THESIS

, LIBRARY
Oi 0 Michigan State
University

 

 

 

 

This is to certify that the
thesis entitled

THE ACQUISITION OF PLURAL MORPHEMES IN KOREAN

presented by

Hye Sun Park

has been accepted towards fulfillment
of the requirements for the

MA degree in Linguistics

 

 

Cit Ll

Major'lgrotessfi’s Signature
AdétKI' it” .QolO

Date

 

 

MSU is an Affirmative Action/Equal Opportunity Employer

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:IProleoc&Pres/ClRC/DateDue.indd

 

THE ACQUISITION or PLURAL MORPHEMES IN KOREAN
By

Hye Sun Park

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS
Linguistics

2010

ABSTRACT

THE ACQUISITION OF PLURAL MORPHEMES IN KOREAN
By
Hye Sun Park

Korean is one of the languages that has non-obligatory, semantically and
syntaxtically complex pluralizers. There are two pluralizers —deul and —ne in Korean.
They are both considered as pluralizers since they deliver a more-than-one interpretation
of a noun that it is attached to. However, while —deul is claimed as having universal
quantification properties (Park 2008) which can generate exhaustive reading or
specificity properties (Kim 2008), —ne has been known as allowing associative
interpretation.

In this study, we first discuss the different interpretation of —deul and -ne by
showing the two distinct syntactic positions of the two Korean pluralizers, and then ask
questions if Korean children have a more-than-one interpretation, exhaustive
interpretation and associative reading. Regarding the questions, the three experiments
were conducted with Korean children from age 4:0 to 5:0 and 5:0 to 6:0 and adults. Our
study results give empirical evidence that shows the properties of —deul and —ne have and
shows how 4 to 6 years old Korean children treat the morphemes, —deul and —ne. It

supports the claim that optionality plays an important role delaying acquisition.

ACKNOWLEDGEMENTS

First of all, I sincerely thank my advisor, Cristina Schmitt for her continuous
support and guidance throughout the project. She gave me the opportunity to start this
project and encouraged me to develop it into a thesis. The procedure was definitely long
and hard. Without her, I would not be able to progress and be here with this
accomplishment. I also want to thank Nao Nakano who performed such an important role
for this project. The discussion with her and the advice she gave me helped me to go

through all the hard works.

I am also grateful to the other members in my committee, Professor Alan Munn
and Professor Yen-Hwei Lin. Professor Alan Munn inspired me in many ways during my
years in graduate school and his teaching and comments guided me to make such a result.
I thank Professor Lin for giving me valuable advice and support. Her guidance also
encouraged me and lead me to make such a progress. I also want to thank to Professor
Woohyeok Chang and Professor Heij in Ahn at Dankook University who kept giving me
precious advice whenever I have doubts and troubles about my works and decisions. And
I appreciate Keunbyeol Kindergarten, Hyewon Kindergarten and Dankook University
students for their cooperation collecting the data for this project and all others who helped

me to finish the experiment successfully.

Finally, my gratitude and blessings goes to my parents who have always stood by

my side and support me in any respect during the completion of the project.

iii

TABLE OF CONTENTS

LIST OF TABLES ............................................................................ vi
LIST OF FIGURES .......................................................................... vii
INTRODUCTION ........................................................................... 1
CHAPTER 1
LIN GUISTIC DESCRIPTION ............................................................. 3
1.1 Korean pluralizers ............................................................... 3
1.2 Semantic Property of -deul and -ne ........................................... 6
1.2.1 -deul .................................................................... 6
1.2.2 -ne ...................................................................... 11
1.3 Syntactic positions of -tu1 and -ney ........................................... 13
CHAPTER 2
PLURAL MORPHEME ACQUISITION ................................................. 26
2.1 Acquisition Background ........................................................ 26
2.2 Hypothesis and Prediction ...................................................... 28
CHAPTER 3
STUDIES ....................................................................................... 30
3.1 Participants ........................................................................ 31
3.2 Experiment 1 ...................................................................... 32
3.2.1 Materials and Methods ............................................... 32
3.2.2 Predictions ............................................................. 34
3.2.3 Result ................................................................... 34
3.2.4 Discussion .............................................................. 37
3.3 Experiment 2 ...................................................................... 38
3.3.1 Materials and Methods ............................................... 38
3.3.2 Predictions ............................................................. 40
3.3.3 Result ................................................................... 41
3.3.4 Discussion .............................................................. 43
3.4 Experiment 3 ..................................................................... 44
3.4.1 Materials and Methods ............................................... 44
3.4.2 Predictions ............................................................. 47
3.4.3 Result ................................................................... 47
3.4.4 Discussion .............................................................. 49
CHAPTER 4
CONCLUSION ................................................................................ 52

APPENDIX 1 .................................................................................. 55

APPENDIX 11 ................................................................................. 61
APPENDIX III ................................................................................ 64
APPENDIX IV ................................................................................ 68
APPENDIX V ................................................................................. 70
APPENDIX VI ................................................................................ 71
REFERENCES ................................................................................ 72

LIST OF TABLES

Table 1 Singular and plural personal pronouns ........................................... 1.5
Table 2 Subjects ............................................................................... 31
Table 3 An example of a more-than-one interpretation task ............................. 33

Table 4 Descriptive statistics for subject’s performance correctly interpreting
-deul and —ne ................................................................................... 35

Table 5 An example of maximal presupposition task ..................................... 40

Table 6 Descriptive statistics for subject’s exhaustive reading of —deul and
—deul + all ....................................................................................... 41

Table 7 An example of associative reading task ........................................... 46

Table 8 Descriptive statistics for subject’s associative reading of —ne and cardinal
number .......................................................................................... 48

vi

LIST OF FIGURES

Figure 1 Percentage of —deul and —ne interpretation (False answer) for adults and
children .......................................................................................... 36

Figure 2 Percentage of —deul and -deul +all exhaustive reading for adults and
children .......................................................................................... 43

Figure 3 Percentage of —ne and cardinal number associative interpretation for
adults and children ............................................................................. 49

vii

INTRODUCTION

Studies of child language acquisition have shown that the basic interpretation of
plural morphology is mastered early in English and Spanish. (Bybee 1985; Dressler 1989,
Brown 1973, Barner and Snedeker 2005, Kouider et al. 2006, Miller 2007). Less is
known about acquisition of non-obligatory plurals. In this thesis we describe the Korean
pluralizers which are distinct from obligatory pluralizers and present three experiments
testing children’s ability to interpret various properties of the Korean —deul and —ne
pluralizers. To our knowledge no previous study of the acquisition of plurals in Korean
has been done.

—deul and —ne are the only plural morphemes in Korean and they are interesting
from the acquisition point of view because they have complex prOperties and many
restrictions in their distribution. First of all, unlike English and Spanish pluralizers,
Korean pluralizers are optional, and bare noun phrases can mean either singular or plural.
In addition to this, with their tendency to appear mainly with animate nouns, Korean
pluralizers are consequently not as frequent when compared to obligatory pluralizers.
Secondly, researchers have argued that —deul has distributive and universal quantification
properties (Park 2008) or specificity (Kim 2008) properties and —ne has been known as
allowing an associative interpretation (Madigan, Yamada and Peng 2008). In other words
—ne produces the reading that takes the form: x and those associated with x in some
context c.

Based on these properties of Korean pluralizers, it is possible that children’s
acquisition of plural marking systems such as Korean will pose special problems because

these morphemes are semantically and syntactically complex and distinct from plural

marking in languages that have semantically regular, obligatory and generally applicable
plural morphology.

In this study, we ask the following questions: (i) do children treat —deul and —ne
as associated to a more-than-one interpretation? (ii) do children and adults interpret -deul
as associated to an exhaustive interpretation? (iii) do children have associative readings
of -ne? (iv) do children have more difficulty learning one pluralizer morpheme than the
other?

To investigate these questions, we propose a set of three experiments testing
children’s interpretation of the properties associated to these two morphemes. The results
will contribute to give us a better understanding of the acquisition of complex pluralizers
and complex morphology in general.

