ArticlePDF Available

Predicate classes in Yucatec Maya

January 1993

Authors:

University of Erfurt

Abstract The purpose of a classification of predicates is to provide the basisfor any explanations of different grammatical behavior of predicates that may be grounded in differences of lexical semantics. The scope of such a classificationhas to comprise not only verbs, but all words that can function as predicates, because the criteria must not be structural. The criteria of such a classificationare therefore such as characterize situations, participants and their roles in situations. The features form a hierarchy and are defined operationally

Content uploaded by Christian Lehmann

Content may be subject to copyright.

CLIPP

Christiani Lehmanni inedita, publicanda, publicata

titulus

Predicate classes in Yucatec Maya

huius textus situs retis mundialis

http://www.uni-erfurt.de/

sprachwissenschaft/personal/lehmann/CL_Publ/

pred_classes_ym.pdf

dies manuscripti postremum modificati

29.07.1995

occasio orationis habitae

–

volumen publicationem continens

Función 13/14

annus publicationis

1993

paginae

195-272

Predicate classes

in Yucatec Maya

Christian Lehmann

Abstract

The purpose of a classification of predicates is to provide the basis for any explanations

of different grammatical behavior of predicates that may be grounded in differences of

lexical semantics. The scope of such a classification has to comprise not only verbs, but

all words that can function as predicates, because the criteria must not be structural. The

criteria of such a classification are therefore such as characterize situations, participants

and their roles in situations. The features form a hierarchy and are defined operationally

so that tests for the assignment of a given predicate to some class can be formulated.

Some 450 stems of spoken Yucatec Maya are classified in this way. Generalizations about

the structural behavior of the semantic classes are proposed. The grammatical and lexical

operations that transfer a predicate from one class into another are reviewed. Finally, an

attempt at formulating the underlying typological principles is made.

Christian Lehmann

The term `participant' will be treated as an inanimate noun.

although Owen 1969 does make it appear as if this were the case in the object area under consideration

1. Introduction

The purpose of classifying the predicates of a language will not be immediately obvious; the enterprise

might even seem obsolete from the start. Aren't there as many different classifications of predicates,

or verbs, as there are linguists, or at least, as there are criteria by which verbs can be distinguished?

So what is the possible fruitfulness of such a classification?

A classification of a set of objects gains significance to the extent that the criterion employed correlates

with other criteria that differentiate the same set. If we classified the verbs of Yucatec Maya (YM)

according to their first phonemes, we should find litte, or no, connections of the resulting classification

with one that resulted from the application of any non-phonological criteria. If we classified them by

their number of syllables, the case would be a bit better, because some conjugation classes are generally

not open to polysyllabic roots. (Thus, there are no polysyllabic transitive base verbs.) However, the

best correlations are achieved if we employ semantic and structural criteria. For instance, the class

of YM verbs that in incompletive aspects bears the suffix -ik (morphological criterion) is essentially

the same as the one that accepts a direct object (syntactic criterion) and as the class of verbs that

designate situations involving (at least) two participants between which

there is a gradient of control

(semantic criterion).

If the correlation between structural and semantic classifications were complete, this would mean that

there was a biunique association mapping each structural feature on one semantic feature. We would

then have a combined semantic-structural classification. However, things are not that easy.

As we

will see in more detail, there are verbs carrying the suffix -ik which are not transitive and designate

a process involving only one participant (e.g. íichin-t `take a bath'); there are others that do not carry

the suffix, are not transitive, but designate a process involving two participants between which there

is a gradient of control (e.g. tsikbal `tell' and ts'íib `write'); and there are yet other cross-associations.

It is therefore wise not to mix structural and semantic criteria from the start, but instead do a structural

and a semantic classification separately and, in a second step, compare the two.

Some of the classes that can be established on a semantic basis are well known in Mayan studies and

have, indeed, been well known for centuries in Classical Philology. Among them are perception verbs

(verba sentiendi), communication verbs (verba dicendi), verbs of feeling and affect (verba affectuum),

motion verbs, and others. One class of verbs that has received much attention in recent typological

linguistics, in connection with studies of ergativity, are verbs of bodily impingement, such as `beat',

`kick', `kill'. The reason why such semantic classes are well-established in linguistics is, apparently,

that they have been found to be of morphological or syntactic relevance, i.e. to correlate with structural

classes, in well-studied languages. Any descriptive study that tries not only to represent the structural

facts, but also, to understand them, will be interested in the potential relevance of such semantic

classes.

For the reasons mentioned, a couple of semantic classes have been much used in the description of

both Mayan and other languages. Statements such as the following are very common: in this language,

perception verbs take an accusativus-cum-infinitivo construction; verbs of communication take a

subordinate clause introduced by a certain conjunction as a direct object; stative verbs are not used

in the progressive aspect; and so forth. The usefulness of such semantic classes is thus beyond doubt.

However, they have mostly been taken for granted, understood on intuitive grounds, and have seldom

been defined. Even where some such class has been studied as a lexical field – as there have been

well-known studies of verbs of communication, of motion, of cooking –, attention has generally been

Predicate classes in Yucatec Maya

See Lehmann 1989 for the descriptive theory and Lehmann 1991 (with further references) for a theory

of predicate classes.

Fieldwork in 1991 was supported by the Deutsche Forschungsgemeinschaft, whose help is gratefully

acknowledged.

Cordial thanks are due to Ramón May Cupul, my main informant, for his keen interest in my problems,

his inexhaustable patience and good humour.

devoted more to the internal structure of the field than to its position in the overall verbal lexicon and

its delimitation against other lexical classes. The reason for this situation is that comprehensive

classifications of the predicates of a language on a semantic basis have seldom been undertaken. For

YM, I am aware of no such study.

In an article such as the present one, only a first step towards such a classification can be taken. My

purpose here is

- to present the semantic criteria that are relevant to such a classification, introducing, at the same

time, an appropriate terminology,

- to combine the criteria in a systematic way such that they yield a comprehensive classification,

- to fill the major resulting classes with example stems of YM – without aiming at exhaustiveness

–, and

- to show some ways in which these lexical classes are structurally relevant, i.e. behave in a

homogeneous way as to morphology or syntax.

The classes found should be usable in studies of verbal morphology or syntax. The primary purpose

is, thus, a purely descriptive one: to provide some conceptual and terminological tools for the grammar

of YM. At a secondary level, the study represents the application of a general model of grammatical

description to a particular language and, thus, serves to test the model and to provide a specimen of

how it may be applied to different languages.

The variant of YM described here is the language currently spoken in Quintana Roo. The data were

collected in 1988 and 1991, during two fieldwork stays in the small village named Yaxley, município

Felipe Carrillo Puerto, Quintana Roo.

This variety differs in a number of respects from the one that

underlies most of the work of the Chicago and New Orleans schools of Mayan studies. Informants

were mainly male adult Maya-Spanish bilinguals.

2. The domain of classification

2.1. Preliminaries

If we were to do a structural classification, a proper domain of application would be the verbs of YM.

These could be defined as those lexemes which may be directly preceded by one of a set of auxiliaries

that consist of an aspect morpheme and a personal clitic of the subject/possessive paradigm, so that

the combination, either by itself or supplied with some suffixes of a well-defined set, yields a complete

sentence. We could then proceed to establish derivation and conjugation classes on the basis of the

available affix positions and the paradigms that may fill them; or we could establish syntactic classes

on the basis of the combinatorial potential of verb forms in the verb phrase and the clause. Thanks

to the work of, among others, Blair 1964, McClaran 1969, Owen 1969, Po'ot Yah 1981, Stefflre 1972,

the grammar of YM is relatively far advanced on this track.

Given the purpose of this study, the verbs of the language are not the proper domain of classification,

because they cannot be identified on semantic grounds. A prototypical action predicate such as `hit'

will emerge as a verb not only in YM, but probably in all languages. A predicate such as `know', on

Christian Lehmann

Both ohel and k'ahóol are relational nouns; cf. §§ 2.2 and 4.3.2.

The primary function of a linguistic sign is the one which requires least grammatical, especially

morphological, apparatus; cf. Kury1owicz 1936.

The -t suffix on a class of transitive verbs is not an inflectional, but a derivational suffix and thus cannot define a conjugation

class. More on this in § 6.3.1.4.

the other hand, is a verb in English and several other languages, but not in YM.

However, we want

our classification to comprise both `hit' and `know'.

Reference and predication are basic notions of linguistic theory which will not be defined here. The

predicates of a language will be conceived as those words whose primary function

it is to predicate,

rather than to refer. In the case of YM, the case is a bit complicated by the fact that – given the lack

of a copula – most referring expressions can also function as predicates. Despite this complication,

the above definition correctly identifies the predicates as those words which may function in

predication but not, without further ado, in reference.

Words such as `become', `turn into' etc. will be considered predicates. Semantically speaking, it is

their predicate nominal which is the relevant predicate here, the verbs mentioned serving only as more

or less grammatical verbalizers. The decision taken here is motivated by the structural fact that these

verbalizers do, in fact, fit into some of the major word classes which form the object area under

investigation. In consonance with this decision, their predicate nominal will be considered a participant,

which – it should be clear – again is a gross oversimplification.

2.2. Form classes of Yucatec Maya

The definition excludes adverbs and most grammatical words from the domain of this study.

Something needs to be said about the word classes and subclasses which are included. In order to avoid

circularity, these need to be defined here by their formal characteristics.

A noun (n) can be recognized by two properties: 1) it can be inserted in the frame le -o' `that ';

2) it cannot directly modify another noun (as adjectives can). Examples are x ch'úuppal `girl' or ta'n

`ash'. A relational noun (n/rel) is one which must either take a possessive attribute (which entails

a possessive clitic) or else must undergo some morphological modification in order to avoid this.

Examples include suku'n `elder brother' and p'éek `hate' (lit. `hater').

An adjective (adj) is a word which can directly modify a noun as an attribute, being either preposed

or postposed. Examples are k'oha'n `sick' and úuchben `old'. Most adjectives can also be used

predicatively and then take absolutive suffixes. Almost all of them can undergo derivation by the suffix

-tal (s. §6.2.4). There is a small class of words, with bèey `such' and he'l `voici/-lá' among them, which,

although sharing with adjectives their morphological properties, cannot be used attributively. They

will be treated as a subclass of adjectives.

A modal (mod) is a word which can only function as the predicate of a sentence but which does not

inflect. Examples are yàan `exist' and k'abéet `necessary'.

A verb is a word which can be preceded by an auxiliary (an auxiliary is a complex of a tense/aspect

word and a subject clitic; both of the latter sets are defined by enumeration). A transitive verb (v/tr)

is one which is preceded by a subject clitic and followed by an abolutive suffix. Examples include

kon `sell' and áan-t `help'.

A reflexive verb (v/refl) is a transitive verb which takes the relational noun

báah `self' (with its possessive clitic agreeing with the subject) as a direct object. An intransitive verb

Predicate classes in Yucatec Maya

The personal enclitic morphemes which precede verbs are called subject clitics because their reference

comprises the actor of the transitive verb and - in incompletive TAM categories - the sole actant of the

intransitive verb. The personal suffixes are called absolutive suffixes because their reference comprises the

undergoer of the transitive verb and - in completive TAM categories - the sole actant of the intransitive verb.

TAM categories

class label

incompletive completive

examples

v/itr -0 -0 -n(-ah/-ak) alkab `run', xiknal `fly'

v/itr -Vl -Vl -0 hàan `eat', ch'íih `grow'

v/itr -tal -tal -chah 1 -lah xol `kneel', kux `live'

F1. Intransitive conjugation classes

(v/itr) is one which only takes one of the two devices defining a transitive verb at a time.

Three major

conjugation classes of intransitive verbs will be distinguished according to suffixes which appear in

certain tense/aspect/mood (TAM) categories. They are displayed in F1.

Apart from these, there are a few irregular verbs such as bin `go'. It is profitable to assign the three

conjugation classes to two major groups, namely those that take a suffix in incompletive TAM

categories (and it is no coincidence that that suffix ends in vowel plus l in both of the classes

concerned), and those that do not. All of these classes are productive. However, some facts indicate

that the first conjugation class is the unmarked one for intransitive verbs. Polysyllabic roots, among

them all Spanish loan verbs, generally join this class.

3. The functional criteria

3.1. Classificatory features

We may conceive of the meaning of a predicate as a bundle of semantic features, some of which relate

to the situation designated, others to the participants involved and yet others to the specific roles that

the participants have in the situation. Let us call these situation features, participant features and

role features. As has long been known, the features in each of these classes differ as to their

classificatory potential within the lexicon: some are very general and applicable to all or most

predicates, others are more specific and differentiate only among members of one lexical field. In a

study such as the present one, which aims at a comprehensive high-level classification, only the

generic, or classificatory, features will play a role.

These differences in generality are reflected in the language system in the following way:

- The classificatory features have grammatical relevance. E.g., situation features such as

[+/-dynamic] (state vs. process) are reflected in different word classes and in processes of

word-formation. Participant features such as [+/-human] determine the selection of different

pronouns. Role features such as [+/- control] manifest themselves in different syntactic functions

of actants.

