by YMSL signers.

b. Identification of Yucatec Maya size-and-shape gestures in natural discourse

In order to find out whether the height-specifier gestures produced by hearing speakers of Yucatec Maya under the conditions of an elicitation task also occur in spontaneous speech, i.e. along with spoken Yucatec Maya, I looked at videos of 3 conversations (2 recorded in Kopchen, 1 recorded in Chemax) and 4 interviews (recorded in Kopchen). The conversations had already been transcribed and translated by Olivier Le Guen (who is a fluent speaker of Yucatec Maya) and coded for different types of manual and non-manual gestures (time gestures, spatial gestures, etc.). I identified instances of height-specifier gestures in the annotated video files, specified which entity they referred to and additionally coded them for

74 Since all participants of the study are right-handed, man-der and man-izq are equivalent to the tiers Gloss_DH and Gloss_NonDH used by myself. The tier cara was used to annotate non-manual markers.

124 the features Height and Eye gaze (see Table 1). Furthermore, I determined the stretches of Yucatec Maya speech that overlapped with the production of height-specifier gestures. The interviews recorded in Kopchen were transcribed and translated into Spanish by Rodrigo Petatillo Chan (who is a native speaker Yucatec Maya). I identified instances of height-specifier gestures in the video files, specified which entity they referred to and coded them for the features Height and Eye gaze. Furthermore, I determined the stretches of Yucatec Maya speech that overlapped with the production of height-specifier gestures.

References

Tkachman, Oksana & Irit Meir. 2018. Novel compounding and the emergence of structure in two young sign languages. Glossa: A Journal of General Linguistics 3(1). 136. http://doi.org/10.5334/gjgl.632.

125 Appendix B: List of handshape names

Bent-B

F m

Index

Y f

126 Sammanfattning på svenska

Josefina Safar A comparative study of Yucatec Maya Sign Languages

(Översättning från engelska av Calle Börstell och Pernilla Hallonsten Halling)

Inledning

Ämnet för denna avhandling är inhemska teckenspråk i yukatekmayasamhällen i Yucatán, Mexiko. Inom ramen för mitt avhandlingsprojekt har jag genomfört fältstudier i fyra byar med en stor andel döva: Chicán, Nohkop, Trascorral och Cepeda Peraza. Döva födda i dessa samhällen har aldrig haft tillgång till dövskolor eller liknande, och har heller aldrig varit i kontakt med det mexikanska dövsamhället. På grund av detta har de inte haft möjlighet att tillägna sig det nationella mexikanska teckenspråket (Lengua de Señas Mexicana; LSM) och har således utvecklat sina egna lokala teckenspråk för att kommunicera med varandra och sina hörande släktingar. I dessa mindre samhällen är det många av invånarna som har döva släktingar, grannar eller vänner, och yukatekiska teckenspråk (Yucatec Maya Sign Languages; YMSLs) delas därmed av döva och hörande. De yukatekiska teckenspråken är unga teckenspråk med tre generationer döva i Chicán, men endast en generation i Nohkop, Trascorral och Cepeda Peraza.

De yukatekiska teckenspråkssamhällena

De yukatekiska teckenspråkssamhällena påminner om varandra i många avseenden som är karakteristiska för byar i det lantliga Yucatán. Folket odlar traditionellt majs och andra grödor, har bin, jagar och tillverkar hängmattor samt andra hantverk. På senare tid har nya inkomstkällor tillkommit och många människor i dessa samhällen söker arbetstillfällen på närliggande orter. I yukatekmayasamhällen bor familjer tillsammans i flera hus på en gemensam släktgård kallad solar. Det talade språk som används av hörande är yukatekiska, och numera kan de flesta hörande i den yngre generationen även spanska. Trots likheterna skiljer sig de fyra bysamhällena åt i fråga om såväl befolkningsmängd som antalet och åldersfördelningen bland döva. I byn Chicán bor 16 döva som är mellan 16 och 67 år gamla. Den äldsta döva i Chicán var i 80-årsåldern när han gick bort i januari 2020. I Nohkop föddes fyra av fem syskon döva i en familj. De är idag mellan 17 och 24 år gamla. I

127 Trascorral bor en familj med 13 syskon, varav sex föddes döva och idag är mellan 9 och 27 år gamla. I Cepeda Peraza bor tio döva, från olika familjer, som är mellan 28 och 47 år gamla. Yukatekmaya har en i allmänhet positiv inställning till dövhet och teckenspråk. Döva ses som intelligenta och kompetenta samhällsmedlemmar och teckenspråk ses som ett fullvärdigt och funktionellt kommunikationsmedel. Döva fyller liknande sociala roller som hörande: de gifter sig med döva eller hörande, skaffar barn, arbetar och tar sig an olika sysslor i samhället. Till skillnad från andra teckenspråksgemenskaper, till exempel i Sverige, finns det inget särskilt dövsamhälle, utan döva och hörande umgås och har skapat ett teckenspråk tillsammans. En stor skillnad mellan döva och hörande i yukateksamhället är tillgängligheten till utbildning: utbildning på teckenspråk finns inte i dessa samhällen och undervisning sker enbart på spanska. Till följd av detta får döva ingen eller högst grundläggande utbildning. Teckenspråksanvändarna i de fyra olika yukatekiska teckenspråkssamhällena har aldrig varit i kontakt, och deras respektive teckenspråk har utvecklats oberoende av varandra. Men trots att språken saknar historiska kopplingar uppvisar de påtagliga likheter i såväl teckenförråd som grammatik. Detta kan förklaras med en gemensam kulturell bakgrund, och genom att språken har kontakt med de gester som hörande använder när de talar yukatekiska. Döva tar till sig dessa gester och införlivar dem i teckenspråket. Exempelvis tecknas ’mycket’ på samma vis som en gest i allmänt bruk bland hörande, och tecknet ser identiskt ut i de fyra samhällena. Andra delar av teckenförrådet, till exempel färger, saknar motsvarande gester, och färgtecken såsom ’vit’ ser helt olika ut i respektive by. Hittills har språkkontakt med mexikanskt teckenspråk varit minimal: några yngre yukatekiska teckenspråksanvändare känner till det mexikanska handalfabetet och några enstaka tecken från mexikanskt teckenspråk. Med ökad rörlighet och tillgång till internet och sociala medier ökar dock även kontakten med det mexikanska dövsamhället och det är oklart hur detta kommer att påverka den yngre generationens tecknande i framtiden.

Varför är det viktigt att undersöka yukatekiska teckenspråk?

● Yukatekiska teckenspråk är hotade minoritetsspråk och fram tills idag har vi vetat väldigt lite om deras teckenförråd och grammatik.

● Yukatekiska teckenspråk ger oss en inblick i hur nya teckenspråk uppstår och förändras över generationer.

● Yukatekiska teckenspråk visar på hur hörandes gester kan införlivas i teckenspråk.

