MUTUAL DEPENDENCIES OF NOMINAL AND
CLAUSAL SYNTAX IN CH’OL
A Dissertation
Presented to the Faculty of the Graduate School
of Cornell University
in Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy
by
Carol Roseman Little
August 2020
©c 2020 Carol Roseman Little
ALL RIGHTS RESERVED
MUTUAL DEPENDENCIES OF NOMINAL AND CLAUSAL SYNTAX IN CH’OL
Carol Roseman Little, Ph.D.
Cornell University 2020
This dissertation investigates nominal and clausal structure in Ch’ol, a Mayan lan-
guage of southern Mexico. I take as an empirical basis the interpretation of bare nouns
and extraction asymmetries in Ch’ol to argue for a base-generated account of verb-initial
syntax. I propose that (in)definiteness is reflected structurally, demonstrating that only
bare nominals within the VPmay receive an indefinite interpretation. Any nominals gen-
erated or moved outside the VP must be definite or take wide scope. I capture this em-
pirical generalization by linking semantic interpretation to syntactic structure. Data on
elements that can or cannot be displaced from nominal expressions also bear on under-
lying syntactic structure. I argue that the islandhood of nominals provides evidence for
whether they are contained in a moved constituent. When extraction of an element is pos-
sible, it is not in a movement-derived position—having implications for the underlying
structure of verb-initial word order.
With these conclusions in mind, I support a syntax for VOS word order which is base-
generated in Ch’ol. VSO is derived from VOS by moving the object. This analysis offers a
straightforward account of the observed semantic properties of nominal arguments and
extraction asymmetries. I model how a base-generated account of VOS is derived through
linearization rules in the narrow syntax. The syntactic structure argued for here results in
testable predictions about the semantic and syntactic properties of nominal arguments in
other languages. I conclude that verb-initial languages do not constitute a uniform syn-
tactic class, and further investigation into structural dependencies of nominal and clausal
syntax can provide us with a window into the underlying similarities and differences
across languages.
DEPENDENCIAS MUTUAS DE LA SINTAXIS NOMINAL Y CLAUSAL EN CH’OL
Carol Roseman Little, Ph.D.
Universidad Cornell 2020
En esta tesis se investiga la estructura nominal y clausal en ch’ol, una lengua mayense
del sur de México, perteneciente a la rama cholana-tseltalana. La interpretación de los
sustantivos escuetos y las asimetrı́as de extracción en ch’ol, forman la base empı́rica para
desarrollar un análisis generado por la base (base-generated) de dichos datos hacia una
sintaxis de verbo-inicial. Se propone que la (in)definitud está reflejada en la estructura
sintáctica, lo cual demuestra que sólo los nominales escuetos que están dentro de la frase
verbal pueden tener una interpretación indefinida. Cualquier nominal generado o de-
splazado fuera de la frase verbal tiene que ser definido, o bien, puede tomar un alcance
amplio con respecto a otros elementos cuantificacionales. Se captura esta generalización
empı́rica al conectar la interpretación semántica con la estructura sintáctica. Los datos
acerca de los elementos que pueden o no ser desplazados de las expresiones nominales,
proporcionan evidencia que apoya la propuesta de la estructura sintáctica subyacente
presentada en este trabajo. Se sostiene que los nominales que son islas sintácticas, proveen
pruebas de que están contenidos dentro de un constituyente desplazado—cuando se
puede extraer un elemento de una frase nominal, la frase nominal no está en una posición
derivada por el movimiento, lo cual tiene implicaciones con respecto a la estructura del
orden verbo-inicial.
Ante estas conclusiones, se propone un análisis sintáctico del orden VOS, el cual está
generado por la base de dichos datos en ch’ol. El orden VSO se deriva del orden VOS
al mover el objeto. Dicho análisis proporciona una manera sencilla y directa de capturar
las propiedades observadas en los argumentos nominales y en las asimetrı́as sintácticas.
Se proponen reglas de linearización que permiten formalizar el análisis de VOS dentro
de la sintaxis estrecha (narrow syntax). La estructura sintáctica propuesta en esta tesis
brinda predicciones comprobables acerca de las propiedades semánticas y sintácticas de
los argumentos nominales en otras lenguas. Se concluye que las lenguas de verbo-inicial
no forman una categorı́a homogénea, por lo que una investigación adicional con respecto
a las dependencias estructurales de la sintaxis nominal y clausal, nos permitirá indagar
más a fondo las similitudes y diferencias subyacentes en las lenguas del mundo.
BIOGRAPHICAL SKETCH
Carol-Rose Little was born in Durham, North Carolina. She attended McGill University
and gradated with a B.A. (Hons) in Linguistics and Russian Studies. She is graduating
from Cornell with a Ph.D in linguistics and graduate minors in American Indian and In-
digenous Studies and Cognitive Science. Her research program brings together syntax,
semantics and morphology, rooted in a strong commitment to fieldwork and language
documentation. She investigates possible structural variations crosslinguistically and
how these structures interface with semantic computation. Her theoretical analyses draw
on data collected from fieldwork with understudied languages, such as Ch’ol (Mayan)
and Mi’gmaq (Algonquian). The members of her dissertation committee are Miloje De-
spić and Sarah Murray (co-chairs), Molly Diesing, and Jessica Coon (McGill University).
During her time as a Ph.D. student, Carol-Rose presented at distinguished conferences
such as the North East Linguistic Society (NELS), the West Coast Conference on Formal
Linguistics (WCCFL), and Sinn und Bedeutung (SuB). She has also published in the peer-
reviewed journal Glossa: A Journal of Generative Linguistics and co-edited volumes for the
proceedings of the Semantics and Linguistic Theory conference. Since 2015, Carol-Rose
has been conducting fieldwork in Chiapas, Mexico with speakers of Ch’ol.
iii
RESEÑA BIOGRÁFICA
Carol-Rose Little nació en Durham, Carolina del Norte, EEUU. Cursó la licenciatura
en la Universidad McGill en Montreal, Canadá en lingüı́stica y estudios rusos. Presenta
esta tesis doctoral a la Facultad de la Escuela de Postgrado de la Universidad Cornell
para optar al grado de doctora en lingüı́stica con dos especializaciones, una en estudios
amerindios e indı́genas y la otra en ciencias cognitivas. Su programa de investigación
reúne los campos de sintaxis, semántica y morfologı́a, basándose en un firme compro-
miso al trabajo de campo y la documentación lingüı́stica. Investiga posibles variaciones
estructurales en las lenguas del mundo y como estas estructuras interactúan con la com-
putación semántica. Sus análisis teóricos se basan en datos recopilados de su trabajo de
campo con lenguas poco estudiadas y subrepresentadas, como el ch’ol (lengua mayense)
y el mi’gmaq (lengua algonquina). Los miembros de su comité de la tesis doctoral son
Miloje Despić, Sarah Murray (co-directores), Molly Diesing y Jessica Coon (Universidad
McGill). Durante su tiempo como estudiante de doctorado, Carol-Rose presentó en con-
gresos distinguidos como el North East Linguistic Society (NELS), el West Coast Con-
ference on Formal Linguistics (WCCFL), y Sinn und Bedeutung (SuB). También ha publi-
cado en la revista revisada por paresGlossa: A Journal of Generative Linguistics y co-editado
volúmenes de las actas de la conferencia Semantics and Linguistic Theory (SALT). Desde
2015, Carol-Rose ha estado realizando trabajo de campo en Chiapas,México con hablantes
de ch’ol.
iv
Cha’añ xShenia
v
ACKNOWLEDGEMENTS
Completing a Ph.D. is not something accomplished alone. I have many, many people
to thank for supporting me, encouraging me and being there for me throughout all these
years. The following list is by no means exhaustive.
First and foremost, I would like to thank the Arcos López family for hosting me dur-
ing my stays in San Miguel since 2015. Specifically, Nicolás Arcos López, Evelina Arcos
López, Celia Álvaro Méndez, Celia Arcos Gúzman, Rosa López Pérez, Enrique Arcos
López, Shenia Arcos Álvaro, Marı́a Mendoza Pérez, and Nely Arcos Álvaro, for open-
ing up their homes to me and teaching me everything I know about Ch’ol. None of this
would have been possible without your support. I look forward to many more years of
collaboration. Of course, thanks to alä xQuique and little xCarolita for always keeping
things interesting. ¡Tyi kpusik’al mi ksu’beñetyla wokolix la’wälä! ¡Mi kweñ k’uxbiñetyla!
I would like to thank my dissertation committee, with whom I feel very fortunate
to have worked: Miloje Despić, Sarah Murray, Molly Diesing, and Jessica Coon. I am
honored to have had you all as advisors, mentors, and teachers throughout these years.
I am truly indebted to my co-chair Miloje Despić. On many occasions we have met
for hours talking through data puzzles and connections to theoretical analyses. Miloje’s
enthusiasm for linguistics is infectious. After meetings with him, I always feel invigo-
rated, motivated, and happy to be in this field. Miloje always made sure to congratulate
me on any conference or paper acceptance—usually being the first one to do so! I admire
Miloje’s continual and unwavering commitment to his research, teaching, and advising—
an example I strive for in the future.
I first met my co-chair, Sarah Murray, at an Algonquian conference in 2012 and again
in 2013. She encouraged me to apply to the Cornell Ph.D. program, and I am very happy
I did. I was fortunate to be her student in many seminars and classes, as well as her TA,
her RA, and her advisee. From research to teaching to fieldwork, Sarah sets an example to
be followed. Her many pieces of advice throughout the years have guided me to become
vi
a better researcher, fieldworker, teacher, and person. I often find myself relying on her
advice when I am conducting research, doing fieldwork, crafting handouts, giving talks,
writing papers, and developing courses. (This list is nonexhaustive!) I thank her for
inviting me to collaborate with her in the Cheyenne community of Lame Deer, Montana.
Of course, I cannot forget to thank her for the Ithaca-to-Montana road trip in the summer
of 2019 and for teaching me to drive manual—a skill that has come in handy many times!
Sarah is truly a role model for me in linguistics and in life.
I feel fortunate to have been Molly Diesing’s advisee and student. I always look for-
ward to our meetings and discussions of data and theory. Her questions and comments
push me to probe deeper into the data and theory, always yielding a better outcome. As
a student in her seminars, I was often left surprised at how quickly the time passed. I
would count it as a success if one day I could be even just half as engaging as Molly is in
the classroom.
Beginning when I was an undergraduate student at McGill University, Jessica Coon
has made a huge impact on my linguistic trajectory. Her field methods class in 2011
introduced me to what was to become the foundation for my future research—linguistic
fieldwork. Jessica is truly exemplary in her strong commitment to bridging theory and
linguistic fieldwork. Through Jessica, I first met Nicolás Arcos López and many other
Ch’ol speakers. Jessica has opened many doors for me; without her I would not be where
I am today. Thank you for everything, Jessica!
My time at Cornell and in Ithaca has been very rewarding—I couldn’t have picked a
better place to call my home base for these past years. Thank you to my cohort: Shohini
Bhattasali, Ryan Hearn, Mary Moroney, and Yanyu Long, and, of course, to our honor-
able cohort member Binna Stefánsdóttir. I am also indebted to Ekarina Winarto, whose
friendship I value very much. Mary, here’s to many more collaborations! Thank you also
to Andrea Hummel, Naomi Enzinna, Todd Snider, and Eszter Ótott-Kovács. Thank you
to MiaWindhearn for our discussions on all sorts of topics and trusting me when I invited
vii
you to come with me to Chiapas. I have enjoyed interactions with other faculty members
at Cornell, particularly John Whitman, John Bowers, and Sally McConnell-Ginet. Read-
ing groups at Cornell such as the Semantics Group and SynCirc were always productive
venues to talk about recent work and give practice talks. Thank you to their many mem-
bers! I would also like to thank Holly Boulia, Jenny Tindall and Gretchen Ryan for all
their help with my administrative questions.
I have benefited greatly from conversations with Ch’ol linguists and scholars, Juan
Jesús Vázquez Álvarez and Nicolás Arcos López. Juan Jesús’s dissertation on the gram-
mar of Ch’ol has beenmy guide to everything Ch’ol—I even have a shortcut onmy phone
to the pdf! In addition to providing a thorough description of classifiers in Ch’ol, Nicolás’s
thesis also contains many gems of transcribed spontaneous speech from people in San
Miguel to whom I have become close—always enjoyable to come across especially dur-
ing these last few months of writing. I was honored you both came to my dissertation
defense.
I am also deeply indebted to many of my friends in Chiapas who welcomed me
into their homes and shared their language and culture with me. Thank you to More-
lia Vázquez Martı́nez for all your hard work. Our collaborations have been extremely
fun and fruitful. We had some great adventures traveling around Chiapas, Guatemala,
and NewOrleans. Let’s hope for manymore! I also would like to thankMorelia’s mother,
Virginia Martı́nez in El Campanario, for working with me as well and making sure I had
enough empanadas to eat. Many thanks to Juan Javier Álvaro López for taking me on var-
ious adventures and for introducing me to literally everyone in Tumbalá. ¡Wokox awälä
cha’añ ta’ awäk’eyoñ kñijkañ akaru! Also thanks to his family in Ignacio Allende for hosting
me several times. I am indebted to Don Lucio Meneses Dı́az in La Ilusión for sharing
his language with me and letting me record him in his picturesque oasis of a house. I
also would like to thank Juana Karen Peñate Montejo for letting me include some of her
beautifully written poetry in this dissertation.
viii
My friends in San Cristóbal de las Casas always made my stays there so wonderful.
Thank you, Silvestre Gómez Jiménez and Daniel Ochoa Nájera at CELALI. Thank you
especially to SophieWalters, for always helpingmewith lodging arrangements etc. in San
Cristóbal. Thank you, Esteban Marván Mirón, for making the map of the Ch’ol speaking
region (available in three languages, by the way, English, Ch’ol, and Spanish!) and for
your continued efforts to support the involvement of Ch’oles in your archaeological work.
I was very fortunate to have spent time as a visiting scholar at Heinrich-Heine-
Universität, Düsseldorf in the fall of 2018. Thank you to Sebastian Löbner for encouraging
me to apply to the position and then hosting me. I also benefited from conversations with
Peter Sutton, Curt Anderson, and Hana Filip. Thanks, Peter, for organizing our 2019
European tour of Finnish counting and measuring constructions!
I have benefited from conversations with manymembers of the linguistics community
at conferences or other linguistics events. These people include Alan Bale, Conor Quinn,
Lauren Clemens, Justin Royer (my fellow Mayanist!), Dan Kaufman, Will Oxford, Ben
Bruening, Suzi Lima, Ryan Bennett, Robert Henderson (who was the one who put the
idea in my head to go and study a Mayan language), Mike Hamilton, Jon Ander Mendia,
Laura Kalin, Scott AnderBois, Danny Law, Michelle Yuan, Violeta Vázquez-Rojas Mal-
donado, Ana Aguilar Guevara, Roberto Zavala, José Armando Fernández Guerrero, and
many, many more. This list is by no means exhaustive. The work presented in this disser-
tation has also been influenced by the comments and questions of audiences at TbiLLC
13, NELS 50, WSCLA 24, FAMLi 5, SSILA, the 2020 LSA as well as audiences at Concordia
University, McGill University, SUNY Albany, and Syracuse University.
Many of my friends have kept me sane and made this journey possible. Thank you to
Thea Knowles for always encouraging me to strive for the best. I couldn’t have made
it without you and our writing retreats. Here’s to many more! Also thanks to the
Kirkkomäet (Silva, Annika, Jussi, Niklas) and Raija Pilvinen. My year in Finland with
you all was just the beginning of my interest in linguistics, kiitoksia kaikesta! Thank you
ix
also to my dear friends throughout all these years: Ayşe Bursalı, Abby Cryan, Angela
Musliner.
I would also like to thank my family for always supporting me and encouraging me
to follow what I love. Without you I would truly not be where I am today. Thank you
to my loving parents, Brown and Mark Little, my favorite sister Grace Little, and my
ever-supportive grandparents, Carol and Griffith. Your support has been immeasurable
throughout my Ph.D. I love you very much! Thank you also to Janice, Cameron, Trena,
Richard, Bonnie, and Gordon. And, of course, I can’t forget Ellie.
Financial support for the research in this dissertation is thanks to a National Science
Foundation Doctoral Dissertation Research Improvement Grant under grant no. BCS-
1852744 and an Engaged Cornell graduate student research grant as well as grants from
the American Indian and Indigenous Studies Program, Latin American Studies Program,
the Cognitive Science Program, and the linguistics department at Cornell.
x
TABLE OF CONTENTS
Biographical Sketch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Dedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi
Table of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiv
List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
List of Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvi
1 Introduction 1
2 Background on Ch’ol and verb-initial word order 6
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Linguistic and fieldwork context . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2.1 Ch’ol in the Mayan language family . . . . . . . . . . . . . . . . . . . 6
2.2.2 Where Ch’ol is spoken . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2.3 Status of Ch’ol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2.4 Orthographic conventions . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2.5 Previous descriptive work with Ch’ol . . . . . . . . . . . . . . . . . . 12
2.2.6 Fieldwork context and general methodology . . . . . . . . . . . . . . 13
2.3 Clause structure of Ch’ol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.3.1 Aspect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.3.2 Person marking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.3.3 Verbs in Ch’ol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.3.4 Alignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.3.5 Split ergativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.3.6 Clausal word order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.3.7 Negation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.3.8 The preverbal position . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.3.9 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.4 Ch’ol nominal basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.4.1 Nouns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.4.2 The noun class prefixes x- and aj- in Ch’ol . . . . . . . . . . . . . . . . 41
2.4.3 Plural markers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.4.4 Demonstratives and determiners . . . . . . . . . . . . . . . . . . . . . 44
2.4.5 Possessive constructions . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.4.6 Modifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.4.7 Relative clauses and =bä . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.4.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.5 Conclusion and a brief overview of verb-initial word order . . . . . . . . . . 49
xi
3 Structurally reflected (in)definiteness 53
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.2 Bare nouns in argument position . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.2.1 Absolutive arguments . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.2.2 Ergative subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.2.3 Scope interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.2.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.3 Deriving the Ch’ol pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.3.1 Semantic preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.3.2 The Mapping Hypothesis and indefinite bare arguments . . . . . . . 61
3.3.3 Existential closure of the event and negation . . . . . . . . . . . . . . 62
3.3.4 Deriving definite bare nominal subjects . . . . . . . . . . . . . . . . . 68
3.3.5 Moved absolutive subjects . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.3.6 Numerals in object position . . . . . . . . . . . . . . . . . . . . . . . . 72
3.3.7 Definite interpretations of bare internal arguments . . . . . . . . . . 75
3.3.8 Dimensions of definiteness and effects on word order . . . . . . . . . 77
3.3.9 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
3.4 Incorporated nominals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
3.4.1 Morphosyntactic properties of incorporation antipassives . . . . . . 82
3.4.2 New data: Semantics of incorporated nominals in Ch’ol . . . . . . . 87
3.4.3 Implications for an analysis . . . . . . . . . . . . . . . . . . . . . . . . 90
3.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4 Movement-derived islands 96
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
4.2 Subextraction data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4.2.1 Interrogative possessor extraction . . . . . . . . . . . . . . . . . . . . 98
4.2.2 Numeral extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
4.2.3 Demonstratives and non-interrogative possessors cannot extract . . 105
4.2.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.3 Evidence for object shift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
4.3.1 Adjunct placement as a diagnostic for object shift . . . . . . . . . . . 107
4.3.2 Interpretational differences of numerals and adjunct placement . . . 108
4.3.3 Constituent order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
4.3.4 Summary of object restrictions . . . . . . . . . . . . . . . . . . . . . . 111
4.4 Capturing the subextraction facts . . . . . . . . . . . . . . . . . . . . . . . . . 113
4.4.1 Proposal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
4.4.2 Numerals may not extract from moved constituents . . . . . . . . . . 115
4.4.3 Moved absolutive subjects block subextraction . . . . . . . . . . . . . 116
4.4.4 Shifted ditransitive objects block subextraction . . . . . . . . . . . . . 117
4.4.5 Banning demonstrative and non-interrogative possessor extraction . 118
4.4.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
4.5 Licensing subextraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
4.5.1 Agree and unlocking . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
4.5.2 Banning subextraction from specifiers . . . . . . . . . . . . . . . . . . 123
xii
4.6 Implications for theories of word order . . . . . . . . . . . . . . . . . . . . . 124
4.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
5 A base-generated account of verb-initial syntax 127
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
5.2 Core data and base generating the VOS/VSO distinction . . . . . . . . . . . 129
5.2.1 VOS objects = NPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
5.2.2 VSO objects = DPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
5.2.3 Other factors influencing word order . . . . . . . . . . . . . . . . . . 131
5.2.4 Word order restrictions to be captured . . . . . . . . . . . . . . . . . . 132
5.2.5 Base-generating verb-initial syntax in Mayan . . . . . . . . . . . . . . 133
5.3 Phrasal-fronting and head movement accounts of VOS/VSO . . . . . . . . 134
5.3.1 Coon’s (2010b) predicate-fronting account . . . . . . . . . . . . . . . 135
5.3.2 Clemens and Coon’s (2018) head movement account . . . . . . . . . 136
5.3.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
5.4 Additional empirical considerations . . . . . . . . . . . . . . . . . . . . . . . 140
5.4.1 Order of morphemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
5.4.2 Nonperfectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
5.4.3 Semantic and syntactic properties of objects (Chapters 3 and 4) . . . 147
5.4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
5.5 Base-generating verb-initial syntax . . . . . . . . . . . . . . . . . . . . . . . . 155
5.5.1 Order of morphemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
5.5.2 Nonperfective complements and nominals . . . . . . . . . . . . . . . 157
5.5.3 VSO objects: Semantic and syntactic features . . . . . . . . . . . . . . 158
5.5.4 Extension to ditransitives . . . . . . . . . . . . . . . . . . . . . . . . . 163
5.5.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
5.6 Formalizing right specifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
5.7 Implications for Mayan languages with fixed VSO word order . . . . . . . . 168
5.7.1 Reflexive objects in fixed VSO languages . . . . . . . . . . . . . . . . 171
5.7.2 Syntactic ergativity and fixed VSO Mayan languages . . . . . . . . . 171
5.7.3 VSO and syntactic ergativity . . . . . . . . . . . . . . . . . . . . . . . 175
5.7.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
5.8 VSO Mayan languages and the NP/DP hypothesis . . . . . . . . . . . . . . . 180
5.9 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
6 Conclusion 184
A Glossed narrative: Xi’ba 187
xiii
LIST OF TABLES
2.1 Ch’ol consonants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2 Ch’ol vowels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3 Aspect markers in Ch’ol. Tila lexical variants in parentheses. . . . . . . . . 14
2.4 Ch’ol person markers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.5 Ch’ol plural markers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.6 Participant plural set B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.7 Ch’ol verb roots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.8 Ch’ol stem moprhology (Coon, 2013, 28) . . . . . . . . . . . . . . . . . . . . 40
2.9 Demonstratives and determiners in Ch’ol . . . . . . . . . . . . . . . . . . . . 44
3.1 Available interpretations of bare arguments in Ch’ol . . . . . . . . . . . . . 59
3.2 Semantic notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.3 Examples of incorporated nominals with transitive verb roots . . . . . . . . 83
4.1 Subextraction summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.2 Summary of object restrictions and word order . . . . . . . . . . . . . . . . 112
4.3 Summary of object restrictions and adverbials . . . . . . . . . . . . . . . . . 112
5.1 Ch’ol object restrictions (Coon, 2010b, 363) . . . . . . . . . . . . . . . . . . . 132
5.2 Ch’ol object restrictions (Coon, 2010b, 363) (repeated) . . . . . . . . . . . . . 134
xiv
LIST OF FIGURES
2.1 Map of Ch’ol communities, courtesy of Esteban Mirón Marván. . . . . . . . 10
2.2 Photo of San Miguel, December 2019, by the author. . . . . . . . . . . . . . . 13
xv
LIST OF ABBREVIATIONS
* ungrammatical construction
# infelicitous construction
% speaker variation attested with respect to grammaticality
? odd construction
( ) optional
1 first person
2 second person
3 third person
A set A markers in Ch’ol, marks ergative/possessive
ABS absolutive
AGR agreement marker
AP antipassive
AF agent focus
AFF affirmative
ASV assertive
B set B markers in Ch’ol, marks absolutive
CLF classifier
COMP complementizer
GEN genitive
DEINAL deinalienable suffix
DEM demonstrative
DEP dependent suffix
DER derived adjective
DET determiner
DIR directional
xvi
DO direct object
DTV derived transitive verb
D.NML derived nominal suffix
ENCL enclitic
EP epenthesis
ERG ergative
EXCL exclusive
EXT existential
FOC focus
INTER interrogative
IO indirect object
IMP imperative
IPFV imperfective aspect
IV intransitive verb suffix
IRR irrealis
LV light verb
NEG negation
NOM nominative
NC noun prefix
N.CLF noun classifier
PART participant
PL plural
PREP preposition
POS positional
PRED predictive
PROG progressive
PRON pronoun
xvii
PFV perfective aspect
PLIN indeterminate plural
PRT particle
PSV passive
REA realis
RED reduplication
REP reportative
REL relative clause suffix
RS relational suffix
SG singular
SS status suffix
STAT stative suffix
TV transitive verb
xviii
Kty’añ I
Jiñi kty’an che’ bajche’ ek’tyak mi imos jiñi pañchañ,
che’ bajche’ xk’äk’äs yäpyäpña che’ ik’yoch’añ.
Jiñi kty’añ che’ bajche’ k’äñk’äñ ixim che’ mi ilok’el tyi pañumil,
chäkchab ixim mi isujtyel che’ mi yochel tyi ap’ätyälel.
Ili kty’añ ach’päk’añbä lum ya’ baki mi ityejchel iyälas xpejpemtyak.
Juana Karen Peñate Montejo
Printed with permission of the poem’s author.
xix
CHAPTER 1
INTRODUCTION
The structural position that a given nominal occupies has links to both its possible semantic in-
terpretations and syntactic properties. This fact is generalizable across a number of typologically
diverse languages. However, the way in which nominals are linearized can vary. One obvious
manifestation of this syntactic variation is the diversity of word orders across languages. The goal
of this dissertation is to take a closer look into the structural position of nominal arguments and
their dependencies on clausal syntax. I investigate the syntactic and semantic properties of nom-
inal arguments in a verb-initial language and the implications these properties have for analyses
of verb-initial word order. At a higher level, this dissertation makes the claim that while certain
dependencies between constituents hold across many languages, the way in which a language
linearizes these constituents can differ.
The mutual dependencies of nominal and clausal syntax can be seen in a number of domains
and are exhibited across typologically diverse languages. For instance, the movement of nomi-
nal arguments from their base-generated position has been connected to different semantic and
syntactic properties. On the semantic side, it has been shown that movement of nominals out-
side the VP is directly correlated to their interpretation: for instance, object shift and a specific
interpretation of the object has been shown to hold in a number of languages (Diesing, 1992;
Diesing and Jelinek, 1995; Kelepir, 2001; Aldridge, 2004). On the syntactic side, this movement
also induces freezing effects; that is, these constituents are movement-derived islands. This has
been shown before in, e.g., English (Chomsky, 1973; Lasnik, 2001), German (Diesing, 1992), and
Ukrainian (Mykhaylyk, 2010).
Specifically, I investigate the syntactic and semantic properties of nominal arguments in Ch’ol,
a verb-initial Mayan language of southern Mexico, with data from my fieldwork gathered over
the course of several years. Verb-initial word orders make up 12–19% of the world’s languages
(Tomlin, 1986; Dryer, 2013; Van Everbroeck, 2003; Clemens and Polinsky, 2017). Various propos-
als of verb-initial syntax have been put forth including a base-generated account of VOS, pursued
1
for Mayan languages (England, 1991; Aissen, 1992; Otaki et al., 2019), Austronesian languages
(Chung, 1998, 2006), and the Wakashan language Nuu-chah-nulth (Wojdak, 2007). A phrasal-
fronting account of VOS word order has been posited for Mayan (Coon, 2010b) and Austronesian
languages (Aldridge, 2002; Chung, 2006; Massam, 2000). To derive verb-initiality, head move-
ment of the verb to a position higher than the subject has been posited for Afro-Asiatic languages
(Fassi Fehri, 1993; Ouhalla, 1994) and Celtic languages (Sproat, 1985; McCloskey, 1991). Clemens
(2014, 2019) and Clemens and Coon (2018) adopt a head movement account and posit an addi-
tional prosodic constraint that reorders an NP object with the verb that selects it, generating VOS.
In this dissertation, I compare these three families of proposals in light of new observations on
the semantic and syntactic properties of nominal arguments in Ch’ol, ultimately favoring a base-
generated account. I conclude that verb-initial languages do not form a unified syntactic class and
offer diagnostics for probing structural structures.
This dissertation has three main conclusions: (i) the interpretation of bare argument nomi-
nals is structurally reflected; (ii) (non)islandhood has implications for whether a constituent has
moved; (iii) verb-initial word order can be base-generated in ergative languages. I elaborate on
each of these points below.
First, I argue that in Ch’ol (in)definiteness is reflected structurally. In particular, only bare
nominals remaining within the VP may be indefinite. Other bare nominals, which are generated
outside the VP or which have moved to a VP-external position, may not receive an existential
indefinite interpretation. Bare nouns in ergative subject position, for instance, can only be inter-
preted as definite in Ch’ol as they are generated outside the VP. This supports a syntactic account
of (in)definiteness with implications for the list of type shifters across languages.
Second, I argue that movement-derived islands explain extractability asymmetries. Under the
Freezing Principle, moved constituents are islands for extraction—they do not allow elements to
be extracted from them. This has implications for accounts of verb-initial word order. Under a
phrasal-movement account of verb-initial word order, a phrase containing the verb and object
moves to a position above the subject, generating VOS. However, extraction from VOS objects is
possible. A phrasal-movement account would rule out extraction from VOS objects as the VOS
2
object is contained within a moved constituent.
Third, these findings support a particular analysis for word order, namely base-generating
VOS, with transitive subjects in a right specifier. VSO is derived from VOS bymoving the object to
an outer projection of the verbal domain. This analysis capitalizes on the link between object shift
and the semantic and syntactic properties of VSO objects. I also posit that verb-initial word order
is not derived in a uniform fashion across languages, offering testable predictions for particular
analyses.
The dissertation is organized into five main chapters and an appendix. The contents of each
chapter are detailed below.
In Chapter 2, I provide background on the morphosyntactic properties of Ch’ol, based on
descriptions in Warkentin and Scott (1980), Coon (2010a) and Vázquez Álvarez (2011) as well as
my own fieldwork from 2015 to 2020. I further present information on where Ch’ol is spoken
and the methodology I used to collect the data in this dissertation. I lay out my assumptions for
how case assignment, agreement markers, and split ergativity are derived, following Coon (2013,
2017). This chapter concludes with a brief introduction on the three families of verb-initial word
order analyses that I compare in Chapter 5.
In Chapter 3, I investigate the interpretational possibilities of bare nominals in Ch’ol and their
implications for clausal structure. I argue that the Mapping Hypothesis from Diesing (1992) can
capture the fact that only bare nominals which are next to the verb can be indefinite. This is
because they are within the domain of existential closure, namely the VP. The analysis I propose
is based on the notion that indefinite internal arguments are instances of pseudo-incorporation.
Ergative subjects or nominals that have moved outside the VP cannot be interpreted as narrow
scope indefinites. This chapter supports analyses of word order where VSO is derived from VOS,
in part, via movement of the object, capturing that quantifiers, such as numerals in VSO object
position, obligatorily take wide scope. I also discuss implications for other constructions with bare
nominals, namely those of incorporated nominals. Unlike bare nominals in argument position,
these incorporated nominals do not refer—that is, they do not introduce possible antecedents for
agreement or pronominal anaphora.
3
Chapter 4, based on Little (2020), presents data on subextraction from nominals, focusing on
the extraction of interrogative possessors and numerals from different argument positions. Subex-
traction is possible from VOS objects, but not VSO ones. I argue that VSO objects have moved,
thus creating an island for extraction, following from the Freezing Principle. In this chapter, I also
present a way to formalize these freezing effects under Agree-based frameworks for extraction. I
end this chapter with implications for theories of verb-initial word order and a brief discussion of
dialectal variation with respect to subextraction.
Chapter 5 ties together the findings of Chapters 3 and 4 to argue for a base-generated account
of verb-initial syntax. I propose that there is a link between object shift and word order, in that
VSO is derived from VOS via movement of the object. I first compare the following analyses for
verb-initial word order: the predicate-fronting analysis from Coon (2010b) and the head move-
ment account in Clemens and Coon (2018). While both of these analyses can account for core data
with respect to VOS/VSO word order, certain questions arise when other empirical patterns are
taken into consideration. The issues that are discussed concern the order of morphemes, nominal-
ized complements, the islandhood of fronted phrases, and interpretational differences of VOS and
VSO objects. I consider potential modifications to both these accounts in order to capture the ad-
ditional empirical patterns from Chapters 3 and 4, but ultimately favor a base-generated analysis,
based on England (1991) and Aissen (1992). I provide a formalization of right specifiers, drawing
on a suggestion from Otaki et al. (2019). This formalization connects facts about VOS languages
and ergativity: only ergative languages may parameterize specifiers on the right. I end this chap-
ter with implications for rigidly VSO Mayan languages, syntactic ergativity, and areas for future
investigation with respect to nominal syntax in Mayan languages.
Chapter 6 is the conclusion. I discuss the implications of the analyses proposed in this disserta-
tion as well as specific points of future investigation. I conclude that properties such as alignment
are of particular interest in determining the syntax of verb-initial languages.
In Appendix A, I have included a fully glossed narrative from a speaker of the Tumbalá dialect
of Ch’ol. This narrative and its audio along with additional narratives and audio may be found in
the Archive of the Indigenous Languages of Latin America. Throughout the dissertation, I refer-
4
ence excerpts from this narrative by including the name of the narrative and corresponding line
number. Excerpts from other transcriptions are referenced only by the name of the transcription.
5
CHAPTER 2
BACKGROUNDON CH’OL AND VERB-INITIALWORDORDER
2.1 Introduction
In this chapter, I provide background on Ch’ol morphosyntax, highlighting relevant points for
theories of verb-initial word order. The descriptions below are based on previous work such as
Warkentin and Scott (1980), Vázquez Álvarez (2002, 2011), and Coon (2010a), as well as my field
notes from 2015 to 2020. I detail the analyses I adopt for agreement, case and split ergativity in
Ch’ol based on previous work (e.g., from Coon (2013, 2017)). I first discuss the linguistic context
of Ch’ol as well as where the fieldwork for this dissertation was conducted (Section 2.2). I then
move on to the clausal domain, providing background on person-marking, alignment, split erga-
tivity, and word order (Section 2.3). In Section 2.4, I present basics on the structure of nominals
and elements occuring within nominals. I conclude in Section 2.5, with with a brief summary of
previous work on verb-initial syntax.
2.2 Linguistic and fieldwork context
2.2.1 Ch’ol in the Mayan language family
There are currently about 30 Mayan languages spoken today by over 6 million people in Mex-
ico, Guatemala, Belize, and Honduras, as well as diaspora communities in the United States and
Canada. Ch’ol belongs to the Greater Tseltalan family, in the Ch’olan subgroup. A list of Mayan
languages is given in (1), with Ch’ol underlined. The asterisk indicates that the language is no
longer spoken; names in parentheses indicate other ways of referring to the language; semi-
colons signal subgroupings within families. (1) lists 33 languages; note that sometimes Achi,
6
Akatek, and Chalchitek are described as dialects of K’iche’, Q’anjob’al and Awakatek, respectively
(Bennett et al., 2015).
(1) Mayan family classification (Bennett et al. 2015, 4)
a. Huastecan: Wastek (Teenek) and Chicomulceltec (Kabil)∗
b. Yukatekan: Yukatec (Maya), Lacandon (Lakatun); Mopan, Itza’
c. Greater Tseltalan/Ch’olan-Tseltalan:
i. Ch’olan: Ch’ol, Chontal (Yokot’an); Ch’orti’, Ch’olti’∗
ii. Tseltalan: Tseltal, Tsotsil
d. Greater Q’anjob’alan:
i. Q’anjob’alan: Q’anjob’al, Akatek, Popti’ (Jakaltek); Mocho’ (Motocintlek)
ii. Chujean: Chuj, Tojol-ab’al
e. K’ichean–Mamean:
i. K’ichean: Q’eqchi’; Uspantek; Poqomchi’, Poqomam; Achi, K’iche’, Kaqchikel,
Tz’utujil, Sakapultek, Sipakapense
ii. Mamean: Tektitek (Teko), Mam; Awakatek, Chalchitek, Ixil
The Ch’ol language and the people are both referred to as Ch’ol, sometimes written as ‘Chol’.
Throughout this dissertation, I use the spelling ‘Ch’ol’. Ch’ol speakers refer to their language as
lakty’añ ‘our language’ and the people as lakpi’älob ‘our friends’. The name ‘Ch’ol’ is of unknown
origin. Some have reported that Ch’ol comes from chol ‘cornfield’. However, this is generally
agreed to be a folk etymology without basis in regular historical sound change: there is no ev-
idence that would suggest that ch in chol would become ch’ in the name for the language and
people (D. Mora-Marı́n, personal communication, April 2, 2020 and N. Hopkins, personal com-
munication, April 26, 2020).
2.2.2 Where Ch’ol is spoken
Ch’ol is spoken in southernMexico in communities in the Mexican states of Chiapas, Tabasco and
Campeche, as shown in the map in Figure 2.1. Locations with large populations of Ch’ol speakers
include Tila, Tumbalá and Palenque (these are all cabeceras municipales or ‘municipal capitals’).
There are also communities of Ch’ol speakers in the Mexican state of Campeche, who moved
there during the 1960s (Rodrı́guez Ceja, 2012), as well as some recently founded communities in
7
southeastern Chiapas (see map in Figure 2.1 and discussion in Vázquez Álvarez (2019) on the
sociolinguistic profile of the variant of Ch’ol spoken in Frontera Corozal).
Ch’ol has three mutually intelligible dialects: Sabanilla, Tila and Tumbalá. The Sabanilla di-
alect is sometimes described as a subvariety of the Tila dialect (Vázquez Álvarez, 2011, 5). The
Tumbalá dialect is spoken in the municipalities of Tumbalá, Salto de Agua and Palenque. The Tila
dialect is spoken in the municipality of Tila and parts of Tabasco. The Sabanilla dialect is spoken
in Sabanilla and Huitiupán (Nicolás Arcos López, personal communication, April 26, 2020). These
varieties differ slightly in their phonology and morphology, but the greatest variation is seen in
the lexicon. For instance in Tumbalá the progressive aspect marker is woli, whereas Tila speakers
use choñkol. Sabanilla speakers use yäkel, which Vázquez Álvarez (2011, 5) reports to be borrowed
from the Tseltal progressive form yakal. The data in this dissertation come from the Tumbalá vari-
ant of Ch’ol. I note dialectal differences when relevant.
2.2.3 Status of Ch’ol
There are about 222,000 speakers of Ch’ol including about 40,000 monolinguals (Ethnologue,
2018).1 While Ch’ol is still used by multiple generations, language shift towards Spanish is ob-
served, especially among younger generations of speakers. Gómez Jiménez (2015, 19) reports that
Spanish is displacing Ch’ol in many contexts, resulting in a loss of both the Ch’ol language and
culture. For example, Ch’ol is no longer used as the main language of communication in local
governments. This is especially visible in younger generations of Ch’ol speakers as many harbor
negative attitudes towards their language due to feelings of “marginalization, discrimination, and
other conditions imposed by the Mexican state” (Gómez Jiménez, 2015, 18).
That being said, there are many ongoing efforts by Ch’ol teachers, authors and scholars to fur-
ther foster the preservation and retention of the Ch’ol language and culture. These include those
working at institutions such as the Centro Estatal de Lenguas, Arte y Literatura Indı́genas (The Center
for Indigenous Languages, Art and Literature of Chiapas, CELALI), Casas de la Cultura (Cultural
1This number comes from around 2010. I suspect the number of Ch’ol speakers to be much greater.
8
Houses) in municipalities such as Salto de Agua and Tumbalá, and universidades interculturales
(intercultural universities) in Tabasco and Chiapas (UIET and UNICH, respectively).
9
10
Figure 2.1: Map of Ch’ol communities, courtesy of EstebanMirónMarván.
2.2.4 Orthographic conventions
Throughout this dissertation I will use the orthographic system of Ch’ol agreed upon in 2010
by Ch’ol writers and bilingual teachers. This orthography is based on the one proposed by the
author and native speaker of Ch’ol, José Dı́az Peñate in Dı́az Peñate (1992) and updated in INALI
(2011). Table 2.1 shows Ch’ol consonants written in the International Phonetic Alphabet with the
orthographic variant in parentheses. Table 2.2 shows the vowel inventory of Ch’ol. Ch’ol has a
sixth vowel, the high central unrounded vowel [1], orthographically ä. Historically, there was a
long and short vowel contrast in Mayan languages but long and short vowels merged with their
counterparts in the Ch’olan branch, except for *aa and *a, where *aa became *a, and *a became *ä
(Kaufman and Norman, 1984, 85).
Table 2.1: Ch’ol consonants
Labial Alveolar Post-Alveolar Palatal Velar Glottal
Implosive á (b)
Plosive p c (ty) k P(’)
> >
Ejective p’ ts’ (ts’) tS’ (ch’) c’ (ty’) k’
> >
Affricate ts (ts) tS (ch)
Fricative s S (x) h (j)
Nasal m ñ (ñ)
Approximant w l j (y)
11
Table 2.2: Ch’ol vowels
Front Central Back
High i 1 (ä) u .
Mid e o
Low a
The vowels in Table 2.2 contrast with their aspirated counterpants, represented orthographi-
cally as -Vj. For instance, sak’ ‘itchy’ and sajk’ ‘grasshopper’ constitute a minimal pair where the
minimal contrast is the aspiration of the vowel (Coon, 2017, 649). Vowel aspiration is also a pro-
ductive process used to derive passives, e.g., the root mäñ ‘buy’ is mäjñ ‘be bought’ in the passive,
where the minimal contrast is vowel aspiration.
2.2.5 Previous descriptive work with Ch’ol
There exists relatively extensive documentation of Ch’ol as well as earlier forms of Ch’ol. For
instance, Proto-Ch’olan served as the basis for the Classic Lowland Mayan writing (i.e., Mayan
glyphs) (Mora-Marı́n, 2009a), receiving much attention in the epigraphical literature (e.g., Schele
(1982); Kaufman and Norman (1984); Bricker (1986); Houston et al. (2000)). Early grammars
and documentation on Ch’ol from the 20th century include Bercerra (1937), Aulie and Aulie
(1978) (the Ch’ol-Spanish dictionary), Attinasi (1973, 1976), Warkentin and Scott (1980), and
Warkentin and Brend (1974). Whittaker and Warkentin (1965) and Hopkins et al. (2016) contain
collections of stories and narratives in Ch’ol. Josserand (2001) provides a detailed description of
Ch’ol ritual language. A comprehensive grammar of Ch’ol was published by Vázquez Álvarez
(2011) as his doctoral dissertation, and Coon (2017) provides a recent grammar sketch.
12
2.2.6 Fieldwork context and general methodology
Unless otherwise noted, the data in this dissertation come from working with Ch’ol speakers
over the course of five years, beginning in May of 2015. I worked mostly with speakers of the
Tumbalá dialect spoken in San Miguel, Salto de Agua, Chiapas, Mexico, about 16 miles from the
city of Palenque, see map in Figure 2.1 and photo in Figure 2.2. I traveled to San Miguel about
twice a year and resided there for one to three months at a time. The data in this dissertation
come from transcribed narratives, spontaneous speech, and elicited speech that I collected during
my stays in San Miguel. For elicited speech, I asked for grammaticality and felicity judgements,
following methodology and discussion in, e.g., Matthewson (2004), Murray (2015), and Gillon
(2015). Contextswere provided in Spanish or Ch’ol depending on the task. After 2016, I conducted
fieldwork with speakers mainly in Ch’ol and sometimes in Spanish. I primarily worked with
groups of speakers, rather than one-on-one interviews. Elicitation sessions occurred in the houses
of the language consultants, so speakerswithin that householdwould often discuss questionswith
each other before giving me a judgement. I primarily worked with a group of five women who
were monolingual Ch’ol speakers and understood Spanish, but did not speak it. When asking
for translations of sentences into Spanish, I consulted younger bilingual speakers who learned
Spanish in school.
13
Figure 2.2: Photo of San Miguel, December 2019, by the author.
Ch’ol is still used daily for most communication within the community of San Miguel. Com-
munity announcements, meetings and church services are all conducted in Ch’ol. Ch’ol is the first
language learned by children in the community; they are introduced to Spanish through school
and media. In addition to Spanish, some speakers may have some knowledge of Tseltal, another
Mayan language spoken locally. In San Miguel, Spanish is used for doing business with outsiders
and in educational settings. Education at the local preschool, primary school and secondary school
is all in Spanish. Students will use Spanish when talking to teachers, but use Ch’ol when talking
amongst themselves. Writing in Ch’ol is not explicitly taught in school, but some speakers have
taught themselves to write in the language. The Arcos López family, who hosted me during my
stays in San Miguel, use only Ch’ol to communicate with one another and with me. This is to say,
my fieldwork experience was quite immersive.
14
2.3 Clause structure of Ch’ol
Ch’ol is a synthetic, predicate-initial, head-marking language, as can be seen in example (2). In
(2) there is a high morpheme-to-word ratio, arguments follow the aspect marker and verb, and
agreement, such as the A3 prefix, is marked on the verb.
(2) [ASP Wol=ix=me=ku=ta’ ] [V i-ch’oj-ch’oj-tsep-tyak-ob ] [O tye’ ]
PROG=ASP=MIR=AFF=REA A3-ONO-RED-cut-PLIN-PL wood
[S x-mel bij-tyak-ob ].
NC-make road-PLIN-PL
‘The various construction workers are already chopping up trees, I’m telling you.’
Below, I detail background on the elements in (2) such as aspect, person marking, verbs and
verbal affixes, alignment, and word order, as well as outline my assumptions about how case,
agreement and split ergativity are derived in Ch’ol.
2.3.1 Aspect
As in other Mayan languages, Ch’ol clauses are marked for aspect, rather than tense. Ch’ol has
three aspects: perfective, imperfective and progressive.
Table 2.3: Aspect markers in Ch’ol. Tila lexical variants in parentheses.
Short form Long form
perfective ta (tyi) tsa’/ta’
imperfective mi muk’, mu’
progresive wo (choñ) woli (choñkol)
The perfective aspect indicates viewing “a situation as a single whole, without distinction of
the various separate phases that make up that situation” (Comrie, 1976, 15). Perfective forms are
translated into past tense in English as in (3). The perfective aspect can also appear in antecedent
of nonpast conditionals as in (3d).
15
(3) a. jiñi k-ty’añ tsa’=bä i-käñ-ty-es-a-yoñ k-tyaty k-ña.
DET A1-language PFV=REL REL-learn-DT-CAU-DT-B1 A1-father A1-mother
‘the language that my parents taught me’ Don Lucio
b. Ya’ ta’ majl-i-y-ob tyi puts-’el ili wiñik-ob ba’añ ixim
there PFV go-IV-PL PREP flee-NML DET man-PL LOC corn
‘These men fled to where the corn was.’ Xi’ba.
c. Ta’=tyo=bi i-wejl-a a k’ajk a x-kuj.
PFV=still=REP A3-blow-DTV PAR fire PAR NC-owl
‘The owl blew on the fire.’ Xi’ba.46
d. Mi tsa’ chojk-i-yety lok’el tyi aw-e’tyel. . .
if PFV throw.PSV-IV-B2 DIR:out PREP A2-work
‘If you have lost employment. . . ’ (Lit. ‘If you have been thrown out of your work.’)2
In casual speech, the short form of the aspect marker is often pronouncedwith the set Amarker
as in (4). Orthographic forms are in parentheses on the right.