- Mid-level features are reflected in subregularities of the combinatory potential of words. The

difference between bodily vs. mental attainment of an undergoer is brought out partly in YM by

the fact that many predicates of mental impact do not combine directly with their undergoer noun

phrase, but instead with u yóol NP, `NP's mind'. Such features may also be reflected in derivational

categories.

Christian Lehmann

stative dynamic

class

inclusion

property state durative

process

terminative

process

ingressive

event

punctual

event

F2. Situation features

- The low-level features, e.g. the difference between sa'sakal `cough' and he'síin `sneeze', have no

structural relevance.

Although I have been speaking of high, mid and low level features and have been exemplifying with

binary features, it will be understood that the differences considered here are gradient and shiftable

rather than abrupt and steady, and that the classes to be found are fuzzy and open rather than mutually

exclusive and jointly exhaustive. Insofar as we are dealing with lexical features which have varying

degrees of grammatical relevance, the whole study of predicate classes is one in the border area

between lexicon and grammar.

The situation features to be analyzed can be conceived of as a gradient leading from the pole of

maximal time-stability, stativity to the pole of maximal transiency, dynamicity. The predicate classes

relevant in this context are displayed in F2. On the side of dynamic situations, a process may be atelic

(durative) or telic. A telic process may be bounded at the end or at the start. If the former, it is

terminative; if the latter, it is an event. If the event is only bounded at the start, it is ingressive; if it

is also bounded at the end, it is punctual.

Obviously, F2 provides the basis for aktionsarten as lexical and, in particular, derivational categories.

It is, at the same time, the semantic basis of the noun-verb continuum. At the left pole, we only find

nouns; at the right pole, only verbs. However, the primary function of nouns is to refer, not to predicate.

In accordance with what was said above on the scope of this study, class inclusion will only be

mentioned in passing.

The following participant features will be of relevance: A participant may be propositional or

non-propositional. In the latter case it may be a place (location) or an entity. An entity may be an

inanimate thing or an animate being. The latter may be an animal or human.

The role features can be represented in a two-dimensional space, as in F3. The vertical dimension

relates to the degree of involvement, of centrality vs. marginality of the participant to the situation.

Sources are less directly involved than goals, experiencers less directly than patients. The horizontal

dimension differentiates between directly involved participants according to the extent to which they

control or are controlled by the situation, i.e., to which they are actors rather than undergoers. At the

pole of the prototypical actor we find the agent; at the pole of the prototypical undergoer, the patient.

Between them, there is a gradience of control where, among others, the participants of a predicate

such as k'áat `ask' find their places. A situation that has a controlling participant is an action.

Predicate classes in Yucatec Maya

To be honest, I did, in fact, partly proceed in this way. But this was a heuristic device which was to help

me to find representatives of the conceivable classes as fast as possible. The methodological justification of

the results is, in any case, as follows above.

F3. Involvement of participants

What is universal and what is language-specific about a classification with this basis? It is assumed

that the dimensions of classificatory features introduced above have a cognitive basis and are, therefore,

universal, while the low-level features may have a cognitive or a cultural basis and will be increasingly

language-specific. Moreover, while the classificatory features themselves are universal, they admit

many different combinations among each other and with the low-level features. They thus form

complex bundles wich may or may not turn up in the form of a predicate in a specific language. Thus,

while English has both lie and cheat, thus differentiating between a purely linguistic act between two

human participants and one which involves, in addition, the transfer of a value between them, YM

only has tus, which covers both situations.

For this reason, a classification of the predicates of a language on a universal semantic basis cannot

proceed in an aprioristic manner. I.e., we do not start from a presumedly universal concept such as

`burn', classify this as representing a terminative process whose central participant is an undergoer

which is a physical object, then look for the YM word representing it, say èel,

and thus know the

correct class allocation of èel beforehand. Instead, we do a semasiological analysis of YM predicate

words, come up with the roots èel, tóok and tàab corresponding to the notion of `burn' (each with their

derivatives), and examine how these behave with respect to our dimensions, so that we can allocate

them to their appropriate classes. The result is, in this case, that èel does come close to the presumed

universal concept, although it is not necessarily terminative. Tóok has the animate being causing the

fire as its central participant. And tàab has both the actor and the undergoer as central participants

and is ingressive.

3.2. Operationalizing the features

3.2.1. Testing semantic properties

The semasiological procedure alluded to presupposes the availability of tests by which we find out

the class allocations. Such a test consists essentially in inserting the predicate in question in a

prefabricated sentential frame. The behavior of the predicate under examination is assessed accordingly

as the resulting sentence is fine or not. It goes without saying that the application of such a test is hardly

ever a simple matter; it is not a discovery procedure. For example, in order to test whether a human

participant has control, the imperative test is usually applied. However, this does not reduce to looking

Christian Lehmann

It should be clear that this is an oversimplification. For a multi-layered model distinguishing between

speech act, propositional content and state of affairs, see Hengeveld 1992, ch. 1.5.

whether the verb in question has an imperative form, or whether a sentence containing it is

well-formed. Most non-control verbs can, in fact, be used in such a frame; cf. receive this present as

an expression of our gratitude! What is to be tested, instead, is, whether such a sentence can be used

as a true command (cf. Dik 1978). This involves, among other things, that the speaker has no magical

forces and that he may sanction non-compliance (thus excluding, e.g., dream nicely!). As a result of

such complicated conditions, the tests yield many borderline cases.

If there is a paradigmatic relationship between A and B such that B is derived by a regular

morphological process from A, then it is often the case that B possesses a semantic feature for which

A is not specified. For instance, English verbs equipped with the particle up, like eat up, use up (B),

are terminative, while eat and use (A) are durative-terminative. One may say that only derived verbs

have aktionsarten, while roots may have a `verbal character' (cf. Leiss 1990 with references). While

these two categories are coextensive with respect to their subcategories, aktionsarten are much more

pronounced than verbal characters and therefore easier to diagnose by linguistic tests. Although this

must be granted, the conclusion cannot be that lexical semantic features can only be set up for derived

categories. We will assign basic and derived YM words equally to the predicate classes. However,

it must be understood that the class membership (the verbal character) of the former is much more

shiftable than that of the latter.

A limitation of many tests is that they can differentiate only between members of the same word class.

Thus, the imperative test can be used to find out about control properties of verbs, but not of adjectives.

This is because the test uses a syntactic frame with a slot suitable for a given word class, but not for

others. In general, however, this limitation is not too serious, since the question to be tested usually

arises only for members of one word class. The question of control, e.g., is settled independently for

adjectives.

Despite such shortcomings, the tests provide a valuable means of exteriorizing, objectivizing and

controlling speakers' and linguists' intuitions.

3.2.2. Participant features

In a study of predicate classes, only such participants need be considered which are inherent in a

predicate. To the extent they are exteriorized, they will be represented as actants. See §3.2.4 on the

problem of determining whether an NP is an actant. A predicate may have from zero to three inherent

participants; in YM, however, it needs to have at least one.

The tests for determining participant properties are straightforward: combine the predicate in question

with participants of different lexical classes and see whether the result makes sense and remains inside

the range of natural polysemy. The latter corollary goes by itself: we do not want to class túun k'áaxal

le ha'o' `(the water is falling =) it is raining' and túun k'áaxal le ba'lo'bo' `the things are getting stuck'

as uses of one predicate (k'áax is homonymous).

The participant features will be interpreted as follows. By proposition is meant an abstract object

that may be represented by a – possibly reduced – clause. The ontological status of the proposition

will not be considered; kah `begin' and a'l `say' will likewise be assumed to take a propositional

participant.

Participant features of predicates do not simply replicate semantic classes of referential expressions;

selection restrictions of verbs do not replicate noun classes. What is of relevance for a predicate is

not, strictly speaking, absolute properties of the participants, but rather, certain qualitatively different

Predicate classes in Yucatec Maya

a. Máaxih?. `Who is it?'

b. Ba'nih? `What is it?'

F4. Class inclusion test

a. (-en). `I am/it is .'

b. yàanil(-en). `My/its current condition is .'

F5. Property test frame

kinds of involvement that the participants are capable of. As a predicate feature, the feature [human]

has the interpretation `requiring intellectual power of the participant'. The feature [animate] has the

interpretation `requiring life-force (thus, an autonomous source of energy granting the possibility of

independent motion and/or of exerting force) of the participant'. Typically, participants linked to

predicates with these features will be human and animate, resp. But any thinking entity, like a computer

or an agency, or any entity exercising force, like the wind, will be acceptable participants of predicates

with the respective features. Similarly, the participant feature [place] allows for any participant that

can function as a place; and the feature [thing], for anything that can be treated as a thing, including

animate beings.

3.2.3. Situation features

In principle, all types of situation may have up to three participants. However, the more stative

situations tend to have less participants. There is no class inclusion and no property with three

participants and relatively few with two participants.

Class inclusion predicates a substance over the subject. Such a predicate responds to the questions

in F4, while none of the others do.

I assume this test differentiates between class inclusion and property predicates also if both are

represented by nouns. Thus, huntúul hòoykep `a lazy-bones' should not be among the expectable

answers to F4.a. However, this remains to be investigated in detail.

The difference between a property and a state is one between an essential characteristic vs. an

accidental condition of the participant(s). It is easily exemplified from Spanish with soy moreno `I

am brown' (by nature) vs. estoy moreno (since I have had much sun). However, as this example shows,

often the same lexical predicate may be used either in property or in state predications, so that this

feature need not necessarily constitute predicate classes. On the other hand, there are examples of

predicates that neatly fall into either class; e.g. chowak `long' (property) vs. su'lak `ashamed' (state).

We will therefore have to differentiate between predicates that are assignable to one of the classes,

vs. such predicates which are indifferent to the distinction between property and state.

Since in YM there are no verbs designating properties (such as Engl. resemble), no test for verbs is

necessary. For adjectives and relational nouns, the pair of frames in F5 may be used.

If the variants F5.a and b yield no difference in meaning or only one of emphasis, we are faced with

a state predicate; e.g. k'oha'n `sick'. If version F5.a signifies an essential characteristic, and b a current

situation, this means that the predicate by itself signifies a property, but may be converted into a state

predicate if equipped with yàanil. If version F5.b is impossible, this allows, in principle, of two

interpretations. Either we are faced with a class inclusion predicate (which cannot be converted into

a state); e.g. a watan `your wife'. Or we are faced with a state predicate that cannot be made `more

stative', so to speak, by the addition of yàanil; e.g. biha'n `gone'. These cases may be kept apart by

applying F4 or F6.

The second test for property vs. state is the frame F6.

Christian Lehmann

Leti'e', huntúul máak.

hump'éel ba'l.

`That is a person

thing.'

F6. Property test frame

a. Ts'o'k u óoxp'éel minùutoh/òorah/k'ìin.

`It has -ed for three minutes/hours/days

b. Táantik u . `It ed a moment ago.'

F7. Process test frame

a. H óoxp'éel minùutoh/òorah/k'ìin.

`It -ed (for) three minutes/hours/days

b. H ichil óoxp'éel minùutoh/òorah/k'ìin.

`It -ed within three minutes/hours/days.'

F8. Terminativity test frame

The alternatives in F6 have, of course, to be chosen according to participant properties. Instead of máak

and ba'l, it is often advisable to use more specific class nouns. If the sentence is impossible, this is

a hint that the predicate (normally, adjective) in question is a state predicate and therefore cannot be

used to characterize the nature of a person or thing (e.g. wi'h `hungry').

Although the term `durativity' is usually not applied to states, they are as durative as atelic processes.

The difference between a state and a process resides in the fact that something changes during the

process. This may entail that afterwards the participant is in a different state than before. This may

be a different condition of its body (e.g. ch'íih `grow') or mind (e.g. kan `learn'), or it may find itself

in a different place (e.g. xíimbal `walk'). Or else the change may be limited to the time span during

which the process goes on, so that afterwards everything is essentially as it was before (e.g. k'ay `sing').

In any case, a process requires the investment of energy, while a state does not.

If predicates to be tested are not verbs, they can safely be assumed to designate states rather than

processes. If they are verbs – very few state predicates are –, then if they form the resultative participle

in -a'n (cf. §6.2.3.1), they designate processes (e.g. kah `live, dwell'). Similarly, if the verb in question

fits in the frame F7 (where an appropriate time unit has to be chosen), it designates a process rather

than a state.

It is this test that makes us classify páahtal `can' as a state, but kul `sit' and wen `sleep' as process

predicates.

Every process will start at one point and end at another point. Thus, the difference between atelic

(durative) and telic processes is not that the former do not start or end. Instead, a telic process has

a built-in boundary beyond which it does not proceed. If it has a built-in endpoint, it is terminative;

if it has a built-in start point, it is ingressive. Atelic process concepts contain no information about

start or end; they can in principle continue indefinitely.