128 ● Att jämföra olika teckenspråk med ursprung i samma geografiska och kulturella kontext underlättar forskning som försöker identifiera hur olika sociolingvistiska faktorer (t.ex. antal generationer av döva, dövas ålder, fördelning av döva över olika familjer, osv.) påverkar språkliga strukturer.

De fyra delstudierna

Avhandlingen består av fyra delstudier, i vilka jag undersökt olika aspekter av yukatekiska teckenspråk. I delstudie I undersökte jag interaktion mellan döva och hörande i yukatekiska teckenspråkssamhällen i olika sammansättningar: interaktion mellan döva och hörande från samma by, döva i interaktion med hörande från andra byar i Yucatán, och interaktion mellan döva från olika byar. Till denna analys använde jag teorin om transspråkande (eng. translanguaging), enligt vilken människor inte bara använder olika språk (t.ex. svenskt teckenspråk, finskt teckenspråk, svenska eller engelska) när de kommunicerar med varandra, utan även tillämpar många andra strategier för att uttrycka sig. Här visar jag att döva och hörande med olika färdigheter i yukatekiska teckenspråk i interaktion använder sig av talad yukatekiska och/eller spanska, gester och ritande på papper, i luften eller på marken samt pekar på och hanterar olika föremål i omgivningen. Min analys visar att vissa faktorer underlättar för döva och hörande att kommunicera med varandra, även i situationer där de inte behärskar varandras språk: 1. Förekomst av konventionella gester, 2. Positiv inställning till dövhet och teckenspråk, 3. Delad kulturell bakgrund. Dessa faktorer leder även till likheter mellan teckenspråk i samhällen utan historisk kontakt. I delstudie II undersökte jag siffertecken (tecken för räkneord) i de yukatekiska teckenspråken från Chicán, Nohkop och Cepeda Peraza. Jag upptäckte att siffertecknen har vissa motsvarigheter hos de gester som hörande använder för att räkna, men att de olika yukatekiska teckenspråkssamhällena utvecklat sina egna räkneordssystem. Vissa strategier för att uttrycka siffror är unika för ett av de yukatekiska teckenspråkssamhällena, och har inte beskrivits tidigare i något annat teckenspråk i världen: i Chicán är det möjligt att summera olika tal för att konstruera ett högre tal, så som FEMTIO+TJUGO+TJUGO+NIO för ‘99’. Andra räknestrategier återfinns i alla tre samhällen och har även dokumenterats tidigare i andra teckenspråk. Resultaten från delstudie II visar också att vissa tal kan tecknas på många olika sätt, och att ålder, grad av läskunnighet och familjetillhörighet hos teckenspråksanvändare

129 påverkar deras val av siffertecken. Dessutom kan kombinationer av siffertecken och andra tecken påverka siffertecknens handform och vridning. I delstudie III undersöks huruvida användare av yukatekiska teckenspråk uttrycker skillnaden mellan substantiv och verb. Delstudien är uppdelad i två: i del 1 undersöker jag hur form- och storleksspecificerare (eng. size -and-shape specifiers; SASSes) används för att markera substantiv. Det är en vanligt förekommande strategi i yukatekiska teckenspråk att verb står för sig själva, medan substantiv utgörs av verbtecknet kombinerat med en form- och storleksspecificerare. Till exempel kan tecknet SKÄRA översättas som ’att skära’, medan tecknet

SKÄRA+SASS_AVLÅNGT-FÖREMÅL betyder ’kniv’. Denna strategi är vanligare i yukatekiska teckenspråk än i andra teckenspråk beskrivna i tidigare forskning, och yngre användare, hörande användare, samt användare från en specifik familj i Chicán använder strategin mer än andra yukatekiska teckenspråksanvändare. I del 2 undersöks skillnader i handform mellan substantiv och verb. Gällande andra teckenspråk har det påvisats att samma rörelse med olika handformer kan uttrycka skillnaden mellan ett föremål (t.ex. ’tandborste’) och dess handling (t.ex. ’att borsta tänderna’). Resultaten från min studie visar att yukatekiska teckenspråksanvändare inte använder sig av denna strategi för att särskilja substantiv och verb, men att andra faktorer ligger bakom valet av specifik handform i tecken för föremål respektive deras associerade handling. Sammantaget visar resultaten från delstudie III att de båda strategierna för att skilja substantiv från verb beskrivna ovan har motsvarigheter i de gester som hörande yukateker använder, men att de införlivats i yukatekiska teckenspråk på olika sätt. I delstudie IV fokuserar jag på en särskild gest som används av hörande yukateker, närmare bestämt en flat hand med handflatan nedåt. Normalt sett används denna gest för att specificera en persons kroppslängd, vanligtvis ett barn, även om den också kan användas för att beskriva icke-mänskliga ting, såsom växter. Jag analyserar hur denna höjdspecificerande gest blivit en del av teckenförrådet och grammatiken i de fyra yukatekiska teckenspråkssamhällena. Jämfört med gesten som används jämte talad yukatekiska används höjdspecificeraren på yukatekiska teckenspråk endast för mänskliga referenter och har många fler funktioner än gesten. Tecknet kan fungera som ett substantiv för ’barn, dotter/son, yngre syskon, far eller (mor)mor’, det kan användas som verb för ’att växa’, det kan användas för att lokalisera en person i rummet, det kan användas som del av ett persontecken, och det kan helt enkelt förtydliga att den som omtalas är en människa. Jag upptäckte dessutom att höjdspecificerarens form när den används av tecknare skiljer sig något från när den används som gest av hörande. En detaljerad jämförelse av höjdspecificerande gester och tecken

130 illustrerar den komplexa process under vilken en gest kan införlivas i ett teckenspråk, och visar på viss variation mellan de fyra samhällena.

Slutsatser

● De yukatekiska teckenspråken är likartade men obesläktade teckenspråk. De överlappar varandra i fråga om teckenförråd och grammatik som en följd av sin gemensamma grund i gester och en delad kultur.

● Den sociala och kulturella kontext i vilken ett språk utvecklas ger form till dess språkliga strukturer. Av den anledningen påträffar vi en hel del variation mellan och inom yukatekiska teckenspråkssamhällen.

● Många språkliga strukturer återfinns redan hos första generationen yukatekiska teckenspråksanvändare, men språken förändras över tid. Nya lexikala och grammatiska konventioner uppstår när teckenspråket överförs till nästa generation.

● Gester som används jämte talad yukatekiska fungerar som en grund för yukatekiska teckenspråk. Döva teckenspråksanvändare tar till sig dessa gester och förändrar och anpassar dem till sina teckenspråk.

131 Bibliography

Adone, Dany, Anastasia Bauer, Keren Cumberbatch & Elaine L. Maypilama. 2012. Colour signs in two indigenous sign languages. In Ulrike Zeshan & Connie de Vos (eds.), Sign languages in village communities: Anthropological and linguistic insights, 53–86. Berlin: De Gruyter Mouton.

Bauer, Anastasia. 2014. The use of signing space in a shared sign language of Australia. Berlin: De Gruyter Mouton.

