Ang marks the what?

:
An analysis of noun phrase markers in Cebuano

Senior Essay
Department of Linguistics, Yale University
Advised by Claire Bowern and Robert Frank

Ava Tattelman Parnes

10 May 2011
 Acknowledgements:

Words can barely express the gratitude I feel toward those who have helped me tackle this
daunting project. I would like to thank my advisors, Professors Claire Bowem and Robert Frank,
for guiding me through the research process, suggesting possible analyses, asking me the tough
questions, and making me smile. They always made me feel supported and made what could
have been a frustrating experience an enjoyable one. I could not have completed this essay
without their incredible teaching skills and their dedication to working through the tough spots
with me. I would also like to thank Professor Larry Hom, the Director of Undergraduate
Studies, and the rest of the senior linguistics majors for their support and ideas in the senior essay
class. Lastly, lowe my understanding of Cebuano to one woman, Ms. Threese Serana. Her
warm personality and connection to Cebuano are what helped me to become interested in the
intricacies of the language two years ago. Her dedication to the discovery process that is
elicitation made this project a joy. The sentences, stories, and excitement we shared while
discovering things about Cebuano will remain with me long after this project is completed.
Table of Contents
1 Introduction 1
1.1 Overview of Essay 1
1.2 Cebuano Language Information and History 1
1.3 Voice in Austronesian Languages 3
1.4 Noun Phrase Markers in Austronesian Languages 6
1.5 Cebuano Noun Phrase Markers 10
2 Background 11
2.1 Outline of Cebuano Grammar 11
2.2 Cebuano Voice System 13
2.3 Analyses of Tagalog and Applications to Cebuano 16
3 Noun Phrase Markers as Topic Markers 17
3.1 Previous Analysis in Cebuano 17
3.2 Evidence in Favor of the Topic Marker Analysis 18
3.3 Evidence Against the Topic Marker Analysis 19
3.4 Tagalog Noun Phrase Markers as Topic Markers 20
3.5 Interim Summary 24
4 Noun Phrase Markers as Determiners 25
4.1 Previous Analysis in Cebuano 25
4.2 Implications of Analyzing Noun Phrase Markers as Determiners 26
4.3 Analysis of Relevant Data from Cebuano 28
4.4 Tagalog Noun Phrase Markers as Determiners 33
4.5 Interim Summary 35
5 Noun Phrase Markers as Case Markers 36
5.1 Previous Analysis in Cebuano 36
5.2 Implications of Analyzing Noun Phrase Markers as Case Markers 39
5.3 Analysis of Relevant Data from Cebuano 40
5.4 Tagalog Noun Phrase Markers as Case Markers 45
5.5 Interim Summary 49
6 Summary and Conclusion 50
References 54
Appendix 57
1 Introduction

1.1 Overview of Essay

Cebuano is an Austronesian language spoken in the Philippines by about 15 million

people. Like many other Western Austronesian languages, it has a complex agreement system

between verbs and noun phrases whose characterization has been the subject of much

disagreement over at least the last fifty years (Himmelmann 2005). As a result of the varied

characterizations of the voice system, Cebuano noun phrase markers have also received many

different classifications, ranging from 'introducing particles' and 'construction markers' to

'determiners', 'topic markers', and 'case markers' (Reid 2002). This paper attempts to arrive at

a classification of the Cebuano noun phrase markers by looking at their distribution, function,

and semantic characteristics, both as described in previous analyses and in newly collected data.

Section 1 offers a basic history of the Cebuano language as well as a summary of the

issues in analyzing noun phrase markers in Austronesian languages. Section 2 contains an

outline of relevant features in the grammar of Cebuano, including the voice system, as well as an

explanation of the relevance of Tagalog to analysis in Cebuano. In Sections 3 through 5, I

examine three different analyses of Cebuano noun phrase markers, noting possibilities suggested

by my own evidence and relevant analyses in Tagalog. Section 3 covers the topic marker

analysis, Section 4 covers the determiner analysis, and Section 5 covers the case marker analysis.

Section 6 provides a summary of the possibilities and conclusions.

1.2 Cebuano Language Information and History

Cebuano is a member of the Bisayan languages, a subgroup of the Central Philippine

languages (Lewis 2009). Other Central Philippine languages include Tagalog and Bikol.

1
Cebuano is spoken in the Southern Philippines on the islands of Cebu and Bohol and on parts of

the islands of Leyte, Negros, and Mindanao (Lewis 2009).

....
....
"""
...
- ....

--DON

...

The first written work on the language was done in the 16th and 17th centuries by the

Spanish, and the first English-Cebuano dictionary was published in 1900 (Tanangkingsing 2009:

10). Some of the most important early analysis of the language was done by John U. Wolffin

his dissertation and subsequent publications in the 1960s and 1970s. Due to the colonization of

the Philippines by the Spanish in the 16th century, the language has a good deal ofloanwords

from Spanish, including much of the numeral system above ten, and has also incorporated words

from English and Chinese, for example sipir for 'zipper' and bakya from the Chinese for

2
'wooden slippers' (Tanangkingsing 2009: 14). Since the first written work on the language,

Cebuano has changed considerably, both in its morphosyntax (specifically verb forms) and in its

lexical inventory. At least one third of the vocabulary from the mid-17th century was unknown

to a group of 100 speakers ofCebuano in the 1960s (Wolff2008: 197).

Cebuano is similar to Tagalog in its sound system, verb system, affix system, and some

of its lexical inventory. However, Cebuano has a different tense-aspect system, a more complex

system of deictics, and a slightly different set of noun phrase particles. Unlike Tagalog, Cebuano

has what Frank Blake (1905: 127) calls "a sort of indefinite accusative particle...ug." Tagalog has

the most speakers of any Philippine language, with over 21 million recorded in the Philippines

during the 2000 census, while Cebuano is the second most spoken language, with over 15

million speakers on the Philippines (Lewis 2009).

1.3 Voice in Austronesian Languages

Cebuano is a member of the Austronesian language family, possibly the largest language

family in the world, with about 1200 languages. It is also an incredibly diverse and widespread

family, stretching from Easter Island to Madagascar, with at least 80 separate languages in

Vanuatu alone (Crowley 2008: 97).

While the linguistic structures of these languages vary greatly, many languages in Taiwan

and the Philippines are known for their elaborate voice systems, sometimes called 'focus'

systems, which mark verbs for actor focus, undergoer focus, locative focus, and instrumental

focus (Adelaar & Himmelmann 2005: xvi; Crowley 2008: 100).

3
The following two examples illustrate the voice alternation in Cebuano.'

(1) ang liyon nikaon sa iro

ANG lion AV-eat SA dog
'The lion ate the dog'

(2) ang liyon gikaon sa iro

ANG lion NAV-eat SA dog
'The dog ate the lion'

In example (1), the verb nikaon carries the actor voice affix ni- and the ang-marked NP ang liyon

is the actor ofthe sentence while the sa- marked NP sa iro is the patient. Example (2) features

the same noun phrases but a different verb affix. The verb gikaon carries the patient voice affix

gi- and the ang-marked NP is the patient of the sentence while the sa-marked NP assumes the

agent role. While sentences like (2) have sometimes been analyzed as passives (Guilfoyle,

Hung, & Travis 1992), there is significant evidence that this is not the case (see Section 5.3), and

that the above alternation represents two equally viable structures in the language, rather than a

frequently used active form and its infrequently used passive cousin.

The classification of this system is complicated by the varied terminology used to refer to

the verb types and to the ang-marked noun phrase. Most significantly, the use of the terms

'topic' and 'focus' in Austronesian linguistics have an interesting history that obscures their

relationship to traditional focus and topic. Traditionally, pragmatic focus is defined as referring

to information that is new and of high interest, often marked by stress (Talmy 1985) while

pragmatic topic refers to given information. Levinson (1983) posits that "a major function of

topic marking is precisely to relate the marked utterance to some specific topic raised in the prior

1 Due to the various analyses of the classifications of the noun phrase markers, I gloss them with capital letter
equivalents, unless I am citing an example from the literature, in which case I use the same gloss as the author. I
gloss verb affixes as 'actor voice' and 'non-actor voice' to maintain clarity in the examples; there are also many
other analyses of these affixes. Again, when citing an example from the literature, I use the same affix gloss as the
author. A full list of the gloss abbreviations used can be found at the end of this paper.

4
.,
discourse." The most often used example of topic marking is Japanese, as in (3), an example

taken from Nakanishi (2001).

(3) John-wa gakusei desu

JoOO-TOP student is
'Speaking of John, he is a student'

In Austronesian linguistics, these terms were used beginning in the 1950s to distinguish

subjects in European languages from patterns found in Austronesian languages (Himmelmann

2005). The term 'topic' was used instead of 'subject' and 'focus' was used instead of 'voice' to

talk about noun-verb agreement. Ang was thought to mark this 'topic' and the 'topic' matched

the 'focus' of the verb. Thus, an 'agent focus verb' marks the agent as the 'topic', while an

'obj ect focus verb' marks the patient as the 'topic'. It is unclear whether Austronesian 'topic'

and 'focus' are compatible with the traditional senses of the terms.

The issue of the basic alignment system in these languages remains highly contested as

well. The voice or 'focus' systems describe the agreement between the verb and a particular

noun phrase marker, essentially determining which noun phrase is the 'subject' of the sentence.

The relation of the marking of these 'subjects' to the valency of the verb determines the

classification of the alignment system. While it is generally agreed that the most of the

languages are not nominative-accusative - that is, subjects of transitive and intransitive clauses

are marked the same while objects of transitive clauses are marked differently - some claim the

languages show ergative alignment or split-S (active) alignment while others suggest an entirely

different system. Himmelmann (2005: 158) notes that "in Western Austronesian languages

clear-cut cases of ergative alignment are restricted to person marking systems, with little or no

evidence of an ergative distribution of noun phrase markers." Furthermore, since the difference

between ergative and nominative-accusative alignment is based on how markers group with

5
respect to transitive and intransitive clauses, the analysis of alignment is heavily reliant

determining which verbs are transitive in the language and which noun phrases are core

arguments of those verbs. These distinctions are not always clear, complicating the analysis of

the voice system (see Section 5.3). Additionally, according to Dixon (1994: 219) it is unclear

exactly what a language being ergative would entail, since there does not seem to be much that

correlates with it. Because the voice system interacts with noun phrase markers, the lack of

consensus on the analysis of the voice system complicates the classification of those markers.

Another characteristic of Western Austronesian languages is that they seem to lack a

definite unmarked word order: in many of these languages, word order is more flexible,

employing, for example, both SVO and VXS order. However, not all word order is free. Most

Western Austronesian languages have prepositions as opposed to postpositions and auxiliaries

generally precede main verbs. Constituents of noun phrases and prepositional phrases cannot be

distributed discontinuously across the clause, and there are generally only two or three phrase

ordering options within a clause. Placement of adjuncts, though, is much less restricted

(Himmelmann 2005: 142). Due to the flexible word order of arguments, noun phrase markers

playa key role in determining agreement with the verb and assignment of theta roles to noun

phrases.