A test commonly used to differentiate between atelic and telic, in particular, terminative processes

would take the form of F8 in YM.

Here the verb has to be inserted in the simple past tense (e.g. meyahnah `worked'), and an appropriate

time unit has to be chosen. If version F8.a is fine, then the verb is durative. If version b is fine, then

it is terminative. This test has been applied with satisfactory results to various languages, including

Predicate classes in Yucatec Maya

There is another hint that F8.b is, in fact, the preferred form of duration adverbials for many speakers.

In their Spanish, they will say things such as trabajé en tres oras, which is something they cannot have learnt from Spanish

monolinguals.

Táantik u -e', ba'le' mika'h (u) séegir u .

`It has just -ed, but it will go on to .'

F9. Terminativity test frame

"Ku " u k'aat ya'le' "ku chúunul u ".

`"It -s" means "it starts to "'.

F10. Ingressivity test frame

Ts'o'k u ; le méetik táan u beòoraha'.

`He -ed it; therefore it is -ing now.'

F11. Ingressivity test frame

English and Spanish. It does not seem to work for the dialect of YM that I investigated. Most

informants saw no difference between versions a and b, one even claimed never to use F8.a. Only

one informant could be trained to make the difference the Spanish way, which is obviously

methodologically unsatisfactory.

For this reason, a different, though admittedly less elegant test was

employed to distinguish terminative from durative predicates, F9.

The appropriate forms of the verb in question have to be inserted in the slots. If the resulting sentence

is fine, the verb may be atelic (e.g. báaxal `play'); if not, the verb is terminative (e.g. xup' `use up').

This test must be applied with care. If a transitive verb is tested, its object must be definite, because

if there is an indefinite quantity of objects to be treated, most processes can be iterated indefinitely.

Consequently, if F9 with a transitive verb is fine, it has to be made sure that the same object is being

referred to in both clauses and that no repetition of the process (after meanwhile undoing its result)

is being referred to. Only if the very same process continues in the second clause can the verb be taken

to be atelic.

An ingressive verb designates the start of a process. There are, accordingly, periphrastic relationships

of the form F10.

Here, there has to be a pair of lexically related verbs, the first ingressive and to be inserted in the first

slot, the second durative and to be inserted in the second slot. If an appropriate sentence of the form

F10 can be found, this is a hint that the first verb is indeed ingressive (e.g. tàab `lighten' – tóok `burn').

F11 does similar service.

Here an ingressive and a durative verb have to be inserted as in F10. Both tests have their

shortcomings. For one thing, care must be taken not to insert the resultative participle of the putative

ingressive verb in the second slot (which would yield an ungrammatical sentence, anyway). A genuine

ingressive verb is either morphologically unrelated to its atelic counterpart or derived from it, not vice

versa. For another, the tests identify only such ingressive verbs which have a durative counterpart;

no similar condition was imposed on the other situation features.

If a process is bounded both at the start and at the end, it is bilateral telic or punctual. First of all,

punctual predicates have to be distinguished from durative ones. Given this alternative, if a predicate

does not allow any specification of duration, like in F8.a, it is punctual (e.g. tus `lie'). If, on the other

Christian Lehmann

hand, a predicate is possible in a frame such as F8 (a or b), then if this means that the event lasted

so much time, the predicate is durative. However, if it means either that the event is repeated during

the time (e.g. ho'p `dig up' in F8.a) or that so much time elapsed before the event took place (e.g. úuch

`happen' in F8.b), then it is an event predicate.

A more intricate problem is how punctual predicates can be distinguished from ingressive predicates.

If sentences of the form F10 or F11 can be found, the verb in question does not designate a punctual

event. On the other hand, if sentences of the form F9.a are possible (on an iterative reading), the verb

in question cannot be ingressive; it can only designate an event, as in óoxp'éel minùutoh síit'nahih

`he hopped for three minutes'.

3.2.4. Role features

The degree of involvement of a participant is the extent to which it is either centrally engaged in the

happening or, on the contrary, marginal and inessential to it. In YM as in any other language, clause

structure mirrors the degree of involvement of participants. That is, in general central participants have

no case markers (as in YM) or only grammatical ones, peripheral participants are adjoined by

adpositions or similar devices (prepositions in YM). Central participants are represented by actants

(constituents governed by verbal valency) and also in the personal affixes of the verb, peripheral ones

not. Moreover, the grammatical preposition ti' signals a more central involvement than the more

concrete prepositions such as áanal `under'.

The problems of determining the degree of involvement of a participant are familiar from valency

grammar. Here it will be assumed that a constituent is governed through verbal valency if its

grammatical properties, including its mere presence, are determined by the verb. This usually leaves

no doubt about beneficiaries, instruments, time specifications and the like; these are normally

peripheral participants. Well-known problems arise with specifications of locations (essive, ablative,

allative, perlative) in connection with local situation, motion and transport predicates, such as kah

`live, dwell', hóok' `leave, go out', ts'a' `put'. However, these are usually represented by NPs which

are either obligatory or constructed directly (without an intervening preposition), so that they are

covered by the above criteria. When they are neither, as with síit' teh tùunicho' `hop onto the stone',

they will not be considered actants.

Just as optionality of an actant representing a participant does not change the number of participants,

incorporation vs. exteriorization of an actant with respect to the verb also does not change it. Thus,

kin t'okik (tr.) bu'l and kin t'ok bu'l (intr.) `I harvest beans', although having two and one actants,

respectively, both have two participants. The case is different for such intransitive counterparts of

transitive verbs which have no incorporated actant, in pairs such as háan (v/itr) and háan-t (v/tr), both

`eat'. The onomasiological framework presupposed in a study such as this must be fine enough, i.e.

sufficiently tuned to distinctions made in languages, to enable us to say that the first has only one

participant.

For a participant to control a situation means for it to have the power to let the situation realize or

not. This presupposes that the participant may be animate. It will, in fact, generally be animate. The

lifeless forces of the kind of the lightnings and falling stones that kill people are conceived as

controllers in analogy with animate beings. Cf. also §3.2.2 above on the interpretation of such

participant features. Moreover, the situation has to be dynamic. Properties allow of no control, states

only marginally.

One of the tests of control is the imperative test, already mentioned in §3.2.1. It consists in inserting

an intransitive verb in the frame F12.a, and a transitive one, in F12.b.

Predicate classes in Yucatec Maya

The contradictory results yielded for kul by F12 and F13 vs. F14 are typical of posture verbs; cf. §6.4.

a. -én! ` !'

b. -eh! ` it!'

F12. Imperative test frame

Tu path u báah u . `He dared/tried to .'

F13. Control test frame

--a'n. `It is -ed.'

F14. Controlledness test frame

Hach chich u -ah. `He -ed it very hard.'

F15. Affectedness test frame

If the resulting sentence can designate a true command, the participant represented by the subject has

control in the situation.

Another test is usually more reliable than the imperative test, and this is the frame `John hesitated

to .' However, YM does not seem to have the verb `hesitate', so that this test is inapplicable.

However, F13 is similar.

As in F12, if the resulting sentence makes sense, the verb imputes control to its subject (e.g. kul `sit

(down)'), otherwise not (e.g. ah `wake up'). In particular for the latter verb, F13 is clearly more reliable

than F12.

A participant is controlled by a situation if this happens to it, rather than the participant bringing it

about. For a subject, failure on the tests F13 or F12 usually implies that the participant represented

by it is at least mildly controlled by the situation. For intransitive verbs, the resultative participle, i.e.,

F14, can serve as a test frame.

If the form is possible (as, e.g., with ah `wake up', kul `sit'), this means that the participant is more

controlled by, than controller of, the situation.

With transitive verbs, the subject usually represents

the actor, the object the undergoer. Cases such as mùuk'yah-t (v/tr)`suffer, stand' are no exceptions:

application of F12 and F13 shows that the subject does retain some control. Therefore, the question

of the degree of being controlled generally arises for the direct object.

The first distinction to be made here is the one between affected and effected objects. For instance,

the predicates xik `crack, split' and tsíik `comb' have an affected undergoer; mèet `make' and a'l `say'

have an effected undergoer. Verbs such as chuy `sew' are ambiguous between an affective and an

effective reading.

Among affected undergoers, there is a scale reaching from very weak mental affectedness (e.g. k'áat

`ask') over stronger mental affectedness (tus `lie, deceive'), mild physical affectedness (méek'

`embrace') down to strong physical affectedness (ch'ak `hack'). While these degrees are as yet largely

a matter of intuition, F15 is at least an example of the sort of test that may be applied.

A transitive verb has to be inserted. The resulting sentence is fine with predicates that strongly affect

their undergoers (e.g. hats' `beat', `hit'), and awkward with those that only weakly control them (e.g.

áan-t `help').

Christian Lehmann

There are quite a few pairs of words that differ only in their selection restriction, such as polok `fat (animal)'

vs. pìim `thick (thing)', or táankelem `juvenile (male)' vs. lóobbayan `juvenile (female)'.

Finally, a useful distinction can be made between total and partial affectedness. Examples are kìin-s

`kill' (total) vs. lòobit `damage' (partial). All the predicates that entail some sort of movement of their

undergoer (e.g. hok `pull out', k'ex `change') affect it totally.

3.3. The classificatory hierarchy

Before we start on the classification, we have to settle on the question of the mutual relations of our

criteria: Do they form a hierarchy or do they intersect to yield a cross-classification; and to the extent

that they do the latter, what is the proper order of the classification? Inside each of the three feature

spaces of situation, participant and role features, the features are hierarchically or conventionally

ordered the way they have been introduced above.

The logical relations among the three feature spaces are as follows: Full control is possible only in

dynamic situations; therefore role features are subordinate to situation features. With respect to

situation and participant features, we observe the following: Predicates are represented by nouns,

adjectives, modals and verbs. These word classes are most directly related to the situation features

of class inclusion, property, state and process/event. The other two kinds of features are more closely

related to the participants and to the devices connecting these with the predicate. They are reflected

in the predicate itself only in the form of subclasses inside the grammatical classes mentioned.

Therefore situation features should occupy the highest level in the hierarchy.

A related observation lends additional weight to this one. While there are many word-formation

processes in predicates that change situation features, just as there are others that change role features,

there are hardly any that change participant features. Apparently the various kinds of participants are

cognitively so unbridgeably different that two situations that differ only in a participant property are

either not conceivable or are related in so idiosyncratic a way that there can be no process of

word-formation converting one into the other.

This appears to argue that participant features are

closer to the lower level of classification evinced by lexical structure itself.

We will therefore start with the feature space that determines the most far-reaching grammatical

divisions, viz. situation features, and then come in turn to role and participant features.

The above decisions have been devised so as to minimize class overlap. What is commonly connected

by lexical or grammatical processes will be classed together in one class or two neighbouring low-level

classes; a difference that leads to the selection of an altogether different predicate has a high position

in the precedence hierarchy of classificatory features.

4. Classification of Yucatec Maya predicates

The following classification of YM predicates will be detailed in each of the subclasses to the extent

that there are relevant words. Wherever possible, each class will be exemplified by root words. Derived

words will be added for the sake of completeness. Morpheme boundaries relevant to subsequent

discussion will be indicated. Beyond this, no special attention will be paid to derivational relations

in the present section, as they will be dealt with systematically in §6. The sample vocabulary

enumerated below has been gathered randomly; no particular measures have been taken to secure its

representativeness. Finally, it should be clear that the glosses cannot in all cases reflect the semantic

characteristics of the YM words properly.

Predicate classes in Yucatec Maya

4.1. Class inclusion

4.1.1. One participant

ko'lel (n) `woman'

che' (n) `wood'

táanxel (adj) `foreign'.

4.1.2. Two participants

kìik (n/rel) `elder sister'

chúumuk (n/rel) `center'

tséel (n/rel) `side'.

4.1.3. Summary

Just as in any other language, most predicates that designate class inclusion in YM are nouns. There

is no need for extensive exemplification. However, some of these words are adjectives.

4.2. Properties

4.2.1.One participant

4.2.1.1. Human

táankelem `youthful (boy)'

lóobbayan `youthful (girl)'

ayik'al `rich'

óotsil `poor'

tòot `mute'

ts'u't `selfish, avaricious'.

4.2.1.2. Animate

mehen `small, young'

sùuk `tame'

ki'chpam `pretty'

mùuk'náal `strong'

ch'óop `blind'

kóok `deaf'

móoch `lame'.

4.2.1.3. Thing

ya'b `much, many'

nohoch `big'

chàan `small'

chichan `small'

ts'úuts' `tiny'

chowak `long'

kóom `short'

Christian Lehmann

ka'nal `high, tall'

kàabal `low'

pìim `thick, big, voluminous'

tòoh `straight'

wóolis `round'

hobon `hollow'

túulis `entire, whole, complete'

polok strong, fat,

bek'ech `thin'

àal `heavy'

sáal `light'

úuchben `old'

túumben `new'

uts `good'

k'àas `bad'

ma'lòob `good'

lòob `bad'

hats'uts `nice'

ko'h `expensive'

sak `white'

bòox `black, brown'

chak `red'

ya'x `green'

k'an `yellow'

ch'óoh `blue'

k'a'n `loud'

chich `hard'

ts'u'y `hard'

ch'óoch' `salty'

ki' `tasty'

páap `piquant, spicy hot'

su'ts' `sour'

k'áah `bitter'

ch'uhuk `sweet'.