Bayley, Robert, Ceil Lucas & Mary Rose. 2002. Variation in American Sign Language: The case of DEAF. Journal of Sociolinguistics 4(1). 81–107.

Berez-Kroeker, Andrea L., Lauren Gawne, Susan Smythe Kung, Barbara F. Kelly, Tyler Heston, Gary Holton, Peter Pulsifer, David I. Beaver, Shobhana Chelliah, Stanley Dubinsky, Richard P. Meier, Nick Thieberger, Keren Rice & Anthony C. Woodbury. 2018. Reproducible research in linguistics: A position statement on data citation and attribution in our field. Linguistics 56(1). 1–18.

Berlin, Brent & Paul Kay. 1969. Basic color terms: Their universality and evolution. Berkeley & Los Angeles: University of California Press.

Bohnemeyer, Jürgen, Melissa Bowerman & Penelope Brown. 2001. Cut and break clips. In Stephen C. Levinson & Nick J. Enfield (eds.), Manual for the field season 2001, 90–96. Nijmegen: Max Planck Institute for Psycholinguistics.

Börstell, Carl. 2017. Object marking in the signed modality: Verbal and nominal strategies in Swedish Sign Language and other sign languages. Stockholm: Stockholm University dissertation.

Braithwaite, Benjamin (in press). Emerging sign languages in the Caribbean. In Olivier Le Guen, Josefina Safar & Marie Coppola (eds.), Emerging sign languages of the Americas. Berlin: De Gruyter Mouton.

Brentari, Diane & Marie Coppola. 2012. What sign language creation teaches us about language. WIREs Cognitive Science 4. 201–211.

Brentari, Diane, Marie Coppola, Laura Mazzoni & Susan Goldin-Meadow. 2012. When does a system become phonological? Handshape production in gesturers, signers, and homesigners. Natural Language and Linguistic Theory 30(1). 1–31.

Brentari, Diane & Susan Goldin-Meadow. 2017. Language emergence. Annual Review of Linguistics 3. 363–388.

Brinton, Laurel J. & Elizabeth Closs Traugott. 2005. Lexicalization and language change. Cambridge: Cambridge University Press.

132 Bybee, Joan. 2011. Usage-based theory and grammaticalization. In Bernd Heine & Heiko Narrog (eds.), The Oxford handbook of grammaticalization. Oxford: Oxford University Press. Online: doi: 10.1093/oxfordhb/9780199586783.013.0006.

Bybee, Joan, Revere Perkins & William Pagliuca. 1994. The evolution of grammar. Tense, aspect and modality in the languages of the world. Chicago: University of Chicago Press.

Chelliah, Shobhana. 2001. The role of text collection and elicitation in linguistic fieldwork. In Newman, Paul & Ratliff, Martha (eds.), Linguistic fieldwork, 152–165. Cambridge: Cambridge University Press.

Cooperrider, Kensy, Natasha Abner & Susan Goldin-Meadow. 2018. The palm-up puzzle: Meanings and origins of a widespread form in gesture and sign. Frontiers in Communication 3.

Copland, Fiona, Sara Shaw & Julia Snell (eds.). 2015. Linguistic ethnography: Interdisciplinary explorations. London: Palgrave Macmillan.

Coppola, Marie (in press). Gestures, homesign, sign language: Cultural and social factors driving lexical conventionalization. In Olivier Le Guen, Josefina Safar & Marie Coppola (eds.), Emerging sign languages of the Americas. Berlin: De Gruyter Mouton.

Coppola, Marie & Ann Senghas. 2010. Deixis in an emerging sign language. In Diane Brentari (ed.), Sign languages, 543–569. Cambridge: Cambridge University Press.

Coppola, Marie, Elizabet Spaepen & Susan Goldin-Meadow. 2013. Communicating about number without a language model: Number devices in homesign grammar. Cognitive Psychology 67. 1–25.

Cormier, Kearsy, David Quinto-Pozos, Zed Sevcikova & Adam Schembri. 2012. Lexicalisation and de-lexicalisation processes in sign languages: Comparing depicting constructions and viewpoint gestures. Language & Communication 32(4). 329–348.

Currie, Anne-Marie, Richard P. Meier & Keith Walters. 2002. A crosslinguistic examination of the lexicons of four signed languages. In Richard P. Meier, Kearsy Cormier & David Quinto- Pozos (eds.), Modality and structure in signed and spoken languages, 224–236. Cambridge: Cambridge University Press.

Dachkovsky, Svetlana, Rose Stamp & Wendy Sandler. 2019. Time will tell: Dynamics of grammaticalization & lexicalization of TIME-PASS in Israeli Sign Language. Paper presented at TISLR13, Hamburg.

De Meulder, Maartje. 2018. “So, why do you sign?” Deaf and hearing new signers, their motivation, and revitalisation policies for sign languages. Applied Linguistics Review 10(4). https://doi.org/10.1515/applirev-2017-0100.

De Meulder, Maartje, Annelies Kusters, Erin Moriarty & Joseph J. Murray. 2019. Describe, don’t prescribe. The practice and politics of translanguaging in the context of deaf signers. Journal of Multilingual and Multicultural Development 40(10). 892–906.

133 De Meulder, Maartje, Joseph J. Murray & Rachel McKee (eds.). 2019. The legal recognition of sign languages. Advocacy and outcomes around the world. Bristol: Multilingual Matters. de Vos, Connie. 2011. Kata Kolok color terms and the emergence of lexical signs in rural signing communities. The Senses and Society 6(1). 68–76. de Vos, Connie. 2012. Sign-spatiality in Kata Kolok: How a village sign language in Bali inscribes its signing space. Nijmegen: Radboud University dissertation. de Vos, Connie. 2016. Sampling shared sign languages. Sign Language Studies 16(2). 204–226. de Vos, Connie & Ulrike Zeshan. 2012. Introduction: Demographic, sociocultural and linguistic variation across rural signing communities. In Zeshan, Ulrike & de Vos, Connie (eds.), Sign languages in village communities: Anthropological and linguistic insights, 2–23. Berlin: De Gruyter Mouton. de Vos, Connie & Roland Pfau. 2015. Sign language typology: The contribution of rural sign languages. Annual Review of Linguistics 1. 265–288. de Vos, Connie & Victoria Nyst. 2018. The time depth and typology of rural sign languages. Sign Language Studies 18(4). 477–487.

Dikyuva, Hasan. 2012. Mardin Sign Language: Signing in a “deaf family”. In Ulrike Zeshan & Connie De Vos (eds.), Sign languages in village communities: Anthropological and linguistic insights, 395–400. Berlin: De Gruyter Mouton.

Dikyuva, Hasan, Cesar Ernesto Escobedo Delgado, Sibaji Panda & Ulrike Zeshan. 2012. Working with village sign language communities: Deaf fieldwork researchers in professional dialogue. In Ulrike Zeshan & Connie De Vos (eds.), Sign languages in village communities: Anthropological and linguistic insights, 313–344. Berlin: De Gruyter Mouton.