1.4 Noun Phrase Markers in Austronesian Languages

The classification of noun phrase markers is an issue across all Western Austronesian

languages. As noted by Reid (2002), there have been a great number of classifications of the

monosyllabic words that precede nouns. Since they are often translated to English as articles or

prepositions, several analyses have simply used this classification (Vanoverbergh 1955;

Lambrecht 1978). Other linguists have simply labeled them based on their distribution, e.g.

6
r

"introducing particles", "noun-marking particles", or "noun phrase markers" (Reid 2002: 296).

Others have labeled them "construction markers" in order to denote that they identify the

construction of which they are a constituent. There have also been many analyses which attempt

to denote the grammatical function of the words, e.g. "nominalizing particles", "determinants of

relations", "determiners", or, most commonly, "case-marking particles" or "topic/goal-markers"

(Reid 2002: 298).

Himmelmann (2005: 133) notes that the noun phrase markers in Austronesian languages

are all clitics. Reid (2002: 295) notes that in some cases, authors classify the markers simply as

proclitics. The labeling of these particles as phrase-marking clitics does little to illuminate their

role in the Cebuano phrase. Clitics are neither affixes nor words that stand alone, but none of

these wordhood classifications require a particular syntactic classification. Thus, whether or not

the particles are clitics, there are still questions as to their function within the sentence.

However, their status as clitics is helpful in determining the relevance of other observations.

According to Zwicky (1977), clitics have some properties of affixes, such as being bound

morphemes and being stressless, and some properties of independent words, such as resisting

phonological rules that do not apply across word boundaries and attaching to words of more than

one syntactic category. In Cebuano, the noun phrase markers ang, sa, and ug as well as the

personal name markers si, ni, and kang are not clearly independent words, as they cannot stand

alone, even as the answer to a question. Furthermore, they do not bear stress.

Yet the markers are not affixes either, as they can attach to more than one type of

syntactic category. For example, ang can attach to nouns, adjectives, and verbs.

(4) ang auto kay dako

ANG car was big
'The car was big'

7
 (5) ang pula nga auto kay dako
ANG red NGA car was big
'The red car was big'

(6) tulo ka mga iro ang nipaak sa tawo

three KA PL dog ANG AV-bite SA man
'Three dogs were the ones who bit the person'

In some languages, orthography also supports the analysis of the particles as clitics, in

that they are separate words in the orthography, similar to the English word 'the'. However, this

evidence does not speak to the function of the particles in the sentence.

Even classifications which do focus on the function of the noun phrase markers within

the sentence vary greatly. For example, in Cebuano, ang and sa mark common noun phrases.

The example below shows a few different characterizations of the noun phrase markers.

(7) ang liyon nikaon sa kame

DET lion ACT-eat OBL meat (Huang & Tanangkingsing 2005)
NOM/TOP lion ACT-eat ACC meat (Guilfoyle, Hung, & Travis 1992)
ABS lion ANTIP-eat OBL meat (Walters 1994)
TOP lion AV-eat OBL meat (Ghazali 1990)
NOM/ABS lion AV-eat OBL meat (Tanangkingsing 2009)
'The lion ate the meat'

In this example, the ang-marked noun phrase is classified by different authors as a noun marked

by a determiner, a noun bearing nominative case, a noun bearing absolutive case, and a topic

marked noun. In the first two cases, the verb is analyzed as active, while in the third example, it

is an antipassive, and in the fourth and fifth it is an actor voice verb. The sa-marked noun is

analyzed as accusative or oblique, depending on the framework.

Reid points out several problems with the above categorizations, particularly the popular

case marker analysis. One is that although the words do introduce noun phrases, they can also

precede predicate nouns as in (8) and fronted noun phrases as in (9). He suggests that fronted

8
NPs cannot receive case by virtue of moving to a fronted position, since A'-movement does not
. 2
assign case.

(8) ang babae ang pumasok (Tagalog)

ANG woman ANG entered
'The one who entered was the woman'

(9) ang babae, ay pumasok (Tagalog)

ANG woman TOPIC.LINKER entered
'As for the woman, she entered'

Additionally, Reid states that the various labels are functional but do not specify syntactic

category, pointing out that discussions of terminology and labels do not illuminate the role of

these markers in syntax. He chooses to examine the idea that they are determiners based on the

criteria that they are dependents of head nouns, occur at the outer edge of the noun phrase,

cannot be modified by other forms, and agree in semantic features with the head noun (Reid

2002: 298). His theory of determiners maintains that determiners are dependents of head nouns,

so evidence of markers followed by verbs and prepositional phrases leads him to conclude that

the markers are not determiners. Yet his proposed analysis of the 'nominative' noun phrase

bears a strong resemblance to a Determiner Phrase with a D head, suggesting that perhaps he too

focuses on terminology and labeling, rather than syntactic analysis. Thus, while Reid highlights

the lack of standard analysis of these markers, and surveys their previous analyses, he does not

provide a convincing cross-linguistic analysis. Given the differences in voice/focus structure

among the Austronesian languages and the fact that noun phrase markers are deeply related to

voice structure, it is difficult to generalize across the entire language family. As evidenced by

previous research, attempts to devise a single analysis of phrase marking systems in the entire

2 Reid's rejection of the case marker analysis is based on the assumption that fronted NPs move to an A' position.
However, it is possible that fronted NPs move via A-movement, or that they receive case before moving.
Additionally, predicate NPs are not barred from showing case cross-linguistically. Thus Reid's rejection does not
preclude a consideration of noun phrase markers as case markers.

9
language family leads to imprecise labeling with a focus on terminology rather than syntactic

function. In order to develop a deeper understanding of noun phrase markers, it is necessary to

analyze them within a single language and to test the implications of each analysis in depth.

1.5 Cebuano Noun Phrase Markers

This essay will focus on the classification of the noun phrase markers in Cebuano. As

shown above, there have been many analyses of these markers in the literature, but there is no

standard analysis (Himmelmann 2005: 146). Previous work on Cebuano has tended to focus on

the voice/'focus' system of the language, and has chosen an analysis for the noun phrase markers

that best fits the larger voice and phrase structure analysis without much independent analysis.

For example, Tanangkingsing & S. Huang (2007) employ the case marker analysis while H.

Huang & Tanangkingsing (2005) treat the noun phrase markers as determiners, prepositions, and

case markers. I will examine the noun phrase markers independently, and determine whether it

is possible to develop an independent analysis, or whether the function of the markers is

inextricably bound up in the analysis of the voice/'focus' system. If the markers are case

markers, then the case features of the noun phrase must be assigned in view of the argument

structure of the verb and the theta roles it assigns to the noun phrases. In this case, an analysis of

the voice system is required for a complete understanding of the noun phrase markers. However,

if the markers are topic markers, then verb agreement is simple: the verb agrees with the topic-

marked noun phrase and the analyses of the voice system as nominative-accusative or ergative-

absolutive do not have bearing on the noun phrase markers.

Data for my analysis comes from several publications on Cebuano, including Michael

Tanangkingsing's 2009 dissertation, a reference grammar ofCebuano. Additionally, data has

been taken from elicitation done by members of Claire Bowem's field methods class (LING

10
,
241b) at Yale University in the spring of 2009. Additional data was collected in the spring of

2011 through elicitation sessions with Ms. Threese Serana, who was also the consultant for the

class.

Elicitation with Ms. Serana was conducted on a weekly basis, focusing mainly on

complete sentences of differing levels of transitivity and with different voice structures. An

effort was made to obtain various grammatical sentences with different word orders and different

uses of noun phrase markers, as well as to determine ungrammatical sentences whose

ungrammaticality was based solely on noun phrase marker usage.

2 Background

2.1 Outline of Cebuano Grammar

The Cebuano phoneme system is very similar to Tagalog. Cebuano has sixteen

consonants, shown in the table below along with their orthographic equivalents, if they differ.

T a bi e 2-1 : C eb uano Ph oneme S sem

t (B unye & Y ap 1971)
bilabial dental palatal velar glottal
voiceless stops p t k t = - or no letter
voiced stops b d g
nasals m n 1) = ng
fricatives s h
liquids I I r
glides J=y w

Cebuano has three main vowels, /iI, lui, and la/. However, due to the influence of Spanish and

numerous Spanish loanwords, lei and 101 also occur and have made their way into the

orthography.

Although Cebuano is often characterized as a verb-first language (Billings 2005), word

order tends to be extremely flexible, with VSo, vas, and SVO orders possible in simple
sentences (assuming traditional characterization of subject and object, see Section 2.2 below).

11
 (10) ang tawo nikaon ug kanon
ANG man AV-eat UG nee
'The man ate rice'

(11) nikaon ang tawo ug kanon

AV-eat ANG man UG nee
'The man ate rice'

(12) nikaon ug kanon ang tawo

AV-eat UG nee ANG man
'The man ate rice'

However, there is less flexibility in sentences with more than two arguments, or in

sentences involving possession, where the possessor always follows the possessee.

(13) nihatang ang tawo sa libra sa iyang higala

AV-give ANG man SA book SA his friend
'The man gave the book to his friend'

(14) *nihatang sa libra ang tawo sa iyang higala

AV-give SA book ANG man SA his friend

(15) si Inday ang nipaak sa iro ni Threese

SI Inday ANG AV-bite SA dog NI Threese
'It was Inday who bit Threese's dog'

(16) *si Inday ang nipaak ni Threese sa iro

SI Inday ANG AV-bite NI Threese SA dog

In addition to the noun phrase markers discussed in this paper, Cebuano also has a few

important feature markers and linkers. The plural marker mga precedes nouns it modifies.

(17) ang iro nipaak ni Inday

ANG dog AV-bite NI Inday
'The dog bit Inday'

(18) ang mga iro nipaak ni Inday

ANG PL dog AV-bite NI Inday
'The dogs bit Inday'

12
The numeral linker ka, corresponding roughly to 'units of is used with cardinal numbers in

addition to the plural marker.

(19) ang tufo ka mga iro nipaak ni Inday

ANG three KA PL dog AV-bite NI Inday
'The three dogs bit Inday'

Clauses and phrases are often connected with the conjunction ug (different from the noun

phrase marker ug) or with the relativizer nga.

(20) mupadafa ug sulat si Inday kang Perla nga taga Sugbu

AV-send UG letter SI Inday KANG Perla NGA from Cebu
'Inday will send a letter to Perla who is from Cebu'

Nga is also used to link adjectives to nouns.

(21) ang tawo nga taas nikaon sa puti nga kanon

ANG man NGA tall AV-eat SA white NGA nee
'The tall man ate the white rice'

In some cases, nga attaches to the word before it, such as a demonstrative or pronoun. In this

case, the final vowel is dropped.