(4) a. [ ta-k mel-e tamal ] (ta’ kmele tamal)
PFV-A1 make-TV tamale
‘I made tamales.’
b. [ ci kutS-u siP ] (ta’ ikuchu si’)
PFV.A3 carry-TV wood
‘She carried wood.’
Throughout this dissertation, I will label the projection that hosts the perfective aspectual
marker IP (i.e., InflP). Attinasi (1973) and Warkentin and Scott (1980) describe tsa’/ta’ as a past
tense marker, but as demonstrated below, many past tense clauses appear without the perfective
aspect marker. Coon (2013, 39) notes that tsa’/ta’ could bundle both past and perfective features to-
gether. Relevant for the empirical pattern in this dissertation is that tsa’/ta’ contrasts with the woli,
the progressive aspect marker andmi, the imperfective aspect marker. I refer to the group contain-
ing progressive and imperfective aspects as nonperfectives, to contrast them with the perfective
aspect.
The two nonperfective aspects marked in Ch’ol are the imperfective (IPFV) and progressive
(PROG) (6). The imperfective aspect marks habitual and continuous nonprogressive readings, such
as in (5a). In (5b), from a narrative, the speaker relates how her mother used to take a certain type
2From the General Consulate of Mexico in Phoenix, AZ: https://consulmex.sre.gob.mx/
phoenix/index.php/servicios-consulares-en-lenguas-indigenas/ch-ol
16
of medicine. The imperfective is used as this was a habitual action, even though it was in the past,
indicated by wajalix ‘back then’.
(5) a. Mi i-weñ jap lembal.
PFV A3-a.lot drink liquor
‘He drinks a lot.’
b. Mu’=tyo i-weñ-k’ux k-mama, wajali=x
IPFV=still A3-very-eat A1-mother back.then=already
My mother used to eat it a lot back then.’ Xi’ba.65
The progressive, on the other hand, is used for events which are ongoing, as in the examples
in (6).3 As in (6b), it can be used in past tense contexts as well.
(6) a. x-wi’il-kol-el-ob woli=x i-ñajäy-el i-cha’añ i-k’äñ jiñi
NC-after-grow-NML-PL PROG=already A3-forget-NML A3-PREP A3-use DET
lakty’añ.
PART.PL=A1-language
‘As for the younger generations, they are forgetting to use Ch’ol.’ Don Lucio
b. Woli i-juch’ sa’ x-Nely, che’ ta’ k’oty-i x-Norma.
PFV A3-grind masa NC-Nely COMP PFV arrive-IV NC-Norma
‘Nely was grinding masa when Norma arrived.’
In contrast to the perfective aspect marker, Coon (2013) analyzes the nonperfective markers as
complement-taking predicates. I discuss this aspectual difference more in Section 2.3.5.
2.3.2 Person marking
Ch’ol has two sets of personmarkers, referred to in theMayan literature as Set A and Set Bmarkers.
Set A markers in Ch’ol are prefixes that index possessive- or ergative-marked nominals. Set B
markers are suffixes, which index absolutive arguments, as summarized in (7).
(7) PERSON MARKING LABELS IN MAYAN
a. Set A: ergative, possessive
b. Set B: absolutive
3The subject xwi’ilkolelob ‘younger generation/teenagers’ is likely in a topicalized position, accounting
for why the second position clitic appears on the progressive aspect, after the topic in (6).
17
Ch’ol person markers and their allomorphs are given below in Table 2.4. Pronouns are also pro-
vided in the third column. Ch’ol pronouns are based on the Proto-Mayan form *ha’-. The forms
joñoñ and jatyety are probably derived from *ja’-oñ and *ja’-ety (Hopkins et al., 2008; Mora-Marı́n,
2009b). These forms are related to the determiner jiñ(i) in Ch’ol.
Table 2.4: Ch’ol person markers.
Set A (ergative/possessive) Set B (absolutive) Pronouns
1st k-/j- -(y)oñ joñoñ
2nd a(w)- -(y)ety jatyety
3rd i(y)- Ø jiñ
The glides y and w in Table 2.4 appear to break up vowel hiatus; the first person marker k- is
realized as j- ([h]) before velar consonants. There is no overt realization of the set B marker for
third person, so I therefore do not include set B markers in the glosses for third person. There is
no gender distinction in Ch’ol, so throughout the dissertation, I gloss third person forms with ‘he’
or ‘she’ (when not apparent through context), with the understanding that either interpretation is
possible.
I follow previous work (Woolford, 2000; Coon, 2013; Preminger, 2014) in analyzing the set B
markers as clitics. They are generated via an Agree relation, as schematized in (8). A low func-
tional head, labeled as Voice in (8), enters into an Agree relation (in the sense of Chomsky (2000,
2001)) with the object in the complement of the verb, triggering set B morphology, which attaches
to the verbal status suffix in SS. Under a movement approach in (8), clitics are D elements that
move from the DP in the complement of the verb and attach to the functional head SS. Following
conventions in previous work (Vázquez Álvarez 2011, Coon 2013), I will indicate set B markers
with a hyphen ‘-’, instead of ‘=’. The status suffix projection (SSP) indicates the edge of the verbal
projection. I label the projection that introduces the transitive subject as VoiceP. I note that the
projection I label as VoiceP has been described as bundling features of vP and VoiceP together; see
Coon and Preminger (2013) and Harley (2017) for more details on the mixed properties of Voice in
Ch’ol.
18
(8) a. Ta’ a-k’el-e-yoñ.
PFV arrive-TV-B2
‘You saw me.’
b. SSP
SS=oñ VoiceP
Voice VP
cli V DPtic doub D pro (1SG)ling
For absolutive subjects, also marked with set B clitics, I(nfl), not Voice, licenses and Agrees
with the absolutive subject and the clitic attaches to the edge of the verbal projection.
Set A markers are agreement markers that I take to be the result of an agreement relation with
a low functional head and a nominal in its specifier. This is schematized in (9).4 The transitive
verb suffix is hosted SSP.
(9) a. Ta’ i-mek’-e nene aj-Rosa.
PFV A3-hug-TV baby NC-Rosa
‘Rosa hugged the baby.’
b. SSP
SS VoiceP
-e
-TV
DPk Voice’
aj-Rosa Voice VP
NC-Rosa ik- mek’
A3- hug V NP
nene
baby
Plural markers are given in Table 2.5 and may appear on nouns and verbs. They can reflect
plural arguments indexed by set A and set B markers. The third person plural marker is used with
4This agreement configuration is what Legate (2014, 37) refers to as “inherent” agreement, analogous
to inherent ergative case assignment. Under an approach where agreement and case licensing go hand in
hand, this type of agreement is expected. What is important is that set A markers are the result of a local
agreement relation between a low functional head and the subject or possessor in a specifier position.
19
human-denoting and some animate-denoting nominals. The participant pluralizer la appears with
first and second person markers.
Table 2.5: Ch’ol plural markers
[+participant] la
[–participant] -ob
Plural markers are not obligatory for a plural interpretation in Ch’ol (Little, 2018b). The plural
marker -ob patterns similarly to the non-inflectional plural marker described in Halkomelem Sal-
ish (Wiltschko, 2008) as well as the optional plural marker in Yucatec Maya, which Butler (2011)
analyzes as being adjoined to the DP. Further investigation is required to better describe the plural
marker’s distribution in Ch’ol, but for now, I assume that the plural is noninflectional and is an
adjoined modifier.5 See, however, additional details on plurality in Chapter 3, Section 3.3.7.
Participant plurals are composed of the first or second personmarker and la. Table 2.6 provides
set B markers of participant plurals, with la.
Table 2.6: Participant plural set B
First person plural -oñ=la
First person plural exclusive -oñ=laj-oñ
Second person plural -ety=la
I analyze the exclusive first person plural as being morphologically complex, composed of the
participant pluralizer and first person set B marker as per discussion in Little (2018a).6,7
5The plural -ob can index both subjects and objects in transitive clauses, but it may not only reference
the object, as indicated by the translations in (1). This points towards an analysis where the plural marker
is adjoined to a position above the subject and object and can be ‘omnivorous’ (as per Nevins (2011))—that
is, it can pluralize both subject and object in some cases.
(1) Tyi i-k’el-e-yob.
PFV A3-see-PL
‘They saw them, They saw him, *He saw them.’ (Tila Ch’ol)
Note that ‘he saw them’ is not a banned construction in the language—with overt pronouns this interpre-
tation can be specified. I thank Jessica Coon for pointing this out.
6=La oftentimes undergoes vowel assimilation when -oñ attaches to it in the exclusive form (laj+oñ →
lojoñ).
7Ch’ol has been described as having an inclusive and exclusive distinction (e.g. Kaufman and Justeson
20
2.3.3 Verbs in Ch’ol
Ch’ol verbal roots are generally CVC in shape and can be transitive or intransitive. Examples are
given in Table 2.7. These verbal roots contrast with derived verbs, further discussed below.
Table 2.7: Ch’ol verb roots
Transitive Intransitive
män ‘buy’ majl ‘go’
choñ ‘sell’ tyäl ‘come’
k’ux ‘eat’ jub ‘descend’
k’el ‘see’ lets ‘ascend’
kuch ‘carry’ wäy ‘sleep’
ch’il ‘fry’ kol ‘grow’
mek’ ‘hug’ chäm ‘die’
bek’ ‘pour’ ajñ ‘run’
xul ‘break’ käl ‘stay’
Below, I provide background on the various status suffixes that appear with the verbal roots
in Table 2.7. I additionally provide information on derived transitive verbs and valency changing
morphemes.
Intransitive roots
Intransitive roots appear with the intransitive verb suffix -i and set B markers with the perfective
aspect marker, as shown in (10). Intransitive verb suffixes are hosted in a status suffix projection.
2003, Law 2009, Coon 2010a, Vázquez Álvarez 2011). But, as I describe in Little (2018a), the exclusive form
is morphologically more complex than the inclusive and is formed by adding the first person marker -
oñ to the inclusive form. Effectively, the exclusive form contains the inclusive form. In Little (2018a), I
provide evidence that the inclusive is not a traditional inclusive form (i.e., including speaker and hearer in
its denotation), but in fact a general first person plural form, unspecified for the inclusion of the hearer. The
exclusive, on the other hand, is morphologically and semantically more specified to exclude the hearer.
21
(10) a. Ta’ ju’b-i-yoñ.
PFV descend-IV-B1
‘I descended.’
b. Ta’ lets-i.
PFV ascend-IV
‘She ascended.’
c. Ta’ wäy-i-yety.
PFV sleep-IV-B2
‘You slept.’
With nonperfective aspect markers, intransitive verbs appear with the nominalizing suffix -el
and set A personmarkers, as shown in (11). I discuss these aspectual differences further in Section
2.3.5.
(11) a. Woli k-jul-el.
PROG A1-arrive-NML
‘I am arriving.’
b. Mi i-lets-el.
IPFV A3-ascend-NML
‘She ascends.’
c. Woli a-wäy-el.
PROG A2-sleep-NML
‘You are sleeping.’
Root transitive verbs
With the perfective aspect marker, root transitive verbs take verbalizing suffixes, also referred to
as status suffixes, which are fully harmonic with the root vowel, as in (12). These verbs are CVC
in phonological form, the minimum word length in Ch’ol.
(12) a. Ta’ k-päk’-ä ixim.
PFV A1-plant-TV corn
‘I planted corn.’
b. Ta’ a-k’el-e-yoñ.
PFV A2-see-TV-B1
‘You saw me.’
c. Ta’ i-boñ-o otyoty.
PFV A3-paint-TV house
‘She painted a house.’
22
In nonperfective aspects, there is either no suffix, such as in (13a), or the dependent clause
suffix -e’ (DEP) may optionally appear with third person objects as in (13b) and (13c).
(13) a. Mi a-k’el-oñ.
IPFV A2-see-B1
‘You see me.’
b. Woli k-päk’(-e’) jiñi ixim.
PROG A1-grow-DEP DET corn
‘I am planting the corn.’
c. Woli i-boñ(-e’) otyoty.
PROG A3-paint-DEP house
‘She is painting a house.’
Non-root transitive verbs
Root transitive verbs contrast with derived or non-root transitive verbs. Derived transitive verbs
are generally not CVC in shape and appear with a derived transitive suffix -V or -Vñ. Unlike
root transitives, the derived transitive verbalizer is not a harmonic vowel, as seen in the perfective
forms in (14).8
(14) a. Ta’ k-pejk-a juñ.
PFV A1-read-DTV paper
‘I read a document.’
b. Ta’ i-ts’ijb-u tsol-ts’ijb.
PFV A3-write-DTV align-letter
‘He wrote the alphabet.’
c. Ta’ i-mul-a a-bujk.
PFV A3-like-DTV A2-shirt
‘She liked your shirt.’
d. Ta’ j-k’uxb-i-yety.
PFV A1-love-DTV-B2
‘I loved you.’
With nonperfective aspects, these derived transitives appear with the derived nominal suffix
-Vñ, as in (15). The -Vñ suffix can be decomposed further into -V, the derived transitive vowel,
and -ñ, a nominalizer.
8In examples such as (14a) and (14c) the derived transitive suffix is realized as -ä in the Tila dialect.
23
(15) a. Woli k-pejk-añ juñ.
PROG A1-read-D.NML paper
‘I read a document.’
b. Mi i-ts’ijb-uñ tsol-ts’ijb.
IPFV A3-write-D.NML align-letter
‘He wrote the alphabet.’
c. Woli i-mul-añ a-bujk.
PROG A3-like-D.NML A2-shirt
‘She likes your shirt.’
d. Mi j-k’uxb-iñ-ety.
IPFV A1-love-D.NML-B2
‘I love you.’
Unergative roots
Unergative structures are expressed with the light verb cha’l and a nominalized unergative root.
Sometimes the verb bajb- ‘hit/strike’ is also used as a light verb with unergative roots. Unergative
roots such as soñ ‘dance’ and k’ay ‘sing’ are nominal in nature and appear in the object position
of the light verb cha’l in in (16), with the result being that these constructions are structurally
transitive.9
(16) a. Ta’ k-cha’l-e k’ay.
PFV A1-LV-DTV song
‘I sang.’
b. Ta’ i-cha’l-e soñ x-Wañ.
PFV A3-LV-DTV dance NC-Juan
‘Juan danced.’
9While the light verb can be passivized in certain cases, it is unclear if the form in (16b) can be pas-
sivized, in the sense described for certain impersonal passives in, e.g., Dutch (Perlmutter, 1978). Forms
such as (1) are possible with the meaning that ‘there was dancing’ where soñ ‘dance’ is the subject of the
intransitive verb ujty- ‘finish’.
(1) Ta’ ujty-i soñ.
PFV finish-IV dance:NML
‘There was dancing here.’
Passive morphology may appear on the light verb cha’l such as in the construction in (2a). The con-
struction in (2a) is common when the object of the PP is borrowed from Spanish such as publicar ‘publish’.
Compare the passive construction in (2a) to the active in (2b). In both cases publicar appears as the comple-
ment to the preposition.
(2) a. Ta’ cha’l-eñ-ty-i tyi publicar. b. Ta’ i-cha’l-e tyi publicar.
PFV LV-DTV-PSV-IV PREP SP:publish PFV A3-LV-DTV PREP SP:publish
‘It was published.’ ‘She published it.’
24
Some verbal roots are ambiguous and may surface in unergative constructions as well as with
intransitive morphology. One such form is the root wäy ‘sleep’ in (17). When wäy appears in
unergative constructions, the nominalizer -el is suffixed to it, cf. the intransitive construction in
(17b) where wäy appears with the intransitive verbalizer -i.
(17) a. Ta’ k-cha’l-e wäy-el.
PFV A1-LV-DTV sleep-NML
‘I slept (on purpose).’
b. Ta’ wäy-i-yoñ.
PFV sleep-IV-B1
‘I slept.’
Compare the unergative construction in (18) to the transitive construction in (19).
(18) Unergative structure
a. Ta’ k-cha’l-e [NP soñ ].
PFV A1-LV-TV dance
‘I danced.’
b. SSP
SS VoiceP
-e
-DTV
DP Voice’
pro Voice VP
1PRON
V NP
cha’l
LV soñ
dance
25
(19) Transitive verb construction
a. Ta’ j-kuch-u [NP si’ ].
PFV A1-carry-TV wood
‘I carried wood.’
b. SSP
SS VoiceP
-u
-TV
DP Voice’
pro
Voice VP
1PRON
V NP
kuch
carry si’
wood
I discuss other constructions with the light verb cha’l in more detail in Chapter 3, Section 3.4.
Positionals
A third class of roots specific to Mayan languages are positional roots. These are a large and
distinct class of roots that convey information about the position or configuration of an object (see
e.g. England 1983, Haviland 1994, and Henderson 2019). Positionals are eventive predicates and
appear with aspect markers such as in (20) and (21). With the perfective aspect, the positional
suffix -li/-le appears, glossed as the positional intransitive verb suffix (POS.IV). With nonperfective
aspects, -tyäl appears, glossed as the positional nominal suffix (POS.NML).
26
(20) Perfective positionals
a. Ta’ buch-le-yoñ tyi kosina.
PFV seated-POS.IV-B1 PFV kitchen
‘I sat in the kitchen.’
b. Ta’ wa’-li ch’ityoñ tyi bij.
PFV stand.on.4.legs-POS.IV boy PREP road
‘The boy stood on the road.’
(21) Nonperfective positions
a. Woli k-ñol-tyäl tyi a’b.
PROG A1-lie.down-POS.NML PREP hammock
‘I am lying in the hammock.’
b. Mi i-buch-tyäl tyi kolem xajlel.
IPFV A3-seated-POS.NML PREP big rock
‘He sits on a big rock.’
See Arcos López (2009) and Vázquez Álvarez (2011) for more discussion of these roots in
Ch’ol.10
Valency changing operations
Passives For root transitives, passives are formed by aspirating the vowel (CVC → CVjC) and
the addition of the intransitive verbalizer, as in (22a) and (22b). If the verb root ends in a fricative
consonant (j, s, x, or [h], [s], [S] in IPA ), the verb is passivized by adding the suffix -le as in (22c).
These passive forms appear with set B person markers. Compare the passives in (22) to their
transitive counterparts in (23).
(22) Perfective passives (23) Perfective transitives
a. Ta’ chojñ-i wakax. a. Ta’ i-choñ-o wakax.
PFV sell.PSV-IV cow PFV A3-sell-TV cow
‘Cows were sold.’ ‘He sold cows.’
b. Ta’ mäjñ-i bu’ul. b. Ta’ i-mäñ-ä bu’ul.
PFV buy.PSV-IV beans PFV A3-buy-TV beans
‘Beans were bought.’ ‘She sold beans.’
c. Ta’ k’ux-le-yoñ. c. Ta’ i-k’ux-u-yoñ.
PFV eat-PSV.IV-B1 PFV A3-eat-TV-B1
‘I was bitten.’ ‘It bit me.’
10Arcos López (2009), for instance, discusses how numeral classifiers are derived from transitive verb
roots and positional roots.
27
With nonperfective aspects in (24), root transitives are passivized by aspirating the root vowel
as well, and adding a nominalizing -el suffix. These forms appear with set A personmarkers. With
nonperfective aspect markers, if the vowel of the root transitive ends in a fricative, it appears with
-tyäl in (24c). Interestingly, the same morphology found on passive roots that end in fricatives is
found on the positionals from above. See Coon and Preminger (2011) for discussion. Compare the
passive form in (24) to their active forms in (25).
(24) Nonperfective passives (25) Nonperfective transitives
a. Woli i-chojñ-el wakax. a. Woli i-choñ wakax.
PFV A3-sell.PSV-NML cow PROG A3-sell cow
‘Cows are being sold.’ ‘He is selling cows.’
b. Mi i-mäjñ-el bu’ul. b. Mi i-mäñ bu’ul.
PFV A3-buy.PSV-NML beans IPFV A3-buy beans
‘Beans are bought.’ ‘She buys beans.’
c. Woli j-k’ux-tyäl. c. Woli i-k’ux-oñ.
PFV A1-eat-PSV.NML PFV A1-eat-B1
‘I am being bitten.’ ‘It is biting me.’
Derived transitive verbs are passivized by adding the suffix -ty to the derived transitive form
in (26) and (28). With the perfective aspect marker, the intransitive verb suffix -i appears, as in (26).
With nonperfective aspect markers, the nominalizing suffix -el appears after the passive marker
-ty. Compare (26) and (28) to their active forms in (27) and (29).
(26) Perfective passives (27) Perfective transitives
a. Ta’ kolty-äñ-ty-i aläl. a. Ta’ i-kolty-a aläl.
PFV help-DTV-D.PSV-IV child PFV A3-help-DTV child
‘The child was helped.’ ‘He helped the child.’
b. Ta’ xujch’-iñ-ty-i tya’k’iñ. b. Ta’ i-xujch’-i tya’k’iñ.
PFV steal-DTV-D.PSV-IV money PFV A3-steal-DTV money
‘Money was stolen.’ ‘She stole money.’
28
(28) Nonperfective passives (29) Nonperfective transitives
a. Mi i-kolty-äñ-ty-el aläl. a. Mi i-kolty-añ aläl.
PFV A3-help-DTV-D.PSV-NML child IPFV A3-help-D.NML child
‘The child is helped.’ ‘He helps the child.’
b. Woli i-xujch’-iñ-ty-el b. Woli i-xujch’-iñ tya’k’iñ.
PFV A3-steal-DTV-D.PSV-NML PROG A3-steal-D.NML money
tya’k’iñ. ‘She is stealing money.’
money
‘Money is being stolen.’
Ditransitives Ditransitive constructions are formed by adding the applicative -b suffix to tran-
sitive verbs as in (30). Indirect objects are marked with set B suffixes. Direct objects must be third
person in ditransitive constructions.
(30) a. Ta’ k-choñ-b-e-yety karu.
PFV A1-sell-APPL-DTV-B2 car
‘I sold you a car.’
b. Ta’ i-ts’ijba-b-e-yoñ i-k’a’ba.
PFV A1-write-APPL-DTV-B1 A3-name
‘He wrote down his name for me.’
The verb ak’ ‘give’ is the only verb that is ditransitive but does not appear with the applicative
suffix -b, as seen in (31). The vowel in ak’ is realized as äwith set A prefixes.11
(31) a. Ta’ k-äk’-e-yety bu’ul.
PFV A1-give-DTV-B2 beans
‘I gave you some beans.’
b. Ta’ aw-äk’-e karu k-chich.
PFV A2-give-DTV car A1-sister
‘You gave my sister a car.’
Ditransitives are passivized with the tyi/tyel suffix as can be seen in (32). Set B suffixes in-
dex the indirect object in the perfective forms and set A prefixes index the indirect object in the
nonperfective forms. Only the indirect object can be passivized in these forms.
11Compare the ditransitive form of ak’ ‘give’ with its transitive form meaning ‘place’ or ‘put’ in (1). The
status suffix of ak’ in its transitive form is ä.
(1) Ta’ k-äk’-ä latyu tyi mesa.
PFV A1-give-TV plate PREP table
‘I placed the plate on the table.’
29
(32) a. Ta’ choñ-b-eñ-ty-i-yety karu.
PFV sell-APPL-DTV-D.PSV-IV-B2 car
‘You were sold a car.’
b. Woli y-äk’-eñ-ty-el karu k-chich.
PROG A3-give-DTV-D.PSV-NML car A1-sister
‘My sister is being given a car.’
Causatives Causatives are realized with the suffix -es,12 which attaches to predicates such as
mich’ ‘angry’ in (33a) or intransitive verb roots such as ju’b ‘descend’ in (33b).
(33) a. Ta’ i-mich’-es-a-yoñ yumäl.
PFV A3-angry-CAU-DTV-B1 leader
‘The president made me angry.’
b. Woli i-ju’b-es-añ pok’ aj-Rosa.
PROG A1-descend-CAU-D.NML bowl NC-Rosa
‘Rosa took down a bowl.’
These constructions may also be passivized with -tyi/tyel, as in (34).
(34) a. Ta’ mich’-es-äñ-ty-i-yoñ.
PFV angry-CAU-DTV-D.PSV-IV-B1
‘I was angered.’
b. Woli i-ju’b-es-äñ-ty-el pok’.
PROG A1-descend-CAU-DTV-D.PSV-NML bowl
‘The bowl is being taken down.’
2.3.4 Alignment
As mentioned above, Ch’ol is a head-marking language with grammatical relations marked on
the verb. Transitive objects and intransitive subjects are indexed by absolutive markers (set B) in
(35). Transitive subjects are marked with set A prefixes, as in (36).
(35) Absolutive markers = set B
a. Ta’ majl-i-yoñ. b. Ta’ i-k’el-e-yoñ aj-Maria.
PFV go-IV-B1 PFV A3-see-TV-B1 NC-Maria
‘I went.’ ‘Maria saw me.’
12In the Tila dialect, the causative suffix is sometimes realized as -is and the derived transitive suffix as
-ä.
30
(36) Ergative markers = set A
Ta’ i-juch’-u sa’ aj-Maria.
PFV A3-grind-TV masa NC-Maria
‘Maria ground masa.’
I assume, following Coon (2017), that a low functional head, labeled as Voice, licenses ergative
case. This is consistent with the inherent ergative approach to case (Aldridge, 2004; Legate, 2008;
Coon et al., 2014). Transitive Voice also licenses case for the absolutive object.
(37) Case licensing in transitive clauses in Ch’ol
VoiceP
DP Voice’
SUBJ Voice VP
ERG
V NP
ABS OBJ
Absolutives in intransitive constructions are licensed by Infl (labeled as I) (Coon, 2017), as
shown in (38).13
(38) Absolutive case licensing for intransitive verbs in Ch’ol
IP
I SSP
SS vP
v VP
V NP
ABS
SUBJ
13Under Legate’s (2008) terms, absolutive is the default case in Ch’ol.
31
2.3.5 Split ergativity
In Ch’ol, intransitive subjects and transitive objects are indexed via set B markers, as in (39) where
both the first person intransitive subject and first person transitive object are indexed by -oñ. With
nonperfective aspects, intransitive subjects are indexed with set A markers, not with set B, as in
(40). In (40), the first person intransitive subject is indexed by k-. This same marker also indexes
ergative subjects, as in (41).
(39) a. Ta’ majl-i-yoñ. b. Ta’ i-k’el-e-yoñ.
PFV go-IV-B1 PFV A1-see-TV-B1
‘I went.’ ‘He saw me.’
(40) a. Mi k-majl-el. b. Woli k-majl-el.
IPFV A1-go-NML PROG A1-go-NML
‘I go.’ ‘I am going.’
(41) Ta’ k-tyuk’-u kajpe’.
PFV A1-cut-TV coffee
‘I harvested coffee.’
At first glance, it seems as if Ch’ol exhibits a split ergative system based on aspect where it
is ergative-absolutive with the perfective aspect marker, but nominative-accusative aligned with
nonperfective aspect markers. This is due to the fact that with perfective aspects, set B markers
mark subjects of intransitive verbs and objects of transitive verbs in (39). However, with nonper-
fective aspects, set A markers index subjects of intransitive verbs in (40), seemingly patterning
with the ergative subject of the transitive verb in (41). However, Coon (2010a, 2012, 2013) argues
that this aspect split is epiphenomenal and the verbs in (40) are nominalized complements of the
predicative nonperfective aspect. The set A markers in (40) are possessive markers. Recall that
ergative and possessive markers are identical: compare (40) with (42).
(42) k-otyoty
A1-house
‘my house’
That is, the set A markers in (40) and (42) are both genitive prefixes, indexed on a nominal-
ized verb. Additional evidence that the verbs in (40) are nominalized is that they appear with a
32
nominalizing suffix -el, and not the intransitive verb suffix -i, which appears in (39a).
The two aspect markers mi and woli are analyzed as predicates that take nominalized com-
plements. The analysis for (40a) is schematized in (43). The imperfective aspect mi in (40a) takes
the nominalized kmajlel ‘my going’ as its complement. The verb root majl has the nominalizing
suffix -el and possessive marker k- on it. The possessive marker k- is co-indexed with a PRO in the
nominalized complement.
(43) Woli [DP ki- [ majl-el PROi ] ].
PROG A1- go-NML
‘I am going.’ Lit. ‘My going is happening.’
The structure of (43) is provided in (44) from Coon (2013).
(44) VP
V DP
woli
PROG
D PossP
DPi Poss’
pro
Poss nP
1PRON
k-majl-el PROi
When possessor subjects are overt in nominalized complements they appear after the verb, as
in (45a). This parallels possessive phrases, where possessors also appear after their possessees, as
shown in (45b) (and discussed more in Section 2.4.5 below).
(45) a. i-majl-el aj-Rosa
A3-go-NML NC-Rosa
‘Rosa’s going’
b. i-wakax aj-Rosa
A3-cow NC-Rosa
‘Rosa’s cow’
33
Transitive verbs with nonperfective aspects are also analyzed as nominalized complements
where the subject is structurally a possessor. For transitive verbs with nonperfective aspects, mor-
phological differences also provide evidence for their nominal status. Compare the transitive verb
in the perfective aspect in (46b) and the transitive verb with the progressive aspect in (46a). In
(46b) the transitive verbalizer -u appears with the verb juch’ ‘grind’. In (46a) the suffix is null or
the dependent clause suffix -e’ can optionally appear with third person objects.
(46) a. Woli i-juch’(-e’) sa’ aj-Rosa.
PROG A3-grind-DEP masa NC-Rosa
‘Rosa is grinding masa.’
b. Ta’ i-juch’-u sa’ aj-Rosa.
PFV A3-grind-TV masa NC-Rosa
‘Rosa ground masa.’
The structure of (46a) according to Coon (2010a, 2012, 2013) is given in (47). The subject, Rosa,
is expressed as the possessor in Spec,PossP and is co-indexed with a PRO in the nominalized com-
plement (see also Coon and Royer (To Appear)). The verbal stem raises to n, where set A marking
appears. The set A agreement morpheme is analyzed as the result of an agreement relation be-
tween Poss and the possessor in its specifier. The φ features are then inherited by n, triggering the
spell out of set A markers on the nominalized stem. To achieve the correct word order, I argue in
Chapter 5 that the possessor is on the right.
(47) DP
D PossP
DPi Poss’
aj-Rosa Poss nP
n VoiceP
PROi Voice’
Voice VP
e’
DEP V NP
juch’
grind sa’
masa
34
The aspect marker woli takes (47) as its complement, as shown in (48).
(48) VP
V DP
woli
PROG
i-juch’(-e’) sa’ aj-Rosa
A3-grind masa NC-Rosa
Further evidence that the complements of nonperfective aspects are nominal comes from data
such as in (49a). In the expression for ‘it is raining’, the noun ‘rain’ appears as the complement
to the progressive aspect maker woli in (49a). This nominal may not appear with the perfective
aspect marker as in (49b).
(49) a. Woli [NP ja’al ].
PROG rain
‘It is raining.’
b. * Ta’ [NP ja’al ]
PFV rain
Intended: ‘It rained.’
In sum, the seemingly split ergative pattern is in fact epiphenomenal. The aspectual-based
split of agreement comes down to the fact that progressive markers are verbal predicates that take
nominalized complements. With the perfective aspect marker, the verb appears with a verbal
suffix. Structurally, however, the perfective aspect is a ‘true’ aspect marker in an inflectional head,
whereas nonperfective aspects are predicates that take nominal complements.
2.3.6 Clausal word order
Ch’ol is verb-initial with alternating VOS/VSO word order (Vázquez Álvarez, 2002, 2011; Coon,
2010b; Clemens and Coon, 2018). This alternation is connected to structural properties of the ob-
ject. VOS word order is associated with NP objects, and VSO with DP objects. Bare object NPs
such as waj ‘tortilla’ in (50a) must appear next to the verb, generating the pragmatically neutral
VOS word order. VSO word order is ungrammatical with NP objects, as shown by (50b). Other
35
modifiers, such as numerals, may also appear with objects in VOS position. I use the term NP
descriptively to refer to nominals without any overt elements in or above D.
(50) a. Ta’ i-k’ux-u [O waj ] [S aj-Maria ].
PFV A3-eat-TV tortilla NC-Maria.
‘Maria ate a tortilla.’
b. * Ta’ i-k’ux-u [S aj-Maria ] [O waj ] .
PFV A3-eat-TV NC-Maria tortilla.
Intended: ‘Maria ate a tortilla.’
DP objects are banned from appearing in VOS object position. I use DP as a descriptive term in
this dissertation, meaning that they are nominals that contain overt material in or above D.14,15 In
Tumbalá Ch’ol, VSO word order with a full DP object was judged felicitous in a context in which
the speaker is looking at a dead deer and says (51a). DP objects are judged to be ungrammatical
in VOS object position, as in (51b).
(51) a. Ta’ i-jul-u [S aj-Ariañ ] [O jiñi me’ ].
PFV A3-shoot-TV NC-Adrian DET deer
‘Adrian shot the deer.’
b. * Ta’ i-jul-u [O jiñi me’ ] [S aj-Ariañ ].
PFV A3-shoot-TV DET deer NC-Adrian
Intended: ‘Adrian shot the deer.’
The word order patterns in (50) and (51) are discussed more in Chapter 5 of the dissertation.
Other predicates such as adjectives or nouns also appear before any overt subjects as in (52a).
Predicates appear with set B markers, as in (52b) and (52c). Data such as in (52) lead to the charac-
terization of Ch’ol as generally a predicate-initial language.
(52) a. Säsäk ili waj.
white this tortilla
‘This tortilla is white.’
b. Pek’-ety.
short-B2
‘You are short.’
14While VSO word order is marginal (indeed, two full postverbal nominals across Mayan languages are
uncommon (England, 1991)), VSO sentences are still judged felicitous and are spontaneously produced, as
also Coon (2010b, 362) notes for the Tila dialect of Ch’ol.
15See discussion in Chapter 5, Section 5.8, for some outstanding issues regarding nominal syntax and
Mayan languages.
36
c. X-k’el juñ-oñ.
NC-see book-B1
‘I am a student.’
2.3.7 Negation
The negation marker ma’añ precedes aspectual markers, as in (53). The negation marker mach ap-
pears with stative predicates as in (54). In the Tumbalá dialect, the irrealis enclitc =ik also appears
cliticized to the negation marker or predicate. The negation markerma’añ is most likely composed
of the negative markermach and the existential predicate añ; see Chapter 3, Section 3.3.3 and Coon
(2006) for discussion.
(53) Verbal predicate negation
a. Ma’añ=ik ta’ k-tyaj-a.
NEG=IRR PFV A1-find-TV
‘I didn’t find it.’
b. Ma’añ=ik mi i-mul-añ.
NEG=IRR IPFV A3-like-D.NML
‘She doesn’t like it.
(54) Stative predicate negation
a. Mach k-om=ik bu’ul.
NEG A1-want=IRR beans
‘I don’t want any beans.’
b. Mach loktor-oñ=ik.
NEG doctor-B1=IRR
‘I am not a doctor.’
2.3.8 The preverbal position
The preverbal position in Ch’ol, as in many other Mayan languages, is where focused,wh-marked
or topicalized constituents appear, as shown in (55). Topicalized constituents often appear with
the =i enclitic, as in (55c). In many cases, the particle a precedes the topicalized constituent as in
(61), but it is not obligatory.
37
(55) a. Wh-word
[wh Chuki ]i ta’ i-kuch-u ti ch’ityoñ?
what PFV A3-carry-TV boy
‘What did the boy carry?’
b. Focus
[FOC Lakña’-ob ]i mi i-majl-el i-tyuk’-b-eñ-ob i-wuty ti.
women-PL IPFV A3-go-NML A3-pick-APPL-NML-PL A3-fruit
‘The women go and pick its fruit.’ Don Lucio
c. Topic
[TOP Jiñi muty=i ] mi i-mel iy-otyoty.
DET chicken=ENCL IPFV A3-make A3-house
‘As for the bird, it makes its house.’ Don Lucio
As Ch’ol is an obligatory wh-fronting language, interrogative words such as majki ‘who’ and
chuki ‘what’ must appear preverbally in (56a) and (57a).16 In situ wh-words are ungrammatical, as
per (56b) and (57b).
(56) a. Chukii ta’ i-k’ux-u ti aj-Maria?
what PFV A3-eat-TV NC-Maria
‘What did Maria eat?’
b. * Ta’ i-k’ux-u chuki aj-Maria?
PFV A3-eat-TV what NC-Maria
Intended: ‘What did Maria eat?’
(57) a. Majkii ta’ i-k’ux-u waj ti?
who PFV A3-eat-TV tortilla
‘Who ate a tortilla?’
b. * Ta’ i-k’ux-u waj majki?
PFV A3-eat-TV tortilla who
Intended: ‘Who ate a tortilla?’
The data in (56) and (57) also display that Ch’ol is not syntactically ergative. That is, Ch’ol does
not restrict the A-bar extraction of ergative arguments. Further exemplifying this point is that,
without context, the sentence in (58) is ambiguous between extraction of the ergative subject or
extraction of the absolutive object.
(58) Majki ta’ i-kel-e?
who PFV A3-see-TV
‘Who saw her?’ or ‘Who did she see?’
16Question words could be further broken up into an indefinite pronoun majch ‘someone’ and chu
‘something’ and the question enclitic =ki, though I do not make this distinction in the glosses. See
Vázquez Álvarez and Coon (To Appear).
38
Multiplewh-questions are not permitted in Ch’ol, as displayed in the ungrammatical examples
in (59).
(59) a. * Majki j ta’ i-k’ux-u chuki t j?
who PFV A3-eat-TV what
Intended: ‘Who ate what?’
b. * Majki j chukii ta’ i-k’ux-u ti t j ?
who what PFV A3-eat-TV
Intended: ‘Who ate what?’
Focused-marked constituents in Ch’ol may also appear in the clause-initial position
(Vázquez Álvarez, 2011), as shown in (60), though in situ focus has also been shown to be pos-
sible in Ch’ol (Clemens et al., 2017).
(60) a. Aj-Mariai ta’ i-juch’-u sa’ ti.
NC-Maria PFV A3-grind-TV masa
‘Maria ground masa.’
b. Sa’i ta’ i-juch’-u ti aj-Maria.
masa PFV A3-grind-TV NC-Maria
‘Maria ground masa.’
Topicalized and focused constituents may co-occur, where the topicalized constituent must
precede the focused constituent. An example is given in (61) from the Tila dialect of Ch’ol.
(61) [TOP a li aj-Wañ=i ] [FOC ixim ] tyi i-mäñ-ä.
TOP DET NC-Juan=ENCL corn PFV A3-buy-TV
‘As for Juan, it is corn that he bought.’ (Vázquez Álvarez, 2011, 342) (with my bracketing)
Depictive secondary predicates may also appear in the preverbal position as in (62). See
Vázquez Álvarez (2010) for more discussion.
(62) a. Buch-ul ta’ majl-i-yoñ tyi Saltu.
seated-STAT PFV go-IV-B1 PREP Salto
‘I went to Salto seated.’
b. Loñ-li-li-ña tyal la=k-tyaty.
in.vain-RED-walk-ATTR come PART.PL=1-father
‘The man was coming towards us walking.’ Bats’
39
2.3.9 Summary
In sum, Ch’ol is a predicate-initial, head-marking ergative-absolutive language. I assume that set
A (ergative/genitive) person prefixes are the spell out of an agreement relation between a low
functional head and a nominal in its specifier. Set B markers are analyzed as clitics, derived via
Agree. Ergative case is licensed in Ch’ol by a low functional head (in this dissertation labeled as
Voice). Absolutive objects are licensed by a low functional projection whereas absolutive subjects
are licensed by INFL, throughout this dissertation labeled as I. Ch’ol constituent order within the
clausal domain is summarized in (63).
(63) Topic/Focus/Wh- Neg TAM Predicate
The morphology of transitive and intransitive verbs are summarized in Table 2.8 for perfective
and nonperfective aspects. With the perfective aspect marker, root transitives take a vowel suf-
fix harmonic with the vowel of the root. Nonperfective aspects take nominalized complements.
Root transitives can appear with a dependent suffix -e’ with third person objects. The intransitive
verbalizer with the perfective aspect is always -i; in nonperfective aspects -el surfaces. There is
no overt nominalizing morpheme for root transitives; for nonroot transitives this morpheme is
spelled out as -ñ.
Table 2.8: Ch’ol stem moprhology (Coon, 2013, 28)
Perfective Nonperfective
Root transitive A-root-V-B A-root-(e’)-B
Nonroot transitive A-root-V-B A-root-Vñ-B
Intransitive root-i-B A-root-el
Positional root-li-B root-tyäl-B
Following Coon (2010a, 2012, 2013), I take the apparent aspect split in Ch’ol to be epiphenom-
enal. Perfective aspects take ‘true’ verbs whereas nonperfective aspect markers are predicates that
take nominalized complements.
40
2.4 Ch’ol nominal basics
In this section, I cover background on the nominal domain in Ch’ol. While fully extended nominal
expressions are rare in casual speech, they are possible in elicited contexts. For instance, in (64),
determiners, numerals, and adjectives are prenominal while the (noninterrogative) possessor is
postnominal.
(64) jiñi cha’-kojty i-säsäk wakax aj-Rosa
DET two-CLF A3-white cow NC-Rosa
‘the two white cows of Rosa’s’
I discuss each one of these elements below.
2.4.1 Nouns
Nouns are not marked for case, rather agreement markers appear on the verb, as discussed above.
Bare nouns can serve as arguments and may be interpreted as definite or indefinite, as in (65).
(65) a. Ta’ juli wiñik.
PFV arrive-IV man
‘A/the man arrived.’
b. Ta’ i-boñ-o otyoty ch’ityoñ.
PFV A3-paint-TV house boy
‘The boy painted a/the house.’
I further discuss the interpretation of bare nouns with respect to (in)definiteness in Chapter 3.
Nouns in Ch’ol are furthermore unspecified for number—that is, they have general num-
ber in the sense of Corbett (2000), or are number neutral in the sense of Chierchia (1998a) and
Rullmann and You (2006). This is supported by data in (66), where the bare noun ts’i’ ‘dog’ is not
specified for number.
41
(66) Tsa’ j-k’el-e ts’i’.
PFV A1-see-TV dog
‘I saw a dog/dogs.’ Speaker comment: We do not know how many dogs you saw. Little
(2018b)
See discussion in Chapter 3, Section 3.3.7 for more on plurality in Ch’ol.
2.4.2 The noun class prefixes x- and aj- in Ch’ol
There are two noun class prefixes in Ch’ol, x- and aj-, which appear with certain nouns and proper
names. Historically aj- is related to the Proto-Mayan masculine prefix, and x- the Proto-Mayan
feminine prefix. In Ch’ol, however, their uses do not correspond to gender. In the Tumbalá dialect
x- is used to derive demonyms for places as in (67a) and (67b) and can attach to N-V compounds
or nominalized antipassives as in (67c) and (67d) to mean ‘someone who does X’.17
(67) a. x-Palenke c. x-choñ waj
NC-Palenque NC-sell tortilla
‘someone from Palenque’ ‘tortilla seller’
b. x-tyejklum d. x-mäñ-oñ-el
NC-city NC-buy-AP-NML
‘someone from the city’ ‘shopper’
The prefixes x- and aj- can also attach to names, as in (68). According to Arcos López (2009,
Chapter 3) both prefixes can codify respect or friendship of the speaker towards the referent. In
certain contexts, aj- can sometimes be used to indicate disagreement between participants and x-
can be used to make fun of someone when prefixed to a name.
(68) a. x-Wañ b. aj-Rosa
NC-Juan NC-Rosa
‘Juan’ ‘Rosa’
17In the Tila dialect of Ch’ol, the prefix aj is used to derive the meanings in (67), e.g., aj-Palenke ‘someone
from Palenque’, aj-choñ waj ‘tortilla seller’, etc.
42
There are also dialectal differences: INALI (2011) indicates that x- is used more often with
names in the Tumbalá dialect, whereas Tila speakers use aj-.
These noun class prefixes may appear on names in argument position (69a) as well as in voca-
tives (69).
(69) a. Ta’=ix majl-i x-Wañ.
PFV=already go-IV NC-Juan
‘Juan left already.’
b. X-papito, la’ jap-ä a-kajpe’
NC-papito come drink-PFV A2-coffee
‘Papito18, come drink your coffee.’ (Arcos López, 2009, 125)
2.4.3 Plural markers
The plural suffix -ob can optionally attach to human-denoting nouns and some animal-denoting
nouns, as in (70). This plural marker also can appear on verbs, as discussed above in Section 2.3.2.
I analyze the plural marker as adjoined to the nominal, as per Wiltschko (2008) for Halkomelem
Salish and Butler (2011) for Yucatec Maya, though further investigation is needed in order to de-
scribe its precise distribution.
(70) a. x-ch’ok-ob b. k-erañ-ob c. wiñik-ob
NC-girl-PL A1-sibling-PL man-PL
‘girls’ ‘my siblings ‘men’
This plural marker can also attach to proper names to derive an associative plural meaning, as
in (71).
(71) aj-Celiaj-ob
NC-Celia-PL
‘Celia and her friends’19
18Term of endearment.
19This can also mean a group of women all named Celia.
43
The plural suffix -tyak can also attach to nouns and adjectives as in (72) and (73), with the
meaning of ‘various’.20
(72) a. ixim-tyak b. otyoty-tyak
corn-PLIN house-PLIN
‘various types of corn’ ‘various houses’
(73) Context: After seeing pictures of different pairs of shoes
Uts’aty-tyak=jax.
nice-PLIN-EXCL
‘They are very nice.’ (WhatsApp, May 25th, 2019)
The plural markers -ob and -tyak may also co-occur, as in the excerpt from a pamphlet in (74).
(74) Jiñ mi i-tyaj i-kolty-äñ-tyel pejtyelel-ob ch’ityoñ x-ch’ok-tyak-ob k’äläl
DET IPFV A3-receive A3-help-DTV-PSV.NML all-PL boy NC-girl-PLIN-PL until
tyi 22 años i-ja’bil-el-ob.
PREP 22 SP:years A3-year-RS-PL
‘All boys and girls up to 22 years of age will receive this support.’21
2.4.4 Demonstratives and determiners
Ch’ol demonstratives and determiners are listed in Table 2.9. The Tila dialect has an additional
determiner li, not used by Tumbalá speakers (Vázquez Álvarez, 2011).22
20The plural marker tyakmay also attach to verbs as in (1).
(1) Jiñ=jach k-poj-k’ajty-iñ-tyak=loñ.
DET=EXCL A1-HON-ask-D.NML-PLIN=PL.EXCL
‘This is all we ask.’ (Vázquez Álvarez, 2011, 80)
21Guı́a cha’añ programas sociales [Guide for social programs], 2014, Ministry of Social Develop-
ment https://www.gob.mx/cms/uploads/attachment/file/30800/GPS_2014_Chol_del_
Noroeste.pdf
22Little and Vázquez Martı́nez (2018) note that Tumbalá speakers sometimes use a particle a before
nouns. They suggest that a possibly comes from a shortened a jiñi, the topic marker a plus the determiner
jiñi. More investigation is needed, however, to figure out exactly what the particle a is in the Tumbalá
dialect. Arcos López (2009) glosses a as a filler word.
44
Table 2.9: Demonstratives and determiners in Ch’ol
ili ‘this’
jiñ(i) ‘that’
ixä ‘that (distal)’
The demonstratives ili and ixä encode deixis with respect to the speaker/hearer; jiñi, on the
other hand, is often used in anaphoric contexts (Little and Vázquez Martı́nez, 2018).23 Bare nouns
may also be interpreted as definite: a determiner is not needed for a definite interpretation, as
discussed further in Chapter 3. VOS objects in Ch’ol may not appear with the determiners or
demonstratives in Table 2.9.
2.4.5 Possessive constructions
In the nominal domain, set A person markers, bolded in (75), are prefixed to the possessed noun
and overt possessors follow possessees. The possessive structure I adopt is given in (76).