Emmorey, Karen. 1999. Do signers gesture? In Lynn S. Messing & Ruth Campbell (eds.), Gesture, speech and sign, 133–159. Oxford: Oxford University Press.

Emmorey, Karen, Helsa B. Borinstein & Robin Thompson. 2005. Bimodal bilingualism: Code- blending between spoken English and American Sign Language. In James Cohen, Kara T. McAlister, Kellie Rolstad & Jeff MacSwan (eds.), Proceedings of the 4th International Symposium on Bilingualism, 663–673. Somerville, MA: Cascadilla Press.

Enfield, Nick. 2009. The anatomy of meaning: Speech, gesture, and composite utterances. Cambridge: Cambridge University Press.

Ergin, Rabia. 2017. Central Taurus Sign Language: A unique vantage point into language emergence. Medford: Tufts University dissertation.

Escobedo Delgado, Cesar Ernesto. 2012. Chican Sign Language: A sociolinguistic sketch. In Ulrike Zeshan & Connie De Vos (eds.), Sign languages in village communities: Anthropological and linguistic insights, 377–380. Berlin: De Gruyter Mouton.

134 Fenlon, Jordan, Kensy Cooperrider, Jon Keane, Diane Brentari & Susan Goldin-Meadow. 2019. Comparing sign language and gesture: Insights from pointing. Glossa 4(1), 2.

Fischer, Susan. 1978. Sign language and creoles. In Patricia Siple (ed.), Understanding language through sign language research, 309–333. New York: Academic Press.

Foster, George. 1948. Empire’s children. Mexico: Nuevo Mundo.

Fox Tree, Erich. 2009. Meemul Tziij: An indigenous sign language complex of Mesoamerica. Sign Language Studies, 9(3). 324–366.

Fusellier-Souza, Ivani. 2006. Emergence and development of sign languages: from a semiogenetic point of view. Sign Language Studies 7(1). 30–56.

García, Ofelia & Li Wei. 2014. Translanguaging: Language, bilingualism and education. London: Palgrave MacMillan.

Goldin-Meadow Susan. 2003. The resilience of language: What gesture creation in deaf children can tell us about how all children learn language. New York: Psychology Press.

Goldin-Meadow, Susan. 2005. Watching language grow. Proceedings of the National Academy of Sciences of the United States of America 102(7). 2271–2272.

Goldin-Meadow, Susan & Diane Brentari. 2017. Gesture, sign, and language: the coming of age of sign language and gesture studies. Behavioral and Brain Sciences 40, 1–17.

Green, Elizabeth Mara. 2014. The nature of signs: Nepal’s Deaf society, local sign, and the production of communicative sociality. Berkeley: University of California dissertation.

Groce, Nora Ellen. 1985. Everyone here spoke sign language: Hereditary deafness on Martha's Vineyard. Cambridge, MA: Harvard University Press.

Hanks, William F. 2010. Converting words: Maya in the age of the cross. Berkeley: University of California Press.

Haviland, John B. 2013. The emerging grammar of nouns in a first generation sign language: Specification, iconicity, and syntax. Gesture 13(3). 309–353.

Haviland, John B. 2016. ‘But you said “four sheep” . . . !’: (Sign) language, ideology, and self(esteem) across generations in a Mayan family. Language and Communication 46. 62–94.

Haviland, John B. (in press). Zinacantec family homesign (or “Z”). In Olivier Le Guen, Josefina Safar & Marie Coppola (eds.), Emerging sign languages of the Americas. Berlin: De Gruyter Mouton.

Himmelmann, Nikolaus. 2004. Lexicalization and grammaticalization: Opposite or orthogonal? In Bisang, Walter, Nikolaus Himmelmann & Björn Wiemer (eds.), What makes grammaticalization? : a look from its fringes and its components, 41–42. Berlin: De Gruyter Mouton.

135 Hopper, Paul J. 1991. On some principles of grammaticization. In Elizabeth Closs Traugott & Bernd Heine (eds.), Approaches to grammaticalization, Volume I: Focus on theoretical and methodological issues, 149–187. Amsterdam: John Benjamins.

Hopper, Paul J. & Elizabeth Closs Traugott. 2003. Grammaticalization. Cambridge: Cambridge University Press.

Horton, Laura (in press-a). Representational strategies in shared homesign systems from Nebaj, Guatemala. In Olivier Le Guen, Josefina Safar & Marie Coppola (eds.), Emerging sign languages of the Americas. Berlin: De Gruyter Mouton.

Horton, Laura (in press-b). A sociolinguistic sketch of deaf individuals and families from Nebaj, Guatemala. In Olivier Le Guen, Josefina Safar & Marie Coppola (eds.), Emerging sign languages of the Americas. Berlin: De Gruyter Mouton.

Hou, Lynn. 2016. “Making hands”: Family sign languages in the San Juan Quiahije community. Austin: University of Texas dissertation.

Hou, Lynn. 2018. Iconic patterns in San Juan Quiahije Chatino Sign Language. Sign Language Studies 18(4). 570–611.

Hunger, Barbara. 2006. Noun/verb pairs in Austrian Sign Language (ÖGS). Sign Language & Linguistics 2(2). 71–94.

INEGI. 2010. Censo de población y vivienda 2010. Mexico: Instituto Nacional de Estadística y Geografia.

Iseli, Jacqueline. 2018. Deaf Ni-Vanuatu and their signs: A sociolinguistic study. Wellington: Victoria University MA thesis.

Israel, Assaf. 2009. Sublexical variation in three sign languages. Haifa: University of Haifa MA thesis.

Janzen, Terry. 2012. Lexicalization and grammaticalization. In Roland Pfau, Markus Steinbach & Bencie Woll (eds.), Sign language – An international handbook, 816–841. Berlin: De Gruyter Mouton.

Janzen, Terry & Barbara Shaffer. 2002. Gesture as the substrate in the process of ASL grammaticalization. In Richard P. Meier, Kearsy Cormier, & David Quinto-Pozos (eds.), Modality and structure in signed and spoken languages, 199-223. Cambridge: Cambridge University Press.

Jewish Deaf Association. 2003. Sign language in Judaism. London: Jewish Deaf Association.

Johnson, Robert E. 1991. Sign Language, culture, and community in a traditional Yucatec Maya Village. Sign Language Studies 73. 461–474.

Johnston, Trevor. 2001. Nouns and verbs in Australian Sign Language: An open or shut case? Journal of Deaf Studies and Deaf Education 6(4). 235–257.

136 Johnston, Trevor. 2018. A corpus-based study of the role of headshaking in negation in Auslan (Australian Sign Language): Implications for signed language typology. Linguistic Typology 22(2). 185–231.

Johnston, Trevor & Adam Schembri. 2007. Australian Sign Language (Auslan): An introduction to sign language linguistics. Cambridge: Cambridge University Press.

Johnston, Trevor, Myriam Vermeerbergen, Adam Schembri & Lorraine Leeson. 2007. ‘Real data are messy’: Considering cross-linguistic analysis of constituent ordering in Auslan, VGT and ISL. In Pamela Perniss, Roland Pfau & Markus Steinbach (eds.), Visible variation: Comparative studies on sign language structure, 163–206. Berlin: De Gruyter Mouton.