4.2.1.4. Place

nu't' `narrow'

kóoch `wide'

táax `even, smooth, flat'

k'o'lamak `rocky'

tàam `deep, profound'.

4.2.1.5. Proposition

hàah (adj) `true'

Predicate classes in Yucatec Maya

unah (mod) `should'

k'a'náan (mod) `necessary'

talam (adj) `difficult'

xàan (adj) `slow'

séeb(an) (adj) `fast'.

4.2.1.6. Summary

Except for unah and k'a'náan, which are modals, these are all adjectives.

4.2.2. Two participants

4.2.2.1. Animate

4.2.2.2. Thing

bèey (adj) `such, similar'

hel (adj) `other'

hela'n (adj) `different'

yàanal (adj) `other'

-íilih (suffix) `same'

ti'bil (adj) `fit'.

These are mostly adjectives.

4.2.2.3. Place/proposition

4.3. States

4.3.1. One participant

4.3.1.1. Human

su'lak (adj) `ashamed'

ki'mak (adj) `glad'

náay (adj) `in a good mood'

k'ùux `angry'

ts'íik `furious, angry'.

These are adjectives. They are not directly attributable to humans, but are restricted to óol `soul'. Thus,

instead of `NP is glad', YM has ki'mak u yóol NP `NP's soul is glad'.

4.3.1.2. Animate

wi'h `hungry'

uk'ah `thirsty'

k'oha'n `sick'

k'íilkab `warm, sweaty'

kimen `dead'.

Christian Lehmann

These are all adjectives. A set of predicates that one might expect in this group, namely those of bodily

positions, are basically ingressive in YM and therefore treated in §4.6.1.1.1.

4.3.1.3. Thing

ke'x `where'

he'l `voici/-la'

yanil, yanik `situated; in a position/condition'

biha'n `gone'

hùun `alone'

chup `full'

chúumuk `half'

kul `turned'

chíin `upside down'

chokol `hot, warm'

ke'l `cool'

síis `cold'

k'íin `tepid'

ch'úul `wet'

tikin `dry'

sáas `light'

chíik-a'n `light, clear'

éek' `dark, dirty'

e'hoch'e'n `dark'

áak' `fresh'

yih `ripe'

tak'an `ripe, done'

mùun `tender, unripe'

pàah `spoilt (food)'

pa'x `leaky'

yàalab `remaining'.

These are adjectives, although some of them can only be used predicatively.

4.3.1.4. Place

háanil `clean, free of herbs'.

This an adjective.

4.3.1.5. Proposition

páah (v/itr -tal) `be possible'

bey (v/itr -tal) `be possible'

sùuk (mod) `be accustomed'.

While sùuk is a modal, the other two are intransitive verbs.

4.3.2. Two participants

Predicate classes in Yucatec Maya

4.3.2.1. Human - proposition

òoksah-óol-t `believe'

tukul-t `believe'.

There appear to be no roots in this group.

4.3.2.2. Animate – animate

yáamah (n/rel) `love'

p'éek (n/rel) `hate'.

These are relational nouns. Yáamah is a perfect participle of a verb which is no longer in use. Complex

predicates include ts'o'ka'n u bèel (yéetel) `married (to)'; cf. §4.7.2.1.1.

4.3.2.3. Animate - thing

k'áat (n/rel) `wish'

k'ahóol (n/rel) `know'

sahak (ti') (adj) `afraid (of)'.

4.3.2.4. Animate - place

4.3.2.5. Animate - proposition

ohel (n/rel) `know'

k'áat (n/rel) `wish'

tàak (mod) `eager, want'

sahak (ti') (adj) `afraid (of)'

bíinèet (v/tr) `lack'.

This group is mixed, as indicated for each word. The last verb takes what needs to be done as a subject

and the being needing it as an indirect object. Its morphological transitivity is syntactically spurious.

4.3.2.6. Thing – thing

ti'a'l (n/rel) `belong'

bíinèet (v/tr) `lack'

yàan (ti') (mod) `have'

mina'n (ti') (mod) `lack'.

4.3.2.7. Thing - place

yàan (ti') (mod) `exist, contain'

mina'n (ti') (mod) `not exist'

p'áat (v/itr -Vl) `stay'

nàats' (ti') (adj) `near'

náach (ti') (adj) `far'.

4.3.2.8. Summary

Christian Lehmann

All of the major word classes are represented in §4.3.2; even most of the subgroups are mixed in

themselves.

4.3.3. Three participants

One predicate that one might expect here, `owe', has ingressive aktionsart; s. §4.6.3.

4.4. Durative processes

4.4.1. One participant

4.4.1.1. Control

4.4.1.1.1. Human

xíimbal `walk'

óok'ot `dance'

meyah `work'

báaxal `play'

che'h `laugh'

pak'ach `make tortillas'

páak `weed'

ts'íib `write'

ts'ùu'ts' `smoke'.

These are all intransitive verbs of the -0 suffix class. Several of them have an inner effected undergoer,

which does not count as a participant.

4.4.1.1.2. Animate

péek (v/itr -0) `move'

áalkab (v/itr -0) `run'

hu'k' (v/itr -0) `glide on one's backside'

báab (v/itr -0) `swim'

xiknal (v/itr -0) `fly'

ho'p (v/itr -0) `burrow'

pa's (v/itr -0) `scrape'.

These are, without exception, intransitive verbs of the -0 suffix class.

4.4.1.2. Non-control

4.4.1.2.1. Human

náay (v/itr -0) `dream'

wayáak' (v/itr -0) `dream'

ts'íikil (v/itr -0) `be annoyed'.

These are all intransitive verbs of the -0 suffix class.

4.4.1.2.2. Animate

wen (v/itr -Vl) `sleep'

Predicate classes in Yucatec Maya

kux (v/itr -tal) live'

háak (v/itr -Vl) `slide'

òok'ol (v/itr -0) `cry, weep'.

These are intransitive verbs of diverse classes.

4.4.1.2.3. Thing

máan (v/itr -0) `pass'

ki'ki-lankil (v/itr -0) `tremble'

kíilbal (v/itr -0) `thunder'

húum (v/itr -0) `roar'

k'i'nam (v/itr -0) `hurt'

k'áax (v/itr -Vl) `fall, rain'

tùul (v/itr -0) `run over (liquid)'

níik (v/itr -Vl) `ripple, purl'

báan (v/itr -Vl) `drizzle'

léem-bal (v/itr -0) `shine'

ch'íih (v/itr -Vl) `grow'

lòol-lankil (v/itr -0) `flourish'.

These are all intransitive verbs. Most of them are either of the -Vl class or contain another suffix

ending in -Vl.

4.4.1.2.4. Place

4.4.1.2.5. Proposition

t'íil (v/itr -Vl) `last'

sáan (v/itr -tal) `last, take a while'.

4.4.2. Two participants

4.4.2.1. Control

4.4.2.1.1. Human – animate

xahche'-t `comb'

ts'iik-il-t `be angry at'

ki'ki't'àan-t `praise'

tsik `respect'.

These are transitive verbs.

4.4.2.1.2. Human - thing

balak'-t `roll'

yóot' `massage'

yóot'k'a'-t `press'

le'p' `scratch'

ho'ch `shave'

Christian Lehmann

meyah-t `work at, process'

báax-t `play'

lap' `hold in one's fist'

us-t `blow'

kanáan-t `care'

kax(an)-t `search'

pàakat `look at'

cha'n-t `watch'

pa't `wait for, expect'.

All of these are transitive verbs.

4.4.2.1.3. Human - proposition

tukul (v/tr) `think'

pa't (v/tr) `expect'.

These are transitive verbs.

4.4.2.1.4. Animate – animate

pay `avoid; bullfight'

ts'ak `treat'

meyah-t `serve'

tséen-t `feed'

áan-t `help'

tsay-pàach-t `follow'

t'àan `speak, call'

àawat `call, cry'.

These are all transitive verbs.

4.4.2.1.5. Animate - thing

il `see, look'

u'y `hear, listen, feel, smell'

úutsben-t `smell at, sniff'

mùuk'yah-t `tolerate, stand'

kóol `pull'

tul `push'

tíit `shake'

la'ch `scratch'

k'áak'-t `barbecue'

ta' (v/itr -0) `shit'

wíix (v/itr -0) `urinate'.

These are mostly transitive verbs. The class contains the perception verbs, which allow potential

control. YM is among the languages that distinguishes only two basic perception predicates, `see' and

all the rest (cf. Viberg 1984). The two intransitive verbs take effected objects.

Predicate classes in Yucatec Maya

4.4.2.1.6. Animate - place

p'áat (v/itr -Vl) `stay'

kah (v/itr -tal) `live, dwell'.

These are intransitive verbs.

4.4.2.1.7. Animate - proposition

kan (v/tr) `learn' (also terminative)

pat (v/refl) `dare, try'.

4.4.2.2. Non-control

All durative situations involving an animate participant appear to be controlled by this.

4.4.2.2.1. Thing – thing

ch'ul (v/tr) `wetten'.

4.4.3. Three participants

4.4.3.1. Human – animate – thing

e's (v/tr) `show'.

4.4.3.2. Human – human – proposition

tsol (v/tr) `explain'

tsikbal-t (v/tr) `tell'.

4.4.3.3. Summary

All three-participant durative situations inolve a human participant who has control. They are all

symbolized by transitive verbs.

4.5. Terminative processes

4.5.1. One participant

4.5.1.1. Control

hàan (v/itr -Vl) `eat'

míis (v/itr -0) `sweep'

múuch' (v/itr -Vl) `gather'.

These are intransitive verbs. The first and possibly also the second has an internal second participant.

4.5.1.2. Non-control

4.5.1.2.1. Animate

kim (v/itr -Vl) `die'

Christian Lehmann

ka'n (v/itr -Vl) `tire'

ts'óoy (v/itr -Vl) `get weak'.

4.5.1.3. Thing

tàal (v/itr/irr) `come'

chíik-pah (v/itr -Vl) `appear, come out'

èel (v/itr -Vl) `burn'

lòok (v/itr -0) `boil'

tah (v/itr -Vl) `ripen'

nun-tah (v/itr -Vl) `overripen'

ts'úum (v/itr -Vl) `get slack'.

4.5.1.4. Proposition

ts'o'k (v/itr -Vl) `end'

xúul (v/itr -Vl) `cease'.

4.5.1.5. Summary

With the exceptions of lòok and the irregular tàal, all the words in §4.4.1 are intransitive verbs of the

-Vl or -tal classes.

4.5.2. Two participants

4.5.2.1. Control

4.5.2.1.1. Human - animate

ts'ak `cure'

ch'úuk-t `spy'.

4.5.2.1.2. Human - thing

bah `nail, hammer'

xot `cut, saw'

k'os `cut short'

kol `clear (bush)'

póol `trim'

p'ay `chop'

súus `carve, peel'

ts'íil `peel'

sap' `evaporate'

tóok `burn'

chak `cook'

tsah `fry'

wol `round'

yéets' `squeeze'

k'it `strew'

tos `strew'

Predicate classes in Yucatec Maya

t'oh `pour'

pak' `sow'

t'óot' `sow (into a bed)'

hoch `harvest'

t'ok `pick'

chuy `sew'

k'ax `tie, tighten'

to' `wrap'

p'o' `wash'

bon `paint, tint'

p'íis `measure'

xok `count, read'

ts'íib-t `write'

k'ay `sing'

pax `play (music)'.

These are all transitive verbs.

4.5.2.1.3. Human - place

em (ti') (v/itr -Vl) `go down (from)'

míis-t (v/tr) `sweep'.

4.5.2.1.4. Animate – animate

chuk-pàach-t `catch up with, reach'.

4.5.2.1.5. Animate - thing

mèet/bèet `make'

hob `undo, take apart'

hut `destroy'

hàan-t `eat'

uk' `drink'

lak `take off'

ch'a' `fetch'

wek `spill'

xup' `use up'

chup `fill'

ba'pàach-t `surround'.

These are transitive verbs.

4.5.2.1.6. Animate - place

bin (v/itr/irr) `go (to a place)'

náats' (v/itr -Vl) `approach'

na'k (v/itr -Vl) `climb'.

This group is mixed as indicated.

Christian Lehmann

4.5.2.1.7. Animate - proposition

kan (v/tr) `learn' (also durative)

xul (v/tr) `end'.

4.5.2.2. Non-control

4.5.2.2.1. Animate - thing

chuk-pah (v/itr -Vl) `become complete, sufficient for'.

