POSITIONAL ROOTS IN KANJOBAL (MAYAN)
LAURA ELLEN MARTIN
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA IN PARTIAL
FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIOA
_ _ ~ __
O Copyright by
Laura Ellen Martin
Many people have been of great help during the
several years of work which led eventually to the completion
of this dissertation. My deepest debt is to Diego de Diego
Antonio and Francisco Pascual who worked with me on Kanjobal
during my field stays in Santa Eulalia. They were always
careful and willing assistants even for the most tiresome
tasks and their interest in understanding the beauties and
complexities of Kanjobal were a constant inspiration to me-
Many other people in Santa Eulalia were also very helpful
and are gratefully acknowledged here: Maria Pascual, Juan
Lorenzo, and Juana Francisco who helped collect and check
data on positional roots; doia Adelivia Soto and her family
who made me welcome in their home and cared for me like a
sister; Jose Pablo Luis of San Rafael who first introduced
me to Santa Eulalla; the municipal administration of Santa
Eulalia who encouraged my work there; and all the Kanjobal
people, many of whom I never met, who recorded texts and
were so willing to share their language with me.
During the time spent working on this dissertation,
many people at the University of Florida were especially
i I II
supportive. I want to thank particularly, the chairman of
my committee, Dr. M. J. Hardman-de-Bautista, both for her
support during my last quarter and for her early example,
encouragement, and training which first led to my commitment
to field-based linguistics. Dr. D. Gary Miller is also
gratefully acknowledged here for the great time and care he
took to read and comment on my dissertation. Both his
criticism and his praise have been especially welcomed. Dr.
Murray Lasley, Dr. Alan Burns, and Dr. Norman Markel also
deserve considerable recognition for the criticisms they
offered on the manuscript and for their willingness to take
on the reading of it under difficult circumstances and with
limited time. Bohdan Saciuk also read much of the disserta-
tion and made many useful suggestions. I want to thank him
for that as well as for his constant, friendly personal sup-
port. Nora C. England also contributed a great deal to my
work first by getting me started in Mayan linguistics, then
by sharing her insights during the time I worked on Kanjobal,
and finally by adding many helpful comments on the draft of
this dissertation. Thanks also to Nancy McDavid, Randall
Martin, and Howard Tupper who made invaluable technical con-
tributions to the preparation of the manuscript.
While on leave to do field work and attend graduate
school as well as on the job, I have been fortunate to have
both the material and moral support of Cleveland State Uni-
versity In Cleveland, Ohio. I thank especially Dr. Bruce
Beatie, Chairman of the Department of Modern Languages; Dr.
Willis Sibley, Chairman, Department of Anthropology; and
Deans Jack Soules and Leslie Armour, College of Arts and
Sciences. Acknowledgment is also made here of Jean Imm,
Eileen Cornez, and Linda Zimmer who handled a lot of admin-
istrative problems for me and of Anita Stoll and Phillipa
Yin who took up slack for me in my absence and wished me well.
There are also my students, too numerous to mention, who
challenged and inspired me and will no doubt make it all
worthwhile. Special thanks to Clive Fetzer who worked with
me on Kanjobal and who had many insights into the language
and the people which are reflected in this work.
During 1972-73, the Proyecto Linguistico Francisco
Marroquin provided me with material support while I worked
on Kanjobal. I want to acknowledge them and also thank the
Indian students at the PLFM who were always so delightful to
be with. I am also grateful to Karen Dakin for sharing some
of her private Kanjobal materials, Chris Day for his quick
response to my requests for Jacaltec data, and to Terrence
Kaufman for his initial instruction about Mayan.
And finally, there is no way to properly thank the
many friends who saw me through the final stages and actual
completion of this dissertation, especially my parents,
Audrey and Sanford Martin; Kathrine Sturgen, Pat Kwachka,
Bill England, Nora England, and Lawrence Carpenter. I could
never have done it without you.
1 __ ~IPI
TABLE OF CONTENTS
ACKNOWLEDGMENTS ............... .. .. iii
CONVENTIONS AND ABBREVIATIONS . . . . . . x
ABSTRACT. ...... . . . . . . . .xii
0 INTRODUCTION . . . . . . . .. .. . 1
0.1 Purpose . . . . . . . . . . 1
0.2 Cultural and Linguistic Background. . . . 6
0.3 Data Base .. .. . . . . . . 24
NOTES. . . . . . . . . ... . .29
1 GRAMMATICAL SKETCH . . . . . . . ... 30
1.1 Phonology . . . . . . . . . 31
1.1.1 Phonemic Inventory . . . . . 32
126.96.36.199 Obstruent consonants. . . 35
188.8.131.52 Sonorant consonants . . .. 47
184.108.40.206 Vowels . . . . . .. . 50
1.1.2 Phonological Processes . . . ... 51
220.127.116.11 Stress assignment . . .. 52
18.104.22.168 Vowel harmony . . . . 57
22.214.171.124 h-insertion . . . ... 59
126.96.36.199 Reduplication . . . .. 60
188.8.131.52 Glottalization phenomena . 62
184.108.40.206.1 Vowel glottalization 62
220.127.116.11.2 Glottal stop dele-
tion . . . . 65
18.104.22.168 Nasal assimilation. . . . 67
22.214.171.124 Palatalization . . . . 68
126.96.36.199 Consonant reduction phenomena 69
188.8.131.52.1 Cluster reduction.. 69
184.108.40.206.2 Sibilant reduction 71
220.127.116.11 Vowel reduction . . . . 73
18.104.22.168.1 Vowel lowering . 73
22.214.171.124.2 Vowel devoicing. . 74
126.96.36.199.3 Vowel deletion . 76
1.1.3 Distributional Constraints . . ... 77
188.8.131.52 Morpheme structure. . . . 78
184.108.40.206 Cooccurrence restrictions . 80
TABLE OF CONTENTS (continued)
220.127.116.11 Clusters . . . . .. 81
18.104.22.168.1 Vowel clusters . 81
22.214.171.124.2 Consonant clusters 82
126.96.36.199 Frequencies . . . . .. 84
1.1.4 Nativization and Loan Phonology. ... 85
1.2 Grammar . . . . . . . . . . 90
1.2.1 Grammatical Categories . . . . . 90
188.8.131.52 Person and possession ... 91
184.108.40.206.1 Person. . . .. 91
220.127.116.11.2 Possession. .. . 97
18.104.22.168 Number . . . . . .. . 107
22.214.171.124 Categorization. . . . .. 111
126.96.36.199 Transitivity and agency . . 120
188.8.131.52 Aspect and tense . . ... 126
184.108.40.206 Direction and location. . .. 138
1.2.2 Grammatical Processes . . . .. . 148
220.127.116.11 Morphological processes . 149
18.104.22.168.1 Morphological word
classes. . . 149
22.214.171.124.2 Derivational affixes 155
126.96.36.199 Compounding ...... .. 166
NOTES. . . . . . . . ... ..... 169
2 POSITIONAL ROOTS: PHONOLOGICAL PATTERNS . . .. .174
2.1 Segment Distribution. . . . . . ... 174
2.1.1 Cooccurrence Patterns. . . . . 176
2.1.2 Frequencies. . . . . . . ... 176
2.2 Phonological Processes. . . . . . .. .179
2.2.1 Reduplication. . . . . . ... 179
2.2.2 Metathesis . . . . . . ... 185
2.3 Sound Symbolism . . . . . . . . 186
2.3.1 Vowel Alternations . . . . . 190
2.3.2 Consonant Patterns . . . . ... 199
NOTES. . . . . . . . . . . . 207
3 POSITIONAL ROOTS: SYNTACTIC PATTERNS. . . .208
3.1 Positionals Within the Verbal System. . ... 209
3.1.1 Positionals and Oirectional/Locational
Verbs . . . . . . . . 210
3.1.2 Transitive Derivation. . . . .. 219
188.8.131.52 Direction transitivization. . 219
184.108.40.206 Causative . . . . .. 223
220.127.116.11 Self-causative. . . . .. 229
18.104.22.168 Indirect transitivization 235
3.1.3 Intransitive Derivation . . . . 238
22.214.171.124 Inchoative. . . . . .. .239
126.96.36.199 Iterative . . . . .. 242
188.8.131.52 Progressive . . . . . 244
TABLE OF CONTENTS (continued)
184.108.40.206 Completive. . . . . ... 248
220.127.116.11.1 Agentless completion
with -ay . . 248
18.104.22.168.2 Implied causation
-Coj . . . . 253
3.2 Positional Roots Within the Nominal System. . 264
3.2.1 Nominal Derivation . . . . ... .264
3.2.2 Numeral Classifiers. . . . . ... 269
3.3 Positionals Within the Attributive System . 278
3.3.1 Adjectival Derivation. . . . ... 278
22.214.171.124 Redup -VC2i intensitives. . 279
126.96.36.199 Redup -Cl statives. . . .. 291
188.8.131.52 Stative atributives in -naj 294
184.108.40.206 Miscellaneous adjectival
derivation. . . . . ... 296
220.127.116.11.1 Plural derivation
with -kixhtaq. . 297
18.104.22.168.2 Emphatic derivation
with -taq. ... .298
22.214.171.124.3 Diminutive with -ich
or -ix . . .. . 299
126.96.36.199.4 ResuT-ant derivation
with -inaq ... .300
3.3.2 Attributive Phrases . . . . ... .302
NOTES. . . . . . . . .... .. . .307
4 POSITIONAL ROOTS: SEMANTIC PATTERNS . . . .309
4.1 Class-specific Semantic Features . . ... .311
4.1.1 Conflation . . . . . . ... 311
4.1.2 Direction/Location . . . . .. 320
4.2 Root-specific Semantic Features . . ... .332
4.2.1 Humanness and Animateness . . ... .333
4.2.2 Size and Shape . . . . . ... .337
4.2.3 Position . . . . . . .. 339
4.2.4 Quantity and Distribution. . . .. .341
4.2.5 Texture. . . . . . . . . 343
4.3 Context-specific Semantic Features. . . ... 344
4.3.1 Metaphor and Semantic Extension. ... .346
4.3.2 Joking, Insults, and Language Play . 355
NOTES. . . . . . . . ... ..... 363
5 KANJOBAL LANGUAGE AND CULTURE . . . . .. .364
5.1 Direction and Location. . . . . . ... 377
5.2 Time and Space. . . . . . . . .380
5.3 Perceptual Categorization . . . . ... .383
5.4 Ouality and Illusion .. . . . . .. 386
NOTES . . . . . . . . . . . 389
TABLE OF CONTENTS (continued)
A FREQUENCY DATA. ........ . . 392
B LOANWORDS . . ..... . . 395
C INVENTORY OF POSITIONAL ROOTS . . ... 399
D TEXT ........... . . . 412
REFERENCES .............. . . 441
BIOGRAPHICAL SKETCH ........... . . 449
CONVENTIONS AND ABBREVIATIONS
KSE Kanjobal of Santa Eulalia
[ ] phonetic transcription
/ / phonemic transcription (practical orthography)
divides morphemes in phonemic transcriptions
GNumC general numeral classifier
NCI noun classifier
P positional root
comp completive aspect
form stem formative
income Incompletive aspect
iv intransitive verb stem
CONVENTIONS AND ABBREVIATIONS (CONTINUED)
PIV pseudointransitive suffix
pot potential aspect
prog progressive aspect
Hyphens before or after a morpheme indicate that it is bound.