Kakumasu, Jim. 1968. Urubu Sign Language. International Journal of American Linguistics 34. 275–281.

Kegl, Judy, Ann Senghas, Marie Coppola. 1999. Creation through contact: Sign language emergence and sign language change in Nicaragua. In Michel DeGraff, (ed.), Language creation and language change: Creolization, diachrony, and development, 179–237. Cambridge: MIT Press.

Kendon, Adam. 1980. A description of a deaf-mute sign language from the Enga Province of Papua New Guinea with some comparative discussion (Part I-III). Semiotica 32. 1–34 (Part I), 81– 117 (Part II), 245–313 (Part III).

Kendon, Adam. 1988. How gestures can become like words. In Fernando Poyatos (ed.), Cross- cultural perspectives in nonverbal communication, 131–141. Toronto: Hogrefe & Huber Publishers.

Kendon, Adam. 1992. Some recent work from Italy on quotable gestures (emblems). Journal of Linguistic Anthropology 2(1). 92–108.

Kendon, Adam. 2004. Gesture: Visible action as utterance. Cambridge: Cambridge University Press.

Kendon, Adam. 2008. Some reflections on the relationship between “gesture” and “sign.” Gesture 8(3). 348–366.

Kimmelman, Vadim. 2009. Parts of speech in Russian Sign Language: The role of iconicity and economy. Sign Language & Linguistics 12(2). 161–186.

Kisch, Shifra. 2008. “Deaf discourse”: the social construction of deafness in a Bedouin community. Medical anthropology 27(3). 283–313.

Kisch, Shifra. 2012. Demarcating generations of signers in the dynamic sociolinguistic landscape of a shared sign language: The case of the Al-Sayyid Bedouin. In Ulrike Zeshan & Connie De Vos (eds.), Sign languages in village communities: Anthropological and linguistic insights, 87–126. Berlin: De Gruyter Mouton.

Kocab, Annemarie, Jennie Pyers & Ann Senghas. 2015. Referential shift in Nicaraguan Sign Language: A transition from lexical to spatial devices. Frontiers in Psychology 5. 1540.

137 Kusters, Annelies. 2010. Deaf utopias? Reviewing the sociocultural literature on the world’s “Martha’s Vineyard situations”. Journal of Deaf Studies and Deaf Education 15(1). 3–16.

Kusters, Annelies. 2015. Deaf space in Adamorobe: An ethnographic study in a village in Ghana. Washington, DC: Gallaudet University Press.

Kusters, Annelies. 2017. Gesture-based customer interactions: Mumbaikars’ multimodal and metrolingual practices. International Journal of Multilingualism 14(3). 283–302.

Kusters, Annelies, Massimiliano Spotti, Ruth Swanwick & Elina Tapio. 2017a. Beyond languages, beyond modalities: transforming the study of semiotic repertoires. International Journal of Multilingualism 14(3). 219–232.

Kusters, Annelies, Maartje De Meulder & Dai O’Brien. 2017b. Innovations in Deaf Studies: Critically mapping the field. In Annelies Kusters, Maartje De Meulder & Dai O’Brien (eds.), Innovations in Deaf Studies. The role of deaf scholars, 1–53. New York: Oxford University Press.

Kusters, Annelies & Sujit Sahasrabudhe. 2018. Language ideologies on the difference between gesture and sign. Language & Communication 60. 44-63.

Labov, William. 1972. Sociolinguistic patterns. Philadelphia: University of Pennsylvania Press.

Lanesman, Sara & Irir Meir. 2012. The survival of Algerian Jewish Sign Language alongside Israeli Sign Language in Israel. In Ulrike Zeshan & Connie de Vos (eds.), Sign languages in village communities: Anthropological and linguistic insights, 153–179. Berlin: De Gruyter Mouton.

Leeson, Lorraine & Carmel Grehan. 2004. To the lexicon and beyond: The effect of gender on variation in Irish Sign Language. In Mieke Van Herreweghe, Mieke & Myriam Vermeerbergen (eds.), To the lexicon and beyond. Sociolinguistics in European Deaf communities, 39–73. Washington: Gallaudet University Press.

Le Guen, Olivier. 2012. An exploration in the domain of time: from Yucatec Maya time gestures to Yucatec Maya Sign Language time signs. In Ulrike Zeshan & Connie de Vos (eds.), Sign languages in village communities: Anthropological and linguistic insights, 209–250. Berlin: De Gruyter Mouton.

Le Guen, Olivier. 2018. La importancia de la gestualidad en el estudio del maya yucateco actual. Cuadernos de Lingüística de El Colegio de México 5(1). 331–373.

Le Guen, Olivier, Rebeca Petatillo Balam & Rita Kinil Canché (in press-a). Yucatec Maya multimodal interaction as basis for Yucatec Maya Sign Language. In Olivier Le Guen, Josefina Safar & Marie Coppola (eds.), Emerging sign languages of the Americas. Berlin: De Gruyter Mouton.

Le Guen, Olivier, Marie Coppola & Josefina Safar (in press-b). Introduction: How emerging sign languages in the Americas contributes to the study of linguistics and (emerging) sign languages. In Olivier Le Guen, Josefina Safar & Marie Coppola (eds.), Emerging sign languages of the Americas. Berlin: De Gruyter Mouton.

138 Le Guen, Olivier & Safar, Josefina (in prep.) Intergenerational evolution of the use of the signing space in Yucatec Maya Sign Languages.

LeMaster, Barbara & John Dwyer. 1991. Knowing and using female and male signs in Dublin. Sign Language Studies 73. 361–396.

Lightfoot, Douglas. 2011. Grammaticalization and lexicalization. In Narrog, Heiko & Bernd Heine (eds.), Grammaticalization from a typological perspective, Oxford: Oxford University Press. Online: DOI: 10.1093/oxfordhb/9780199586783.013.0035.

Loon, Esther van, Roland Pfau, & Markus Steinbach. 2014. The grammaticalization of gestures in sign languages. In Cornelia Müller, Alan Cienki, Ellen Fricke, Silva Ladewig, David McNeill, & Jana Bressem (Eds.), Body – language – communication: An international handbook on multimodality in human interaction, Vol. 2, 2133–2149. Berlin: De Gruyter Mouton.

Lucas, Ceil, Robert Bayley & Clayton Valli. 2001. Sociolinguistic variation in American Sign Language. Washington: Gallaudet University Press.

Lucas, Ceil, Robert Bayley, Ruth Reed & Alyssa Wulf. 2001. Lexical variation in African American and white signing. American Speech 76(4). 339–360.

Lucas, Ceil, Robert Bayley & Clayton Valli. 2003. What’s your sign for PIZZA? An introduction to variation in American Sign Language. Washington D.C.: Gallaudet University Press.