(75) a. k-waj b. i-waj aj-Maria
A1-tortilla A3-tortilla NC-Maria
‘my tortilla’ ‘Maria’s tortilla’
23The determiner jiñi is also used in focus contexts and sometimes glossed as a focus marker
(Vázquez Álvarez, 2011). Additionally, it appears in copular constructions along with the affirmative clitic
=äch as in (1).
(1) K-tema=loj-oñ jiñ=äch i-ñusa-k’iñ x-mayaj-ob y-ik’oty i-ty’añ.
A1-topic-PART.PL-B1 DET=AFF A3-pass-day NC-maya A3-COM A3-language.
‘Our topic was the history of the Mayan people and their language.’
http://chol.lingspace.org/ctu/tallertyak-tyi-tumbala-yikoty-san-miguel/
Zavala (2017) notes a similar pattern of grammaticalization in Olutec (Mixe-Zoquean), where demonstra-
tives become pronouns, copula morphemes, and focus markers. See also Li and Thompson (1977) for dis-
cussion on anaphoric pronouns becoming copula morphemes.
45
(76) a. i-wakax aj-Rosa
A3-cow NC-Rosa
‘Rosa’s cow’
b. DP
D PossP
Poss’ DPk
Poss nP aj-Rosa
NC-Rosa
ik-wakax
A3-cow
I take the set A prefix to be the spell out of an agreement relation between the possessor and a
low functional head, Poss. The φ features on Poss are inherited by n, triggering the set A agreement
morpheme on nP. The agreement morpheme is spelled out on nP, but below any adjuncts. That is,
it appears on adjectives but not on numerals, discussed further in Section 2.4.6.
Similar to ergative case assignment above, I adopt the view that genitive case is assigned by
Poss to the nominal in its specifier, schematized in (77).
(77) Genitive case assignment
DP
D PossP
Poss’ DP
GEN
Poss nP Possessor
n NP
possessee
When the possessor is interrogative, it precedes the possessee, generating the possessor–
possessee order in (78).
46
(78) a. majkii i-wakax ti?
who A3-cow
‘whose cow?’
b. DP
DPk D’
majki
who D PossP
Poss’ tk
Poss nP
ik-wakax
A3-cow
As will be discussed further in Chapter 4, overt possessors are banned from appearing in VOS
object position.
2.4.6 Modifiers
Modifiers, such as adjectives, numerals, and quantifiers, appear before the noun as in (79). Nu-
merals in Ch’ol appear with an obligatory numeral classifier.24
(79) a. cha’-kojty wakax b. säsäk wakax c. ka’bäl wakax
two-CLF cow white cow many cow
‘two cows’ ‘a white cow’ ‘many cows’
Possessive markers attach to adjectives as in (80a), but not numerals or other quantifiers, as in
(80b).25
24See work by Arcos López (2009); Bale and Coon (2014); Bale et al. (2019); Little et al. (2020) for more on
numerals and classifiers in Ch’ol.
25The meaning of (80b) does not have the maximality entailments as English does in the phrase ‘my two
white cows.’ To indicate maximality, a possessed numeral is used as in (1), discussed in detail in Little
(2018b).
(1) tyi i-cha’-kojty-lel k-säsäk wakax
TYI A3-two-CLF-RS A1-white cow
‘my two white cows’
47
(80) a. k-säsäk wakax b. cha’-kojty k-säsäk wakax
A1-white cow two-CLF A1-white cow
‘my white cow’ ‘two of my white cows’
The structure of the nominal phrase is given in (81). I follow Bale et al. (2019) in labeling the
phrase containing a numeral and its obligatory classifier as MP, adjoined to the nP. Possessive
marking is the result of agreement between the possessor in Spec,PossP and Poss, the same head
that assigns the possessor genitive case. Agreement is spelled out on the nP, excluding adjuncts
such as numerals.26
(81) DP
D PossP
Poss’ DP
possessor
Poss nP
MP n’
num CLF
n NP
AP NP
adjective noun
2.4.7 Relative clauses and =bä
Relative clauses usually follow the head noun and are marked with the clitic =bä (a borrowing
from the Mixe-Zoque language, Zoque (Vázquez Álvarez, 2011)). The relative clause marker =bä
is a second position clitic appearing on the first element of the relative clause. The morpheme =bä
is bolded in the examples in (82).
26The specifier of PossP and nP have both been proposed to be the position where possessors are base-
generated (see discussion in Alexiadou et al. (2007)). An alternative approach is that the possessor occupies
Spec,nP and numerals are adjoined to a specifier position above the possessor. However, under this type
of account, it would be unclear how to capture structural differences with respect to objects in VOS posi-
tion. As will be discussed further in Chapter 4, overt possessors in object position trigger object shift, but
numerals do not. If numerals occupied a position above possessors, an explanation would be needed as to
why possessors trigger object shift, but numerals (structurally higher under this approach) do not.
48
(82) a. Ta’ j-k’el-e wiñiki [ ta’=bä i-choñ-b-e-yoñ bu’ul ti ]
PFV A1-see-TV man PFV=REL A3-sell-APPL-B1 beans
‘I saw the man who sold me beans.’
b. Ta’ wejl-i majlel x-ch’e’i [ ta’=bä i-chuk-u ti ch’ityoñ ].
PFV fly-IV DIR:away NC-bird PFV=REL A3-catch-TV boy
‘The bird that the boy caught flew away.’
Modifiers may also appear with the relative clause enclitic =bä. Adjectives without =bä are
obligatorily prenominal as in (83a), whereas modifiers with =bä may appear before or after the
nominal, as in (83b) and (83c). Unlike in (83a), possessivemorphology cannot appear onmodifiers
with =bä, displayed in the ungrammatical example (83d).
(83) a. i-säsäk wakax c. säsäk=bä i-wakax
A3-white cow white=REL A3-cow
‘her white cow’ ‘her white cow’
b. i-wakax säsäk=bä d. * i-säsäk=bä wakax
A3-cow white=REL A3-white=REL cow
‘her white cow’ intended: ‘her white cow’
With some modifiers, especially derived modifiers, =bä is obligatory, as in (84) (see also dis-
cussion in Coon (2018)).
(84) a. wel-el*(=bä) tye’ b. tye’ wel-el*(=bä)
flat-STAT=REL wood wood flat-STAT=REL
‘flat wood’ (Martı́nez Cruz, 2007) ‘flat wood’
As mentioned above, these modifiers with =bä may appear either prenominally or postnomi-
nally, schematized in (85) (see also Bale et al. (2019, 23)).
(85) nP
{modifier=bä} nP {modifier=bä}
Noun
49
2.4.8 Summary
In sum, the order of constituents within the nominal domain is given in (86).
(86) Nominal constituent order
Dem/Det Numeral set A- Adjective Noun Possessor
Demonstratives, determiners, numerals, and adjectives are all prenominal and possessors are
postnominal. Interrogative possessors are prenominal. The genitive set A marker is spelled out on
the nP (including any adjectives). As briefly discussed in this section, demonstratives, determin-
ers, and overt possessors are banned from appearing in VOS object position.
2.5 Conclusion and a brief overview of verb-initial word order
I have given an overview of Ch’ol morphosyntax, which I will reference throughout this disserta-
tion. Additionally, throughout this chapter, I have highlighted some empirical points of Ch’ol mor-
phosyntax that form the basis for the analyses of the following chapters. In Chapter 3, I provide
data on the interpretation of bare nouns and a semantic analysis of how the syntax reflects seman-
tic interpretation. In Chapter 4, I argue that movement-derived islands account for subextraction
asymmetries. Chapter 5 ties together the conclusions from Chapter 3 and 4 to argue for a base-
generated account of verb-initial syntax. Three possibilities for verb-initial syntax are summarized
in (87). The base-generated account in (87a) has been pursued for Mayan languages in England
(1991), Aissen (1992) and Otaki et al. (2019), as well as for Austronesian languages (Chung, 1998,
2006) and the Wakashan language Nuu-chah-nulth (Wojdak, 2007). A phrasal-fronting account of
VOS word order in (87b) has been posited for Mayan (Coon, 2010b) and Austronesian languages
(Aldridge, 2002; Chung, 2006; Massam, 2000); while head movement of the verb above the subject
has been posited for Afro-Asiatic languages (Fassi Fehri, 1993; Ouhalla, 1994) and Celtic languages
(Sproat, 1985; McCloskey, 1991). Recent work by Clemens (2014, 2019) and Clemens and Coon
(2018) adopt a head movement account and posit an additional prosodic constraint that reorders
an NP object with the verb that selects it, generating VOS in (87c). Clemens and Polinsky (2017)
50
provide a recent overview of verb-initial languages with a special focus on Mayan and Austrone-
sian.27
(87) VOS word order
a. Base-generated (proposal)
IP
I SSP
SS VoiceP
Voice’ DP
Voice VP SUBJ
V NP
OBJ
b. Phrasal-fronting (Coon, 2010b)
TP
vP j T’
v VP
T VoiceP
V NP
DP Voice’
OBJ
SUBJ Voice t j
c. Syntactic input and prosodic output of VOS order with NP objects (Clemens and Coon,
2018, 258)
IP ! ι
I SSP ϕ ϕ
TAM
SUBJ
ř ř DP
V+Voice+SS VoiceP TAM=V-Voice-SS OBJNP
SUBJDP
tVoice VP
tV OBJNP
The relationship between the underlying syntax of each of these accounts differs. For a base-
27I am faithful to the labels of the original analyses. For instance, what I label IP, Coon (2010b) labels TP.
51
generated account, the underlyingword order is VOS, where subjects, in the specifier of VoiceP, are
linearized to the right. Phrasal-fronting and head movement accounts begin by base-generating
SVO. For phrasal-fronting accounts, such as in Coon (2010b), VOS surfaces via the fronting of a
phrasal projection containing both the verb and object. Head movement of the verb generates
VSO in (87c), and VOS is derived via certain prosodic constraints.
Each account therefore differs in how VSO word order is derived. For a base-generated ac-
count, VSO is derived from VOS by moving the object to a position above the subject in (88a). For
phrasal-fronting accounts, the object moves out of the VP, to a position below the subject, in (88),
before the vP moves to Spec,TP, as in (88b). Under a prosodic account, after headmovement of the
verb, VSOword order is generated. Prosodic reordering constraints do not apply to the DP object,
as the DP is considered to be a phase and already sent to spell out.
(88) VSO
a. Base-generated (proposal)
IP
I SSP
SS VoiceP
VoiceP DPi
OBJ
Voice’ DP
Voice VP SUBJ
V ti
52
b. Phrasal-fronting (Coon, 2010b)
TP
vP j T’
v VP
T VoiceP
V ti
DP Voice’
SUBJ
Voice AbsP
DPi Abs’
OBJ Abs t j
c. Syntactic input and prosodic output of VSO order (Clemens and Coon, 2018)
IP ! ι
I SSP ř ϕ
TAM=V-SS
TAM
ϕ ϕ
V+Voice+SS VoiceP SUBJDP OBJDP
SUBJDP
tVoice VP
tV OBJDP
The rest of this dissertation investigates the syntactic and semantic properties of nominal argu-
ments and their implications for clausal syntax. The properties exhibited of Ch’ol, such as struc-
turally reflected (in)definiteness and movement-derived islands, are not unique to Ch’ol. What
is particular to Ch’ol, I argue, is how the constituents are linearized with respect to each other,
most straightforwardly captured by base-generating VOS. I conclude that while structural depen-
dencies across language pattern similarly, (i) the way in which they are underlyingly linearized in
the syntax is a point of crosslinguistic variation; and (ii) verb-initial languages do not comprise a
uniform syntactic class.
53
CHAPTER 3
STRUCTURALLYREFLECTED (IN)DEFINITENESS
3.1 Introduction
In a language such as English, overt morphemes like ‘the’ or ‘a’ signal a definite or indefinite inter-
pretation. Not all languages, however, have overt markers of (in)definiteness. A question arises
as to how definiteness is expressed in a language lacking obligatory articles where, depending
on the context, bare nouns can be interpreted as definite or indefinite. Stemming from Carlson
(1977), much work has investigated the interpretations of bare nouns and how their possible in-
terpretations connect to their structural position. Many authors have shown that in languages that
allow bare nouns to be arguments, their possible interpretations are constrained. That is, while
definite and kind-level interpretations are freely available, indefinite interpretations arise in a set
of restricted environments (e.g., Dayal (2004) for Hindi and Geist (2010) for Russian and recently
Deal and Nee (2017) for Teotitlán del Valle Zapotec).
In this chapter, I investigate the relationship between the structural position of bare nominal
arguments in Ch’ol and their availability to be interpreted as indefinite. I propose a structurally
motivated account of indefiniteness, based on Chung and Ladusaw’s (2004) Restrict operator as
well as confining existential closure to the VP domain, as per Diesing (1992). This is based on the
notion that in Ch’ol, indefinite arguments are instances of pseudo-incorporation, in the sense of
Massam (2000, 2001). My analysis provides further support for crosslinguistic generalizations on
the constraints of semantic interpretation.
As seen in (89) and (90), bare nouns can be arguments of verbs in Ch’ol and can be definite or
indefinite. Both a definite interpretation and indefinite interpretation is available for the intran-
sitive subject wiñik ‘man’ in (89) and the bare object noun waj ‘tortilla’ in (90). However, the bare
transitive subject wiñik ‘man’ in (90) can only be definite.
54
(90) Ta’ i-mäñ-ä otyoty wiñik.
(89) Ta’ jul-i wiñik. PFV A3-buy-TV house man
PFT arrive-IV man
(i) ‘The man bought a/the house.’
‘A/the man arrived.’
(ii) NOT: A man bought a/the house.
I propose that these interpretations follow naturally from adopting existential closure at the
VP, as per Diesing (1992). Restricting existential closure to this structural domain accounts for the
fact that only internal arguments, i.e. the subject in (89) and the object in (90), may be interpreted
as indefinite. External arguments, such as the subject in (90), may only be definite, as they are
generated above existential closure. A definite interpretation is derived via standard assumptions
about type shifting operations for bare argument languages (i.e., Chierchia (1998b); Dayal (2004)).
The placement of existential closure for intransitive and transitive verbs is modeled in (91).
(91) a. Intransitive verb in Ch’ol b. Transitive verb in Ch’ol
IP IP
I SSP I SSP
SS VP ∃-closure SS VoiceP
V SUBJ Voice’ SUBJ
Voice VP ∃-closure
V OBJ
While I analyze the indefinite interpretation of absolutive subjects and objects as instances of
pseudo-incorporation, I also discuss instances of full nominal incorporation, such as in so-called
‘incorporation antipassive’ constructions, as in (92).
(92) Ta’ k-cha’l-e [NP kuch si’ ].
PFV A3-LV-DTV carry wood
‘I carried wood.’ Lit. I did wood-carrying.
As I will demonstrate below, in contrast to pseudo-incorporated nominals, bare nouns as incorpo-
rated nominals display a different pattern from the bare nouns in (89) and (90). Unlike the argu-
ments above, incorporated nouns may not serve as antecedents for agreement markers or pronomi-
nal anaphora. Additionally, kuch si’ ‘carry wood’ is in argument position of the light verb cha’l—it
55
is nominal. This syntactic fact has implications for a semantic analysis, as I will detail in Section
3.4.
This chapter is structured as follows. In Section 3.2, I present relevant data on bare nominals
as arguments in Ch’ol, providing evidence that a definite interpretation is freely available for bare
nouns. Indefinite interpretations are only possible for subjects of unaccusative verbs and objects
of transitive verbs. In Section 3.3, I implement an analysis based on Chung and Ladusaw’s (2004)
Restrict and existential closure at the VP (Diesing, 1992) for the indefinite interpretation of bare
nouns. This is based on the notion that the indefinite interpretations arise as they are instances
of pseudo-incorporation. I adopt the ι operator to derive a definite meaning for bare nouns and,
as ergative subjects are generated above existential closure, they may only be definite, achieved
via ι. I further discuss how a definite interpretation can be derived for internal arguments as well
as different dimensions of definiteness in Ch’ol. In Section 3.4, I discuss grammatical properties
of incorporation antipassives and list desiderata for a semantic analysis of these structures. Sec-
tion 3.5 concludes the chapter, detailing some crosslinguistic implications of this analysis for type
shifters in natural language as well as for analyses of word order.
3.2 Bare nouns in argument position
As has been reported before, bare nouns may be arguments in Ch’ol, as in the sentence in (93).1
(93) Tyi y-il-ä wiñik x-ixik.
PFV A3-see-TV man NC-woman
‘The woman saw the man.’ (Vázquez Álvarez, 2011, 21)
Vázquez Álvarez (2011) translates both bare nouns as definite in (93). As previewed in the intro-
duction, while a definite interpretation is available for all bare arguments, where nouns may be
interpreted as indefinite is syntactically restricted. Below, I provide evidence for this with data
from transcribed narratives and elicited speech.
1The x- prefix on the noun ixik is a noun class prefix; it does not contribute to whether the noun is
definite or indefinite. See Chapter 2, Section 2.4.2 for more details on noun class prefixes.
56
3.2.1 Absolutive arguments
Absolutive subjects
Bare absolutive subjects can be definite or indefinite, as shown in the examples from narratives in
(94). For instance in (94a), the woman is an established protagonist in the story and the sentence
was uttered towards the end of the story. In (94b), this is the first mention of the man.
(94) a. Context: The woman is an established protagonist in the story. (Definite)
Ta’ puts’-i lok’el x-ixik.
PFV flee-IV DIR:out NC-woman
‘The woman fled away.’ Bajlum
b. Context: First mention of man in the story. (Indefinite)
Ta’=bi chäm-i wiñik.
PFV=REP die-IV man
‘A man had died.’ Xi’ba.11
In both sentences in (94), the bare nouns are next to the verb complex. In (94a), x-ixik is next to
the verb puts’i lok’el ‘fled away’, which is what Aissen (2009) analyses as a serial verb construction.
In (94b) the bare noun wiñik ‘man’ is also next to the verb but receives an indefinite interpretation.
Bare absolutive subjects may appear next to the verb, or there may be intervening material,
such as a PP adjunct or adverb. When next to the verb, wiñik ‘man’ can be interpreted as indefinite
as in (95a). When there is an intervening adverb between the verb and bare subject, as in (95b),
the only possible interpretation is definite.
(95) Context A: What happened yesterday? Context B: When did the man arrive?
a. Ta’ jul-i wiñik ak’bi. b. Ta’ jul-i ak’bi wiñik .
PFV arrive-IV man yesterday PFV arrive-IV yesterday man
‘A man arrived yesterday.’ ‘The man arrived yesterday.’
A ✓; B ✗ A ✗; B ✓
Absolutive objects
Absolutive objects may also be definite as per (96a), or indefinite, as in (96b).
57
(96) a. Previous sentence in story ‘The people entered to where some corn was.’
Ta’=bi i-jok’-o-yob ixim a wiñik-ob.
PFV=REP A3-dig-TV-PL corn PRT man-PL
‘The men dug into the corn.’2 Xi’ba.86
b. Story about an evil godmother who lures a young girl to look for snails. This is the first
mention of the rock in the story.
Ta’=bi i-ch’äm-ä xajlel. . .
PFV=REP A3-take-TV rock
‘She took a rock.’ Kumale
In contrast to subjects in (95), bare objects must always appear adjacent to the verb, such as in
(97). These facts have also been reported by Coon (2010b) and Clemens and Coon (2018).
(97) a. Ta’ k-mäñ-ä otyoty ak’bi.
PFV A1-buy-TV house yesterday
‘I bought a house yesterday.’
b. * Ta’ k-mäñ-ä ak’bi otyoty.
PFV A1-buy-TV yesterday house
Intended: ‘I bought the house yesterday.’
3.2.2 Ergative subjects
Bare nouns may also appear as ergative subjects, but are always interpreted as definite, as indi-
cated by the translations in (98).
(98) Ta’ i-mäñ-ä otyoty wiñik.
PFV A3-buy-TV houseman
(i) ‘The man bought a/the house.’
(ii) NOT ‘A man bought a/the house.’
Speakers reject (ii) as a translation of (98) and instead provide (99) as the Ch’ol version for the
translation in (ii) where the numeral juñtyikil ‘one’ is used to force an indefinite interpretation of
the subject.
2It is unclear based on context whether the object receives a maximal interpretation. The point here is
that the noun is being used to refer back to the previously mentioned entity. See discussion in Section 3.3.7
for more on how definite interpretations for internal arguments arise as well as Section 3.3.8 for more on
anaphoric definites.
58
(99) Ta’ i-mäñ-ä otyoty juñ-tyikil wiñik.
PFV A3-buy house one-CLF man
‘A man bought a/the house.’
Furthermore, in a survey of eight transcribed narratives from the Tumbalá dialect of Ch’ol (see
Little and Vázquez Martı́nez (2018) and Vázquez Martı́nez and Little (2020)), every bare ergative
argument was interpreted as definite, such as in example (100).
(100) Context: The fox is an established protagonist in the story.
Ta’=bi iy-ajñ-es-a wax.
PROG A3-run-CAU-DTV grey.fox
‘The grey fox was running after them.’ Xi’ba.77
3.2.3 Scope interactions
Scope interactions provide evidence that indefinite objects must take low scope with respect to
other quantifiers, as in (101a). For instance, in (101a), juñ ‘book’ in the object position takes narrow
scope with respect to jujuñtyikil xch’ok ‘every girl’. Similarly in (101b), the bare nominal ixim ‘corn’
takes low scope with respect to the negation marker ma’añik.3
(101) a. Ta’ i-pejk-a juñ ju-juñ-tyikil x-ch’ok.
PFV A3-read-TV book RED-one-CLF NC-girl
‘Every girl read a book.’ ∀ > ∃
b. Ma’-añ=ik ta’ k-mäñ-ä ixim.
NEG-EXT=IRR PFV A1-buy-TV corn
‘I didn’t buy any corn.’ ¬ > ∃
3In addition to nominal modifiers, there are other quantificational adverbial elements that can appear
prefixed to the verb such as in (1).
(1) a. Mi i-ñoj-choñ waj.
PFV A3-always-sell tortilla
‘He always sells tortillas.’
b. Ta’ i-laj-jap-ä-yob sa’.
PFV A3-all-drink-TV-PL pozol
‘They drank all the pozol.’ (Adapted from Aulie and Aulie 1978, 48)
Preliminary evidence suggests that these prefixes take scope over internal arguments such as absolutive
objects and subjects, but not external arguments. These verbal prefixes need further investigation especially
with respect to the differences in A(verbial) quantification and D(eterminer) quantification, which has, for
instance, been discussed in detail for typologically diverse languages in Bach et al. (1995).
59
3.2.4 Summary
In sum, indefiniteness is not licensed in every argument position. Ergative-marked bare nouns
are always interpreted as definite. Objects of transitive verbs and subjects of unaccusative verbs
(i.e., absolutive arguments) may be definite or indefinite. While bare objects must remain next to
the verb and can be interpreted as definite or indefinite, bare absolutive subjects can be next to the
verb, or other material may intervene. But, as shown by the differences in (95), the bare absolutive
subject is interpreted as definite if there is intervening material between it and the verb. Table 3.1
summarizes the available interpretations of bare arguments in Ch’ol.
Table 3.1: Available interpretations of bare arguments in Ch’ol
Definite Indefinite
Subjects of intransitive verbs (absolutive) ✓ ✓
Objects of transitive verbs (absolutive) ✓ ✓
Subjects of transitive verbs (ergative) ✓ ✗
3.3 Deriving the Ch’ol pattern
For my analysis, I adopt a Neo-Davidsonian framework where verbs are predicates of events.
I also assume, following much previous work, that external arguments are not arguments of
verbs, but are introduced by a Voice head (e.g., Marantz (1984); Kratzer (1996)). To derive the
indefinite interpretation of internal arguments, verbs and predicative nouns combine via Restrict
(Chung and Ladusaw, 2004), paralleling Chung and Ladusaw’s (2004) analysis of low scope in-
definites in Maori (Austronesian), and capitalizing on the notion that indefiniteness in Ch’ol is an
instance of pseudo-incorporation. The theme argument is then existentially closed at the VP, as
per the Mapping Hypothesis (Diesing, 1992). Structurally constraining existential closure to the
VP predicts the interpretational possibilities of bare nouns in Ch’ol.
60
3.3.1 Semantic preliminaries
I use a Neo-Davidsonian framework to model my analysis of the facts in Table 3.1. I give a sample
of my notational conventions in Table 3.2. Nouns are type 〈et〉, sentences are type t, verbs are
predicates of events, etc.
Table 3.2: Semantic notation
Notation Type Metalanguage Object language Syntax
e individual a, b, c. . . ajRosa ‘Rosa’ DP
t truth value 1,0 ta’ ik’uxu waj IP
ajRosa ‘Rosa ate a
tortilla’
s eventuality (event) s1, s2 . . .
〈et〉 function from indi- λx.[TORTILLA(x)] waj ‘totilla’ N
viduals to truth val-
ues
〈e〈st〉〉 function from in- λxλe.[SELL(x)(e)] choñ ‘choñ’ V
dividuals to func-
tions from eventu-
alities to truth val-
ues
〈st〉 function from even- λe.[SEE(r)(e)] k’el ajRosa ‘see VP,
tualities to truth Rosa’ VoiceP
values
All verbs in Ch’ol take internal arguments as per Coon (2012, 2013). This is encoded semanti-
cally for the unaccusative verb jul ‘arrive’ and the transitive verb choñ ‘sell’ in (102a) and (102b),
repsectively.4 (See also Coon (2019) for verbs in Chuj, a Mayan language of the Q’anjob’alan
branch.) I analyze Ch’ol nouns as being type 〈et〉, as in (102c).5
(102) a. ~jul = λxλe.[ARRIVE(x)(e)]
b. ~mäñ = λxλe.[BUY(x)(e)]
c. ~otyoty = λx.[HOUSE(x)]
4Unergative verb constructions in Ch’ol are formally transitive, where the subject is introduced as an
agent to a light verb and the light verb takes a nominal such as ‘dance’ as its internal argument. See details
from Chapter 2 Section 2.3.3 and some discussion in Section 3.4 below.
5Little (2018b) analyzes nouns as being number neutral, with the denotation λx.[∗TORTILLA(x)]. See
Section 3.3.7 below for more discussion.
61
3.3.2 The Mapping Hypothesis and indefinite bare arguments
In order to account for low scope indefinites, I adopt Diesing’s (1992) Mapping Hypothesis in
(103).
(103) Mapping Hypothesis Diesing (1992)
a. Material from VP is mapped into the nuclear scope.
b. Material from IP (above VP) is mapped into a restrictive clause.
More specifically, existential closure is structurally constrained to the VP level as per (103a)
and modeled in (104).
(104) . . .
. . . VP ∃-closure
V NP
In Ch’ol, absolutive arguments, that is, subjects of intransitive verbs and objects of transitive
verbs, are both introduced as complements to the verb, as shown in (105). Voice introduces the
agent in Ch’ol, and the transitive verbalizer is in a status suffix above VoiceP. For ease of expo-
sition, I assume that subjects of transitive verbs are in a right specifier of VoiceP, as per Aissen
(1992). I further defend this view in Chapter 5.6
6Though not represented in (105a), there is additionally an intransitive vP projection between the SSP
and VP. This projection would host, e.g., passivizing material. Since it does not factor into the semantics
here, I do not include it in the tree.
62
(105) a. Intransitive verb in Ch’ol b. Transitive verb in Ch’ol
IP IP
I SSP I SSP
SS VP ∃-closure SS VoiceP
V SUBJ Voice’ SUBJ
Voice VP ∃-closure
V OBJ
3.3.3 Existential closure of the event and negation
As mentioned in Chapter 2, I label the projection that hosts the perfective aspect marker as IP
(i.e., for INFL). I assume that this is the same projection as INFL/T from Coon (2017) or INFL
from Clemens and Coon (2018). This is where the event is existentially closed, below any eventive
negation markers. Negation, which appears before the aspect marker, is composed of the negative
particle mach and the existential predicate añ, as in (106). The negation marker in Tumbalá Ch’ol
additionally appears with the irrealis enclitic -ik.
(106) a. Ma’-añ=ik ta’ i-k’el-e-yoñ.
NEG-EXT=IRR PFV A3-see-TV-B1
‘She did not see me.’
b. Ma’-añ=ik woli i-wäy-el.
NEG-EXT=IRR PROG A3-sleep-NML
‘He is not sleeping.’
I take the event to be existentially closed below negation—a robust crosslinguistic generaliza-
tion (Herburger, 2001; Zeijlstra, 2004; Zimmermann, 2007). In fact, Ch’ol provides evidence in
support of overt existential closure of events in negative sentences, as has also been described in
the Central Chadic language Bura in Zimmermann (2007). Indeed, the eventive negative mor-
pheme can be decomposed into negation particle mach and the existential predicate añ, glossed as
such in (106).7
7Zimmermann (2007) analyzes the distribution of the existential predicate adi in Bura, a Central Chadic
63
With these pieces in mind, I first derive an indefinite interpretation for the intransitive subject.
Derivation for bare indefinite absolutive subjects
Consider the sentence in (107), where the subject wiñik ‘man’ is indefinite.
(107) Ta’ jul-i wiñik.
PFV arrive-IV man
‘A man arrived.’
The verb ‘arrive’ in (102a) of type 〈e〈st〉〉 cannot yet combine with the subject ‘man’ of
type 〈et〉 as there is a type mismatch. To remedy this, I adopt the operation Restrict from
Chung and Ladusaw (2004), stated in (108). This composition rule combines a verb and a noun,
where the end result is that the event denoted by the verb becomes more restricted in meaning.
(108) Restrict = λP〈e〈st〉〉λQ〈et〉λxλe. [P(x)(e) ∧ Q(x)] (Chung and Ladusaw, 2004)
Restrict is commonly used in the semantics for noun incorporation as Chung and Ladusaw
(2004, Chapter 3) do for Chamorro (Austronesian) and Coon (2019) does for the Mayan language
Chuj. Chung and Ladusaw (2004) also use Restrict to derive the semantics for the indefinite article
he in Maori. Using Restrict also capitalizes on the notion that the indefinites in Ch’ol are instances
of pseudo-incorporation.
To illustrate how this works for Ch’ol, take (109). The verb and noun in (107) combine first via
Restrict in (108), deriving (109). Here the event of arriving is restricted to events of men arriving.
language. In Bura, the existential predicate occurs in negative eventive clauses and can introduce new
discourse references. The negation marker in Ch’ol for eventive clauses is composed of the existential and
the negative particle. The existential predicate in Ch’ol can also used to introduce indeterminate pronouns
as in (1).
(1) Añ chu kom.’
EXT what A1-want
‘I want something.’
The existential predicate can also be used with tensed clauses as in (2).
(2) Añ ta’ k-jap-ä lembal.
EXT PFV A1-drink-TV alcohol
‘I have drunk alcohol.’ Lit. ‘There is an event of me drinking alcohol.’
64
(109) VP〈e〈st〉〉
jul wiñik
λxλe.[ARRIVE(x)(e) & MAN(x))] (via Restrict)
V〈e〈st〉〉 NP〈et〉
jul wiñik
λxλe.[ARRIVE(x)(e)] λx′.[MAN(x′)]
The result in (109) is not yet complete as the theme argument is not yet saturated. This is where
existential closure comes in, given in (110) and structurally constrained to the VP level.
(110) Existential closure at VP
λP〈e〈st〉〉λe.∃y[P(y)(e)]
After existential closure, the result is (111).
(111) VP〈st〉
jul wiñik
λe.∃y[ARRIVE(y)(e) & MAN(y)]
λP〈e〈st〉〉λe.∃y[P(y)(e)] VP〈e〈st〉〉
jul wiñik
λxλe.[ARRIVE(x)(e) & MAN(x))] (via Restrict)
V〈e〈st〉〉 NP〈et〉
jul wiñik
λxλe.[ARRIVE(x)(e)] λx′.[MAN(x′)]
Putting everything together, we derive (113) where (113) is true if and only if there is an arriv-
ing event and there is a man who is the theme of the arriving event. I assume that the intransitive
verbalizer -i, glossed as IV in (107), is obligatory in the syntax but does not factor into the compo-
sitional semantics. The aspect marker encodes existential closure of the event, given in (112).8
(112) ~ta’ = λR.〈st〉∃e[R(e)]
8Carlson (2003) takes a different view where the event is existentially closed at the VP level, not at the
clause level. Importantly for the data here, existential closure of the event takes place below negation, as
discussed above.
65
(113) IPt
ta’ jul-i wiñik
∃e∃y[ARRIVE(y)(e) & MAN(y)]
I〈st〉 SSP〈st〉
ta’ jul-i wiñik
λR.〈st〉∃e[R(e)] λe.∃y[ARRIVE(y)(e) & MAN(y)]
SS VP〈st〉
-i jul wiñik
∅ λe.∃y[ARRIVE(y)(e) & MAN(y)]
λP〈e〈st〉〉λe.∃y[P(y)(e)] VP〈e〈st〉〉
jul wiñik
λxλe.[ARRIVE(x)(e) & MAN(x))] (via Restrict)
V〈e〈st〉〉 NP〈et〉
jul wiñik
λxλe.[ARRIVE(x)(e)] λx′.[MAN(x′)]
Derivation for bare indefinite absolutive objects
As shown above for absolutive subjects, absolutive objects may also be indefinite, as in (114).
(114) Ta’ i-mäñ-ä otyoty aj-Rosa.
PFV A3-buy-TV house NC-Rosa
‘Rosa bought a house.’
The derivation for the VP proceeds as seen above for absolutive subjects, where the verb mäñ
‘buy’ and otyoty ‘house’ combine via Restrict and the theme argument is existentially closed at the
VP, as in (115).
66
(115) VP〈st〉
mäñ otyoty
λe.∃y[BUY(y)(e) & HOUSE(y)]
λP〈e〈st〉〉λe.∃y[P(y)(e)] VP〈e〈st〉〉
mäñ otyoty
λxλe.[BUY(x)(e) & HOUSE(x))] (via Restrict)
V〈e〈st〉〉 NP〈et〉
mäñ otyoty
λxλe.[BUY(x)(e)] λx′.[HOUSE(x′)]
As per Kratzer (1996), external arguments are not arguments of the verb but rather introduced
in the specifier of VoiceP where Voice has the semantics in (116). See also Coon and Preminger
(2013) for arguments that the subject is generated externally in Ch’ol as well.
(116) ~Voice = λxλe.[Agent(x)(e)] (Kratzer, 1996, 121)
There is no overt morphological realization of Voice in the transitive clause in (114). (Overt Voice
morphemes do exist in Ch’ol, such as for causative structures, detailed more in Chapter 5 Section
5.4.1.) Voice introduces the ergative subject, aj-Rosa, and semantically encodes it as the agent of
the sentence. The VP in (115) and Voice head combine via Event Identification, given in (117).
(117) Event Identification (Kratzer, 1996)
λP〈e〈st〉〉λQ〈st〉λxλe.[P(x)(e) ∧ Q(e)]
The agent, aj-Rosa is combined, via Functional Application. The transitive verbalizer -ä is
needed in the syntax and sits in SSP, but does not factor into the semantic derivation. The event is
existentially closed by the aspect marker ta’, as in (118). I assume that the third person agreement
prefix i-, glossed as A3, is the reflex of an agreement relation with the subject aj-Rosa and does not
factor into the compositional semantic analysis. See Chapter 2 Section 2.3.2 for more details on
agreement.
67
(118) IP〈t〉
ta’ i-mäñ-ä otyoty aj-Rosa
∃e[Agent(r)(e) & ∃y(BUY(y)(e) & HOUSE(y))]
I〈〈st〉t〉 SSP〈st〉
ta’ i-mäñ-ä otyoty aj-Rosa
λR〈st〉.∃e[R(e)] λe.[Agent(r)(e) & ∃y(BUY(y)(e) & HOUSE(y))]
SS VoiceP〈st〉
-ä i-mäñ otyoty aj-Rosa
∅ λe.[Agent(r)(e) & ∃y(BUY(y)(e) & HOUSE(y))]
Voice’〈e〈st〉〉 DP〈e〉
mäñ otyoty aj-Rosa r
λxλe.[Agent(x)(e) & ∃y(BUY(y)(e) & HOUSE(y))]
(via Event Identification)
Voice〈e〈st〉〉 VP〈st〉
λxλe.[Agent(x)(e)] mäñ otyoty
λe.∃y[BUY(y)(e) & HOUSE(y)]
λP〈e〈st〉〉λe.∃y[P(y)(e)] VP〈e〈st〉〉
mäñ otyoty
λxλe.[BUY(x)(e) & HOUSE(x))] (via Restrict)
V〈e〈st〉〉 NP〈et〉
mäñ otyoty
λxλe.[BUY(x)(e)] λx′.[HOUSE(x′)]
The sentence in (118) is true if and only if there exists an event of buying with Rosa as the agent
and there exists a house and the house is the theme of the buying event.
The analysis in (118) correctly predicts a low scope indefinite reading of otyoty ‘house’, as the
object is existentially closed at the VP level—lower than the subject or negation, consistent with
Diesing’s (1992) analysis of existential indefinites. This is exhibited in the data on the scope of
bare indefinite internal arguments from (101) repeated in (119). Existential closure of the internal
argument takes place at the VP level, therefore capturing that the internal argument will take low
68
scope with respect to the subject, as in (119a), or negation, in (119b).
(119) a. Ta’ i-pejk-a juñ ju-juñ-tyikil x-ch’ok.
PFV A3-read-TV book RED-one-CLF NC-girl
‘Every girl read a book.’ ∀ > ∃
b. Ma’añ=ik ta’ k-mäñ-ä ixim.
NEG=IRR PFV A1-buy-TV corn
‘I didn’t buy any corn.’ ¬ > ∃
3.3.4 Deriving definite bare nominal subjects
To illustrate why bare nominal transitive subjects are always definite, I consider the sentence in
(120).
(120) Ta’ i-mäñ-ä otyoty wiñik.
PFV A3-buy-TV houseman
‘The man bought a house.’
Given existential closure at the VP, consider again the syntactic structure of the sentence in
(120), given in (121).
(121) VoiceP
Voice’ SUBJ
Voice VP ∃-closure
V OBJ
Now given the denotation of wiñik ‘man’ in (122) there is a type mismatch: Voice, repeated in
(123), takes an individual as its first argument. Voice in (123) cannot combine with (122), as (122)
as is type 〈et〉.
(122) ~wiñik = λx.[MAN(x)]
(123) ~Voice = λxλe.[Agent(x)(e)] (Kratzer, 1996, 121)
69
Furthermore, as existential closure is syntactically restricted, if Voice and (122) were to com-
bine via Restrict, it would happen above the domain of existential closure, meaning that the agent
argument cannot be existentially closed and the agent argument would be left unsaturated.
As Ch’ol does not have a definite article in the standard sense (see Vázquez Álvarez (2011);
Little and Vázquez Martı́nez (2018); Vázquez Martı́nez and Little (2020) for discussion), I propose,
following standard assumptions about bare argument languages (Chierchia, 1998b; Dayal, 2004),
that bare nouns may be type shifted via the ι operator given in (124). This type shifter takes the
predicate in (122) and shifts it to an individual, requiring the predicate to be true of one entity.
(124) ι type shifter9
λP〈et〉.ι(z)[P(z)]
The full derivation is given in (125). Voice combines with the VP from (115) above. The agent is
merged in a position above VP, and above existential closure; thus, for it to be an argument of the
verb, the only option is for it to be type shifted via ι in (124). It may then combine via Functional
Application and be interpreted as the agent.
9In order to capture definite plurals ι can be substituted for the supremum operator, where definiteness
with pluralities can come through a maximality condition (Landman, 2004) as in (1).
(1) λP.sup(P) if defined; undefined otherwise (Solt, 2011, 2)
where sup(P) = ιx[P(x) & ∀y[P(y) → y ≤ x]]
70
(125) IP〈t〉
ta’ i-mäñ-ä otyoty wiñik
∃e.[Agent(ιz(MAN(z)))(e) & ∃y(BUY(y)(e) & HOUSE(y))]
I〈〈st〉t〉 SSP〈st〉
ta’ i-mäñ-ä otyoty wiñik
λR〈st〉.∃e[R(e)] λe.[Agent(ιz(MAN(z)))(e) & ∃y(BUY(y)(e) & HOUSE(y))]
SS VoiceP〈st〉
-ä i-mäñ otyoty wiñik
∅ λe.[Agent(ιz(MAN(z)))(e) & ∃y(BUY(y)(e) & HOUSE(y))]
Voice’〈e〈st〉〉 NP〈e〉
mäñ otytoy wiñik
λxλe.[Agent(x)(e) & ∃y(BUY(y)(e) & HOUSE(y))] ιz(MAN(z))
(via Event Identification)
λP〈et〉.ι(z)[P(z)] NP〈et〉
Voice〈e〈st〉〉 (115) wiñik
λxλe.[Agent(x)(e)] λx.[MAN(x)]
Existential closure may only apply at the VP level, accounting for the fact that only nominals
in the complement position of a verb may be indefinite. The ι type shifter, on the other hand, is
available for all nouns regardless of their syntactic position—accounting for why bare nouns as
complements to verbs may also be definite. When nominal complements of verbs are type shifted
via (124), the verb and nominal combine via Functional Application, where the complement satis-
fies the individual argument of the verb.
3.3.5 Moved absolutive subjects
Certain moved absolutive subjects also are interpreted as definite.10 Take again the data in (126)
where absolutive subjects next to the verbmay be interpreted as indefinite, as in (126a), but moved
subjects may not, as in (126b).
10Note that focused absolutive subjects can be indefinite or definite and appear in a preverbal position.
It is movement within the verbal domain that is relevant for the availability of a definite or indefinite
interpretation.
71
(126) Context A: What happened yesterday? Context B: When did the man arrive?
a. Ta’ jul-i wiñik ak’bi. b. Ta’ jul-i ti ak’bi wiñiki .
PFV arrive-IV man yesterday PFV arrive-IV yesterday man
‘A man arrived yesterday.’ ‘The man arrived yesterday.’
A ✓; B ✗ A ✗; B ✓
Data with the predicate añ display a similar pattern. Añ is typically glossed as an existential
predicate but is also used in locative constructions. (See also Polian (2004) for a similar pattern in
the closely related language Tseltal regarding the existential/locative predicate.) When the absolu-
tive subject is next to the existential predicate, as in (127a), it receives an indefinite interpretation.
This again follows from the placement of existential closure at the VP level. When the subject is
no longer next to the verb, indicated by the intervening PP in (127b), an indefinite interpretation
is no longer available.
(127) a. Añ wiñik tyi otyoty.
EXT man PREP house
‘There is a man in the house.’
b. Añ tyi otyoty wiñik.
EXT PREP houseman
‘The man is in the house.’
Returning to the examples in (126), restricting existential closure at the VP level predicts the
available interpretations. The structure for (126a) is given in (128b).
(128) a. Ta’ jul-i wiñik ak’bi.
PFV arrive-IV man yesterday
‘A man arrived yesterday.’
b. IP
I SSP
ta’
PFV
SS VP
-i
∃
VP AdvP
ak’bi
yesterday
V NP
jul wiñik
arrive man
In (128b), the absolutive subject is introduced in the complement of V and is in the domain of
existential closure.
72
Once the subject moves out of the VP, the only available interpretation is a definite one, via ι,
as modeled in the structure in (129b).11,12 In (129b), the subject moves to a right specifier of SSP,
the projection demarcating the edge of the verbal domain. As mentioned above, I do not represent
intransitive v in the structure, but the specifier of v can also be a conceivable landing site for the
moved absolutive subject. What is important is that the subject wiñik ‘man’ has moved out of the
VP.
(129) a. Ta’ jul-i ak’bi wiñik.
PFV arrive-TV yesterdayman
‘The man arrived in Salto.’
b. IP
I SSP
ta’
PFV
SS NPi
wiñik
man
SS VP
-i
∃
VP AdvP
ak’bi
yesterday
V ti
jul
arrive
3.3.6 Numerals in object position
Data from shifted and unshifted objects with numeral modifiers provide further evidence for
where existential indefinite interpretations are available. For instance, NPs with numeral mod-
ifiers receive a low scope interpretation in (130a), with respect to the adverbial prefix ñoj ‘always’.
But when there is an intervening adjunct such as the PP in (130b), the NP with the numeral mod-
ifier is interpreted as specific. In (130a), there are two possible readings for the object: (i) the
speaker always sees two men on the road, but not necessarily the same men, or (ii) the speaker
11This proposal is similar to recent work in Collins (2019). Collins proposes that in Tagalog, all nominals
which move outside the VP receive definite interpretations, derived via an ι type shifter. Nominals within
the VP must receive a low scope indefinite interpretation. This is where the current account differs from
Collins’s (2019). In Ch’ol, bare nominals inside VP can also be type shifted via ι.
12Bare objects must appear next to the verb, but absolutive subjects do not need to, as seen in (126b).
I propose that this is due to the different licensors for these arguments. As per Coon (2017), objects are
licensed by Voice and absolutive subjects are licensed by INFL.
73
always sees two specific men. However, in (130b), there is an intervening PP between the verb
and the object and the only possible interpretation is that the speaker always sees two specific
men on the road. These differences in interpretations provide clues to where the objects in (130)
are structurally.
(130) Context A: There is a road behind my house. I always see two men walking there, some-
times it is Marcos and Rogelio, sometimes Juan and Eladio and sometimes Jorge and Javier.
Context B: There is a road behind my house. I always see Nicolás and José walking there.
a. Mi k-ñoj-k’el [ cha’-tyikil wiñik ] tyi bij.
IMF A1-always-see two-CLF man PREP way
‘I always see two men on the road.’ ✓A ✓B
b. Mi k-ñoj-k’el tyi bij [ cha’-tyikil wiñik ].
IMF A1-always-see PREP way two-CLF man
‘I sometimes see on the road two men.’ ✗A ✓B
It has been noted before in the literature that numerals can be both quantificational or predica-
tional (e.g., Milsark (1974); Diesing (1992); de Hoop (1996); Dayal (2013)). The two interpretations
in (130) with respect to the adverbial prefix ñoj ‘always’ can be analyzed as following. In (130a)
the numeral is type 〈et〉, combining with the noun via predicate modification, deriving an object
of type 〈et〉. The object remains next to the verb, within the domain of existential closure; the verb
and object combine via Restrict and then the theme is existentially closed at the VP level, below the
attachment of the adverbial prefix ñoj- ‘always’. This is schematized in (131a). In (130b), the quan-
tificational object moves out of the VP, outside the scope of the adverb ñoj ‘always’, as in (131b).13
The subject is a PRO as the verb is a nominalized complement. The PRO is controlled by a DP in
Spec,PossP. See Chapter 2, Section 2.3.5 for more details.
13An additional prediction of the analysis in (131b) is that the numeral should take wide scope with
respect to the subject. This remains to be tested.
74
(131) a. The object can take low scope with respect to the VP adverb
VoiceP
PRO Voice’
Voice VP
VP PP
tyi bij
AdvP VP on road
ñoj- V NP
always k’el
see cha’tyikil wiñik
two men
b. The object must take wide scope with respect to the VP adverb
VoiceP
VoiceP QPi
cha’tyikil wiñik
PRO Voice’ two men
Voice VP
VP PP
AdvP VP tyi bij
on road
ñoj- V ti
always k’el
see
In (130b), the numeral does not appear next to the verb, as there is an intervening PP. I take
the intervening PP to be indicative that the object has moved out. The presuppositional reading
may be accounted for by positing that the numeral is quantificational (i.e., it is 〈〈et〉〈〈et〉t〉〉). It
combines with the nominal, saturating its first argument, deriving a QP of type 〈〈et〉t〉. As the
object is a generalized quantifier, it may not remain in object position, as this would result in a
75
type mismatch. Rather it moves out of the VP, undergoing Quantifier Raising, leaving behind a
trace and taking scope over the sentence at LF. (The QP in (131b) undergoes additional movement
at LF outside the VoiceP.) This, in turn, accounts for the wide scope of the numeral in (130b). Note
that the quantificational interpretation is not ruled out in (130a)—if the numeral in (130a) is a QP,
it undergoes additional movement at LF. Crucially, however, the low scope interpretation of the
object is ruled out in (130b).