The thesis is organized as follows. In chapter I, we discuss the semantic
properties of Korean pluralizers, —deul and —ne, and their syntactic position in a structure
of DP. In chapter II, we present an overview of plural acquisition studies and then
introduce the acquisition model and the hypotheses tested in the three experiments.
Chapter III presents the three experiments, and finally in chapter IV we re-examine our

hypotheses and predictions and draw our conclusions.

CHAPTER 1

LINGUISTIC DESCRIPTION
In this chapter, we will first introduce the Korean plural morphemes, discuss the

semantic properties of the morphemes —deul and —neI have and then propose an analysis

of the positions of -deul and —ne in a DP structure.

1.1 Korean pluralizers

Chierchia (1998) distinguishes languages with and without a mass-count
distinction. He argues that all nouns are mass in languages without a mass-count
distinction such as Japanese, Chinese and Korean (Classifier languages), and derives the
following facts from this: first, languages without a mass-count distinction lack a plural
marking system; and second, they obligatorily require classifiers for counting. However,
it has been argued that the existence of (optional) plural marking in classifier languages
challenges Chierchia’s analysis (e.g., Chung 2000). In Korean, plural is optionally
marked by —deul and/or —ne. These morphemes can give a more-than-one interpretation

to the noun phrases they are attached to as in (1).

(1) a. Hakseng-deul-i iss-da.
Student-PL-NOM EX-DEC
‘There are students.’

 

l The transcription of Korean follows the revised romanization of Korean (National Academy of the
Korean Language. Seoul, Ministry of Culture and Tourism, 2000). Abbreviations used in the glosses of the
Korean examples: ACC=accusative ; CL= classifier ; DEC=declarative ; Q=question ; EX=existential verb ;
LOC=locative ; NEG=negation ; NOM=nominative ; GEN=genitive ; PL=plural ; PRS=present ;PST=past;
REL=relative marker; SG=singular ; TOP=topic. Hyphens in the Korean examples indicate suffixation.

#: syntactically well-formed but infelicitous in the discourse context.

However, although Korean has these morphemes which allow a more-than—one
interpretation like in English or Spanish, —deul and —ne are not considered as regular
plurals in a number of ways. One reason is the fact that they tend to pluralize animate
nouns and not inanimate nouns. Among the two morphemes, —deul is mainly associated

to animate nouns as shown in (2), and the distribution of -ne is strictly restricted to

Hakseng-ne-ga iss-da.
Student-PL-NOM EX-DEC

‘There are students.’

‘There is a student and those associated with that student.’

animate nouns as in (3).

(2) a.

(3) a.

Hakseng-deul-i man-da.
People-PL-NOM many-DEC
‘There are many students.’

# Moga—deul-i man-da.
Hat-PL-NOM many-DEC
‘There are many hats.’
Hakseng-ne-ga man-da.
People-PL-NOM many-DEC
‘There are many students.’
*Moga-ne-ga man-da.
Hat-PL-NOM many-DEC

‘There are many hats.’

Because of the animacy restriction, they are less frequent than English plural. On
top of that, bare nouns in Korean are neutral in number and can be interpreted as either
plural or singular as presented in (4). The diagram in (5) shows schematically the

ambiguity between singular and plural.

(4) Haksaeng-i iss-ta.
Student-NOM EX—DEC
‘There is a student.’

‘There are students.’

(5)

Bare Noun

/\

Plural Singular

However, a bare noun phrase in Korean is ambiguous not just in number. It is
also ambiguous between a definite, a specific indefinite, and nonspecific indefinte. There
are no articles to give a definite reading in Korean and the source of a definite or a
specific reading has been controversial. Some researchers argued that the Korean plural —
deuI is associated to specificity (Kim, 2008) and some argued that —deul is associated to
distributivity (An 2007, Park 2008, Joh 2009) and universal quantification (Park, 2008)
which can give an exhaustive reading. —ne has not been considered as a morpheme that
gives specificity or exhuastivity but known as having an associativity (Madigan, Yamada

and Peng 2008).

In the following section, we will discuss these semantic properties of Korean
plural morphemes in detail and then we will ask questions about the plural morpheme

acquisition concerning these properties.

1.2 Semantic Property of -deul and -ne
1.2.1 -deul

Kim (2005) argues that the Korean noun denotation universally includes both
singularities and pluralities. As in (6), his analysis of the plural marker —deul explains the
fact that bare plurals in Korean are ambiguous between a plural and singular reading. In
Kim (2005), —deu1 is a marker whose denotation serves to filter out the atomic (i.e.
singular) entities from the extension of the noun with which it combines and this is
illustrated in (7). The dots in (7) indicate the elimination of the possible semantic

interpretation of the noun.

(6) Pati-e haksayng-eul codae-hae-ss-da.
Party-to student-ACC invite-do-PST—DC
‘(We) invited a student/students to the party.’
(Song 1975, 20)

(7)
Bare Noun-PL

\
§
\
\
S
\
\
\
\
\
\
\

\

Plural Singular

We believe his claim is correct and can be supported by the examples in (8). The
sentence (8a) is ambiguous since the bare noun haksaeng ‘student’ can refer to either
singular or pluralized entities. However the sentence (8b) unambiguously delivers the
meaning that ‘Inho saw more than one student’ since the plural morpheme —deuI filters
out the singular interpretation. The sentence (8c) shows the negation of the sentence in
(8b) and it shows that the negation only removes the plural reading. In English, the
sentence ‘John didn’t see students.’ will be considered false if John saw one student. In
Korean, however, (8c) is true in the situation when John saw one student. This
interpretation of (8c) is due to the fact that —deul filters out the singular interpretation and

it tells us that —deul gives a more-than-one interpretation.

(8) a. Inho-ga haksaeng-eul bo-ass-da.
Inho-NOM student -ACC see-PST-DEC
‘Inho saw (a) student(s).’

b. Inho-ga haksaeng-deul -eul bo-ass-da.
Inho-NOM student-PL -ACC see-PST-DEC

‘Inho saw more than one student.’

c. Inho-ga haksaeng-deuI-eul bo-jian-ass-da.
Inho-NOM student-PL-ACC see-NEG-PST-DEC

‘Inho did not see more than one student.’

In Park (2008), the plural morpheme —deul is also analyzed as optional since the
bare noun can express the property that the plural morpheme has. However, she claims

that the plural morpheme —deul has more than a mere pluralizing effect and therefore its

properties cannot be fully captured under the simple plurality analysis. She argues that
—deuI has a link to distributivity, which involves universal quantification in the course of
interpretation and for that reason she claims that nouns with -deul exhibit difierent
semantic properties from those of English—like plural noun phrases. The following
examples from Park (2008) in (9) support this idea. In (9b) —deul is attached to the noun
phrase in the subject of the sentence and the interpretation is that all the faculty members
'took part in' the gathering event. Sentence (9a) does not have the reading that requires all
the faculty members to take part in the event. Sentence (9a) only means that
‘mathematics department professors gathered in the classroom as a group’, which is

vague as to whether all math professors participated.

(9) a. Suhakkwa gyosu—ga gyosil-e mo-yess-da.
Math-dept. professor-NOM classroom-DAT gather-PST-DEC

‘Professors of a math-department gathered in the classroom.’

b. Suhakkwa gyosu-deul-i gyosil-e mo-yess-da.
Math-dept. professor-PL-NOM classroom-DAT gather-PST-DEC
‘(All) the professors of a math-department gathered in the classroom.’

(Park 2008, 282)

However, this claim about —deul is controversial. One of the conflicting analyses
comes from Kim (2008). Her arguments accord with Park’s (2008) in that plural markers
in classifier languages mark more than the plurality of referents of the nouns they attach
to, but she claims that the markers in classifier languages not only mark plurality but also

definiteness/specificity of their base nouns.

In her analysis, the concept of the specificity has been used to describe a
situation ‘when the speaker refers to a particular entity in the universe of discourse,
which may be identifiable or non-identifiable.’ ‘Identifiable’ here is used as denoting a
pragmatic concept of the definiteness. Therefore, specificity implies that a specific
expression can be definite or indefinite and it follows that all definite NPs are specific.
Definite expressions are used when the referent is identifiable to both the speaker and the
bearer; indefinite specific expressions are used when the referent is identifiable only to
the speaker; and indefinite non-specific ones are used when the referent is identifiable to

neither the speaker nor the hearer. This notion is represented schematically in (10).