Lutzenberger, Hannah. 2018. Manual and nonmanual features of name signs in Kata Kolok and Sign Language of the Netherlands. Sign Language Studies 18(4). 546–569.

MacBurney, Susan. 2012. History of sign languages and sign language linguistics. In Pfau, Roland, Steinbach, Markus & Woll, Bencie (eds.), Sign language: An international handbook, 909– 948. Berlin: De Gruyter Mouton.

MacDougall, Jennifer Paige. 2012. Being deaf in a Yucatec Maya Community: Communication and identity negotiation. Montreal: McGill University dissertation.

Makoni, Sinfree & Alistair Pennycook (eds.). 2006. Disinventing and reconstituting languages. Clevedon, UK: Multilingual Matters.

Marsaja, Gede. 2008. Desa Kolok – A Deaf village and its sign language in Bali, Indonesia. Nijmegen: Ishara Press.

Martinod, Emmanuella, Brigitte Garcia & Ivani Fusellier (in press). A typological perspective on the meaningful handshapes in the emerging sign languages on Marajó Island (Brazil). In Olivier Le Guen, Josefina Safar & Marie Coppola (eds.), Emerging sign languages of the Americas. Berlin: De Gruyter Mouton.

McKee, Rachel, David McKee, Kirsten Smiler & Karen Pointon. 2007. “Maori signs”: The construction of indigenous Deaf identity in New Zealand Sign Language. In David Quinto- Pozos (ed.), Sign languages in contact, 31–81. Washington: Gallaudet University Press.

139 McKee, David, Rachel McKee & George Major. 2011. Numeral variation in New Zealand Sign Language. Sign Language Studies 12(1). 72–97.

McKee, Rachel & David McKee. 2020. Globalization, hybridity, and vitality in the linguistic ideologies of New Zealand Sign Language users. Language & Communication 74. 164–181.

McNeill, David. 1992. Hand and mind: What gestures reveal about thought. Chicago: University of Chicago Press.

McNeill, David. 2000. Introduction. In David McNeill (ed.), Language and gesture, 1–10. Cambridge: Cambridge University Press.

McWhorter, John H. 2018. Is grammaticalization in creoles different? In Narrog, Heiko & Bernd Heine (eds.), Grammaticalization from a typological Perspective, Oxford: Oxford University Press. Online: 10.1093/oso/9780198795841.001.0001.

Meier, Richard, Kearsy Cormier & David Quinto-Pozos. 2002. Modality and structure in signed and spoken languages. Cambridge: Cambridge University Press.

Meir, Irit, Wendy Sandler, Carol Padden Mark & Aronoff. 2010a. Emerging sign languages. In Marc Marschark & Patricia Elizabeth Spencer (eds.), Oxford Handbook of Deaf Studies, Language and Education, Vol. 2, 267–280. Oxford: Oxford University Press.

Meir, Irit, Mark Aronoff, Wendy Sandler & Carol Padden. 2010b. Sign languages and compounding. In Sergio Scalise & Irene Vogel (eds.), Cross-disciplinary issues in compounding, 301–322. Amsterdam/Philadelphia: John Benjamins.

Meir, Irit, Assaf Israel, Wendy Sandler, Carol Padden & Mark Aronoff. 2012. The influence of community on language structure. Evidence from two young sign languages. Linguistic Variation 12(2). 247–291.

Meir, Irit & Wendy Sandler. 2020. Variation and conventionalization in young sign languages. In Edit Doron, Malka Rappaport Hovav, Yael Reshef & Moshe Taube (eds.), Linguistic Contact, Continuity and Change in the Genesis of Modern Hebrew, 337–363. Amsterdam: John Benjamins.

Meo-Zilio, Giovanni & Silvia Mejía. 1980. Diccionario de gestos. España e Hispanoamérica, Volume 1. Bogotá: Instituto Caro y Cuervo.

Meyerhoff, Miriam. 2006. Introducing sociolinguistics. New York: Routledge.

Mesch, Johanna & Lars Wallin. 2015. Gloss annotations in the Swedish Sign Language Corpus. International Journal of Corpus Linguistics 20(1). 102–120.

Mesh, Kate. 2017. Points of comparison: What indicating gestures tell us about the origins of signs in San Juan Quiahije Chatino Sign Language. Austin: University of Texas dissertation.

Mesh, Kate & Lynn Hou. 2018. Negation in San Juan Quiahije Chatino Sign Language: The integration and adaptation of conventional gestures. Gesture 17(3). 330–374.

140 Mudd, Katie, Hannah Lutzenberger, Connie de Vos, Paula Fikkert, Onno Crasborn & Bart de Boer. 2020. The effect on sociolinguistic factors on variation in the Kata Kolok lexicon. Asia- Pacific Language Variation 6(1). 53–90

Müller Cornelia. 2018. Gesture and sign: Cataclysmic break or dynamic relations? Frontiers in Psychology 9:1651. Online: doi: 10.3389/fpsyg.2018.0165.

Narrog, Heiko & Bernd Heine (eds.). 2018. Grammaticalization from a typological Perspective. Oxford: Oxford University Press.

Neveu, Grace Kathleen. 2019. Lexical conventionalization and the emergence of grammatical devices in a second generation homesign system in Peru. Austin: University of Texas dissertation.

Nonaka, Angela. 2004. The forgotten endangered languages: Lessons on the importance of remembering from Thailand’s Ban Khor Sign Language. Language in Society 33. 737–767.

Nonaka, Angela. 2007. Emergence of an indigenous sign language and a speech/sign community in Ban Khor, Thailand. Los Angeles: University of California dissertation.

Nonaka, Angela. 2012. Language ecological change in Ban Khor, Thailand: An ethnographic case study of village sign language endangerment. In Ulrike Zeshan & Connie de Vos (eds.), Sign languages in village communities: Anthropological and linguistic insights, 277–312. Berlin: De Gruyter Mouton.

Nyst, Victoria. 2007. A descriptive analysis of Adamorobe Sign Language (Ghana). Utrecht: LOT.

Nyst, Victoria. 2012. Shared sign languages. In Roland Pfau, Markus Steinbach & Bencie Woll (eds.), Sign Language: An international handbook, 552–574. Berlin: De Gruyter Mouton.

Nyst, Victoria. 2013. Towards a new model of sign language affiliation. Paper presented at TISLR11, London.

Nyst, Victoria. 2015. Sign Language Fieldwork. In Orfanidou, Eleni, Bencie Woll & Gary Morgan (eds.), Research methods in sign language studies: A practical guide, 108–118. Oxford: Wiley-Blackwell.

Nyst, Victoria. 2016a. The depiction of size and shape in gestures accompanying object descriptions in Anyi (Côte d’Ivoire) and in Dutch (The Netherlands). Gesture 15(2). 156– 191.

Nyst, Victoria. 2016b. Size and shape depictions in the manual modality: a taxonomy of iconic devices in Adamorobe Sign Language, Semiotica 210. 75–104.

Nyst, Victoria. 2018. Cross-linguistic variation in space-based distance for size depiction in the lexicons of six sign languages. Sign Language & Linguistics 21(2). 350–379.