3.3.7 Definite interpretations of bare internal arguments
While I focused on deriving an indefinite interpretation of internal arguments and a definite inter-
pretation of external arguments, I discuss here how definite interpretations can arise with internal
arguments and the factors that play a role in their definiteness. Consider again the sentence in
(132).
(132) Ta’ i-mäñ-ä otyoty aj-Rosa.
PFV A3-buy house NC-Rosa
‘Rosa bought a/the house.’
Without context, the object can be interpreted as definite or indefinite. However, there are
factors that constrain this intepretation. Crosslinguistically, objects tend to be indefinite and inan-
imate while subjects are definite and animate (Keenan, 1976). While a definite interpretation may
arise for the object in (132), the animacy of the object can affect its interpretation. While the an-
imate object in the VOS sentence in (133a) is more likely to be interpreted as definite without
context, the inanimate VOS object in (133b), is more likely to be interpreted as indefinite out of the
blue.
(133) a. Tyi y-il-ä wiñik x-ixik.
PFV A3-see-TV man NC-woman
‘The woman saw the man.’ (Vázquez Álvarez, 2011, 21)
b. Ta’ i-tyuk’-u kajpe’ x-Wañ.
PFV A3-cut-TV coffee NC-Juan
‘Juan picked coffee.’
76
It has been noted in theMayan literature that animacy influences word order (England, 1991), and
here I suggest that it also influences whether it is interpreted as definite.
Above we saw how ι may type shift bare subjects—this is the only way they may combine
with transitive verbs to be arguments. The definite interpretation of the object can be type-shifted
via ι, derived in (134) for the VP ‘buy the house’.
(134) Derivation for the VP ‘buy the house’
VP〈st〉
mäñ otyoty
λe.[BUY(ιz(HOUSE(z)))(e)]
V〈e〈st〉〉 NP〈e〉
mäñ otyoty
λxλe.[BUY(x)(e)] ιz(HOUSE(z))
λP〈et〉.ι(z)[P(z)] NP〈et〉
otyoty
λx.[HOUSE(x)]
There are no morphological differences between the analysis of indefinites above and definite in-
ternal objects in (134), but certain factors such as animacy and context influence its interpretation.
Another area worth exploring is that of number marking and definiteness in Ch’ol. As men-
tioned in Chapter 2, bare nouns appear to be number neutral in Ch’ol—they can combine directly
with numerals, as in (135), without extra plural morphology.
(135) a. cha’-tyikil wiñik b. ux-k’ej waj
two-CLF man three-CLF tortilla
‘two men’ ‘three tortillas’
Ch’ol does, however, have a plural marker, -ob, used mainly for human-denoting nouns. Nev-
ertheless, the bare noun in (136) is interpreted as definite and singular out of the blue, which is
why the meaning of this sentence is odd.14
14Further comparison is needed if a similar construction results in an infelicitous interpretation when
the subject is inanimate.
77
(136) #Woli i-k’oty-el wiñik tyi ju-jum-p’ej tyejklum.
PROG A3-arrive-NML man PREP RED-one-CLF city
‘The man is arriving in every city.’
Speaker comment: Sounds like he is an omnipresent man, as if from a story.
Examples such as (136) for Dayal (2004) suggest that it is not possible for a number-marking
language to have a purely indefinite reading outside of (pseudo)incorporation structures.15 In-
definiteness must come through the grammar in other ways (such as the pseudo-incorporation
analysis given above). Out of the blue, (136) shows that a definite interpretation is more readily
available for the bare noun wiñik ‘man’. Further work is needed to ascertain whether the struc-
tural position of the subject in (136) influences its interpretation. As discussed in Chapter 2, non-
perfective aspects take nominalized complements. Structurally, the subject of (136) is a possessor
(indicated by the appearance of set A genitive morphology on k’otyel ‘arrive’). Nevertheless, this
example does bring up a few lines of inquiry, such as whether Ch’ol really is a number neutral
language after all, or if plurality must be overtly marked in certain contexts.
In the next section, I turn to overt markers of definiteness in Ch’ol and discuss syntactic impli-
cations for when definiteness is overtly marked.
3.3.8 Dimensions of definiteness and effects on word order
Overt morphemes, such as jiñi, can also signify a definite interpretation in Ch’ol. I argue that
jiñi indicates anaphoric definiteness; in contrast, bare nouns correspond to a unique definite in-
terpretation. Furthermore, determiners are prohibited from appearing in VOS object position. I
tie this into the syntactic structures proposed by Jenks (2018) for unique and anaphoric definites,
provided in (137a) and (137b).
15See also discussion in Despić (2019) for how Dayal’s (2004) analysis relates to kinds and anaphoricity.
78
(137) Jenks (2018, 515)
a. Unique definite b. Anaphoric definite
NP DP
ι NP
Pred(1) DP
NP
D NP
ιx s NP
Important for the discussion at hand is that anaphoric definites have a DP structure, but unique
definites are NPs. This relates to the word order facts in (138): NPs can appear in VOS object
position, but DPs are banned from this position, instead surfacing in VSO object position.
(138) a. VOS objects = NPs
Ta’ i-mäñ-ä [O otyoty ] [S aj-Maria ].
PFV A3-buy-TV house NC-Maria
‘Maria bought a/the house.’
b. VSO objects = DPs
Ta’ i-mäñ-ä [S aj-Maria ] [O jiñi otyoty ].
PFV A3-buy-TV NC-Maria DET house
‘Maria bought the house.’
While the ban on DP objects in VOS object position is construed as a syntactic difference, I
relate the syntactic restriction in (138) to the semantics of bare definite nouns and nouns marked
with jiñi. That is, the DP object in (138b) is anaphoric. I begin with a brief overview of unique
versus anaphoric definiteness.
Recent work has argued that definiteness can be split up into different dimensions. Support
for this comes from languages that mark these dimensions in distinct ways. For instance, Schwarz
(2009, 2013) presents data from the Frisian dialect of Fering (Germanic) in (139), which marks
unique and anaphoric definites with morphologically distinct determiners. The first context in
(139a) is an instance of a unique definite determiner. Unique definites make reference to salient
entities in larger or global contexts—for instance ‘the king’ in a country with one king or ‘the pi-
ano’ in a room with one piano. Anaphoric definites, such as in (139b) on the other hand, reference
previously mentioned entities, such as the horse in the first sentence.
79
(139) Fering
a. A köning kaam to bischük.
theweak king came to visit
‘The king came for a visit.’ (Schwarz (2013, 541) citing Ebert (1971a, 82–83))
b. Oki hee an hingst keeft. *A / Di hingst haaltet.
Oki has a horse bought theweak thestrong horse limps
‘Oki has bought a horse. The horse limps.’ (Schwarz (2013, 538) citing Ebert (1971b,
161))
While in Fering, different articles are used in (139a) and (139b), in English, as evidenced by
the translations, ‘the’ is used in both contexts. Though languages such as English do not make a
morphological distinction, recent work has shown that other languages do (Schwarz, 2009, 2013;
Arkoh and Matthewson, 2013). Some languages mark unique definites such as (139a) with bare
nouns but use a determiner for anaphoric reference (Arkoh and Matthewson, 2013; Jenks, 2018).
Other languages mark both unique and anaphoric definites with bare nouns (Moroney, 2020). I
provide support here that Ch’ol mostly falls into the category where unique definintes, such as in
the example in (139a), are indexed by bare nouns but anaphoric definites are marked with an overt
determiner. For instance, in (140a), ‘moon’ is referred to by a bare noun, whereas in the anaphoric
usage in (140b), a determiner appears.
(140) a. Unique context (first mention of uw ‘moon’ in the story):
Mi tsa’ mäjk’-i uw. . .
COMP PFV hide.PSV-IV moon
‘When the moon is hidden. . . (Whittaker and Warkentin, 1965, 71)
b. Anaphoric context: After previous mention of x-k’aläl ‘girl’ twice before:
K’uñtya k’uñtya mi y-ust-es-äñ-tyel i-tyaty i-ña’ jiñi
slow slow IPFV A3-convince-CAU-DTV-PSV-NML A3-father A3-mother DET
x-k’aläl.
NC-girl
‘Slowly the parents of the girl become convinced.’ (Cuentos Cultura Chol: 14)
In elicited contexts, to refer to the sun, speakers prefer bare nouns as in (141a). If the determiner
jiñi appears with ‘sun’, speakers comment that it is better without jiñi. This supports the argument
that bare nouns can refer to unique definites, but jiñi indexes anaphoricity.
80
(141) a. Ta’ lok’-i k’iñ.
PFV come.out-IV sun
‘The sun came out.’
b. # Ta’ lok’-i jiñi k’iñ.
PFV come.out-IV DET sun
Intended: ‘The sun came out.’
Speaker comment: This sentence is better without jiñi
This pattern is further corroborated by a corpus study by Little and Vázquez Martı́nez (2018)
and Vázquez Martı́nez and Little (2020). They found that unique definites are referred to
with bare nouns, but determiners usually appear when reference is made anaphorically. In
Vázquez Martı́nez and Little (2020), 64% of the instances of anaphoric definites appeared with
a determiner, whereas unique definites were referred with bare nouns 100% of the time.
Based on work from Schwarz (2009, 2013), Jenks (2018) posits different semantic denotations
for unique and anaphoric definites. The denotations for the two types of definiteness are given in
(142).
(142) Jenks (2018, 513)
a. UNIQUE DEFINITE ARTICLE
~ι = λsr.λP〈e〈st〉〉. : ∃!x[P(x)(sr].ιxP(x)(sr)
b. ANAPHORIC DEFINITE ARTICLE
~ιx = λsr.λP〈e〈st〉〉.λQ〈et〉 : ∃!x[P(x)(sr ∧ Q(x)].ιxP(x)(sr)
Relevant to the discussion here is that for Jenks (2018), the two syntactic structures for unique
definites and anaphoric definites are distinct, repeated in (143). Unique definites are NPs as in
(143a), but anaphoric definites are DPs (143b).
(143) Jenks (2018, 515)
a. Unique definite b. Anaphoric definite
NP DP
ι NP
Pred(1) DP
NP
D NP
ιx s NP
81
The determiner jiñi, I propose, corresponds to Jenks’s anaphoric definites. While unique defi-
nites, NPs for Jenks (2018), are compatible within the VP, anaphoric definite DPs are not and must
move out. This is reflected syntactically as objects with jiñi surface in VSO object position. This is a
syntactic restriction—a prohibition on DPs to remain in the VP—but relates to the semantics: DPs
marked with jiñi refer anaphorically, unlike unique definite NPs. The semantics is thus reflected
in the syntax: anaphoric definites have overt markers in D triggering movement out of the VP.
3.3.9 Summary
In this section, I developed a pseudo-incorporation analysis for indefinite bare nouns in Ch’ol, con-
straining existential closure to the VP, as per Diesing’s (1992) Mapping Hypothesis. Bare nouns
may be indefinite only if they remain in their syntactically base-generated position inside the VP.
Nominals combine with verbs via Restrict and are subsequently existentially closed at the VP
level. Ergative subjects are generated outside the VP in Spec,VoiceP, and therefore are not in the
domain of existential closure. The only option is for external subjects to have a definite interpre-
tation, which is done by assuming that ι can type shift the predicative nominal to an individual.
By adopting the ι type shifter, we can also account for the fact that bare nouns within the VP can
also be definite. Arguments in the complement of VP may be type shifted and combine with the
verb via Functional Application, though certain caveats on this were discussed in Section 3.3.7.
I ended with a discussion of unique and anaphoric definiteness in Ch’ol with consequences for
word order.
The analysis for indefinite internal objects developed here predicts that they can be possible
antecedents for agreement or anaphora. Indeed, this is borne out below in (144). The first line
of (144a) introduces xcholelob ‘farmers’ as the complement of the existential predicate añ. The
next line refers back to the farmers via third person marking (B3), showing that bare nouns are
accessible to anaphoric elements in subsequent discourse.
82
(144) a. Añ abi x-cholel-ob.
EXT REP NC-cornfield
‘There were some farmers.’ Xi’ba.1
b. Ta’=bi sujty-i-Ø-yob tyilel, i’ik’=ix abi.
PFV=REP return-IV-B3-PL DIR:towards black=already REP
‘They were returning at night.’ Xi’ba.2
In the next section, I consider another instance of bare nouns—those that are incorporated
nominals. The nominals in these constructions exhibit a number of syntactic and semantic differ-
ences from bare nouns in argument position. One ofwhich is that, unlike bare nouns as arguments,
incorporated nominals are not accessible for agreement or pronominal elements in subsequent
discourse.
3.4 Incorporated nominals
Nominals may also be incorporated with verb stems in Ch’ol, in a construction referred to as “in-
corporation antipassives” (Gutiérrez Sánchez, 2004; Coon, 2010a; Vázquez Álvarez, 2011; Coon,
2019). First, I provide morphosyntactic background on these constructions. I then give new se-
mantic data displaying evidence that incorporated nominals are number neutral, may not be an-
tecedents for pronominal anaphora, and that incorporation is a productive process in Ch’ol. I then
list desiderata for a semantic analysis of these constructions.
3.4.1 Morphosyntactic properties of incorporation antipassives
Incorporation antipassives with root transitive verbs
As has been discussed by Gutiérrez Sánchez (2004), Coon (2010a) and Vázquez Álvarez (2011),
nouns can be integrated with verbal stems. Coon (2010a) and Vázquez Álvarez (2011) call this
structure an incorporation antipassive. Ch’ol incorporation antipassives are an instance of Mithun
(1984, 849)’s composition by juxtaposition where “the V and the N are simply juxtaposed to form
83
an especially tight bond”. No phonological changes are triggered with this juxtaposition and no
extra morphology appears with the incorporated noun and root transitive verb.16 Examples of
incorporation antipassives are given in Table 3.3 for root transitive stems.17
Table 3.3: Examples of incorporated nominals with transitive verb roots
Incorporation antipassive (V N) Translation
choñ waj ‘sell tortilla’
mel pan ‘make bread’
k’el juñ ‘watch book’ (i.e., study)
boñ otyoty ‘paint house’
mäñ muty ‘buy chicken’
These incorporation antipassives surface as nominals (Coon, 2012). They can appear as an
argument to the light verb cha’l in (145a), as the complement of PPs with nonperfective aspects
(145b), in PPs with predicates of motion (145c), and as the subject of predicate adjectives (145d).
They can also appear in object position with other verbs such as mulañ ‘like’ in (145e). In (145d)
and (145e), the incorporation antipassive occurs with the determiner jiñi, but without jiñi these
constructions remain acceptable.
(145) a. Ta’ k-cha’l-e [NP kuch si’ ].
PFV A1-LV-DTV carry wood
‘I carried wood.’ Lit. ‘I did wood-carrying.’
b. Woli-yety [PP tyi kuch si’ ].
PROG-B2 PREP carry wood
‘You are carrying wood.’
c. Sam-oñ [PP tyi kuch si’ ].
go-B1 PREP carry wood
‘I am going to carry wood.’
16The suffix oñ is used to derive antipassives from root transitives, with the addition of the nominalizer
-el as in mäñ-oñ-el ‘buying’ in (1a). Without this extra morphology, the verb root may not appear as the
complement to the light verb as in (1b).
(1) a. Ta’ k-cha’l-e män-oñ-el. b. * Ta’ k-cha’l-e mäñ.
PFV A1-LV-TV buy-AP-NML PFV A1-LV-TV buy
‘I shopped.’ Intended: ‘I shopped.’
17Derived transitive verbs take the derived antipassive suffix -Vyaj in incorporation antipassive con-
structions, i.e., pejk-ayaj juñ ‘read book’, ts’ujts’-uyaj nene ‘kiss baby’.
84
d. Weñ wokol [DP jiñi kuch si’ ].
very difficult DET carry wood
‘Carrying wood is very difficult.’
e. Ma’añ=ik mi k-mul-añ [DP jiñi kuch si’ ].
NEG=IRR IPFV A1-like-DTV.NML DET carry wood
‘I do not like carrying wood.’
These incorporation antipassives have different syntactic properties from regular transitive
constructions. For instance, observe the transitive construction in (146a) and the incorporated
construction with the light verb in (146b).
(146) a. Ta’ j-kuch-u si’.
PFV A1-carry-TV wood
‘I carried wood.’
b. Ta’ k-cha’l-e [NP kuch si’ ].
PFV A1-LV-DTV carry wood
‘I carried wood.’ Lit. ‘I did wood-carrying.’
In (146a), the verb kuch ‘carry’ has the transitive verbalizing suffix -u on it. This suffix does not
appear in (146b), where kuch is part of the incorporation antipassive structure. Furthermore, in
(146a) the verb kuch is the predicate of the sentence, whereas the incorporation antipassive appears
as the nominal complement to the transitive light verb cha’l.
Unlike with transitive constructions as in (147), the incorporation antipassive object may not
be extracted, either via focus (148a) or wh-movement (148b).
(147) a. Si’i ta’ j-kuch-u ti.
wood PFV A1-carry-TV
‘I carried wood.’
b. Chukii ta’ a-kuch-u ti?
wood PFV A2-carry-TV
‘What did you carry?’
(148) Based on data from Coon (2010a, 66 ex 34)
a. * Si’i ta’ k-cha’l-e [NP kuch ti ].
wood PFV A1-LV-TV carry
Intended ‘I carried wood.’
b. * Chukii ta’ a-cha’l-e [NP kuch ti ]?
what PFV A2-LV-TV carry
Intended: ‘What did you carry?’
85
As discussed in Coon (2010a, 66), incorporated nominals may not appear with determiners
(149a), be set B markers (149b), be proper names (149c), or be a null pro (149d).
(149) Based on Coon (2010a, 66)
a. * Ta’ k-cha’l-e [NP jap jiñi kajpe’ ].
PFV A1-LV-TV drink DET coffee
Intended: ‘I drank the coffee.’
b. * Ta’ k-cha’l-e [NP k’el-ety ].
PFV A1-LV-TV see-B2
Intended: ‘I saw you.’
c. * Ta’ k-cha’l-e [NP k’el x-Maria ].
PFV A1-LV-TV see NC-Maria
Intended: ‘I saw Maria.’
d. * Ta’ k-cha’l-e [NP jap pro ].
PFV A1-LV-TV drink
Intended: ‘I drank it.’
Incorporated nominals are, for the most part, bare nominals. There seems to be variation with
respect to whether incorporated nominals can be modified, also noted by (Coon, 2010a, 67). For
instance, sumuk ‘delicious’ cannot modify waj ‘tortilla’ in (150a), but pulem ‘burnt’ can modify si’
‘wood’ in (150b). Speaker variation is attested—not all speakers readily accept modified incorpo-
rated nominals such as in (150b).
(150) a. * Ta’ k-cha’l-e [NP mel sumuk waj ].
PFV A1-LV-TV make delicious tortilla
‘I did delicious tortilla making.’
b. Ta’ k-cha’l-e [NP kuch pulem si’ ].
PFV A1-LV-TV carry burnt wood
‘I did burnt-wood-carrying’.
Syntactic structure of incorporated nominals
In (151a) the incorporation antipassive is the internal nominal argument of the light verb, as shown
by its structure in (151b). A transitive verb with a NP internal object is given in (152) for compar-
ison. Recall that in addition to appearing as complements to the transitive light verb, incorpora-
tion antipassives also appear with determiners and in subject position of adjectival predicates—
lending further support to their status as nominals.
86
(151) Incorporated antipasstive construction Based on Coon (2010a, 68)
a. Ta’ k-cha’l-e [NP kuch si’ ].
PFV A1-LV-TV carry wood
‘I carried wood.’ Lit. ‘I did wood-carrying.’
b. SSP
SS VoiceP
-e
-DTV
DP Voice’
pro
Voice VP
1PRON
V NP
cha’l
LV kuch si’
carry wood
(152) Transitive verb construction Based on Coon (2010a, 68)
a. Ta’ j-kuch-u [NP si’ ].
PFV A1-carry-TV wood
‘I carried wood.’
b. SSP
SS VoiceP
-u
-TV
DP Voice’
pro
Voice VP
1PRON
V NP
kuch
carry si’
wood
With these syntactic properties in mind, I turn to new data on the semantics of incorporated
nominals.
87
3.4.2 New data: Semantics of incorporated nominals in Ch’ol
Incorporated nominals are number neutral
One property of incorporated nominals that holds crosslinguistically is that they are number neu-
tral (Baker, 1988; Dayal, 2011). This is the case in Ch’ol in (153).
(153) Ta’ k-cha’l-e [NP mel otyoty ].
PFV A1-LV-TV make house
‘I built houses.’ Lit. I did house-buidling.
Speaker comment: We don’t know many houses you worked on. Could be one, but
wouldn’t be surprised if you worked on three.
Incorporated nominals cannot be antecedents for pronominal anaphora
There is crosslinguistic variation as to whether incorporated nouns can be antecedents for sub-
sequent pronominal anaphora or agreement (Mithun, 1986; Baker, 1996). In languages such
as Kalaallisut (Van Geenhoven, 1998; Bittner, 1994), Mohawk (Baker, 1988), and Chamorro
(Chung and Ladusaw, 2004), incorporated nominals may serve as antecedents. An example from
Chamorro (Austronesian) is given in (154), where an incorporated nominal serves as an antecedent
for a pro. In (154), the nominal kareta ‘car’ is incorporated into the ‘have’ construction. It is then
referred back to in the next sentence by the pro, evidence for Chung and Ladusaw (2004) that in-
corprated nominals in Chamorro can introduce antecedents for agreement/pronominal anaphora.
(154) Käda taotao ni gäi-[karetai] ha-diséseha na siña, ha-bendi proi
each person COMP WH[nom].agr-have-car agr-wish.Prog Comp can agr-sell
‘Each person who owns a cari wishes that he could sell iti.’ Chamorro
Chung and Ladusaw (2004, ex. 85a)
On the other hand, Dayal (2011) provides evidence from Hindi that incorporated nouns can-
not be antecedents for pronominal anaphora as in (155). The noun kitaab ‘book’ may not be an
antecedent for the pronoun vo.
88
(155) anu-ne kitaabi paRhii. # voi bahut acchii thii
anu book read-PFV it very good be-PST
‘Anu book-read (read a book). It was very good.’ Hindi (Dayal, 2011, 134)
In Hindi, a full nominal is possible, as in (156), where the anaphoric bare noun laRkii ‘girl’ in
(156b) refers back to the incorporated nominal of (156a).
(156) Hindi: Dayal (2003, 21) with my bolding added
a. anu apne beTe ke-liye laRkii dekh rahii hai
anu self’s son for girl is-looking-at
‘Anu is looking at prospective brides for her son.’
b. vo #uskaa/laRkii-kaa swabhaav jaanaa caahtii hai
she her/girl’s nature to know wants
‘She wants to know #her/the girl’s temperament.’
The Ch’ol data pattern like the Hindi data above in that incorporated nominals cannot serve
as antecedents for agreement. For instance in (157a), the incorporated nominal may not serve as
an antecedent for set B agreement markers on the predicate sumuk ‘delicious’. Nominal anaphora,
however, are possible as in (157b) where jiñi waj ‘the food’ is an instance of an associative anaphor.
(157) a. Ta’ k-cha’l-e mel waji. # Weñ sumuk-Øi proi.
PFV A1-LV-TV make tortilla. very elicious-B3
‘I did tortilla-making. They were delicious.’
b. Ta’ k-cha’l-e mel waj. ✓ Weñ sumuk-Ø jiñi waj.
PFV A1-LV-TV make tortilla. very delicious-B3 DET food
‘I did tortilla-making. The tortillas were delicious.’
Incorporation is productive
Unlike in other languages, in Ch’ol there is no restriction placed on which verbs can incorporate
and which nouns can be incorporated. For instance, in some languages, the actions denoted by
incorporation structures must refer to ‘proto-typical’ events. In the Danish data in (158), while
incorporating ‘pig’ with butcher is fine, incorporating ‘ostrich’ is not.
89
(158) Danish (Dayal 2015, 54 data credited to Line Mikkelsen, p.c.)
a. gris-slagte b. ? struds-slagte
pig-butcher ostrich-butcher
‘pig butcher’ Intended: ‘ostrich butcher’
Dayal (2015) refers to this property as institutionalization/enrichment and likens it to restrictive-
ness. Presumably, there are no ostriches native to Denmark so the butchering of them is not institu-
tionalized. In languages such as Danish and Hindi, incorporation is limited to a certain paradigm
conditioned by proto-typical events.
In Ch’ol, however, no restrictions have been found with respect to the incorporation antipas-
sive structures. All verbs and all nounsmay be incorporating or incorporated, respectively. For in-
stance, mel otyoty ‘build house’ is a commonplace event in (159a), whereas k’el ajiñ ‘see crocodiles’
in (159b) may not be. However, both these nominals may be incorporated.
(159) a. Ta’ k-cha’l-e [NP mel otyoty ].
PFV A1-LV-TV build house
‘I built houses.’
b. Ta’ k-cha’l-e [NP k’el ajiñ ].
PFV A1-LV-TV watch crocodile
‘I looked at crocodiles.’
There is also no animacy restriction for the incorporated nominal. Human-denoting nominals
may also be incorporated, as in (160). For instance, in (160b) the noun xchoñ-waj ‘tortilla seller’ has
been incorporated with the verb k’el ‘watch’. Note that x-choñ-waj itself is an instance of incorpo-
ration of waj ‘tortilla’ with the verb choñ ‘sell’. The prefix x- turns the incorporation antipassive
nominal to mean ‘one who does X’, as per discussion in Chapter 2, Section 2.4.2.
(160) a. Ta’ k-cha’l-e [NP mek’ nene ].
PFV A1-LV-TV hug baby
‘I hugged babies.’
b. Ta’ k-cha’l-e [NP k’el x-choñ waj ].
PFV A1-LV-TV see NC-sell food
‘I watched tortilla sellers.’
90
The lack of restrictions on what can incorporate/be incorporated is further supported by the
fact that nouns borrowed from Spanish may appear in incorporation antipassives, as in (161) for
repeyo ‘plaster’ and komputadora ‘computer (spelling changed to conform to the orthographic con-
ventions of Ch’ol).
(161) Spanish-borrowings are also possible as incorporated nominals
a. Ta’ k-cha’l-e [NP choñ repeyo ].
PFV A1-LV-TV sell SP:plaster
‘I sold plaster.’
b. Ta’ k-cha’l-e [NP tyoj-’es-ayaj komputadora ].
PFV A1-LV-TV straight-CAU-D.AP SP:computer
‘I fixed computers.’
3.4.3 Implications for an analysis
Moving forward, a semantic analysis of incorporation antipassives must capture the three points
in (162).
(162) Empirical generalizations for a semantic analysis of incorporated antipassives
a. Incorporated nominals cannot be antecedents for pronominal anaphora.
b. Incorporation is productive: there are no restrictions on the proto-typicality of events
denoted by the incorporation antipassives.
c. The incorporation antipassive constructions are formally nouns.
Below, I consider previous analyses and how they may capture the points in (162). Dayal’s
(2011) analysis can account for (162a), intuition from Chung and Ladusaw’s (2004) composition
rule can account for (162b). A challenge is that both analyses fall short of predicting that these
forms must surface as nominals (162c). I elaborate below.
Dayal (2011)
Dayal’s (2011) analysis of incorporating verbs may be on the right track in order to capture
the fact that incorporated nominals do not introduce antecedents for agreement or pronominal
91
anaphora.18 Using data from Hindi nominal incorporation, Dayal (2011) provides the following
semantics for incorporating verbs (163a) and regular transitive verbs that take individuals as argu-
ments (163b) in a Neo-Davidsonian framework. Incorporating transitive verbs in (163a) for Dayal
differ from transitive verbs in (163b) in that the incorporating verb in (163a) takes a property, not
an individual, as its first argument. This analysis assumes that some verbs are lexically ambiguous
between incorporating and transitive.
(163) Incorporating and transitive ‘catch’ (Dayal, 2011, 146)
a. Incorporating verbs
~ catchINC-V  = λP〈et〉λyλe[P-catch(e) & Agent(e) = y],
where ∃e[P-catch(e)] 1 iff ∃e′= [catch(e′)&∃x[P(x) & Theme(e) = x]]
b. Standard transitive verb in a Neo-Davidsonian semantics
~ catchTV  = λxλyλe.[catch(e) & Agent(e) = x & Theme(e) = y]
The incorporation semantics in (163a) “takes the property to be a modifier of the event, yield-
ing in effect a sub-type of the event” (Dayal, 2011, 146). That is, the incorporated nominal does
not refer independently; rather it modifies the event, in this case it modifies catching-events. Ad-
ditionally in (163a), the theme argument is suppressed and instead expressed in a presupposition.
This captures the fact that incorporated nominals cannot be antecedents to pronominal anaphora.
In contrast, the transitive verb ‘catch’ in (163b) takes an individual as its first argument, which is
interpreted as the theme of the event.
Directly implementing the analysis from Dayal (2011) would mean that all transitive verbs
in Ch’ol would have the additional lexical semantics in (163a), and thus be lexically ambiguous
between the incorporating verb semantics in (163a) and a transitive verb denotation. This lexi-
cal ambiguity is a welcome result for Hindi, as Hindi has restrictions on the types of verbs and
nominals which may incorporate—i.e., incorporation in Hindi is restricted to proto-typical events.
For Ch’ol, however, there is no restriction on the proto-typicality of incorporation. Proto-typical
events as well as not so proto-typical events may be expressed via incorporation antipassives. A
lexical ambiguity account does not quite do justice to the productivity of incorporation in Ch’ol.
18Dayal (2011) also makes some observations about telicity and incorporation, paralleling observations
by Kratzer (2004) on objective case. Further investigation of telicity and incorporation is needed in Ch’ol to
determine how these incorporation structures relate to telicity.
92
Rather than make all verbs lexically ambiguous, let us consider a Chung and Ladusaw 2004-
style composition rule in the next section.
Chung and Ladusaw (2004)-style analysis for incorporated nominals?
In addition to being used as a way to derive narrow scope indefinites, Restrict
(Chung and Ladusaw, 2004) has also been implemented for incorporated nominals. As discussed
above, under Chung and Ladusaw’s (2004) analysis, the composition rule Restrict combines verbs
and nouns by restricting the event, but without placing any restrictions on what can be incorpo-
rated or which verbs can incorporate. The fact that Restrict may be used to combine any verb
and predicative noun is an advantage—as we saw above, there are no restrictions on the types of
verbs that can incorporate as well as the types of nouns that may be incorporated. However, the
theme argument is then existentially closed—a desirable result for Chung and Ladusaw (2004),
as in Chamorro this type of incorporation allows for the incorporated nominal to be available for
subsequent discourse anaphora. However, as Dayal (2003, 2011) points out for Hindi, and as I
have shown above for Ch’ol, incorporated nominals may not serve as antecedents for anaphoric
elements.
A composition rule, inspired by Restrict, could be developed in order to account for the fact
that incorporated nominals may not serve as antecedents for pronominal anaphora in Ch’ol. Un-
der such an account, the verb and noun combine, but the rule also suppresses the theme argument
of the verb, as Dayal (2011) has done in (163a). This possibility is given in (164).
(164) An incorporation composition rule?
λP〈e〈st〉〉λQ〈et〉λe.[Q-P(e)],
where ∃e[Q-P(e)] = 1 iff ∃e′[P(e′) & ∃x[Q(x) & Theme(e) = x]]
(164) turns Dayal’s (2011) incorporating verb semantics into a composition rule that combines
a verb and noun (rather than making the verb lexically ambiguous) and suppresses the theme
argument, making it inaccessible for certain anaphoric elements in subsequent discourse. This
can capture the desired results of (i) making the incorporated nominals inaccessible to anaphoric
elements; and (ii) the productivity of the incorporation antipassives.
93
On the nominal nature of incorporation antipassives
Yet, both these semantic analyses face challenges accounting for point (162c)—that all incorpora-
tion antipassive structures are nominal. Syntactically, Coon (2019) posits that Ch’ol lacks an agen-
tive intransitive Voice head, which is why these antipassives surface not as verbs but as nouns.
Semantically, it is still necessary to determine why incorporation antipassives must surface as
nominal arguments as well as what their semantic type is. One possibility is that the light verb
cha’l- must take arguments of type 〈st〉, making the incorporated antipassives formally different
from 〈et〉 nominals such as ‘boy’ or ‘corn’. This does not immediately explain, however, the fact
that incorporation antipassives may be modified by a determiner or be in complement position
of other verbs, such as ‘like’ above in (145e). For now, I leave these questions open for future
research.
3.5 Conclusion
I began with the data from (89)–(90), repeated in (165)–(166).
(166) Ta’ i-mäñ-ä otyoty wiñik.
(165) Ta’ jul-i wiñik. PFV A3-buy-TV house man
PFT arrive-IV man
(i) ‘The man bought a/the house.’
‘A/the man arrived.’
(ii) NOT: A man bought a/the house.
I accounted for the possible interpretations of the nouns in (165) and (166) by following a Diesing-
style approach to existential indefinites, namely that existential closure happens at the VP layer.
This allows for complements of verbs to be indefinite but excludes an existential indefinite inter-
pretation for argumentsmoved or generated outside the VP. For indefinite interpretations, Restrict
from Chung and Ladusaw (2004) combines the verb and its internal object together, capitalizing
on the notion that these are instances of pseudo-incorporation.
Bare nouns may also appear in other constructions, such as in incorporation antipassives in
(167).
94
(167) Ta’ k-cha’l-e [NP kuch si’ ].
PFV A3-LV-DTV carry wood
‘I carried wood.’ Lit. I did wood-carrying.
Unlike the bare arguments in (165) and (166), the noun si’ ‘wood’ in (167) does not refer. Incor-
poration antipassives are productive in Ch’ol with no apparent restrictions on which verbs can
incorporate nor restrictions on the type of bare nominals that may be incorporated. I discussed
possible analyses for incorporation such as Dayal’s (2011) incorporating verb semantics and a
Chung and Ladusaw (2004)-style composition rule. Where both these analyses fall short is that
they do not predict that incorporation antipassives always surface as nouns, and not verbs.
The analysis of indefinites in Ch’ol in this chapter has implications for theories of type shifters
available to natural language as well as for analyses for VOS word order. For instance, Partee
(1986) proposed type shifters such as in (168).
(168) Partee 1986 type shifters (nonexhaustive)
a. ι = λP.ιx[P(x)]
b. ∃ = λPλQ.∃x[P(x) ∧ Q(x)]
Chierchia (1998b) and Dayal (2004) proposed a ranking of type shifters, given in (169). Impor-
tantly, ι and ∩, which turns predicates into kind-terms, are ranked higher than ∃.
(169) Ranking of type shifters Dayal (2004)
{ ι,∩ } > ∃
This chapter supports a structural approach to existential indefiniteness (i.e., (Diesing, 1992) and
recent work by Dayal (2013)), in that there does not need to be a ranking of type shifters. Rather,
existential indefinite interpretations are governed by the syntax, with the VP being the domain of
existential closure. That is, rather than the type shifters in (168b) being available to type shift nouns
in any position, the data from Ch’ol provide support that there are certain structural constraints
as to where bare nouns may be interpreted as indefinite.
Finally, as this chapter has argued that the interpretation of nouns is reflected by their struc-
tural positions, analyses of word order should capture this. For instance, bare object nouns in
95
Ch’ol must appear in VOS object position. VOS objects are precisely objects that can receive an
existential indefinite interpretation. Therefore, any analysis for Ch’ol word order should capture
that VOS objects are syntactically within the VP. It should also capture that an existential indefinite
interpretation for constituents moved or generated outside the VP is not possible. This includes
VSO objects and objects non-adjacent to verbs (i.e., in V-XP-(S)-O position). I will further develop
this implication for word order in Chapter 5.
96
CHAPTER 4
MOVEMENT-DERIVED ISLANDS
4.1 Introduction
The previous chapter investigated the semantic properties of VP-internal and VP-external nomi-
nals. In this chapter, I investigate the syntactic properties of nominals that are within the VP in
contrast to those that have moved outside the VP. Specifically, I look into which elements can and
cannot be subextracted from nominal arguments. To preview, Ch’ol exhibits asymmetries in what
is available for subextraction. Interrogative possessors may extract from absolutive subjects in
(170a), but not from absolutive objects, as in (170b). Numerals, on the other hand, may extract
from both absolutive subjects (171a) and absolutive objects (171b).
(170) a. Majkii ta’ yajl-i [ i-wakax ti ] ?
who PFV fall-IV A3-cow
‘Whose cow fell?’
b. * Majkii ta’ a-k’el-e [ i-chich ti ]?
who PFV A2-see-TV A3-sister
Intended: ‘Whose sister did you see?’
(171) a. Cha’-kojtyi ta’ yajl-i [ ti wakax ].
two-CLF PFV fall-IV cow
‘Two cows fell.’
b. Cha’-kojtyi ta’ i-k’el-e [ ti wakax ] aj-Wañ.
two-CLF PFV A3-see-TV cow NC-Juan
‘Juan saw two cows.’
To capture this asymmetry, I provide evidence that objects with overt possessors always move out
of the VP and for this reason subextraction is blocked. This follows from the Freezing Principle
(Ross, 1974; Wexler and Culicover, 1977), or a ban on extraction from a moved constituent. Nu-
meral modifiers in object position do not obligatorily trigger object shift and therefore may extract
from the object. That is, the nominals in (170a) and (171a) are inside the VP, but the object in (170b)
97
has moved outside the VP. I conclude that movement-derived islands block the extraction of ele-
ments from nominal arguments. When extraction is possible, it is not from a moved constituent.
This work has implications for deriving verb-initial word order. That is, analyses that posit
that a verbal projection fronts with the object (e.g., Massam (2000); Chung (2005); Coon (2010b)),
face challenges in capturing that extraction from the object in (171b) is possible. Following from
the Freezing Principle, the moved verbal projection containing the verb and object should be an
island for extraction. However as shown in (171b), numerals may extract from object position, a
surprising fact if it is contained in a moved constituent. I argue that this points to an analysis of
verb-initial word order where VOS objects are not contained in a moved constituent; rather VOS
is base-generated and extraction from objects is possible from their their base position.
This chapter is structured as follows. In Section 4.2, I present the extraction asymmetry facts
previewed in the introduction. I include data on other elements, such as demonstratives and pos-
sessors, which are banned from extracting from any argument position. In Section 4.3, I present
evidence that DP objects in Ch’ol undergo obligatory object shift, with evidence from their se-
mantic interpretation as well as the ordering of constituents. Evidence for object shift comes from
the ordering of DP objects and adverbs as well as certain interpretational differences, building
on the analysis from Chapter 3. New data provide evidence that objects with overt possessors
obligatorily shift, while objects with numerals do not. In Section 4.4, I posit that object shift can
account for the extraction asymmetries of Section 4.2: shifted objects are movement-derived is-
lands and therefore subextraction from them is banned. In Section 4.5, I develop an idea from
Little (2020) on how subextraction is licensed, predicated on Agree-based frameworks for extrac-
tion (e.g., Rackowski and Richards (2005); Van Urk and Richards (2015)). Section 4.6 discusses im-
plications of this chapter’s proposal with respect to phrasal-fronting analyses of verb-initial word
order. I conclude with some data on dialectal variation and subextraction in Section 4.7.
98
4.2 Subextraction data
Subextraction, which I also refer to as left branch extraction1, is when a constituent embedded in a
larger constituent is moved out via A-bar movement. In this chapter, I focus on the subextraction
of elements from nominals. While the subextraction of interrogative possessors has been noted
before for the Tila dialect of Ch’ol (Coon, 2009) and the closely related language Tsotsil (Aissen,
1979, 1987, 1996), this study adds new data on the possibilities for extracting a variety of elements
from different positions in Ch’ol.
4.2.1 Interrogative possessor extraction
In the nominal domain, possessors follow possessees, as in (172a). When the possessor is inter-
rogative, it precedes the possessee, generating the possessor–possessee order in (172b).
(172) a. i-wakax aj-Maria b. majkii i-wakax ti?
A3-cow NC-Maria who A3-cow
‘Maria’s cow’ ‘whose cow?’
I present data on the extraction of interrogative possessors in (172b). The inability of posses-
sors, such as in (172a), to undergo subextraction is discussed further in Section 4.2.3 and Section
4.4.5. For simplicity of exposition, I assume the nominal structures from Aissen (1992) for (172a)
and (172b) in (173a) and (173b), but see Chapter 5 for discussion motivating these structures. Re-
call that the possessive set A markers are derived via local agreement between a low functional
head and the possessor (see discussion in Chapter 2 Section 2.4.5). The interrogative possessor
moves to the edge of the DP, driven by wh-features.2
1All elements that may subextract in Ch’ol are located in a left branch.
2Under a phrasal-fronting analysis proposed by Coon (2010b), the order Possessor-Possessee is first
generated and nP, containing the possessee, fronts to a position above the possessor, between DP and
PossP as in (1). The nominal structure in (1) is also compatible with the analysis of this chapter.
99
(173) a. DP b. DP
D PossP DP j D’
majki
Poss’ DP who D PossP
Poss nP aj-Rosa Poss’ ti
NC-Rosa
i-wakax Poss nP
A3-cow
i-wakax
A3-cow
Interrogative possessorsmay extract from absolutive subjects, as in (174a), where the interrog-
ative possessormajki appears before the verb. The trace ti in the glosses is where the interrogative
possessor originates, in Spec,PossP. The interrogative possessor subextracts from its position in
the left branch of Spec,DP, as schematized in (173b). Pied-piping is also an available option, as
shown by (174b), where the constituent containing the interrogative possessor and its possessee
fronts to the preverbal position. Semantically, subextraction is associated with focus on the ex-
tracted constituent. In (174b), a speaker may be asking whose cow fell, but in (174a), the speaker
may not have properly heard the name of the cow’s owner and be asking for the interlocutor to
repeat it.3
(1) Predicate-fronting account of possessive structures (Coon, 2010b)
a. DP b. DP
D IP DPk D’
majki
nP j I’ who D IP
i-wakax
A3-cow I PossP nP j I’
i-wakax I PossP
DPk Poss’ A3-cow
aj-Rosa Poss t tk Poss’j
NC-Rosa Poss t j
In Chapter 5, I present more details comparing analyses for the nominal domain in Ch’ol.
3With nonperfective aspects, subextraction is possible:
(1) Majkii woli i-yajl-el [ i-wakax ti ]?
who PROG A3-fall-NML A3-cow
‘Whose cow is falling?
As discussed in Chapter 2, the subject majki i-wakax ‘whose cow’ in (1) is structurally a possessor.
100
(174) Absolutive subject
a. Majkii ta’ yajl-i [ i-wakax ti ] ?
who PFV fall-IV A3-cow
‘Whose cow fell?’
b. [ Majki i-wakax ]i ta’ yajl-i ti ?
who A3-cow PFV fall-IV
‘Whose cow fell?’
For some speakers, extraction of an interrogative possessor out of ergative subjects in (175a)
is marginally acceptable, though most speakers find this ungrammatical. Pied-piping is also an
option available to all speakers, as in (175b).4
(175) Ergative subject
a. ??/* Majkii ta’ i-k’el-e-yety [ i-chich ti ] ?
who PFV A3-see-TV-B2 A3-sister
‘Whose sister saw you?’
b. [ Majki i-chich ]i ta’ i-k’el-e-yety ti ?
who A3-sister PFV A3-see-TV-B2
‘Whose sister saw you?’
Extraction of an interrogative possessor from the subject of an unergative verb is grammatical
for some speakers (and slightly better than (175a)), as shown in (176a). Pied-piping, again, is an
available option for all speakers, as in (176b).
(176) Unergative subject
a. % Majkii ta’ i-cha’l-e soñ [ i-chich ti ] ?
who PFV A3-dance-TV dance.NML A3-sister
‘Whose sister danced?’
b. [ Majki i-chich ]i ta’ i-cha’l-e soñ ti ?
who A3-sister PFV A3-dance-TV dance.NML
‘Whose sister danced?’
Extraction of an interrogative possessor from an absolutive object, on the other hand, is un-
grammatical, as in (177a).5 In this case, pied-piping is the only available option as in (177b).
4In the Tila dialect of Ch’ol, extraction from ergative subjects is ungrammatical (Coon, 2009).
5With the applicative suffix -be, possessor extraction from object position is possible, as in (1). Following
Coon and Henderson (2010), (1) is an external possessive construction, indicated by the applicative marker.
For Coon and Henderson (2010), external possessors are located in the specifier of ApplP. Thus, under this
analysis the external possessor in (1) is actually extracted from Spec,ApplP and not from the direct object.
Therefore, it is not an instance of subextraction.
101
(177) Absolutive object
a. * Majkii ta’ a-k’el-e [ i-chich ti ] ?
who PFV A2-see-TV A3-sister
Intended: ‘Whose sister did you see?’
b. [ Majki i-chich ]i ta’ a-k’el-e ti ?
who A3-sister PFV A2-see-TV
‘Whose sister did you see?’
Paralleling the pattern for objects of transitive verbs, extraction of an interrogative possessor
out of the direct object or indirect object of a ditransitive verb is also ungrammatical, as in (178)
and (179).6 In these examples, the interrogative possessor must pied-pipe its possessee.
(178) Ditransitive DO
a. * Majkii ta’ a-choñ-be aj-Rosa [ i-karu ti ] ?
who PFV A2-sell-APPL NC-Rosa A3-car
Intended: ‘Whose car did you sell to Rosa?’
b. * Majkii ta’ a-choñ-be [ i-karu ti ] aj-Rosa?
who PFV A2-sell-APPL A3-car NC-Rosa
Intended: ‘Whose car did you sell to Rosa?’
Only possible meaning: ‘Who did you sell Rosa’s car to?’
(179) Ditransitive IO
a. * Majkii ta’ a-choñ-be karu [ i-chich ti ] ?
who PFV A2-sell-APPL car A3-sister
Intended: ‘Whose sister did you sell a car to?’
b. * Majkii ta’ a-choñ-be [ i-chich ti ] karu ?
who PFV A2-sell-APPL A3-sister car
Intended: ‘Whose sister did you sell a car to?’
Finally, interrogative possessors inside PPs must pied-pipe the entire PP, as in (180a). They
may not extract from PPs, as per (180b).
(1) Majkii ta’ a-k’el-be i-chich ti?
who PFV A2-see-APPL A3-sister
‘Whose sister did you see?’
6I provide two possible word orders for the ditransitive constructions, as multiple overt postverbal
nominals are rare in Mayan languages. It seems that when the direct object is bare, the pragmatically
neutral ordering of postverbal objects is V-DO-IO.
102
(180) PP adjunct
a. [PP Majki tyi y-otyoty ]i ta’ k’oty-i-yety ti ?
who PREP A3-house PFV arrive-IV-B2
‘To whose house did you go?’
b. * Majkii ta’ k’oty-i-yety [PP tyi y-otyoty ti ] ?
who PFV arrive-IV-B2 PREP A3-house
Intended: ‘To whose house did you go?’
In fact, extraction of the object of a preposition is impossible, as in (181).7
(181) * [ Majki y-otyoty ]i ta’ k’oty-i-yety [PP tyi ti ]?
who A3-house PFV arrive-IV-B2 PREP
Intended: ‘Whose house did you go to?’
I follow previous work, which argues that PPs in Ch’ol are adjuncts (Coon, 2013;
Coon and Preminger, 2013), and therefore adjunct islands. This claim is further supported by the
data above, where the extraction of the PP object in (181) and subextraction of the interrogative
possessor in (180b) are banned. Additional support for the islandhood of PPs comes the fact that
PPs are never selected as complements in Ch’ol. For instance, picture-of-NPs are expressed with
possessive phrases—‘a picture of Rosa’ in Ch’ol would be ‘Rosa’s picture’.
4.2.2 Numeral extraction
Before moving on to the numeral extraction data, I provide the structure I adopt for nominals
with numerals, following Bale et al. (2019), in (182). Numerals are adjoined to the nP in a MP
(or measure phrase) projection. The MP contains both the numeral and the numeral’s obligatory
classifier.