4.5.2.2.2. Animate - proposition

náay-t (v/tr) `dream'

wayáak'-t (v/tr) `dream'

ka'n (v/itr -Vl) `tire'.

4.5.2.2.3. Thing – thing

4.5.3. Three participants

4.5.3.1. Animate – thing – place

nats' (v/tr) `bring close'.

4.6. Ingressive process

4.6.1. One participant

4.6.1.1. Control

4.6.1.1.1. Animate

wa'l (v/itr -tal) `stand up'

kul (v/itr -tal) `sit down'

chil (v/itr -tal) `lie down'

pek (v/itr -tal) `lie down'

xol (v/itr -tal) `kneel'

p'uk (v/itr -tal) `squat'

t'ùuch (v/itr -tal) `squat'

mòot' (v/itr -tal) `cower'

xak (v/itr -tal) `walk on four legs'

tsel (v/itr -tal) `lie on one's side'

nok (v/itr -tal) `lie face down'

haw (v/itr -tal) `lie on one's back'

líik' (v/itr -Vl) `get up'

k'uch (v/itr -Vl) `arrive'

he'l (v/itr -Vl) `stop, rest'.

This group contains the verbs of body position or so-called posture verbs. They will be treated in detail

in §6.4.

4.6.1.2. Non-control

Predicate classes in Yucatec Maya

4.6.1.2.1. Animate

síih (v/itr -Vl) `be born'

ah (v/itr -Vl) `wake up'

léek (v/itr -Vl) `have an attack'.

4.6.1.2.2. Thing

wáach' (v/itr -Vl) `burst into blossom'

yàan (v/itr -tal) `evolve'.

4.6.1.2.3. Place

4.6.1.2.4. Proposition

káah (v/itr -Vl) `begin'

chúun (v/itr -Vl) `start'

chúum-pah (v/itr -Vl) `start over'

hóop' (v/itr -Vl) `set in step by step'

k'am (v/itr -Vl) `show effect'.

4.6.1.3. Summary

All the words in §4.6.1 are intransitive verbs of the -Vl or the -tal class.

4.6.2. Two participants

4.6.2.1. Control

4.6.2.1.1. Animate – animate

4.6.2.1.2. Animate – thing

mach `seize, hold'

náach `seize/hold with one's teeth'

ti'al-t `acquire'

ti'l-int `acquire'

nahal-t `earn'

p'at `leave, abandon'

tàab `light'.

These are all transitive verbs.

4.6.2.1.3. Animate – place

òok (v/itr -Vl) `enter'.

Verbs of body position are considered one-participant predicates in consonance with §3.2.4. and have

therefore been treated above in §4.6.1.1.1.

4.6.2.1.4. Animate – proposition

Christian Lehmann

chun (v/tr) `start'

kah (v/tr) `start'

na't (v/tr) `understand, guess'.

These are all transitive verbs.

4.6.2.2. Non-control

4.6.2.2.1. Animate - thing

k'am `get'

k'ahóol-t `get to know'.

4.6.2.2.2. Animate - proposition

ohel-t `get to know'.

4.6.2.2.3. Summary

These are transitive verbs.

4.6.3. Three participants

mahan-t `lend'

p'ax `borrow, get in debt'.

These are transitive verbs with a controlling human being, a receiver and a thing as participants.

4.7. Event

4.7.1. One participant

4.7.1.1. Control

e'l (v/itr -0) `put eggs'

pankal (v/itr -0) `fling out'

síit' (v/itr -0) `hop'.

These are intransitive verbs of the -0 suffix subclass. The first has a non-exteriorized effected

undergoer.

4.7.1.2. Non-control

4.7.1.2.1. Animate

sa'sak (v/itr -Vl) `cough'

he'síin (v/itr -0) `sneeze'

hàayab (v/itr -0) `yawn'

háak' (v/itr -Vl) `get frightened'

t'i'l (v/itr -Vl) `get stuck'.

4.7.1.2.2. Thing

lemléem (v/itr -0) `lighten'

wáak' (v/itr -Vl) `bang'

Predicate classes in Yucatec Maya

xíik (v/itr -Vl) `burst'

káach (v/itr -Vl) `break'

pa'x (v/itr -Vl) `break'

síit' (v/itr -0) `spring'

lúub (v/itr -Vl) `fall'

ch'úuk (v/itr -Vl) `slide down'

ch'áah (v/itr -Vl) `drop'

ts'óop (v/itr -Vl) `sink (into mud)'

búul (v/itr -Vl) `sink'

k'áal (v/itr -Vl) `shut'

múuts' (v/itr -Vl) `shut oneself, wither'

sùut (v/itr -0) `turn round, return'.

4.7.1.2.3. Place

4.7.1.2.4. Proposition

úuch (v/itr -Vl) `happen'

sùut (v/itr -0) `happen/do again'

ch'éen (v/itr -Vl) `cease, stop'.

4.7.1.3. Summary

All the words in §4.7.1.2 are intransitive verbs, almost all of them of the -Vl class.

4.7.2. Two participants

4.7.2.1. Control

4.7.2.1.1. Human – human

lox (v/tr) `box'

tus (v/tr) `lie to, cheat'

ts'o'k (v/itr -Vl) poss. bèel (yéetel) `marry'.

4.7.2.1.2. Human – animate

ts'on `shoot'

ts'ol `hit'

méek' `embrace'.

These are transitive verbs.

4.7.2.1.3. Human – thing

pul `throw'

xab `(trans-)plant'

òokol `steal'

hel `replace'

k'ex `change'

Christian Lehmann

bul `gamble'

kuch `load (on one's back)'

k'óoch `load on head or shoulders'

walk'es `turn over'

sut `turn'

wak `explode'

xik `blow up, burst open'

ch'ak `fell'

hak `pull down'

ho'p `dig out'

lóoch' `scoop'

ch'in `throw at'

k'ol `hit'

ts'ay `hit'

ts'ol `strike'

boh `knock'

max `pinch'

lah `slap'

lèep' `scratch with thumb-nail'

loch `bend'

p'oh `bend'

buts' `bend'

nah-k'ab-t `touch (with hand)'

nah-che'-t `touch (with foot)'

hay `extend, display'

t'it' `spread'

ta'k `hide'

k'al `shut, close'

pix `cover'

mak `cover'

hok' `cinch up'

k'at `block'

tup' `turn off'

hets' `calm down'

cha' `let loose'.

These are all transitive verbs.

4.7.2.1.4. Human – proposition

chuk `grasp, find out'

cha' `let'

óot `agree, comply'

ch'en `stop'.

These are transitive verbs.

Predicate classes in Yucatec Maya

4.7.2.1.5. Animate – animate

chi' `bite'

hats' `beat, hit'.

These are transitive verbs.

4.7.2.1.6. Animate – thing

ch'a' `take, fetch'

huts' `displace'

lak `remove'

lik' `lift'

ch'uy-kint `lift, hang up'

chuk `catch'

kach `break'

pa' `break (compact object)'

p'ik `crack'

óop' `smash'

hok `tear out'

ho'p `dig out'

he' `open'

hap' `open (mouth)'

p'il `open (eyes)'

leb `take apart'

luk' `swallow'

bal `conceal'

muts' `close (eyes)'.

These are transitive verbs.

4.7.2.1.7. Animate – place

tàal (v/itr/irr) `come (from)'

u'l (ti') (v/itr -Vl) `come (to)'

luk' (ti') (v/itr -Vl) `go off'

hóok' (ti') (v/itr -Vl) `go out (from)'.

4.7.2.2. Non-control

4.7.2.2.1. Human - thing

xeh (v/tr) `vomit'

k'a'h (v/itr -Vl) `occur, remember'

tu'b (v/itr -Vl) `forget'.

These are transitive and intransitive verbs as indicated.

4.7.2.2.1.1. Human - proposition

k'a'h (v/itr -Vl) `occur, remember'

Christian Lehmann

tu'b (v/itr -Vl) `forget'.

These are intransitive verbs.

4.7.2.3. Animate - thing

kax-t `find'

sat `lose'

lòobit `hurt'.

These are transitive verbs.

4.7.2.3.1. Animate – place

háak-che'-t (v/tr) `slide, slip'.

4.7.2.3.2. Thing - thing

sùut (v/itr -0) `become, turn into'.

4.7.3. Three participants

4.7.3.1. Human – human – thing

ts'a' `give'

k'ub `hand in'

síih `present'

sut `give back'

man `buy'

kon `sell'

bo'l-t `pay'.

4.7.3.2. Human – human – proposition

a'l `say'

k'áat `ask'

núuk `answer'.

4.7.3.3. Human – thing – thing

sut `turn into'

nak-kunt `lean against'

but' `put into'

ch'ik `stick into'.

4.7.3.4. Human – thing – place

ts'a' `put'

púut `unload'

túux-t `send'.

4.7.3.5. Summary

Predicate classes in Yucatec Maya

All the words in §4.7.3 are transitive verbs with a human controlling participant.

5. Grammatical properties of Yucatec Maya predicate classes

5.1. Preliminaries

It was said in §2 that four word classes of YM come under the heading of predicate in the sense of

this study, viz. nouns, verbs, modals and adjectives. The verbs in turn can be subdivided into transitive

and intransitive. These are definable as distribution classes, i.e. they are distinguished by their structural

behavior. Since the criteria of the above classification of predicates are semantic in nature, it would

be logically possible for the word classes to exhibit a stochastic distribution over the predicate classes.

On this background, any tendencies of correlation between predicate classes and word classes are

noteworthy.

It is true that many of our test frames presuppose the appurtenance of the members of the predicate

class under examination to a particular grammatical class; insofar, the criteria used were not purely

semantic. However, on the one hand, this is inevitable if semantic intuitions are to be operationalized

by test frames; and on the other hand, the tests have not been used to exclude members of a

grammatical class that does not fit them from the predicate class in question; they have only been used

to differentiate among the members of one word class. Insofar, although the allocation of members

of distinct word classes to one predicate class has not always been made according to explicit criteria,

there has been no circularity in the use of the available criteria.

Below a set of generalizations will be presented which are based on the sample vocabulary appearing

in §4. Although this is not statistically representative, the regularities observed in the sample will be

taken to be indicative of general tendencies.

5.2. Main situation classes

The structural results of the above classification are the following:

Apart from some adjectives such as táanxel `foreign', words designating class inclusion are nouns.

This includes a sizeable group of relational nouns, among them kinship terms. YM here follows the

far majority of languages in the world.

Words designating properties are overwhelmingly adjectives. Only properties of propositions can be

expressed by modals; and one relational property without concrete meaning, `same', is expressed by

a suffix. There are no verbs designating properties, as there are in many other languages.

The group of predicates of state is structurally very heterogeneous. All the word classes considered

are represented. There is even one exceptional transitive verb, bíinèet `lack', signifying a relational

state. However, it may be generalized that one-participant states are represented by adjectives if they

do not relate to propositions, that the entire distributional variation may be found inside the group

of two-participant states, and that nouns are confined to this subgroup, too.

All the dynamic concepts (processes and events) are expressed by verbs. There is not even a modal

in this group. Moreover, while there is a pervasive tendency of intransitive verbs to associate with

one-participant concepts, and of transitive verbs to associate with two- and three-participant concepts,

this tendency is least pronounced in states, and still suffers many exceptions in durative processes,

while it is most purely represented in events. The two pseudo-transitive verbs found, bíinèet and ichint,

designate a state and a durative process.

It is thus very clear that the sequence `property – state – durative process – terminative process –

ingressive event – punctual event' of F2, which is constituted by increasing dynamicity and

Christian Lehmann

It should be clear that this is not necessarily so. Cf. the Latin trivalent verbs of the kind of pudet/ taedet me

`I am ashamed/bored', which designate two-participant states.

time-relativity, is mirrored, on the structural side, by a transition from non-verbal to increasingly, and

finally, exclusively verbal word classes, and that the differential use of transitive and intransitive verbs

correlates most closely with multi- vs. one-participant predicates inside that predicate class which

is most dynamic and therefore most verbal, viz. the events. Furthermore, if a transition area is to be

identified in this gradience, i.e. an area of greatest fuzziness, heterogeneity and cooccurrence of

opposite semantic forces and structural devises, it is the class of state predicates, which alone exhibits

all the word classes involved.

5.3. Role features

The distribution of roots of different morphological classes over the predicate classes is noteworthy.

While many of the more stative multi-participant situations are expressed by verbs that have been

transitivized with the -t suffix, most of the basic transitive roots are found in the most dynamic

predicate classes. There is not a single basic transitive root among stative predicates and relatively

few among durative predicates. This means, evidently, that multi-participant predicates are

prototypically more dynamic, while one-participant predicates are prototypically more stative.

In one-participant predicates, there is a clear association of control properties with intransitive

conjugation classes. If absence of control is to be expressed, either the -Vl class is chosen or, under

circumstances to be discussed in §6, the -tal class, i.e. the other incompletive suffix class. If lack of

control is not to be expressed, the incompletive suffixless class comes in. As was said in §2.2, this

functions as the default class. As a consequence, it does contain all the control (agentive) verbs, but

also many non-control verbs that for some reason were not assigned to either of the incompletive suffix

classes.