(22) kini nga liyon mukaon ug kame

this NGA lion AV-eat UG meat
'This lion eats meat'

(23) kining liyon mukaon ug kame

this NGA lion AV-eat UG meat
'This lion eats meat'

Relative clauses headed by nga are important in determining the role of noun phrase markers in

contexts other than simple transitive sentences (see Section 5.3).

2.2 Cebuano Voice System

The Cebuano voice system is the agreement system between verbs and their arguments.

While Cebuano verbs are not marked for person, number, or gender, affixes do indicate tense and

aspect, as well as the theta roles of the arguments. Arguments are marked with one of the noun

13
phrase markers ang, sa, or ug, or one of the personal name markers si, ni, or kang' The verbal

morphology indicates the role of the noun phrase marked with ang or si. Other arguments fill in

appropriate roles. Actor voice (AV) verbs agree with an ang-marked agent, while non-actor

voice verbs agree with an ang-marked patient (PV), instrument (IV), or location (LV). Below is

a selection of verbal affixes along with examples of their use in sentences.

- : C eb uano V erbAffiixes (T ananzikinzsmz 2009)

Ta bl e 22
affix VOICE TENSE ASPECT/MOOD
mag(a)- AV Future Imperfective; Volitional
man-Z mang- AV Future Imperfective
mo- /mu- AV Future Volitional
na- AV Non-Future Potential-Spontaneous
ni- AV Non-Future Volitional
ei- NAV Non-Future Perfective; Volitional
gika- IV Non-Future Potential-Abilitative
na- PV;IV Non-Future Potential-Spontaneous
ma- PV;IV Future Potential-Spontaneous
i- IV Future Volitional
-un PV Future Volitional
-an LV Future Volitional

(24) ang lalaki mupadala sa sulat ngadto sa babayi

ANG boy MU-send SA letter toward SA woman
AV-send
'The boy will send the letter to the woman'

(25) ang lalaki nipadala sa sulat ngadto sa babayi

ANG boy NI-send SA letter toward SA woman
AV-send
'The boy sent the letter to the woman'

(26) gipadala sa lalaki ang sulat ngadto sa babayi

GI-send SA boy ANG letter toward SA woman
PV-send
'The boy sent the letter to the woman'

In the above three examples, it is clear that while AV verbs mark the agent with ang, the

PV verb gipadala marks the patient with ang. In the next three examples, it is clear that

3 I will focus mainly on the common noun markers. Some observations are also applicable to the personal name
markers. However, the personal name marker kang does not match any common noun marker.

14
,
changing either the verb affix or the noun phrase markers will alter the theta role distribution in

the sentence. In a simple transitive sentence, if the verb is changed from AV to PV, the agent

becomes the patient and the patient becomes the agent. If the verb affixes remain constant and

the noun phrase markers are switched (making required adjustments for personal names), the

agent and patient also switch.

(27) ang iro nipaak ni Inday

ANG dog NI-bite NI Inday
AV-bite
'The dog bit Inday'

(28) ang iro gipaak ni Inday

ANG dog OI-bite NI Inday
PV-bite
'Inday bit the dog'

(29) si Inday gipaak sa iro

SI Inday OI-bite SA dog
PV-bite
'The dog bit Inday'

In some cases, the agent can even be omitted from the sentence with aNAV verb.

(30) ipadala ni Inday ang sa/at ngadto sa Sugbu

I-send NI Inday ANG letter toward SA Cebu
IV-send
'Inday will send the letter to Cebu'

(31) ipadala ang sa/at ngadto sa Sugbu

I-send ANG letter toward SA Cebu
IV-send
'The letter will be sent to Cebu'

However, the omission of agents is not consistent enough to justify calling (31) a passive

according to Tanangkingsing & Huang (2007), and generally, agents are not demoted to non-

arguments in NAV clauses. They also note that most instances of agent omission are fixed

expressions of naming and identifying.

15
,
Once one knows which affixes correspond to which voice, it is not difficult to pick out

the agent in an actor voice sentence (it is the ang- or si-marked NP) or a patient voice sentence

(it is the sa- or ni- marked NP). However, it is not clear what type of agreement system Cebuano

possesses, or what is driving this agreement. An independent look at the properties of the noun

phrase markers may help to clarify the possible systems.

2.3 Analyses of Tagalog and Applications to Cebuano

Although Cebuano is the second most spoken language in the Philippines, the literature

tends to be more descriptive than analytic of the syntax. Tagalog has been studied more

frequently, in more depth, and with greater formal rigor than Cebuano, and therefore I include

some analyses of Tagalog in order to provide some insight into possibilities for Cebuano

analysis. While there are differences between Cebuano and Tagalog, major similarities exist in

terms of the set of noun phrase markers, their morphology, and their use in sentences. Therefore,

conclusions drawn about Tagalog may be applicable to Cebuano. However, the extent of

syntactic differences between Cebuano and Tagalog is not well studied, so no definitive

conclusions can be drawn without further tests.

As in Cebuano, the analysis of noun phrase markers in Tagalog has depended heavily on

analysis of the voice and argument structure of the language. There have been several analyses,

ranging from its classification as an ergative language (Aldridge 2004), an accusative language

with extensive use of passive constructions (Guilfoyle, Hung, & Travis 1992), a hybrid between

accusative and ergative languages, to a member of its own distinct class.

Below, I examine previous analyses of Cebuano noun phrase markers, some of which are

independent and some of which are part of an analysis of the voice system. I note the

implications of these analyses and then present my own data and what it adds to the discussion. I

16
then look at some similar analyses in Tagalog to see what insight they might provide for an

analysis of the Cebuano noun phrase markers.

3 Noun Phrase Markers as Topic Markers

3.1 Previous Analysis in Cebuano

While the use of the term 'topic' was initially introduced in order to separate

Austronesian linguistics from European linguistics, the analysis of noun phrase markers as

'topic' or pragmatic topic markers has continued even recently. Authors who note that the

choice of noun phrase markers tend to have some influence on the perceived importance of the

noun in the sentence continue to posit that noun phrase markers are topic markers. Ghazali

(1990: 52) describes the topic as the part of the sentence with the most emphasis, and notes that

it can be the actor, goal, instrument, or location, and that it is always marked with ang. Below,

examples (32) - (35) demonstrate, respectively, actor focus, object focus, instrumental focus, and

locative focus."

(32) magluto ang babaye ug bugas sa kulon

ACT-cook NOM woman OBL rice OBL ricepot
'The woman will cook rice in the ricepot'

(33) luto 'on sa babaye ang bugas sa kulon

cook-OBI GEN woman NOM rice OBL ricepot
'The rice will be cooked in a ricepot by the woman'

(34) isulat ni Linda ang lapis ug sulat

INS-write GEN Linda NOM pencil OBL letter
'The pencil will be used by Linda to write a letter with'

(35) luto 'an sa babaye ang kulon ug bugas

cook-LOC GEN woman NOM ricepot OBL nee
'In the ricepot is where the woman will cook rice'

Ghazali recognizes that there is a close relationship between the voice marking of the

verbs and the noun phrase marking of their arguments. In each of the above examples, the affix

4 The following examples employ the glossing system used by Ghazali (1990).

17
,
of the verb must match the noun phrase marked with ang. She also notes that the locative and

instrumental voices sometimes match topics other than location and instrument: locative voice is

sometimes used for the indirect object or benefactive, and instrumental voice is sometimes used

for the time of an action.

3.2 Evidence in Favor of the Topic Marker Analysis

There are several reasons for viewing ang-marked noun phrases as pragmatic or

discourse topics. The first is Ghazali's observation that ang seems to mark the phrase with the

most emphasis, that is, what the sentence is 'about'. Furthermore, ang seems to mark

information previously given in the discourse. For example, the following minimal pair differs

only in the use of ang or sa in one instance. The sentences are translated identically; the only

difference relates to what the discourse preceding the sentence would have been.

(36) walay pagkaon nga gikaon ang iro

NEG food NGA NAV-eat ANG dog
'There was no food that the dog ate' (previous discourse about the dog)

(37) walay pagkaon nga gikaon sa iro

NEG food NGA NAV-eat SA dog
'There was no food that the dog ate' (previous discourse about the food)

Lastly, there are some properties of ang-marked noun phrases that are inconsistent with

classification as subjects. For example, ang-marked nouns must be specific or definite.

According to Schachter & Otanes (1972), this is a strange requirement for a subject. In English

and other European languages, the subject of a sentence can be indefinite and non-referential. In

Cebuano, this is not possible:

(38) kaonon sa bata ang saging

eat-NAV SA child ANG banana
'The child will eat the/*a banana'

18
Since pragmatic topics mark given information, it is expected that they are specific and definite,

and so classifying the ang-marked noun phrase as a pragmatic topic provides a satisfying

explanation for the definiteness restriction.

3.3 Evidence Against the Topic Marker Analysis

There are some issues, however, with the classification of ang as a topic marker. If ang

marks pragmatic topic in the traditional sense, then ang-marked noun phrases are the given

information in a sentence. Pragmatic focus refers to new information, and thus pragmatic topic

and focus are mutually incompatible (Kroeger 1993b). In the answer to a WH-question, the

phrase corresponding to the WH-word should carry focus and thus not be marked for pragmatic

topic. However, an ang-marked noun phrase can be used to answer a WH question.

(39) unsa ang imung gipalit sa merkado

what ANG you NAV-buy SA market
'What did you buy at the market?'

(40) gipalit nako ang sinena

NAV-buy ISG ANG dress
'I bought the dress'

Since the focus-carrier is marked with ang, ang does not mark pragmatic topic.

Another possibility is that ang marks discourse topic, a noun phrase that is salient in context.

In this case, it would mark what the discourse is 'about' at the level beyond the sentence. One

way to measure this is Topic Persistence, "the number of contiguous subsequent clauses in which

the participant NP remains a semantic argument of the clause." (Cooreman, Fox & Givan 1984)

Topical arguments have higher values for Topic Persistence, so if ang marks discourse topic,

ang-marked nouns should have higher values for Topic Persistence. Data on Topic Persistence

from Walters' (1994) research on ergativity in Cebuano suggests that this is not the case.

19
 -
Tabl e 3 1 : T OPIC . t ence
. P ersis 0 r c eb uano NP S (W a Iters 1994)
# Clauses Agents Patients
'Actor focus' (ang-marked agents) 8 TP = 1.9 TP = 0.8
'Object focus' (ang--marked patients) 24 TP = 2.7 TP = 0.5

If ang marked discourse topic, topic persistence would be highest for agents in 'actor focus'

clauses and for patients in 'object focus' clauses. Yet in Walters' data, agents have higher values

for topic persistence both in actor focus clauses and in object focus clauses. Even when the

patient is ang-marked, the agent has higher topic persistence. This suggests that ang does not

mark discourse topic, because ang-marking does not have an effect on the topicality of the noun

phrase.