(10) Bare Noun .
/Pl< Singular
Specific Non-specific Specific Non-specific

/\ l /\

Definite Indefinite Indefinite Definite Indefinite Indefinite

 

Given this distinct interpretation of specific and definite above, Kim (2008)

claims that specificity in Korean can be marked through the presence of the plural marker

—deul (11).

(11) a. Mary-neun gae-leul kileugo-sipohan-da.

Mary-TOP dog-ACC raise-want-DEC
‘Mary wants to raise (a) dog(s).’ (non-specific)

b. Mary-neun gae-deul-eul kileugo-sipohan-da.
Mary-TOP dog-PL-ACC raise-want-DEC
‘Mary wants to raise certain dogs.’ (specific)

0. Mary-neun keun gae-deul-eul kileugo-sipohan-da.
Mary-TOP big dog-PL-ACC raise-want-DEC
‘Mary wants to raise big dogs.’ (specific)

(Kim 2008, 168)

In her analysis, (11a) illustrates that the bare singular noun gae ‘dog’ can be
construed as a singular or plural. Whether it gets a singular reading or a plural reading,
this bare nominal is interpreted in a non-specific way only. When the plural marker —deul
is sufiixed to the bare noun gae ‘dog’ as in (11b), however, it is construed to be plural
specific, i.e, ‘some specific dogs’. Therefore, (11c) cannot be used as an answer to a
question, ‘what kind of dog do you want to have?’ and it supports the indication that -

deul gives a specific reading. This schematic of—deul are shown in (12).

(12)
Bare Noun-PL
Plural Singular
Specific Non-specific Specific Non-specific

10

From the claims about semantic properties of —deul, we found that a more-than-
one interpretation of —deul has not been controversial. Its pluralizing property is agreed
by all (Kang 1994, Irn 2000, Back 2002, Kwak 2003, Kim 2005, Park 2008). So we ask a
question if a more-than-one interpretation of —deul shows up in children’s language and if
the acquisition of these pluralizers is slower in comparison to the plurals that have more
semantically regular, obligatory and generally applicable plural morphology.

The definite/specific interpretation of —deul, however, has been controversial.
However these claims about semantic properties of —deul can converge to a question of
whether —deul generates an exhaustive interpretation or not. If universality or definiteness
is one of the properties that —deul has, it will require an exhaustive interpretation since
both universality and definiteness require a set that contains all and only individuals with
a certain property. However, if —deul has only specificity as Kim (2008) claims, it would
not demand exhaustive interpretation. Therefore, to have more empirical evidence to
support the claims about -deul, we ask a question to both adults and children. Do adults

and children interpret —deu1 as having an exhaustive interpretation?

1.2.2 -ne

Not much has been written about the pluralizer —ne. —ne is also a pluralizer
which gives a more-than-one interpretation and it is also known to be semantically
distinct from additive plurals like English —s (Corbett 2000, Moravcsik 2003). Inforrnally,
an associative plural is comprised of a focal individual and his or her associates

(Nakanishi and Ritter, 2008). In Vassilieva (2008), an associative plural is defined as a

11

nominal expression that refers to a group by naming its most salient member. An

example of this interpretation is shown in (13).

(13) a. Inho-ne-ga belsso ttona-ass-da.
Inho-PL-NOM already leave-PST-DEC
‘Inho and those associated with him already left.’

b. Geu ai-ne-ga belsso ttona-ass—da.
That child-PL-NOM already leave-PST-DEC
‘That child and those associated with him/her already Iefi.’

In (13a), Inho—ne refers to a group consists of a focal individual named Inho and
his associates. —ne only allows the associative reading, even when it combines with a
common noun. ai-ne ‘child-PL’ in (13b) is interpreted as ‘that child and those associated
with him/her’. The construction with an associative plural is used to name a new group
into discourse, a group that is understood to be contextually or inherently associated with
its named protagonist (V assilieva, 2008). ‘Contextually associated’ means that the group
represented by the protagonist is determined from the context and ‘inherently associated’
‘ means that the group will be interpreted as, for example, x and x’s family. Studies of the
acquisition of the associative morpheme in Korean have not been done as far as we know.
The pluralizer —ne which has different semantic properties from —deul or any other
regular plurals raise the following questions. First, do children have a more-than-one

interpretation of —ne? Second, do they have an associative reading for —ne?

12

1.3 Syntactic positions of —deul and -ne

Before we discuss the syntactic position of the pluralizers, —deul and —ne, we
want to point out that there is another morpheme which appear to be the same on the
surface as -deul. However, although the two morphemes appear to be the same, their
distribution is not the same. The morpheme that we have been discussing up until now is
also known as the Intrinsic Plural Marker (IPM) or Nominal deul. The other morpheme is
known as the Extrinsic Plural Marker (EPM) or Non-Nominal deul. Example (14) shows
the position of the IPM and (15) shows all the possible positions of the EPM. The IMP is
a morpheme that is immediately preceded by a noun and pluralizes the noun it attaches to
whereas the EPM is a morpheme that can be optionally concatenated with any phrase
such as adverbial, verbal, or prepositional phrases regardless of the number of its
appearance. But the difference between the IPM and the EPM is not limited to their
distribution. They have different interpretations. While IPM delivers a more-than-one
interpretation to the noun it is attached to, the EPM does not give rise to the pluralizing
effect for the phrase it attaches to. No matter where the EPM appears, it indicates the
plurality of the subject of the matrix clause. As a result, we consider the EPM as an
agreement marker rather than a pluralizer. The IPM —deul is the pluralizer that we study
and therefore the EPM will not be included to our discussion since it is irrelevant to our

study.

(14) Haksaeng-deul-i gongbu-leul han-da.
Student-PL-NOM study-ACC do-DEC

‘More than one student are studying.’

13

(15) Haksaeng-deul-i gongbu-leuI(-deul) yeolsimhi(-deul)

Student-PL-NOM study-ACC hard
gyosileseo(-deul) han-da(-deul).
in classroom do-DEC

‘Students are studying hard in the classroom.’

In this section, we argue that the two morphemes, —deul and —ne should be
distinguished and be considered as morphemes occupying different syntactic positions in
a DP structure even though they both are the pluralizers that mark plurality. This
argument is based on the following descriptions.

Madigan, Yamada and Peng (2008) suggest that —ne is a simple associative plural
marker and the distribution of —ne is different from —deul.

As for —deul, it appears that it is not an associative plural marker since it cannot
yield an associative reading when attached to a proper noun or a common noun as shown
in (16). The only reading we can obtain from Inho-deul in (16a) is one in which there are
multiple people who are all named Inho. The only reading available for ai-deul ‘child-PL’
in (16b) is one where the children, who do not necessarily need to be closely associated

to one another, already left.

(16) a. Inho-deul-i beolsso ttona-ass-da.
Inho-PL—NOM already leave-PST-DEC
*‘Inho and those associated with him already left.’

‘More than one person named lnho already lefi.’

14

b. Geu ai-deul-i beolsso ttona-ass-da.

That child-PL-NOM already leave-PST-DEC
*‘That child and those associated with him/her already left.’
‘The children already left.’

The distribution of —deul and —ne with personal pronouns also gives a piece of
evidence to say that they have a different function. Since plural personal pronouns are
interpreted as a group of a focal individual and his/her associates, the interpretation of
them is said to be similar to that of associative nominals. For example, the second person
plural pronoun ‘you (pl)’ is interpreted as the listener and the associates who cannot be
all referred as ‘you’. As shown in Table 1, the second person pronoun no has a pluralized
form nohi, however, while -deul cannot pluralize the second person pronoun, no, —ne can

pluralize it.