Apart from bíinèet and ichint, an actant slot stands for a real participant.

On the other hand, there

are several intransitive verbs designating two-participant concepts. However, this is possible only if

one of the participants is a non-entity, viz. a proposition or a place. Such situations typically evince

a less pronounced control cline. There are some intransitive verbs with two entities as participants.

But they are confined to non-control predicates, i.e. such predicates as k'a'h `remember', tu'b `forget',

which involve less than complete control on the part of one participant, or less than complete

affectedness on the part of the other participant. These predicates, instead, are represented by

intransitive verbs which require the use of a preposition (mostly ti' or éetel) with the second participant.

YM thus confirms the well-known rule (cf. Hopper & Thompson 1980) that the prototypical transitive

verb is one that represents a maximally dynamic event controlled by one participant and completely

affecting the other participant. The prototypical intransitive verb, on the other hand, represents a

durative one-participant predicate.

In YM, transitive verbs are more confined to the typical "transitive" situation, i.e. one with a clear

control cline, than in many other languages. Strong evidence for this generalization is provided by

the phasal predicates, such as `start', `continue', `end'. While in English and other European languages,

these are normally constructed personally (e.g. I start working), this is not so in YM. They are

impersonal intransitive verbs; what one says is ku chúunul in meyah `my working starts'. This means

that the semantic role properties of the syntactic subject are more narrowly defined in YM than in

English. This is in consonance with the ingredient of ergativity to be found in YM grammar.

6. Recategorization operations

Predicate classes in Yucatec Maya

base meaning action noun meaning

xok read xòok study (n)

tus lie tùus lie (n)

k'ay sing k'àay song

pax play pàax music

F16. Examples of action noun derivation

6.1. Introduction

The above predicate classes do not represent an immobile category inventory of the language. They

are related to each other by operations of recategorization, which transfer a given predicate into a

different class. These manifest themselves as lexical and grammatical processes, i.e. processes of

word-formation, of inflectional morphology and of syntax. They produce complex predicates which

belong to one of the existing predicate classes. Thereby they both feed existing classes with further

material and, at the same time, provide the operational basis of these classes. From the methodological

point of view, they provide independent evidence that the predicate classes found by the other criteria

are actually operative in the language.

YM abounds in derivational processes of all sorts. Most of them are in the verbal sphere; even more

are suffixal. There is no question of reviewing them all here. The emphasis is, instead, on such

processes of inflection and word-formation that provide the operational basis of the predicate classes

we have seen before.

We will subdivide the following discussion according to whether relevant operations change situation

or role features. As we saw in §3.3, operations that change participant features in systematic ways

are not to be expected.

6.2. Change of situation features

6.2.1. Complex class inclusion predicates

It would lead us to far to review here the noun-forming derivational morphology of YM. As a feature

of especial importance in the present context, it may be mentioned that action nouns may be derived

from transitive verbs by low tone in the root syllable. Observe the pairs in F16:

By this device, a particular instance may be subsumed under a class of comparable situations, rather

than being predicated as a process or event.

6.2.2. Complex property predicates

There are no special operations converting a state predicate into a property predicate. However, a state

predicate may be used in the frame F6, thus producing, e.g., a phrase such as hùuntúul wi'h máak `a

hungry person'. This would force a property interpretation of the predicate. It does not, however,

involve any special recategorization operation, but merely consists in using a state predicate as a

property predicate.

A number of derivational processes yield property adjectives. They are briefly reviewed in the

following subsections.

Christian Lehmann

This entails that the stem is recategorized as an adjective; cf. Lehmann 1991[R], §4.2.2.2.

Reduplication plays an eminent role in diverse processes of YM word-formation. Formally, it is a unitary

process. Given a root of the phonological structure C

V(C

, where C

can only be /'/, the reduplicated stem has the form

RED-ROOT, where RED is identical to ROOT if C

= 0 and C

is a nasal, and otherwise it is C

V'.

base

verb

meaning potentiality

adjective

meaning

háak (v/itr -Vl) slide háakchalak slippery

hats (v/tr) separate hatschalak separable

k'ax (v/tr) tie k'axchalak tieable

F17. Examples of potentiality adjective derivation

6.2.2.1. Potentiality adjectives

Base: bivalent verb root.

Morphological change: suffixation of -chalak to base.

Syntactic change: actor role and direct object slot of base disappear; subject slot of base is converted

into modifying slot;

undergoer role is mapped on modifying slot.

Semantic change: situation is represented as property of the undergoer which consists in the possibility

of undergoing the process.

F17 shows some examples.

The derivational process is morphologically and semantically related to the extended deagentive

derivation in -chah mentioned in §6.3.1.1.2; cf. Owen 1969:42.

6.2.2.2. Display adjectives

Base: verb root.

Morphological change: base is reduplicated

and suffixed with -kil.

Syntactic change: if base is bivalent verb, actor role and direct object slot disappear; subject slot of

base is converted into modifying slot; undergoer role is mapped on modifying slot.

Semantic change: situation is represented as property of undergoer whereby this exhibits strong

features of undergoing it.

F18 shows some examples.

Predicate classes in Yucatec Maya

base

verb

meaning display

adjective

meaning

háak (v/itr -Vl) slide ha'háahkil slippery

léem (v/itr -bal) shine lemléemkil brilliant

síit' (v/itr -0) hop si'síit'kil elastic

míis (v/itr -0) sweep mi'míiskil cleanly swept

tak' (v/tr) stick, glue ta'ta'kil sticky

F18. Examples of display adjective derivation

base

verb

meaning disposition

adjective

meaning

síit' (itr -0) hop síit'kalak elastic, agile

sats' (tr) extend sats'k'alak extensible, elastic

ts'un press ts'unk'alak soft, yielding

pak (tr) weed pahk'alak easily weeded

F19. Examples of disposition adjective derivation

The reduplication lends vividity and emphasis to the meaning of these adjectives and renders the

semantics of the process slightly irregular (cf. Owen 1969:40 for a different account). Many adjectives

are derived from diverse bases by mere reduplication: k'a'k'àas `eminently bad' from k'àas (adj) `bad',

xi'xib `macho' from xib (n) `man', lo'lòoch `crooked' from loch (v/tr) `bend', etc.

6.2.2.3. Disposition adjectives

Base: verb root.

Morphological change: suffixation of -k'alak to base.

Syntactic change: if base is bivalent verb, actor role and direct object slot disappear; subject slot of

base is converted into modifying slot; undergoer role is mapped on modifying slot.

Semantic change: situation is represented as property of undergoer or only participant which consists

in its disposition or propensity of undergoing situation.

F19 shows some examples.

The derivational process is morphologically and semantically related to the spontaneous derivation

reviewed in §6.3.2.1; cf. Owen 1969:44.

6.2.3. Complex state predicates

Christian Lehmann

I have only scant evidence for the derivation of positional adjectives from transitive roots as described

in Owen 1969:46.

base

verb

meaning resultative

adjective

meaning

ch'íih (itr -Vl) grow ch'íiha'n aged

ka'n (itr -Vl) tire ka'na'n tired

kul (itr -tal) sit kula'n at home

kah (itr -tal) live, dwell kaha'n resident

chuy (tr) sew chuya'n sewn

hets' (tr) fix, calm down hets'an fixed, secure

F20. Examples of resultative derivation

There are no morphological processes converting a property predicate into a state predicate. However,

as explained in §F3, a property predicate may be used in a syntactic frame like F5, which turns the

situation into a state. Using yanik `in a situation/position' as the copula has the same effect, as shown

by pairs such as tòoh `it is straight' vs. tòoh yanik `it is upright'.

All derived words expressing states are adjectives. There are a couple of derivational processes yielding

such adjectives. They are briefly reviewed in the following subsections.

6.2.3.1. Resultative adjectives

Base: any dynamic verb stem except v/itr -0.

Morphological change: base is suffixed with -a'n.

Syntactic change: if base is bivalent verb, actor role and direct object slot disappear; subject slot of

base is converted into modifying slot; undergoer role is mapped on modifying slot.

Semantic change: situation is represented as state of undergoer which results from a terminated more

dynamic situation.

F20 shows some examples.

The derivation of resultative adjectives in -a'n is one of the most regular and productive morphological

processes of YM. If it only applied to transitive stems, one would consider it the formation of a past

passive participle. However, it applies with equal regularity to intransitive verbs, except those of the

incompletive -0 suffix class, which contains all the control verbs. As the examples show, the resulting

form is lexicalized in some cases.

6.2.3.2. Positional adjectives

Base: verb root of v/itr -tal class.

Morphological change: base is suffixed with -Vkbal.

Allomorphy: V is harmonic with the root vowel.

Syntactic change: subject slot of base is converted into modifying slot.

Semantic change: situation is represented as position of undergoer resulting from entrance into

situation.

Predicate classes in Yucatec Maya

I have insufficient evidence for Owen's (1969:44f) claim that transitive verbs also regularly undergo the

derivation and that it has a positional meaning.

base

v/itr -Vl

meaning positional

adjective

meaning

ch'eb become oblique ch'ebekbal oblique

kul sit kulukbal sitting

chil lie chilikbal lying, horizontal

xok kneel xolokbal on the knees

kah live, dwell kahakbal living

F21. Examples of positional adjective derivation

F21 shows some examples.

This adjective is derivable from all of the posture verbs described in §6.4 and other verbs denoting

situations which conceivably put their undergoer into some position (cf. Owen 1969:46).

6.2.4. Complex process predicates

6.2.4.1. Processive verbs

Base: stem of noun, adjective, modal, v/tr class.

Morphological change: stem joins v/itr -tal class.

Allomorphy: in the past, the suffixes -chah and -lah are in complementary distribution for some verbs,

in free variation for others.

Syntactic change: if base is bivalent verb, actor role and direct object slot disappear; for noun bases,

a subject slot is created; for adjective bases, their slot is converted into a subject slot; the only

remaining role is mapped on subject slot.

Semantic change: situation is represented as durative process applying to participant.

F22 shows some examples.

Christian Lehmann

base meaning processive

v/itr -tal

meaning

áak'ab (n) night áak'abtal become night

cha'n (n) spectacle cha'ntal be on display

uts (adj) good utstal improve

ya'b (adj) much ya'btal become much

su'lak (adj) ashamed su'laktal be ashamed

síis (adj) cold síistal get cold

bèey (adj) such bèeytal be possible

yàan (mod) exist yàantal be, become

ch'eb (v/tr) tilt ch'ebtal be oblique

tak' (v/tr) stick, glue tak'tal attach oneself

F22. Examples of processive derivation

This operation is of extreme generality and regularity. It can convert practically any less dynamic

predicate into a process predicate and, therefore, all word-classes except intransitive verbs. As the

examples show, there is also no difference between property (uts, ya'b) and state (su'lak, síis) adjectives

as bases. Derived stems such as e'hoch'e'n `dark' or chúunk'ìin `morning' may as well undergo

processive derivation. As a consequence, its semantics are rather empty. It is essentially a process of

verbalization, which is put to work when one needs to express such categories as tense, aspect and

mood.

6.2.4.2. Durative verbs

Base: adjective or intransitive non-action verb.

Morphological change: stem gets suffix -lankil and joins v/itr -0 class.

Syntactic change: for adjective bases, their slot is converted into a subject slot.

Semantic change: situation is represented as protracted non-terminative process undergone by

participant, typically a vain or unsuccessful movement in the manner designated by the root.

F23 shows some examples.

Predicate classes in Yucatec Maya

base meaning durative

v/itr -lankil

meaning

háak (v/itr -Vl) slide háaklankil move sliding

xol (v/itr -tal) kneel xollankil slide on one's

knees

híil (v/itr -tal) stretch out híillankil creep

núuch (adj) knock-kneed núuchlankil stumble about

knock-kneed

p'ùux (adj) humpbacked p'ùuxlankil walk with drawn-

in head

F23. Examples of durative derivation

role affected

valency change

actor undergoer

detransitivization deagentivization introversion

transitivization agentivization extraversion

F24. Operations of valency change

Bases are typically posture and motion verbs. They may already have undergone the spontaneous

derivation described in 6.3.2.1; e.g. núuchk'alankil `stumble about knock-kneedly and miserably'.

6.3. Change of role features

6.3.1. Verbal valency change

YM grammar is extremely sensitive to role features of participants. It makes a decisive difference for

verbal morphology and syntax whether a verb is transitive or intransitive, and if it is intransitive, what

the role of the actant is. Many of the verbs that have been used as examples §4 are in fact derived so

that their control properties are morphologically explicit. In English and many other languages, it is

possible to use a given verb in different valency frames, as, e.g., the verb sell may be used with both

actor and undergoer specified (Helen sells books), or only with actor (Helen is selling), or only with

undergoer (the books sell well). This is impossible in YM. In reviewing the derivational morphology

for the change of role features, we first limit ourselves to verbal derivation. By valency increase, an

intransitive verb may get an actor or an undergoer added. Similarly, by valency decrease, a transitive

verb may loose its actor or its undergoer slot. The relevant operations are summarized in F24.