3.4 Tagalog Noun Phrase Markers as Topic Markers

In Tagalog as well, several analyses classify ang as a topic marker. Richards (2000)

examines the following alternation in Tagalog and claims that it has nothing to do with case, but

with which item is marked as the topics:

(41) bumili ang lalaki ng bigas

ACTORToPIC-bought TOPIC man GOAL nee
'The man bought rice'

(42) binili ng lalaki ang bigas

GOALTOPIC-bought ACTOR man TOPIC nee
'A man bought the rice'

He states that the ang-marked topic is in an external subject position, while the actor

(lalaki in both examples) is base-generated in an internal subject position. The uniqueness of

Tagalog stems from its ability to fill both the external and internal subject positions with

different nominals without either being demoted to an adjunct. Example (43), a tree for (42)

shows the actor (ng lalaki) generated in an internal subject position while the topic (ang bigas)

occupies an external subject position (Richards 2000).

5 In the following examples I use Richards' (2000) gloss system.

20
 (43) IP

I' NP
/~
I VP ang big-as
I

~J
binili NP

1191alaki

Richards also states that one of the major problems in theories of Tagalog is the

assumption that the topic moves to Spec IP, a position where subjects in other languages receive

case. He instead argues for an A'-specifier above Spec IP (labeled nP) which is occupied by the

Tagalog topic. This position has been argued for in Icelandic, and Richards compares the

Tagalog alternation to the following alternation in Icelandic:

(44) eg he! aldrei hit Mariu

I have never met Maria

(45) Mariu he! eg aldrei hit

Maria have I never met

In (45), eg is generated in Spec VP and moves to Spec IP while Mariu is generated under V' and
moves by A'-movement to Spec nP, a non-ease-assigning position. This movement is shown in

(46).

21
 (46) IIP

NP II'

I
Mariu II IP
I
hef NP l'

I
eg I VP

U
~
P V
IAV NP
j I
hitt

Richards suggests that Tagalog structure is very similar, the only difference being that

while Icelandic movement to Spec rrP is overt, Tagalog movement occurs at LF, as shown in

example (47).

(47) IIP

NP III

I
II IP
I
nakilala N P l'

I
ko I VP

I~NP
r~~
Si Maria
,:. .J
,

nakilala ko si Maria
AV-met ISG SI Maria
'I met Maria'

22
 Richards has several reasons for proposing this structure in Tagalog. First, Tagalog

topicalization behaves like A'-movement for Binding Theory, rather than case driven movement,

suggesting that ang cannot be a case marker. Movement of an anaphor by A-movement to a

position c-commanding its binder is not possible, but topicalization of an anaphor is possible in

Tagalog:

(48) nagmamahal si Juan sa kanyang sarili

ACTORTOPIC-loves TOPIC Juan Loc his self
'Juan loves himself

(49) minamahal ni Juan ang kanyang sarili

GOALTOPIC-loves ACTOR Juan TOPIC his self
'Juan loves himself

Topicalization of an anaphor above its binder is also possible in Icelandic, suggesting Tagalog

topicalization is movement to Spec nP, rather than case-driven movement (Richards 2000: 107).

(50) Jon elskar sjalfan sig

John loves himself

(51) sjalfan sig elskar Jon

himself loves John

Further evidence stems from the fact that Tagalog has a ban on extraction of non-topics.

(52) [cp sino ang [rrP hinalikan [IP ni Maria]]]

who DIRECTIONTOPIc-kissed ACTOR Maria
'Who did Maria kiss?'

(53) "Icr sino ang [rrP humalik [IP si Maria]]]

who AcronTcnc-kissed TOPIC Maria

This looks similar to the Icelandic restriction in that only elements in the external subject

position can be extracted.

(54) [cp hvern [rrP hefur [IP Maria kysst]]]?

whom has Maria kissed

(55) *[cp hvern [rrP Maria hefur [IP kysst]]]?

whom Maria has kissed

23
In both languages, extraction can take place only if the extracted element is in Spec nP. Since

the topic occupies Spec nP, it can be extracted, but extraction of other elements is blocked by the

topic in Spec nP. In (55), Maria, the topic, blocks extraction through Spec nP. In (53), si Maria

has moved to Spec nP at LF, and therefore blocks extraction of the object of kissing.

Richards addresses another objection to Tagalog topichood by noting that although

Tagalog 'topics' do not need to be especially discourse-prominent, Icelandic can also have less

discourse-prominent topics. Additionally, the restriction that Tagalog topics be specific or

definite is also similar to Icelandic topic, and fits much better with the topic classification than

with the subject classification cross-linguistically (Richards 2000: 110). Based on the

similarities between Icelandic and Tagalog, Richards concludes that ang is a topic marker rather

than a case marker, and that the question of whether or not Tagalog is ergative-absolutive is

irrelevant.

3.5 Interim Summary

It seems then, that although the term 'topic' has been used in the Cebuano literature for

years, this term does not line up with either pragmatic or discourse topic. However, as Richards'

analysis suggests, it may be necessary to consider topichood outside the usual senses of given

information or 'aboutness' and to view it as an internally or externally generated noun phrase

that moves to an AI position at LF. In order to determine whether this type of analysis is

desirable, it is necessary to consider other possibilities for the noun phrase markers, and to

reconsider the similarities between Cebuano 'topics' and European subjects.

24
4 Noun Phrase Markers as Determiners

4.1 Previous Analysis in Cebuano

Due to some of the differences in function between ang, sa, and ug, some have

questioned whether the noun phrase markers truly form a paradigm. In particular, they point out

that ang behaves differently from both sa and ug. Huang and Tanangkingsing (2005: 585) note

that while sa can be preceded by a demonstrative pronoun in a noun phrase, ang cannot.

(56) kining sa iyang mama kay pilo-on

this SA her mother is fold-NAY
'This, which is her mother's, is folded.'

(57) *kining ang lalaki nipaak sa iro

this ANG boy AV-bite SA dog

While some use the differences between ang and other noun phrase markers to claim it is a topic

marker, Huang and Tanangkingsing classify ang as a determiner and sa as a case marker.

Himmelmann (2005: 146) notes that the split between common and personal markers is typical

for determiners. The noun phrase marker sa, on the other hand, behaves somewhat like a

preposition or case marker in that it is used for adjuncts and is an obligatory constituent of

complex prepositions such as para sa 'for' and sulod sa 'inside'.

(58) mupadala si Inday ug sulat sa batang lalaki

AV-send SI Inday UG letter SA young boy
'Inday will send a letter to the boy'

(59) nagdula-dula ang iro sa bola sulod sa balay

AV-play ANG dog SA ball inside SA house
'The dog is playing with the ball inside the house'

However, while authors often use the term 'determiner' to describe ang and to separate it

from sa and ug, there is little discussion as to what the term actually means and what criteria

have been used in the classification. For example, Huang and Tanangkingsing's paper is focused

on repair, and does not provide any analysis for their choice of classification.

25
4.2 Implications of Analyzing Noun Phrase Markers as Determiners

Classifying ang as a determiner references an extended literature on determiners

developed initially in Germanic and Romance languages (Ghomeshi et al. 2009). Determiners

have traditionally been associated with turning a predicate into an argument (Stowell 1989).

However, there is disagreement as to whether the category 'determiner' denotes a syntactic

position, word class, or both, a distinction which has bearing on the analysis of bare nominals

and on the classification of Cebuano NP markers. Ghomeshi et al. (2009) outline three

possibilities:

1. Determiners are not required for argumenthood:

(60) NP

apples

11. Determiners are required for argumenthood, but a null determiner can occupy the

D head. The existence of the DP is dependent on the determiner, be it overt or

null:

(61) DP

D NP
0/~
apples

26
 111. D is a syntactic position that exists separately from the morphological determiner,

and does not require an overt or null determiner to exist. The empty D head alone

can license the bare nominal:

(62) DP

D NP

apples

The three analyses of bare nominals present three ways to interpret why overt determiners are

not necessary in some cases in English. They also provide possible frameworks for assessing the

presence or absence of determiners in Cebuano. If determiners are not always necessary to tum

predicates into arguments, or if empty or null determiners are possible, then it is not necessary to

use the label determiner for ang simply because it occupies a similar linear position to the word

'the'. The label should only be used if ang shares significant characteristics with the determiner

class and words that occupy the D head.

Another property associated with determiners is that of definiteness. Ghomeshi et al. (2009)

note that it is unclear whether definiteness is a property of lexical entries or the D head itself. If

it is a property of lexical entries, different words may carry different values for definiteness. If,

however, it is a property of the syntactic position, it is less clear from where variation would

stem.

While turning predicates into arguments and denoting definiteness are the two major

features associated with determiners, there are other features that determiners may encode. For

example, in German, determiners are marked for gender, number, and case (Wiltschko 2009). It

is therefore possible that all noun phrase markers are determiners and have case-marking

properties as well as determiner properties. If all noun phrase markers are D heads, then ang, sa,

27
,
and ug would be glossed as determiners with a case feature, while if only ang is a D head, sa

could be a preposition or a case marker, and ang would not necessarily carry a case feature. To

determine whether noun phrase markers are indeed determiners or whether ang alone is a

determiner, it is necessary to look at the argument-creating function of the particles as well as

their impact on the understanding of the definiteness of the nouns they mark.

4.3 Analysis of Relevant Data from Cebuano

In most cases, it seems that a noun phrase marker such as ang (or sometimes sa) is

necessary to tum a noun phrase into an argument in a sentence.

(63) *mga iro nipaak sa tawo

PL dog AV-bite SA man
'Some dogs bit the man'

(64) *ubang mga iro nipaak sa tawo

some PL dog AV-bite SA man
'Some dogs bit the man'

(65) ang mga iro nipaak sa tawo

ANG PL dog AV-bite SA man
'The dogs bit the man'

(66) ang ubang mga iro nipaak sa tawo

ANG some PL dog AV-bite SA man
'Some dogs bit the man'

In these cases, ang marks a definite noun phrase, and sentences without an ang-marked noun

phrase are ungrammatical. While the notion of definiteness is not entirely well defined, many

agree that it encompasses a sense of familiarity with the referent and uniqueness of that referent

or group (Paul 2009: 221).

28
,
However, there are some cases where ang marks a non-definite noun phrase, for example,

in its use in sentences that contain generics."

(67) ang mga liyon mukaon ug kame

ANG PL lion AV-eat UG meat
'Lions eat meat'

There are some cases where ang is not necessary to turn a noun phrase into an argument,

for example with demonstratives,"

(68) kining liyon mukatkat sa kahoy

this lion AV-climb SA tree
'This lion climbed the tree'

Much like the bare nominals in English, there are several possibilities for this data. If

demonstratives are not determiners, then the same possibilities that apply to English bare

nominals apply here; that is, either determiners are not necessary to tum predicates into

arguments, there is an empty D head, or there is a null determiner in the D head. The other

possibility is that demonstratives do occupy the D head in the DP, and themselves tum the noun

phrases into arguments.

However, if demonstratives occupy the D head, one would expect that they could not co-

occur with other determiners. In English, for example, '*the this lion' is ungrammaticaL This is

not the case in Cebuano. In fact, ang and the demonstrative can co-occur.