Underlined morphemes In discussive passages are always in
Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degrae of Doctor of Philosophy
POSITIONAL ROOTS IN KANJOBAL (MAYAN)
Laura Ellen Martin
Chairman: Dr. M. J. Hardman-de-Bautista
Major Department: Linguistics
Kanjobal is a Mayan language spoken by several
thousand people in northwestern Guatemala; it has been
relatively little described. Like most Mayan languages,
Kanjobal possesses a distinct root class, known as position-
als. These roots have special phonological, syntactic, and
semantic characteristics. This study presents a description
of the properties of these roots within the context of the
overall grammatical and semantic organization of the Santa
Eulalla variety of Kanjobal (KSE). The description is the
result of fifteen months of field work in Guatemala.
An introductory chapter provides linguistic and ethno-
graphic background on the KSE language and its speakers, and
a description of the nature of the data base. This is fol-
lowed by a sketch of KSE phonology and grammar. The phono-
logical description includes a phonemic analysis, a discus-
sion of major phonological processes, and information on dis-
tributional patterns and nativization. The grammatical
description is organized into two parts: grammatical cate-
gories, in which the important underlying categories of
person, number, categorization, transitivity, tense/aspect,
and direction/location are defined and illustrated, and
grammatical processes, in which derivational morphology,
compounding, and sentence formation are treated.
The next three chapters are devoted to the detailed
description of the nature of positional roots. First, they
are considered phonologically in Chapter 2 and examined for
special distribution patterns and phonological processes.
The possibility of sound symbolism operating in the position-
al root class is also investigated.
The next chapter describes the morphological and
syntactic processes affecting KSE positional roots. The
roots show special relationships with directional/locational
verbs and clitics; furthermore, they are easily derived as
transitive or intransitive verb stems. They occur as syn-
tactically defined nouns in certain types of numeral phrases
and have semantic ties with attributives. All these deriva-
tional and phrasal functions are described and illustrated
in Chapter 3.
In Chapter 4, positional roots are examined for their
semantic patterns. The class is defined in terms of features
such as direction/location and conflation properties. Indi-
vidual roots are specified for features such as humanness
size, distribution, shape, and position. Positional roots
also have extremely productive semantic roles in metaphori-
cal expressions in the language. Several types of semantic
play are therefore described.
The final chapter returns to a larger view of the
language and to those patterns found within the broad gram-
matical system and reinforced within a single area of the
grammar, the positional class. These concluding remarks
deal with the questions of linguistic and cultural overlap.
Several general categories such as direction, time and space,
humanity and duality which are important to KSE grammatical
organization may also be reflected in non-grammatical
cultural behavior. Ethnographic data are considered in
order to determine to what degree such overlapping reflec-
tions may be present in KSE language and culture.
Appendices include an inventory of positional roots
used in the study, data on phoneme frequencies, a list of
common loanwords and a text taken from a monolingual KSE
speaker. This text gives a morphemic analysis with a
literal interlinear and a free translation.
The purpose of this study is to examine in detail
the linguistic characteristics of a single root class in
the Kanjobal language. The members of this class have
traditionally been called positionals, and, in one form
or another, are a common feature of most languages in the
Mayan language family to which Kanjobal belongs. The term
positional is somewhat misleading since only some members
of the class refer to the position of bodies in space.
The majority refer to the distribution, shape, size, and
aspect of objects. The term is in very wide use, however
inappropriately, and that designation is used here. The
positionals are phonologically, syntactically, and seman-
tically distinct within the Kanjobal language system and
have an important role in metaphorical usage.
Positional roots occur In most, if not all, of the
Mayan languages, sometimes as a subclass within another
grammatical class and sometimes as a separate morphological
class. In all cases they have distinct derivational
properties and syntactic patterns. References to them in
larger grammatical studies are common, but as a class they
have received relatively little linguistic attention. The
following descriptive examples have been collected from
some readily available sources.
All positional roots are CVC and are dis-
tinguished by patterns of reduplication which occur
with no other class. (Hopkins 1967:76) Positional
roots in Chuj are not a class of verb roots, but
a set of roots which are derivationally and in-
flectionally distinct from verb roots. Chuj
positional roots may be derived to form verb
stems (or derived to form adverbial or noun stems)
S. ut positionals do not occur as verb stems
without derivation. (Hopkins 1970:29)
Positional verb roots are recognized in two
1) they do not occur as simple CVC unin-
2) upon derivation with -an 'transitive stem
formative' or -ah 'intransitive stem
formative', they take an infixed -h-.
All positional roots are CV or CVC. (Day
1973a:15) Many positional roots are also transi-
tive verb roots. (25) Adjectives and positionals
occur only as stative verb complements and in com-
pound nouns (27) '-an 'positional stem formative'
occurs on positional teams (which are always
roots) in all environments' (except when redupli-
Positional roots are CV or CVC (Kaufman 1971:
35), and are derived in various ways as verbs, both
transitive and intransitive (46-49, 51-53), as
adverbs (60-62), nouns (73-76), adjectives
(84), and as specific numeral classifiers (88).
Positional roots are bound forms which must
be derived to form words, always with a change
in class. Some of the particular derivational
affixes which form words from positional roots
are specific to this root class, and most commonly
form verbs or adjectives. The adjective thus
formed indicates that something has the position,
form, or state described by the root, while the
verb indicates that something is becoming like
or is placed like that described by the root.
Positional roots have a semantic element in common;
they generally describe position, form or state
of an object, and imply absence of movement.
Positionals can be distinguished from other
root classes by three characteristics: 1) All
positlonals have the shape CVC. 2) Positionals
do not occur without derivational affixes as simple
uninflected stems. 3) Positionals can be derived
to form adjectives, intransitive verbs, or causa-
tives by means of special suffixes occurring only
with roots of this class. Positional roots also
possess the rather peculiar semantic property of
categorizing objects in highly specific ways.
As is obvious from these citations, positionals are defined
phonologically (both by their shape and by the phonological
processes which apply to them), syntactically (by morphology
and by syntactic role), and semantically (both as a class
and as individual items).
In this study, a substantial collection of Kanjobal
positional roots are considered in precisely these three
linguistic contexts. Their phonological characteristics
I ----~---~LI-- - -r
are surveyed for distribution patterns of various kinds.
The morphological and syntactic characteristics of posi-
tional roots are analyzed in considerable detail, both as
a class and as a set of elements within the larger grammati-
cal system. Further, positional roots are examined seman-
tically, and the semantic features which organize the class
are given preliminary definition. It is believed that clues
may be found in the internal semantics of the positional
class which are relevant to the broader semantic categories
of the larger language system. These language categories
may in turn find parallel expression in cultural categories
which are manifest in Kanjobal behavior which is other than
linguistic. Some attention is, therefore, given to the
questions of semantic parallels in language and culture.
The organization of the study is straightforward.
First, a brief background of the Kanjobal language and
people is given as Chapter 0 in order to place them in their
proper space and time context and to provide enough ethno-
graphic data to facilitate the understanding of certain
semantic references made in later chapters. Also contained
in the introductory chapter is a description of the data
base used for this study. This is followed in Chapter 1
by a brief overall grammatical sketch of the Kanjobal
language and a review of the relevant linguistic literature.
The main body of this study is contained in Chapters 2,
I --- -- I F- ii
3,and 4 which describe the phonological, syntactic, and
semantic nature of positional roots. The last chapter
considers the possibility of finding language and culture
parallels in Kanjobal based on the semantic data presented
in earlier chapters.
This work is unique in that it is the first full-
length investigation of positional roots, a class which has
an important place in the grammatical systems of Mayan
languages. In fact, this investigation is one of the few
attempts made in Mayan linguistics which goes beyond surface
structural description. Hopkins (1970:19) remarked on the
lack of such studies on Mayan languages and called for
'syntactic studies which illuminate the relationships be-
tween syntactic classes'. The present study attempts to
begin to fill this gap, at the same time going even further
by investigating the semantic relationships as well. While
Hopkins notes the importance of such studies within Mayan
linguistics for both historical and synchronic purposes,
such a detailed analysis is important for general linguis-
tics as well. Linguistic science is always in need of
analyzed bodies of data against which current theory can
be tested since the best test of a theory is in its account
of new data not used in its original formulation. Further,
this analysis of positional roots may be expected to
contribute a great deal to ongoing cultural studies in
the Mayan area since it will provide at least a partial
assessment of semantic categories which underlie the
grammar. Used in conjunction with a cultural study that
defines the cultural semantic categories, these data could
provide an important testing ground for theories of language
and culture correlations as well as for theories which are
more purely linguistic in nature.
0.2 Cultural and Linguistic Background
Kanjoball is one of the more than twenty-five extant
Mayan languages spoken in southern Mexico and Guatemala.
A number of classification schemes have been proposed for
Mayan languages. The one adopted here is that of Kaufman
as explicated in Kaufman 1974; it is the most recent and
is based on far more data than that available to previous
According to this analysis (Kaufman 1974:85), the
Mayan family has two great divisions, Eastern and Western.
The Western division, to which Kanjobal belongs, has two
branches, the Cholan and the Kanjobal, which have a time
depth of some twenty centuries. The Cholan languages In-
clude Chol, Chontal, and Chorti in one group and Tzeltal-
Tzotzil In another. The Kanjobal branch Is divided Into
the Chuj-Tojolobal group; Motozintlec, about which little
is known; and the Kanjobalan complex. There is still some
- -- II
controversy over the exact genetic relationships among the
three varieties in the Kanjobalan complex: Jacaltec, Acatec,
and Kanjobal. The details of the controversy are not rele-
vant here but the positional data can help resolve the ques-
tion. The arguments for the separation as adopted here are
found in Kaufman 1976. Information on the phonemic inven-
tories of the three languages is giver in Kaufman (1975:
76-84). These languages are mutually intelligible to some
degree and probably have been separated no more than one
thousand years (Kaufman, personal com-unication).
This study is based on the Kanjobal variety spoken
in Santa Eulalia (henceforth abbreviated KSE). The terns
'Kanjobal' or 'Kanjobal proper' are used here to refer to
the language of all four Kanjobal-speaking towns and the
designation 'Kanjobalan' refers to the three-language com-
plex which includes Jacaltec and Acatec as well. 'Greater
Kanjobalan' refers to the languages of the Kanjobal group
of western Mayan: Kanjobal, Acatec, Jacaltec, Chuj, and
The speakers of Kanjobal number around forty
thousand and reside in four towns high in the Chuchumatan
mountains in the northwest Guatemalan department of
Huehuetenango: San Juan Ixcoy, San Pedro Soloma, Santa
Eulalia, and Santa Cruz Barillas. There are some minor
dialect variations among these towns which are outside the
scope of this investigation. These variations are primarily
phonological and some of them are summarized in Kaufman
Kanjobal speakers are primarily in contact with the
Chuj speakers of San Mateo IxtatAn and San Sebastian CoatAn
to the northwest, with Acatec speakers in San Miguel AcatAn
and San Rafael la Independencia to the west, and with
Jacaltec speakers from Jacaltenango and other towns to
the far west. The Kanjobal region itself borders on the
Mam area to the southwest and on Ixil country to the south-
east (see Map I for location of towns and linguistic regions
in Huehuetenango Department).
Current archeological, linguistic, and other evi-
dence suggest that this very area was the home of the
early proto-Mayan groups whose migrations ultimately took
them to Yucatan and the Peten where they developed a very
high level of civilization (McQuown 1964:69; Vogt 1964a:17;
Vogt 1969:11; Kaufman 1976:82-86). Kanjobalan as well as
Mamean speakers, therefore, probably live in much the same
region their ancestors occupied as early as 3500 B.C.
Contact between Northern Mam and some members of the
Kanjobalan branch has been close, resulting in many simi-
larities (primarily through phonological convergence and
lexical borrowings) even though they belong to opposing
branches of the primary Mayan split which may have occurred
/r 3- 4 -
' I *1
around 2500 B.C. (Kaufman 1974:85). These phonological
correspondences have been responsible for the occasional
grouping of Kanjobal proper with Mamean. (See McQuown 1956
for review of early classifications. For details on early
Mayan diversification, see McQuown 1964, Kaufman 1969 and
The central portion of the region in question, which
includes the Kanjobal-speaking towns, is among the most
inaccessible in Guatemala. It has been little visited and
little studied even in Conquest times. Steep slopes and
ravines do not favor north-south travel, and the towns have
been opened to motor travel from the department capital at
Huehuetenango only within the last two or three decades.