Table 1. Singular and plural personal pronouns

 

 

 

 

1St person pronoun 2ud person pronoun 3rd person pronoun
Singular No No Geu
xi Uli V Nohi
\l Geu-deul
Plural * Na-deul *No-deul
\l Geu-ne
* Na-ne \/ No-ne

 

 

 

 

 

 

The possibility for —ne to pluralize second person pronoun would be because of
the fact that —ne creates associative reading. The second person plural pronoun nohi and

no-ne basically deliver the same meaning but show slight difference in their usage as

15

shown in (17). The derived second person plural pronoun no-ne as in (17a) is preferred
when compared to nohi as in (17b) when the referred group is significantly distinguished

from other groups or has more tight relationship among the associates of the group.

(17) a. No-ne-ui tim-eun ig—yess-ni?
You-PL-Gen team-Top win-PST-Q?

‘Did your team win?’

b. Nohi-ui tim-eun i g-yess-ni?
You(pl.)-Gen team-Top win-PST—Q?

‘Did (each of) your team(s) win?’

As for the third person pronoun geu, —deul and —ne both can be suffixed to
pluralize the bare form. For a plural interpretation to be obtained the third person has to
have a plural morpheme to be pluralized. However, the first person pronoun cannot
appear with pluralizers. Both —ne and —deul cannot be suffrxed to pluralize the bare form.
We assume that uli blocks —ne to pluralize the first person pronoun, na.

Now, according to the different behavior of —deul and —ne described so far, we
propose a basic syntactic analysis for the two morphemes.

First, we propose that -—deul and —ne occupy different syntactic positions. The

co-occurrence of —deul and —ne in (18) supports this claim.

(18) a. Geu salam-deul-ne-neun chaek-eul ilg-oss-da.
That person-PL-PL-TOP book-ACC read-PST-DEC

‘Those people and their associates read (a) book(s).’

16

b. Geu salam— ne-deul-neun chaek-eul ilg-oss-da.2
That person-PL-PL-TOP book-ACC read-PST-DEC

‘Those people and their associates read (a) book(s).’

From this, we assume that there must be two positions for the two morphemes.

Among the two positions, first we propose the position of —deul. There is a piece
of syntactic evidence that leads us to assume that —deul takes the same syntactic position
as the Classifier Phrase (CLP). The following structure in (19) represents the position of —

deul that we are proposing.

(19)
DP
NumP D
CLP Num
NP CL
{CL/— deul}

As has been introduced in the previous section, Korean is a language that has a
classifier system. Therefore a particular classifier is required for counting nouns. Borer
(2005) argues that both plurals and classifiers serve to create count nouns from
unstructured stuff. That is, plural inflection in non-classifier languages such as -s in

English, for instance, is classifier inflection which corresponds to classifiers in classifier

 

2 The two morphemes can be reversed but it is not clear if this reversed order tells anything about their
syntactic position. Madigan, Yamada and Peng (2008) say that the use of multiple plural markers is also
grammatical with no reported change in meaning.

l7

languages such as Korean. Her analysis and the fact that —deul exhibits complementary
distribution with a classifier such as in (20) lead us to claim that —deul and classifiers in

Korean belong to the same functional category.

(20) a. Inho-ga haksaeng-deuI-eul bo-ass-da.
Inho-NOM student-PL-ACC see-PST-DEC

‘Inho saw more than one student.’

b. Inho-ga hakseng du-myeong-eul bo-ass-da.
Inho-NOM student two-CL-Acc see-PST-DEC
‘Inho saw two students.’

c. *Inho-ga haksaeng-deul-du-myeong-eul bo-ass-da.
Inho-NOM student-PL-two-CL-ACC see-PST-DEC

‘Inho saw two students.’

The sentences in (20a, b) are good since —deul and the classifier did not occur in
a single DP structure. (20c) is ungrammatical if we treat haksaeng-deul—du-myeong-eul
‘two students’ as a single DP. However it can be acceptable if we have two separately

generated DPs, haksaeng—deuN—eul) and du—myeong-eul, as in (21).

(21) Inho—ga haksaeng-deul(-eul) du-myeong-eul bo-ass-da.

Inho-NOM student-PL(-ACC) two-CL-ACC see-PST—DEC

‘Inho saw two students.’

18

Because case on haksaeng—deul(—eul) can be dropped in Korean, it is possible
that (20c) could also have the structure of (21). However, if case is dropped in (20c), it
appears as a pause between the two DPs (Kim, 2005) and there is evidence to show they
are not a single DP. The sentence in (22a) shows the same sentence in (21) but with two
separately generated DPs with overt cases. To support that there are two separately
generated DPs, we give evidence of double object construction and a topic-comment
relation construction. (22b) shows a double object construction and it shows the
appearance of an adverb between the direct and indirect object. The adverb joyonghi
‘quietly’ between the two DPs in the example in (22c) confirms that they are the separate
DPs. These two DP constructions have been claimed as having a topic-comment relation
‘ and therefore Kim (2005) says the interpretation of a sentence with two DPs such as in
(21) is more appropriate if we interpret it as ‘Inho — as for students — saw two (of them).’

(23) is the structure of the two DPs that she is proposing.

(22) a. Inho-ga haksaeng-deul-eul du—myeong-eul3 bo-ass-da.
Inho-Nom student-PL- ACC 2-CL-Acc see-PST—DEC

‘Inho saw students and they were two.’

b. Inho-ga haksaeng-deuI-ege joyonghi
Inho-Nom student-PL- DAT quiet
cheak-eul ju-ass-da.
book-ACC give-PST—DEC

‘Inho gave students books silently.’

 

3 Kim (2005) suggested that the apparent ‘Accusative Case-doubling’ construction contains two separately
base-generated DPs, which form a Topic-Comment structure, and give rise to a partitive interpretation.

19

c. Inho-ga haksaeng-deul-eul joyonghi

Inho-Nom student-PL- ACC quiet
du-myeong-eul bo-ass-da.
2-CL-ACC see-PST—DEC

‘Inho saw students silently and they were two.’

(23) TopP

The structure in (23) was proposed by Kim (2005) to explain the two DPs and
their relationship. This structure also can explain the appearance of adverbial material
between the first and the second DP shown in (22b).

The syntactic position of —deuI occupying the position of a classifier is also
supported by Park (2008). She argues that —deul is incompatible with a classifier and they
have an identical function. She argues that if —deul appears with classifier, it gives rise to

a conflict with the complex of number and classifier which already resides in CL. The

structure of (24) is shown in (25) and (26). (25) and (26) show the co-occurrence of —deul
and classifier and their conflict due to the violation of the Head-Movement Constraint.
This shows that —deul cannot be allowed in a single DP structure with classifier and
supports our claim that —deul and classifier occupies the same position. To have a proper

structure of an example (24), the structure should be as in (27).

(24) * Sakwa-deul-se—kay.
Apple-PL-three-CL

‘Three apples.’
Vconflict
(25) [DPINumPICLPINP salewedeallwp *sakwa-deul se-kaylcm se-kayINumplnp
apple-PL apple-PL three-CL
(26)
DP
NumP D
CLP Num
NP CL

sakwa *deul, se-key
‘apple’ PL, 3-CL

vP
/\ /\
NumP D AgrO v
/\ /\
CLP Num DP A grO’
/\ /\
NP CL
sakwa deul NumP D
‘apple’ PL /\
CLP Num
NP/\CL
se-key
3-CL

For the position of the associative morpheme —ne, we adopt the analysis of
Nakanishi and Ritter (2008) and propose that ——ne belongs to a category GROUP which is

merged above D. (28) is the tree structure of DP with —deul and -ne that we are proposing.

(28)
Ger
/\
DP Grp
/\ _ne
NumP D
CLP Num
NP CL
{ CL/— deul}

The first piece of evidence that supports the position of —ne as in (26) can be
found from its occurrence with proper names. Longobardi (1994) shows N-to-D
movement in a DP structure of Italian and argues that nouns that are proper names can
move to a D position. D has been considered as an operator position which binds an NP
and turns it into an argument (Stowell 1989, 1991, Longobardi 1994, 1996). This
assumption amounts to the claim that D is obligatorily present in argumental noun
phrases across all languages if arguments are to be treated as entities or generalized
quantifiers. In the Korean case, based on the fact that —ne only selects the entities of type
<e>, as the example below show (29), we assume that the proper name in Korean
occupies a D position. Consequently —ne must be the head of the GrP since its
interpretation requires the entity to form a group out of entities. This is why it can appear
with proper names and pronouns. And this structural notion explains why the entities
selected by this associative pluralizer —ne are interpreted as ‘a group with the focal

individual’ as in (30).