(69) ang kining mga liyon mukatkat sa kahoy

ANG these PL lion AV-climb SA tree
'These lions climbed trees'

Since both can alone change the NP into an argument, they each must be associated with

whatever structure or feature does this. Yet since both can occur together, they do not both

6 It is unclear, though, whether generics are actually non-definite. In this case, there seems to be no particular
familiar referent. However, in Romance languages, definite determiners accompany plural generics, so this type of
result is not limited to Austronesian languages.
7 However, in these cases, the demonstrative appears in the 'ang form' and not the sa or kang form.

29
" 8
occupy the D head. This suggests that there is some sort of null or empty determiner that turns

predicates without an overt determiner into arguments, or that no determiner is necessary to tum

noun phrases into arguments. But in (64) it is clear that there are some cases in which ang is

necessary to obtain a grammatical NP. Therefore, it is most likely that there is a DP structure

introduced or licensed by some but not all noun modifiers, and that this DP structure, occupied or

not, turns the NPs into arguments. As to which word actually occupies the D head, ang is the

best candidate, since it is the leftmost word in the sentence. Demonstratives would occupy lower

XP positions which carry or license DP structure.

(70) DP

D NP
I~
ang mga liyon
ang occupies D and turns the phrase into an argument

(71) DP

~
D XP

X NP
I ~
kining liyon
kining introduces or licenses DP structure which turns
the phrase into an argument

(72) DP

~
D ZP

ang
I Z NP
I ~
ubang mga liyon
ang occupies D and turns the phrase into an argument

8 In this paper, I do not address the possibility of multiple stacked DPs.

30
 (73) ZP

Z NP
I /~
*ubang mga liyon
ubang does not introduce or license DP structure, thus
without ang, the phrase is ungrammatical

Yet if kining can occupy an XP which introduces DP structure, it is also possible that ang

does the same, since it does not possess all of the definiteness characteristics usually associated

with determiners. For example, it marks generics. In this situation, ang would not be a

determiner, but would introduce or license structure necessary to tum a predicate into an

argument, much like kining.

(74)

D
.------, DP

YP

Y XP
I NP
ang X
I ~"
kining mga liyon

Analyzing ang as introducing or occupying a D head raises questions about sa and ug,

often considered to form a noun phrase marker paradigm. While sa possesses definiteness

characteristics distinct from ug, it has a different distribution from ang and is not only associated

with arguments of the verb, but also with adjuncts and prepositions. These differences are

discussed in more detail below.

31
 The definiteness distinction between sa and ug is clear and repeatedly attested. Like ang,

sa marks mainly definite nouns with known referents, while ug only marks indefinite nouns.

(75) nagdala ug sikbut ang batang lalaki

AV-carry DG net ANG young boy
'The boy was carrying a/*the net'

(76) nagdala sa sikbut ang batang lalaki

AV-carry SA net ANG young boy
'The boy was carrying the/*a net'

While there is generally only one ang-marked noun phrase in a sentence, there can be several sa-

marked noun phrases, one sometimes appearing to function as an argument while others are

adjuncts or prepositions. For example, in (77), sa marks 'the letter', which is taken to be an

argument of 'Inday sent the letter to Perla'. In (78), sa marks not only the indirect object 'to the

boy', but also marks the location of the boy 'in the city', and a characteristic of the city, 'the city

ofCebu'.

(77) padalhan ni Inday si Perla sa sulat

send-NAY N1 Inday S1 Perla SA letter
'Inday sent the letter to Perla'

(78) mupadala ang babayi ug sulat ngadto sa lalaki sa siyudad sa Sugbu

AV-send ANG woman UG letter toward SA boy SA city SA Cebu
'The woman will send a letter to the boy in Cebu City'

However, Himme1mann (2005: 147) notes that the variety of functions of a sa-like marker is a

common feature of Philippine languages, and that there is cause for doubt about whether

Cebuano even has a core-peripheral argument distinction.

While the question of whether ang, sa, and ug are determiners is one that has been

debated in the literature, one must ask what bearing the answer to this question has on the

understanding of the rest of Cebuano sentence structure. While other languages have agreement

between verbs and noun phrases based on person, number, gender, or word order, Cebuano

possesses none of this. Agreement is based solely on the interaction between the verb

32
morphology and the noun phrase markers. Thus in order to label the noun phrase markers

determiners, this agreement must be able to be obtained within the determiner modeL

If, following Huang and Tanangkingsing (2005), ang is a determiner while sa and ug are

case markers, this agreement could be obtained by noting that the verb agrees with noun phrases

marked with a determiner. This recalls the topic marker classification, where ang marks topic,

and the verb agrees with topic marked noun phrases.

It is also possible that ang, sa, and ug as well as the personal name markers si, ni, and

kang do form a paradigm. This is a tempting analysis, since ang and sa do share some

definiteness characteristics, and since demonstratives and pronouns have three forms that

correspond to ang, sa, and kang (since names cannot be indefinite). It seems unlikely that the

determiner characteristic alone would be responsible for driving agreement with the verb. Yet

there is no reason why ang cannot be marked for case like sa and ug. As discussed above,

German determiners are marked for case, and in Cebuano it is possible that agreement is case

based.

4.4 Tagalog Noun Phrase Markers as Determiners

In Tagalog, since ang-marked NPs have definiteness restrictions that are not usually

associated with subjects, one proposal is that ang is actually a determiner, and definiteness

restrictions stem from its determiner status, rather than its characterization as a case or topic

marker.

Himmelmann (2006: 2) notes that the markers form a paradigm, all non-pronominal

argument and adjunct expressions must have one of the markers, and personal pronouns and

demonstratives, which are not marked, have three different forms whose distribution is roughly

equivalent to the distribution of the markers, differing from the common markers because of

33
r
their required definiteness. He claims that the markers are actually syntactic heads and that sa

heads prepositional phrases while ang and ng head determiner phrases. In classifying sa as a

preposition, he notes that adjuncts are introduced with sa and that it marks many non-subject

arguments, including beneficiaries and recipients. He points out that sa is different from

European prepositions in that it is the only preposition in Tagalog, and is thus obligatory in all

prepositional phrases. In order to express different meanings, sa is combined with a specifier.

Additionally, while in English, the complement ofa preposition is a DP, this is not true in

Tagalog ifDPs are considered to be ang or ng marked phrases.

The assumption that ang, ng, and sa form a paradigm breaks down because while ang

and ng are in complementary distribution, they each can immediately precede a phrase marked

by sa (Himmelmann 2006: 4) and there are no phrases where sa immediately precedes ang or ng.

(79) yamang ang sa pagong ay tumubo' hanggang sa magbunga

while SPEC Loc turtle PRED AV- until Loc AV-flower
MRK growth
,... while that of the turtle grew until it bore fruit'

He states that ang and ng must be in c-commanding position over sa, since their phrase

structural position is always 'higher' and no empty categories are assumed (Himmelmann 2006).

Both ang and ng occur at the left edge of the phrase and can be replaced by demonstratives. Yet

while ang marks subject phrases, topic phrases, and predicates, ng marks non-subject

complements (where it alternates with sa) and possessors. Since the syntactic distribution is

determined by the marker, Himmelmann concludes that they are heads of their phrases. Yet

unlike demonstratives, they are not distributional equivalents of their phrases because they

cannot stand alone (Himmelmann 2006: 7). However, this is most likely a result of their clitic

status, rather than their syntactic structure.

34
 Himmelmann notes that sa is widely accepted as a preposition which allows for a

specifier, and that the category of ang and ng is more disputed. He analyzes ang as a specific

article and ng as its genitive form, and considers them determiners heading DPs. There is some

support for this in that phrases with both ang and a demonstrative are considered ungrammatical

in all but informal registers. This suggests that demonstratives occupy the same position as ang

and ng. However, there is also evidence that there are multiple types of demonstratives,

including adjuncts that occur in rightmost position accompanying another non-adjunct

demonstrative. Therefore, while there is evidence suggesting similarities between determiners,

demonstratives, and ang and ng, there are also issues with the analysis, and it does not entirely

account for the voice alternations that are prevalent in the language.

4.5 Interim Summary

Due to the varying possibilities of the structure of determiners, it is difficult to tell

whether ang is a D head or whether it simply introduces or licenses DP structure. If ang is a

determiner, deeper analysis reveals one of two possibilities: if ang is the only determiner, that is,

the only noun phrase marker to introduce DP structure, then it is possible that ang-marked NPs

are the only real arguments in the sentence, and that sa- and ug-marked NPs are all adjuncts.

Determining agreement would be simple: the verb simply agrees with the only DP in the

sentence. This would require positing that all Cebuano verbs are intransitive. However, if ang,

sa, and ug are all determiners, there must be case marking on these determiners, and agreement

with the verb is based on case.

The ability of ang to tum noun phrases into arguments points toward it being a

determiner, but questions still exist due to ambiguities about definiteness. The ability of

demonstratives both to tum noun phrases into arguments and to co-occur with ang points to the

35
existence of null or empty D heads in Cebuano, and thus whether ang occupies the D head or

merely introduces or licenses a functional D head is unclear.

5 Noun Phrase Markers as Case Markers

5.1 Previous Analysis in Cebuano

While no Western Austronesian language has case affixes, Himmelmann (2005) notes

that languages like Cebuano have phrase-marking clitics which are often called case markers. In

his reference grammar, Tanangkingsing (2009: 105, 107) classifies noun phrase markers as case

markers, and claims that every noun phrase takes a case marker, though he notes that ang can be

omitted in colloquial speech. He analyzes 'actor focus' constructions as intransitive, and 'object

focus' constructions as transitive, thus determining that Cebuano is morphologically ergative:

actors in 'actor focus' clauses and patients in 'object focus' clauses take the same case marker, si

or ang, while agents in 'object focus' clauses take a separate maker, ni or sa. Patients in 'actor

focus' clauses are taken to be extended arguments and are marked with kang, sa, or ug

(Tanangkingsing 2009: 106). Rather than use the terms 'ergative' and 'absolutive' however, he

chooses to use the term 'nominative' for the 'actor focus' construction and the patient of the

'object focus' construction, and the term 'genitive' for the agent of the 'object focus'

construction. Below is his table of case markers.