The Acatec towns are still without year-round transportation
except on foot. Only scattered references to this area ap-
pear in early documents and there are no early linguistic
materials known, such as those which exist in substantial
quantities for other Mayan language groups.
According to Vogt 1964a, the region now occupied
by speakers of languages of the Greater Kanjobalan group
is the least culturally disturbed of Mayan areas. This
no doubt results in part from the little attention the zone
received in the years after the Conquest. The area was
undoubtedly conquered by the mid-16th century (Virvez et
al. 1968:7 quoting municipal records; LaFarge 1947:x-xi).
Records of the Spanish incursions into Guatemala reveal
that Don Jacinto de Barrios, while chasing still unconverted
Lacandones in 1695, was in the area and probably visited
Santa Eulalia and other Kanjobal towns. Raids by the savage
and cannibalistic Lacadones are an extremely common motif
in Kanjobal folk tales and undoubtedly refer to real events
about the time of Don Jacinto's expedition (Stone 1932).
Town records for Santa Eulalia also indicate how
little attention was paid to the small central towns of the
northwest. For Santa Eulalia, there is information on the
founding of the town around 1550 and the arrival of Mer-
cederian missionaries soon after. There are a few later
documents relating to boundary disputes, tax records, and
lists of town officials since 1873, but there is little
else (LaFarge and Byers 1931:7; LaFarge 1947:x-xii; and
Virvez et al. 1968:7-10).
The first serious studies made in the region were
sponsored by the Middle American Research Institute at
Tulane University. Their expedition to Jacaltenango and
nearby towns is described in LaFarge and Byers 1931 wherein
Is documented the discovery in this area of the active use
of the old Mayan calendar although in somewhat reduced
form. Later, Oliver LaFarge, who had been in Santa
Eulalia for a day or two on the earlier trip, returned
there for a stay of several months. He made careful
observations of Santa Eulalia life, particularly the re-
ligious observances. His monograph, finally published in
1947, is a rare and valuable work which confirms the
presence in these Kanjobalan towns of a religious system
of Prayermakers which is in many ways very different from
the cofradia system so often described for other Mayan
towns. This ritual system may reflect, therefore, a less
assimilated variety of Mayan religion. Nevertheless, like
the religious system found in any Mayan town, it shows
great evidence of syncretism of Christian elements into a
complex organization which permeates most of daily life.
While it is not within the scope of this study to detail
the nature of Kanjobal religion, reference will be made
later to the presence of certain elements within the seman-
tic system which seem to underlie the ritual.
Ethnographic work in Central America has exploded
in the last four decades and a great many studies have been
produced for Yucatec, Quich6, and the Chiapas highland areas;
but the Cuchumat&n towns have still been somewhat neglected.
The most recent works for Kanjobal-speaking towns include
a land tenure study in Santa Eulalia (Davis 1970) and,
under the direction of the anthropologist of that study, a
community monograph (Virvez et al. 1968). This extremely
useful monograph, written by the town leaders, summarizes
the early documents found in Santa Eulalia, and describes
the old customs and traditions. It reviews the contemporary
state of the community and, in their own words, gives the
people's hopes for their future.
Cuchumatgn people live today much as they must have
done since pre-Colombian times (see Vogt 1964a and h for
detailed discussion of pre-Colombian cultural patterns).
Santa Eulalia is typical of Kanjobal-speaking towns. Of
course, there are differences which affect the character
of a town. For example. Protestantism has had greater
success in Barillas than in Santa Eulalia; the non-Indian
Guatemalan (ladino) population is larger in San Juan Ixcoy;
San Mateo Ixtatin has a valuable commercial resource in its
salt exports. The Ixtatanecos have a terrible reputation
throughout the area for fierceness and dishonesty (probably
due in part to that very salt monopoly). Feuds between
San Miguel and San Rafael have dated from the separation of
the latter from its former municipal center early in this
century. But in spite of these differences, the area
is, for the most part, relatively homogeneous. In fact,
in terms of daily life and overall cultural concerns, life
as lived in the Kanjobalan area is not unlike that de-
scribed for Chiapas (see Vogt 1969 and sources cited there).
This homogeneity of culture will be an important considera-
tion in later chapters as we attempt to examine the lin-
guistic reflections of these pervasive cultural interests.
The people of this area, as in most of Guatemala,
live primarily in settlements which are grouped into hamlets
called aldeas and attached to a town center (see Map II
for aldeas of Santa Eulalia). These hamlets often have
their own chapel and school. For the most part an Indian's
life revolves around these small settlements, which are
generally endogamous and may reflect early patrilineal clan
centers such as those described by Vogt for Zinacantan and
called by him the sna (Vogt 1964i:23-30 and 1g69:140). In
Kanjobal, the root na is still a common general term for
house or household as it is in Tzotzil. Santa Eulalia has
nine such aldeas which vary in distance from twenty minutes
to two days walk away from the town center. Since municipal
lands spread over a considerable distance a town may, as
Santa Eulalia does, have aldeas in several different cli-
matic zones, from low (tierra caliente) to temperate
(tierra templada) to highland or cold country (tierra fria).
The Santa Eulalia municipal center itself and the nearby
aldeas of Nancultac, Pett, and Temux2 are in cold country
at more than 8,000 feet in altitude. The more distant
aldeas such as Chojzunil and Cocald are in lower, hotter
country. This variety allows the cultivation of crops year
round and almost constant food supplies to the town center.
Topographically, the region is one of slopes, ridges, and
valleys which impede travel. The temperatures are mild to
cold and the climate is damp with heavy mists daily through-
out the rainy season (June to late November) and not uncommonly
during the rest of the year in the town center and high aldeas.
Aldeas, together with even smaller subdivisions
called caserios, are part of the larger municipal center
headed by the township which bears its name.3 Even when
new towns are formed from aldeas which grow to sufficient
size, these new centers still retain their ceremonial de-
pendence on the older township. This is the case, for
example, with Barillas which separated from Santa Eulalia
in 188g. Many independent towns in the area seem to look
to Santa Eulalia, however, and there is fair evidence that
Santa Eulalia exercised a general hegemony over the entire
Kanjobal area until very recent times (LaFarge 1947:131f).
Adult status among the Kanjobal is achieved by mar-
riage which usually takes place in the late teens. Children
are heartily desired, but infant mortality is so high that
it is not uncommon for fewer than half a woman's children
to survive to adulthood. The activities of a typical day
are strictly divided by sex, and by an early age children
join their parent of the same sex for instruction in the
duties of adulthood. Women are responsible for the care
of the house and the very young children and for the daily
preparation of food, especially the grinding of corn and
the cooking of large quantities of tortillas, the dietary
staple. Women also sell in local markets; and, in towns
where small item weaving is practiced, it is done by women.
Other craft specialties are the province of men. These
include the weaving of blankets in San Rafael, the making of
pile rugs in San Sebastian Coatan (a Chuj-speaking town),
the plaiting of hats in Jacaltenango, and the production
of the heavy wool-felt jacket worn by all men of the region.
These short jackets are known as capixays and are made in
Santa Eulalla and San Pedro Soloma,which supply them for
the whole region. A few men specialize in making marimbas
(a wooden musical instrument much like a xylophone) and blow-
guns for small game hunting. Most fabric, tools, and house-
hold utensils in Santa Eulalia today are machine-made
imports; there is little market outside the area for the
few craft items made here. This is in contrast to the situ-
ation in modern Cakchiquel towns, where craft items are
part of a thriving Guatemalan tourist industry. Apparently,.
except for wool products, Santa Eulalia and surrounding
towns have never, even in pre-Colombian times, engaged in
much craft production for export. They are outside the main
Guatemalan trade routes and their ancient trading patterns
have probably been primarily with Mexico before the days of
national boundaries and customs tariffs. Any marketing done
between towns is usually handled by men, occasionally with
their wives to accompany them. Women do not travel much,
and never alone, remaining by custom and by preference in
the neighborhood of their families. They are, as a result,
nearly universally monolingual.
The main occupation of most Indian men is subsist-
ence farming. They are responsible for all the phases of
food production, from the burning and cleaning of the fields
to the sowing of corn and some wheat, caring for the fields
with devotion until the harvest. This annual cycle is the
basis for all life in the Cuchumatanes as it is throughout
the Mayan area and has been for Mayan peasants since before
the building of the great Mayan cities. Corn, the mainstay
of civilization in this part of the world, is found in every
meal. Together with the other members of the well-known
Mesoamerican culinary triumvirate--beans and squash--corn
Is the primary food supply. Meat is scarce here although
families raise some chickens and pigs in addition to the
sheep which supply the wool. They sometimes eat meat pro-
ducts but most meat is sold to Ladinos. In Santa Eulalia,
fresh fruits and vegetables are very limited in supply during
part of the year and, in any case, are often too expensive
for Indian families.
Life in Santa Eulalia is somewhat precarious.
Dysentery, parasites, and other chronic ailments take
their toll among both children and adults. Childhood
diseases are still common although active immunization
programs are bringing them under control. Many people
become ill from diseases contracted in the lowlands during
periods of migrant labor on coastal plantations. Occa-
sional epidemics of flu or other virus result in large
numbers of deaths. There is no permanent medical care in
the town and standards of sanitation are low. Neverthe-
less, Kanjobal people are progressive, eager to learn about
ways to improve their lives, and give strong support to
those who have opportunities for instruction in fields that
seem to them to be relevant to their needs, insisting only
that such information be shared with the community. Al-
though they may appear to be shy or suspicious with strangers,
they are, upon acquaintence, very conscientious and cheerful
people, generous and hard-working.
While most Kanjobal-speakers live in the aldeas. the
town center is important to them in a variety of ways.
Probably as they did before the Conquest, they come to
the municipal center to market, to court, to resolve dis-
putes, to provide civil service, to perform religious
duties, and to participate in ceremonies and festivals.
The town center Is the residence for some Indian families
and for all the ladinos, or non-Indian Guatemalans, who
live in Santa Eulalla. In the town center are all the
municipal buildings, the main market, the church, a public
school, a parochial school with its convent, and a number
of commercial establishments owned mostly by Ladlnos.
These include two large general stores, two public dining
rooms (comedores) which cater to Ladinos and strangers, a
new pension which provides rooms for travelers and permanent
housing for out-of-town residents such as school teachers.
There are locally owned truck lines which transport Indians
to the coast as migrant laborers. Ladino families also
supply the wax candles used in great quantities by Indians
in ritual observance. They own the only ovens in Santa
Eulalia and produce several kinds of wheat breads, including
the dark whole wheat bread known as xeka which is very nour-
ishing and greatly prized by Indians. There is a Ladino
tailor and several seamstresses.
Ladino immigration to Santa Eulalia began around the
turn of the century and their dominance in political and
commercial affairs was immediate and certainly out of pro-
portion to their numbers. In 1964, there were about 350
Ladinos in Santa Eulalia while the Indian population was
over ten thousand (Virvez et al. 1968:1). Because many
young Ladinos are leaving home to attend school or find
work in larger cities such as Huehuetenango and Quetzal-
tenango, and because adult men also tend to be away a great
deal, the Ladino population is largely made up of women
and children living in large extended families. Many of
the middle-aged women have never been away from the town,
but their identification as part of the larger Guatemalan
society gives them a very different world view from the
Indians with whom they have almost daily contact. There is
tension between Indian and Ladino elements within Santa
Eulalia and Ladino economic dominance is being challenged,
particularly by one Protestant Indian who owns a large
general store, several trucks and a bus line; but to a large
degree, the society is characterized by mutual tolerance and
a profound sense of separateness. Many Ladinos are bilin-
gual and have a superficial knowledge of Indian custom, but
they do not pretend to share Indian beliefs about the world
in spite of sharing their space within it.