(29) a. *Manheun salam-ne-ga wa-ss-da.
Many person-PL-NOM come-PST-DEC

Many people came.

b. Geu salam-ne-ga wa-ss-da.
That person-PL—NOM come-PST—DEC

That person and his/her associates came.

23

(30) Inho-ne
Inho-PL

‘Inho and his associates.”

This claim can be further supported by the discussion of personal plural pronouns.
The interpretation of plural personal pronouns is said to be similar to that of associative
nominals and therefore it has been differentiated from other plural nominals. Like the
associative nominals, plural personal pronouns are interpreted as a group of a focal
individual and his/her associates. This is not true for other plural nominals because they
can only refer to every identically named element. For example, the second person plural
pronoun ‘you (pl)’ indicates a listener and the associates. But it is impossible for plural
nominal ‘cups’ to refer any other item other than ‘cup’. For this, Panagiotidis (2002)
proposed a structure of personal pronoun which supports our analysis of the position of
associative pluralizer. In his analysis, all pronouns consist of two functional shells (a DP
layer and a NumP layer) and one lexical NP layer. This is shown in (31). His claim is that
all pronouns are definite descriptions and [person] is a special type of deictic (definite)

feature.

(31) [DP D° [person] [NumP Num° [number] [NP N° [gender]]]]

Adopting this analysis, the interpretation of pluralization of personal pronouns
with the associative pluralizer —ne can be explainable with the structure that we are
proposing. Among the three personal pronouns in Korean, the associative pluralizer —ne

is allowed to be attached and pluralize the second and third person pronoun no and geu.

And when —ne is attached to the second or third person pronoun, it is interpreted
‘associatively’ in a context; No-ne as ‘you and your associates.’ and geu-ne as ‘he/her
and his/her associates.’ The second person pronoun has a lexicalized plural form, nohi
‘you (pl)’, and nohi also means ‘you and your associates.’ Based on the analysis of

personal pronoun, the structure would be as in (32).

(32) [DP [NumP [CLP [NP fieh‘til CLP ti lNumP ti IDP n0hii I

The personal pronoun with an associative pluralizer such as no-ne ‘you (pl)’, a
derived plural pronoun, which gives the same meaning as nohi can be analyzed as shown

in (33).

(33) [(3er [DP [NumP [CLP [NP 99:] CLP 1: lNumP ti IDP 1101(3er 116 I

In sum, the two different position of -deu1 and —ne in the structure of DP
correctly captures the semantic differences of the two morphemes and it was supported
by the two—DP structure and the structure of proper names and personal pronouns. —deul
only can appear with third person pronouns while —ne can appear with second and third
person pronouns.

In the next section, we will present the overview of the acquisition of plurals and

the hypotheses based on the previous plural acquisition studies.

25

CHAPTER 2

PLURAL MORPHEME ACQUISITION

The experimental studies presented in this thesis examine the acquisition of the
two different plural morphemes, -deul and ——ne. In this chapter, to be able to test subjects’
specific knowledge or lack of knowledge of the two morphemes, we first overview the
previous studies of plural morpheme acquisition and present our hypotheses based on the

current findings.

2.1 Acquisition Background

There have been no studies about acquisition of Korean pluralizers but studies of
plural morpheme acquisition in other languages have shown how children learn plural
morphemes.

Studies of the acquisition of the plural morpheme in English (F erenz and Prasada,
2002; Wood, Kouider, and Susan 2009) have found that English-learning toddlers begin
to mark the singular-plural distinction after 20 months of age and before 24 months of
age. Furthermore, Koider, Halberada, Wood and Careyet 2006, using a preferential
looking paradigm, suggest that most three-year-olds (but not two-year—olds) comprehend
the morpheme -s on novel word forms as indicating more than one.

Recent studies of children learning plurals in a language with inconsistent input
such as Chilean Spanish showed a different acquisition pattern from that of English
speaking children. Miller and Schmitt 2009, Miller 2007 show that Chilean Spanish
children’s comprehension of plural morphology has not been completed even at around

age five. The plural morpheme in Chilean Spanish is not consistently produced on all

26

elements within the noun phrase and therefore its presence is more variable. The results
from Chilean Spanish confirm the hypothesis adapted from Yang’s (2002) Variation
Model of language acquisition which says variability in the input will delay child
comprehension of grammatical morphemes when the variability causes unreliability in
the input.

Little is known about the acquisition of optional pluralizers like —deul but there is
one study of the acquisition of the Chinese pluralizer —men which has properties similar
to those of —deul (Munn, Zhang and Schmitt, 2009). —men is a morpheme which is
interpreted as both definite and plural and can also create associative meanings when
attached to a proper noun. Importantly, it cannot appear with numerals or classifiers. In
this study, the experiment tested whether children know the plurality, and definiteness of
—men noun phrases and whether there are differences in the learning of the semantic parts
that —men encodes. The experiment was done with 3 to 10-year old children and the
results showed that Mandarin speaking children do not fully acquire definiteness of -men
until 7 to 10 years of age. However the different behavior between the S-6-year-olds and
the 7-10-year-olds indicated that there is a clear developmental pattern. While 7-10—year-
olds behaved like adults treating —men as plural and maximal, the 5-6-year-olds barely
treated -men as plural or as maximal. They only distinguished between the singular and
plural conditions. The 3-4-year-olds did not seem to distinguish the singular from the
plural, nor treat —men as maximal. The fact that the 5-6-year-old group distinguished the
singular from the plural, but did not treat —men as maximal supports the idea that the
component parts of portmanteau morphemes are learned separately and that plurality is

learned before definiteness. This study suggests that the different properties of a

27

morpheme are likely to be Ieamed at different times by children and the children’s
deviation from the adult patterns reflects a different preference of interpretation rather
than some property of the linguistic representation that is learned very late.

No study has been conducted about associative pluralizers such as —ne which
include exceptional associates as a member of the set. However Zapf and Smith (2009)
showed that two-year-old English speaking children had knowledge of the plural
requirement when two identical instances of an object were presented but had difficulty
in generating labels when presented two of a different kind. This study suggests that
young children may have difficulty in pluralizing non identical sets of objects, which

might extend to difficulties with associative plural, in languages such as Korean.

2.2 Hypotheses and Prediction
Given the previous studies of plurals and the behavior of Korean pluralizers, we

propose the following hypotheses.

1. If variable and unreliable input delays the acquisition of the morpheme and a
non obligation of plurality also hinders acquisition, the mastery of Korean pluralizers will

be protracted in comparison with the mastery of obligatory plurals such as English —s.

2. If age differences affect the ability to use language during the earliest stages of

language acquisition, young children will have more errors interpreting the pluralizers

than older children.

28

3. If children make a distinction between one and more-than-one of the same
kind prior to learning the rule that pluralizes, the acquisition of the pluralizer —deu1 will
be faster than the acquisition of —ne since —deul refers to the identical elements while -ne

creates a group of associates of the referred elements which may not be identical.

CHAPTER 3

STUDIES

The study is comprised of three experiments and we used the Truth Value
Judgment Task (TVJT) as described in Gordon (1996) and Crain and Thornton (1998).
The TVJT was used to test both children and adult's knowledge of pluralizers on the
interpretation of referring expressions such as gore-deal or gore-ne. For examples, the

pluralizer —deul rules out certain interpretations of sentences like:

(34) Gore-deul-i mul-eul ppum-ess-da.
Whale-PL-NOM water-ACC blow-PST-DEC

‘Whales blew water.’

Since —deul is a morpheme which gives a more-than-one interpretation gore-deal
in (34) cannot refer to a singular whale.

During the test, the subject must decide whether a statement such as in (34) is
true or false as a description of a particular situation. The sentences will be asked to both
adults and children and their answers will let us to tap into the participants in a different
age group’s knowledge or lack of knowledge of the properties of the pluralizers, —deul
and —ne.