7It is also not possible to front the interrogative possessor and preposition, leaving behind the possessee:
(1) *Majki tyi ta’ k’oty-i-yety otyoty ?
who PREP PFV arrive-IV-B2 house
Intended: ‘To whose house did you arrive?’
103
(182) Position of numerals for cha’kojty wakax ‘two cows’
nP
MP nP
num M wakax
cha’- kojty cow
two CLF
The MP constituent, containing the numeral and classifier, may extract from its position ad-
joined to the nP in (182). Extraction of only the numeral or only the classifier is ungrammatical.
Numerals may extract from absolutive subject position in (183a) and absolutive object position
in (183b).8
(183) a. Cha’-kojtyi ta’ yajl-i [ ti wakax ] tyi potreru.
two-CLF PFV fall-IV cow PREP field.
‘Two cows fell in the field.’ Absolutive subject
b. Cha’-kojtyi ta’ i-k’el-e [ ti wakax ] aj-Wañ.
two-CLF PFV A3-see-TV cow NC-Juan
‘Juan saw two cows.’ Absolutive object
When only the numeral extracts, the numeral alone is focused (rather than the whole nominal
argument). For instance, (183a) may be a response to someone who mistakenly said that only one
cow fell.9
As shown for interrogative possessors, extraction out of ergative subjects receives mixed
judgements, as in (184).
(184) % Cha’-kojtyi ta’ i-k’ux-u-yoñ [ ti mis ].
two-CLF PFV A3-eat-TV-B1 cat
‘Two cats bit me.’ Ergative subject
8The subextraction facts are the same for nonperfective aspects and numerals:
(1) a. Cha’-kojtyi woli i-yajl-el [ ti wakax ] tyi potreru.
two-CLF PROG A3-fall-NML cow PREP field.
‘Two cows are falling in the field.’
b. Cha’-kojtyi woli i-k’el [ ti wakax ] aj-Wañ.
two-CLF PROG A3-see cow NC-Juan
‘Juan saw two cows.’ (This can also mean: ‘Juan is looking after two cows.’)
9For reasons of space, I do not give data on other quantifiers, but the numeral in (183) can be substituted
with a quantifier such as pejtyel(el) ‘all’ and the extraction facts remain the same.
104
Numeral extraction is possible from the direct object of a ditransitive verb, as in (185a), but
not from the indirect object (185b). The data in (185a) contrast with what was reported above for
objects of ditransitives with interrogative possessors.
(185) a. Cha’-kojtyi ta’ k-choñ-be [ ti karu ] aj-Wañ.
two-CLF PFV A1-buy-APPL car NC-Juan
‘I sold two cars to Juan.’ Ditransitive DO
b. ?? Cha’-tyikil-obi ta’ k-choñ-be karu [ ti x-’ixik-ob ] .
two-CLF-PL PFV A1-sell-APPL car NC-woman-PL
Intended: ‘I sold cars to twowomen.’ Ditransitive IO
Additional data also reveal a difference between extracting from the direct object and extract-
ing from the indirect object in Ch’ol. In (185b), the classifier usedwith ‘two’ is the human classifier
(-tyikil) so it is only possible to associate it with the indirect object ‘women’ and not ‘car’—but even
then it is still judged to be strange. When both objects take the same classifier, as in (186) with chili
peppers and cats, the extracted numeral unambiguously modifies the direct object ‘chili peppers’.
(186) Cha’-kojty ta’ k-äk’-e ich mis.
two-CLF PFV A1-give-APPL chili cat.
‘I gave two chili peppers to the cat.’
NOT: ‘I gave chili peppers to two cats.’
The example in (186) provides evidence that extracting from the direct object is the only option in
ditransitive constructions. I return to why this may be the case in Section 4.5.10
10While ich ‘chili pepper’ is linearly closer, it is structurally further away, given the structure in (1).
(1) ApplP
Appl’ NP
mis
Appl VP cat
V NP
ak’
give ich
chili
It is therefore not the case that the numeral in (186) can only be associated with the direct object because
it is structurally closer (i.e., the Minimal Link Condition). In fact, given the structure in (1), moving from
the direct object is a longer move.
Evidence for the structure in (1) comes from binding relations between the direct and indirect objects.
The indirect object in (2) may bind into the reflexive direct object.
105
Finally, it is not possible to extract numerals from PPs, as shown by the ungrammatical (187).
I take (187) as further evidence for the islandhood of PPs in Ch’ol.
(187) * Cha’-p’eji ta’ k’oty-i-yoñ [PP tyi ti otyoty ].
two-CLF PFV arrive-IV-B1 PREP house
Intended: ‘I went to two houses.’
4.2.3 Demonstratives and non-interrogative possessors cannot extract
Elements that may not subextract from any argument position are demonstratives and non-
interrogative possessors. Data on demonstratives are given in (188) and data on non-interrogative
possessors are given in (189).
(188) a. * Ixäi ta’ yajli [ ti wakax ].
that PFV fall-IV cow
Intended: ‘That cow fell.’ Absolutive subject
b. * Ixäi ta’ i-k’el-e-yoñ [ ti wiñik ].
that PFV A3-see-TV-B1 man
Intended: ‘That man saw me.’ Ergative subject
c. * Ixäi ta’ k-mäñ-ä [ ti karu ].
that PFV A1-buy car
Intended: ‘I bought that car.’ Absolutive object
d. * Ixäi ta’ k-choñ-be [ ti karu ] aj-Rosa.
that PFV A1-sell-APPL car NC-Rosa
Intended: ‘I sold that car to Rosa.’ Ditransitive DO
e. * Ixäi ta’ k-choñ-be karu [ ti wiñik ].
that PFV A1-sell-APPL car man
Intended: ‘I sold that man a car.’ Ditransitive IO
(2) Ta’ k-päs-b-e i-bä x-Quique tyi espejo.
PFV A1-show-APPL-DTV A3-self NC-Quique PREP SP:mirror
‘I showed Quique himself in the mirror.’
106
(189) a. * aj-Rosai ta’ yajl-i [ i-wakax ti ].
NC-Rosa PFV fall-IV A3-cow
Intended: ‘Rosa’s cow fell.’ Absolutive subject
b. * aj-Rosai ta’ i-k’el-e-yoñ [ i-chich ti ].
NC-Rosa PFV A3-see-TV-B1 A3-sister
Intended: ‘Rosa’s sister saw me.’ Ergative subject
c. * aj-Rosai ta’ j-k’el-e [ i-chich ti ].
NC-Rosa PFV A1-see-TV A3-sister
Intended: ‘I saw Rosa’s sister.’ Absolutive object
d. * aj-Rosai ta’ k-choñ-be-yety [ i-karu ti ].
NC-Rosa PFV A1-sell-TV-B2 A3-car
Intended: ‘I sold you Rosa’s car.’ Ditransitive DO
e. * aj-Rosai ta’ k-choñ-be karu [ i-chich ti ].
NC-Rosa PFV A1-sell-TV car A3-sister
Intended: ‘I sold a car to Rosa’s sister.’ Ditransitive IO
I return to these ungrammatical instances of subextraction in Section 4.4.5.
4.2.4 Summary
The subextraction facts from this section are summarized in Table 4.1. Crucially, extraction of
interrogative possessors is banned from the object position of transitive and ditransitive verbs.
Numerals, on the other hand, can extract out of object position. Extraction from subjects of tran-
sitive verbs and indirect objects is ungrammatical or significantly degraded. The extraction of
demonstratives and possessors is banned from all positions.
Table 4.1: Subextraction summary
Subj of Subj of Obj of DO of IO of
IV TV TV ditrans ditrans
Interrogative possessor ✓ % * * *
Numeral ✓ % ✓ ✓ *
Dem * * * * *
Possessor * * * * *
107
4.3 Evidence for object shift
In this section, I provide diagnostics for object shift from adjunct placement and the semantic
properties of the object, following from the analysis in Chapter 3. I will show below that objects
modified by (predicative) numerals pattern like NP objects; objects with overt possessors pattern
like DP objects. When numerals are quantificational (QPs), they pattern syntactically like DP ob-
jects. I will use data on object shift to account for the asymmetries exhibited between the extraction
of numerals versus interrogative possessors. In short, object shift bleeds subextraction.
4.3.1 Adjunct placement as a diagnostic for object shift
The placement of adjuncts has been used as a test in Germanic languages for object shift (i.e.
Holmberg (1986); Diesing (1992, 1996)) and can provide evidence that the object has moved. As
previously discussed in Coon (2010b), temporal adverbs and prepositional phrases can intervene
between a DP object and verb, generating V-XP-(S)-O order, as in (190a). An adjunct may not
intervene between an NP object and verb, as in (190b).11 The adverb/PP may only appear after
the NP object, as in (190c).
11While temporal adverbs and PPs display similar distributional behavior with respect to NP and DP
objects, other adverbial elements may attach directly to the VP, such as the prefix cha’ ‘again’ in (1).
(1) Tyi i-cha’-boñ-o otyoty jiñi wiñik.
PFV A3-again-paint-TV house DET man
‘The man painted the house again.’ (Coon, 2010b, 373)
As suggested by Coon (2010b, 373), I take the adverb in (1) to be a VP-level adverb. Evidence for this
comes from the fact that (1) means that the house has been painted before, but not necessarily by the same
man, suggesting the adverb does not have scope over the subject. These VP modifiers can have scope over
the object and thus seem to attach lower in the verbal domain than the adverbs and PPs discussed above.
Directionals are another set of elements that seem like adverbs in Ch’ol and may appear between a bare
noun object and verb. However, as Coon (2013) notes, directionals are different from other adverbials, as
they must appear next to the verb, without anything else intervening. Directionals have been reported to
be a closed class of items in a number of Mayan languages (England 1976, Mam; Haviland 1993; Aissen
1994, Tsotsil; Mateo Toledo 2004, Q’anjob’al). Mateo Toledo (2004) argues that directionals in Q’anjob’al are
clitics to the verb. Aissen (2009) proposes that the directionals in Tsotsil are serial verb constructions. For
these reasons, I do not use directionals as a diagnostic for object shift.
108
(190) Data based on Coon (2010b, ex (40) and (41))
a. Ta’ j-k’ux-u ak’bi/tyi k-otyoty [DP jiñi waj ].
PFV A1-eat-TV yesterday/PREP A1-house DET tortilla.
‘I ate the tortilla yesterday/in my house.’12
b. * Ta’ j-k’ux-u ak’bi/tyi k-otyoty [NP waj ].
PFV A1-eat-TV yesterday/PREP A1-house tortilla.
Intended: ‘I ate a tortilla yesterday/in my house.’
c. Ta’ j-k’ux-u [NP waj ] ak’bi/tyi k-otyoty.
PFV A1-eat-TV tortilla yesterday/PREP A1-house
‘I ate a tortilla yesterday/in my house.’
As in (190a) for DP objects, an adjunct may intervene between the verb and an object with an
overt possessor as in (191).
(191) Ta’ j-k’ux-u ak’bi [DP i-waj aj-Eve ].
PFV A1-eat-TV yesterday A3-tortilla NC-Eve
‘I ate Eve’s tortilla yesterday.’
The placement of adjuncts with respect to the object has implications for interpretational dif-
ferences. I discuss data from Chapter 3 below on the interpretation of numerals, which provides
further support for object shift.
4.3.2 Interpretational differences of numerals and adjunct placement
As per data from Chapter 3 Section 3.3.6, numerals and their relative position to the verb lend
support that some objects move out of the VP. When they do move, semantic differences are ob-
served. Adjuncts may appear after an object modified by a numeral and a verb as in (192a). The
adjunct may also intervene between the verb and the object in (192b). The interpretations of the
numerals in (192) differ. Evidence for this comes from their scopal relations with the adverb prefix
ñoj- ‘always’. While (192a) is felicitous in a context in which the speaker sees twomen on the road,
12It is also possible for the adverb to appear after the verb and DP object in (190a), as in (1).
(1) Ta’ j-k’ux-u jiñi waj ak’bi.
PFV A1-eat-TV DET tortilla yesterday.
‘I ate the tortilla yesterday.’
109
but not necessarily the same two men, (192b) is felicitous in a context in which the speaker often
sees two specific men.
(192) Context A: There is a road behind my house. I always see two men walking there, some-
times it is Marcos and Rogelio, sometimes Juan and Eladio and sometimes Jorge and Javier.
Context B: There is a road behind my house. I always see Nicolás and José walking there.
a. Mi k-ñoj-k’el [ cha’-tyikil wiñik ] tyi bij.
IMF A1-always-see two-CLF man PREP way
‘I always see two men on the road.’ ✓A ✓B
b. Mi k-ñoj-k’el tyi bij [ cha’-tyikil wiñik ].
IMF A1-always-see PREP way two-CLF man
‘I sometimes see on the road two men.’ ✗A ✓B
As discussed in Chapter 3, these interpretational differences are predicted by Diesing’s (1992)
Mapping Hypothesis, whereby unshifted objects such as in (192a) can receive an existential inter-
pretation, as they are within the domain of existential closure, (i.e. within the VP). Presupposi-
tional or specific objects as in (192b) shift out of the domain of existential closure (i.e. outside the
VP), indicated by the intervening adverb. In (192b), the numeral is quantificational so must move
out of the VP domain (as discussed in Chapter 3, Section 3.3.6). The interpretational differences in
(192) constitute semantic evidence of object shift, when material intervenes between the verb and
object.
4.3.3 Constituent order
Recall from Chapter 2 that VOS word order is possible with NP objects as in (193), and VSO
appears with DP objects, given below in (194).
(193) VOS = NP objects
a. Ta’ i-k’ux-u [O waj ] [S aj-Maria ].
PFV A3-eat-TV tortilla NC-Maria.
‘Maria ate a tortilla.’
b. * Ta’ i-k’ux-u [S aj-Maria ] [O waj ] .
PFV A3-eat-TV NC-Maria tortilla.
Intended: ‘Maria ate a tortilla.’
110
(194) VSO = DP objects
a. Ta’ i-jul-u [S aj-Ariañ ] [O jiñi me’ ].
PFV A3-shoot-TV NC-Adrian DET deer
‘Adrian shot the deer.’
b. * Ta’ i-jul-u [O jiñi me’ ] [S aj-Ariañ ].
PFV A3-shoot-TV DET deer NC-Adrian
Intended: ‘Adrian shot the deer.’
The data in (195) provide evidence that objects with overt possessorspattern like theDP objects
from above. As seen with objects with determiners, VOS word order is prohibited for objects with
overt possessors, as in (195a). Rather, objects with overt possessors appear in VSO object position
in (195b).
(195) a. * Ta’ i-k’ux-u [O i-waj aj-Wañ ] [S aj-Maria ].
PFV A3-eat-TV A3-tortilla NC-Juan NC-Maria
Intended: ‘Maria ate Juan’s tortilla.’
b. Ta’ i-k’ux-u [S aj-Maria ] [O i-waj aj-Wañ ].
PFV A3-eat-TV NC-Maria A3-tortilla NC-Juan
‘Maria ate Juan’s tortilla.’
A possessed nominal without an overt possessor, on the other hand, may appear in VOS object
position, as in (196).
(196) Ta’ i-boñ-o [O iy-otyoty ] [S wiñik ].
PFV A3-paint-TV A3-house man
‘The man painted his house.’
The possessive structure in (197a) is banned from appearing in VOS object position, but the
structure in (197b), with no overt possessor, is possible in VOS object position.13
13As discussed in Coon and Henderson (2010), the possessor in (197b) could be a minimal pronoun in
the sense of Kratzer (2009), whose features are valued by a nominal which c-commands it.
111
(197) a. Banned in VOS object position
DP
D PossP
Poss’ DP
Poss nP aj-Rosa
NC-Rosa
i-wakax
A3-cow
b. Allowed in VOS object position
PossP
Poss’ DP
Poss nP pro
3PRON
i-wakax
A3-cow
Numerals can appear in VOS object position, as shown in (198a). The example in (198b)
demonstrates that numerals can also occur in VSO object position.
(198) a. Ta’ i-k’ux-u [O cha’-k’ej waj ] [S aj-Rosa ].
PFV A3-eat-TV two-CLF tortilla NC-Rosa
‘Rosa ate two tortillas.’
b. Ta’ i-k’ux-u [S aj-Rosa ] [O cha’-k’ej waj ].
PFV A3-eat-TV NC-Rosa two-CLF tortilla
‘Rosa ate two tortillas.’
Numerals and constituent order will prove to be an important diagnostic for the proposal that
movement of the object blocks subextraction. As I will show below, the numeral may extract from
(198a) but not (198b).
4.3.4 Summary of object restrictions
A summary of object restrictions and constituent order is given in Table 4.2; a summary of the
ordering of objects and adjuncts is given in Table 4.3.
112
Table 4.2: Summary of object restrictions and word order
VOS VSO
Bare object ✓ *
Object with determiner * ✓
Object with overt possessor * ✓
Object with numeral ✓ ✓
Table 4.3: Summary of object restrictions and adverbials
V-XP-(S)-O
Bare object *
Object with determiner ✓
Object with overt possessor ✓
Object with numeral ✓
In Tables 4.2 and 4.3, we see that bare objectsmust appear next to the verb in VOS positionwith
nothing intervening. Objects with determiners appear in VSO position, and can have intervening
material. This pattern holds for objects with possessors and numerals. As detailed in Section
4.3.2, objects modified by numerals are special in that they provide evidence for interpretational
differences of objects in V-XP(-S)-O and V-O-XP(-S) position. When adjacent to the verb, an object
modified by a numeral can receive an existential indefinite interpretation; an intervening adjunct
is indicative that the numeral takes wide scope.
I argue that in the contexts where objects appear in VSO object position, the object has moved
out of the VP. Thus, when the object remains in its position as the complement to the verb in
(199a) subextraction is possible. When it has undergone movement outside the VP, as in (199b),
subextraction from it is not longer possible.
113
(199) a. Subextraction from nominal is possible
. . .
. . . VP
V NP
NOMINAL
b. Subextraction from nominal is not possible
YP
YP DPi
Y . . . NOMINAL
. . . VP
V ti
4.4 Capturing the subextraction facts
4.4.1 Proposal
I argue that object shift, discussed in Section 4.3, freezes the object, creating an island. Elements
within that object are subsequently prohibited from extracting out. This follows from the Freezing
Principle in (200), which bans moving something out of an already moved element (i.e. Ross
(1974); Wexler and Culicover (1977); see also discussion in Corver (2006)).
(200) The Freezing Principle (Wexler and Culicover, 1977, 17)
Moved constituents are islands to extraction.
While the Freezing Principle in (200) may be too strong14—indeed, not all moved constituents
are islands for extraction (see e.g., Konietzko et al. (2018))—object shift has been shown to cre-
ate islands in a number of languages, including English (Chomsky, 1973; Lasnik, 2001), German
14See also discussion on the Revised Extraction Constraint from Diesing (1992, 128).
114
(Diesing, 1992), and Ukrainian (Mykhaylyk, 2010). Here I argue that Ch’ol, too, bans extraction
from shifted objects.
Recall that nominals with overt possessors are banned from VOS object position. As with non-
interrogative possessors, interrogative possessors also trigger object shift before the interrogative
phrase fronts to the preverbal position. It then follows that for interrogative possessors in object
position, the object first moves, creating an island as per the principle in (200). This accounts for
the ungrammaticality of interrogative possessor extraction from object position in (201a).
(201) a. * Majkii ta’ a-k’el-e [ i-chich ti ] ?
who PFV A2-see-TV A3-sister
Intended: ‘Whose sister did you see?’
b. CP
C’
C IP
I SSP
ta’
PFV
SS VoiceP
-e
-TV
VoiceP DPi
✗
DP Voice’ DPk D’
pro Voice VP majki D PossP
2PRON who
V ti ik-chich tk
k’el A3-sister
see
Given the principle in (200), extraction from, not extraction of, the shifted object is banned. As
noted above, the interrogative possessor can pied-pipe the possessee as in (202), which is predicted
on this analysis. As Coon (2010b, 368) notes, the entire object constituent may extract from its
moved position, and indeed there is no ban on the movement of object constituents in Ch’ol.
(202) [ Majki i-chich ]i ta’ a-k’el-e ti ?
who A3-sister PFV A2-see-TV
‘Whose sister did you see?’
115
4.4.2 Numerals may not extract from moved constituents
Further support for the current proposal comes from shifted objects with numeral modifiers. As-
suming that VOS is derived from VSO via object shift (I motivate this further in Chapter 5), the
extractability facts from VOS versus VSO objects can be captured. While objects with overt pos-
sessors always shift, numerals may appear in VOS or VSO object position. It is therefore possible
to test whether extraction of the numeral is grammatical when the object shifts. Extraction of a nu-
meral from the object position is only possible in VOS word order, as shown by (203a). Extraction
of a numeral from a shifted object, indicated by VSOword order, is ungrammatical, as in (203b).15
(203) a. Cha’-kojtyi ta’ i-k’el-e [ ti wakax ] aj-Rosa.
two-CLF PFV A3-see-TV cow NC-Rosa
‘Rosa saw two cows.’
b. * Cha’-kojty ta’ i-k’el-e aj-Rosa [ ti wakax ].
two-CLF PFV A3-see-TV NC-Rosa cow
Intended: ‘Rosa saw two cows.’
15Adjectival extraction may constitute further evidence that extraction from the object is only possible
from VOS objects in (1a), not VSO, as in (1b).
(1) Context: Which car did Rosa buy?
a. Säsäki ta’ i-mäñ-ä [ ti karu ] aj-Rosa.
white PFV A3-buy-TV car NC-Rosa
‘Rosa bought the white car.’ Unshifted absolutive object
b. * Säsäki ta’ i-mäñ-ä aj-Rosa [ ti karu ].
white PFV A3-buyTV NC-Rosa car
Intended: ‘Rosa bought the white car.’ Shifted absolutive object
However, further investigation is needed to determine if the adjectives in (1) are true instances of subex-
traction. Adjectives may not, however, extract from absolutive subjects or ergative subjects (recall that
numerals can extract from absolutive subjects). When adjectives appear discontinuous from nouns in these
positions, they are interpreted as secondary predicates or depictives, as shown in (2).
(2) a. * Pek’ ta’ i-k’el-e-yoñ wiñik.
short PFV A3-see-TV-B1 man
Intended: ‘The short man saw me.’ Ergative subject
OK if means ‘The man saw me as short.’
b. * Säsäk ta’ wejl-i maj x-ch’e’.
white PFV fly-IV DIR NC-bird
Intended: ‘The white bird flew away.’ Absolutive subject
Speaker comment: A translation of this would be ‘the bird flew away white’, however, speak-
ers comment that this does not make sense.
See Vázquez Álvarez (2010) and Martı́nez Cruz (2007) for further discussion on secondary predication,
depictives, and adjectives in Ch’ol. I note this as an issue for future research.
116
While it is impossible to test whether unshifted objects with overt possessors allow extraction
of an interrogative possessor, as they always shift, data from shifted objects with numerals can
provide support for the bleeding effects of object shift.
4.4.3 Moved absolutive subjects block subextraction
Absolutive subjects in Ch’ol originate as complements to the verb. Unlike objects, absolutive sub-
jects with overt possessors do not always move (see also Coon (2010b, 362)). Nevertheless, if they
do move, indicated by the placement of adjuncts, extraction from the subject is ungrammatical.
In (204), the subject has an overt possessor, and the PP adjunct may appear between the verb and
subject, or after the verb and subject.
(204) Ta’ yajl-i (✓ tyi potreru ) i-wakax aj-Rosa (✓ tyi potreru ).
PFV fall-IV PREP field A3-cow NC-Rosa PREP field
‘Rosa’s cow fell in the field.’
In (205), an interrogative possessor may extract when the subject remains next to the verb, as
in (205a). However, when the subject has moved, indicated by the placement of the intervening
PP in (205b), interrogative possessor extraction is blocked.16
(205) a. Majkii ta’ yajl-i [ i-wakax ti ] tyi potreru?
who PFV fall-IV A3-cow PREP field
‘Whose cow fell in the field?’
b. *Majkii ta’ yajl-i tyi potreru [ i-wakax ti ] ?
who PFV fall-IV PREP field A3-cow
Intended: ‘Whose cow fell in the field?’
16Speakers have similar judgements for (204) and (205) if the PP is replaced with the temporal adverb
ak’bi, as in (1) and (2). Extraction is degraded in (2) when the subject has moved.
(1) Ta’ yajl-i (✓ ak’bi ) i-wakax aj-Rosa (✓ ak’bi ).
PFV fall-IV yesterday A3-cow NC-Rosa yesterday
‘Rosa’s cow fell yesterday.’
(2) Majkii ta’ yajl-i (?? ak’bi ) [ ti i-wakax ] (✓ ak’bi )?
who PFV fall-IV yesterday A3-cow yesterday
‘Whose cow fell in the field?’
117
4.4.4 Shifted ditransitive objects block subextraction
Recall that extraction of interrogative possessors is ungrammatical from objects of ditransitive
verbs, but the extraction of numerals is grammatical for direct objects of ditransitive verbs. The
argument for ditransitive objects proceeds in a similar fashion: when the object shifts, extraction
is blocked. I again use adjunct placement as a diagnostic for object shift. While interrogative
words must always front to the preverbal position, it is possible to test object shift with objects
that have overt noninterrogative possessors. Data from direct objects with noninterrogative pos-
sessors reveal that they, too, undergo object shift, as in (206). In (206), the adverb ak’bi ‘yesterday’
comes between the verb and direct object ikaru ajRosa, indicating that the direct object has shifted.
Following the line of argument laid out here, it is correctly predicted that extraction of interroga-
tive possessors from objects of ditransitive verbs is ungrammatical: overt possessors trigger object
shift with ditransitive objects as well, bleeding extraction from object position.
(206) Ta’ k-choñ-be-yety ak’bi [ i-karu aj-Rosa ].
PFV A1-sell-APPL-B2 yesterday A3-car NC-Rosa
‘I sold you Rosa’s car yesterday.’
(207) *Majkii ta’ k-choñ-be-yety ak’bi [ i-karu ti ]?
who PFV A1-sell-APPL-B2 yesterday A3-car t
Intended: ‘Whose car did I sell you yesterday?’
Numerals provide further evidence that object shift bleeds extraction: when an object with
a numeral has shifted, the numeral cannot extract out. When an adverb intervenes between a
ditransitive object and verb in (208a), extraction of the numeral is blocked. When the verb and
object are adjacent, however, extraction may proceed as expected in (208b).
(208) a. * Cha’-kojty ta’ k-choñ-be-yety ak’bi [ ti karu ].
two-CLFi PFV A1-sell-APPL-B2 yesterday car
Intended: ‘I sold you two cars yesterday.’ Shifted DO
b. Cha’-kojty ta’ k-choñ-be-yety [ ti karu ] ak’bi.
two-CLFi PFV A1-sell-APPL-B2 car yesterday.
‘I sold you two cars yesterday.’ Unshifted DO
118
4.4.5 Banning demonstrative and non-interrogative possessor extrac-
tion
Demonstratives and non-interrogative possessors may never extract from any position. I sug-
gest that demonstratives project in Ch’ol17 and that there is a ban on the A-bar extraction of the
demonstrative head (see discussion on movement in, e.g., Travis (1984)).18 Numerals, adjectives,
and interrogative possessors, on the other hand, form their own constituents, and therefore have
the possibility of extracting out.
To rule out possessor extraction, I propose that this is due to the inability of (non-interrogative)
possessors to move to the edge of DP. Under a phase-based theory of movement (Chomsky, 2000,
2001, 2008), in order to extract out of a phase, constituents must move to the edge of that phase.
Adopting the proposal that DPs are phases (Adger, 2003; Svenonius, 2004; Chomsky, 2008; Despić,
2015), a possessor must be able to reach the edge of the DP phase to extract out. To illustrate, I
repeat the syntactic structure of possessives in (209).
(209) a. i-wakax aj-Rosa
A3-cow NC-Rosa
‘Rosa’s cow’
b. DP
D PossP
Poss’ DPk
Poss nP aj-Rosa
NC-Rosa
ik-wakax
A3-cow
Interrogative possessors always appear before their possessees, as in (210a). This involves
additional movement of the interrogative possessor to Spec,DP, as shown in (210b). As the inter-
17See, e.g., Nicolae (2015) for an analysis of weak demonstratives in Romanian as heads.
18The determiner jiñ(i) may appear discontinuous to the noun (Vázquez Álvarez, 2011, 337), but as dis-
cussed by Vázquez Álvarez (2011) and Little and Wiegand (2018), it also acts as a focus marker. In Ch’ol
and many Mayan languages, the preverbal position is the focus position. For the present paper, I follow
Vázquez Álvarez (2011) and Little and Wiegand (2018) in assuming that jiñ(i), when discontinuous from
nouns, is the spell outof focus features in the preverbal focus position, rather than an extracted constituent.
119
rogative possessor is at the edge of the DP phase, it has the possibility to subextract.
(210) a. majkii i-wakax ti?
who A3-cow
‘whose cow?’
b. DP
DP j D’
majki
who D PossP
Poss’ ti
Poss nP
i-wakax
A3-cow
I posit that while wh-features obligatorily trigger movement in (210), other features, such as
focus, do not. It may be expected that if a possessor fronts within the DP, it gives rise to a focus in-
terpretation, paralleling focusmovement in the clausal domain. However, as shown by (211), even
when prosodically stressed, a non-interrogative possessor may never occur before its possessee.
In Ch’ol, the possessor in (211) is unable to move to the edge of the DP phase, evidenced by its
inability to appear prenominally. As it cannot move to the edge of the DP, it cannot subextract.
(211) * aj-Rosai i-wakax ti
NC-Rosa A3-cow
Intended: ‘Rosa’s cow’
Interrogative possessors, on the other hand, do move to the edge of DP, driven by wh-features,
as in (209), and therefore have the ability to extract out. The differences in wh- and focus-driven
movement within the nominal domain have parallels in the verbal domain. While wh-features
always trigger movement within the verbal and nominal domains, focus features do not: in situ
focus seems possible in Ch’ol (Clemens et al., 2017).
120
4.4.6 Summary
Taking VSO word order and the distribution of adjuncts as indications of object shift, we see that
the subextraction asymmetries follow naturally from the restrictions on VOS objects. When ob-
jects shift, they become frozen. For objects with overt interrogative possessors, object shift always
occurs, which means that extraction from the object is always blocked. For objects with numerals,
extraction of the numeral is only blocked if the object has moved. Under my account, demonstra-
tives may not extract from any position due to a ban on the A-bar movement of heads. Possessors
may not subextract, as they cannot move to the edge of the DP to do so. Next, I turn to licensing
subextraction and why subextraction from specifiers is ungrammatical or strongly dispreferred.
4.5 Licensing subextraction
4.5.1 Agree and unlocking
In this section, I propose a way to capture the subextraction freezing effects under Agree-
based approaches to extraction. Under Agree-based analyses of extraction, Agree, in the sense
of Chomsky (2000, 2001), unlocks nominals for extraction (i.e., Rackowski and Richards (2005);
Van Urk and Richards (2015); Branan (2018)). Agreement morphemes in this context are taken to
be the result of an Agree relation between a probe with unvalued φ features and a goal (an NP or
DP) with valued φ features. This is modeled in (212). The probe on X in (212) searches in its do-
main and finds a nominal (WP) with valued φ features. An agreement morpheme is then spelled
out on X.
(212) Agree (Chomsky, 2000, 2001)
XP
X YP
[uvφ]
Y WP
Agree [valφ]
121
Rackowski and Richards (2005) implement an Agree-based theory for extraction in that Agree
in (212) unlocks nominals, making extraction possible. In Ch’ol, Agree takes place between a
probe on Voice and the object in the complement of V, modeled in (213).
(213) Voice licenses and Agrees with the complement of V
VoiceP
Voice VP
V NP
Agree OBJ
Recall that Voice licenses and Agrees with the object (see Chapter 2, Section 2.3.2). This licens-
ing allows the nominal to be unlocked and therefore elements may extract from it.
In Little (2020), I proposed that once the object moves out of the domain of the head that Agrees
and licenses it, it is frozen. In the word order analysis adopted in Little (2020), the object shifts
from vP to a position between the subject and vP, in a projection labeled as Abs(olutive)P. This is
schematized in (214).
(214) Object shift in Coon (2010b) and the analysis adopted in Little (2020)
VoiceP
DP Voice’
SUBJ
Voice AbsP
DPi Abs’
OBJ Abs vP
v VP
V ti
Under the analysis adopted in Little (2020), v in (214) Agrees and licenses the object. By Agree-
ing with the object, it is unlocked for extraction. Once the object moves out of the vP, and conse-
quently outside the domain that licenses it, it is no longer unlocked for extraction.
122
The object shift analysis adopted in this dissertation and further elaborated on in Chapter 5
posits that objects shift to a position above the subject in VoiceP. However, the object is in the same
domain of Voice, the head that Agrees and licenses the object, schematized in (215).
(215) Object movement adopted in this dissertation:
VoiceP
VoiceP DPi
OBJ
Voice’ DP
Voice VP SUBJ
V ti
The idea from Little (2020)—that objects outside of the domain of their licensors are no longer
unlocked for subextraction—can still be maintained by adopting the feature inheritance analysis
from Richards (2007)/Chomsky (2008). Features on a phase head, such as transitive v or Voice, are
inherited by a nonphase head such as V. Voice, under the proposal adopted here, is the locus of a
probe with uninterpretable features. The Agree probe belongs to Voice, but is ultimately valued
on V as these features are then inherited by V, as discussed in Richards (2007, 567). Therefore, if
the object in (215) has moved outside V, it is no longer unlocked for extraction. This also connects
to the ban on extraction from specifiers. The object in (215) is in the specifier of VoiceP after
movement and, as seen above, subextraction from specifiers is banned or dispreferred.19
19This raises some questions regarding extraction from intransitive subjects in nonperfective aspects,
which are analyzed as structural possessors, in the specifier of PossP. These possessors control a covert
internal argument to the intransitive verb, schematized in (1).
(1) Woli [DP ii- [ yajl-el PROi ] ].
PROG A3- fall-NML
‘He is falling.’ Lit. ‘His falling is happening.’
Subextraction is possible from the subject of (1), as noted in footnote 3 of this chapter. It is possible
that the control structure in (1) licenses subextraction from the subject-possessor, but this should be further
explored.
123
4.5.2 Banning subextraction from specifiers
Extraction from an ergative argument or indirect object is mostly banned, as indicated in (216).20
(216) ??/* Majkii ta’ i-k’el-e-yety [ i-chich ti ] ?
who PFV A3-see-TV-B2 A3-sister
‘Whose sister saw you?’
Unlike absolutive subjects and objects, both ergative subjects and indirect objects are introduced
in a specifier position: ergative subjects in Spec,VoiceP and indirect objects in Spec,ApplP. Under
the current proposal, ergative subjects are not prima facie islands for extraction, given the Freezing
Principle in (200). In fact, there is no evidence that subjects of transitive verbs move from their
base position in Spec,VoiceP (Coon, 2010b, 2013, 2017; Clemens and Coon, 2018). Nevertheless,
variation in judgements is unsurprising, given crosslinguistic facts on the variability of extraction
out of transitive subjects (i.e. Chomsky (2008); Polinsky et al. (2013), inter alia).
Recall that set A agreement is the result of a specifier-head configuration, what Legate (2014)
terms ‘inherent agreement’, as described in Chapter 2, Section 2.3.2, schematized in (217). The
configuration in (217) triggers set A agreement markers on the verb.
(217) VoiceP
DPk Voice’
majki i-chich Voice VP
who A3-sister ik- k’el
A3- see V DP
pro
2PRON
20The variability in judgement may connect to processing factors. The subject is possessed, so the logical
parse ofmajki ‘who’ is as the possessor. Similarly, in (1), the numeral appears with the classifier for animals,
making it easier to associate with subject ‘cat’.
(1) % Cha’-kojtyi ta’ i-k’ux-u-yoñ [ ti mis ].
two-CLF PFV A3-eat-TV-B1 cat
‘Two cats bit me.’ Ergative subject
124
If Agree between a probe and a goal only happens with internal arguments, only internal
arguments may be unlocked so that elements may extract out of them. Agreement derived via the
configuration in (217) does not permit the unlocking of that argument so that elementsmay extract
out. Therefore, under this proposal, arguments in specifiers cannot be unlocked for subextraction
because they have not been Agreed with, in the sense of (213) above.
4.6 Implications for theories of word order
As I will discuss in Chapter 5, the data here have implications for the syntax of verb-initial lan-
guages. Firstly, VSO word order must be derived in part via object shift since object shift, I argue,
is crucial in capturing the subextraction asymmetries. Secondly, a potential challenge arises for
theories that posit phrasal-fronting of a constituent containing the verb and object to a higher
position in the clause. A phrasal-fronting account is modeled in (218).
(218) A phrasal-fronting account of VOS (Coon, 2010b)
TP
vP j T’
v VP
T VoiceP
V NP
DP Voice’
OBJ
SUBJ Voice t j
Recall that extraction out of VOS objects is possible as in the example below with the numeral in
(219).
(219) Cha’-kojtyi ta’ i-k’el-e [ ti wakax ] aj-Rosa.
two-CLF PFV A3-see-TV cow NC-Rosa
‘Rosa saw two cows.’
As pointed out by Chung (2005), a moved constituent should constitute an island, given the
Freezing Principle. Thus, the vP in (218) is predicted to be an island for extraction. A question
125
arises as to the structure of (219). If the object cha’kojty wakax ‘two cows’ is in the fronted vP, how
is the numeral able to extract from it?
A simpler solution, which I motivate in the next chapter, is modeled in (220). By base-
generating VOS word order, objects in their VOS position are not within moved constituents,
and extraction from them is possible.
(220) Base-generating VOS
IP
I SSP
SS VoiceP
Voice’ DP
Voice VP SUBJ
V NP
OBJ
4.7 Conclusion
In this chapter, I argued that movement-derived islands capture asymmetries exhibited in what is
available for subextraction. Specifically, overt possessors always trigger object shift, which is why
they may never extract from object position. Numerals, on the other hand, do not obligatorily
trigger object shift and may extract from unmoved objects, but not frommoved ones. This follows
from the Freezing Principle (Ross, 1974; Wexler and Culicover, 1977), or a ban on extracting from
a moved constituent.
The Freezing Principle raises some challenges for phrasal-fronting proposals of VOS word
order. For phrasal-fronting accounts, to derive VOS, a phrase containing the verb and object fronts
to a position higher than the subject. I showed data where extraction from VOS objects is possible.
If the object is inside amoved constituent, howdoes extraction proceed from the object? In the next
126
chapter, I favor a base-generated account, in part because the subextraction facts can be captured
in a more straightforward way.
Finally, while this dissertation focuses on the Tumbalá dialect of Ch’ol, some interesting dif-
ferences arise with the mutually intelligible Tila dialect of Ch’ol. The Tila dialect of Ch’ol is less
permissive in what may extract. For instance, the Tila speakers I consulted find extraction of the
numeral in (221) degraded.
(221) ?/% Cha’-kojty tyi yajl-i [ ti wakax ].
two-CL PFV fall-IV cow
‘Two cows fell.’ Tila Ch’ol
Tila Ch’ol has been reported to have an additional definite determiner li, which Tumbalá
speakers do not use (Vázquez Álvarez, 2011; Vázquez Martı́nez and Little, 2020). This has im-
plications for the left branch extraction implication in (222).21
(222) Left Branch Extraction Implication:
If a language permits left branch extraction, it lacks articles. (Uriagereka, 1988; Corver,
1992; Bošković, 2005)
I suggest that Tila Ch’ol speakers are less likely to find (221) acceptable because the determiner
li is a grammaticalized definite article. This provides an additional data point for the implicational
relationship in (222). Tumbalá Ch’ol allows bare nouns to be definite, as per Chapter 3, but Tila
Ch’ol appears to mark definiteness overtly. Nevertheless, more investigation is needed on the
differences with respect to left branch extraction and definiteness in each dialect to ascertain its
connection with the implication in (222).
21The implication in (222) is formulated as a one-way implication: it does not mean that all languages
without articles allow left branch extraction.
127
CHAPTER 5
A BASE-GENERATEDACCOUNTOF VERB-INITIAL SYNTAX
5.1 Introduction
The conclusions from Chapters 3 and 4 are the following: (i) (in)definiteness is reflected struc-
turally; and (ii) the islandhood of nominals provides evidence for whether they are contained in a
moved constituent. While these results are not unique to Ch’ol, they point to a particular analysis
of verb-initial word order, developed further in this chapter. In Ch’ol, I argue that verb-initial
word order is base-generated with subjects linearized to the right. This captures the observed em-
pirical patterns while also having implications for the derivation of word order in other Mayan
languages. I additionally claim that verb-initial languages do not form one syntactic class. The
diagnostics in this dissertation can thus provide insight into further investigation on how con-
stituents are linearized in a given language.
Mayan languages are generally analyzed as being verb-initial (Norman and Campbell, 1978;
England, 1991; Aissen, 1992; Clemens and Coon, 2018). They fall into three main categories: VOS
word order, fixed VSO word order and VOS/VSO alternating. Ch’ol is in the last category as it
is VOS/VSO alternating (Coon, 2010b; Vázquez Álvarez, 2011). Certain syntactic properties of
the object govern whether the object surfaces in VOS or VSO object position. Specifically, if the
object has material in or above D, it is prohibited from appearing in VOS object position (Coon,
2010b). I refer to VOS objects descriptively as NPs, shown in (223a). If there is material in D, such
as the determiner jiñi in (223b), the object appears in VSO object position. I refer to these objects
descriptively as DP objects.
128
(223) a. VOS objects = NPs
[V Ta’ i-k’ux-u ] [O waj ] [S aj-Rosa ].
PFV A3-eat-TV tortilla NC-Rosa
‘Rosa ate a tortilla.’
b. VSO objects = DPs
[V Ta’ i-k’ux-u ] [S aj-Rosa ] [O jiñi waj ].
PFV A3-eat-TV NC-Rosa DET tortilla
‘Rosa ate the tortilla.’
In this chapter, I favor a base-generated account for VOS syntax and argue that the data from
the previous two chapters can be more simply captured under this analysis. The proposal advo-
cated for is given in (224), where VOSword order in (223a) is the underlying order with the subject
in a right specifier in (224a). VSO word order in (223b) is derived via movement of the object to a
position above the subject (224).
(224) A base-generation account of verb-initial syntax
a. VOS b. VSO
IP IP
I SSP I SSP
SS VoiceP
SS VoiceP
Voice’ DP
VoiceP DP
Voice VP SUBJ
OBJ
V NP Voice’ DP
SUBJ
OBJ Voice VP
V tOBJ
The structures in (224) are updates of the analyses originally proposed by England (1991) and
modeled in Aissen (1992).
This chapter makes a claim on the relationship between VOS and VSOword orders. Under the
analysis motivated here, VOS is the underlying word order from which VSO is derived. VSO is
derived via movement of the object to a higher position above the subject, capturing links between
object movement, properties of VSO objects, and word order.
To motivate the approach in (224), I compare two previous word order analyses: a
129
predicate-fronting analysis from Coon (2010b), which has its roots in other phrasal-fronting ac-
counts (e.g., Kayne 1994; Massam 2000; Aldridge 2002), and a head movement account from
Clemens and Coon (2018) (which has also been developed in other verb-initial languages, e.g.,
Sproat (1985); McCloskey (1991); Ouhalla (1994)). While certain modifications could be made to
these analyses, I argue that the analysis in (224) can account for word order and other syntactic
and semantic properties of objects in Ch’ol in a more parsimonious way.
I begin by presenting core data on NP and DP objects in Ch’ol (Section 5.2). I then present
two recent analyses for the word order contrast in (223). That is, I provide background on Coon’s
(2010b) predicate-fronting analysis and Clemens and Coon’s (2018) head movement account (Sec-
tion 5.3). In Section 5.4, I present additional data, including data from Chapters 3 and 4, and
consider how these two analyses may be modified to account for these data. In Section 5.5, I mo-
tivate the analysis previewed in (224), which can account for the word order differences in (223)
as well as the additional data from Chapters 4 and 5. I argue that the analysis in Section 5.5 can
be extended to ditransitives. Section 5.6 presents a way to formalize rights specifiers through lin-
earization rules based on a suggestion in Otaki et al. (2019). In Section 5.7, I discuss implications
for the structure of Mayan languages with fixed VSO word order as well as syntactic ergativity in
VSOMayan languages. Section 5.8 provides some remarks on future areas of investigation within
the nominal domain across Mayan languages. Section 5.9 summarizes and concludes the chapter.
5.2 Core data and base generating the VOS/VSO distinction
Norman and Campbell (1978) and England (1991) analyze Proto-Mayan basic word order as VOS,
but VSO when the object is marked for definiteness or higher on the animacy hierarchy. Factors
influencing postverbal arguments include phonological weight, specificity, definiteness, and ani-
macy (though see Clemens and Coon (2018) for more nuances, such as animacy being a processing
effect). I review these factors influencing word order for Ch’ol, with a focus on overt markers of
definiteness. At the end of this section, I discuss earlier syntactic work that base-generates VOS
word order (i.e., England 1991; Aissen 1992).
130
5.2.1 VOS objects = NPs
As noted by England (1991) for Mayan languages in general and Coon (2010b) for Ch’ol, two
postverbal arguments are not very common in spontaneous speech1, but they do occur, as in the
VOS examples taken from transcribed narratives in (225). In (225a) the VOS object is bare; in (225b)
the object appears with a possessive prefix, which is co-indexed with the subject.
(225) a. Ta’=bi i-jok’-o-yob [O ixim ] [S a wiñik-ob ].
PFV=REP A3-take-TV-PL corn DET man-PL
‘The men dug into the corn.’ Xi’ba.86
b. Wo i-kuch-ob tye [O i-we’el ] [S a Xi’ba ].
PROG A1-carry-PL DIR:come A3-meat DET demon
‘The xi’bas were carrying their meat.’ Xi’ba.13
Modifiers and numerals may also appear in VOS object position, as shown by (226).
(226) a. Ta’ i-mäñ-ä [O chächäk karu ] [S aj-Rosa ].
PFV A3-buy-TV red car NC-Rosa
‘Rosa bought a red car.’
b. Ta’ i-choñ-o [O cha’-kojty wakax ] [S aj-Rosa ].
PFV A3-buy-TV two-CLF cows NC-Rosa
‘Rosa sold two cows.’
VSO word order with NP objects is not possible, as in (227a), cf. (227b).
(227) a. * [V Ta’ i-k’ux-u ] [S aj-Rosa ] [O waj ].
PFV A3-eat-TV NC-Rosa tortilla
Intended: ‘Rosa ate a tortilla.’
b. [V Ta’ i-k’ux-u ] [O waj ] [S aj-Rosa ].
PFV A3-eat-TV tortilla NC-Rosa
‘Rosa ate a tortilla.’
5.2.2 VSO objects = DPs
Objects with determiners are banned from appearing in VOS object position, and instead surface
in VSO object position in (228a). I refer to these objects descriptively as DP objects.
1See also Brody (1984) for some discussion surrounding conceptual problems with the notion of basic
word order.
131
(228) a. [V Ta’ i-k’ux-u ] [S aj-Rosa ] [O jiñi waj ].
PFV A3-eat-TV NC-Rosa DET tortilla
‘Rosa ate the tortilla.’
b. * [V Ta’ i-k’ux-u ] [O jiñi waj ] [S aj-Rosa ].
PFV A3-eat-TV DET tortilla NC-Rosa
Intended: ‘Rosa ate the tortilla.’
Proper names are also banned from appearing in VOS object position (229).
(229) * Ta’ i-mek’-e [O x-Wañ ] [S jiñi aläl ].
PFV A3-see-TV NC-Juan DET boy
Intended: ‘The child hugged Juan.’2
As also demonstrated in Chapter 4, overt possessors are not possible in VOS object position,
as in (230). They must appear in VSO object position (230b), patterning like the DPs above.