This separation of the two groups is perhaps most
obvious during the celebration of the annual titular festi-
val, the most important celebration of the village. The
period from the ninth to the twelfth of February, in the
height of the dry season, is set aside to commemorate the
patron saint of the town, Santa Eulalia of Barcelona. The
town is the scene of tremendous activity and preparation
in the weeks before and of great excitement throughout.
Although many of the activities planned for the festival
are the same for both Indians and Ladinos, the two groups
do not participate in them at the same time or in the same
place though they often observe each other. There are
dances--for Indians, traditional dancing with a small local
marimba in the parochial school and for Ladinos, modern
social dancing with a capital orchestra imported to the
town hall for the occasion. There are sports--Indian
school children against other Indian school children in
the aldea of Pett and Ladino community leaders and teachers
against other communities in tne town center. There are
new clothes--Indian women wear new white silk overblouses
(huipils), and colorful new hair ribbons, bought locally
in the general stores or from traveling salesmen. Ladinas
buy modern dresses made privately by a seamstress from
designs in a magazine, or bought in another town. There
are processions--Indians carry sacred images on a ritual
circuit and Ladinos convey the town queen to the big soccer
game. There is a native queen as well, but she is crowned
before a different audience and has little role in the
festivities. There is market for everyone, with out-of-
town hucksters drawing huge crowds. There are games of
chance and fortune-telling birds. Ladino relatives from
other towns come for a visit and Indians from outlying
aldeas come too. It is a time for drinking, for gossiping,
for watching costumed dancers, and for town renewal. The
festival serves primarily to bind people to the community,
while reinforcing their places within it. When it is over,
Indians go about the task of planting in the high country,
but there is a period of great depression for Ladinos before
they settle back into their daily lives. Plans for the
next festival begin at once.
While the festival has little religious significance
for the Ladinos, it is a period of intense religious ac-
tivity among the Indians. The prayer leaders pray con-
stantly and make sacrifices in the sacred cave under the
town (Yalan Na), foreseeing the fate of Santa Eulalia in the
coming year. Worship at the great cross in front of the
church increases dramatically and large numbers of wax
candles are burnt there and at other shrines. But reli-
gious concerns are not limited to the period of the town
festival. Religious and ritual elements are extremely im-
portant in Kanjobal life which, after all, is connected
historically to a theocracy. Old religious ways have sur-
vived here although fewer and fewer learn what they mean.
Increasingly, outside forces erode the old customs. Change
has been accelerated since LaFarge described Santa Eulalia's
religious life because Catholic reform movements as well
as Protestant missionary efforts have combined to reduce
the impact of the 'pagan' religion called costumbre. Until
a new community study is done, the amount of change cannot
be known but even if it is very great, many elements from
the old cosmology still shape daily life and still mold
Kanjobal world view. Among these elements are the pre-
occupation with the passing of time, a belief in animal
companion spirits and hill deities, a strong sense of
hierarchy, a special attention to the location and
direction of bodies in space, a subtle use of metaphor, and
a rich folkloric tradition.
0.3 Data Base
Linguistic work for the Kanjobal area includes a
basic description for most of the major dialects in the
Kanjobal group except Kanjobal proper: San Mateo Ixtatin
Chuj (Hopkins 1967); Jacaltenango Jacaltec (Day 1973a and
Craig 1975); and Tojolobal (Furbee-Losee in press). Many
of these works owe a great debt to one of the earliest
descriptions of languages in this branch, a description
of Aguacatenango Tzeltal contained in Kaufman 1971. Work
continues in all the languages of the Greater Kanjobalan
group especially at the Proyecto LingUistico Francisco
Marroquin (PLFM), a linguistic study center now under an
Indian administration which is dedicated to the preparation
of materials by Indians about and in indigenous languages.
In addition to the longer grammars noted above,
several papers and other materials have become available
In recent years on Jacaltec (Craig 1973 and 1976;
Day 1973b), Kanjobal and Acatec (Dakin 1976), Chuj (Max-
well 1976a,Hopkins 1973) and Tojolobal (Furbee-Losee 1973
and 1976). An as yet unpublished computer-stored dictionary
has been prepared for Jacaltec (Day 1971) and is described
in Day 1976. A set of basic teaching materials is available
for Kanjobal (Martin Barber et al. 1973).
As a result of this recent interest in Mayan lin-
guistics generally, a good overall knowledge is now avail-
able about the phonological and morphological systems of
languages of the Greater Kanjobalan group. The differences
between the languages of the Kanjobalan complex appear to
be relatively slight. Since some basic materials are al-
ready available it seems even more useful to take a dif-
ferent approach in newer studies like this one. Reference
will be made to these other works as they are relevant to
the grammatical description in this analysis.
The data for this analysis were gathered in Santa
Eulalia during two field stays. During the first period,
I was affiliated with the Proyecto Linguistico Francisco
Marroquin (PLFM), then in Antigua Guatemala. I was con-
tracted to them from September, 1972, until August, 1973,
to produce teaching materials in Kanjobal, train teachers,
direct the production of literacy materials (see, for
example, Martin Barber 1973a and 1973b), collect texts,
and do general linguistic analysis. During the second
period, from January to March, 1975, I lived in Santa
Eulalia and collected data on positional roots.
In total this study is based on some ten hours of
taped, transcribed, and translated narratives, several
thousand items of vocabulary and elicited materials, a set
of basic teaching materials, an inventory of over 270
positional roots with extensive information on their semantic
characteristics and syntactic distribution, and additional
miscellaneous materials such as examples of children's games
which use positionals and notes on metaphoric extensions and
joking contexts for positional roots. Some unpublished
notes made by K. Dakin on Kanjobal and Acatec have also been
available to me.
The primary informants for both periods were two
exceptionally able Kanjobal speakers: Diego de Diego An-
tonio and Francisco Pascual F. Mr. Pascual was eighteen
years of age at the time we first began working together.
He was living then in the town center but has recently re-
turned with his wife and child to his native aldea of Nan-
cultac. He is functionally bilingual in Spanish, has com-
pleted a sixth grade education, and is literate in both
Spanish and Kanjobal. His only living parent is his mother
who is completely monolingual in Kanjobal. He has worked
as a teacher at the PLFM from time to time since 1973. He
has a great interest in old folk tales and was responsible
for the collection of many very interesting and useful ones.
Mr. de Diego is in his early thirties and lives in
the aldea of Pett with his parents and wife, all of whom are
monolingual in Kanjobal. He himself is functionally bi-
lingual and literate and has also completed six years of
education. He is highly respected in his aldea and in the
community and has held political office as well as serving
onthe board of the credit cooperative. He has been an ex-
tremely dedicated student of linguistics and is devoted to
the study of Kanjobal. Both he and Mr. Pascual are known
in their communities for their work in language maintenance
and in literacy training.
Although Mr. de Diego has been the primary informant
for most of the data for positional roots, many other people
have participated in the collection of data either during
the first or the second field stay. Some have recorded
texts for Mr. de Diego, Mr. Pascual, or for me. Some have
checked data, especially the positional roots, or in other
ways have rendered valuable assistance to this research.
Together, these informants reflect several degrees of bi-
lingualism and all adult age groups; the materials collected
from them include extemporaneous anecdotes, accounts of
daily life, religious stories, animal fables and histories.
A list of these people follows with aldeas of residence,
ages, and degree of bilingualism.
Jose Andr6s Pett 91 monolingual
Eulalia Bernabe Conob 15 bilingual
Lorenzo Caio Juan Temux 27 monolingual
Francisco Diego Ramirez Temux 63 monolingual
Juana Francisco Pett 32 monolingual
Marfa Francisco Conob 58 monolingual
Eulalia Garcia M.
Juan Garcia Miguel
Juan Lorenzo Diego
Alfonso Nicolas A.
Diego Pedro Primero
It is important to
very interested in the
and cultural identity.
Tziquina 60 monolingual
Conob 26 monolingual
Tziquina 38 bilingual
Temux 84 monolingual
Pett 31 very bilingual
Conob 22 bilingual
Nancultac 38 bilingual
Conob 18 very bilingual
Pett 52 monolingual
note here that Kanjobal speakers
maintenance of their own linguis-
They encourage the study of
their language especially when it can result in materials
for their own use. Interest in literacy training in Kanjobal
is very high despite the fact that there is little opportun-
ity for it and that is only by volunteer efforts. It is
Intended that the results of this study and many of the
materials which made it possible be returned to the people
of Santa Eulalia in a form which will be useful to them.
Kanjobal is the Hispanized version of the language
name and is used here because it is the only form which is
common in the literature. In the language itself, the
language is called 'anjob'al, derived from q'an 'word' by
way of the intransitive verb g'anjab'i 'to speak' with the
instrumental suffix -b'al. Literally therefore the term
means 'for speaking' and does not, as LaFarge (1947:iv)
asserts, mean 'four-way or straight language' deriving
from kan 'four'.
The standard national alphabet is used for place
names of official units such as aldeas. The primary dif-
ference between the national alphabet and the alphabet used
in this study and designed by Terrence Kaufman is the use of
c and qu for the velar stop which is represented in the
Kaufman alphabet by k. A detailed description of the
Guatemalan alphabet controversy is given in Kaufman 1975.
Some aldea names have obvious etymologies. These
are some examples: Nancultac from nan 'in the middle' and
t'ulta 'woods (Sp. monte)'; Paicono from pay 'ancient' and
konob 'town (centerJ Tziguin from tz'ikin 'bird' and
na 'house'. Hopkins 1973 presents a very complete analysis
of place nomenclature for Chuj and many other Mayan lan-
guages including Kanjobal. He documents a large number of
borrowings among the languages of the Kanjobal branch.
31n Kanjobal, the town center is usually not
referred to by the Spanish name Santa Eulalia but rather
by the native word konob' (national orthography Conop)
1. GRAMMATICAL SKETCH
Chapter 1 is not intended to provide a complete
description of the Kanjobal linguistic system, but rather
to present in overview the outlines of the system, par-
ticularly those aspects of it which have bearing on the
class of positional roots. These roots function within an
interlocking system of sounds, meanings, and grammatical
processes; they cannot be described adequately unless they
are understood in that context. The grammatical sketch
presented in this chapter provides the framework for under-
standing positional roots and their place in the larger
grammatical system of Kanjobal; this system, in turn, is
characteristic of most other Mayan languages in its group
Chapter 1 is organized into two parts. In the first,
the outlines of the phonology are given, including a pho-
nemic inventory and a set of common phonological processes.
Some remarks are made concerning phoneme frequencies,
morpheme structure, and the processes of nativization which
affect the large number of Spanish loanwords in Kanjobal.
Phonological features and patterns which are especially
relevant to the positional class are treated in Chapter 2.
The second part of Chapter 1 presents a sketch of
the grammar proper. Rather than present a detailed list of
morphemes or organize the sketch in terms of word classes,
the material is given in terms of grammatical categories.
This presentation offers a better opportunity for understand-
ing the kinds of semantic categories which function in
Kanjobal grammar. Such categories are basic to the semantic
organization of Kanjobal and therefore to its grammatical
organization as well. They underlie the organization of the
positional root class as they do the rest of the grammar.
Categories such as person and agency are probably
universally relevant in human language organization. Other
categories, while still optionally present in any language,
seem to be of special importance in Mayan Kanjobal. The
most striking of these include matter classification and
location/direction. The second part of Chapter 1 sketches
the overall relevance of such categories to Kanjobal grammar
while Chapters 3 and 4 will demonstrate in greater detail
the ways in which these categories interact while focusing
on a particular part of the grammar, the positional class.