Using the TVJ T, we created experimental protocols (stories + test sentences) that

can be used to test whether adults and children know the following:

(a) —deul/—ne must be used to refer a plural entity.

(b) —deul/—ne cannot be used to refer a singular entity.

30

(c) —deul is interpreted as having an exhaustive reading.

((1) —ne is interpreted as having an associative reading.

The first experiment tested (a) and (b), the second experiment tested (c) and the

third experiment tested ((1).

3.] Participants

For the present study, 58 children and 60 adults were tested. The children were 4
to 6-year-old monolingual Korean speakers. Adults were also monolingual Korean
speakers and they were first year college students at Dankook University, Korea. Table 2
shows the number of participants, the range, mean and standard deviations. Adults were
tested in university classrooms and children were tested individually in kindergarten
classrooms. All the children and adults participated in the three experiments. Child
participants’ performance was videotaped. Adults participated in a paper and pencil

version of the children’s task.

 

 

 

 

Table2.

Subjects
Group 3 Range Mean Desigtion
4 year-olds 31 4;O-5;0 4.1 .31
5 year-olds 27 5;1-6;0 5.1 .18
Adults 60 20 20 .00

 

31

3.2 Experiment 1
The first experiment was designed to determine whether children interpret the

morphemes, —deul and —ne as associated to a more-than-one interpretation.

3.2.1 Materials and Methods

In the first experiment, four stories tested the interpretation of -—deul and four
stories tested the interpretation of —ne. There were three pictures for each story. The first
picture shows two animals of the same species performing the same activity, and the
second picture shows one of the animals leaving. In the third picture, the animal that is in
the picture is performing a different activity from the animals’ activity in the first picture.
Example (35) is one example story that was used in the first experiment. See Appendix I

for all the stories.

(35)

Bada-e gore du—mari-ga suyeong-eul hago-iss-oss-da.
Ocen-LOC whale two-CL-NOM swim-ACC do-PRE-PST-DEC

‘In the ocean, two whales swam.’

Gore han-mari-nun nasgam-eul-garo tto-nass-da.
Whale one-CL-TOP nap-ACC-for leave-PST-DEC

9

‘One of them went away to take a nap.

Nameun gore-nun mul-eul ppum-ess-da.
Lefi whale-TOP water-ACC blow-PST-DEC

‘The other one blew water.’

32

After the story, the target (i) or (ii) was given. Table 3 shows one example of the
testing materials. To the child participants, the targets were given by a puppet and the
children were told to answer if the puppet was correct or not. Adults were told to mark as
true or false the targets on an answer sheet based on what they had heard and seen in the
story. No subject heard both targets with the same story. Group A heard (i) and group B

heard (ii) so that all the materials were counterbalanced.

 

 

 

 

Table 3.

An example of a more-than-one interpretation task.

Condition Stimulus Study sentence

£32511“ Gore-deul-i suyeong-eul ha-ess-da.

. . . Whale-PL-NOM swim-ACC do—PST-DEC
IndIcatIng ‘Th h l ,

more than e w a es swam.

one entity

(ii) using

-deul Gore-deul-i mul-eul ppum-ess-da.
indicating Whale-PL-NOM water-ACC blow-PST-DEC
singular w ‘The whales blew water.’

entity "

 

 

The adult participants heard one additional filler sentence after or before they
heard a target sentence (i) or (ii). For one story, there are two filler sentences, true and
false, and their distribution was counterbalanced. The filler sentences were not given to

the child participants. See Appendix IV for all the filler sentences.

33

3.2.2 Predictions

In this experiment, the participants heard a story as above. When the story ended,
the participants were asked to judge a target sentence based on what they have heard and
seen. If participants heard the target (i) ‘Whale-deul blew water.’ and if they know the
morpheme —deul has more-than-one interpretation, they should reject sentence (i)
because there was only one whale that blew water. Target (ii) is ‘Whale-deul swam.’ If
the participants heard this sentence and know the plural morpheme —-deul has a more-
than-one interpretation, they should accept sentence (ii) since there were two whales
swimming in the story. The test for a more-than-one interpretation of -ne is the same
kind as the test of —deul. In this experiment, we expect adults to reject (i) but accept (ii).
For 4-year-old children, we expect their correct answers to be significantly worse than
adults and 5-year-old children’s correct answers to be significantly better than 4-year-old

children’s answers.

3.2.3 Results

We compared the responses of three different age groups, 4 year-olds, 5 year-olds
and adults to see the developmental pattern of the children. We separated the
experimental items that had true as the target answer from the false ones. The target
sentence with a true answer had a pluralizer —deul to describe a more-than-one animals in
the picture story and the target sentence with a false answer had a pluralizer —deul to
describe a singular animal in the picture story. Descriptive statistics for the adults and

children’s more-than-one interpretation of —deul and —ne are presented in Table 4.

34

Table 4.

Descriptive statistics for subject’s performance correctly interpreting —deu1 and —ne

 

 

 

Age Group

More-than-one interpretation Adult 4 year-olds 5 year-olds Children

M (SD) M (SD) M (SD) M (SD)
DEUL (i. True answer) 0.95 (0.17) 0.91 (0.22) 0.90 (0.19) 0.91 (0.21)
DEUL (ii. False answer) 0.94 (0.18) 0.44 (0.40) 0.38 (0.40) 0.40 (0.40)
NE (i. True answer) 0.67 (0.43) 0.80 (0.35) 0.83 (0.36) 0.81 (0.35)
NE (ii. False answer) 0.77 (0.31) 0.48 (0.42) 0.40 (0.42) 0.43 (0.41)
N 60 31 27 58

 

 

First, a between-subject one way ANOVA analysis was conducted to see the

difference between children’s age groups. The results indicated that the pattern of

responses of 4 year-olds and 5 year-olds did not show significant main effect in both

targets with true answer (—deul: F(1, 56) = 0.45, p = .833/ —ne: F(1, 56) = 0.124,p = .726)

and with false answer (—deul: F(1, 56) = 0.81,p = .776/ —ne: F(1, 56) = 1.230, p = .272).

Therefore, for the further analysis of children’s more-than-one interpretation of -deul and

—ne in comparison with adults’ interpretation, we combined 4 and 5 year-olds’ responses.

The analysis of children’s responses shows that the mean of target (i), true

answer, is (0.91) for —deul and is (0.81) for —ne. Children’s mean of target (ii), false

answer, however, has a mean of (0.40) for —deul and is (0.43) for —ne.

For adults the mean for the true answer is (0.95) and the mean for the false

answer is (0.94) for -deul. For —ne, the mean values are lower than —deul: (0.67) for the

targets, true answer, and (0.77) for the false answer.

A between-subjects one way ANOVA indicated that children’s responses of

target (i), true answer, were not significantly different from adult’s responses (—deul: F (1 ,

116) = 1.015, p = .316/ —ne: F(1, 116) = 3.584, p = .061). A between-subjects one way

ANOVA result for children’s responses of target (ii), false answer, however, were

35

significantly different from adult’s responses (-—deul: F (l, 116) = 87.337, p < .001 / —ne:
F(l,116)= 26.220,p < .001).

Next, a 2 x 2 mixed-design ANOVA was conducted to see if children acquired
one morpheme faster than the other. Age group is the between subjects variable with two
levels (adults and children); Type of morpheme in the experimental items where the
answer was false is the within subjects variable with two levels (—deul and —ne). For this
analysis, we chose target (i), false answer, since the true answer target (ii) reveals a yes-
bias.

Figure 1 shows the percentage of adults and children’s correct interpretation of
both —deul and —ne as associated to more-than-one interpretation. A between-subjects
ANOVA of children revealed no significant effect of type of morpheme (F (1 , 57) = 0.269,
p = .606). However, the results of the adults showed a significant effect between the type

ofmorpheme (F(1, 59) = 13.409,p < .001).

Figure 1.
Percentage of —deul and —ne interpretation (False answer) for adults and children.