(230) a. * Ta’ i-k’ux-u [O i-waj x-Wañ ] [S aj-Maria ].
PFV A3-eat-TV A3-tortilla NC-Juan NC-Maria
Intended: ‘Maria ate Juan’s tortilla.’
b. Ta’ i-k’ux-u [S aj-Maria ] [O i-waj x-Wañ ].
PFV A3-eat-TV NC-Maria A3-tortilla NC-Juan
‘Maria ate Juan’s tortilla.’
5.2.3 Other factors influencing word order
As noted by Norman and Campbell (1978), England (1991), Aissen (1992), and Vázquez Álvarez
(2011), other factors that influence the order of postverbal arguments are animacy and phonolog-
ical weight. For instance, Vázquez Álvarez (2011) notes that in (231) for Ch’ol, postverbal argu-
ments higher on the animacy hierarchy are more likely to be interpreted as subjects. This accounts
for the strangeness of (231). A woman is higher in animacy than a snake, but is in object position,
so the sentence in (231) sounds as if the woman bit the snake.
(231) * Tyi i-k’ux-u [O x-ixik ] [S lukum ].
PFV A3-bite-TV NC-woman snake
Intended: ‘The snake bit the woman.’ Tila Ch’ol (Vázquez Álvarez, 2011, 22)
2This sentence is grammatical with the meaning ‘Juan hugged the child.’
132
To express the intended meaning of (231), a focus construction is used where the subject is
fronted to the preverbal focus position in (232).
(232) [S Jiñ lukum ]i tyi i-k’ux-u [O x-ixik ] ti.
FOC snake PFV A3-bite-TV NC-woman
‘The snake bit the woman.’ Tila Ch’ol Vázquez Álvarez (2011, 22) with my bracketing
As noted by Clemens and Coon (2018, 247), phonological weight of the NP can influence word
order, such as in (233) where the object is an NP in VSO object position but is also modified by the
relative clause.
(233) Ta’ i-mäñ-ä [S aj-Rosa ] [O karu ta’=bä k-choñ-o ].
PFV A3-buy-TV NC-Rosa car PFV=REL A1-sell-TV
‘Rosa bought the car that I sold.’3
5.2.4 Word order restrictions to be captured
The word order restrictions that I will be focusing on are summarized in Table 5.1.
Table 5.1: Ch’ol object restrictions (Coon, 2010b, 363)
VOS VSO
NP ✓ *
DP * ✓
The restriction on VOS objects is also captured in (234).
(234) Nominals with material in or above D0 are banned from VOS object position in Ch’ol.
(Coon, 2010b, 361)
The restriction in (234) is syntactic in nature, though see Chapter 3, Section 3.3.7 for more on
the relationship between DPs and the dimension of definiteness they express. Demonstratives,
3Some preliminary consultations with speakers suggest that the interpretation of the object must be
definite. This relates to the fact that it has undergone movement, however, further investigation is needed
to determine if it is related to its movement-derived position or properties related to relative clauses in
Ch’ol.
133
overt possessors, determiners, and proper names are rejected in VOS object position. This is not
necessarily a semantic restriction—bare nouns may receive definite interpretations in VOS object
position, as demonstrated in Chapter 3. However, as presented in Chapters 3 and 4, the position of
the object does have an effect on the available semantic interpretations. I will discuss this further
in Section 5.4.
5.2.5 Base-generating verb-initial syntax in Mayan
Early analyses of Mayan word order include England (1991) and Aissen (1992), developed on
the intuition that VOS is base-generated and VSO is derived from VOS by reordering the object
to the right of the subject. England (1991) posits the base word order in (235). Topicalized and
focus-marked constituents are preverbal.
(235) TOPIC FOCUS [ V O S ] REORDERED O England (1991, 484)
Aissen (1992) posits the syntactic structures in (236) for Tsotsil, where subjects are in right
specifiers. VOS word order is base generated in (236a) and VSO is derived by moving the object
to a position above the subject in the specifier of XP as in (236b). The structures from Aissen (1992)
are updated with current assumptions about phrase structure.4
4The original trees posited by Aissen (1992), under a Government and Binding framework (i.e.,
Chomsky (1981)), are given in (1), where the subject is in a right specifier, generated VP-internally.
(1) a. VOS b. VSO
IP IP
I VP
I XP
V’ SUBJ
V OBJ X’ OBJ
X VP
V’ SUBJ
V tOBJ
134
(236) a. IP b. IP
I VoiceP
I XP
SUBJ
Voice VP OBJ
V OBJ X VoiceP
SUBJ
Voice VP
V ti
Aissen (1992) posits that specifiers of lexical categories are to the right (i.e., V, N and Adj) and
functional specifiers are to the left. Subjects are therefore generated in right specifiers, but the
specifiers of CPs and DPs are to the left. Left specifiers of CPs and DPs are the landing site for wh-
or focus-marked constituents. Ultimately, I will support an analysis akin to (236), with updates
to account for the ordering of morphemes. For now, I set aside a base-generated analysis and in
the next section focus on a phrasal-fronting and a head movement account for verb-initial word
order.
5.3 Phrasal-fronting and head movement accounts of VOS/VSO
Two recent accounts for VOS/VSO word order in Ch’ol include a predicate-fronting account
(Coon, 2010b) and a head movement account from Clemens and Coon (2018). Both these analy-
ses derive VOS by first base-generating SVO. I review how each account captures the word order
restrictions from Section 5.2.4, repeated in Table 5.2.
Table 5.2: Ch’ol object restrictions (Coon, 2010b, 363) (repeated)
VOS VSO
NP ✓ *
DP * ✓
135
5.3.1 Coon’s (2010b) predicate-fronting account
A phrasal-fronting syntax has been pursued in Austronesian languages (Aldridge, 2002; Chung,
2006; Massam, 2000), as well as Mayan (Coon, 2010b). Here I provide background on the analysis
from Coon (2010b), who derives Ch’ol word order.
Coon (2010b) derives VOS word order by fronting the vP containing the verb and object to
Spec,TP. Coon (2010b, 355) posits that strong agreement features on T require the verb to move
overtly to T. This is modeled in (237). TP appears here, in order to remain faithful to the original
work, however TP is what I label IP in this dissertation.
(237) Deriving VOS in a predicate-fronting account (Coon, 2010b)
TP
vP j T’
v VP
T VoiceP
V NP
DP Voice’
OBJ
SUBJ Voice t j
To derive VSO word order with DP objects, the DP object first shifts out of the vP to a position
Coon labels Abs(olutive)P, following Massam (2000); then the vP fronts to Spec,TP. Coon’s VSO
account is modeled in (238).
136
(238) Deriving VSO in a predicate-fronting account (Coon, 2010b)
TP
vP j T’
v VP
T VoiceP
V ti
DP Voice’
SUBJ
Voice AbsP
DPi Abs’
OBJ Abs t j
5.3.2 Clemens and Coon’s (2018) head movement account
Head movement accounts of verb-initial syntax have been pursued in a number of languages,
i.e., Afro-Asiatic languages (Fassi Fehri, 1993; Ouhalla, 1994) and Celtic languages (Sproat, 1985;
McCloskey, 1991). Clemens (2014, 2019) provides such an analysis in Niuean (Austronesian), and
Clemens and Coon (2018) apply this account to Mayan. For Clemens and Coon (2018), VSO word
order is achieved by base-generating SVO word order then moving the verb via head movement
to a projection above the subject. This is modeled in (239).
(239) Clemens and Coon (2018, 241)
IP
I SSP
TAM
SS Voice
Voice SS SUBJ
tvoice VP
V Voice
tV OBJ
VOS word order is achieved by postsyntactic reordering of the object in order tomeet prosodic
137
well-formedness constraints. These constraints reorder NP objects to be pronouncedwith the verb
that selects them. The prosodic phrasings of VSO and VOS in Ch’ol are given in (240).
(240) Prosodic phrasings of VSO and VOS clauses in Ch’ol Clemens and Coon (2018, 252)
a. (V)ϕ(S)ϕ(O)ϕ (DP Object)
b. (V O)ϕ (S)ϕ (NP Object)
The prosodic analysis from Clemens and Coon (2018) is based onMatch Theory (Selkirk, 2011)
given in (241), under an Optimality Theoretic framework (Prince and Smolensky, 1993). Match
constraints, according toMcCarthy and Prince (1995), match syntactic input with prosodic output.
(241) Selkirk (2011) via Clemens and Coon (2018, 256)
a. MATCH-ι: CPs/IPs correspond to intonational phrases (ι-phrases) and vice versa.
b. MATCH-ϕ: Syntactic XPs correspond to phonological phrases (ϕ-phrases) and vice
versa.
c. MATCH-ř: Syntactic X0s correspond to prosodic words (prosodic-řs) and vice versa.
Important for deriving the VOS/VSO distinction is MATCH (ϕ,XP), which matches an XP to a
ϕ-phrase. The constraints in (241) force isomorphic syntactic and prosodic structure. That is, if the
syntactic structure outputs SVO, the prosody favors it to be pronounced as SVO.
Recall, however, for Clemens and Coon (2018) the verbmoves to SSP via headmovement, gen-
erating VSO order in the syntax. Clemens and Coon (2018) posit another constraint, outranking
MATCH, (ϕ,XP), which accounts for the fact that the verb and object are both pronounced as one
ϕ-phrase. This constraint, ARG-ϕ in (242), is highly ranked, capturing that the syntactic input and
the prosodic output of VOS sentences are not isomorphic.
(242) Argument condition on phonological phrasing (ARG-ϕ): A head H0 with a categorial fea-
ture [C] and head C0 with the same [C] feature must constitute a ϕ-phrase.
(Clemens and Coon 2018, 258)
The constraint in (242) says that heads should be phrased together with their arguments, and
can force a nonisomorphic syntax and prosody. The constraint in (242) applies to NP objects;
DPs, for Clemens and Coon (2018), are phases and spelled out before (242) applies. The argument
condition on phonological phrases connects the verb and object via C-selection. This way the verb
138
and its argument can interact after the verb undergoes head movement. For Clemens and Coon
(2018), the syntactic input and subsequent prosodic output of VOS order is given in (243).
(243) Syntactic input and prosodic output of VOS order5 (Clemens and Coon, 2018, 258)
IP ! ι
ϕ ϕ
I SSP
TAM
ř ř SUBJDP
V+Voice+SS VoiceP TAM=V-SS OBJNP
SUBJDP
tVoice VP
tV OBJNP
The table in (244) demonstrates how the order of constraints and violation of ARG-ϕ results in
a VOS ordering rather than VSO for NP objects. ARG-ϕ shifts the NP to be pronounced with the
verb. An isomporphic VSO ordering is dispreferred, as the verb and its argument are not phrased
together.6
(244) Clemens and Coon (2018, 258)
[vP Verb [VoiceP [DP Subject ][VP [NP Object ]]]] ARG-ϕ MATCH (ϕ,XP) MATCH (XP,ϕ)
a. (Verb (Subject)ϕ (Object)ϕ)ι ∗!
 b. ((Verb Object)ϕ(Subject)ϕ)ι ∗ ∗
Now, ARG-ϕ only applies to NP arguments as DP objects are phases, and sent to spell out
before the constraint in (242) applies. VOS ordering is generated in (244) as the verb V categorizes
for a nominal [N] feature. ARG-ϕ therefore applies as the head V categorizes for the nominal
5While Voice is represented in the syntactic input, it is not always computed in the output if there is
no overt Voice morpheme. Overt manifestations of Voice, of course, would not be deleted in the prosodic
output. Examples of overt manifestations of Voice are causative morphemes, discussed further below.
6An additional requirement is the need for the verb and object to satisfy STRONG START, in (1).
(1) STRONG START (Selkirk, 2011): A prosodic constituent optimally begins with a leftmost daughter
constituent that is not lower in the prosodic hierarchy than the constituent that immediately follows.
STRONG START ensures that the verb and object are phrased together as one prosodic constituent, as in
(Verb Object)ϕ and not (Verb(Object)ϕ)ϕ. The phrasing of (Verb(Object)ϕ)ϕ violates STRONG START because
it contains two prosodic subconstituents (Clemens and Coon, 2018, 259). The leftmost daughter ends up
being lower in the prosodic hierarchy—the verb is phrased as a prosodic word (ř), ranked lower than a
prosodic constituent (ϕ), violating (1).
139
feature, constituting a ϕ-phrase, which is why the object is pronouncedwith the verb in the output
in (243). NP objects, unlike DP objects, are not phases, therefore according to Clemens and Coon
(2018, 261) “both instantiations of the categorial feature relevant to ARGUMENT-ϕ (one associated
with the object and one with the verb root) are visible to PF during the same spell-out cycle.” If
the object is a DP object, the categorial feature is not visible to ARG-ϕ, so prosodic reordering does
not apply.
The prosodic ordering of a VSO sentence with a DP object under the head movement analysis
in Clemens and Coon (2018) is given in (245). In contrast to (243), as ARG-ϕ does not apply to DPs,
the DP object remains in its base position as the complement to the verb and therefore VSO is the
linear output.
(245) Syntactic input and prosodic output of VSO order
IP ! ι
I SSP ř ϕ
TAM=V-SS
TAM
ϕ ϕ
V+Voice+SS VoiceP SUBJDP OBJDP
SUBJDP
tVoice VP
tV OBJDP
5.3.3 Summary
Under the predicate-fronting analysis of Coon (2010b), the vP fronts to Spec,TP to satisfy strong
agreement features on T, generating VOS word order in the syntax. For Clemens and Coon (2018),
VOS word order is derived first via head movement of the verb to a position above the subject,
generating VSO. Then via prosodic constraints, an NP object is reordered to be pronounced with
the verb, generating VOS in the prosody.
One clear difference between the predicate-fronting account and head movement account is
140
how VSO is derived. For the predicate-fronting account, VSO is derived in the syntax via object
shift before the vP fronts. Under a head movement account, VSO objects remain in situ in the VP.
In the next section, we will see that both accounts can capture different data in Ch’ol but need to
be modified in order to explain data from the previous chapters. These differences also bring into
question the link between object shift and its effects on word order. A predicate-fronting account
derives VSO word order in part via shifting the object. While it is possible to add object shift to a
head movement account, this movement does not affect the order of constituents.
5.4 Additional empirical considerations
In this section, I consider data on morpheme order, nominalized complements and properties of
VSO objects in light of the two analyses presented above. I highlight the link between object shift
and its effects on word order. Clemens and Coon’s (2018) account can be modified to accommo-
date object shift, but this movement is vacuous.
5.4.1 Order of morphemes
I begin with how each approach derives the morphology of transitive verb stems. Perfective
verbs take a verbalizing suffix, in v for Coon (2010b), and labeled as a status suffix (SS) in
Clemens and Coon (2018). For root transitive verbs, this is a harmonic vowel suffix, assimilat-
ing to the vowel of the verbal root. For instance in (246a) the verbal root is kuch ‘carry’ and the
verbal suffix is -u, as in (246b); the verb mek’ ‘hug’ appears with the harmonic vowel verbalizer -e.
(246) Transitive verbs
a. Ta’ j-kuch-u si’.
PFV A1-carry-TV wood
‘I carried wood.’
b. Ta’ i-mek’-e nene aj-Eve.
PFV A3-hug-TV baby NC-Eve
‘Eve hugged a baby.’
141
How each account derives the order of morphemes for the sentence in (246a) is modeled in
(247).7
(247) a. Predicate-fronting
TP
vP T’
v VP T VoiceP
V v tV NP
DP Voice’
j-kuch -u
A1-carry si’TV
wood pro
Voice tvP
1PRON
b. Head movement
IP
I SSP
ta’
PFV
SS Voice
Voice SS DP
-u tVoice VP
V Voice -TV pro
j-kuch 1PRON tV NP
A1-carry
si’
wood
Both the predicate-fronting and head movement analyses can account for the placement of the
transitive verbalizer -u. In the predicate-fronting analysis the vP, containing the verbalizer, fronts
to Spec,TP and the verb undergoes head movement to v. In the head movement analysis, the verb
moves through Voice (in this case there is no overt realization of Voice) and then to the verbalizer
in SSP.
The placement of the aspect marker in (246a) raises a challenge for the predicate-fronting but
not head movement account, as noted by Clemens and Coon (2018). If the vP occupies Spec,TP
7For Coon (2013, 54), under the predicate-fronting account, Voice and the subject enter into an agree-
ment relation and the features on Voice are inherited by v, and consequently spelled out on the vP.
142
in (250a), it is unclear where the aspect marker appears8—it cannot be in T as that is below the
landing site of the verb. Under a headmovement account, on the other hand, this is not a problem:
the aspect marker is in IP, above the landing site of the verb.
Another challenge to the predicate-fronting account, pointed out by Clemens and Coon (2018),
is what happens when voice morphemes, such as causatives, are present, as in (248).
(248) Ta’ k-ju’b-es-a pok’.
PFV A3-descend-CAU-DTV bowl
‘I took the bowl down.’
The morphological structure of the verb in (248) is given in (249).
(249) Morphological template of (248)
k- ju’b -es -a
A1- descend -CAU -DTV
Agreement V Voice SS
As the verbalizer -a is below VoiceP for Coon (2010b), it is unclear what maximal projection
fronts. Causatives are associated with Voice and if we place Ch’ol’s causative suffix in Voice, how
do we derive the correct order of morphemes in (249)? As seen in the tree in (250a), the vP fronts
to Spec,TP without the Voice suffix. There is no maximal projection in (250a) that contains both
the verb, object and Voice morpheme to the exclusion of the subject. The head movement account
from Clemens and Coon (2018), on the other hand, can capture the correct order of morphemes as
the verb moves through Voice to the status suffix in (250b).
(250) a. Predicate-fronting ✗
TP
vP T’
v VP T VoiceP
V v tV NP
ki-ju’b -a
DPi Voice’
descend TV pok’
bowl pro Voice tvP
1.PRON -es
CAU
8Coon (2010b) suggests a possible solution where a low Asp(ect)P rather than vP fronts to Spec,TP.
143
b. Head movement ✓ based on the structure from Clemens and Coon (2018, 241)
IP
I SSP
ta’
PFV
SS VoiceP
Voice SS DPi
-a tVoice VP
V Voice pro
ki-ju’b -es 1.PRON tV NP
descend CAU
pok’
bowl
While the order of morphemes is not a problem for the head movement account, it poses chal-
lenges for a predicate-fronting account. As suggested in Coon (2010b), a possible way to derive
the correct order under a predicate-fronting account would be to assume that some morphemes
attach via a Morphological Merger from (Marantz, 1984, 1988). However, the canonical use of
Morphological Merger is to account for second position clitic effects and it is often applied to
lexical heads (Embick and Noyer, 2001). In this instance, Voice is a functional head, so further in-
vestigation would be needed to motivate a Morphological Merger-based account of the causative
suffix in Ch’ol.
5.4.2 Nonperfectives
While the head movement account can derive the order of morphemes, a modification is needed
so that it can capture nominalized complements, discussed in this section. Ch’ol is an ergative–
absolutive language; however, as Coon (2010a, 2012, 2013) has demonstrated, nonperfective as-
pects behave differently. For instance the subject of the intransitive verb is marked with the set B
first person suffix -(y)oñ in (251). The nonperfective forms in (252) mark the intransitive subject
with the first person marker k-.
(251) Ta’ majl-i-yoñ.
PFV go-IV-B1
‘I left.’
144
(252) a. Mi k-majl-el. b. Woli k-majl-el.
IPFV A1-go-NML PROG A1-go-NML
‘I go.’ ‘I am leaving.’
Coon (2010a, 2012, 2013) argues the verbs in (252) are in fact nominalized complements to
an aspectual predicate and the set A markers are possessive markers, as discussed in Chapter 2,
Section 2.3.5. The aspect markers mi and woli differ from the perfective aspect marker ta’ in that
they are predicates and take nominal complements. That is, kmajlel in (252) translates to ‘my going’
and is the complement to the nonperfective aspect markers in (253). Additionally, the verb root
majl has the nominalizing suffix -el and possessive marker k- on it.
(253) Woli [DP ki- [ majl-el PROi ] ].
PROG A1- go-NML
‘I am going.’ Lit. ‘My going is happening.’
The structure of (253) is provided in (254) from Coon (2013).
(254) VP
V DP
woli
PROG
D PossP
DPi Poss’
pro
Poss nP
1PRON
k-majl-el PROi
When possessor subjects are overt in nominalized complements they appear after the verb.
This parallels possessive phrases, where possessors also appear after their possessees, as shown
in (255).
(255) a. i-majl-el aj-Rosa b. i-wakax aj-Rosa
A3-go-NML NC-Rosa A3-cow NC-Rosa
‘Rosa’s going’ ‘Rosa’s cow’
145
Nominalized transitive clauses have a similar structure. The verb is nominalized and the sub-
ject is structurally a possessor, co-indexed with a PRO in the nominalized complement. For tran-
sitive verbs with nonperfective aspects, morphological differences also provide evidence for their
nominalized status. Compare the transitive verb with the perfective aspect in (256b) and the tran-
sitive verb with the progressive aspect in (256a). In (256b) the verbalizer -u, which is a harmonic
vowel suffix with the verb juch’ ‘grind’, appears. In (256a) with the progressive aspect, -u does not
appear—the dependent clause marker -e’ may optionally surface.
(256) a. Woli i-juch’(-e’) sa’ aj-Rosa.
PROG A3-grind-DEP masa NC-Rosa
‘Rosa is grinding masa.’
b. Ta’ i-juch’-u sa’ aj-Rosa.
PFV A3-grind-TV masa NC-Rosa
‘Rosa ground masa.’
The structure of (256a) according to Coon (2010a, 2012, 2013) is the following. The subject,
Rosa, is expressed as the possessor in Spec,PossP and is co-indexedwith a PRO in the nominalized
complement, as in (257).
(257) DP
D PossP
DPk Poss’
aj-Rosa
Poss nP
PROk i-juch’ sa’
The structure in (257), however, does not capture the correct word order, where aj-Rosa appears
after the nominalized verb.
Under a predicate-fronting analysis, Coon (2010b) accounts for the word order of VO-
Possessor by fronting the nP to a position between DP and PossP (labeled as IP, following work
on the nominal domain by Sobin (2002)). This parallels vP fronting in the verbal domain and cap-
146
tures the correct word order of nominalized complements, as the subject (structurally a possessor)
appears after the verb and object.
(258) Predicate-fronting account of nominalized complements based on (Coon, 2013, 145)
DP
D IP
nP j I’
PRO i -juch’ sa’ I PossPk k
A3-grind masa
DPk Poss’
aj-Rosa Poss t j
While Clemens and Coon (2018) do not consider nominalized complements, their analysis
could be extended as follows. The verb moves to a position above the possessor, landing at INFL,
below DP, as modeled in (259).
(259) Extension of head movement account for nominalized complements
DP
D IP
I’ PossP
I I DPk Poss’
Poss Poss aj-Rosa tPoss nP
Voice n
tn Voice
V Voice
i-juch’ PROk
A3-grind tVoice VP
tV NP
sa’
masa
To achieve the correct word order of VOS, ARG-ϕ applies and the object sa’ ‘masa’ is re-
ordered to be pronounced with its complement, the verb. My extension of the account from
Clemens and Coon (2018) parallels their account of the clausal domain; in (259), the verb moves
via headmovement to a high positionwithin the DP.With this modification, it is possible to derive
147
the correct order within the nominal domain. It is unclear, however, what could be the motivating
factor for movement to the IP projection within the nominal domain. There are no overt mor-
phemes corresponding to IP in the DP domain.9 In the clausal domain, the verb moves to the
SS head, which is spelled out as the verbal status suffix. A similar issue arises for the predicate-
fronting account—IP is needed as a landing site for nP in (258), but no morphemes or other syn-
tactic properties are associated with it. While these two proposals can account for the word order
in nominalized complements, some questions remain, such as if the projection between DP and
PossP is really needed. In the analysis supported in Section 5.5, there is no need for an additional
projection between DP and PossP.
5.4.3 Semantic and syntactic properties of objects (Chapters 3 and 4)
I now compare how each analysis fares with respect to the data on the interpretation of objects
from Chapter 3 as well as the subextraction asymmetries from Chapter 4. To begin, I take as a
basis for my analysis the proposal from Chapter 3 that existential closure happens at the VP level.
For both accounts, VOS objects remain within the VP, so both accounts predict the availability of
low scope indefinite interpretations of bare nominals, as argued for in Chapter 3. The difference
between both accounts is where VSO objects are interpreted. Under a predicate-fronting account,
objects move out of the VP before the vP fronts. For the head movement account, the VSO object
remains in its base position within the VP, below the subject. The trees in (263) model existential
9There is evidence for overt morphemes of the Possessor head. For instance, the -Vl suffix, glossed here
as a relational suffix, turns nonrelational nouns such as ‘cloth’ into relational ones, such as ‘clothing (that
one is wearing)’, as shown in the examples in (1). With the addition of this suffix, the relationship between
possessor and possessee is intrinsic, in the words of Barker (1995, 2011), in that (i) these nouns must be
obligatorily possessed; and (ii) the relationship between possessor and possessee is ‘tighter’. This suffix,
hypothesized to be in a Possessive head, also derives part-whole constructions, as in (2).
(1) Coon (2010a, 86)
a. i-pisil aj-Rosa b. i-pisl-el aj-Rosa
A3-cloth NC-Rosa A3-cloth-RS NC-Rosa
‘Rosa’s cloth’ (e.g., laundry) ‘Rosa’s clothing (e.g., that she is wearing)
(2) i-tye’-el otyoty
A3-wood-RS house
‘the wood of the house’
148
closure and VSO word order for both accounts.
(260) VSO output with existential closure at the VP
a. Head movement
IP
I SSP
TAM
SS Voice
Voice SS DP
tVoice
V Voice
SUBJ ∃ VP
tV DP
OBJ
b. Predicate-fronting
TP
vP T’
v
∃ VP T VoiceP
V tOBJ
DP Voice’
SUBJ
Voice AbsP
DP Abs’
OBJ Abs tvP
Below, I provide evidence in favor of moving the object out of the VP to derive VSO word
order. I begin with how each account models the placement of adverbs with respect to objects. I
make a modification to the head movement account in order to add in object shift.
Adverb placement
To begin, I review facts from adjunct placement fromChapter 4, which were first observed in Coon
(2010b). Adjuncts may intervene between DP objects as in (261a) but not NP objects as in (261b).
The verb and NP object must appear adjacent to each other, as in (261c).
149
(261) a. Ta’ j-k’ux-u ak’bi jiñi waj.
PFV A1-eat-TV yesterday DET tortilla.
‘I ate the tortilla yesterday.’
b. * Ta’ j-k’ux-u ak’bi waj.
PFV A1-eat-TV yesterday tortilla.
‘I ate a tortilla yesterday.
c. Ta’ j-k’ux-u waj ak’bi.
PFV A1-eat-TV tortilla yesterday
‘I ate a tortilla yesterday.
Adjunct placement with respect to VOS and VSO objects in Ch’ol is summarized in (262).
(262) Adverb Placement in Ch’ol
a. NP/VOS objects: V ONP XP (S)
b. DP/VSO objects: V XP (S) ODP
Both a predicate-fronting analysis and a head movement analysis can account for the place-
ment of adverbs. If adverbs appear as left adjuncts to the VoiceP, then it is predicted that the
adverb can intervene between the subject and DP object. This is given in (263) for VSO adverbs.
(263) DP Objects and adjunct placement
a. Predicate-fronting
TP
vP T’
v VP
T VoiceP
V tOBJ
AdvP VoiceP
ak’bi
yesterday DP Voice’
SUBJ
Voice AbsP
DP Abs’
OBJ Abs tvP
150
b. Head movement
IP
I SSP
TAM
SS VoiceP
Voice SS
AdvP VoiceP
V Voice
ak’bi
DP
yesterday tVoice VP
SUBJ
tV DP
OBJ
The placement of adverbs connect to interpretational differences of objects, as I will demon-
strate next.
Interpretation of the object (Chapter 3)
As discussed in Chapter 3, the structural position of a numerals reflects the possible semantic
interpretations available to it. In (264a), the object and verb are adjacent and the numeral can take
narrow scope with respect to the adverbial prefix ñoj ‘always’. When there is an intervening PP
in (264b), the numeral takes wide scope. More concretely, in (264a) it is possible that the speaker
always sees two men on the road but they do not necessarily need to be the same two men. Here,
the numeral ‘two’ takes low scope with respect to the adverb ñoj. In (264b) the interpretation is
that the speaker always sees two specific men on the road. The numeral ‘two’ takes wide scope
with respect to the adverbial prefix on the verb.
(264) Context A: There is a road behind my house. I always see two men walking there, some-
times it is Marcos and Rogelio, sometimes Juan and Eladio and sometimes Jorge and Javier.
Context B: There is a road behind my house. I always see Nicolás and José walking there.
a. Mi k-ñoj-k’el [ cha’-tyikil wiñik ] tyi bij.
IPFV A1-always-see two-CLF man PREP way
‘I always see two men on the road.’ ✓A ✓B
b. Mi k-ñoj-k’el tyi bij [ cha’-tyikil wiñik ].
IPFV A1-always-see PREP way two-CLF man
‘I sometimes see on the road two men.’ ✗A ✓B
151
Both accounts can capture the low scope indefinite interpretation of (264a) as the object is
adjacent to the verb, and within the VP, i.e., within the domain of existential closure. Where the
accounts differ is with (264b). Recall that adverb placement is indicative of object shift for Coon
(2010b).
A predicate-fronting account can capture the semantic differences in (264). Before vP fronting,
the object cha’tyikil wiñik ‘two men’ is shifted out of the VP, and out of the domain of existential
closure. I label the projection containing the numeral and noun in (264b) a QP, reflecting that in
(264b) the numeral is quantificational (see Chapter 3 Section 3.3.6 for more discussion). Recall
that with imperfective aspects, the subject is structurally a possessor, which controls a PRO in the
nominalized complement.
(265) . . .
. . . VoiceP
PP VoiceP
tyi bij
PREP road PRO Voice’
Voice AbsP
QP Abs’
cha’-tyikil wiñik Abs vP
two-CLF man
ñoj-k’el
always-see
For the analysis in Clemens and Coon (2018), it is possible to posit that VSO objects always
undergo object shift to a position under the subject, accounting for these semantic differences.
An updated account of VSO word order for Clemens and Coon (2018) is given in (266) where
the object shifts to the specifier of a projection labeled as AbsP. (Clemens and Coon (2018) do not
consider object shift in their paper.)
152
(266) Head movement account updated with object shift
. . .
. . . SSP
SS VoiceP
Voice SS
PP VoiceP
V Voice
tyi bij
ñoj-k’el
PREP road PRO
always-see
tVoice AbsP
QP Abs’
cha’-tyikil wiñik Abs VP
two-CLF man
tV tOBJ
With this modification, the semantic facts from above can be captured, though this raises some
questions about the relationship between VSO and VOS word orders. Under a predicate-fronting
account, VSO is derived in part by shifting the object overtly in the syntax—linking word order
and these semantic differences. By adding in object shift as in (266) for Clemens and Coon (2018),
the semantic facts are captured, but this movement is vacuous. That is, the linear order before
object shift (after head movement of the verb) is VSO and after object shift it is also VSO. So while
this movement is semantically driven, it does not effect constituent order.
Subextraction asymmetries (Chapter 4)
VOS and VSO objects also exhibit differences with respect to subextraction. As argued for in
Little (2020) and in Chapter 4, object shift bleeds subextraction. Objects with overt possessors
behave like DP objects: they are banned from appearing in VOS object position. These objects are
movement-derived islands, making subextraction from them impossible. This follows from the
Freezing Principle (Ross, 1974; Wexler and Culicover, 1977), or a ban on extraction from a moved
constituent. This can be seen clearly with numerals. Numeral modifiers are acceptable in VOS
153
and possible in VSO object position. The numeral may extract from (267a), as in (268a), but not
from (267b), as per (268b).
(267) Numerals may appear in VOS and VSO object position
a. Ta’ i-k’ux-u [O cha’-k’ej waj ] [S aj-Rosa ].
PFV A3-eat-TV two-CL tortilla NC-Rosa
‘Rosa ate two tortillas.’
b. Ta’ i-k’ux-u [S aj-Rosa ] [O cha’-k’ej waj ].
PFV A3-eat-TV NC-Rosa two-CL tortilla
‘Rosa ate two tortillas.’
(268) Numerals may only extract from VOS objects
a. Cha’-k’eji ta’ i-k’ux-u [O ti waj ] [S aj-Rosa ].
two-CLF PFV A3-eat-TV tortilla NC-Rosa
‘Rosa ate two tortillas.’
b. * Cha’-k’ej ta’ i-k’ux-u [S aj-Rosa ] [O ti waj ].
two-CLF PFV A3-eat-TV NC-Rosa tortilla
Intended: ‘Rosa ate two tortillas.’
Object shift could be added (motivated by the semantic facts stated above) to account for the
observed subextraction asymmetries for the analysis in Clemens and Coon (2018). Under the
predicate-fronting account from Coon (2010b), VSO objects do shift, predicting they are islands
for extraction, explaining the asymmetries in (268).
As I discussed in Chapter 4 Section 4.6, there is a caveat regardingwhen extraction from objects
is possible. Under a predicate-fronting account, such as the case for VOS numerals, an explanation
is needed for why extraction from moved vPs is possible. Given the Freezing Principle, the vP in
(269) should constitute an island, but as shown in (268a), extraction is possible from the object NP.
(269) A phrasal-fronting account does not predict (268a)
TP
vP T’
v VP T VoiceP
-u
TV V NP DP Voice’
k’ux
eat cha’-k’ej waj aj-Rosa Voice tvP
two-CLF tortilla
154
As will be presented in Section 5.5, by base-generating VOS, objects are not embedded in
moved constituents and extraction may proceed from them.
Interim summary and discussion of object shift
In this subsection, I compared Coon’s (2010b) predicate-fronting analysis and Clemens and Coon’s
(2018) head movement account with respect to the semantic and syntactic properties of objects
from Chapter 3 and 4. Semantically, VSO objects take wide scope, which was exhibited with the
interpretation of numerals. Syntactically, VSO objects are islands for extraction. The main take
away from this section is that object shift can explain both these empirical patterns. Although
object shift can be added to the head movement account, this raises the larger question of the role
of object shift in deriving VSO word order in Ch’ol.
Object shift has been linked to word order differences across a number of typologically differ-
ent languages (e.g., Diesing 1992: German; Gallego 2013: Romance languages; Diesing and Jelinek
1995: Egyptian Arabic; Woolford 2017: Eskimo-Aleut languages; Bhatt and Anagnostopoulou
1996: Hindi; Aydemir 2004; Kelepir 2001: Turkish, Holmberg 1986; Vikner 1994: Scandinavian,
etc.). In Ch’ol, object shift explains certain syntactic and semantic properties of VSO objects. This,
along with the other studiesmentioned, points to a link between object shift and word order in the
syntax. While it is possible to modify Clemens and Coon’s (2018) VSO account and add in object
shift, this movement is vacuous: the linear order of constituents is unaffected. I believe this misses
a more generalizable connection between object shift and its effect on word order.
5.4.4 Summary
In this section, I began with how Coon’s (2010b) and Clemens and Coon (2018) accounts capture
morpheme order in Ch’ol. A head movement account captures the order of morphemes in Ch’ol,
but certain questions arise with respect to the placement of the aspect marker and the causative
morpheme for the predicate-fronting account. A predicate-fronting account can capture nomi-
155
nalized complements by moving nP; under a head movement account, certain additions can be
made so that it can also capture nominalized complements. I ended with some comments on the
relationship between object shift and word order. While Clemens and Coon’s (2018) account can
accommodate object shift, the linear order of constituents is unaffected.
Next, I develop a base-generated account of verb-initial word order, which, I argue, parsimo-
niously derives the morphological, syntactic, and semantic properties discussed in this section.
5.5 Base-generating verb-initial syntax
In this section, I propose an analysis supporting a base-generated account of verb-initial word
order, first developed in England (1991) and Aissen (1992), and more recently argued for
in Otaki et al. (2019) for the Mayan language Kaqchikel. I update Aissen’s (1992) approach
with current assumptions about phrase structure and the ordering of syntactic heads from
Clemens and Coon (2018). I explain how this approach captures the data above and I then ex-
tend the analysis to account for nominals and ditransitive verbs. The analysis for VOS is given
in (270a) and VSO in (270b). VOS is base generated with the subject in a right specifier. VSO is
derived from VOS where the object shifts to a position above the subject.
(270) Proposal
a. VOS b. VSO
IP IP
I SSP I SSP
SS VoiceP
SS VoiceP
Voice’ DP
VoiceP DP
Voice VP SUBJ
OBJ
V NP Voice’ DP
OBJ Voice VP SUBJ
V tOBJ
156
In addition to being applied to Mayan languages (England, 1991; Aissen, 1992; Otaki et al.,
2019), other authors have explored base-generating verb-initial syntax. Chung (1998) proposes a
base-generation account of VOS word order in Chamorro (Austronesian) and Wojdak (2007) for
Nuu-chah-nulth, a Wakashan language. Below, I provide reasons for why base-generating VOS in
Ch’ol is a straightforward way of accounting for the aspects of Ch’ol morphosyntax and semantics
discussed above.
5.5.1 Order of morphemes
Apredicate-fronting analysis faced challenges accounting for the order of morphemes in causative
constuctions. Under the head movement account from Clemens and Coon (2018) the order of
morphemes in (271a) can be captured. I adopt the order of phrases from Clemens and Coon’s
(2018) in that the head moves to the status suffix, via Voice. This achieves the correct morpheme
ordering as well as word ordering, as shown in (271b).
(271) a. Ta’ k-ju’b-es-a pok’.
PFV A3-descend-CAU-TV bowl
‘I took the bowl down.’
b. IP
I SSP
ta’
PFV
SS Voice
Voice SS
-a DPi
V Voice tVoice VP
pro
ki-ju’b -es tV NP 1PRON
descend CAU
pok’
bowl
The verbal head moves along the projections of the verbal spine to the status suffix, in line with
standard assumptions of the Mirror Principle (Baker, 1985). The perfective aspect marker is in I.
157
5.5.2 Nonperfective complements and nominals
For both the predicate-fronting and head movement analyses, additional structure between DP
and PossP is needed to account for word order in nominalized complements. Under an account
with right specifiers, the ordering of the possessor with respect to nominalized complements is
base-generated. Take, for instance, the nonperfective clause in (272a) where the nominalized tran-
sitive verb is the complement of the progressive aspect woli. The postverbal order is O-S, where
the subject is structurally a possessor. As Aissen (1992) posits, possessors are also structurally on
the right. The tree in (272b) captures the structure in (272a). The verb moves to the nominalizing
head n, below aj-Rosa and the correct word order is achieved.10
(272) a. Woli [DP [nP i-juch’ sa’ PROi ] aj-Rosai ].
PROG A3-grind masa NC-Rosa
‘Rosa is grinding masa.’
b. VP
V DP
woli
PROG
D PossP
Poss’ DPi
aj-Rosa
Poss nP
n Voice
Voice n
PROi
V Voice tvoice VP
juch’ tV NP
grind
sa’
masa
A base-generated structure also captures the correct order for other possessive constructions,
as in (273a), with the syntactic structure in (273b). The possessor is base-generated in a right
10The noun may actually move as high as Poss. See footnote 9 above.
158
specifier above its posessee, generating the order possessee–possessor.
(273) a. i-wakax aj-Rosa
A3-cow NC-Rosa
‘Rosa’s cow’
b. DP
D PossP
Poss’ DPi
Poss nP aj-Rosa
NC-Rosa
i-wakax
A3i-cow
This account does not need an additional IP projection between DP and PossP, unlike for Coon
(2010b) and the modification I made to Clemens and Coon (2018). This extra projection served as
a landing site for nP under Coon (2010b) and the nominalized verb for Clemens and Coon (2018).
Otherwise, IP in the nominal domain was not related to any other morphological or syntactic
function.
5.5.3 VSO objects: Semantic and syntactic features
To derive VSO word order, VSO objects are reordered in the sense of England (1991), or moved to
the right, as in Aissen (1992). The structure for VSO is given in (274).
159
(274) VSO
IP
T SSP
SS VoiceP
VoiceP DP
OBJ
Voice’ DP
Voice VP SUBJ
V tOBJ
Additional evidence that the DP object moves to a position above the subject comes from the
sentence in (275) where a VSO object may bind into a subject. The animate subject iyum ‘its owner’
is bound by a definite object ixä machity ‘that machete’ in (275).11
(275) Ta’ i-xul-u ti [S ii-yum ] [O ixä machity ]i.
PFV A3-break-TV A3-owner that machete
‘Itsi owner broke [ that machete ]i.’
(275) provides further evidence for the analysis of VSO in that the object shifts to a position
above the subject. This also contrasts with Coon (2010b), where the object shifts to a position below
the subject before vP-fronting. Under Coon’s (2010b) account, the shifted object is syntactically
lower than the subject, which would not straightforwardly explain the binding facts in (275). In
Clemens and Coon (2018), the VSO object also does not c-command the subject.
As I explained in Chapter 2, absolutive case is licensed in situ for objects (Coon, 2017). This is
11Speakers judge this sentence acceptable and translate it as ‘Its owner broke that machete’ (Spanish:
Su dueño quebró ese machete, with the understanding that ‘that machete’ binds the possessor in the subject).
However, one potential confounding factor that should be further investigated is the fact that the order of
the subject and object is the same order expected of possessees and possessors. That is, the linear order of
S-O in this sentence corresponds to the Possessee–Possessor order exhibited in the nominal domain
([DP iyum ixä machity ] = the machete’s owner). The translations into Spanish of this sentence lead me to
believe that these do indeed act as subject and object, however, further tests are needed to confirm this.
This analysis also predicts that quantifiers in object position should be able to bind into the subject. For
instance in the sentence ‘its owner broke every machete’, ‘every machete’ should be able to bind into the
subject given the structure in (274).
160
to say, the movement of the object in (274) is not case related; rather it is motivated by a need for
overt markers of definiteness to vacate the VP, i.e., the domain of existential closure. I discuss this
type of movement with respect to syntactic ergativity in Section 5.7.
Adjuncts
I propose that adjuncts, such as adverbs, are right-adjoined to the VP.12 This accounts for the
ordering of V-O-XP-S for NP objects in (276a) and V-XP-S-O for DP objects in (276b).
(276) Proposal for the placement of adjuncts
a. VOS b. VSO
IP IP
I SSP I SSP
SS VoiceP
SS VoiceP
Voice’ DP
VoiceP DP
SUBJ
Voice VP OBJ
Voice’ DP
VP AdvP
SUBJ
V NP Adv Voice VP
OBJ VP AdvP
V tOBJ Adv
The proposals for adjuncts in (276) are important for data related to objects, discussed below.
Interpretation of objects
Recall the data on numerals in object position from (264), repeated in (277). When the object with
a numeral modifier is next to the verb, a low scope indefinite interpretation is possible, in (277a).
12Further support for the placement of adverbs at the VP can come from ellipsis. For instance, if it is
possible to elide a VP in Ch’ol including the VP adjunct to the exclusion of the subject, this would provide
evidence that the adjunct attaches lower than the subject.
161
When the object shifts out in (277b), it is specific.
(277) a. Mi k-ñoj-k’el [ cha’-tyikil wiñik ] tyi bij.
IPFV A1-always-see two-CLF man PREP way
‘I always see two men on the road.’ ñoj > 2; 2 > ñoj
b. Mi k-ñoj-k’el tyi bij [ cha’-tyikil wiñik ].
IPFV A1-always-see PREP way two-CLF man
‘I always see on the road two men.’ ñoj > 2
Under a base-generated account, the object moves out of the VP to Spec,VoiceP. It is there-
fore predicted that the numeral in (277b) does not take low scope with respect to the adverbial
quantifier.
(278) Object moves out of the domain of existential closure
Mi [VP k-ñoj-k’el ti ] tyi bij [ cha’-tyikil wiñik ]∃ i.
IPFV A1-always-see PREP way two-CLF man
‘I often see on the road two men.’
The structure is also provided in (279). Recall that (277b) is marked with the imperfective
aspect marker, so the verb is analyzed as being nominalized, hence the PRO in subject position.
(279) VoiceP
VoiceP QPi
cha’-tyikil wiñik
two-CLF man
Voice’ PRO
Voice VP
VP AdvP
AdvP VP tyi bij
PREP road
ñoj- V ti
always k’el
see
162
Subextraction asymmetries (Chapter 4)
This account can straightforwardly capture the subextraction asymmetries from the previous
chapter. For instance, in (280a) the interrogative possessor may not extract from object position as
it has been moved and is therefore an island for extraction. This is modeled in the tree in (280b),
where wh-movement is always to the left. (I discuss how specifier direction is parameterized in
Section 5.6.)
(280) a. * Majkii ta’ a-k’el-e [ i-chich ti ] ?
who PFV A2-see-TV A3-sister
Intended: ‘Whose sister did you see?’
b. Subextraction blocked from moved objects
CP
C’
C IP
T SSP
ta’
SS VoiceP
Voice SS
-e VoiceP DPi
V ✗ Voice TV
ah-k’el
Voice’ DP DPk D’
A2-see h
Voice VP pro majki D PossP
2PRON
tV ti ik-chich tk
A3-sister
Recall that VOS objects do permit extraction of numerals, as in (281a). Under a predicate-
fronting account the object is within a moved phrase. Given the Freezing Principle, the moved
phrase would be an island for extraction, predicting (281a) to not be possible. The account here
provides a simpler solution to extraction from VOS objects in (281a): VOS objects remain in their
base position and therefore are not islands, so subextraction from them is possible.
163
(281) a. Cha’-k’eji ta’ i-k’ux-u [O ti waj ] [S aj-Rosa ].
two-CLF PFV A3-eat-TV tortilla NC-Rosa
‘Rosa ate two tortillas.’
b. Subextraction possible from objects in their base position
CP
C’
C IP
T SSP
ta’
SS VoiceP
Voice SS
-e
V Voice TV Voice’ DPh
ah-k’el
A2-see protVoice VP 2PRON
tV nP
MP nP
cha’-k’ej waj
two-CLF tortilla
5.5.4 Extension to ditransitives
This account can also be extended to ditransitive constructions in Ch’ol. When the direct object
is bare, the most natural ordering of the verb, direct object and indirect object is V-DO-IO, as in
(282).13 Ditransitives are additionally marked with the applicative suffix -be on the verb.
(282) Ta’ k-choñ-b-e karu aj-Rosa.
PFV A1-sell-APPL-DTV car NC-Rosa
‘I sold a car to Rosa.’
Ditransitives in Ch’ol are what Pylkkänen (2008) categorizes as low applicatives, as discussed
13I have not found any examples of three postverbal verbal arguments in texts. Speakers judged sen-
tences with three postverbal arguments as unacceptable.
164
in Coon and Henderson (2010). I follow Georgala et al. (2008) and Georgala (2012) as well as sug-
gestions in Bruening (2010) that some low applicatives are in fact syntactically above the VP, di-
verging from the low applicative structure posited in Pylkkänen (2008). Georgala et al. (2008) call
these structurally high low applicatives ‘thematic applicatives’. The structure for ditransitives in
Ch’ol is given in (283).14 This derives the correct order of morphemes as the verb moves via head
movement through Appl in (283). Importantly for this analysis, the specifier of ApplP is on the
right, deriving the correct postverbal ordering of DO-IO.
(283) IP
I SSP
ta’
PFV
SS VoiceP
Voice SS
-e
Voice’ DPi
Appl Voice -DTV
pro
V Appl
tVoice ApplP 1PRON
ki-choñ b
A1-sell
Appl’ DP
tAppl VP aj-Rosa
tV NP
karu
car
14Under a predicate-fronting account, it is unclear how to achieve the correct word order for the objects
of ditransitives. Assuming that ApplP is generated below vP, after vP fronting and head movement of the
verb, the order that is derived is V-IO-DO in (1), but the attested word order is V-DO-IO in (282).
(1) Ditransitives under a predicate-fronting account?