Nyst, Victoria, Sylla, Kara & Moustapha Magassouba. 2012. Deaf signers in Douentza, a rural area in Mali. In Zeshan, Ulrike & de Vos, Connie (eds.), Sign languages in village communities: Anthropological and linguistic insights, 251–276. Berlin: De Gruyter Mouton.

141 Occhino, Corrine, Benjamin Anible, Erin Wilkinson & Jill Morford. 2017. Iconicity is in the eye of the beholder: How language experience affects perceived iconicity. Gesture 16(1). 100–126.

Ortega, Gerardo & Aslı Özyürek. 2016. Generalisable patterns of gesture distinguish semantic categories in communication without language. In Anna Papafragou, Dan Grodner, Dan Mirman & John Trueswell (eds.), Proceedings of the 38th annual meeting of the Cognitive Science Society (CogSci 2016), 1182–1187. Austin: Cognitive Science Society.

Ortega, Gerardo, Annika Schiefner & Aslı Özyürek. 2019. Hearing non-signers use their gestures to predict iconic form-meaning mappings at first exposure to signs. Cognition 191:103996.

Otheguy, Ricardo, Ofelia García & Wallis Reid. 2015. Clarifying translanguaging and deconstructing named languages: A perspective from linguistics. Applied Linguistics Review 6(3). 281–307.

Pacheco Cahuich, Maribel (forthcoming). Trayectorias de vida de los sordos de Chicán, usuarios de la lengua de señas maya yucateca. La sordera indígena frente a los retos actuales en un contexto intercultural. Mexico City: UNAM BA thesis.

Padden, Carol & Tom Humphries. 1988. Deaf in America: Voices from a culture. Cambridge: Harvard University Press.

Padden, Carol, Irit Meir, So-One Hwang, Ryan Lepic, Sharon Seegers & Tory Sampson. 2013. Patterned iconicity in sign language lexicons. Gesture 13(3). 287–305.

Padden, Carol, So-One Hwang, Ryan Lepic & Sharon Seegers. 2015. Tools for language: Patterned iconicity in sign language nouns and verbs. Topics in Cognitive Science 7(1). 81–94.

Perniss, Pamela, Robin Thompson & Gabriella Vigliocco. 2010. Iconicity as a general property of language: Evidence from spoken and signed languages. Frontiers in Psychology 1. Online: doi: 10.3389/fpsyg.2010.00227.

Pfau, Roland. 2015. The grammaticalization of headshakes: From head movement to negative head. In Andrew D.M. Smith, Graeme Trousdale & Richard Waltereit (eds.), New directions in grammaticalization research. Amsterdam: John Benjamins, 9–50.

Pfau, Roland & Markus Steinbach. 2011. Grammaticalization in sign languages. In Bernd Heine & Heiko Narrog (eds.), The Oxford handbook of grammaticalization, 683–695. Oxford: Oxford University Press.

Pfau, Roland & Markus Steinbach. 2013. PERSON climbing up a tree (and other adventures in sign language grammaticalization). Sign Language & Linguistics, 16(2). 189–220.

Pyers, Jenny, Anna Shusterman, Ann Senghas, Elizabeth S. Spelke & Karen Emmorey. 2010. Evidence from an emerging sign language reveals that language supports spatial cognition. Proceedings of the National Academy of Sciences of the United States of America 107(27). 12116–12120.

Reed, Lauren. 2019. Sign languages of Western highlands, Papua New Guinea, and their challenges for sign language typology. Canberra: Australian National University MA thesis.

142 Reed, Lauren W. & Alan Rumsey. 2020. Sign Languages in Papua New Guinea and the Solomon Islands. In Adam Kendon, Sign Language in Papua New Guinea. A primary sign language from the Upper Lagaip Valley, Enga Province, 141–183. Amsterdam: John Benjamins.

Richie, Russell, Matthew L. Hall, Pyeong Whan Cho & Marie Coppola (in press). Converging evidence: Enhanced conventionalization of gestural referring expressions in richly-connected networks. Language Dynamics and Change.

Safar, Josefina. 2015. Eine Analyse von Diskursen über Chicán Sign Language (Mexiko) [An analysis of discourse about Chican Sign Language (Mexico)]. Das Zeichen 101(1). 434–446.

Safar, Josefina. 2018. Review of Sarvasy, Hannah & Diana Forker (eds.). “Word hunters: Field linguists on fieldwork.” Linguistic Typology 22(3). 505–516.

Safar, Josefina. 2019. Translanguaging in Yucatec Maya Signing Communities. Applied Linguistics Review 10(1). 31–53.

Safar, Josefina. 2020a. Translanguaging in Yucatec Maya signing communities. Data and coding. OSF. osf.io/yn5rg.

Safar, Josefina. 2020b. Numeral variation in Yucatec Maya Sign Languages. Data and coding. OSF. osf.io/tjgqb.

Safar, Josefina. 2020c. Noun-verb distinction in Yucatec Maya Sign Languages. Data and coding. OSF. osf.io/q2jyu.

Safar, Josefina. 2020d. Height-specifiers in Yucatec Maya Sign Languages. Data and coding. OSF. osf.io/njfmw.

Safar, Josefina (forthcoming-a). What’s your sign for TORTILLA? Documenting lexical variation in Yucatec Maya Sign Languages. Language Documentation & Conservation.

Safar, Josefina (forthcoming-b). “When you were that little…” – From Yucatec Maya height- specifier gestures to Yucatec Maya Sign Language person-classifier signs. Gesture 18(1).

Safar, Josefina, Olivier Le Guen, Geli Collí Collí & Merli Collí Hau. 2018. Numeral variation in Yucatec Maya Sign Languages. Sign Language Studies, 18(4), 488–516.

Safar, Josefina & Rodrigo Petatillo Chan (in press). Strategies of noun-verb distinction in Yucatec Maya Sign Languages. In Olivier Le Guen, Josefina Safar & Marie Coppola (eds.), Emerging sign languages of the Americas. Berlin: De Gruyter Mouton.

Safar, Josefina & Olivier Le Guen (in press). Yucatec Maya Sign Language. A sociolinguistic overview. In Olivier Le Guen, Josefina Safar & Marie Coppola (eds.), Emerging sign languages of the Americas. Berlin: De Gruyter Mouton.

143 Sandler, Wendy, Irit Meir, Carol Padden & Mark Aronoff. 2005. The emergence of grammar: Systematic structure in a new language. Proceedings of the National Academy of Sciences of the United States of America 102(7). 2661–2665.

Sandler, Wendy, Mark Aronoff, Irit Meir & Carol Padden. 2011. The gradual emergence of phonological form in a new language. Natural Language and Linguistic Theory 29(2). 503– 543.

Sarvasy, Hannah & Diana Forker (eds.). 2018. Word hunters: Field linguists on fieldwork. Amsterdam: John Benjamins.