 

 

lOO‘V
° 95%
80% 800
60%
o H deul
40% lief: ._. n,
20%
0%
Adult Children

36

3.2.4 Discussion

The adults’ results in the first experiment showed that the Korean morphemes
—deul and —ne are associated to a more-than-one interpretation and this result is consistent
with the claims about the properties of Korean pluralizers (Kang 1994, Im 2000, Baek
2002, Kwak 2003, Kim 2005, Park 2008). Although the mean of a more-than-one
interpretation of the morpheme —ne was not as high as —deul, we still consider —ne as
pluralizer morpheme in adults’ language since we regard the lower mean value of —ne as
due to interference from the ‘inherent associated’ interpretation. Vassilieva (2008)
suggested that in an associative interpretation, a group can be understood by the context
or a group can be understood to be inherently associated with its named protagonist, such
as x and x’s family. For example, in the story, there were two whales in the ocean. Later
the participants saw a singular whale which blew water. The context did not give any
evidence to make the participants recognize the whales as a family or friends. But
because of the fact that they are the same species the participants might have made an
association between the whale which blew water with the other whale which also blew
water in some place or other. This explanation is compatible with the meaning of the
morpheme —ne itself which requires an association. If subjects made such an association
between the two animals, they would be more likely to judge the singular —ne items as
True, thus lowering the mean plural responses for that morpheme.

From the children’s result in this experiment, we did not find any developmental
pattern of morpheme acquisition and could not confirm Hypothesis 2, whether the age

impacts the acquisition of a more-than—one interpretation. However, our results show that

37

4 and 5-year-olds do not interpret the morphemes —deul and —ne as having a more-than-
one interpretation. This result is very much like what was found for Chinese pluralizer
—men that 5-6-year-olds barely treated —men as plural (Munn, Zhang and Schmitt, 2009).
The results support Hypothesis 1; mastery of Korean plurals will be protracted in
comparison with the acquisition of obligatory plural morphemes.

Finally, we did not find any different behavior in children comparing —deul and

—ne and therefore Hypothesis 3 was not confirmed.

3.3 Experiment 2

The second experiment was designed to test if the participants interpret -deul as
associated to an exhaustive interpretation. The adult data will provide evidence to argue
for or against the claims in the theoretical literature about the meaning of —deul. In this
experiment, Hypothesis 2, allows us to predict a more adult-like performance of 5 years

old participants than 4 years old participants.

3.3.1 Materials and Methods

There were four stories in this experiment and each story included two pictures.
The first picture showed five animals of the same species. In the second picture, four
animals possess the same object each and only one animal possesses a different object
from the other four. Example (36) is one example story that was used in Experiment 2.

See Appendix II for all the stories.

38

(36)

Narngeuk-e paengguin dasos-mali-ga iss-oss-da.
Antarctica-LDC penguin five-CL-NOM EX-PST—DEC

‘In Antarctica, there were 5 penguins.’

Oneu gyeoul paengguin ali-ga al-eul pum—oss-da.
One winter penguin four-CL-NOM egg-ACC warm-PST-DEC

‘In one winter, 4 penguins kept their eggs warm.’

Geuronde paengguin han-mali-nun sagwa-leul pum—oss-da.
But penguin one-CL-TOP apple-ACC warm-PST-DEC

‘But 1 penguin didn’t have an egg, so she warmed an apple.’

Table 5 shows one example of the testing materials. After the story, sentence (i)
or (ii) was given. Sentence (i) is the experimental sentence and sentence (ii) is the control
sentence. To the child participants, the sentences were given by a puppet and they were
told to answer if the puppet was correct or not. Adults were told to mark true or false of
the given sentences on the answer sheet based on what they have heard and seen. No
subject heard both the test sentence and the control sentence with the same story. Group A
heard the test sentence and group B heard control sentence so that all the materials were

counterbalanced.

39

Table 5.
An example of the materials from Experiment 2.

 

 

 

 

 

Condition Stimulus Study sentence
(i) using — . .
deul Paengguin-deul-i al-eul pum-oss-da.
indicating Penguin-PL-NOM egg-ACC warm-PST-DEC
five ‘The penguins warmed the eggs.’
animals
(ii) using -
:2“, and Panguin-deul-i modu al-eul pum-oss-da.
. . . Penguin-PL-NOM all egg-ACC warm-PST-DEC
IndIcatIng ,1 g ‘ h . 11 (1 th ,
five «a a. T e pengums a wanne e eggs.
. nun..- “w" —-
anImals

 

Adult participants heard one additional filler sentence after or before they heard a
test sentence (i) or a control sentence (ii). For each story, there were two filler sentences,
true and false, and their distribution was counterbalanced. No filler sentences were used

with the child participants. See Appendix V for all the filler sentences.

3.3.2 Predictions

In this experiment, participants heard 4 stories as above. After they heard each
story, the participants were asked to judge a test sentence (i) ‘Penguin-deul warmed the
eggs.’ or a control sentence (ii) ‘All the penguins warmed the eggs.’ If the participants
interpret —deul as associated to some condition (universal quantification on definiteness)
that forces all members of the set to be included, they should reject the test sentence (i)

because not all the penguins warmed the eggs. But if they do not interpret —deu1 as

40

forcing an exhaustive interpretation, they should accept the test sentence (i). For the
control sentence (ii), participants should reject the sentence because the universal
quantifier all gives the exhaustive reading. We expect the correct response rate of the
control sentences to be near 100% for both adults and children and the results of the

control sentence will be used as baseline that can be compared with —deul alone.

3.3.4 Results
Descriptive statistics for the participants’ exhaustive reading of testing sentences

and control sentences are presented in Table 6.

 

 

Table 6.
Descriptive statistics for subject’s exhaustive reading of —deul and -deul+all
Age Group

Exhaustive interpretation Adult 4 year-olds 5 year-olds Children

M (SD) M (SD) M (SD) M (SD)
DEUL (Test) 0.61 (0.46) 0.40 (0.45) 0.53 (0.46) 0.47 (0.45)
DEUL + ALL (Control) 1.00 (0.00) 0.79 (0.37) 0.60 (0.44) 0.74 (0.40)
N 60 31 27 58

 

 

First, a between-subjects one way ANOVA analysis4 was conducted to see if
there are any significant differences between the age groups. The result indicated that the
pattern of responses of 4 year-olds and 5 year-olds did not show significant main effect

on both testing sentences (F(1, 56) = 0.671, p = .416) and control sentences (F(1, 56) =

 

" In the analysis of this experiment, before the within-subjects analysis, we conducted a one sample t-test
of the very first —deul responses of the two different adult groups, A and B, to see if there was a significant
difference between the groups. Adults in group A got the control sentence (—deul + all condition) prior to
receiving the test sentence (—deul alone condition) and this group’s mean value was significantly higher
than group B’s responses who got the test sentence (—deul alone condition) prior to have control condition
(-deul + all condition). However, since the t-test results of the other responses were not significantly
different, we collapsed the two groups’ data for the further analysis. There were no differences for children.

41

0.206, p = .652). So we combined the two different age groups and compared them with
adults.

The results indicate that adults’ response to ‘—deul alone’ had a mean of (0.61)
and —deul with universal quantifier had a mean of (1.00). The children’s response to ‘—
deul alone’ was (0.47). When there was a universal quantifier in the sentence, children’s
response’s mean was (0.74).

Next, for the analysis of participants’ exhaustive interpretation a 2 x 2 mixed-

design ANOVA was conducted. Age group is the between subjects variable with two
levels (adults and children); The existence of universal quantifier with the morpheme
—deul or not is the within subjects variable with two levels (‘—deul alone’ and ‘—deul +
all’).

Figure 2 shows the percentage of adults and children’s exhaustive interpretation
of both ‘-deul alone’ and —deul with all. For within subject factors, adults’ results
showed a significant effect of the existence of universal quantifier with the morpheme
—deul (F(1, 59) = 41.013, p < .001). For within subject factors, the children’s results also
showed a significant effect of the existence of universal quantifier with the morpheme
—deul ( F(1, 57) = 8.285, p = .006). A between-subjects one way ANOVA indicated that
there was no significant difference (F(1, 116) = 2.849, p = .094) between adults and
children on the ‘—deul alone’ interpretation. On the control sentence, however, there was

significant difference between the two groups, (F (1, 116) = 24.491 , p < .001).