A practical orthography is used in this study as
a phonemic transcription. Periods Indicate morpheme
boundaries in both the phonemic transcription and the
morpheme-by-morpheme gloss. A free translation is
provided where necessary. In the phonology section the
examples are first given in phonetic transcription. All
phonological examples are taken from recorded text. A
sample text is included as Appendix 0.
1.1.1 Phonemic Inventory
The phonemic inventory of Santa Eulalia Kanjobal
(KSE) consists of five vowels and twenty-five consonants
as indicated in Chart I. An accurate account of the
phonemes of this language was first given by Kaufman (per-
sonal communication) based on dialect surveys made during
1971 as part of a dialect identification program for all
Mayan-speaking towns. This comprehensive survey was carried
out by Kaufman and his assistants at the PLFM and is
described in Kaufman 1976. Ultimately, Kaufman prepared
an alphabet suitable for the efficient writing of Mayan
and other Latin American indigenous languages. This
alphabet and its underlying principles are described in
Kaufman 1973. As applied to Kanjobal, the Kaufman alpha-
bet uses the symbols given in Chart II to represent the
phonemic units given in Chart I.
Since it has been widely adopted and is simple to
write and reproduce, this alphabet will be used as the
phonemic transcription for this study, with three changes.
Kaufman's /h-/, the glottal fricative, which he himself indi-
cates as occurring only in certain prefixes and pronominal
Chart I. Phonemic inventory of Santa Eulalia Kanjobal
labialIalveolarialveolar retroflex IvelarluvularIglottal
high i u
mid e o
Chart II. Kaufman alphabet for Kanjobal
k q 7
tz ch tx
tz' ch' tx'
x xh x 3
words, is not considered phonemic in this analysis and is
not written (see 188.8.131.52 for an account of h-insertion).
Secondly, where the two phonemes /t/ and /x/ occur in se-
quence, they are written t-x to distinguish the sequence
from the phoneme /tx/. This notation was suggested by
Kaufman and is used in Mam.l In Kanjobal the sequence only
occurs across morpheme boundaries. In addition to these
changes, there are several problems with the status of the
glottal stop. These are discussed in some detail in sec-
The phonemes of Kanjobal are treated below by sound
class. Examples are given for each phonemic contrast.
Positional roots are a rich source of minimal pairs and
where positionals are used as examples (always with the
positional stem formative suffix -an), they are indicated
by (P). Where necessary in the discussion of Kanjobal
phonology, phonetic transcriptions are given and standard
phonetic symbols are used. Explanations are given in cases
of unusual symbols. Such transcriptions are given in the
conventional brackets, [ ].
184.108.40.206 Obstruent consonants
There are nineteen phonemic contrasts in Kanjobal
which are produced by an obstruction of the air stream be-
fore Its oral release. In addition to the glottal stop,
there are eighttrue stops, six affricates, and four
fricatives. Both the stop and affricate sets are made up
of paired sounds, identical except for the presence of
glottalization on one item of the pair. The obstruent
sounds will be discussed by subclass.
Stops. The true stops in Kanjobal occur at four
points of articulation: bilabial, alveolar, velar, and
uvular. At each point, there is a plain and a glottalized
phoneme. The glottalization is usually simultaneous, but
will be written after the symbol for the glottalized con-
sonant, i.e. [C7]. Except for /q/ which does not occur in
initial position, all stops occur in all positions in the
word. In final position, a simple stop is usually realized
as heavily aspirated, [C h. There may be light aspiration
in any other position. The glottalized stops are usually
unreleased in final position, [C']. The surface phonetic
contrast in final position is one of release and not of
The bilabial stop, /b'/, is realized phonetically
with both implosion and voice, . In a systematic
phonemic analysis it is interpreted as underlying /p'/.
Imploded consonants are frequently voiced as a natural
consequence of their manner of production (see Chomsky and
Halle 1968:323 for detailed discussion). Some other Mayan
languages have imploded realizations of underlying
glottalized voiceless stops and in these cases as well the
imploded stop is voiced (England 1975:18 for examples in
Mam). This identification of voicing with glottalization
probably accounts for the assimilation of Spanish loans
with voiced stops as glottalized in monolingual Kanjobal
speech. Examples include:
[lak'arto] from lagarto 'lizard'
[?arku] from barco 'canoe'
The phenomenon of phonetic ["?] or [b?] for expected
phonemic */p'/ is so pervasive in the Mayan family that
Kaufman reconstructs /b'/ for proto-Mayan (Kaufman, personal
Many Kanjobal speakers, especially young men with
significant experience with Spanish, have a full series of
nonimploded voiced stops used only in Spanish loans. These
are not considered here to be an integrated part of the
Kanjobal sound system. (See section 1.1.4 for further
comments on loan phonology.)
The contrasts among the bilabial, alveolar, velar,
and uvular stops may be Illustrated by the following
pay 'ancient; skunk' b'ay 'at; in'
jopan 'bright (P)' jab'an 'long, thin (P)'
q'ap 'cloth' q'ab' 'hand'
/q/ does not occur
yoqech 'hearth stone'
xiqa7 'cut something'
t'anan 'staring (P)'
q'oj 'trunk (of a
maq'a7 'hit something'
b'aq' 'pit (of fruit)'
yoqech 'hearth stone'
aqgan 'my foot'
k'olan 'small, round (P)' q'ol
lak'an 'stuck to (P)' laq'an
Affricates. The sounds produced by complete
closure followed by a fricative release occur in three posi-
tions of articulation in Kanjobal: alveolar, palatoalveo-
lar, and retroflex. At each point the plain vs. glot-
talized contrast is found as in the stop series. The
affricates are released in all positions and the secondary
glottalization is quite strong. The palatal and palato-
alveolar sounds have a strong fronting effect on preceding
sounds. This palatalization process is described in 220.127.116.11.
The contrasts in the affricate set are exemplified by the
following items (for convenience, the usual phonetic
equivalentof the orthographic symbol is given).
tzul 'gourd dish'
kutzan 'short, round (P)'
'close together (P)'
chipan 'like many little
ich' 'red lowland tick'
patxan 'flat (P)'
txek'an 'having several
litx'an 'flat (P)'
tx'onan 'having an over-
large head (P)'
'a little pile
'full of liquid'
'empty, vacant (P)'
Contrast with occlusives in approximately the same point of
articulation is demonstrated by the following items:
pichan 'rapped up (P)' pitan
chukan 'having a pointed tukan
end stuck into
wech'an 'scattered (P)' wet'an
ch'ojan 'knees drawn up (P)' t'ujan
'watery mass (P)'
'shaped like a
Fricatives. The fricatives. /s. xh, x, j/ (phonetic
S. J, s, j]) occur in four points of articulation: alveolar,
palatoalveolar, retroflex, and velar. Fricative and affricate
clusters are the most susceptible to reduction (see 18.104.22.168
for details). The presence of the palatoalveolar and
retroflex fricatives (phonetic [s] and [i]) in Mayan and
other Mesoamerican languages has left an effect on the
Spanish spoken in that region by way of a large number of
loans containing this sound. This has been documented in
Scavnicky 1975 for Guatemalan urban Spanish and in Martin
1975 for Guatemalan rural Spanish.
The contrasts among the fricatives are demonstrated
by the following items:
tisan 'round, wide (P)'
joxan 'small pile (P)'
pisan 'seated (P)'
xhi '3rd person said
-uxhtaq 'man's brother'
tixhan 'scattered (P)'
jexhan 'having one arm
/x/ and /j/ as well as /ch/ are phonemes of extremely
high frequency in Kanjobal connected discourse because they
occur in important grammatical morphemes such as aspect
markers. One result of this is the distinctive 'hushing'
quality of spoken Kanjobal.
Glottal stop. The analysis of glottal stop in
Kanjobal is somewhat complex and its phonemic status is
uncertain. It undergoes deletion and movement according to
rules which are imperfectly understood (22.214.171.124.1). Native
intuition on this point is contradictory and variable. In
teaching Indian students to write Kanjobal and other Mayan
languages, the glottal stop is the least consistently
Kaufman has reconstructed phonemic /7/ for proto-
Mayan and it has been analyzed as having phonemic status
In at least some positions in Jacaltec (Day 1973a:9),
Tojolobal (Furbee-Losee 1976:171), and Chuj (Hopkins 1967:17).
Proto-Mayan CV7C and CV7VC roots have reflexes in some
Kanjobalan languages which still contain /7/. Hopkins
(1967:53) discusses in some detail the processes affecting
proto-Mayan /7/ roots in Chuj. However, in Kanjobal, all
such proto roots are realized as simple CVC roots (Kaufman
1969:38f). Therefore, in Kanjobal,  occurs in word-
medial position only at morpheme boundaries where It is
the result of the compounding or affixation of roots with
final or initial glottal stop. In medial position as in
other positions it is subject to various deletion
Glottal stops in medial position can also result
from phonological substitution processes. For example,
[?] is a frequent realization of underlying /q'/ in
morpheme-final position. The following apparent submini-
mal pair is attested in a single test:
[masyontoqhis] 'she put it'
*--- o0 0
[Eagyd'onlowis] 'she gave (them) food'
y- 3' 2
-on pseudointransitive (PIV)'
tog 'action away from speaker (dir)'
low- 'eat (iv)
Ix 'woman (NC1)'
In fact, both [ySon] and [yIgon] are realizations of the
root aq'.a 'to give or put'. (The final vowel is a transi-
tive stem formative which is lost in phrasal constructions.)3
The underlying /q'/ is retrievable in careful speech; its
loss is the result of widespread reduction processes present
in KSE rapid and semi-rapid speech (126.96.36.199.2 and 188.8.131.52).
A glottal stop produced by such a process is not to be
considered phonemic. (The [?] in ix is treated below.)
Additional data on the phonemic status of the glottal
stop is available from a consideration of the apparently
II~ICIC-~- -TIJ I --- J
vowel-initial stems in Kanjobal. All such items are regularly
occur with a preceding glottal stop, especially after a
pause, but may lose it in normal transition. When prefixed
for person, these [?]-initial stems also lose the glottal
stop. The person marking prefixes for this class are
different from those required by true consonant-initial
stems. Those accompanying C-initial stems contain vowels
of their own and form separate syllables, i.e. in-, ko-.
But the markers for [?]-initial stems differ in that they
are single sounds which form natural syllables with the
following vowel, i.e. w-, 1- (see section 184.108.40.206 for further
discussion of person marking in KSE). Other investigators
have noted a class of apparent /7/-initial stems which
prefix like true consonant-initial stems (England 1975:31;
Day 1973a:18). No such class has been found in KSE and all
[C7-initial stems, except the numerals o 'five' and uq
'seven', prefix in the same way, at least insofar as the
data have been examined on this point. Obviously, the
description of person marking is simplified if these stems
can be considered to be vowel-initial and the glottal stop
simply a phonetic insertion. Not one of the recent analy-
ses of Kanjobalan languages has considered initial glottal
stop to be phonemic. Kaufman himself has recently stopped
writing initial glottal stop because of its predicability
in these and other Mayan languages (1975:71).
-- 1 -11-11-- 1
Final glottal stops are usually manifested by the
extreme shortness of the preceding vowel. However, the
presence or absence of glottal stop in this position responds
to stylistic pressures, at least to some degree. The same
item may be produced, heard, and transcribed by native
speakers both with or without the final glottal stop. Some
of these cases of deletion may be explained by lexical iden-
tity (see 220.127.116.11.1); however, the data are very contradic-
tory on this point.