6.3.1.1. Deagentivization

6.3.1.1.1. Simple deagentivization

Base: transitive monosyllabic root verb.

Christian Lehmann

base

v/tr

meaning deagentive

v/itr -Vl

meaning

bul submerge búul sink

ch'ul wetten ch'úul get wet

kach break káach break

k'al shut k'áal close oneself

hay extend háay extend oneself

p'at leave p'áat stay

F25. Examples of simple deagentivization

base

v/tr

meaning deagentive

v/itr -Vl

meaning

chun start chúumpah start (intr.)

he' open hée'pah open (intr.)

t'och spit, pick t'óochpah stumble

chuk catch, grasp chúukpah become complete

xa'k'-t mix xáa'k'pah mix oneself

F26. Examples of extended deagentivization

Morphological change: root vowel gets high tone; derived stem joins v/itr -Vl class.

Syntactic change: actor role and direct object slot of base disappear; undergoer role is mapped on

subject slot.

Semantic change: process happens by itself, without agentive control.

F25 shows some examples.

Deagentivization (also called anticausativization) applies most easily to verbs signifying processes

which can conceivably affect their undergoer without the intervention of an actor, which can, so to

speak, happen by themselves. This presupposes, among other things, that no too specific requirements

are made on the part of the actor, esp. not the use of specific instruments. For this reason, for instance,

ho'p `dig out' and xot `cut' are excluded from deagentivization.

6.3.1.1.2. Extended deagentivization

Base: transitive monosyllabic root verb.

Morphological change: root vowel gets high tone, root is suffixed with -pah; derived stem joins v/itr

-Vl class.

Syntactic change: actor role and direct object slot of base disappear; undergoer role is mapped on

subject slot.

Semantic change: process happens by itself, without agentive control.

F26 shows some examples.

Predicate classes in Yucatec Maya

I have derivations in -chah from áak'ab (n) `night', paréeho (adj) `equal' and peóor (adj) `worse', without counterparts in -pah.

base

v/tr

meaning introversive

v/itr -0

meaning

kol clear (wood) kòol clear wood

xok read xòok study

kach break kàach break (things)

chuy sow chùuy sow

k'ay sing k'àay sing

p'at leave p'àat quit

F27. Examples of introversion

The process also applies to some roots which do not directly provide words. As some of the examples

show, the semantic change is sometimes irregular. There seems to be a dialect correspondence between

this process and the derivation in -chah described in Owen 1969:40, although the latter appears to

be more permissive as to base category.

6.3.1.2. Introversion

Base: transitive verb.

Morphological change: root vowel gets low tone; derived stem joins v/itr -0 class.

Allomorphy: some verbs take a -Vl suffix in addition to the tone change, others, including all derived

or polysyllabic stems, take a -bal suffix instead.

Syntactic change: undergoer role and direct object slot of base disappear.

Semantic change: actor acts on unspecified undergoer (the undergoer being the typical one or the action

being habitual).

F27 shows some examples.

Introversion (detransitivization sensu stricto, often called `direct object deletion') applies most easily

to verbs with narrow selection restrictions as to their direct object, as e.g. tak `put on (cloth)'. If the

meaning of a verb is completely open as to the kind of undergoer it may affect, such as cha' `let' or

sat `lose', the action is not easily typified, and the verb does not easily lend itself to introversion.

6.3.1.3. Agentivization

6.3.1.3.1. Causativization

Base: intransitive verb stem, mostly of -Vl suffix class, which is monosyllabic and does not end in

/s/.

Morphological change: root gets -s suffix; derived stem joins v/tr class.

Allomorphy: some verbs suffix -es or -bes instead of -s.

Syntactic change: subject slot with its role is converted into a direct object slot with undergoer role;

subject slot with actor role is added.

Semantic change: situation is controlled by additional participant.

F28 shows some examples (cf. Owen 1969, ch.II.5):

Christian Lehmann

base

v/itr

meaning causative

v/tr

meaning

hóok' (-Vl) go out hó's get sth. out

kim (-Vl) die kins kill

wen (-Vl) sleep wèens put to sleep

ah (-Vl) wake up ahs wake up

kah (-Vl) start kahs start

máan (-0) pass by máans transport

ka'n (-Vl) be learnt ka'ns teach

F28. Examples of causativization

Predicate classes in Yucatec Maya

Often (e.g., in Owen 1969, ch.II.5), no distinction is made between causative and factitive. Derivational

operations are principally goal-determined rather than base determined (i.e., all the stems formed by a given

derivational process share their category, not necessarily the category of their base). Therefore, if factitivization

and causativization were the same derivational operation, differing only in the category of their bases, one

should expect the same formal process in both cases. YM does make a difference here, joining, in this, the

majority of languages.

base meaning factitive

v/tr -kunt

meaning

uts (adj) good utskint improve

su'lak (adj) ashamed su'lakkunt make ashamed

síis (adj) cold síiskunt cool

bèey (adj) such bèeykunt make such

ch'eb (v/tr) tilt ch'ebkunt tilt

hets' (v/tr) calm hets'kunt pacify

F29. Examples of factitivization

This process is severely restricted. Apart from the phonological conditions mentioned, there is a

semantic restriction which excludes action verbs. This has two consequences. First, there are no direct

causatives from transitive verbs. The way of achieving the causativization of a transitive verb is

illustrated by kan `learn': it is first passivized, giving ka'n `be learnt', and from this the causative ka'ns

`teach' may be derived. Second, even intransitive verbs are ineligible if they designate actions rather

than processes. Intransitive verb bases should designate processes which do not require control. This

is why such verbs as bàab `swim', t'àan `speak', e'l `lay eggs' are excluded from causativization.

6.3.1.3.2. Factitivization

Base: adjective.

Morphological change: base gets -kun-t suffix; derived stem joins v/tr class.

Allomorphy: kun after front root vowels (including /a/), kin after back root vowels.

Syntactic change: unique slot with its role is converted into a direct object slot with undergoer role;

subject slot with actor role is added.

Semantic change: situation is controlled by additional participant.

F29 shows some examples (cf. Owen 1969, ch.II.5):

This process is a close counterpart to processive derivation; most bases that can undergo the latter

also can undergo factitivization. There is only a semantic restriction demanding that the process can

conceivably be brought about by outside impact. This is why formations such as pàahkunt `render

moldy' are improbable. The factitivization of transitive bases remains to be investigated. It is possible

that it presupposes their deagentivization. That is, hets'kunt may really be héets'kunt. Otherwise,

hets'kunt and hets' (v/tr) should be synonymous, and one would not see the raison d'être of the former.

6.3.1.4. Extraversion

Base: intransitive verb, mostly of -0 suffix class.

Christian Lehmann

Cf. Owen 1969:33-36 for a different account.

base

v/itr

meaning extraversive

v/tr

meaning

xíimbal walk xíimbat visit

hàan (-Vl) eat hàant eat

tsikbal chat tsikbat tell

páak remove weed páakt remove (weed)

pak'ach make tortilla pak'acht make (tortilla)

ts'íib write ts'íibt write

F30. Examples of extraversion

Morphological change: stem gets -t suffix; derived stem joins v/tr class.

Syntactic change: direct object slot with undergoer role is added.

Semantic change: action extends to undergoer and affects it.

F30 shows some examples.

Intransitive verb bases must designate actions, i.e. they must have an actor role. This is why such verbs

as ah `wake up' or wen `sleep' are excluded from extraversion. Apart from this, the -t suffix forms

transitive verbs affecting their undergoer on a larger scale. Bases need not be intransitive verbs; they

may also be nouns. Compound transitive verbs invariably take this suffix, whatever their composition.

Spanish loans which are used as transitive verbs also have it. It is, thus, a sign of transitivity.

6.3.1.5. Summary

Apart from the differences in productivity, the four valency-changing operations are maximally

symmetric, as is already implied by F24. This applies both to their grammatical and semantic input

restrictions and to their grammatical and semantic effects. In particular, the opposition between inactive

-Vl verbs and active -0 verbs is fed by the detransitivization operations and is mirrored by the

transitivization operations, since agentivization, which adds an actor, prefers inactive input – mostly

-Vl class verbs –, whereas extraversion, which adds an undergoer, prefers active input – mostly -0

class verbs.

Since these operations are so regular and largely in a mirror-image relation, there is hardly any

cumulation of them possible. Relevant examples involve some kind of meaning specialization. From

lech (v/tr) `hang', the introversive lèech `fish (hang up the typical undergoer)' may be derived, and

this may in turn be extroverted to lèech-t `fish (some object which may or may not be a fish)'.

6.3.2. Other role changes

6.3.2.1.Spontaneous verbs

Base: monosyllabic verb or adjective roots.

Predicate classes in Yucatec Maya

base meaning spontaneous

v/itr -Vl

meaning

he' (v/tr) open he'k'ah open oneself

kach (v/tr) break kachk'ah break

sùut (v/itr -0) return sùutk'ah return

unexpectedly

kul (v/itr -tal) sit kulk'ah get into a sitting

position

F31. Examples of spontaneous derivation

Morphological change: base gets -k'ah suffix; derived stem joins v/itr -Vl class.

Allomorphy: there may be some allomorphy between -k'ah, -pah, -chah and -tal.

Syntactic change: in a transitive verb base, actor role and direct object slot disappear; undergoer role

is mapped on subject slot.

Semantic change: situation is represented as process that happens spontaneously, without the

intervention of a controller, potentially even against his volition.

F31 shows some examples.

A base is eligible for spontaneous derivation if the situation designated involves a non-controlling

participant and can conceivably happen spontaneously. Consequently, such verbs as bah `hammer'

or xot `cut' will not undergo this process.

The difference between spontaneous and deagentive derivation resides in various factors. As to type

of base, deagentivization only applies to transitive verbs, while spontaneous derivation applies to verbs

in general. On the other hand, the semantic conditions on the base are not so strict in the case of

deagentivization since the component `quite by itself/on its own' is not present. In some cases,

however, the processes may be in free variation. The exact relationship between deagentivization,

processive and spontaneous derivation in YM remains to be investigated.

6.4. Posture verbs

Body positions are concepts such as `stand', `lie', `sit'. Situations constituted by them may be states

(e.g. `be sitting') or ingressive events (`sit down'). Predicates designating body positions are called

positionals. Verbs representing them are posture verbs.

For the morphological constitution of positionals, either the state or the ingressive event may be taken

as basic. If positionals are basically states, corresponding events may be derived by an ingressive

aktionsart operation, as in English sit - sit down. If they are basically ingressive, corresponding states

may be derived by a stativization operation, as in French s'asseoir - (être) assis. Often, the relevant

verbs in a language are systematically polysemous between the two aktionsarten. Moreover, positionals

are not easily categorized as to their control properties, since body postures may or may not be

controlled.

In YM, positionals are represented as verbal roots of basically stative aktionsart. This can be inferred

from the fact that their conjugation class is v/itr -tal, and agentivization is factitive. If they were

basically ingressive, one would expect their conjugation class to be v/itr -0 and agentivization to be

causative. However, in some verbal categories, such as the imperative, they are ingressive; and in

several aspects there is a polysemy between the two aktionsarten.

Christian Lehmann

meaning posture v/itr -tal

stand wa'l

sit kul

sit (be on one's buttocks) pek'

brood pak

lie chil

lie extended pek

lie face down nok

lie on one's side tsel

lie on one's back haw

kneel xol

cower mòot'

crouch òoch

be on all fours xak

squat t'ùuch

squat p'uk

lean nak

hang ch'uy

live, dwell kah

live kux

F32. Examples of posture verbs

They are singled out here for two reasons. First, posture verbs form a very large and important lexical

field in YM. Second, this field is very homogeneous, especially as regards the members' morphological

potential. Many of the afore-described recategorization operations are applicable to each of the verbs.

Therefore, the field is particularly appropriate for illustrating the interplay of the various processes.

The lexical field of posture verbs comprises at least the verbs enumerated in F32. It is morphologically

very uniform. All verbs are monosyllabic, belong to the same inflection class (v/itr -tal), display the

same kind of allomorphy in certain categories and share the derivational potential exemplified by kul

in F33 (the first process there, the processive, is the base category).

Predicate classes in Yucatec Maya

category derived form meaning

processive kutal sit down

factitive kulkint place, set

durative kullankil sit (around) idle

resultative kula'n seated; at home

positional adjective kulukbal in a (sitting) position

spontaneous kulk'ah fall on one's bottom

F33. Derivational potential of posture verbs

Three kinds of allmorphy are noteworthy here. First, in the past tense, the allomorphs -chah and -lah

appear to be in free variation. Second, root final /l/ disappears before the /t/ of the incompletive suffix

(while generally it is only weakened to /h/). Third, the imperative, which normally has the ending -én,

in posture verbs has the allomorph -lén (e.g. xaklén `down on all fours!'). This is certainly due to

analogical levelling: The central posture verbs wa'l, chil, kul, xol, whose root ends in /l/, have an l-less

allomorph before -tal. Consequently, their imperative forms (e.g. kulén `sit down!') are reanalyzed

with a -lén imperative morph. This then spreads analogically to the other posture verbs.