IP
vP VoiceP
DP Voice’
v ApplP
pro Voice tvP
Appl v 1PRONDP Appl’
-e
V Appl aj-Rosa tAppl VPk-choñ b
A1-sell tV NP
karu
car
Additional movement of the object karu ‘car’ would need to be motivated to account for the ordering of the
postverbal objects in (282), perhaps via roll-up movement.
Under a head movement account, after the verb has moved to SSP from its base position, the direct object
NP would be prosodically reordered via ARG-ϕ and be pronounced with the verb, generating V-DO-IO.
165
5.5.5 Summary
In this section, I have shown how a base-generated account for verb-initial word order can capture
VOS/VSO order as well as additional data on morphological, syntactic and semantic properties
in Ch’ol. Subjects and possessors are in right specifiers of VoiceP and PossP, respectively. This
account capitalizes on the link between object shift and VSO word order, namely that object shift
derives VSO word order, in turn capturing the syntactic and semantic properties of objects that
were presented in Chapters 3 and 4. It also provides a simpler way of capturing why VOS objects
are not islands for extraction. Finally, I extended this account to ditransitive constructions, positing
that the indirect object is in a right specifier of ApplP. I also suggested that low applicatives in
Ch’ol are syntactically above the VP. Next, I posit that the linearization of specifiers to the right is
a parameter setting available to ergative languages and occurs in the narrow syntax.
5.6 Formalizing right specifiers
Under Kayne’s (1994) antisymmetric approach to syntax, the goal is to derive VOS or VSO word
orders from SVO word order, as phrasal-fronting or head movement accounts do. The approach
favored in this chapter is anti-Kayneian in nature, as I have argued for an approach with right
specifiers. For Kayne (1994), specifiers are on the left. As shown in Section 5.4, deriving verb-
initial syntax from a base order of SVO faces certain challenges.15 In this section, I discuss how
right specifiers can be formalized, making a connection between right specifiers and ergative case
assignment.
I follow the work of Aldridge (2004, 2008), Coon (2013, 2017), and Legate (2008), in that a low
functional head assigns ergative case, as outlined in Chapter 2, Section 2.3.4. I propose that the
head that assigns ergative case to its specifier can also encode the direction of the specifier. Under
this analysis only ergative languages have the possibility of parameterizing specifiers on the right.
Case assignment for ergative and genitive is given below in (284).
15See also Davis (1999) for a discussion of a Kayneian approach versus a base-generated account for
verb-initial word order in St’at’imcets (Salish).
166
(284) a. VoiceP b. DP
DP Voice’ D PossP
SUBJ Voice VP
ERG DP Poss’
V NP
Possessor Poss nP
OBJ
GEN n NP
possessee
I begin with outlining my theoretical assumptions under a Minimalist approach to syntax
(Chomsky, 1995), before I present my proposal for parameterizing right specifiers.
In the Minimalist Program Chomsky (1995), Merge is a set operation that joins two syntactic
objects and forms a set as in (285), where L is the label of the set. The label is determined by which
of α or β projects.
(285) Merge(α, β) = {L α, β}
For instance, if V combines with a direct object NP, the label will be VP, as in (286).
(286) Merge (V, NP) = {VP V, NP }
Now, the sets in (285) and (286) are unordered sets. How α is pronounced with respect to β is
not dictated by the set in (285). Linearization of the syntactic objects in (285) is what determines
the order in which they appear. It has been posited that precedence relations are determined by
structural dominance, or whatever syntactic object is structurally higher is pronounced first (i.e.,
as per Kayne’s (1994) Linear CorrespondenceAxiom or Yang’s (1999) Set Linearization Condition).
I depart from this notion and, elaborating on a suggestion in Otaki et al. (2019), formalize a way
to linearize right specifiers in the narrow syntax.
To order a set such as in (285), Takita (2020) proposes that the label L of the set containing
α and β triggers linearization in the narrow syntax. To account for headedness differences across
languages, Takita (2020) motivates an analysis where heads are linearized with their complements
167
prior to spell out. Takita (2020) proposes the following two linearization rules to derive V-O lin-
earization in English and O-V linearization in Japanese. For Takita (2020) the rules in (287) are
parameterized to a given language, accounting for differences in the direction of heads. Lineariza-
tion, thus, takes place before spell out.
(287) Linearization of Heads and their Complements (Takita, 2020)
a. Head-initial rule (e.g., English): {X X, YP}→ <X, YP>
b. Head-final rule (e.g., Japanese): {X X, YP}→ <YP, X>
The order of operations to derive linear order of a set of merged syntactic objects is given in
(288).
(288) a. Merge triggers labeling (Chomsky, 1995, 2000)
b. Labeling triggers linearization (Takita, 2020)
I propose that the low functional head that assigns ergative or genitive case parameterizes the
direction of its specifier, once the subject or possessor is merged. The specifier parameterization
is therefore derived in the narrow syntax. Once the set is linearized, it is sent to spell out where
uninterpretable features are deleted and then the linearized set is transferred to PF. This derives
right specifiers for VoiceP and PossP, i.e., specifiers of case-assigning heads on the right. This
includes movement within the VoiceP, where DP objects move. In (289), [SPEC:R] represents a
right specifier setting for the VoiceP/PossP, triggeredwhen the subject or possessor is merged and
assigned inherent case.
(289) a. VoiceP b. PossP
Voice’ DP Poss’ DP
Voice . . . SUBJ Poss . . . SUBJ
[SPEC:R] [SPEC:R]
ERG GEN
Restricting the subject parameterizing to heads assigning inherent case also buoys another
fact about the Ch’ol data: A-bar movement to the CP or the edge of DP is always leftward. This
168
accounts for the fact that wh-words and focused constituents appear preverbally, in Spec,CP or
Spec,DP.
Asmentioned above, this specifier setting is only possible for ergative languages. Nominative-
accusative aligned languages do not have this option. This raises questions about rare word or-
ders such as OVS. Dryer (2013) reports 11 OVS languages, and at least eight of them are ergative:
Hixkaryana, Tirio and Kuikuro (Cariban), Selknam (Chon), Cubeo (Tucanoan), Asurinı́ (Tupian),
Urarina (Urarinian), and Tuvaluan (Austronesian). A possible way to derive OVS ergative lan-
guages is that they are VSO or VOS underlyingly, but the object moves to a position high in the
clause, such as in a topic position, generating OVS.16 Under this approach, the relative rarity of
OVS can be ascribed to the extra step of topicalizing the object.17 Indeed, this seems to be what
happens in Tuvaluan, one of the languages listed as OVS in Dryer (2013).
While Tuvaluan is listed in Dryer (2013) as an OVS language, Besnier (2000, 133) describes
the basic word order as VSO. He notes however that OVS is more frequent than VSO, attributing
OVS to direct object preposing (Besnier, 2000, 133), as I suggest in my approach above. That is,
Tuvaluan is underlying verb-initial, but by moving the object to a high topic position, OVS word
order is achieved. This puts into question the categorization of the other OVS languages—is OVS
the basic word order or are there other word order patterns attested? With more language-specific
investigation, it will be possible to discern the answer to these questions.
5.7 Implications for Mayan languages with fixed VSO word order
Historically, Proto-Mayan word order was VOS, and VSO when the object was more marked in
terms of definiteness or animacy (Norman and Campbell, 1978). The analysis argued for in this
chapter preserves the idea that VSO in VOS/VSO alternating languages is derived by reordering
the object from a base-generated VOS structure. I now discuss implications of this type of analysis
16The OVS language Ungarinjin is described as not being ergative (Rumsey, 1982, 149); however, as
Whitman (2008) mentions, it has SV order in intransitive clauses, unlike the other OVS languages which
have VS order with intransitive clauses.
17There are also four reported cases of OSV in Dryer (2013); however Whitman (2008) raises questions
about the validity of analyzing these as instances of true OSV basic word order.
169
for Mayan languages with rigid VSO word order.
As per Norman and Campbell (1978) and England (1991), strict VSO word order is an inno-
vation of the Mamean and Q’ajonb’alan subgroups. These languages are all found in one geo-
graphically fixed zone. As such, fixed VSO word order has been characterized as an instance of
diffusion. The languages England (1991) cites with fixed VSOword order are given in (290). There
is dialectal variation in Ixil and Chuj: the dialects which do exhibit fixed VSO order are listed after
the colon in (290).
(290) Mayan languages with fixed VSO England (1991, 451)
a. Mam
b. Tektitek
c. Awakateko
d. Ixil: Nebaj, Chauj
e. Q’anjob’al
f. Popti’ (Jalkatek)
g. Chuj: San Sebastián
An example of VSO in Q’anjob’al is given in (291).
(291) Max y-il-a’ [S ix ix ] [O naq winaq ].
ASP A3-see-TV N.CLF woman N.CLF man
‘The woman saw the man.’ (Q’anjob’al: Coon et al. 2014, 226)
England (1991) posits that fixed VSO languages generalized the more marked VSO order of
Proto-Mayan to be the pragmatically neutral word order. There are two possibilities for how fixed
VSOword order in these Mayan languages came about. The possible structures are given in (292).
170
(292) Two possible structures for fixed VSO in Mayan
a. VSO via head movement of the verb b. VSO via movement of the object
IP IP
I+V+Voice VoiceP
I+V+Voice VoiceP
SUBJ Voice’
VoiceP OBJ
tVoice VP
Voice’ SUBJ
tV OBJ
tVoice VP
tV t
In the structure in (292a), SVO order is generated and the verb moves to I, generating VSO.
This is essentially the head movement account in Clemens and Coon (2018) without prosodic re-
ordering of the object. From a historical perspective, the prosodic constraint which reorders NP
objects with the verb that selects them would be deactivated in fixed VSO languages. The struc-
ture in (292b), on the other hand, is the same structure I developed for VSO in Ch’ol where VOS is
base-generated and the object moves over the subject, generating VSO. From a historical perspec-
tive, these languages would have generalized object shift to occur in all VSO instances.
While the structure in (292a) may be simpler, there is evidence pointing towards the structure
in (292b) where VSO in fixed VSO Mayan languages is derived from VOS via movement of the
object over the subject. This comes from two properties of the VSO Mayan languages in (290).
These two properties are given in (293).
(293) Properties of Mayan languages with fixed VSO word order
a. VOS surfaces with reflexives in object position.
b. Fixed VSOMayan languages are syntactically ergative.
I review these points below, with data from Q’anjob’al, to support the analysis in (292b) for
the languages in (290). I also discuss in Section 5.7.3 why the VSO configuration in (292b) leads to
syntactic ergativity in VSO Mayan languages, but not Ch’ol (which is not syntactically ergative).
171
5.7.1 Reflexive objects in fixed VSO languages
The first piece of evidence that points towards deriving VSO from VOS in fixed VSO languages
comes from the position of reflexive objects. England (1991, 456) notes that for Mam, a fixed VSO
language, VOS word order is observedwhen the object is reflexive. Coon et al. (2014, 226) cite that
the only contexts where Q’anjob’al allows VOS word order are precisely those where the object
must be bound locally, as shown in the Q’anjob’al example in (294a). Compare to the VSO sentence
in (294b).
(294) Q’anjob’al
a. Reflexive: VOS
Max y-il [O s-b’a ] [S ix ix ].
ASP A3-see A3-self N.CLF woman
‘The woman saw herself.’ (Coon et al., 2014, 226)
b. Transitive: VSO
Max y-il-a’ [S ix ix ] [O naq winaq ].
ASP A3-see-TV N.CLF woman N.CLF man
‘The woman saw the man.’ (Coon et al., 2014, 226)
Under an analysis that posits object shift for fixed VSO, this makes sense. If the base word
is VOS, this explains why reflexive objects remain low as they must be bound locally. Under
the analysis in (292a), while the subject still c-commands the object and binds it, additional steps
would be needed to explain VOS word order in these constructions.
5.7.2 Syntactic ergativity and fixed VSO Mayan languages
A property of Mayan languages with fixed VSO word order is that they are syntactically ergative.
That is, there are restrictions on certain types of A-bar movement of the ergative subject. This
property of VSO Mayan languages can be generalized in the implication in (295).
(295) Implication on word order and syntactic ergativity in Mayan:
If a Mayan language has fixed VSO word order, it is syntactically ergative.
172
Note that (295) is formulated as a one-way implication: not all syntactically ergative Mayan
languages have fixed VSO word order.
To briefly illustrate these restrictions, consider the following data fromQ’anjob’al. Q’anjob’al is
syntactically ergative as it exhibits extraction asymmetries with respect to absolutive and ergative
arguments. In (296a) we see that the wh-marked absolutive subject has moved to the preverbal
position. In (296b), the sentence is ungrammatical if the ergative subject is extracted. This sentence
can only mean ‘Who did the man see?’, in which case the absolutive object has been extracted.
(296) Q’anjob’al (Coon et al., 2014, 192)
a. Maktxeli max way-i ti?
who ASP sleep-IV
‘Who slept?’
b. *Maktxeli max y-il-a’ ti naq winaq ?
who ASP A3-see-TV CLF man
Intended ‘Who saw the man?’
Grammatical as ‘Who did the man see?’
In syntactically ergative languages, to extract the ergative argument, different structures sur-
face. In Q’anjob’al, the agent focus construction is used, as exemplified in (297). In addition to the
agent focus morpheme, the verb also appears with the intransitive suffix (-i), indicating that the
structure that surfaces when ergative arguments are extracted is distinct from the syntax in (296).
(297) [ Maktxel ] max-ach il-on-i?
who ASP-2B see-AF-IV
‘Who saw you?’ Q’anjob’al (Coon et al., 2014, 215)
The extraction asymmetries in (296) have led to the characterization of Q’anjob’al as a syntactically
ergative language.
To derive syntactic ergativity, many authors have proposed that the object moves above the
subject so that it may be case licensed by INFL/T. The difference between syntactically ergative
languages and ergative languages that do no have A-bar extraction restrictions is the locus of ab-
solutive case assignment (Aldridge, 2004, 2008; Legate, 2002, 2008; Coon et al., 2014). For instance,
Aldridge (2004, 2008, 2012) posits that in transitive clauses INFL/T licenses case to the object. In
173
order for it to do so, the object must move to a position above the subject, as modeled in (298).18 I
label the projection that introduces the subject as VoiceP for consistency.
(298) Syntactically ergative languages (Aldridge, 2004, 2008, 2012; Coon et al., 2014)
IP
I VoiceP
ABS OBJ VoiceP
SUBJ Voice’
Voice VP
V OBJ
The movement of the absolutive object over the ergative subject in (298) accounts for the “trap-
ping” of the subject, and the extraction restrictions. The reason why syntactic ergativity arises is
because the absolutive object raises above the subject to check case with INFL/T. For this rea-
son, the ergative argument is trapped inside the VoiceP and may not extract out. An additional
stipulation is needed that VoiceP may not project additional specifiers in order to let the ergative
argument extract out. Note that Coon et al. (2014) model the movement in (298) as covert, not
effecting word order.
Ch’ol, on the other hand, is morphologically ergative, and, as discussed in Chapter 2, Section
2.3.4, absolutive case for the object is licensed in situ by a low functional projection, modeled in
(299).
18Aldridge (2004, 2008, 2012) also provides a secondway to derive syntactic ergativitywhere the object is
assigned absolutive case by a low functional head such as v/Voice, but then moves to Spec,VoiceP/Spec,vP
to check an EPP feature.
174
(299) Morphologically ergative languages (Aldridge, 2004, 2008, 2012; Coon et al., 2014)
IP
I VoiceP
SUBJ Voice’
Voice VP
V OBJ
ABS
Returning to Q’anjob’al, I propose that the object movement in (298) is overt, to a right specifier
of VoiceP in fixed VSO Mayan languages. This is modeled in (300). Consequently, this links case
assignment, extraction restrictions, and VSO word order.
(300) Overt object movement and absolutive case assignment in a fixed VSO language
IP
I VoiceP
VoiceP OBJ
ABS
Voice’ SUBJ
Voice VP
V OBJ
Instead of the movement of the object being covert as in (298), the object moves overtly in
Q’ajonb’al and other fixed VSO Mayan languages, deriving both syntactic ergativity and VSO
word order.
Finally, as presented in Coon et al. (2020), constructions with reflexive objects are precisely
where extraction of the ergative argument is possible. Above, I posited that reflexives do not move
over the subject as they must be bound locally in their base VOS object position. Consequently,
since the object has not moved, the subject is not trapped and can extract out. If VSO languages
are underlying VOS, under the verb-initial structure argued for here, extraction restrictions and
reflexive constructions can be explained.
175
Next, I provide two options for why the object-over-subject configuration of VSO sentences
leads to syntactic ergativity in Q’anjob’al, but not in Ch’ol.
5.7.3 VSO and syntactic ergativity
Asmentioned above, a significant difference betweenCh’ol and the rigidly VSOMayan languages
is syntactic ergativity. This leaves the question as to why the configuration in (301) leads to syn-
tactic ergativity in some languages, but not in Ch’ol.
(301) VSO via movement of the object
IP
I+V+Voice VoiceP
VoiceP DP
OBJ
Voice’ DP
tVoice VP SUBJ
tV tOBJ
Below, I propose two possibilities to derive the differences between a syntactically ergative
language, such as Q’anjob’al, and a morphologically, but not syntactically, ergative, language such
as Ch’ol.
The first option to derive this difference concerns (i) the type of movement; and (ii) whether
VoiceP may project multiple specifiers. For Coon et al. (2014), syntactic ergativity arises because
the absolutive argument moves above the ergative subject, driven by case-motivated reasons. A
question then arises as towhy the same configuration proposed for Ch’ol VSO does not ban extrac-
tion of the ergative argument. Ch’ol, as shown in (302a), allows ergatives to extract from transitive
clauses, in contrast to the Q’ajonb’al example, repeated in (302b). The construction in (302a) has
a DP object, as those objects are posited to be in a right specifier of VoiceP, structurally above the
176
subject.
(302) a. Majkii ta’ i-k’el-e ti jiñi wiñik?
who PFV A3-see-TV man
‘Who saw the man?’
Also grammatical as: ‘Who did the man see?’ Ch’ol
b. *Maktxeli max y-il-a’ ti naq winaq ?
who ASP A3-see-TV CLF man
Intended ‘Who saw the man?’
Grammatical as ‘Who did the man see?’
Q’anjob’al
In Aldridge (2004) and Coon et al. (2014), movement of the absolutive argument over the sub-
ject is case-driven movement. Aldridge (2004) and Coon et al. (2014) must stipulate that the phase
edge (in this case labeled as Voice) cannot project multiple specifiers. I assume that VoiceP is a
phase and that in order to extract out of the phase, a constituent must move through the phase
edge, following Chomsky (1986, 2000, 2001, 2008). The configuration in (301) blocks the extraction
of the ergative argument as the Voice phase head may not project another specifier to let the erga-
tive argument extract out. The difference between the VSO configuration in Ch’ol, I propose, is
that in Ch’ol the verbal projection may project multiple specifiers, allowing the ergative argument
to extract out.
Indeed the type of movement undergone by DP objects in Ch’ol and objects in VSO Mayan
languages is different. In Ch’ol, movement is due to the structure of the nominal but in VSO
Mayan languages, object movement is case-driven movement. Recall, that I have assumed that
Voice assigns absolutive case to objects in Ch’ol, whereas in VSOMayan language INFL/T assigns
absolutive case.
The first solution for ±syntactic ergativity is summarized in (303).
(303) Multiple specifiers proposal for ±syntactic ergativity (option 1)
a. Syntactically ergative languages cannot project multiple specifiers on the edge of the
transitive verbal projection so the ergative argument is trapped after the object moves
above it. Movement of the object above the subject is case-motivated.
b. Morphologically, but not syntactically, ergative languages allow multiple specifiers so
the ergative argument may extract out through a specifier on the edge of the transitive
verbal projection. Movement of the object above the subject is not case-motivated.
177
The second option involves changing the head which licenses absolutive case of the transitive
object, adopting the case maraudage analysis of Assmann et al. (2015). In Assmann et al. (2015),
instead of Voice/v, INFL/T always assigns absolutive case.19 Crucial for their analysis is that in
ergative systems Merge occurs before Agree. For Assmann et al. (2015), syntactic ergativity arises
due to the ergative argument marauding case on INFL/T when it tries to extract out. Assuming
phase-by-phase movement and that IP/TP is a phase, what goes wrong is that when the erga-
tive argument moves through the specifier of INFL/T, it steals, or marauds, the absolutive case
feature on INFL/T. This maraudage leaves the internal argument unmarked for case and causes
the derivation to crash. Maraudage of the case feature in INFL/T also implies that DPs can check
more than one case (i.e., ‘case stacking’ (Andrews, 1996; Sadler and Nordlinger, 2006; Richards,
2013; Assmann et al., 2014)). This leads to syntactic ergativity. These steps are schematized in
(304).
19Assmann et al. (2015) use the terms internal case and external case. Internal case is assigned by v to
ergative arguments in a ergative-absolutive system and external case is assigned T to objects of transitive
verbs or subjects of intransitive verbs. For accusative languages, internal case is assigned to objects via v
and external case is assigned to external arguments via T. To streamline the discussion, I utilize the labels
absolutive and ergative case.
178
(304) An Assmann et al. (2015)-approach to syntactic ergativity
a. The ergative subject is first assigned case via Voice
IP
I VoiceP
DP Voice’
SUBJ Voice VP
ERG V DP
OBJ
b. Merge before Agree triggersmovement of the ergative-marked subject through Spec,IP,
where it marauds the absolutive case on I, leaving the object without a case: derivation
crashes
IP
DP I’
SUBJ
I VoiceP
[c:ERG]
[c:ABS]
ABS tSUBJ Voice’
Voice VP
V DP
OBJ
Extending this account to verb-initial Mayan languages, the absolutive case is assigned as
schematized in (305).
(305) IP
I±phase VoiceP
ABS VoiceP DPi
OBJ
Voice’ DP
Voice VP SUBJ
V ti
ERG
179
Under this option, for a language like Ch’ol, it can be assumed that INFL/T is not a phase, so
movement of the ergative argument does not maraud the case on INFL/T. That is, the movement
in (304b) does not happen as INFL/T is not a phase in Ch’ol. In a language such as Q’anjob’al,
INFL/T is a phase, andmovement of the ergative subject through INFL/Tmarauds the absolutive
case. This minimal difference, under the system of Assmann et al. (2015), is an option for why the
configuration in (301) results in syntactic ergativity for Q’anjob’al, but not for Ch’ol. This option
is summarized in (306).
(306) Case maraudage proposal for ±syntactic ergativity (option 2)
a. In syntactically ergative languages, INFL/T is a phase and movement of the ergative
argument through INFL/T marauds the ABS case feature, so the object is left caseless
and the derivation crashes.
b. In morphologically, but not syntactically, ergative languages, INFL/T is not a phase so
movement of the ergative argument does not pass through INFL/T, therefore there is
no maraudage of the ABS case.
Adopting the option in (306), however, raises issues for VOS in Ch’ol—if the object must move
to Spec,VoiceP for case reasons under the account in Assmann et al. (2015), a stipulation would
need to be added that NPs are pronounced in their base position (VOS) whereas DPs are pro-
nounced in Spec,VoiceP (VSO). This extra stipulation may point towards favoring option 1 in
(303).
5.7.4 Summary
In this section, I argued that VSO in fixed VSOMayan languages is derived from VOS by moving
the object overtly. The proposal that VSO is underlyingly VOS can account for the fact that VOS
appears in instances where the object must be bound locally, i.e., when the object is a reflexive.
I also posited that overt movement of the object above the subject links VSO word order with
syntactic ergativity in these languages. Moving forward, a way to test if the word order analysis
for VSO Mayan languages is on the right track is to investigate properties of VSO objects—are
they islands for extraction? Can they bind into subjects? Do they always take wide scope? I leave
these questions for future investigation.
180
I also provided two solutions for why the configuration of object-over-subject leads to syntactic
ergativity for rigid VSO Mayan languages, but not for Ch’ol. The first option relates to whether
multiple specifiers may project at the edge of the verbal domain. The second option involves
changing the case licensing heads and adopting a case mauradage analysis from Assmann et al.
(2015).
5.8 VSO Mayan languages and the NP/DP hypothesis
Before concluding this chapter, I outline areas of future investigation on the nominal and clausal
domains in Mayan, with implications for the underlying syntax of noun phrases across languages.
Within nominal syntax, a point of contention is whether all languages’ nominal expressions have
the structure in (307) or if some languages lack a DP layer.
(307) DP
D NP
N’
N
For instance, the Universal DP Hypothesis states that all languages, no matter if they have
overt definite articles or not, have the structure in (307). That is, the head of the nominal projection
is D and not N. For instance, in English the definite article ‘the’ would be a phonological output
of D. Early proponents of the Universal DP Hypothesis, summarized in Bruening (2009), include
Hellan (1986), Abney (1987), and Szabolcsi (1987). The difference between a languagewith articles,
such as English, and a language without obligatory articles, such as Russian, is phonological.
Under the Universal DP Hypothesis, D is projected but not pronounced in Russian.
Some authors have taken issue with the universality of (307), such as in Payne (1993) and
Bruening (2009). While Payne (1993) and Bruening (2009) have argued against the DP hypothesis,
other authors have argued that the structure in (307) is not universal; rather it is parameterized
across different languages (see discussion in Bošković (2005, 2008); Despić (2011, 2013, 2015)). For
181
the NP/DP hypothesis, a core empirical difference of languages whose nominals are headed by D
is that these languages have overt definite articles. NP languages do not have overt definite articles
and under this hypothesis, Dwould not be projected in a language such as Russian. English, under
this theory, has the structure in (307). I sketch out some research queries relating to this hypothesis
for nominal syntax in Mayan.
As discussed in Chapter 3, Ch’ol does not have obligatory determiners. That is, bare nouns
may be definite depending on the context. Throughout this dissertation, I have labeled nominal
phrases in Ch’ol with determiners or overt possessors as DPs for descriptive purposes, but recall
that Ch’ol does not obligatorily mark definiteness. It is the fact that articles are not obligatory that
preliminarily supports characterizing Ch’ol as an NP language, under the NP/DP hypothesis.
On the other hand, some Mayan languages seem to have obligatory markers of definiteness.
This is exhibited in (308) for Q’anjob’al.
(308) Q’anjob’al
Max-Ø s-tayne-j [S naq unin ] [O te na ].
COM-A3S E3S-take.care-TV N.CLF child N.CLF house
‘The child took care of the house.’ (Mateo Toledo, 2008, 79)
Noun classifiers such as naq and te signal an overt definite interpretation.20,21 As clear for
Ch’ol, objects with overt markers of definiteness appear in VSO object position. Since Q’anjob’al
has obligatory markers of definiteness, it would be classified as a DP language under the NP/DP
hypothesis. If overt markers of definiteness trigger object shift in Ch’ol, is there a connection
between VSO in Q’anjob’al and its possible status as a DP language? That is, if Q’anjob’al objects
project a D, this could be further support for why objects must shift.
20In Chuj, a language related to Q’anjob’al, these nominal classifiers may also appear with numerals and
the indefinite article jun. But in Chuj [ N.CLF NP ] always arises in a definite interpretation (Royer, 2019).
21Extended reflexive constructions such as in (1a) are also an area for further investigation. In the ex-
tended reflexive construction in (1a) in Chuj, the possessor of the object in (1a) must be the subject Xun.
When there is an overt third person pronoun, bolded in (1b), the subject can no longer bind into the object.
(1) Chuj (J. Royer, personal communication, July 1, 2020)
a. Ix-s-chonh s-wakax waj Xun. b. Ix-s-chonh s-wakax winh waj Xun.
PFV-A3-sell A3-cow CLF Xun PFV-A3-sell A3-cow CLF CLF Xun
“Xun sold his (own) cow.’ ‘Xun1 sold his2/∗1 cow.’
182
However, not all rigidly VSO Mayan languages have obligatory definite articles. England
(1983, 151) notes this for Mam, a Mayan language of the Mamean branch:
There is no definite article in Mam. The indefinite article alone marks a noun phrase
as indefinite; its absence implies definiteness. (England, 1983, 151)
An example from Mam is given in (309), with the bare definite nouns bolded.
(309) ma Ø-txi7 t-k’aa7ya-n [S xiinaq ] [O axi7n ].
REC B3-DIR A3-sell-DS man corn
‘The man sold the corn.’ Mam: (England, 1983, 151) (with my bracketing and emphasis)
This contrasts in two ways with Ch’ol. The first is that in Mam bare nouns must be definite.
Recall from Chapter 3 that in Ch’ol bare nouns as objects in VOS word order and absolutive sub-
jects in VS word order can be indefinite. According to England (1983), in Mam, bare nouns are
always definite and indefinites are lexically marked. The second contrasting point is that Mam
is a rigidly VSO language, whereas Ch’ol is VOS/VSO alternating. If VSO objects in Mam can
be bare, but must be definite, then this could constitute further support for the proposal for VSO
languages. If the object in VSO languages is not within the VP, this would capture that bare ob-
jects in Mam can only be definite: it is outside of the domain of existential closure. It would then
be predicted that to express indefiniteness, overt markers must be used, and/or incorporating
structures. Nevertheless, this point contrasts to the one made above for Q’anjob’al. Unlike in
Q’anjob’al, definiteness is not marked in Mam.
Finally, as has been noted in the literature, the presence or absence of an article in a language
can have consequences for other properties of that language (see, e.g., Bošković (2008) or Despić
(2011) for a list of crosslinguistic generalizations). One generalization that wasmentioned in Chap-
ter 4 is that if a language allows left branch extraction, it does not have obligatory articles. This is
another aspect that can be further investigated within the Mayan languages. Without a doubt, the
connections between nominal and clausal structure across Mayan languages promises to be a rich
area for future work.
183
5.9 Conclusion
In this chapter, I compared recent analyses for verb-initial word order in Ch’ol, specifically the
head movement account from Clemens and Coon (2018) and Coon’s (2010b) predicate-fronting
account. I argued for a base-generated approach to verb-initial word order, drawing on analyses
from Chapters 3 and 4. This account, I argue, is able to capture the order of morphemes (adopt-
ing the order of phrases from Clemens and Coon (2018)), nominals, and syntactic and semantic
properties of VOS versus VSO objects. Additionally, this account is extendable to ditransitive con-
structions. I also provided a formalization for parameterizing the direction of specifiers, based on
the head that assigns ergative case. As not all verb initial languages are ergative, it follows that
verb-initial languages do not form a homogenous syntactic class.
I additionally proposed that in fixed VSO Mayan languages, the object moves overtly above
the subject. This is supported by data on reflexives and syntactic ergativity. More investigation
into the syntactic and semantic properties of VSO objects in fixed VSO Mayan languages can
further support this analysis. For instance, are VSO objects islands for extraction? Can they bind
into the subject? Do they always takewide scope? This account would say yes—these are therefore
testable predictions for subsequent work.
184
CHAPTER 6
CONCLUSION
The investigation of mutual dependencies between nominal and clausal syntax can reveal
clues to the underlying structure of a language. These clues also inform us about structural consis-
tencies that hold crosslinguistically. I have examined properties of nominal arguments and their
dependencies on clausal syntax as a window to investigate underlying structure. While these
dependencies may be linearized in different ways, there are robust structural consistencies. This
dissertation presents a case study of Ch’ol arguing that certain mutual dependencies of nominal
and clausal structure favor a particular analysis for verb-initial word order. The dependencies in
Ch’ol are not Ch’ol-specific. Many other languages exhibit similar patterns with respect to object
position and interpretation or movement-derived islands. What is different is that using these
observations in Ch’ol points towards a particular analysis of word order.
The analyses in this dissertation raise some testable predictions with respect to verb-initial
word order and the syntactic and semantic properties of nominals. Some that I listed in Chapter 5
include the scope and islandhood of objects. Verb-initial languages, I argue, do not form a unified
syntactic class. While some properties are quite consistent across verb-initial languages, other
properties are not. One major difference across these languages is case marking. For instance,
VSO languages such as Standard Arabic and Irish are nominative-accusative aligned. I do not
claim that the approach from Chapter 5 for VSO in Q’anjob’al would apply to Standard Arabic or
Irish. Indeed, my formalization of the direction of specifiers is dependent on ergative case.
Moving forward, of particular interest is the investigation of other word order variations
within the Mayan language family. One such language is Chuj, a Mayan language of Guatemala
and Mexico, which, like Ch’ol, exhibits VOS word order as in (310). In (310), however, we see that
there is a full DP object—the proper name Xun, accompanied by an obligatory nominal classifier
waj. This is in contrast to Ch’ol, where it was shown that proper names were not allowed in VOS
object position.
185
(310) Chuj VOS (San Mateo Ixtatán dialect)
Ix-y-il [O waj Xun ] [S ix Malin ].
PFV-A3-see N.CLF Xun N.CLF Malin
‘Malin saw Xun.’ Adapted from Royer (To Appear) with my bracketing
On the one hand, a prosodic account from Clemens and Coon (2018) would face challenges
analyzing (310), as VOS objects are full DPs, and therefore the prosodic reordering of them with
the verb would not be expected. On the other hand, base-generating VOS could account for (310)
though further investigation is needed, especially with respect to the syntactic and structural prop-
erties of the nominal arguments. For instance, does object shift occur in (310) at all? The placement
of adjuncts and the syntactic and semantic properties of the VOS object are important in order to
determine the right syntax for (310). Further examination of typologically diverse languages can
thus provide us with insight on the relationship between syntactic dependencies and their lin-
earization.
186
Kty’añ II
Muk’ikax imejlel jts’ijbu’beñ ityijikñäyel ili kty’añ,
mi kmel yik’oty xpapañichim.
Muk’ikax imejlel ipits’chokoñ ilojwel ili kty’añ,
mi kmel che’ tyi yoralel tsäñal cha’an che’ jiñi mi imek’ob ibä lakbäkel.
Juana Karen Peñate Montejo
Printed with permission of the poem’s author.
187
APPENDIX A
GLOSSEDNARRATIVE: XI’BA
The narrator of this story is Evelina Arcos López, who speaks the Tumbalá dialect of Ch’ol
spoken in San Miguel, Municipio de Salto de Agua, Chiapas, Mexico. This story was recorded in
June 2016 by Carol-Rose Little, who later transcribed, glossed and translated it with the help of
Morelia Vázquez Martı́nez and Shenia Arcos Álvaro. The following is a story about some farmers
who were coming home at night and encounter a pack of xi’bas eating an elderly man they have
just killed. Xi’bas are demons or evil spirits which resemble various animals, making it difficult to
tell whether an animal is really an animal or an evil xi’ba.1
(1) Añ abix xcholelob.
Añ a=bi=x x-cholel-ob.
EXT PRT=REP=already NC-field-PL
‘There were some farmers.’
(2) Ta’bi sujtyiyob tyilel, i’ik’ix abi.
Ta’=bi sujty-i-yob tyil-el, i’ik’=ix a=bi.
PFV=REP return-IV-PL DIR:come-NML black=already PRT=REP
‘They were returning at night.’
(3) I’ik’ix ya ta’ tyiyiyob tyi bij.
I’ik’=ix ya’ ta’ tyiy-i-yob tyi bij.
black=already there PFV come-IV-PL PREP way
‘It was already dark when they were coming down the path.’
(4) Ya’ a ta’ ik’ayob tyi bij a jiñ xcholelob.
Ya’ a ta’ ik’a-y-ob tyi bij a jiñ x-cholel-ob.
There PRT PFV spend.night-IV-PL PREP way PRT DET NC-field-PL
‘The farmers spent the night on the road.’
(5) Che’ abi joch otyoty.
Che’ a=bi joch otyoty.
PRT PRT=REP unoccupied house
‘There was an empty house.’
1Glossing notes: I gloss a as a particle (PRT). It often appears with nouns but also can host clitics, as
seen with abi, the participle with the reportative enclitic. It is unclear if the a which appears with clitics is
the same awhich precedes nouns. I suggest in Chapter 2 that a could come from the topic marker, which is
also a. Arcos López (2009) glosses it as a filler word. Nevertheless, it is of note that the distribution of a, as
in this story, is a distinctive feature of the Tumbalá dialect.
188
(6) Ya’ añ tabla tyi ijol otyotyi.
Ya’ añ tabla tyi i-jol otyoty=i.
there EXT board PREP A3-head house=ENCL
‘There was a board on top of the house.’
(7) A che jiñi ta’ abi letsiyob a uxtyikil wiñik.
A che jiñi ta’ a=bi lets-i-y-ob a ux-tyikil wiñik.
PRT PRT=REP det PFV PRT=REP climb-IV-PL PRT three-CLF.HUM man
‘Then the three men climbed up.’
(8) A che jiñi mach abi a wok iña’tyañob majk tyal a wiñik-ob.
A che jiñi mach a=bi a wo-k i-ña’ty-añ-ob majk tyal a wiñik-ob.
PRT PRT DET NEG PRT=REP PRT prog=IRR A3-think-DTV-PL who come PRT man-PL
‘And, the men weren’t thinking who would be coming.’
(9) Ya’ ta’=bi letsiyob tyi wäyel.
Ya’ ta’=bi lets-i-yob tyi wäy-el.
there PFV=REP ascend-IV-PL PREP sleep-NML
‘They went up it to sleep.’
(10) A che ta’ utyi chämel.
A che ta’ uty-i chäm-el.
PRT PRT PFV finish-IV die-NML
‘A death had just happened.’
(11) Ta’bi chämi wiñik.
Ta’=bi chäm-i wiñik.
PFV=REP die-IV man
‘A man had died.’
(12) Jixmeku a xi’ba wo ikuchob tyi a we’el.
Ji=x=me=ku a xi’ba wo i-kuch-ob tyi a we’el.
DET=AFF=PRED=ASV PRT xi’ba PROG A3-carry-PL DIR:come PRT meat
‘Some xi’bas were coming towards them carrying some meat.’
(13) Wo ikuchob tyi iwe’el a xi’ba.
Wo i-kuch-ob tye i-we’el a xi’ba.
PROG A3-carry-PL DIR:come A3-meat PRT xi’ba
‘The xi’bas are coming towards them carrying some meat.’
(14) Pej xityil abi wo ab ijol a tyi lum.
Pej xity-il a=bi wo ab i-jol a tyi lum.
but mouth.down-STAT PRT=REP PROG PRT.REP A3-head PRT PREP ground
‘They were facing mouth-down with their heads to the ground.’
189
(15) Wo abi ijok’chokob ich’ajañ tyi ikejlab.
Wo a=bi i-jok’-chok-ob i-ch’ajañ tyi i-kejlab.
PROG PRT=REP A3-hang-throw-PL A3-string PREP A3-shoulder
‘They were hanging the string of their bag on their shoulder.’
(16) Iyok aba’ila.
I-yok aba’ila.
A3-leg like.this
‘His leg (leg of the dead man) was like this.’
(17) Ye a päkpäktyalkob a xwäyelob tyi ipam lamina, tabla abi.
Ye a päk-päk-tyak-ob a x-wäyel-ob tyi i-pam lamina, tabla a=bi.
there PRT lying.down-RED-PLIN-PL PRT NC-sleep-PL PREP A3-top roof, board PRT=REP
‘The sleeping men were lying face-down on top of the roof, on the boards.’
(18) Mach abi añik ibojtye’elel a otyotyi.
Mach a=bi añ=ik i-bojtye’-el-el a otyoty=i.
NEG PRT=REP EXT=IRR A3-fence-RS-RS PRT house=ENCL
‘There was no wooden wall around the house.’
(19) Mach abi ibojtye’elel.
Mach a=bi i-bojtye’-el-el.
NEG PRT=REP A3-fence-RS-RS
‘It didn’t have a wooden fence.’
(20) A che jiñi ye’jach a päkpäktyakob a wiñik=ob.
A che jiñi ye’=jach a päk-päk-tyak-ob a wiñik=ob.
PRT PRT DET PRT=EXCL PRT lying.face.down-RED-PLIN-PL PRT man-PL
‘The men were just lying on their stomachs.’
(21) Che jiñi ta’ abi tyiliyob a xi’bajob.
Che jiñi ta’ a=bi tyil-i-yob a xi’ba-job.
PRT DET PFV PRT=REP A3-come-IV-PL PRT xi’ba-PL
‘Then the xi’bas came.’
(22) K’otyi abi ijijsañob.
K’oty-i a=bi i-jij-s-a-ñ-ob.
arrive-IV PRT=REP A3-rest-CAU-DTV-NML-PL
‘They came to put something down.’
(23) Xxxx che abi k’otyel a xi’ba.
Xxxx che a=bi k’oty-el a xi’ba.
Shhh PRT PRT=REP arrive-NML PRT xi’ba
‘Shhh was how the xi’bas arrived.’
190
(24) Polok’ che a k’otyel a wiñiki.
Polok’ che a k’oty-el a wiñik=i.
plop PRT PRT arrive-NML PRT man=ENCL
‘The [dead] man plopped to the ground.’
(25) Päkpäktyakob wo abi ik’elob tyi lum a jiñi
Päk-päk-tyak-ob wo a=bi i-k’el-ob tyi lum a jiñi
lying.down-red-PLIN-PL PROG PRT=REP A3-see-PL PREP ground PRT DET
lakpi’älob.
la-k-pi’äl-ob.
PART.PL-A1-friend-PL
‘The guys were watching the ground as they were lying down.’
(26) Tax a kaj ijochbeñob ibujk a wiñiki.
Ta=x a kaj i-joch-b-eñ-ob i-bujk a wiñik=i.
PFV=already PRT start A3-take.off-APPL-DTV-PL A3-shirt PRT man=ENCL
‘They started to take off the man’s shirt.’
(27) Tax kaji ik’uxob.
Ta=x kaj-i i-k’ux-ob.
PFV=already start-IV A3-eat-PL
‘They started to eat him.’
(28) Ta’ abi i-jochbeyob ibujk.
Ta’ a=bi i-joch-b-e-yob i-bujk.
PFV PRT=REP A3-take.off-APPL-DTV-PL A3-clothes
‘They took off his shirt.’
(29) Ta’ abi ich’ojoyob.
Ta’ a=bi i-ch’oj-o-yob.
PFV PRT=REP A3-munch-TV-PL
‘They munch into him.’
(30) Ich’ojbeyob ijol.
I-ch’oj-b-e-yob i-jol.
A3-bite-APPL-DT-PL A3-head
‘They munched into his head.’
(31) Ta’ ilok’be isotya.
Ta’ i-lok’-b-e i-sotya.
PFV A3-take.out-APPL-DT A3-innards
‘They took out his innards.’
(32) Ya buchul laktatuchi.
Ya buch-ul la-k-tatuch=i.
there sit-STAT PART.PL-A1-grandfather=ENCL
‘The old man was just sitting there.’
191
(33) Buchul aba laktatuch.
Buch-ul aba la-k-tatuch.
seated-STAT PRT PART.PL-A3-grandfather
‘The old man was just sitting there.’
(34) Ta’ abi i-lok’-b-e iyaty bajche’ wo awäli.
Ta’ a=bi i-lok’-b-e iy-aty bajche’ wo aw-äl=i.
PFV PRT=REP A3-take.out-APPL-DTV A3-penis like PROG A2-say=ENCL
‘They took off his penis as you were saying.’
(35) Ta’ abi ilok’be iyatyi.
Ta’ a=bi i-lok’-b-e iy-aty=i.
PFV PRT=REP A3-take.out-appl-dt A3-penis=ENCL
‘They took off his penis.’
(36) Wäle icha’an laktatuch ilili.
Wäle i-cha’an la-k-tatuch ilili.
now A3-PREP PART.PL-A1-grandfather here
‘The old man’s.’
(37) Wo abi ijich’-b-e che’i.
Wo a=bi i-jich’-b-e che’i.
PROG PRT=REP A3-hobble-APPL-DTV PRT
‘They were tying up his feet.’
(38) A che jiñi tyulul abi iwutyob a lakpiälob.
A che jiñi tyul-ul a=bi i-wuty-ob a la-k-piäl-ob.
PRT PRT DET fixed-STAT PRT=REP A3-face-PL PRT PART.PL-A1-friend-PL
‘The men had their eyes fixed on them.’
(39) K’uxux tatuch, cheyob abi.
K’ux-u-x tatuch che-yob a=bi.
eat-IMP=already grandfather PRT-PL PRT=REP
‘Eat the old man, they said.’
(40) Ta’ bä’kaj ik’ux laktatuch.
Ta’ bä’-kaj i-k’ux la-k-tatuch.
PFV then-start A3-eat PART.PL-A1-grandfather
‘They then started to eat the old man.
(41) Ta’ ab ik’uxu a laktatuch.
Ta’ ab i-k’ux-u a la-k-tatuch.
PFV REP A3-eat-TV PRT PART.PL-A1-grandfather
‘They ate the old man.’
192
(42) Jäch jäch jäch jäch ik’ux a laktatuch.
Jäch jäch jäch jäch i-k’ux a la-k-tatuch.
nom nom nom nom A3-eat PRT PART.PL-A1-grandfather
‘Nom, nom, nom, nom they ate the old man.
(43) Weñ sumuk abi!
Weñ sumuk a=bi!
very tasty PRT=REP
‘Very tasty!’
(44) Weñ sumuk che abi laktatuchi.
Weñ sumuk che a=bi la-k-tatuch=i.
very tasty PRT PRT=REP PART.PL-A1-grandfather=ENCL
‘This old many is very tasty they said.’
(45) Wejlañixku a lajk’ajki che abi yälä a
Wejl-añ=ix=ku a la-j-k’ajk=i che a=bi y-äl-ä a
blow-DTV=already=AFF PRT PART.PL-A1-fire=ENCL PRT PRT=REP A3-say-DTV PRT
x-ku.
x-ku.
NC-owl
‘Blow on the fire, the owl said.’
(46) Ta’ abi ibä’wejla a k’ajk a xku.
Ta’ a=bi i-bä’-wejl-a a k’ajk a x-ku.
PFV PRT=REP A3-then-blow-DTV PRT fire PRT NC-owl
‘The owl blew on the fire himself.’
(47) Che jiñixmeku a jiñi bajlum. . . lekojix panmil chejob
che jiñi=x=me=ku a jiñi bajlum. . . lekoj=ix panmil chej=ob
PRT DET=already=PRED=ASV PRT DET jaguar disagreeable=already universe PRT-PL
xi’bajob.
xi’baj-ob.
xi’ba-PL
‘So, the jaguar said, something is off...said the xi’bas.’
(48) Chuk añ?
Chuk añ?
what EXT
‘What’s wrong?’
(49) Lekoj=ix iy-uts’il panmil=i.
Lekoj=ix iy-uts’il panmil=i.
disagreeable=already A3-smell universe=ENCL
‘Something is off.’
193
(50) Wox ab ijamob its’ak a lakpiälobi.
Wo=x ab i-jam-ob i-ts’ak a la-k-piäl-ob=i.
PROG=already REP A3-open-PL A3-medicine PRT PART.PL-A1-friend-PL=ENCL
‘The men opened up some medicine.’
(51) Che’jax a bajche’ iyuts’il pajch’...pajch’i yu’bil.
Che’=jax a bajche’ iy-uts’il pajch’...pajch’=i y-u’bil.
part=EXCL PRT like A3-smell pineapple...pineapple=ENCL A3-feel
‘It smels like pineapple. . . like pineapple.’
(52) Pajch’ cheyob abi a xi’ba.
Pajch’ che-yob abi a xi’ba.
pineapple PRT=PL PRT=REP PRT xi’ba
‘Like pineapple, said the xi’bas.’
(53) Xxxx chejob a xi’ba.
Xxxx che-job a xi’ba.
shhh part-PL PRT xi’ba
‘Shhh said the xi’bas.’
(54) La’ lajk’e bajche’ yilal k’uxtyalob.
la’ la-j-k’e bajche’ yi-lal k’ux-tyäl-ob.
let’s PART.PL-A1-see how A3-seem eat-PSV.NML-PL
‘Let’s see how to eat them.’