The phonemic status of glottal stop in initial and
final position has important consequences for the descrip-
tion of positional root phonology, especially reduplicative
processes (2.2.1). If glottal stops are recognized as
phonemic in final position, this will uncomplicate that
part of the grammar and also account for rare minimal pairs
[?eJ b'e 'road' [9'0] h'e7 'to grind'
[la7] 1a7 'submerged in liquid (P)' [laq'] laq' 'embraced
Application of rules such as glottal stop deletion (18.104.22.168.2)
and /q'/ [(] may result in the production of any of these
forms either with or without the final glottal stop.
Although problems still remain in determining
whether or not a final phonemic glottal stop is present in
any individual item, the establishment of such an underlying
phoneme does allow for greater simplicity in the description.
The problems presented by traveling and disappearing
glottal stops in medial position can then be solved by late
phonetic rules which will apply both to underlying and
derived glottal stops (22.214.171.124.2).
Taking all these data into account, glottal stop
is here considered phonemic in final position in some roots
and is so indicated by written /7/. It is not written
in absolute initial position or as a variant of /q'/ where
its presence is predictable. This analysis also does not
admit root-medial phonemic glottal stop although compounding
and other combinatory processes may produce stem-medial
ones. The analysis adopted here is admittedly incomplete
and not entirely satisfactory from a theoretical perspec-
tive. Further investigation into glottal stop and glot-
talization phenomena in KSE may reveal that many of these
peculiarities are the result of sound change in process.
126.96.36.199 Sonorant consonants
Unobstructed resonating consonants are much fewer
in number than the obstruents and include two nasals, two
liquids, and two glides. They occur in all positions in
Nasals. There are two nasals in KSE: bilabial
and alveolar. They are subject to assimilation pressures
from following consonants (1.1.2). Their phonemic status
is established with the following sets.
mam 'father' nam 'moth'
mu7 'wild herb' nuq' 'neck'
man 'no' nan 'middle'
jon 'gourd bowl' jon 'our avocado'
miman 'big' inat 'seed'
Liquids. The lateral, /1/, is of high frequency in
KSE since it occurs in roots and in very common affixes such
as -laq 'place of' and -V1 'abstract moninalizer'. (Fre-
quency data are found In 188.8.131.52.) It is rarely in contrast
with the flap, /r/, which is of exceedingly low frequency.
The flap is found in only a few native roots such as tur.u7
'to swallow' and t'iran 'bald, naked (P)'. The usual reflex
of *PM */r/ is KSE /y/. (*PM forms taken from Kaufman 1974:
PM *war- KSE way- 'sleep'
PM *ru7x KSE yax 'green'
The flap occurs in medial position as a realization of Span-
ish d, e.g. asaron 'hoe', or of Spanish rr, e.g. wuru
'burro'. It occurs in an assibilated form in initial
position in Spanish loans as in [fos] 'rosa'. (Section
1.1.4 provides further examples of nativization processes
in the assimilation of Spanish loans in Kanjobal.)
Examples of /1/:
wala7 'I say'
Glides. The glides /w/ and /y/ are phonemic in
Kanjobal, occurring in all positions in the word and before
all vowels. /y/ does not occur phonemically in the same
morpheme following the high front vowel /i/ nor does /w/
occur following the high back vowel /u/. Both glides also
occur in consonant clusters. They are high in frequency
In polymorphemic stems since they are the prevocalic re-
alizations of the first person singular (w-) and third person
singular (y-) possessive prefixes as well as person markers
on one class of verbs (184.108.40.206). These are some examples
of phonemic glides:
yet 'of him; of' watx' 'good'
ya 'pain' wonit 'hat'
waykan 'star' xiwl 'many'
b'eyl 'walk' evi 'yesterday'
pay 'ancient' tx'o 'rat'
txay 'fish' xajaw 'month'
poy 'head strap for pojow 'pus'
The sounds [w] and [y] also occur non-phonemically
in KSE as transition sounds resulting from a glide formation
process affecting vowel clusters. For example, /luin/
'Pedro' is normally produced as [luwin]. Native intuition
is of interest here: Indian students objected to writing
luin with the glide (*1uwin) but consistently write the inter-
vocalic glide when it results from underlying iw/ or /y/ as
in the independent pronouns such as ay.on.ti 'we' (see sec-
tion 220.127.116.11.1 for discussion of formation of independent
There are five vowels in Kanjobal, written and
analyzed as /i, e, u, o, a/. Phonetically the cardinal
vowel qualities are not those reflected by the standard usage
of these symbols. The phonetic equivalences of the phonetic
central tendencies for these phonemes are as follows:
/1/ usually realized as [i]
/u/ usually realized as [u]
/e/ realized rather lower than [e], more like [c]
/o/ somewhat more open than [o] but not as open as
[a], here symbolized [o]
/a/ somewhat higher, almost schwa, symbolized as
[aI] or [a-
Various reduction processes affect vowels and produce other
variants, but the tendencies mentioned here seem to be the
basic ones. The factors causing variations in the production
of vowels are discussed in 18.104.22.168.
These are examples of the vowel phonemes:
an 'plant (noun classifier forn)'
in- 'Ist person singular marker'
janan 'uncovered (P)'
jin 'against us'
Jon 'our avocado'
Jun 'our piece of paper'
Jen 'our toasted corn'
mu7 'type of herb'
1.1.2 Phonological Processes
Several processes affect underlying phonemic forms
in KSE and produce variations in the surface phonetic
realizations. Some of these processes are described in
this section and where possible are expressed as phonological
rules, using the common formalisms of modern generative
phonology. This is made easier by the reformulation of the
phonemic inventory in terms of binary distinctive features.
This reformulation is presented in Table I.
The minimum number of features required for the
description of the KSE sound system is ten. The features and
formalisms are those given in Hyman 1974. Some features such
as voice and delayed release may be specified by redundancy
rules or marking conventions. Such rules can, for example,
express the fact that, in KSE, delayed release is a property
only of strident and continuant consonants.
The ordering relationships governing the applica-
tions of the phonological rules discussed below are not com-
pletely understood. Where possible, ordering sequence is
noted with the rule.
22.214.171.124 Stress assignment
Stress assignment for individual lexical items is
a simple matter in KSE. Setting aside Spanish loans, many
of which must be specially marked, and a few clitics and
particles which do not occur in isolation nor receive
word stress, each KSE stem or root carries one primary stress
on its final vowel. A word then may be partially defined
+ + ,+ ,m+ i + +
+ + 1* 11 .+ i +
+ I' 'I . .. I + + *
+ l i i + + l + I, I +
+ ,+ 1 4 i l +~ 7
i f i i + 4 4 + ,
+ 5* 1 + + i + + .5
Si + + + + 111 : i
I i- + I + I I I I 1 + I
* + 1 1+ 1 1 +* 1+ ,1
* I + + + I + I + I
l 1 @ 4 i j. I I i I +
SI + + i 111 111
SS I + + 111+111+
- | U - - -
C 0 <- = c 41a
-* -^ c Uj 4io u
, Us C - i 4 ) ? ,
S 0 C S r-- s
* i 1 C I- U i11
by its occurrence in isolation with final-syllable stress.
(Further discussion of word-formation is found in section
126.96.36.199.) The KSE stress assignment rule may be formalized
V V / -- (C)#
Stress, therefore, need only be written on those exceptional
words for which stress is unpredictable. In every case,
these are Spanish loanwords which are stressed as they are
in Spanish. Common examples include anima 'people' and
marimpa 'marimba (musical instrument)'. Some early borrow-
ings have been relexicalized to follow the native pattern:
keneya, with final stress, from Spanish guineo 'banana' which
has penultimate stress. Native forms which do not take word
stress in insolation include items such as the tense-aspect
marker chi 'incompletive', clitics such as tog 'action away
from the speaker', and particles such as kax 'and' and ti7
'demonstrative (here)'. Some of these may occur in sentence
or word final position, however, and in such cases may take
[winaqh] winaq 'man'
[winaqti?] wlnaq.ti7 'this man'
In the context of sentences stress placement is
a complicated process. Kanjobal, like English, and unlike
Spanish, has a stress-timed rather than a syllable-timed
rhythm pattern. This means that different numbers of
syllables may intervene between the principal stresses
which are distributed by phrase rather than by word. In
general, a phrase is marked by two principal stresses. One
occurs on the final syllable and is followed either by a
sharp drop in pitch, marking the end of a sentence, or by
a gentle, trailing rise in pitch, marking the end of a non-
sentential sequence. Yes-no questions have a similar final-
syllable stress, but there is a sharp rise in pitch. If
the final word is an exceptional item which carries non-final
word stress, the stress is not displaced to the end but the
final vowel is considerably reduced.
The other phrasal stress occurs one to three sylla-
bles after the beginning of a phrase on a major lexical root.
This stress is optional on short phrases. Longer phrases
may carry one or more secondary stresses.
Because stress does not have a phonemic function in
KSE, it is available to be utilized in paralinguistic ways
which are stylistic and idiosyncratic. For example, stress
may be displaced for emphasis; but the rules governing this
aspect of stress placement are not yet well understood.
The following examples illustrate various stress
patterns and phrasal intonation. (Note that stem forma-
tives in KSE are deleted in non-phrase-final position.)
1. Single word
17- carry (tv)'
-i 'iv formative'
'I carry (it)'
2. Sentence-final phrase
ul 'return' -in 'PIV'
w- 'Isg' w- 'lsg poss'
i7 'carry' inat 'seed'
'...1 return carrying my seed'
3. Non-sentential phrase
[kasmasndi6n ?is sq'J']
kax max.0.nub'.on ix s.q'aq'
kax 'and' ix 'woman (NC1)'
max Compp' s- '3 pass'
p '3' q'aq' 'fire'
nub' 'alight with a little light (P)'
'and she lit her fire'
4. Non-final stress
jos- 'work wood (tv)'
'maker of marimbas'
5. Yes-no question sequence
[k 'mcCEj c Entumin kamcoccj unjq and7
k'am chi.9.och.ej ch'en tumin k'am chi.9.och.ej
k'am chi.0.och.ej a.chej a.wakax-
k'am 'neg' tumin 'money' (Sp)
chi 'incomp' jun 'one'
0 '3' -oq partitivee'
och- 'want (tv)' a- '2sg poss'
-ej 'tv form' no7 'animal'
ch'en 'stone (NCI)' chej 'horse'
wakax 'cow' (Sp)
'Don't you want some money? Don't you want an
animal? Don't you want a horse for yourself?--
188.8.131.52 Vowel harmony
Several derivational processes in KSE require that
the affix vowel be sensitive to the height or backness of
the stem vowel. This harmony is probably a productive pro-
cess in Kanjobal word formation. Patterns of vowel harmony
or inverse harmony have been described for many Mayan lan-
guages (Oay 1973a:28f, 47; Furbee-Losee 1976:174) and are
likely to have been part of the repertoire of proto-Mayan
(Kaufman, personal communication).
Vowel harmony by height is displayed by those suf-
fixes with underlying /o/ which is raised to /u/ in the
environment of the high back vowel. No suffix has an under-
lying /u/. An example is /-obt'anej/ which derives transi-
tive verbs from positional roots (184.108.40.206).
patx 'flat thing at rest (P)'
patxobt'anej 'move a flat thing'
jutx 'long thing (P)'
jutx[u]bt'anej 'pull a long thing'
Rounding harmony occurs with one of the formatives
for transitive verb stems. A large class of transitive stems
take a /-V7/ stem formative. The vowel of this suffix is
determined by the backness (roundness) of the stem vowel:
back vowels take identical [o] and front vowels take [a].