T a bl e SIC
- : eb uano N oun Ph rase M ark ers (T ananziki n2S1112 2009)
NEUTRAL NOM GEN DAT EXT LOC
PERSONAL si ni kan~ - -
=y
COMMON ang sa - ug sa

Tanangkingsing (2009: 107) describes nominative case markers as marking nominative

nominal no matter what their semantic role: agent in actor voice clauses and patient in patient

voice clauses In addition, he notes that topicalized nominals are also marked with ang or si. 9

9 The following example uses Tanangkingsing's (2009) glossing system.

36
 (80) ang iya-nga dog gi-uyog ang tree
ANG 3SG.POSS-LK dog PFV.PV-shake ANG tree
'His dog, (it) shook the tree'

Bunye and Yap (1971: 38) also classify the noun phrase markers as case markers, but

rather than use terms such as nominative or ergative, they term ang and si 'topic', sa and ni

'agentive', and sa, ug, and kang 'oblique'. They note that the topic always indicates the subject

or focus of a construction, and that the agentive and oblique have varying roles depending on the

construction. Their notes on topic constructions are similar to Ghazali (1990). They note that

the agentive marker indicates source of action, possession, and location. 10

(81) gikaon sa amahan ang ubas

OBJ-ate AGT father TOP grapes
'The father ate the grapes'

(82) dako ang balay sa babaye

big TOP house AGT woman
'The woman's house is big'

(83) mularga si Ginny sa Pilipinas

ACT-leave TOP Ginny AGT Philippines
'Ginny is leaving for the Philippines'

The oblique markers have several uses including goal, beneficiary, instrument, and occasionally

actor.

(84) guhaton nita ang siya kang Ramon

make-OBJ AGT-3PL TOP chair OBL Ramon
'They'll make the chair for Ramon'

(85) giabri nita ang pultahan sa akong yawe

OBJ-open AGT -3PL TOP door OBL my key
'They opened the door with my key'

(86) kaonon ug bata ang saging

eat-OBJ OBL child TOP banana
'A child will eat the banana'

10 The following examples use Bunye & Yap's (1971) glossing system.

37
The marker sa may also be preceded by an additional preposition or auxiliary word when

referring to recipients of certain actions.

(87) putlon ni Undo ang kahoy alang sa iyang amigo

cut-OBJ AGT Undo TOP tree for OBL his friend
'Undo will cut the tree for his friend'

Bunye and Yap make a couple of other interesting observations. First, when there is a

sequence of constituents marked with sa, the first (the one that follows the verb) is oblique, while

all subsequent ones are agentives.

(88) maghinpyo sila sa mga muebles sa sala sa amahan

ACT-clean TOP- OBL PL furniture AGT living AGT father
3PL room
'They'll clean the furniture in the living room of the father'

They also note that sa marks definite nouns while ug marks indefinite nouns.

(89) mangaon kami ug isda

ACT-eat TOP-1PL OBL fish
'We'll eat fish'

(90) mangaon kami sa isda

ACT-eat TOP-1PL OBL fish
'We'll eat the fish'

However, Bunye and Yap use the terms topic and agentive as types of case, and classify

the noun phrase markers as case markers. Even Ghazali (1990: 55), who uses the term topic,

equates topic with "phrases in nominative nominal position", and notes that topics are always in

nominative case. She also uses the terms 'genitive' and 'oblique' to describe the other noun

phrase markers, effectively using the same terms as the case-marking analysis of Cebuano does.

There is a significant amount of terminology overlap between the case-marking and topic-

marking theories. Thus it is necessary to examine what would be expected of noun phrase

markers that mark case, and compare these expectations to the evidence present in Cebuano.

38
5.2 Implications of Analyzing Noun Phrase Markers as Case Markers

The labeling of Cebuano noun phrase markers as case markers references a long tradition

of case in linguistic theory. While according to Butt (2006), there is no well-defined notion of

case, there is a sense that it marks the relationships between words in a sentence. Blake (200 I)

defines case as "a system of marking dependent nouns for the type of relationship they bear to

their heads." Two main theories of case are structural case within Government-Binding (GB)

and the Minimalist Program (MP) and linking theories of case within Lexical-Functional

Grammar (LFG).

Structural case within GB/MP posits that case assignment depends on the structure of the

verb phrase and the assignment of theta roles. Only abstract case (nominative and accusative

case) is assigned by the verb, while other cases are termed 'inherent' and are not assigned by the

verb. However, all NPs must have Case in order to appear in a sentence. Overt case marking is

considered to be part ofPF, just a matter of spell out of abstract case features (Butt 2006).

Linking theories such as LFG, on the other hand, look at case as a mapping between

thematic roles (agent, beneficiary, patient, etc.), argument structure (the thematic roles of a

particular predicate) and grammatical functions (subject, object, second object, oblique,

complement, adjunct). While subject and object are not very selective about thematic roles,

second objects and obliques are more sensitive to theta information (Butt 2006: 121).

Kiparsky (1997) notes that richness of inflection is correlated with freedom of word

order. That is, languages without much morphology have little word order freedom.

Furthermore, languages with little or no case morphology tend to make use of agreement

morphology or positional cues to identify the subject vs. object (Butt 2006: 109).

39
 Another facet to case-marking, other than how it is assigned, is the question of alignment,

which refers to how case marking is grouped relative to transitive and intransitive verbs.

nominative-accusative alignment marks subjects of intransitive clauses (S) and subjects of

transitive clauses (A) with nominative case, and objects of transitive clauses with accusative

case. By contrast, ergative-absolutive alignment marks Sand 0 with absolutive case and A with

ergative case. In many languages, the absolutive or nominative cases are unmarked

morphologically. An analysis of Cebuano noun phrase markers as case markers must also

hypothesize in which alignment system the language falls.

Based on Kiparsky' s observations, if Cebuano lacked overt case markers, rigid word

order or agreement morphology would be expected to compensate for this. Working within

GB/MP, structure would be expected to determine case, while in LFG, case would be assigned as

a mapping from argument structure to grammatical functions. Additionally, while it is possible

for topics to have case, they do not receive case by virtue of being topics. Thus a case marking

analysis would result in some notion of subject in Cebuano, which might then raise to topic

position. That subject would either be marked with nominative case, or with ergative or

absolutive case depending on the transitivity of the verb.

5.3 Analysis of Relevant Data from Cebuano

Since Cebuano lacks rigid word order or person/number/gender agreement on verbs,

thematic roles of arguments are determined by the interaction between the verb and noun phrase

markers only.

40
 (91) nipaak ang iro sa tawo
AV-bite ANG dog SA man
'The dog bit the man'

(92) nipaak sa iro ang tawo

AV-bite SA dog ANG man
'The man bit the dog'

This suggests that the noun phrase markers are either overt case markers, for example, that ang

marks nominative and sa marks accusative case, or that there is an entirely different system of

agreement occurring in Cebuano, for example topic agreement.

At the very least, it seems that non-argument noun-phrase markers mark inherent case,

such as location. The prepositional nature of these constructions has led some to label them

prepositions rather than case markers. However, they also are present even when a lexical

preposition is used, suggesting they mark the theta role of the noun phrase.

(93) nagdula-dula ang iro sa bola sulod sa balay sa babayi

AV-play ANG dog SA ball inside SA house SA woman
'The dog is playing with the ball inside the woman's house'

In this example, sa also marks possession, which is typical of the genitive case.

In order to believe that the noun phrase markers mark case as opposed to a different sort

of verbal agreement like topic, Cebuano would be expected to have subjects. However, ang-

marked nouns, the best candidate for subjects, have some characteristics that are strange for

subjects, for example the requirement of definiteness. Additionally, due to the flexibility of

Cebuano sentence structure, it is often difficult to find diagnostics for subjecthood.

Himmelmann (2005) lists a few that are often used in determining subjecthood in Western

Austronesian languages, such as quantifier float, and relativization.

Himmelmann notes that floated quantifiers agree exclusively with the ang-marked noun

phrase, and determines that it is therefore the subject. In (94) and (95), the quantifier modifies

41
 the noun phrase of which it is a part. However, when the quantifier is floated, as in (96), the

quantifier modifies only the ang-marked noun phrase.

.\t~'
(94) himuon sa mga bata ang tanang mga sulat
f
do-NAY SA PL child ANG all PL letter
'All the letters will be written by the children'

(95) himuon sa tanang mga bata ang mga sulat

do-NAY SA all PL child ANG PL letter
'The letters will be written by all the children'

(96) himuon tanan sa mga bata ang mga sulat

do-NAY all SA PL child ANG PL letter
'All the letters will be written by the children'
*'The letters will be written by all the children'

Relative clauses also point toward the existence of subjects in Cebuano, since only

subjects can be re1ativized; that is, the head to which the relative clause is attached must be the

subject of the relative clause (Himmelmann 2005: 161).

(97) ang bata nga gilabay niya sa isda

ANG child NGA NAV-throw 3SG SA fish
'The child that he threw the fish to'

(98) ang bata nga naglabay niya sa isda

ANG child NGA AV-throw 3SG SA fish
'The child who threw him the fish'

In the above two examples, ang bata can be taken to be the subject in the relative clause.

In both cases, the verb within the relative clause agrees with the ang-marked noun phrase. In

example (97), the non-actor voice verb agrees with ang bata and the child is the recipient, while

in (98), the actor voice verb agrees with ang bata and the child is the actor. However, a relative

clause where the verb does not agree with the head of the relative clause is ungrammatical:

(99) *ang bata nga gilabay niya ang isda

ANG child NGA GI-throw 3SG ANG fish
*'The child who the fish was thrown to him by'

42
 In this case, ang isda is the subject of the relative clause and agrees with the non-actor voice

verb. However the sentence is ungrammatical because the verb must agree with the head of the

t
';
relative clause. Himmelmann claims this property points to the existence of subjects in Cebuano.

While some tests may point to ang-marked phrases as being subjects, subject itself is a

vague term, and one must question whether subjecthood merely consists of having a certain

number or the right subset of properties. For this reason, some use the term 'syntactic pivot'

rather than subject for ang-marked noun phrases in Austronesian languages (Himmelmann 2005:

156).

Whether or not Cebuano has subjects, if the noun phrase markers are case markers, the

question still remains as to what case they mark. One possibility is that Cebuano follows an

English-like nominative-accusative case marking pattern. Under this schema, ang and si mark

nominative subject while sa, ug, and ni mark accusative object.

(l00) nipaak ang iro sa tawo

AV-bite ANG dog SA man
VERB-TRANS NOM/SUBJ ACC/OBJ
'The dog bit the man'

Sentences with non-actor voice verbs are taken to be passive constructions.

(l 01) ang tawo gipaak sa iro

ANG man NAV-bite SA dog
NOM/SUBJ VERB-TRANS ADJUNCT
'The man was bitten by the dog' /'The dog bit the man'

However, under this analysis, all non-actor voice sentences are assumed to be passives. Cross-

linguistically, passives show defocusing of agents, low text frequency, and distinct word order

from the active construction. Tanangkingsing & Huang (2007) note that clauses with gi- verbs

do not show a tendency toward agent omission and that these clauses account for 51% of clauses

in narratives, too frequent for passives. This suggests that ang does not mark nominative case.

43
,
Another possibility is that ang and si mark abso1utive case while ni and sa mark ergative

case. This requires assuming that actor voice clauses are actually intransitive, this marking their

intransitive subject with absolutive case. Object voice clauses are then taken to be transitive, and

the agent ofthe transitive clause is thus marked with ergative case while the object takes

abso1utive case. In actor voice intransitive clauses, then, patients are not core arguments, but

adjuncts marked with ni, kang, sa, or ug.