(55) Che jiñi lekux abi wo ibä’ñañob a
Che jiñi leku=x a=bi wo i-bä’ñ-añ-ob a
PRT DET disagreeable=already PRT=REP PROG A3-fear-DTV-PL PRT
lakpiälobi año’bä tyi ipam otyoty.
la-k-piäl-ob=i añ-o’=bä tyi i-pam otyoty
PART.PL-A1-friend-PL=ENCL EXT-PL=REL PREP A3-top house
‘And the men on top of house were very afraid of them.’
(56) Wolobix ata ibä’ñañob kome bajche’ aki ju’belob
Wol-ob=ix a=ta i-bä’ñ-añ-ob kome bajche’ a=ki ju’b-el-ob
prog-PL=already PRT=REA A3-fear-DTV-PL because how PRT=INTER get.down-NML-PL
kome i’ik’ix abi.
kome i’ik’=ix a=bi.
because black=already PRT=REP
‘They were very scared as how were they going to get down, as it was already night.’
(57) Mach añobik ipoko.
Mach añ-ob-ik i-poko.
NEG EXT-PL-IRR A3-flashlight
‘They didn’t have a flashlight.’
194
(58) Che jiñix ameku jiñobjax a lakpiälobi ta
Che jiñi=x a=me=ku jiñ-ob-jax a la-k-piäl-ob=i ta’
PRT DET=already PRT=PRED=ASV det-PL=EXCL PRT PART.PL-A1-friend-PL=ENCL PFV
abi ijamäyob its’aki.
a=bi i-jam-ä-yob i-ts’ak=i.
PRT=REP A3-open-DTV-PL A3-medicine=ENCL
‘Then the men opened up the medicine.’
(59) Tamaku iwelchoko... iwelchokota ju’bel.
Ta=ma=ku i-wel-choko... i-wel-chok-o=ta ju’b-el.
PFV=PRE=ASV A3-throwADV-throw-TV A3-throwADV-throw-TV=REA DIR:descend-NML
‘They threw it down.’2
(60) Ta abila yajlyiob ju’bel a xi’ba.
Ta abila yajl-i-yob ju’b-el a xi’ba.
PFV like.this fall-IV-PL DIR:down-NML PRT xi’ba
‘The xi’bas fell down.’
(61) Woxol a jiñi ts’ak mu’bä ik’äñob.
Woxol a jiñi ts’ak mu’=bä i-k’äñ-ob.
round.shape PRT DET medicine IPFV=REL A3-use-PL
‘The medicine they used was ball-like.’
(62) Piäläl, mutyo imel kmama wajali.
Piäl-äl mu-tyo i-mel k-mama wajali.
friend-DEINAL IPFV=still A3-make A1-mother back.then
‘My mother used to make this kind of medicine.’
(63) Woxol mi ijuch’ tyi molino.
Woxol mi i-juch’ tyi molino.
round.shape-STAT IPFV A3-grind PREP grinder
‘She grinds it into a round ball in the grinder.’
(64) Mi ijuch’ che mi iwojxiñ bajche’ ilili.
Mi i-juch’ che mi i-wojx-iñ bajche’ ilili.
IPFV A3-grind PRT PFV A3-ball-DTV.NML like this
‘She grinds it then balls it up, like this.’
(65) Mutyo iweñk’ux kmama wajalix.
Mu=tyo i-weñ-k’ux kmama wajali=x.
IPFV=still A3-very-eat A1-mother back.then=already
‘My mother used to eat it a lot, back then.’
2The adverbial preverb wel relates to the form in which an object is thrown.
195
(66) Che’ bajche’ tyañi.
Che’ bajche’ tyañ=i.
PRT like lime=ENCL
‘Like a lime.’
(67) Che amek ta ikujuyobi ubiyi.
Che a=me=k ta i-kuj-u-y-ob=i ubiyi.
PRT PRT=PRED=ASV PFV A3-hit-DTV-PL=ENCL they.say
‘Then it hit them, they say.’
(68) Che’ ameki ta’ ichokoyob ju’bel.
Che’ a=me=k=i ta’ i-chok-o-yob ju’b-el.
PRT PRT=PRED=ASV=ENCL PFV A3-throw-DTV-PL DIR:down-NML
‘They threw it down.’
(69) Tameki yajliyob a xi’ba.
Ta=me=k=i yajl-i-yob a xi’ba.
PFV=PRED=ASV=ENCL fall-IV-PL PRT xi’ba
‘The xi’bas fell.’
(70) Puts’iyob a wiñiki.
Puts’-i-yob a wiñik=i.
flee-IV-PL PRT man=ENCL
‘The men fled.’
(71) Wuu wuu a wiñiki.
wuu wuu a wiñik=i.
whoa whoa PRT man=ENCL
‘The men were like whoa...’
(72) Puts’iyob amek majlel.
Puts’-i-yob a=me-k majl-el.
flee-IV-PL PRT=PRED=ASV DIR:go-NML
‘They fled away.’
(73) Majliyob abi wiñikob.
Majl-i-yob a=bi wiñik-ob.
go-IV-PL PRT=REP man-PL
‘The men went away.’
(74) Tyo’ochejax abi ta’ majliyobi kome mach añobiki
Tyo’o-che=jax a=bi ta’ majl-i-yob=i kome mach añ-ob-ik=i
EXCL-PART=EXCL PRT=REP PFV go-IV-PL=ENCL because NEG EXT-PL=IRR=ENCL
ipoco.
i-poco.
A3-flashlight
‘They just went away because they did have a flashlight.’
196
(75) Che’eñ cha’ka a tyi ty’an a waxi.
Che’eñ cha’-ka a tyi ty’an a wax=i.
then again-start PRT PREP noise PRT grey.fox=ENCL
‘The grey fox started to make noise.’
(76) Jix ameku a waxi ta’ ka iyajñesañob
Ji=x a=me=ku a wax=i ta’ ka iy-ajñ-es-añ-ob
DET=already PRT=PRED=ASV PRT grey.fox=ENCL PFV start A3-run-CAU-DTV-PL
majlel.
majl-el.
DIR:go-NML
‘The grey fox went after them.’
(77) Ta’bi iyajñesa wax.
Ta’=bi iy-ajñ-es-a wax.
PFV=REP A3-run-CAU-DTV PRT grey.fox
‘The grey fox went after them.’
(78) Wax che’ a baje ts’i’.
Wax che’ a baje ts’i’.
grey.fox PRT PRT like dog
‘Grey foxes are like dogs.’
(79) Che bajche’ mis.
Che bajche’ mis.
PRT like cat
‘Like a cat.’
(80) Juñachich bajche’ mis a wax.
Juña=jach=ix bajche’ mis a wax.
quite=EXCL=already like cat PRT grey.fox
‘These grey foxes look like cats.’
(81) Pero xi’ba abi je wax.
Pero xi’ba a=bi je wax.
but xi’ba PRT=REP DET grey.fox
‘But the grey fox was also a xi’ba.’
(82) Chejiñi ya’ lätsä iximi.
Chejiñi ya’ läts-äl ixim=i.
then there stacked-STAT corn=ENCL
‘There was some stacked corn.’
(83) Ya ta’ majiyob tyi puts’el ili wiñikob ba ixim.
Ya’ ta’ maj-i-yob tyi puts-’el ili wiñik-ob ba’ ixim.
there PFV go-IV-PL PREP flee-NML this man-PL where corn
‘The men fled into the corn.’
197
(84) Che’ bajche’ añbä tyi a-kwarto.
Che’ bajche’ añ=bä tyi a-kwarto.
PRT like EXT=REL PREP A2-room
‘Like there is in your room.’
(85) Ta’ ab ochi a lakpi’äl ba’ iximi.
Ta’ ab och-i a la-k-pi’äl ba’ ixim=i.
PFV REP enter-IV PRT PART.PL-A1-friend where corn=ENCL
‘The men went into the corn.’
(86) Ta’ abi ijok’oyob ixim a wiñikob.
ta’ a=bi i-jok’-o-yob ixim a wiñik-ob.
PFV PRT=REP A3-dig-TV-PL corn PRT men-PL
‘The men dug into the corn.’
(87) Mach ata yomobik k’uxtyäl che’.
Mach a=ta y-om-ob-ik k’ux-tyäl che’.
NEG PRT=REA A3-want-PL-IRR eat-PSV.NML PRT
‘They certainly did not want to be eaten.’
(88) Jok’oyo’ abi a iximi.
Jok’-o-y-o’ a=bi a ixim=i.
dig-tv-PL PRT=REP PRT corn=ENCL
‘They dug into the corn.’
(89) P’uchu aba ixim ta’ k’otyi tyi iwuty a lakpi’älob.
p’uch-ul aba ixim ta’ k’oty-i tyi i-wuty a la-k-pi’äl-ob.
amass-STAT like corn PFV arrive-IV PREP A3-face PRT PART.PL-A1-friend-PL
‘The corn was amassed in a pile and it got into the faces of the men.’
(90) Añ abi yalä machity a ixim.
Añ abi y-alä machity a ixim.
EXT rep A3-DIM machete PRT corn
‘The corn had a machete.’
(91) Añ abi i-ch’ujlel a ixim.
Añ a=bi i-ch’ujlel a ixim.
EXT PRT=REP A3-spirt PRT corn
‘The corn has a spirit (it was alive).’
(92) Ts’inlaw abi bajche’ machity ta’ i-mel-e a ixim.
Ts’inlaw a=bi bajche’ machity ta’ i-mel-e a ixim.
swish PRT=REP like machete PFV A3-do-TV PRT corn
‘The corn went ‘swish’ with the machete.’
198
(93) Ta’ abi i-koty-a-y-ob a jiñ wiñik-ob.
Ta’ a=bi i-koty-a-y-ob a jiñ wiñik-ob.
PFV PRT=REP A3-help-DTV-PL PRT DET man-PL
‘It helped the men.’
(94) Che’ a baj kwentu bajche’ jiñi.
che a baj kwentu bajche’ jiñi.
PRT PRT like.this story like DET
‘That’s the story.’
(95) Che’ bajche’ jiñi.
che’ bajche’ jiñi.
PRT like DET
‘That’s it.’
199
BIBLIOGRAPHY
Abney, S. P. (1987). The English noun phrase in its sentential aspect. Doctoral dissertation, Mas-
sachusetts Institute of Technology, Cambridge, MA.
Adger, D. (2003). Core Syntax. Oxford University Press, Oxford.
Aissen, J. (1979). Possessor ascension in Tzotzil. In Martin, L., editor, Papers in Mayan Linguistics,
pages 89–108. Lucas Brothers, Columbia, MO.
Aissen, J. (1987). Tzotzil Clause Structure. Dordrecht, Reidel.
Aissen, J. (1992). Topic and focus in Mayan. Language, 68(1):43–80.
Aissen, J. (1994). Tzotzil auxiliaries. Linguistics, 32(657–690).
Aissen, J. (1996). Pied-piping, abstract agreement, and functional projections in Tzotzil. Natural
Language & Linguistic Theory, 14(3):447–491.
Aissen, J. (2009). Depictives and serialization in Tzotzil. In Donna B. Gerdts, J. C. M. and Polinsky,
M., editors, Hypothesis A/Hypothesis B: Linguistic explorations in honor of David M. Perlmutter,
Current Studies in Linguistics, pages 1–18. MIT Press, Cambridge.
Aldridge, E. (2002). Nominalization and wh-movement in Seediq and Tagalog. Language and
Linguistics, 3(2):393–426.
Aldridge, E. (2004). Ergativity and word order in Austronesian languages. Doctoral dissertation,
Cornell University, Ithaca, NY.
Aldridge, E. (2008). Generative approaches to ergativity. Language and Linguistic Compass: Syntax
and Morphology, 2(5):966–995.
Aldridge, E. (2012). Antipassive and ergativity in Tagalog. Lingua, 122(3):192–203.
Alexiadou, A., Haegeman, L., and Stavrou, M. (2007). Noun phrase in the generative perspective.
Mouton de Gruyter, Berlin.
200
Andrews, A. D. (1996). Semantic case-stacking and inside-out unification. Australian Journal of
Linguistics, 16(1):1–55.
Arcos López, N. (2009). Los clasificadores numerales y las clases nominales en chol. Master’s thesis,
CIESAS, Mexico.
Arkoh, R. and Matthewson, L. (2013). A familiar definite article in Akan. Lingua, 123:1–30.
Assmann, A., Edygarova, S., Georgi, D., Klein, T., and Weisser, P. (2014). Case stacking below the
surface: On the possessor case alternation in Udmurt. The Linguistic Review, 31(3-4):447–485.
Assmann, A., Georgi, D., Heck, F., Müller, G., and Weisser, P. (2015). Ergatives move too early: On
an instance of opacity in syntax. Syntax, 18(4):343–387.
Attinasi, J. (1973). Lak T’an: a Grammar of the Chol (Mayan) Word. Doctoral dissertation, University
of Chicago, Chicago.
Attinasi, J. (1976). Lexicality and grammaticality in Chol Mayan. In McClaran, M., editor, Mayan
Linguistics Volume One, pages 212–227. University of California, Los Angeles.
Aulie, W. and Aulie, E. (1978). Diccionario Ch’ol-Español, Español-Ch’ol. Summer Institute of Lin-
guistics, México, third edition.
Aydemir, Y. (2004). Are Turkish preverbal bare nouns syntactic arguments? Linguistic Inquiry,
35(3):465–474.
Bach, E., Jelinek, E., Kratzer, A., and Partee, B. H., editors (1995). Quantification in natural languages,
volume 54. Springer Science & Business Media.
Baker, M. (1985). The mirror principle and morphosyntactic explanation. Linguistic Inquiry,
16(3):373–416.
Baker, M. (1996). The Polysynthesis Parameter. Oxford University Press, Oxford.
Baker, M. C. (1988). Incorporation: A Theory of Grammatical Function Changing. University of
Chicago Press, Chicago.
201
Bale, A. and Coon, J. (2014). Classifiers are for numerals, not for nouns: Consequences for the
mass/count distinction. Linguistic Inquiry, 45(4):695–707.
Bale, A., Coon, J., and Arcos, N. (2019). Classifiers, partitions, and measurements: Exploring the
syntax and semantics of sortal classifiers. Glossa: a journal of general linguistics, 4(1).
Barker, C. (1995). Possessive descriptions. CSLI, Stanford, CA.
Barker, C. (2011). Possessives and relational nouns. In Claudia Maienborn, Klaus von Heusinger,
P. P., editor, Semantics: An international handbook of natural language meaning, pages 1109–1130. de
Gruyter, Berlin.
Bennett, R., Coon, J., and Henderson, R. (2015). Introduction to Mayan linguistics. Language and
Linguistics Compass, 10:1–14.
Bercerra, M. (1937). Vocabulario de la lengua chol. Secretarı́a de Educación Pública, Mexico.
Besnier, N. (2000). Tuvaluan: A Polynesian language of the Central Pacific. Routledge, New York and
London.
Bhatt, R. and Anagnostopoulou, E. (1996). Object shift and specificity: Evidence from ko-phrases
in Hindi. In Papers from the 32nd Regional Meeting of the Chicago Linguistics Society, Chicago Lin-
guistics Society, Chicago, volume 1122.
Bittner, M. (1994). Case, Scope, and Binding. Kluwer, Dordrecht.
Bošković, Ž. (2005). On the locality of left branch extraction and the structure of NP. Studia
Linguistica, 59(1):1–45.
Bošković, Ž. (2008). What will you have, DP or NP? In Elfner, E. and Walkow, M., editors, Proceed-
ings of the Northeast Linguistic Society 37, pages 101–114, Amherst, MA. GLSA.
Branan, K. (2018). Attraction at a distance: Ā-movement and case. Linguistic Inquiry, 49(3):409–440.
Bricker, V. R. (1986). A grammar of Mayan hieroglyphs, volume 56. Middle American Research
Institute, Tulane University, Tulane.
202
Brody, J. (1984). Some problems with the concept of basic word order. Linguistics, 22(5):711–736.
Bruening, B. (2009). Selectional asymmetries between CP and DP suggest that the DP hypothesis
is wrong. In MacKenzie, L., editor, Proceedings of the 32nd Annual Penn Linguistics Colloquium,
volume 15, pages 26–35.
Bruening, B. (2010). Ditransitive asymmetries and a theory of idiom formation. Linguistic Inquiry,
41(4):519–562.
Butler, L. (2011). The DP-adjoined plural in Yucatec Maya and the syntax of plural marking.
manuscript, University of Arizona/UC Santa Cruz.
Carlson, G. (2003). Weak indefinites. In Coene, M. and D’hulst, Y., editors, From NP to DP, vol-
ume 1, pages 195–210. John Benjamins, Amsterdam/Philadelphia.
Carlson, G. N. (1977). Reference to Kinds in English. Doctoral dissertation, University of Mas-
sachusetts, Amherst.
Chierchia, G. (1998a). Plurality of mass nouns and the notion of “semantic parameter”. In Roth-
stein, S., editor, Events and grammar, pages 53–103. Kluwer, Dordrecht.
Chierchia, G. (1998b). Reference to kinds across language. Natural Language Semantics, 6(4):339–
405.
Chomsky, N. (1973). Conditions on transformations. In Anderson, S. R. and Kiparsky, P., editors,
A festschrift for Morris Halle, pages 232–286. Holt, New York.
Chomsky, N. (1981). Lectures on government and binding. Foris, Dordrecht.
Chomsky, N. (1986). Barriers. MIT press, Cambridge, MA.
Chomsky, N. (1995). The minimalist program. MIT press, Cambridge, MA.
Chomsky, N. (2000). Minimalist inquiries: The framework. In Martin, R., Michaels, D., and
Uriagereka, J., editors, Step by step: Essays on minimalist syntax in honor of Howard Lasnik, pages
89–156. MIT Press.
203
Chomsky, N. (2001). Derivation by phase. In Kenstowicz, M., editor, Ken Hale: A Life in Linguistics,
pages 1–52. MIT Press, Cambridge, Massachusetts.
Chomsky, N. (2008). On phases. In Freidin, R., Otero, C. P., and Zubizarreta, M.-L., editors,
Foundational issues in linguistics, pages 133–166. MIT press, Cambridge, MA.
Chung, S. (1998). The design of agreement: Evidence from Chamorro. University of Chicago Press,
Chicago.
Chung, S. (2005). What fronts? On the VP-raising account of verb-initial order. In Carnie, A.,
Harley, H., and Dooley, S. A., editors, Verb first: On the syntax of verb initial languages, pages
9–29. John Benjamins, Amsterdam, Philadelphia.
Chung, S. (2006). Properties of VOS languages. In Everaert, M. and van Riemsdijk, H., editors,
The Blackwell companion to syntax, pages 685–720. Blackwell Publishing Ltd, Oxford.
Chung, S. and Ladusaw,W. A. (2004). Restriction and Saturation, volume 42. MIT press, Cambridge,
MA.
Clemens, L. (2019). Prosodic noun incorporation: The relationship between prosody and argu-
ment structure in niuean. Syntax, 22(4):337–377.
Clemens, L. and Coon, J. (2018). Deriving verb-initial word order in Mayan. Language, 94(2):237–
280.
Clemens, L., Coon, J., Little, C.-R., and Martı́nez, M. V. (2017). Focus in Ch’ol. Presented at The
Society for the Study of the Indigenous Languages of the Americas (SSILA) 2017.
Clemens, L. E. (2014). Prosodic noun incorporation and verb-initial syntax. Doctoral dissertation,
Harvard, Cambridge, MA.
Clemens, L. E. and Polinsky, M. (2017). Verb-initial word orders, primarily in Austronesian and
Mayan languages. In Evaraert, M. and van Riemsijk, H., editors, The Wiley Blackwell Companion
to Syntax, pages 1–50. Wiley Online Library, 2 edition.
204
Collins, J. N. (2019). Definiteness determined by syntax: A case study in Tagalog. Natural Language
& Linguistic Theory, 37(4):1367–1420.
Comrie, B. (1976). Aspect: An introduction to the study of verbal aspect and related problems, volume 2.
Cambridge University Press, Cambridge.
Coon, J. (2006). Existentials and negation in Chol (Mayan). In Chang, C., Dugarova, E., Theodor-
opoulou, I., Beltrán, E. V., and Wilford, E., editors, CamLing: Proceedings of the Fourth University
of Cambridge Postgraduate Conference in Language Research, pages 51–58, Cambridge. Cambridge
Institute of Language Research.
Coon, J. (2009). Interrogative possessors and the problem with pied-piping in Chol. Linguistic
Inquiry, 40(1):165–175.
Coon, J. (2010a). Complementation in Chol (Mayan): A Theory of Split Ergativity. Doctoral dissertation,
Massachusetts Institute of Technology, Cambridge, MA.
Coon, J. (2010b). VOS as predicate fronting in Chol. Lingua, 120:354–378.
Coon, J. (2012). Split ergativity and transitivity in Chol. Lingua, 122:241–256.
Coon, J. (2013). Aspects of split ergativity. Oxford University Press.
Coon, J. (2017). Little-v0 agreement and templatic morphology in Ch’ol. Syntax, 20(2):101–137.
Coon, J. (2018). Distinguishing adjectives from relative clauses in Chuj (with help from Ch’ol).
Heading in the right direction: Linguistics treats for Lisa Travis.
Coon, J. (2019). Building verbs in Chuj: Consequences for the nature of roots. Journal of Linguistics,
55(1):35–81.
Coon, J., Baier, N., and Levin, T. (2020). Mayan agent focus and the ergative extraction constraint:
Facts and fictions revisited. manuscript McGill University.
Coon, J. and Henderson, R. (2010). Two binding puzzles in Mayan. In Gutiérrez Bravo, R.,
Mikkelsen, L., and Potsdam, E., editors, Representing Language: Essays in Honor of Judith Ais-
sen.
205
Coon, J., Pedro, P. M., and Preminger, O. (2014). The role of case in A-bar extraction asymmetries:
Evidence from Mayan. Linguistic Variation, 14(2):179–242.
Coon, J. and Preminger, O. (2011). Positional roots and case absorption. In Avelino, H., editor,
New Perspectives in Mayan Linguistics, pages 206–241. Cambridge Scholars Publishing, Newcas-
tle upon Tyne.
Coon, J. and Preminger, O. (2013). Transitivity in Chol: A new argument for the split VP hypothe-
sis. In Fainleib, L., LaCara, N., and Park, Y., editors, Proceedings of the 41st Annual Meeting of the
North East Linguistics Society (NELS), pages 127–142, Amherst, MA. GLSA.
Coon, J. and Royer, J. (To Appear). Nominalization and selection in two Mayan languages. In
Alexiadou, A. and Borer, H., editors,Nominalization: 50 Years on from Chomsky’s Remarks. Oxford
University Press, Oxford.
Corbett, G. (2000). Number. Cambridge University Press, Cambridge.
Corver, N. (1992). On deriving certain left branch extraction asymmetries: A case study in para-
metric syntax. In Broderick, K., editor, Proceedings of the 22nd Annual Meeting of the North East
Linguistic Society (NELS), pages 67–84. GLSA, Amherst.
Corver, N. (2006). Freezing effects. In Everaert, M. and van Riemsdijk, H., editors, The Blackwell
companion to syntax, volume 2, pages 383–406. Wiley Online Library.
Davis, H. (1999). Word order and configurationality in St’át’imcets. In 34th international conference
on Salish and Neighboring languages, pages 61–82.
Dayal, V. (2003). A semantics for pseudo incorporation. manuscript, Rutgers University.
Dayal, V. (2004). Number marking and (in)definiteness in kind terms. Linguistics and philosophy,
27(4):393–450.
Dayal, V. (2011). Hindi pseudo-incorporation. Natural Language and Linguistic Theory, 29(1):123–
167.
206
Dayal, V. (2013). On the existential force of bare plurals across languages. In Caponigro, I. and
Cecchetto, C., editors, From grammar to meaning: The spontaneous logicality of language, pages
49–80. Cambridge University Press, Cambridge.
Dayal, V. (2015). Incorporation: Morpho-syntactic vs. semantic considerations. In Borik, O. and
Gehrke, B., editors, The syntax and semantics of pseudo-incorporation, pages 47–87. Brill, Leiden.
de Hoop, H. (1996). Case configuration and noun phrase interpretation. Garland, New York.
Deal, A. R. and Nee, J. (2017). Bare nouns, number, and definiteness in Teotitlán del Valle Zapotec.
In Cummins, C., Gisborne, N., Heycock, C., Rabern, B., Rohde, H., and Truswell, R., editors,
Proceedings of Sinn und Bedeutung, volume 21, pages 317–334, Edinburgh.
Despić, M. (2011). Syntax in the absence of determiner phrase. Doctoral dissertation, University of
Connecticut, Storrs, CT.
Despić, M. (2013). Binding and the structure of NP in Serbo-Croatian. Linguistic Inquiry, 44(2):239–
270.
Despić, M. (2015). Phases, reflexives, and definiteness. Syntax, 18(3):201–234.
Despić, M. (2019). On kinds and anaphoricity in languages without definite articles. In Aguilar-
Guevara, Ana, Pozas Loyo, J., and Vázquez-Rojas Maldonado, V., editors, Definiteness across
languages, chapter 261–292. Language Science Press, Berlin.
Dı́az Peñate, J. (1992). La’ lakts’ijbuñ lakty’añ tyi ch’ol. Gobierno del Estado de Chiapas, Mexico
City.
Diesing, M. (1992). Indefinites. MIT Press, Cambridge, MA.
Diesing, M. (1996). Semantic variables and object shift. Studies in comparative Germanic syntax,
2:66–84.
Diesing, M. and Jelinek, E. (1995). Distributing arguments. Natural Language Semantics, 3(2):123–
176.
207
Dryer, M. S. (2013). Order of subject, object and verb. In Dryer, M. S. and Haspelmath, M.,
editors, The World Atlas of Language Structures Online. Max Planck Institute for Evolutionary
Anthropology, Leipzig.
Ebert, K. (1971a). Referenz, Sprechsituation und die bestimmten Artikel in einem nordfriesischen Dialekt
(Fering). Doctoral dissertation, Christian-Albrechts-Universität zu Kiel, Kiel.
Ebert, K. (1971b). Zwei formen des bestimmten Artikels. In Wunderlich, D., editor, Probleme und
Fortschritte der Transformationsgrammarik, pages 159–174. Hueber.
Embick, D. and Noyer, R. (2001). Movement operations after syntax. Linguistic inquiry, 32(4):555–
595.
England, N. (1976). Mam directionals and verb semantics. Mayan Linguistics, 1:202–211.
England, N. (1983). A grammar of Mam, a Mayan Language. University of Texas Press, Austin.
England, N. (1991). Changes in basic word order in Mayan languages. International Journal of
American Linguistics, 57:446–86.
Ethnologue (2018). Chol. Ethnologue. https://www.ethnologue.com/language/ctu.
Fassi Fehri, A. (1993). Issues in the structure of Arabic clauses and words. Kluwer, Dordrecht.
Gallego, Á. J. (2013). Object shift in Romance. Natural Language & Linguistic Theory, 31(2):409–451.
Geist, L. (2010). Bare singular NPs in argument positions: restrictions on indefiniteness. Interna-
tional Review of Pragmatics, 2(2):191–227.
Georgala, E. (2012). Applicatives in their structural and thematic function: A minimalist account of
multitransitivity. Doctoral dissertation, Cornell University, Ithaca, NY.
Georgala, E., Paul, W., and Whitman, J. (2008). Expletive and thematic applicatives. In Chang,
C. B. and Haynie, H. J., editors, Proceedings of the 26th West Coast Conference on Formal Linguistics,
volume 26, pages 181–189, Somerville, MA. Cascadilla Proceedings Project.
208
Gillon, C. (2015). Investigating D in languages with and without articles. In Matthewson, L. and
Bochnak, R., editors,Methodology in Semantic Fieldwork, pages 175–205. Oxford University Press,
New York.
Gómez Jiménez, S. (2015). Grupo lingüı́stico lakty’añ, ch’ol. In Ochoa Nájera, J. D., editor, Las
lenguas de Chiapas, volume 1, pages 15–20. Consejo Estatal para las Culturas y las Artes y las
Artes de Chiapas, Tutxla Gutiérrez, Chiapas.
Gutiérrez Sánchez, P. (2004). Las clases de verbos intransitivos y el alineamiento agentivo en el
chol de Tila, Chiapas. Master’s thesis, CIESAS, México.
Harley, H. (2017). The “bundling” hypothesis and the disparate functions of little v. In
D’Alessandro, R., Franco, I., and Gallego, A. J., editors, The verbal domain, pages 3–28. Oxford
University Press Oxford, Oxford.
Haviland, J. (1993). The syntax of Tzotzil auxiliaries and directionals: the grammaticalization of
”motion”. In Peterson, D. A., editor, Proceedings of the Nineteenth Annual Meeting of the Berkeley
Linguistics Society, pages 35–49.
Haviland, J. B. (1994). “Ta xa setel xulem” [The buzzers were circling] – categories of verbal roots
in (Zinacantec) Tzotzil. Linguistics, 32(4-5):691–741.
Hellan, L. (1986). The headedness of NPs in Norwegian. In Muysken, P. and van Riemsdijk, H.,
editors, Features and projections, pages 89–122. Foris, Dordrecht.
Henderson, R. (2019). The roots of measurement. Glossa: A journal of general linguistics, 4(1):1–31.
Herburger, E. (2001). The negative concord puzzle revisited. Natural language semantics, 9(3):289–
333.
Holmberg, A. (1986). Word Order and Syntactic Features in Scandinavian languages and English. Doc-
toral dissertation, Stockholm.
Hopkins, N. A., Cruz Guzmán, A., and Josserand, J. K. (2008). A Chol (Mayan) vocabulary from
1789. International Journal of American Linguistics, 74(1):83–114.
209
Hopkins, N. A., Josserand, J. K., and Guzmán, A. C. (2016). Chol (Mayan) folktales: a collection of
stories from the modern Maya of southern Mexico. Colorado State University, Fort Collins, CO.
Houston, S., Robertson, J., and Stuart, D. (2000). The language of Classic Maya inscriptions. Cur-
rent anthropology, 41(3):321–356.
INALI (2011). Ityoj ts’ijbuñtyel lakty’añ ch’ol [Norma de escritura de la lengua ch’ol]. Instituto Nacional
de Lenguas Indı́genas, México.
Jenks, P. (2018). Articulated definiteness without articles. Linguistic Inquiry, 49(3):501–536.
Josserand, J. K. (2001). Chol Ritual Language. FAMSI.
Kaufman, T. and Justeson, J. S. (2003). A preliminary Mayan etymological dictionary. FAMSI.
Kaufman, T. and Norman, W. M. (1984). An outline of proto-Cholan phonology, morphology, and
vocabulary. In Justeson, J. S. and Campbell, L., editors, Phoneticism in Mayan Hieroglyphic Writ-
ing, pages 77–166. Institute for Mesoamerican Studies, State University of New York, Albany,
NY.
Kayne, R. S. (1994). The Antisymmetry of Syntax. MIT Press, Cambridge, MA.
Keenan, E. L. (1976). Towards a universal definite of ”subject”. In Li, C. N., editor, Subject and
Topic, pages 303–334. Academic Press, New York.
Kelepir, M. (2001). Topics in Turkish syntax: Clausal structure and scope. Doctoral dissertation, Mas-
sachusetts Institute of Technology, Cambridge, MA.
Konietzko, A., Winkler, S., and Culicover, P. W. (2018). Heavy NP shift does not cause freezing.
Canadian Journal of Linguistics/Revue canadienne de linguistique, 63(3):454–464.
Kratzer, A. (1996). Severing the external argument from its verb. In Johan Rooryck, L. A. Z., editor,
Phrase Structure and the Lexicon, pages 109–137. Kluwer, Dordrecht.
Kratzer, A. (2004). Telicity and the meaning of objective case. In Gu´eron, J. and Lecarme, J.,
editors, The Syntax of Time, pages 389–423, Cambridge, MA. MIT Press.
210
Kratzer, A. (2009). Making a pronoun: Fake indexicals as windows into the properties of pronouns.
Linguistic Inquiry, 40:187–237.
Landman, F. (2004). Indefinites and the type of sets. Blackwell Publishing, Malden, MA.
Lasnik, H. (2001). Subjects, objects, and the EPP. In Davies, W. D. and Dubinsky, S., editors,Objects
and other subjects, pages 103–121. Kluwer, Dordrecht.
Law, D. (2009). Pronominal borrowing among the Maya. Diachronica, 26(2):214–252.
Legate, J. A. (2002). Walpiri: Theoretical implications. Doctoral dissertation, Massachusetts Institute
of Technology, Cambridge, MA.
Legate, J. A. (2008). Morphological and abstract case. Linguistic Inquiry, 39(1):55–101.
Legate, J. A. (2014). Voice and v: Lessons from Acehnese, volume 69. MIT Press, Cambridge, MA.
Li, C. N. and Thompson, S. A. (1977). Amechanism for the development of copula morphemes. In
Li, C. N., editor,Mechanisms of syntactic change, pages 419–444. University of Texas Press, Austin,
TX.
Little, C.-R. (2018a). A feature-based analysis of the Ch’ol (Mayan) person paradigm. In Garvin,
K., Hermalin, N., Lapierre, M., Melguy, Y., Scott, T., and Wilbank, E., editors, Proceedings of the
44th Berkeley Linguistics Society, pages 147–161, Berkeley, CA. Berkeley Linguistics Society.
Little, C.-R. (2018b). Possessed numerals in Ch’ol. In Johnson, K. and Goebel, A., editors, Pro-
ceedings of the Tenth Conference on the Semantics of Under-Represented Languages of the Americas,
Amherst, MA. GLSA.
Little, C.-R. (2020). Left branch extraction, object shift, and freezing effects in Tumbalá Ch’ol.
Glossa: a journal of general linguistics, 5(1):26.
Little, C.-R., Moroney, M., and Royer, J. (2020). Classifying classifiers: Two kinds of numeral
classifiers across languages. Presented at the 94th Annual Meeting of the Linguistic Society of
America. New Orleans, LA.
211
Little, C.-R. and Vázquez Martı́nez, M. (2018). La distribución e interpretación de sustantivos en
el ch’ol: Un estudio práctico de corpus. Presented at Form and Analysis in Mayan Linguistics
(FAMLi) 5.
Little, C.-R. and Wiegand, M. (2018). A compositional morphosemantic analysis of exclusivity in
Ch’ol. In Farrell, P., editor, 2018 Annual Meeting of the Linguistic Society of America, volume 3.
Marantz, A. (1984). On the Nature of Grammatical Relations. MIT Press, Cambridge, MA.
Marantz, A. (1988). Clitics, morphological merger, and the mapping to phonological structure. In
Hammond, M. and Noonan, M., editors, Theoretical Morphology: Approaches in Modern Linguis-
tics, pages 253–270. Academic Press, San Diego, California.
Martı́nez Cruz, V. (2007). Los adjetivos y conceptos de propiedad en chol. Master’s thesis, CIESAS,
México.
Massam, D. (2000). VSO and VOS: Aspects of Niuean word order. In Carnie, A. and Guilfoyle, E.,
editors, The syntax of verb initial languages, pages 97–116. Oxford University Press, Oxford.
Massam, D. (2001). Pseudo noun incorporation in Niuean. Natural Language and Linguistic Theory,
19(1):153–197.
Mateo Toledo, B. E. (2008). The family of complex predicates in Q’anjob’al (Maya): Their syntax and
meaning. Doctoral dissertation, University of Texas Austin, Austin, TX.
Mateo Toledo, E. (2004). Directional Markers in Q’anjob’al (Maya); their Syntax and Interaction with
Aspectual Information. Master’s thesis, University of Texas at Austin, Austin, TX.
Matthewson, L. (2004). On the methodology of semantic fieldwork. International journal of Ameri-
can linguistics, 70(4):369–415.
McCarthy, J. and Prince, A. (1995). Faithfulness and reduplicative identity. In Beckman, J., Ur-
banczyk, S., and Dickey, L. W., editors, University of Massachusetts Occasional Papers vol. 18: Pa-
pers in Optimality Theory, pages 249–384. GLSA, Amherst, MA.
212
McCloskey, J. (1991). Clause structure, ellipsis and proper government in Irish. Lingua, 85(2-
3):259–302.
Milsark, G. (1974). Existential sentences in English. Doctoral dissertation, Massachusetts Institute of
Technology, Cambridge, MA.
Mithun, M. (1984). The evolution of noun incorporation. Language, 60:847–894.
Mithun, M. (1986). On the nature of noun incorporation. Language, 62:32–37.
Mora-Marı́n, D. (2009a). A test and falsification of the “Classic Ch’olti’an” hypothesis: A study of
three Proto-Ch’olan markers. International Journal of American Linguistics, 75(2):115–157.
Mora-Marı́n, D. F. (2009b). Reconstruction of the Proto-Ch’olan demonstrative pronouns, deictic
enclitics, and definite articles. Transactions of the Philological Society, 107(1):98–129.
Moroney, M. (2020). Updating the typology of definiteness: Evidence from bare nouns in Shan.
manuscript submitted for publication. Cornell University.
Murray, S. E. (2015). Reciprocity in fieldwork and theory. In Bochnak, R. and Matthewson, L.,
editors,Methodologies in Semantic Fieldwork, pages 287–305. Oxford University Press.
Mykhaylyk, R. (2010). Optional object scrambling in child and adult Ukrainian. Doctoral dissertation,
Stony Brook University, Stony Brook, NY.
Nevins, A. (2011). Multiple agree with clitics: Person complementarity vs. omnivorous number.
Natural Language & Linguistic Theory, 29(4):939–971.
Nicolae, A. (2015). On the syntactic specialization of Romanian demonstratives and the grammat-
icalization of the article cel. Revue roumaine de linguistique, 60:47–70.
Norman, W. M. and Campbell, L. (1978). Toward a Proto-Mayan syntax: a comparative perspec-
tive on grammar. In England, N. C., editor, Papers in Mayan Linguistics, volume Miscellaneous
Publications in Anthropology, no. 6; Studies in Mayan Linguistics, no. 2, pages 136–156. Uni-
versity of Missouri, Columbia, MO.
213
Otaki, K., Sugisaki, K., Yusa, N., and Koizumi, M. (2019). Two routes to the Mayan VOS: From the
view of Kaqchikel. Gengo Kenkyu: Journal of the Linguistic Society of Japan, 159:25–45.
Ouhalla, J. (1994). Verb movement and word order in Arabic. In Lightfoot, D. and Hornstein, N.,
editors, Verb movement, pages 41–72. Cambridge University Press, Cambridge.
Partee, B. H. (1986). Noun phrase interpretation and type-shifting principles. In Groenendijk, J.,
de Jongh, D., and Stokhof, M., editors, Studies in discourse representation theory and the theory of
generalized quantifiers, pages 115–143. Foris, Dordrecht.
Payne, J. (1993). The headedness of noun phrases: Slaying the nominal hydra. In Corbett, G. G.,
Fraser, N.M., andMcGlashan, S., editors,Heads in grammatical theory, pages 114–139. Cambridge
University Press, Cambridge.
Perlmutter, D.M. (1978). Impersonal passives and the unaccusative hypothesis. In Jaeger, J., editor,
Proceedings of the 4th Annual Meeting of the Berkeley Linguistics Society, volume 4, pages 157–189.
Polian, G. (2004). El predicado existential y locativo en tseltal. clasificación de las construcciones
con AY. In Barreras Aguilar, I. and Castro Llamas, M., editors, Memorias del VII Encuentro
Internacional de Lingüı́stica en el Noroeste, pages 256–274. Editorial UniSon, Hermosillo, Sonora.
Polinsky, M., Gallo, C. G., Graff, P., Kravtchenko, E., Morgan, A. M., and Sturgeon, A. (2013).
Subject islands are different. In Sprouse, J., editor, Experimental syntax and Island effects, pages
286–3–09. Cambridge University Press.
Preminger, O. (2014). Agreement and its failures. MIT Press, Cambridge, MA.
Prince, A. and Smolensky, P. (1993). Optimality Theory: Constraint interaction in generative gram-
mar. In RUCCS Technical Report No. 2. Rutgers University Center for Cognitive Science, Piscat-
away, NJ.
Pylkkänen, L. (2008). Introducing arguments. MIT press, Cambridge, MA.
Rackowski, A. and Richards, N. (2005). Phase edge and extraction: A Tagalog case study. Linguistic
Inquiry, 36(4):565–599.
214
Richards, M. D. (2007). On feature inheritance: An argument from the phase impenetrability
condition. Linguistic Inquiry, 38(3):563–572.
Richards, N. (2013). Lardil “case stacking” and the timing of case assignment. Syntax, 16(1):42–76.
Rodrı́guez Ceja, G. (2012). Enfermar y sanar. Persona, cuerpo social y cosmos en la vida cotidiana chol en
Calakmul. Doctoral dissertation, Universidad Nacional Autónoma de México, Mexico City.
Ross, J. (1974). Three batons for cognitive psychology. In Weimer, W. and Palermo, D., editors,
Cognition and the symbolic processes, pages 63–124. Lawrence Erlbaum Associates, Hillsdale, NJ.
Royer, J. (2019). Domain restriction and noun classifiers in Chuj (Mayan). In Baird, M. and Peset-
sky, J., editors, Proceedings of the Forty-Ninth Annual Meeting of the North East Linguistic Society,
pages 87–96, Amherst, MA. GLSA.
Royer, J. (To Appear). An apparent syntax-prosody mismatch as right extraposition: Evidence
from Mayan. In Ayana Whitmal, Yixiao Song, M. A., editor, Proceedings of 50th Annual Meeting
of the North East Linguistic Society NELS. GLSA, Amherst, MA.
Rullmann, H. and You, A. (2006). General number and the semantics and pragmatics of indefinite
bare nouns in mandarin chinese. In von Heusinger, K. and Turner, K., editors, Where semantics
meets pragmatics, volume 16 of Current Research in the Semantics / Pragmatics Interface, pages 175–
196. BRILL.
Rumsey, A. (1982). An Intra-Sentence Grammar of Ungarinjin, North-Western Australia, volume 86 of
Pacific Linguistics, Series B. Australian National University, Canberra.
Sadler, L. and Nordlinger, R. (2006). Case stacking in realizational morphology. Linguistics,
44(3):459–487.
Schele, L. (1982). Maya glyphs: the verbs. University of Texas Press, Austin, TX.
Schwarz, F. (2009). Two types of definites in natural language. Doctoral dissertation, University of
Massachusetts Amherst, Amherst, MA.
215
Schwarz, F. (2013). Two kinds of definites cross-linguistically. Language and Linguistics Compass,
7(10):534–559.
Selkirk, E. (2011). The syntax-phonology interface. In Goldsmith, J., Riggle, J., and Yu, A. C. L.,
editors, The handbook of phonological theory, pages 435–484. Wiley-Blackwell, Hoboken, NJ.
Sobin, N. (2002). Determiners in genitive constructions. Lingua, 112:597–618.
Solt, S. (2011). Attributive quantity words as nonrestrictive modifiers. In Suzi Lima, Kevin Mullin,
B. S., editor, Proceedings of the 39th North East Linguistic Society (NELS), pages 731–744, Amherst,
MA. GLSA.
Sproat, R. (1985). Welsh syntax and VSO structure. Natural Language & Linguistic Theory, pages
173–216.
Svenonius, P. (2004). On the edge. In Adger, D., de Cat, C., and Tsoulas, G., editors, Peripheries:
Syntactic edges and their effects, pages 259–287. Kluwer, Dordrecht.
Szabolcsi, A. (1987). Functional categories in the noun phrase. In István, K., editor, Approaches to
Hungarian 2: Theories and analyses, pages 167–191. Jate Szeged, Szeged.
Takita, K. (2020). Labeling for linearization. The Linguistic Review, 37(1):75–116.
Tomlin, R. (1986). Basic word order: Functional principles. Croom Helm, London.
Travis, L. (1984). Parameters and Effects of Word Order Variation. Doctoral dissertation, Mas-
sachusetts Institute of Technology, Cambridge, MA.
Uriagereka, J. (1988). On Government. Doctoral dissertation, University of Connecticut.
Van Everbroeck, E. (2003). Language type frequency and learnability from a connectionist per-
spective. Linguistic Typology, 7(1):1–50.
Van Geenhoven, V. (1998). Semantic incorporation and indefinite descriptions: Semantic and syntactic
aspects of noun incorporation in West Greenlandic. CSLI publications, Stanford, CA.
Van Urk, C. and Richards, N. (2015). Two components of long-distance extraction: Successive
cyclicity in Dinka. Linguistic Inquiry, 46(1):113–155.
216
Vázquez Álvarez, J. J. (2002). Morfologı́a del verbo de la lengua chol de Tila, Chiapas. Master’s thesis,
Centro de Investigaciones y Estudios Superiores en Antropologı́a Social, San Cristóbal de las
Casas.
Vázquez Álvarez, J. J. (2010). Los depictivos analı́ticos y sintéticos en la lengua chol de Tila,
Chiapas. In y Roberto Zavala, A. J., editor, La predicación secundaria en lenguas de Mesoamérica,
pages 61–85, Mexico. CIESAS.
Vázquez Álvarez, J. J. (2011). A grammar of Chol, a Mayan language. Doctoral dissertation, Univer-
sity of Texas Austin, Austin, TX.
Vázquez Álvarez, J. J. (2019). Perfil sociolingüı́stico de Frontera Corozal, Ocosingo, Chiapas. Re-
vista Pueblos y fronteras digital, 14:1–32.
Vázquez Álvarez, J. J. and Coon, J. (To Appear). Headless relative clauses in Ch’ol. In Caponigro,
I., Torrence, H., and Zavala, R., editors, Headless Relative Clauses in Languages of Mesoamerica.
Oxford University Press, New York.
Vázquez Martı́nez, M. and Little, C.-R. (2020). Dimensions of definiteness in Ch’ol: A dialectal
comparison. Presented at The Society for the Study of the Indigenous Languages of the Ameri-
cas (SSILA) 2020.
Vikner, S. (1994). Scandinavian object shift and West Germanic scrambling. In Corver, N. . and
van Riemsdijk, H., editors, Studies on scrambling, volume 487–517. Mouton de Gruyter, Berlin.
Warkentin, V. and Brend, R. (1974). Chol phonology. Linguistics, 132:87–101.
Warkentin, V. M. and Scott, R. (1980). Gramática ch’ol. Instituto Lingüı́stico de Verano, Mexico.
Wexler, K. and Culicover, P. W. (1977). Some syntactic implications of a theory of language learn-
ability. In Culicover, P., Wasow, T., and Akmajian, A., editors, Formal syntax, pages 7–60. Aca-
demic Press, New York.
Whitman, J. (2008). The classification of constituent order generalizations and diachronic expla-
nation. In Good, J., editor, Linguistic universals and language change, pages 233–252. Oxford Uni-
versity Press, Oxford.
217
Whittaker, A. and Warkentin, V. (1965). Chol Texts on the Supernatural. Instituto Lingüı́stico de
Verano, México.
Wiltschko, M. (2008). The syntax of non-inflectional plural marking. Natural language & linguistic
theory, 26(3):639–694.
Wojdak, R. (2007). The linearization of affixes: Evidence from Nuu-chah-nulth. Springer Science &
Business Media.
Woolford, E. (2000). Ergative agreement systems. University of Maryland Working Papers in Linguis-
tics, 10:157–191.
Woolford, E. (2017). Mainland Scandinavian object shift and the puzzling ergative pattern in
Aleut. In Bailey, L. and Sheehan, M., editors, Order and structure in syntax, pages 117–134.
Language Science Press, Berlin.
Yang, C. D. (1999). Unordered merge and its linearization. Syntax, 2(1):38–64.
Zavala, R. (2017). Dos cadenas de gramaticalización de los demostrativos en oluteco. Presented at
CILLA VII, Austin, TX.
Zeijlstra, H. (2004). Sentential negation and negative concord. Doctoral dissertation, University of
Amsterdam, Amsterdam.
Zimmermann, M. (2007). Overt existential closure in Bura (Central Chadic). In Friedman, T. and
Gibson, M., editors, Semantics and Linguistic Theory, volume 17, pages 333–350, Ithaca, NY. CLC
Publications.
218