When the suffix follows /u/ stems, the o-raising rule de-
scribed above applies to produce [u]. Examples follow:
tek'.a7 'thresh wheat'
il.a7 'see something'
al.a7 'say something'
txon.o7 'sell something'
muq.u7 'bury something'
The most common example of inverse vowel harmony
involves the suffix(es) -il/-al which occurs in all Mayan
languages with several uses. Often suffixes of this shape
serve as a generalizing or abstract noun formative or mark
a type of possessed noun (220.127.116.11). The conditioning is
based on stem vowel height: [i] for roots with stem vowel
/a/ and [a] for all other roots. Examples include:
yax 'green' yaxil 'greenness'
q'in 'festival; time' q'inal 'life'
k'u 'sun; day' k'ual 'day's journey'
te7 'tree' te7al 'tree-full'
son 'marimba' sonal 'marimba-like'
As indicated above (1.1.1) [h] is not considered
to be phonemic in KSE since it occurs only before certain
vowel-initial pronouns and related person markers. It is
more common in the Soloma dialect than in KSE. This
phenomenon is most common in KSE on the clitic eb' '3rd
person, human, plural' which is realized as [?cb7] [hcb7]
and [h6b7]. The rule is a minor one which applies only
to this and a few other lexical items.
Since it is often missing In normal transition or
realized in voiced form ([(]), it may be part of a progres-
sive weakening chain affecting initial glottal stop; [h]
is an occasional realization of  in other environments
as well, for example, between vowels or after /j/. All
attested examples occur across morpheme boundaries.
chi.#.tx'aj ix sek
chi 'income' ix 'woman (NC1)'
0 '3' sek 'dish'
tx'aj- 'wash (tv)'
'she washes dishes'
Several processes of reduplication are productive
in KSE as they are in many Mayan languages (Day 1973a:45;
Hopkins 1967:82f; Berlin 1963). Many of these reduplicative
processes affect only or primarily positional roots and will
be discussed in the section on positional phonology (2.2.1).
There is both complete and partial reduplication in
KSE. The most common complete reduplication is of the
numeral root jun 'one' which functions as an indefinite
article and, in reduplicated form, specifies plural dis-
un.jun k'al on ay jun.jun j.on
k'al 'together' (
ay 'existential particle'
j- 'lpl poss'
'Each of us has his own avocado.'
Complete reduplication also occurs with other roots, es-
pecially with onomatopoeic roots in which Mayan languages
are especially rich:
Emphatic and distributive meanings are also accom-
plished by partial reduplication of -VC2 in CVC nouns
although now this may not be a productive process. There
are only a few examples. The noun root is not always known.
pqoqo 'dust' (
k'oxox 'tosted tortilla' (
chulul 'red sapote'
matzatz 'pineapple' (Barillas form)
This type of reduplication is similar to that found in the
productive derivation of emphatic adjectives from positional
roots: jop.an 'bright, shiny (P)' > jp.opi 'very bright'.
This process and its semantic implications are discussed
Reduplication of the initial C in CVC roots forms
intransitive iterative aspect verbs (always further
derived by -on 'repetitive') from a class of affect words
which imitate sounds or motion.
tz'aq 'sound of making tortillas'
tz'aq.tz'.on.i 'making such a sound'
jach 'sound of papers rustling'
jach.j.on.i 'making such a sound'
Similar reduplicative processes derive positional roots
18.104.22.168 Glottalization phenomena
In section 22.214.171.124, the analysis of the glottal
stop was presented and processes of deletion, weakening,
and movement affecting glottal stop were alluded to there.
In this section, these rules are described insofar as
possible. Also, processes affecting sounds in the environ-
ment of glottal stop and glottalized consonants are
described. The insertion of  before initial vowels has
been described in section 126.96.36.199.
188.8.131.52.1 Vowel glottalization
Vowels occurring in the environment of [?] or [C?]
may themselves be produced with a very slow vibration or
glottal constriction ([v]), sometimes called vocal fry or
'creaky voice'. This process is optional and is variably
applied: more often if the vowel is preceded by a glottal.
most often If the vowel Is both preceded and followed by
glottalized sounds. Some examples are
1. vowel follows a glottal stop
ay.mi jun.oq cuatro cuerda
ay 'existential particle'
mi 'dubitative particle'
cuatro 'four' (Sp)
cuerda 'unit of measure' (Sp)
'it is some four cuerdas (in width)'
2. vowel between glottalized sounds
xiwil.xa no' q'Lq'
no? 'animal (NCI)'
'already many quetzals'
-eb' 'inanimates (GNunC)'
y- '3 poss'
'his four horns'
Note that vowels which precede [+lo] consonants are not
usually glottalized: /u/ in uk'a?.
The effects of glottalized consonants in the
environment may remain after [?] deletion (184.108.40.206.2).
Compare these phrases:
[ci(floj _i1 yftoqh yEth b]
chi.j.tzaloj ix y.Etoq y.etb'i
chi income '
tzaloj- 'be happy (iv)'
ix woman (NC1)'
y- '3 poss'
'She is happy with her husband.'
max.0.tx'aj ix sat
tx'aj- 'wash (tv)'
ix 'woman (NCI)'
'She washed her face.'
In the first sentence the obligatory initial ['] before
vowels remains on [?is] 'woman (NC1)'. In the second
sentence, the glottal stop has been deleted but after
glottalizing the vowel. Note that in the second of the
vowel glottalization examples above, the  in no7
'animal (MC1)' has been deleted but has not glottalized
the preceding vowel.
The rule of vowel glottalization may be formalized
as a variable phonological rule of the type described by
Labov (1971:465-474), which specifies, by angled brackets,
< >, that the glottalization is optional following
glottalized consonants and occurs most often when the vowel
also precedes such a consonant.
V glottaltal friction]>/ C < C >
This rule must be ordered before glottal stop deletion.
220.127.116.11.2 Glottal stop deletion
Glottal stop may be deleted as an optional process
in KSE rapid speech. This may affect either  proper or
[']deriving from underlying /q'/. This deletion especially
affects noun class markers(NC1) and clitics which begin or
end with a vowel. The following examples are part of a
connected text sequence and demonstrate the highly variable
and optional nature of this rule. (The NC1 no' 'animal' is
identical to the generic noun 'animal' and precedes it or
any specific animal or animal product name. See 18.104.22.168
for further details about NCI.)
k'am.xa jun.oq xim awal y.uj [no'n67]
...y.uj [no' nowax no' j6tomawdl] [rnrapa']
k'am 'neg' y- '3 poss'
-xa 'already' -uj 'by'
Jun 'one' wax 'mountain cat'
-oq partitivee' jot- 'dig up (tv)'
xim 'grain (NC1)' -om 'agentive'
awal 'cornfield' mapach 'racoon' (Sp)
'Now there Isrtt any corn (in the field) because of
the animals, because of the mountain cat, animals
that dig up the fields, the racoon.'
---I --- ----i~---P--- ~ ^--q~PLII~i~i~
Obviously, the deletion of glottal stop is governed by compli-
cated minor rules, perhaps marked for class membership or
syntactic function, since it appears that words of certain
syntactic roles or classes are more often subject to them,
e.g. clitics and noun class markers.
In the next example, the deleted glottal is part of
a directional clitic ek' 'passing by' which is a reduced
form of the intransitive motion verb ek'.i (/-i/ is an
intransitive stem formative which is lost in non-phrase-final
b'et- 'go (iv)'
ek' 'pass by (dir)'
Compare with the following example In which ek' functions
as a main verb:
y.et chi.0.ek' k'u?
y- '3 poss' 0 '3'
-et 'of' ek'- 'pass by (iv)'
chi 'income' k'u7 'sun'
'during the day'
9 U~ ----il~
In this instance the initial glottal stop is still present
but has glottalized the preceding consonant after the reduc-
tion and loss of the intervening vowel (22.214.171.124). (The
preceding underlying consonants have undergone palatalization
(126.96.36.199) and consonant reduction (188.8.131.52).)
184.108.40.206 Nasal assimilation
The alvealar nasal, /n/, assimilates to the point
of articulation of following /p, b', k, k', q and q'/. It
may be realized as [m],  or [q]. This rule may be
L+co bacd / C
which says that /n/ assimilates to place of articulation
before non-continuent obstruents. As written this rule will
assimilate /n/ to following palatal affricates and there-
fore overlaps the palatalization rule, discussed in the next
section. These are examples of nasal assimilation.
b'ay jun pak'
'at one side'
chi income in- 'lsg poss'
w- 'Isg' -b'a 'self'
ut- 'do (tv)'
'I do (it) for myself'
a '2sg poss'
-al 'abstract nominalizer'
y- '3 poss'
q'in 'life; festival'
-al 'abstract nominalizer'
'all over the world'
Palatalization occurs in KSE but it is very light.
The palatalization rule affects strid consonants pre-
ceding the co consonants, /ch, ch'. xh/, and the back
vowel /1/; it may be formulated as follows:
( is a palatalized lateral.)
patx 'flat (P)'
-an 'P formative'
yayji 'having shape'
'it is flat'
[to] i ircEqhl ayth Qh]
tol 'that' cheq- 'send (tv)'
chi income -lay 'passivizer'
in 'Isg' teq 'action toward
'that I was sent here'
220.127.116.11 Consonant reduction phenomena
18.104.22.168.1 Cluster reduction
Consonant reduction applies after nasal assimilation
and palatalization and reduces geminate clusters to long
consonants and, in rapid speech, long consonants to single
consonants. The following examples of such progressive
cluster reduction are given with derivations since the
phonological effects in each case are complex.
1. CC underlying:
/chi...b'es.on.kan ix xij/
b'es 'in order (P)'
Nasal Ass 9
'she puts bowls in place'
2. CC resulting from nasal assimilation:
b'ay 'at/with' mam 'father'
Nasal Ass m
Stress a A
with my father'
3. CC resulting from palatalization;
/tzet chi.i.chi7 tzet chi.y.uk'.ej/
tzet 'what (interrogative)'
chi income '
chi7 'eat (tv)'
uk'- 'drink (tv)'
-ej 'tv form'
CC-.C- c- c-
V devoice i
Stress 1 E
'What does (he) eat? What does (he) drink?'
22.214.171.124.2 Sibilant reduction
The sibilant consonants /xh/, phonetic [i], and, to
a greater degree the more complex /x/, phonetic [s], are
subject to point-of-articulation assimilation before /ch/
and the non-back fricatives, /s, xh/. This probably also
takes place before /tz/ although no example has been
attested in the corpus. As a result of this process, /x/ may
lose its retroflexion completely and merge phonetically
with other phonemes, as in the following examples.
kax chi.y.aq'.on ix
kax 'and' aq'- 'give (tv)'
chi income -on 'pseudointransitive'
y- '3' ix 'woman (NCI)'
'and she gives (it)'
a.ix chi.P.low ix
a 'indefinite demonstrative'
low- 'eat (iv)'
'the woman, she eats'
Such a reduction process is even more pronounced in rapid
speech where /xh/ and /x/ will reduce (become less complex)
when not in clusters.
xhi naq ayin
xhi '3 quotative'
naq 'man (NCI)'
'he said to me'
The rules for sibilant reduction are not well understood.
Day (1973a:16) describes a regressive assimilation and re-
duction process which applies cyclically to sibilant and
affricate clusters in Jacaltec. Further analysis in KSE may
reveal that x--reduction and other consonant reduction
phenomena are part of a similar process and may be collapsed
into one rule.
126.96.36.199 Vowel reduction
Vowels in KSE are subject to many reduction pro-
cesses, especially in rapid speech. They may be lowered,
devoiced, or deleted. Their length may also be affected.
They may be shorter than usual (especially before glottal
stop) or they may be lengthened under stress or for para-
linguistic emphasis. Length of vowels appears to be pri-
marily conditioned by extra-linguistic factors and no further
attempt will be made here to describe it accurately. How-
ever, three other rules affecting vowels will be given.