(102) nipaak ang iro sa tawo

AV-bite ANG dog SA man
VERB-INTR ABS/SUBJ ADJUNCT
'The dog bit the man'

(103) ang tawo gipaak sa iro

ANG man NAV-bite SA dog
ABS/OBJ VERB-TRANS ERG/SUBJ
'The man was bitten by the dog' /'The dog bit the man'

Another analysis is that actor voice clauses are antipassives that maintain their patients.

However, much like the passive analysis above, it would be strange for a language to have such a

high frequency of antipassives. Again, one of the problems in determining alignment is that the

phrase marking paradigms "rarely, if ever, provide clear cut evidence for distinguishing core

arguments from peripheral arguments (or adjuncts)." (Himmelmann 2005: 147) Since

determining alignment requires an understanding of which constructions are transitive and which

are intransitive, this issue makes definitive analysis difficult.

However, the existence of a phrase marker paradigm that seems to apply to

demonstratives and pronouns suggests that there is similarity among the noun phrase markers,

which is most easily explained by positing that the markers all share a single feature, such as

case.

44
 - : Cebuano P ronouns
T a bl e 52 (T ananaikingsmg 2009)
si-like ni-like kang-like
lSG ako nako kanako
2SG ikaw nnno kanimo
3SG srya mya kaniya
IPL-EXCL kami namo kanamo
IPL-INCL kita nato kanato
2PL kamo nmyo kaninyo
3PL sila nila kanila

One issue with this analysis is the well-defined definiteness distinction between sa and ug

in (89) and (90), where the use of one or the other changes only definiteness and has no influence

on thematic roles. Some possible solutions are that sa is a determiner that carries case, while ug

is only a case marker, or that sa-marked NPs are arguments while ug-marked NPs are adjuncts.

Unfortunately, both of these possibilities contradict other observations about the determiner-like

distinctions between ang and sa and the argument structure similarities between sa and ug.

The above issues have led some to suggest that Western Austronesian alignment is

wholly different from either nominative-accusative or ergative-absolutive. This, though, is an

unsatisfactory analysis, since one would hope theories could be applied cross-linguistically.

However, if the mapping theory ofLFG is employed, it might be conceivable that actor voice

verbs map agents to subjects, while object voice verbs map patients to subjects, allowing the

LFG account of case to apply both to languages like English and languages like Cebuano.

5.4 Tagalog Noun Phrase Markers as Case Markers

The analysis of noun phrase markers as case markers is also prevalent in Tagalog.

Kroeger (1993a) views the noun phrase markers as case marking clitics which in standard

Tagalog orthography are written as separate words. He classifies them along traditional lines as

nominative, genitive, and dative case markers. Common markers mark all common nouns and

place names, while Personal markers mark personal names.

45
 - :T
T a bIe 53 Noun Phras e Markers (Kroeger 1993a: 13)
agaIo~
NOM GEN DAT
Common ang ng sa
Personal si ni kay

He notes that each verbal clause must contain only one nominative argument and that the voice

marker picks out the role of the nominative argument. Thus, the nominative argument is said to

be 'in focus' (in a different sense from the traditional pragmatic one). Kroeger also points out

that while the nominative argument is usually definite, it can also be generic, or even indefinite if

it is marked by a numeral quantifier.

Rackowski and Richards (2005) suggest that ang and ng are not case markers exactly,

though they carry case information, but that nominative case is checked and valued by T while

accusative case is checked and valued by v, similar to what occurs in English. They characterize

ang as a DP marker which agrees with the verb and copies its case feature to the verb. 11

(104) b-in-ili ng bata ang tela

ASP-Ace-buy CASE child ANG cloth
'The child bought the cloth'

Thus ang-marked DPs undergo case agreement with v and are the highest DP in the vP phase

edge. This conclusion is derived from their notion of locality, and results in the restriction on

extraction to ang-marked noun phrases. Since the ang-marked DP is the highest in vP, it is the

only one that is able to be extracted.

Aldridge (2006) also claims the words are case markers, but differs in the analysis of the

type. She believes that Tagalog is an ergative language, and that ang marks absolutive, rather

than nominative case. She disagrees with Rackowski and Richards' (2005) case agreement

analysis, and argues that the verbal affix is not a copied case feature, but a marker of transitivity.

11 This example is glossed according to Rackowski & Richards (2005).

46
 Aldridge shows that plural agreement does not occur between all ang-marked NPs and

the verb. It occurs in 'nominative agreement' contexts where the agent is marked with ang, but

not in 'accusative agreement' contexts where the patient is marked with ang. 12

(105) nag-si-basa ang mga bata ng liham

NOM.AsP-PL-read ANG PL child CASE letter
'The children read a letter'

(106) "si-ni-basa ng bata ang mga liham

Pt-Acc.Asr-rcad CASE child ANG PL letter
*'The child read the letters'

To account for this distribution, Aldridge proposes that 'nominative agreement' verbs are

actually intransitive and their v does not carry a structural case feature, while 'accusative

agreement' verbs are transitive and their v does carry a case feature. T then, has an absolutive

case feature when it combines with intransitive v. She supports this by noting that sentences

with only one argument have the same verbal morphology as verbs that agree with an ang-

marked agent. (Nag-, above, and -um-, below, both mark verbs that agree with ang-marked

agents.)

(107) d-um-ating ang babae

-INTR.PERF-arrive ABS woman
'The woman arrived'

Under this analysis, sentences with ang-marked patients are transitive. Their agents receive

ergative case and their patients receive absolutive case. Sentences with ang-marked agents are

intransitive, so their agents receive absolutive case and their patients receive inherent oblique

case.

(108) guma-gamit siya ng lalaki

INTR.PROG-USe 3SG.ABS OBL man
'He uses a man'

12 The following examples are glossed according to Aldridge (2006).

47
 Lastly, Aldridge notes that under her analysis, where transitive objects and intransitive
If
t
\
subjects both carry the same (absolutive) case marker, only absolutive noun phrases can be

extracted, a well-known property of ergative languages. Thus treating the noun phrase markers

as case markers requires, for her, an analysis of the voice system of Tagalog as ergative.

While most of the discussion of ergativity in Tagalog is related to agreement of noun

phrases with the verb and relies on an analysis of the noun phrase markers as morphologically

ergative or absolutive, Manning and Sag (1999) claim that what distinguishes Tagalog and other

Philippine languages is not their morphology, but their syntax. Working within LFG, they argue

that syntactic and morphological ergativity are two separate phenomena. Classifications such as

accusativity and ergativity result from differences in the mapping from argument structure to

grammatical relations. Syntactic accusativity is marked by agents mapping to subjects, while

syntactic ergativity is marked by agents mapping to objects, regardless of the morphology

(Manning 1995). Manning claims that Tagalog and other Western Austronesian languages are

neither syntactically ergative nor syntactically accusative. Instead, the voice system allows for

great flexibility in the mapping of argument structure to grammatical relations, allowing both

agents and patients to retain their core argument status; the corresponding grammatical relations

are determined by the voice system, rather than by demoting a core argument to an oblique

argument (Manning 1995: 14). In this way, he is able to explain why Tagalog does not easily fit

into traditional alignment categories without positing an entirely different structure for Philippine

languages.

48
5.5 Interim Summary

Case marker terminology tends to be the default for describing noun phrase markers, yet

there are many consistency issues with this analysis. While the markers appear to form a

paradigm, they do not pattern convincingly with nominative-accusative or ergative-absolutive

languages. Even the presence of subjects in Cebuano cannot be taken for granted. Some

markers possess characteristics expected of inherent case markers, yet ang can also mark

topicalized noun phrases and verbs.

Unfortunately, several of the tests which suggest a case marking analysis in Tagalog are

not available in Cebuano. For example, there is no plural agreement on the verb in Cebuano, and

this lack of additional agreement between verbs and NPs makes an analysis of transitivity more

difficult in Cebuano. Additionally, the two languages have a different array of noun phrase

markers. While both languages have the ang and sa markers, Tagalog has ng, corresponding to

something like a definite genitive while Cebuano has ug, an indefinite oblique. Thus Cebuano

may have different results in areas such as definiteness.

However, perhaps within the LFG case theory, there is enough flexibility to

accommodate Cebuano noun phrase markers as case markers. The idea that Cebuano and

Tagalog are able to map agents to subject or object without demoting either to an adjunct could

accommodate the facts while positing only a slight modification to an existing cross-linguistic

theory.

49
 6 Summary and Conclusion

£
~ As evidenced by the varying analyses presented in this paper, the characterization of the

voice system and the noun phrase markers is still an open question among linguists. There is

evidence both within Cebuano and cross-linguistically for characterization of the noun phrase

markers as something like a topic marker, something like a determiner, and something like a case

marker.

The classification of ang as a topic marker is appealing because it provides a simple

analysis for why ang-marked noun phrases must be definite and why only ang-marked phrases

can be extracted. It also captures a notion of 'aboutness' associated with ang-marked NPs in a

sentence. However, ang-marking does not correlate with 'aboutness' on the level above the

sentence, as evidenced by its lack of correlation with increased Topic Persistence, and so does

not fit into the traditional category of discourse topic. Nor does it align with traditional

pragmatic topichood, since it can mark both new and given information. Thus characterizing

ang as a topic marker is only appropriate if the notion of topichood is expanded cross-

linguistically to include less strict distinctions between given and new information or a different

sense of the meaning of 'aboutness' .

Analyzing all or some of the noun phrase markers as determiners is useful because it

captures well attested differences in definiteness among noun phrases with different markers.

Since noun phrase markers, especially ang, are almost always required in order to form

grammatical sentences, the cross-linguistic observation that determiners tum predicates into

arguments makes the determiner analysis very appealing. The fact that demonstratives can

sometimes take the place of the noun phrase markers is consistent with evidence from other

languages, such as English. Although the co-occurrence of ang and demonstratives necessitates

50
the assumption of null or empty D heads, this is not unprecedented cross-linguistically.

However, there are some instances where ang marks seemingly non-definite noun phrases,

calling into question its status as a definite determiner. Additionally, ifboth ang and sa are

determiners, further marking would be necessary in order to determine agreement with the verb,

suggesting that perhaps the noun phrase markers bear an additional feature, such as case.

The fact that Cebuano has relatively free word order and no person, number, or gender

agreement suggests that overt case marking is what shows agreement with the verb. Because the

agreement between the noun phrase marking and the verbal morphology determines theta roles

of the noun phrases, it is simple to assume that arguments are marked with overt case.

Furthermore, sa assumes roles typical of inherent case markers, such as marking possession and

location. However, there is an implicit assumption that the 'nominative' case marker will denote

the subject, and the subject properties of ang-marked noun phrases are cloudy at best.