Schembri, Adam & Trevor Johnston. 2012. Sociolinguistic aspects of variation and change. In Roland Pfau, Markus Steinbach & Bencie Woll (eds.), Sign language: An international handbook, 788–816. Berlin: De Gruyter Mouton.

Schembri, Adam, Jordan Fenlon, Ramas Rentelis, Sakky Reynolds & Kearsy Cormier. 2013. Building the British Sign Language corpus. Language Documentation & Conservation 7. 136–154.

Senghas, Ann. 2003. Intergenerational influence and ontogenetic development in the emergence of spatial grammar in Nicaraguan Sign Language. Cognitive Development 18(4). 511–531.

Senghas, Ann. 2011. The emergence of two functions for spatial devices in Nicaraguan Sign Language. Human Development 53(5). 287–302.

Senghas, Ann & Marie Coppola. 2001. Children creating language: How Nicaraguan Sign Language acquired a spatial grammar. Psychological Science 12(4). 323–328.

Senghas, Ann, Sotaro Kita & Aslı Özyürek. 2004. Children creating core properties of language: evidence from an emerging sign language in Nicaragua. Science 305(5691). 1779–1782.

Shuman, Malcolm K. 1980. The sound of silence in Nohya: A preliminary account of sign language use by the Deaf in a Maya community in Yucatan, Mexico. Language Sciences 2(1). 144–173.

Shuman, Malcolm K. & Mary Margaret Cherry-Shuman. 1981. A brief annotated sign list of Yucatec Maya Sign Language. Language Sciences 53. 124–185.

Søndergaard, Bent. 1991. Switching between seven codes within one family – a linguistic resource. Journal of Multilingual & Multicultural Development 12:1-2. 85–92.

Spaepen, Elizabet, Marie Coppola, Molly Flaherty, Elizabeth Spelke & Susan Goldin-Meadow. 2013. Generating a lexicon without a language model: Do gestures for number count? Journal of Memory and Language 64(4). 496–505.

Stamp, Rose, Adam Schembri, Jordan Fenlon, Ramas Rentelis, Bencie Woll & Kearsy Cormier. 2014. Lexical variation and change in British sign language. PLoS ONE 9(4). Online: doi:10.1371/journal.pone.0094053.

144 Supalla, Ted & Elissa L. Newport. 1978. How many seats in a chair? The derivation of nouns and verbs in American Sign Language. In Patricia Siple (ed.), Understanding language through sign language research, 91–132. New York: Academic Press.

Swadesh, Morris. 1955. Towards greater accuracy in lexicostatistic dating. International Journal of American Linguistics 21(2). 121–137.

Tano, Angoua. 2016. Etude d'une langue des signes émergente de Côte d'Ivoire: l'exemple de la Langue des Signes de Bouakako (LaSiBo). Leiden: Leiden University dissertation.

Tano, Angoua & Victoria Nyst. 2018. Comparing body-part size and shape constructions in village sign languages with cospeech gesture. Sign Language Studies 18(4). 517–545.

Taub, Sarah F. 2001. Language from the body: Iconicity and metaphor in American Sign Language. Cambridge: Cambridge University Press.

Tkachman, Oksana & Wendy Sandler. 2013. The noun-verb distinction in two young sign languages. Gesture 13(3). 253–286.

Tkachman, Oksana & Irit Meir. 2018. Novel compounding and the emergence of structure in two young sign languages. Glossa: A Journal of General Linguistics 3(1). 136. Online: http://doi.org/10.5334/gjgl.632.

Washabaugh, William. 1979. Hearing and Deaf signers on Providence Island. Sign Language Studies 24. 191–214.

Washabaugh, William. 1986. Five Fingers for Survival. Ann Arbor: Karoma.

Washabaugh, William, James C. Woodward & Susan DeSantis. 1978. Providence Island Sign: A context-dependent language. Anthropological Linguistics 20(3). 95–109.

Webster, Jennifer & Josefina Safar. 2019. Scoring sign language vitality: Adapting a spoken language survey to target the endangerment factors affecting sign languages. Language Documentation & Conservation 13. 346–383.

Webster, Jennifer & Josefina Safar. 2020. Ideologies behind the scoring of factors to rate sign language vitality. Language & Communication 74. 113–129.

Wei, Li. 2011. Moment analysis and translanguaging space: Discursive construction of identities by multilingual Chinese youth in Britain. Journal of Pragmatics 43(5). 1222–1235.

Wilcox, Sherman. 2005. Routes from gesture to language. Revista da ABRALIN – Associação Brasileira de Lingüística 4 (1-2). 11–45.

Wilcox, Sherman, Paolo Rossini & Elena Antinoro Pizzuto. 2010. Grammaticalization in sign languages. In Diane Brentari (ed.), Sign languages, 332-354. Cambridge: Cambridge University Press.

145 Wittenburg, Peter, Hennie Brugman, Albert Russel, Alex Klassmann & Han Sloetjes. 2006. ELAN: a professional framework for multimodality research. In: Proceedings of LREC 2006, Fifth International Conference on Language Resources and Evaluation. 1556–1559.

Woodward, James. 1993. The relationship of sign language varieties in India, Pakistan, and Nepal. Sign Language Studies 78. 15–22.

Woodward, James & Horejes, Thomas. 2016. Deaf/Deaf: Origins and usage. In Genie Gertz & Patrick Boudreault (eds.), The SAGE Deaf Studies Encyclopedia, 285–287. Thousand Oaks, CA: SAGE Publications.

Yano, Uiko & Kazumi Matsuoka. 2018. Numerals and timelines of a shared sign language in Japan: Miyakubo Sign Language on Ehime-Oshima Island. Sign Language Studies 18(4). 640–665.

Zavala, Roberto. 2000. Multiple classifier systems in Akatek (Mayan). In Gunter Senft (ed.), Systems of nominal classification. Cambridge: Cambridge University Press.

Zeshan, Ulrike. 2000. Sign Language in Indo-Pakistan. A description of a signed language. Amsterdam: John Benjamins.

Zeshan, Ulrike. 2003. ‘Classificatory’ constructions in Indo-Pakistani Sign Language: Grammaticalization and lexicalization processes. In Karen Emmorey (ed.), Perspectives on classifier constructions in sign language, 113–142. Mahwah: Lawrence Erlbaum.

Zeshan, Ulrike. 2011. Village sign languages: A commentary. In Gaurav Mathur & Donna J. Napoli (eds.), Deaf around the world: the impact of language, 221–230. Oxford: Oxford University Press.

Zeshan, Ulrike. 2015. “Making meaning”: Communication between sign language users without a shared language. Cognitive Linguistics 26(2). 211–260.

Zeshan, Ulrike, Cesar Ernesto Escobedo Delgado, Hasan Dikyuva & Sibaji Panda. 2013. Cardinal numerals in rural sign languages: Approaching cross-modal typology. Linguistic Typology 17(3). 357–396.

Zeshan, Ulrike & Keiko Sagara (eds.). 2016. Semantic fields in sign languages. Colour, kinship and quantification. Berlin: De Gruyter Mouton.

146