188.8.131.52.1 Vowel lowering
High vowels in KSE are relaxed and lowered (i.e.
/I/ [I] and /u/ [u]) in the environments of following
/1/ or uvular consonants. This lowering occurs even under
stress and even when [+hi] consonants are elsewhere in the
V [-tense] / -] +lat]
y- '3 poss'
icham 'old man'
-il 'abstract nominalizer'
184.108.40.206.2 Vowel devoicing
Devoicing of vowels after a voiceless consonant is
extremely frequent, especially when the consonant is
fricative. This phenomenon is not obligatory and, like
vowel gottalization (220.127.116.11.1), is variably applied. It
is most common when voiceless consonants both precede and
follow the vowel. (In these examples, note additional
cases of consonant reduction.)
kax chi.O.toj naq
chi income '
'and he goes'
xol xim w.awal.tu7
xim 'grain (NCI)'
w- 'Isg poss'
'among my cornfields'
kax chi.0.ek' sut.ut.oq
ek' 'pass by (iv)'
'and it goes round and round'
This rule is easily formulated.
V < [-voice]>
toj 'go (iv)'
naq 'man (NCI)'
tu7 'dem (there)'
'flat, round (P)'
'Redup -VC2 intensive'
C < >
Note that this process operates in precisely the same way
and in the same environment as another case of phonetic
'spread' phenomena, namely vowel glottalization (18.104.22.168.1).
_ I _I
22.214.171.124.3 Vowel deletion
Vowels are optionally deleted in KSE, especially
from unstressed clitics. This process is probably part of
a larger tendency to avoid vowel clusters across morpheme
boundaries, but also results from the stress pattern in KSE.
Vowel clusters are not entirely absent from the grammar
but are rare (126.96.36.199.1). Various patterns of vowel drop-
ping may be illustrated by the single-syllable aspect
markers, which regularly lose vowels even in careful speech.
toj- 'go (iv)'
-i 'iv formative'
q'an- 'help (tv)'
-a7 'tv formative'
'(he) will help me'
ab'- 'heat (tv)'
-ej 'tv form'
'I heard (it)'
Max regularly loses the entire first syllable when the vowel
is lost. The conditioning for this rule is complex and
undoubtedly sytactic in large part.
In rapid speech, the vowel deletion phenomenon is
particularly common and often produces consonant clusters
which would be completely unacceptable in less casual speech.
xal tinani xin
xal 'emphatic particle'
xin 'well, then, pues'4
1.1.3 Distributional Constraints
In this section the restrictions on occurrence of
phonemes in KSE will be discussed. These restrictions in-
clude those on the combination of phonemes into clusters,
those on the combination of phonemes in roots, and those on
the general shape of roots themselves. Information on
frequency phenomena pertaining to KSE phonemes is also
briefly summarized in this section.
188.8.131.52 Morpheme structure
KSE roots of all classes are predominately CVC or
CV in shape. This is the overwhelming pattern in all Mayan
languages and is a hypothesized feature of proto-Mayan.
Nouns and attributive roots are CVC or CV. Some
nouns have rarer shapes such as CVCVC: sanik 'ant', xajaw
'moon', and taqin 'dry'; CVCV: tzima 'gourd bowl'; and VC:
ich 'chile', ix 'woman' and ak' 'new'. A very small number
of noun roots have the shape V7: a7 'water' and -e7 'tooth'.
All intransitive verb roots carry a stem formative,
/-i/, in citation form which is lost in some derivations or
in phrase formation. The roots themselves are primarily
CVC: way- 'sleep', kam- 'die'. Intransitive directional
verbs of motion which also function as verbal clitics are
primarily VC in shape: aj- 'go up', ul- 'return'.
Intransitive verb roots are commonly CVC or CV(7).
The former require one of several possible formative
suffixes. Transitive verb roots include mag'- 'hit some-
thing', 1o7 'eat', xu7 'blow on'. The irregular quotative
verb root chi occurs without a final glottal stop and with-
out a stem formative. Other transitive verbs are VC:
iq- 'carry', och- 'want', ab'- 'feel, sense, hear'.
All but a tiny handful of positional roots are CVC
as are all onomatopoetic roots such as tzin 'sound of a
bell' and soq 'sound of a large snake'. The distributional
patterns observed in the positional class are discussed in
detail in section 2.1.
Particles and clitics in KSE are also CV or CVC:
ta 'if', xa 'already', k'am 'neg part', kax 'and'. All
personal pronoun formatives are VC: in 'I'; on 'we'.
Numerals include the shapes CVC: kan 'four'; VC: ox
'three'; CVCVC: waxag 'eight'; and, in one case, V: 0
'five'. (This morpheme is usually realized with a glide
when followed by a V-initial suffix. Compare [byeb' 'five
inanimates' and l]k'on 'five animals'. In the practical
orthography these glides are written as if they were
Suffixes may be CV, VC, or CVC in shape: -b'i
inchoativee intransitivizer'; -an 'positional formative';
-b'al 'instrumental'. Prefixes may be C, CV, VC, or V in
shape, all Illustrated by possessive markers: w- 'Isg, V-
initial'; ko- 'lpl, C-initial'; in- 'Isg, C-initial'; and
a- '2sg, C-initial'.
There is a large number of longer indivisible forms
in every stem class but these can be assumed to be frozen
compounds, frozen derivatives, or borrowings. In the case
of frozen forms, morpheme cuts are Impossible in a synchronic
grammar although many can be identified historically.
Examples of these longer forms include meltzoj- 'return (iv)'
echb'an- 'wait for (tv), wojb'atz' 'type of large monkey'.
Many older Spanish loans as well as loans from other sources
often have unusual canonical shapes although the item may
be accepted by speakers as completely native. See section
1.1.4 for a discussion of loan phonology and Appendix B
for a list of common loanwords.
184.108.40.206 Cooccurrence restrictions
All vowels occur in root-initial, medial, and final
position. With two exceptions, all consonants occur in all
positions as well: /q/ does not occur in morpheme-initial
position and /r/ has not been found initially in any clearly
native roots. C1VC2 roots in which the consonants differ
only in glottalization are rare In KSE but do occur:
ch'ich- 'scattered (P)', q'eq 'black', kok' 'small (of
animals)'. Such roots are uncommon in all Mayan languages
and absent entirely in some (Hopkins 1967:49).
Affricates do not occur in CVC roots when the other
consonant is also an affricate unless they are both produced
at the same point of articulation. For example, tz'utz'
'coatimundi' occurs but *tzuch does not. The [-strident]
fricatives have a similar restriction: xixoj 'smell of
fresh meat or blood' occurs but *saxh does not.
220.127.116.11.1 Vowel clusters
Clusters of two vowels are permitted in KSE although
they are not common. They often result from j-deletion
(18.104.22.168.2). Most potential clusters are reduced by the
loss of one of the vowels (22.214.171.124.3). For example, the
incompletive aspect marker chi loses its vowel when affixed
to a vowel-initial stem. Compare
b'ey- 'walk (iv)'
-i 'iv formative'
maq'- 'hit (tv)'
-a7 'tv formative'
'we hit (something)'
Vowel clusters are also broken up phonetically by
the formation of transition glides. These have been dis-
cussed elsewhere (126.96.36.199). Identical contiguous vowels
are reduced in discourse to one vowel. If the vowel sequence
results from segment loss the remaining vowel is more likely
to remain long: maya > [ma-1] 'already'.
188.8.131.52.2 Consonant clusters
Clusters of two consonants are common in KSE across
morpheme boundaries. They result from compounding or affixa-
tion. In some cases it is not possible to separate the
morphemes, and the resulting clusters must be considered
root clusters: -uxhtaq 'man's brother'.
No restrictions have been found on what consonants
can cluster. Fricatives as C1 in clusters are very frequent
because of the high frequency of max completivee' and, for
some speakers, s- '3 pl possessive'. Sibilant clusters
however are especially prone to reduction processes (184.108.40.206).
Initial clusters in early Spanish loanwords were
reduced by the insertion of some vowel (1.1.4) but many
more modern loans do have initial consonant clusters.
Initial CC sequences do not occur underlyingly in native
In natural speech, clusters of as many as three
consonants may occur across morpheme boundaries. Examples
taken from text include the following which do not result
from any reduction processes. Examples of clusters which
do result from vowel reduction are given in section
ix 'woman (NC1)'
s- '3 poss'
y-'s- '3 poss'
'food for him'
'each of his thighs'
iJ.an chi.9.tit w.in
ij 'having the point against (P)'
-an 'P form' tit- 'come (iv)'
chi income w- 'lsg poss'
0 '3' -In 'against, at'
'(it) cones against me'
Frequency counts have been taken on KSE phonemes in
running text, based on underlying forms. Each morpheme was
counted only once even if it occurred several times. Bor-
rowed morphemes were eliminated in calculating absolute
phoneme frequencies. Glottal stops were not counted. Ad-
ditional frequency data on phonemic segments in the posi-
tional class are reported in section 2.1.2 and complete re-
sults of frequency counts are included as Appendix A.
In general, the results for running text show, not
unexpectedly, that the most complex consonants are the
least frequent. These include all glottalized consonants
except /b'/. Even less common than these are /xh/ and /r/.
Both are common in loanwords, however. The low occurrence
of /xh/ is quite startling since the more complex retroflex
fricative has an extremely high rate of occurrence. Of the
five most frequent consonant phonemes three are sonorants:
the two nasals and /1/. The others are /t/ and, inter-
estingly, /x/. Unlike other /c'/s, /b'/ is rather common.
Among the vowels, /a/ occurs three times as often
as any of the others; /o/ and /i/ occur an almost equal
number of times. /e/ and /u/ also occur with equal fre-
quency but a third less often than /o/ and /i/.
Frequency data are hard to interpret and are subject
to a large number of variables. Nevertheless, frequency
of sounds is undoubtedly related in some way to relative
naturalness in phonology and other questions. The data for
KSE are made available here in the hope they may be useful
to investigators interested in such questions.
1.1.4 Nativization and Loan Phonology
Although Spanish is the source for the greatest
number of borrowings into KSE and other Mayan languages,
it has not been the only one. Before the Conquest Mayan
speakers were in contact with speakers from language
families such as Uto-Aztecan (Nahuatl), Zapotecan, and
Mixe-Zoque. Kaufman (1964:131-135) discusses non-Spanish
borrowings throughout the Mayan family. Several items that
he mentions are present in modern KSE and include such
forms as kakaw 'cocoa' from Nahuatl and asun 'cloud' and
unin 'child' both from Mixe-Zoque. These together with a
large number of early Spanish loans are not recognized by
native speakers as other than native KSE words. None in-
cludes phonological segments which are non-Mayan and are
only identifiable by their structure or by linguistic
comparisons done with the source languages.
Words of Spanish origin are, of course, the most
frequent loans into KSE as into other Mayan languages.
The borrowing process has been a constant one for over
four hundred years; and, consequently, there are degrees
of assimilation or nativization (cf. Kaufman 1971:12f) as
well as examples of reborrowing. Among the very early
borrowings are a large set of Spanish proper names which
have completely replaced indigenous names. These, together
with the names of animals and products introduced by the
Spaniards, illustrate a series of assimilation processes
affecting the non-Mayan segments. Many of these processes
continue to be observed today in the speech of monolingual
and bilinguals. Even functional bilinguals have many of
these assimilation processes both while speaking Spanish
and on loanwords in KSE. The most common of such processes
are described below.
1. Substitution of /p/ for Sp. (f]: KSE pinka
'plantation' < Sp finca; KSE opisyo 'work'
substitution heard in modern KSE and other
Mayan languages. Its pervasiveness is
probably not unrelated to the high frequency
of [fp] a voiceless bilabial-labiodental
articulation of /f/ in Guatemalan Spanish
(cf. Predmore 1945).