7. Conclusion

Predicate classes are defined on an onomasiological basis. They are, therefore, first and foremost

universal classes of signs for situation cores. Inside a language, they manifest themselves as lexical-

semantic classes. These will be mapped on certain structural classes. The latter may be word-classes,

like noun, adjective, verb etc., and subclasses of these, e.g. property vs. state adjectives or durative

vs. punctual verbs. While word-classes necessarily differ in their grammatical behavior, their

subclasses may be found relevant only in derivational patterns.

The typical features of YM in the relevant domain have emerged on this universal background. The

central generalizations at the typologically relevant level are the following:

- The gamut of situation classes is mapped onto the word-classes of noun, adjective, modal and

verb. While for several situation classes, the association is unique, states are a formally

heterogeneous category, being represented in most of the word-classes.

- The most dynamic subclasses (terminative, ingressive, punctual) do not seem to have structural

relevance.

- Role features, especially involvement and control properties of participants, are structurally highly

relevant and constitute derivational classes as well as inflection classes and syntactic differences

among verbs representing them.

- Grammatical transitivity is confined to verbs with relatively high dynamicity and a relatively clear

control cline between two participants.

References

Blair, Robert W. 1964, Yucatec Maya noun and verb morpho-syntax. Bloomington: Indiana University PhD

diss. Ann Arbor: University Microfilms (65-3466).

Christian Lehmann

Dik, Simon C. 1978, Functional grammar. Amsterdam etc.: North-Holland (North-Holland Linguistic Series,

37).

Hengeveld, Kees 1992, Non-verbal predication. Theory, typology, diachrony. Amsterdam: Universiteit van

Amsterdam (PhD diss.).

Hopper, Paul J. & Thompson, Sandra A. 1980, "Transitivity in grammar and discourse." Language 56:251-299.

Kury2owicz, Jerzy 1936, "Dérivation lexicale et dérivation syntaxique. (contribution à la théorie des parties

du discours)." Bulletin de la Société de Linguistique de Paris 37:79-92.

Lehmann, Christian 1989, "Language description and general comparative grammar." Graustein, Gottfried

& Leitner, Gerhard (eds.), Reference grammars and modern linguistic theory. Tübingen: M. Niemeyer

(Linguistische Arbeiten, 226); 133-162.

Lehmann, Christian 1991, "Predicate classes and PARTICIPATION." Seiler & Premper (eds.) 1991:183-239.

Lehmann, Christian 1991, "Relationality and the grammatical operation." Seiler & Premper (eds.) 1991: 13-28.

Leiss, Elisabeth 1990, Die Verbalkategorien des Deutschen. Ein Beitrag zur Theorie der sprachlichen

Kategorisierung. Erlangen: unveröff. Habilitationsschrift.

McClaran, Marlys 1969, The verbal categories in Yucatec Maya. Paper presented to the 44. annual meeting

of the LSA. Abstract in: Linguistic Society of America 1969:(112).

Owen, Michael G. 1969, The semantic structure of Yucatec verb roots. New Haven: Yale University PhD diss.

(University Microfilms, 69-13,366).

Po'ot Yah, E. 1981, Yucatec Maya verbs. New Orleans: Tulane University Center of Latin American Studies.

Seiler, Hansjakob & Premper, Waldfried (eds.) 1991, Partizipation. Das sprachliche Erfassen von

Sachverhalten. Tübingen: G. Narr (LUS, 6).

Stefflre, Marlys McClaran 1972, Lexical and syntactic structures in Yucatec Maya. Cambridge, Mass.: Harvard

University PhD diss.

Viberg, Åke 1984, "The verbs of perception: a typological study." Butterworth, Brian et al. (eds.), Explanations

for language universals. Berlin etc.: Mouton (Linguistics 21/1); 123-162.

The syntax of non-verbal predication in Yucatec Maya

Article

Full-text available

Dec 2017

Grant Armstrong

The objective of this paper is to fill a void in the formal syntactic literature on Mayan languages by proposing a syntactic structure for clauses with non-verbal predicates in Yucatec Maya. The paper attempts to integrate the rich descriptions of non-verbal predicate constructions found in more functionally-oriented accounts (Lehmann 2002 [1998]; Verhoeven 2007; Vapnarsky 2013) with insights from the generative literature on argument licensing and clause structure in Mayan languages (Coon 2016 for an overview) as well as small clauses (Citko 2011 for an overview). I conclude that non-verbal sentences in Yucatec are matrix small clauses embedded under an Infl node, which is the locus for finiteness/stative aspect rather than tense morphology. This simple structure, coupled with independently motivated operations that have been proposed for small clauses and argument licensing/word order in Mayan languages, is able to account for a range of properties of these sentences that improves upon previous accounts developed in Armstrong (2010), Pye (2011) and Coon (2014).

Semantics and Morphology of Early Adjectives in First Language Acquisition

Book

Jul 2015

Los preverbos en maya yucatecoPreverbos in Yucatec Mayan

Article

Full-text available

Aug 2024

En los estudios de la lingüística maya se suele identificar una clase excepcional de adverbios que se han llamado “adverbios dependientes”, “adverbios incorporados” o “adverbios clíticos”, entre otros nombres. Su característica más llamativa es que aparecen entre la raíz verbal y el pronombre del juego A o el marcador de aspecto o modo. El objetivo de este trabajo es presentar un argumento a favor de la idea de que, en el maya yucateco, tal clase de palabras debe tratarse como su propia categoría léxica, distinta de los adverbios, que denominamos preverbos. Presentamos evidencia morfológica, sintáctica y semántica a favor de distinguir los preverbos de los adverbios por un lado y de raíces incorporadas en bases verbales compuestas por otro. Finalmente, hacemos una clasificación semántica detallada de los preverbos y proponemos un análisis sintáctico que explica su comportamiento.

Control in Yucatec Maya

Article

Full-text available

Dec 1996

Christian Lehmann

Manifestations of control in the verbal grammar of Yucatec Maya are investigated. In the semantic structure of a situation, a distinction between participant features, situation features and role features is made. In general, role features are grammaticalized to a much higher degree in Yucatec Maya than participant or situation features, and also to a higher degree than in other languages. There is a highly regular association of control properties with intransitive conjugation classes. Transitive verbs are more confined to the typical "transitive" situation, i.e. one with a clear control incline, than in many other languages. There are derivational operations which convert control into non-control verbs and vice versa and which are sensitive to control properties of their base.

Experiencer troubles: A reappraisal of the predicate-based asymmetry in child passives

Article

Full-text available

Oct 2022

Children’s understanding of passives of certain mental state predicates appears to lag behind passives of so-called actional predicates, an asymmetry that has posed a major empirical challenge for theories of passive acquisition. This paper argues against the dominant view in the literature that treats the predicate-based asymmetry as theoretically irrelevant. We instead propose a novel account that locates the problem in the syntax of experiencer constructions. Synthesizing theoretical and developmental evidence, we build a case for an early misanalysis of transitive subject-experiencer constructions as unaccusatives – structures that, by design, cannot passivize.

Formas verbales incorporadas transitivas en maya yucateco

Chapter

Full-text available

Jul 2002

Rodrigo Gutiérrez-Bravo

Valency classes in Yucatec Maya

Chapter

Full-text available

Aug 2015

Christian Lehmann

Full description of valency classes of Yucatec Maya in the framework of the Leipzig valency database

Noun incorporation and participation. A typological study on participant association with particular reference to Yucatec Maya

Article

Full-text available

Nov 2022

The present study investigates the function of noun incorporation within the domain of participation. The focus is on the accommodation of peripheral participants in incorporative structures. As many other languages, Yucatec Maya, the Mayan language of the Yucatan peninsula of Mexico, incorporates into the verb especially those participants that reside in the lower part of the control decline, i.e. undergoer, instrument, more rarely a location. The resulting intransitive construction is, however, seldom used in syntax and, instead, immediately retransitivized, opening a new undergoer slot to accommodate a further participant. This participant may be either promoted from a peripheral position (adjunct, possessive attribute), or it may be a new participant bearing no direct semantic relation to the base verb. Yucatec Maya is compared with four other languages, Nahuatl, Guaraní, Mayali and Samoan,1 which also use noun incorporation as a device to change the participant structure. We ask to what extent these languages employ similar or different strategies in their systems of noun incorporation with regard to the association of participants. A comparison with several other Mayan languages (Chontal, Huastec, Kanjobal, Jacaltec, Mam, Quiché) shows that Yucatec Maya is exceptional not only with respect to the importance of incorporative verbs but also with respect to the range of functions fulfilled by incorporation.

Noun incorporation and participation. A typological study on participant association with particular reference to Yucatec Maya

Chapter

Dec 2005

Experiential Constructions in Yucatec Maya: A typologically based analysis of a functional domain in a Mayan language

Book

Mar 2007

Elisabeth Verhoeven

This book combines a fieldwork-based language-specific analysis with a typological investigation. It offers a detailed and comprehensive analysis of the form and semantics of experiencer constructions in Yucatec, the Mayan language of the Yucatecan peninsula in Mexico. Since the linguistic expression of experience is not restricted to a specific grammatical area the study touches a great variety of grammatical fields in the language such as argument structure, grammatical relations, possessive constructions, subordinate constructions, etc. The empirical analysis of the Yucatec data is preceded by a thorough examination of the functional domain and the cross-linguistic coding of experience which until now could not be found in the literature. This study will be of interest to scholars working in the fields of typology and Native American linguistics, and especially to those interested in argument structure and the syntax-semantics interface.

Language Description and General Comparative Grammar

Chapter

Full-text available

Dec 1989

Christian Lehmann

The following postulates are formulated with respect to a scientific description of a language: 1. It is usable for whatever linguistic purpose. 2. It comprises an account of the linguistic system, a lexicon, a text corpus and a statement of the historical situation of the language. 3. The description of the language system accounts not only for core, but also for peripheral subsystems. 4. The linguistic system and the lexicon are presented both in a synthetic and in an analytic form. 5. The description brings out the dynamic character of the language. These postulates can be complied with if the description of the language system instantiates a general comparative grammar. This in itself obeys the postulates. Specific proposals for the implementation of a general comparative grammar, esp. with respect to postulate 4, are made.

Predicate classes and participation

Chapter

Full-text available

Jan 1991

Christian Lehmann

Predicate classes are semantic classes of lexemes that can function as the predicate of a clause. The are formed by the criteria relating to situation cores like dynamicity and telicity, relating to participant roles like degree of involvement, controll, affectedness and by participant properties like animacy. These structuring parameters are defined with examples from many languages. In the final part, numerous predicates are classified by these criteria, to serve as tertia comparationis in onomasiological descriptions of predicate classes of any language.

Non-Verbal Predication: Theory, Typology, Diachrony

Article

Full-text available

Jan 1992

Kees Hengeveld

On Grammatical Relationality

Article

Full-text available

Jan 1985

Christian Lehmann

The article refines Dependency Grammar by explaining the make-up of syntactic and morphological constructions on the basis of the relational properties of their components.

Studies in Functional Grammar

Article

Sep 1982

The verbs of perception: a typological study

Article

Jan 1983
LINGUISTICS

Åke Viberg

Transitivity in Grammar and Discourse

Article

Jun 1980

Transitivity involves a number of components, only one of which is the presence of an object of the verb. These components are all concerned with the effectiveness with which an action takes place, e.g., the punctuality and telicity of the verb, the conscious activity of the agent, and the referentiality and degree of affectedness of the object. These components co-vary with one another in language after language, which suggests that Transitivity is a central property of language use. The grammatical and semantic prominence of Transitivity is shown to derive from its characteristic discourse function: high Transitivity is correlated with foregrounding, and low Transitivity with backgrounding.

Yucatec Maya noun and verb morpho-syntax /

Article

Robert Wallace. Blair

Thesis--Indiana University. Vita. Photocopy of typescript.

Die Verbalkategorien des Deutschen

Article

Jan 1992

Elisabeth Leiss

Dérivation lexicale et dérivation syntaxique. (contribution à la théorie des parties du discours)

Jan 1936
79-92

Kury2owicz
Jerzy

Kury2owicz, Jerzy 1936, "Dérivation lexicale et dérivation syntaxique. (contribution à la théorie des parties du discours)." Bulletin de la Société de Linguistique de Paris 37:79-92.

Predicate classes in Yucatec Maya

Abstract

Recommended publications

ChemInform Abstract: Synthesis and Characterization of New Trifluoromethyl Substituted 3-Ethoxycarbo...

Closure systems, implicational systems, overhanging relations and the case of hierarchical classific...

The Language of Interjections

Computing all integer solutions of a genus 1 equation