Particularly, the requirement of definiteness in an ang-marked noun phrase is a strange

requirement for a subject. The case marker analysis also does not provide a satisfying account of

the definiteness distinction between sa and ug. Another issue with the case marker analysis is

that within GB/MP, it requires an analysis of the language as nominative-accusative or ergative-

absolutive. This is problematic because both agent voice and patient voice phrases seem to

incorporate both agents and patients as core arguments. This data points away from a

nominative-accusative analysis, which requires extensive use of passives with demoted agents

and also points away from an ergative-absolutive analysis, which requires extensive use of

intransitives with oblique patients. The most compelling analysis is the case-marking analysis

within LFG, which allows for the coexistence of two distinct methods of mapping argument

structure to grammatical functions to case.

51
 The issue of how to classify these markers is further complicated by the various

theoretical frameworks that have been used to examine Cebuano and other Austronesian

languages. The same facts are approached differently by authors who subscribe to different

theories, yielding various essentially similar analyses with varying terminology as well as

fundamentally different analyses that use similar terminology. Because their analyses are bound

up in the theories they employ, authors often argue parallel to each other, some in the same

direction and some in opposite directions, complicating the literature. Furthermore, even within

a single theory, there is little consensus as to how each type of marker functions cross-

linguistically. Since commonly used diagnostics in one language family often do not apply to

another, each author decides which properties he or she will use to characterize each possibility,

and then determines which classification most closely matches the set of properties displayed by

Cebuano noun phrase markers. Overlaps in the sets of properties used to distinguish topic

markers, determiners, and case markers lead to conflicting classifications based on the same

I observations.

I
~.
Based on my observations, the most compelling analysis is that the noun phrase markers

f ang and sa license D heads and ang, sa, and ug bear case features. The best analysis of how case
if is assigned is Manning & Sag's (1999) modification ofLFG mapping rules to allow Cebuano to

f map agents to subjects or agents to objects, depending on the voice marker. This analysis

t1 requires neither the extensive use of passives or intransitives nor does it posit a completely
~
different voice system for Western Austronesian languages. The combination of the case marker

analysis with the determiner analysis could also help explain the strange definiteness

requirements of subjects in Cebuano.

52
 While the precise analysis of the noun phrase markers in Cebuano is still unclear, the

close examination in this paper highlights the usefulness of Western Austronesian languages as a

testing ground for theories of noun phrases and argument structure. The questions posed by

these languages offer useful insight into possibilities that should be considered when formulating

theories of topic markers, determiners, or case markers. Rather than claiming that a language

like Cebuano is fundamentally different from a language like English, it makes sense to

reexamine the facts and determine if there is a way of incorporating both sets of evidence under a

single analysis.

53
References

Adelaar, Alexander, & Nikolaus P. Himmelmann (eds.). 2005. The Austronesian languages of
Asia and Madagascar. LondonlNew York: Routledge.

Aldridge, Edith. 2004. Ergativity and word order in Austronesian languages. Ithaca, New York:
Cornell University dissertation.

Aldridge, Edith. 2006. Absolutive Case in Tagalog. Chicago Linguistic Society (CLS) 42(2). 1-
15.

Billings, Loren A. 2005. Ordering c1itics and postverbal R-expressions in Tagalog. In Andrew
Carnie, Heidi Harley, & Sheila Ann Dooley (eds.), Verbfirst: On the syntax ofverb-initial
languages, 303-339. Amsterdam/Philadelphia: John Benjamins Publishing Company.

Blake, Barry J. 2001. Case. Cambridge, United Kingdom: Cambridge University Press.

Blake, Frank R. 1905. The Bisayan Dialects. Journal ofthe American Oriental Society 26. 120-
136. http://www.jstor.org/stable/592885 (15 October, 2010).

Bunye, Maria Victoria R. & Elsa Paula Yap. 1971. Cebuano grammar notes. Honolulu:
University of Hawaii Press.

Butt, Miriam. 2006. Theories ofCase. Cambridge, United Kingdom: Cambridge University
Press.

Cooreman, Anne M., Barbara Fox, & Talmy Given. 1984 The discourse definition of ergativity.
Studies in Language 8(1). 1-34.

Crowley, Terry. 2008. Austronesian Languages. In Keith Brown & Sarah Ogilvie (eds.), Concise
encyclopedia oflanguages ofthe world, 96-105. Amsterdam: Elsevier.

Dixon, Robert M. W. 1994. Ergativity. Cambridge: Cambridge University Press.

Ghazali, Kamila. 1990. Nominative nominals and focus reconstruction in Cebuano. Philippine
Journal ofLinguistics 21(2).51-56.

Ghomeshi, Jila, Ileana Paul, & Martina Wiltschko (eds.). 2009. Determiners: Universals and
variation. Amsterdam/Phi1adelphia: John Benjamins Publishing Company.

Guilfoyle, Eithne, Henrietta Hung, & Lisa Travis. 1992. Spec ofIP and Spec ofVP: Two
subjects in Austronesian languages. Natural Language and Linguistic Theory 10. 375-414.

54
Himmelmann, Nikolaus P. 2005. The Austronesian languages of Asia and Madagascar:
Typological Characteristics. In Alexander Adelaar & Nikolaus P. Himmelmann (eds.), The
Austronesian Languages ofAsia and Madagascar, 110-181. London/New York: Routledge.

Himmelmann, Nikolaus P. 2006. Notes on noun phrase structure in Tagalog. Tenth International
Conference on Austronesian Linguistics (ICAL). January 2006. Puerto Princesa City,
Palawan, Philippines.

Huang, Huei-ju & Michael Tanangkingsing. 2005. Repair in verb-initial languages. Language
and Linguistics 6(4).575-597.

Kiparsky, Paul. 1997. The Rise of Positional Licensing. In Ans van Kemenade & Nigel Vincent
(eds.), Parameters ofmorphosyntactic change, 460-494. Cambridge: Cambridge University
Press.

Kroeger, Paul R. 1993a. Phrase structure and grammatical relations in Tagalog. Stanford,
California: Center for the Study of Language and Information.

Kroeger, Paul R. 1993b. Another look at subjecthood in Tagalog. Philippine Journal of

Linguistics 24(2). 1-16.

Lambrecht, Frans Hubert. 1978. Ifugaw-English dictionary. Baguio City: The Catholic Vicar
Apostolic of the Mountain Province.

Levinson, Stephen C. 1983. Pragmatics (Cambridge Textbooks in Linguistics). Cambridge

University Press.

Lewis, M. Paul. 2009. Ethnologue: Languages ofthe world, Sixteenth edition. Dallas, Texas: SIL
International. http://www.ethnologue.com/show_language.asp?code=ceb (14 March, 2011).

Manning, Christopher D. & Ivan A. Sag. 1999. Dissociations between argument structure and
grammatical relations. In Gert Webelhuth, Jean-Pierre Koenig, & Andreas Kathol (eds.),
Lexical and constructional aspects oflinguistic explanation, 63-78. Stanford, California:
CSLI Publications.

Manning, Christopher D. 1995. Ergativity: Argument structure and grammatical relations.

Linguistic Society ofAmerica (LSA). January 1995. New Orleans, Louisiana.

Nakanishi, Kimiko. 2001. Prosody and information structure in Japanese: A case study of topic
marker wa. In Noriko Akatsuka & Susan Strauss (eds.), Japanese / Korean Linguistics10,
CSLI: Stanford,434-447.

Paul, Ileana. 2009. On the presence versus absence of determiners in Malagasy. In lila
Ghomeshi, Ileana Paul, & Martina Wiltschko (eds.), Determiners: Universals and variation,
215-241. Amsterdam/Philadelphia: John Benjamins Publishing Company.

55
Rackowski, Andrea & Norvin Richards. 2005. Phase edge and extraction: A Tagalog case study.
Linguistic Inquiry 36(4).565-599.

Reid, Lawrence Andrew. 2002. Determiners, Nouns, or What? Problems in the Analysis of Some
Commonly Occurring Forms in Philippine Languages. Oceanic Linguistics 41(2).295-309.

Richards, Norvin. 2000. Another look at Tagalog subjects. In Ileana Paul, Vivianne Phillips, &
Lisa Travis (eds.), Formal issues in Austronesian linguistics, 105-116. Dordrecht: Kluwer.

Schachter, Paul & Fe T. Otanes. 1972. Tagalog reference grammar. Berkeley, Los Angeles &
London: University of California Press.

Stowell, Timothy. 1989. Subjects, specifiers and X-bar theory. In Mark Reuben Baltin &
Anthony S. Kroch (eds.), Alternative conceptions ofphrase structure, 232-262. New York:
Academic Press.

Talmy, Leonard. 1985. Lexicalization patterns: Semantic structure in lexical forms. In Timothy
Shopen (ed.), Language typology and syntactic description: Grammatical categories and
the lexicon, 57-149. Cambridge, England: Cambridge University Press.

Tanangkingsing, Michael & Shuanfan Huang. 2007. Cebuano Passives Revisited. Oceanic
Linguistics 46. 554-584.

Tanangkingsing, Michael. 2009. A Functional Reference Grammar ofCebuano. Taipei, Taiwan:

National Taiwan University dissertation.

Vanoverbergh, Morice. 1955. Iloko grammar. Baguio City: Catholic School Press

Walters, Dennis. 1994. Discourse-based evidence for an ergative analysis ofCebuano. UTA
Working Papers in Linguistics 1. 127-140.

Wiltschko, Martina. 2009. What's in a determiner and how did it get there? In Jila Ghomeshi,
Ileana Paul, & Martina Wiltschko (eds.), Determiners: Universals and variation, 25-66.
AmsterdamJPhiladelphia: John Benjamins Publishing Company.

Wolff, John U. 2008. Cebuano. In Keith Brown & Sarah Ogilvie (eds.), Concise encyclopedia of
languages ofthe world, 197-199. Amsterdam: Elsevier.

Zwicky, Arnold M. 1977. On clitics. Bloomington, IN: Indiana University Linguistics Club.

56
Appendix
Table of Abbreviations
ISG first person singular
2SG second person singular
3SG third person singular
IPL-EXCL first person plural exclusive
IPL-INCL first person plural inclusive
2PL second person plural
3PL third person plural
A agent-like argument of canonical transitive verb
ABS absolutive
ACC accusative
-ACC- accusative verb marker; indicates ang marks patient
ACT- active verb marker; indicates ang marks agent
AGT agent
ANTIP- antipassive
ASP- aspect
AV- actor voice; indicates ang marks agent
DAT dative
DET determiner
ERG ergative
EXT extension
GEN genitive
INS- instrumental verb marker; indicates ang marks instrument
INTR intransitive
IV- instrument voice; indicates ang marks instrument
KA numeral connector; indicates units of
LK linker
LOC locative
-LOC locative verb marker; indicates ang marks location
LV- locative voice; indicates ang marks location
NAV- non-actor voice; indicates ang does not mark agent
NEG negation
NOM nominative
o object-like argument of canonical transitive verb
OBJ object
-OBJ object verb marker; indicates ang marks patient
OBL oblique
PERF perfect
PFV perfective
PL plural
POSS possessive
PROG progressive
PV- patient voice; indicates ang marks patient
S single argument of canonical intransitive verb
SPEC specifier
SUBJ subject
TOP topic
TRANS transitive

57