42

Figure 2.
Percentage of —deul and —deul + all exhaustive reading for adults and children

 

 

100% 100%
80%
73%
60% 61%
47% H Adult

40% O=O Children
20%

0%

deul + ALL deul

3.3.4 Discussion

The results of Experiment 2 from both children and adults show that the
argument that Korean —deul is associated to universal quantification (Park 2008) cannot
be supported. At least from this experimental result, we can say that adults do not require
an exhaustive interpretation of the morpheme —deul. Adults gave 100% exhaustive
reading when the universal quantifier all was present. Without all, only 50% of the time
did they gave an exhaustive reading.

We did not find any significant difference between 4 year-olds and 5 year-olds.
All the children showed an exhaustive reading when the universal quantifier all was
present and they did not show an exhaustive reading without all. The children
significantly differentiated the sentences based on the existence of universal quantifier all.

We therefore conclude that children clearly do not interpret —deul as requiring

43

exhaustivity.
In sum, from the second experiment, we conclude that the Korean morpheme
-deul does not require an exhaustive interpretation and the exhaustive interpretation also

does not exist in children’s language.

3.4 Experiment 3

The third experiment was designed to determine if children have the associative
reading of the morpheme —ne. Hypothesis 3, repeated below, allows us to predict that
children will have difficulties interpreting —ne than —deul. If children make a distinction
between one and more-than-one of the same kind prior to learning the rule that transform
nouns to plural, the acquisition of the pluralizer —deul will be faster than the acquisition

of —ne.

3.4.1 Materials and Methods

In this experiment, there are four stories and two pictures in each story. The first
picture introduces five animals: two animals are of species A and the other three are of
species B. In the second picture, four animals are perfomring the same activity and only
one animal of species A is performing a different activity from the other animals’ activity.
Example (37) is one example story that was used in the third experiment. See Appendix

III for all the stories.

44

(37)

Gorn du—mali-wa mal se-mali-ga nonggang-e iss-oss-da.
Bear two-CL-and horse three-CL-NOM farrn-LOC EX-PST-DEC

‘2 bears and 3 ponies were in the farm.’

Gorn han-mali-wa mal se-mali-neun sule-leul dang-geoss-da.
Bear one-CL-and horse three-CL-NOM cart-ACC drag-PST-DEC
‘1 bear and 3 ponies dragged a cart.

Nameun gom han-mali-neun sule-leul mil-eoss-da.
Left bear one-CL-NOM cart-ACC push-PST-DEC
‘The other bear pushed a cart.’

There was one test sentence and one control sentence for each story: (i) and (ii).

Table 7 shows one example of the testing materials.

45

Table 7.
An example of associative reading task.

 

 

Condition Stimulus Study sentence
3,)! ti: £1313; Gom-ne-ga sule-leul dang-geoss—da.
. . Bear-PL-NOM cart-ACC drag-PST-DEC
of associative , ,
. Bear-ney dragged a cart
reading

 

 

('1) Cardinal Gom du-mali-ga sule-leul dang-geoss-da.

“ml?“ and Bear two-CL-NOM cart-ACC drag-PST-DEC
classrfier , ,

. 2 bear (bare noun) dragged a cart.
Without -ne

 

 

To the child participants, experimental and control sentences were given by a
puppet and children were told to answer if the puppet was correct or not. Adults were told
to mark true or false of the given sentences on the answer sheet based on what they have
heard and seen. No participants heard both test sentences and control sentences with the
same story. Group A heard the test sentence and group B heard the control sentence of
story A so that all the materials were counterbalanced.

The adult participants heard one additional filler sentence after or before they
heard a testing sentence (i) or a control sentence (ii). For each story, there were two filler
sentences, true and false, and their distribution was counterbalanced. The filler sentences

were not given to the child participants. See Appendix V1 for all the filler sentences.

46

3.4.2 Predictions

In this experiment, the participants heard two stories at the same time as above.
Before each story started, the background and the animals in the two stories were
contrasted so that the participants could make a tight association between the animals in
the story. After the story ended, the participants were asked to judge a test sentence (i)
‘Bear—ne are dragging the carts.’ or a control sentence (ii) ‘Two bears are dragging the
carts’. If the participants did not have the associative reading, they would accept sentence
(i) ‘Bear—ne are dragging the carts.’ since ‘bear-ne’ can refer to ‘the bear that is dragging
the cart and his/her associates that is not bear dragging the cart.’ But if the participants do
had the associative reading, they would reject it. For the control sentence (ii) ‘Two bears
are dragging the carts’, the participants should reject it since there is only one bear that is
dragging the cart and no way to create the associative reading present in the sentence.

Therefore, we expect all the participants to reject the control sentences. For the
test sentences, we expect adults to show significantly higher mean value than the mean
value of the control sentences, and 4-year-old children to show no significant difference

between the test sentences and the control sentences.

3.4.3 Results
Descriptive statistics for the participants’ associative reading of the test sentences

and control sentences are presented in Table 8.

47

Table 8.
Descriptive statistics for subject’s associative reading of —ne and cardinal number

 

 

Age Group
Associative reading Adult 4 year-olds 5 year-olds Children
M (SD) M (SD) M (SD) M (SD)
NE (Testing) 0.15(0.23) 0.30 (0.35) 0.37 (0.32) 0.32 (0.34)
Cardinal Number (Control) 0.05 (0.16) 0.14 (0.23) 0.22 (0.32) 0.18 (0.27)
N 60 31 27 58

 

 

The results indicate that adults’ mean response to —ne was (0.15) and to the
cardinal number was (0.05). The children's mean response to —ne was (0.32) and with the
cardinal number, the children’s mean response was (0.18).

First, a between-subjects one way AN OVA analysis was conducted to see if there
were any significant differences between the children’s age groups. The results indicated
that the pattern of responses of 4 year-olds and 5 year-olds did not show a significant
main effect in both testing (F (1, 56) = 0.498, p = .483) and control sentences (F(1, 56) =
2.419, p < .126). We therefore combined the two different age groups and compared them
with adults for further analysis.

For the analysis of associative interpretation, a 2 x 2 mixed design was conducted.

Age group is the between subjects variable with two levels (adults and children); The
existence of the morpheme —ne is the within subjects variable with two levels (—ne and
cardinal number).

The analysis of within subject factors, adults’ results showed a significantly
difference between —ne and the cardinal number (F(1, 59) = 6.268, p = .015). For children,
the responses to —ne and the cardinal number were also significantly different (F (1 , 57) =
7.471,p = .008).

For the between subjects factors, children’s associative interpretation of —ne is

48

significantly higher overall than adult's associative interpretation of —ne (F(1, 116) =
10.869, p < .001). But children’s associative interpretation of the cardinal number is also
significantly higher overall than adult’s (F(1, 116) = 8.737, p < .004). Figure 3 shows
adults and children’s percentage of associative interpretations of —ne and the cardinal

number.

Figure 3.
Percentage of —ne and cardinal number associative interpretation for adults and children.

 

 

100%
80%
60%
H me
40% 330/ o—o Cardinal Number
0
20% 15%? 18%
or 5%
0
Adult Children

3.4.4 Discussion

In the third experiment, the overall participants’ acceptance of the associative
reading was not as high as we expected. For adults, it seems that many adults chose an
alternative interpretation of the associative pluralizer. As has been discussed in chapter 2,
the construction with an associative plural is used to talk about a group already in the

discourse, a group that is understood to be contextually or inherently associated with its

49

named protagonist (V assilieva, 2008). In this experiment, the intention of contrasting the
two different stories was to stimulate the participants to recognize the animals in the same
story as associates to the named animal of the story. However, contrasting the two stories
might not have been strong enough to make them to recognize the animals in the same
story as associates. Vassilieva (2008) says that in the associative interpretation, the
identity of the group represented by the protagonist can be determined from the context,
or, in the absence of contextual evidence, the group will be interpreted as ‘inherently
associated’ with the protagonist (i.e. x and x’s family). Concerning this characteristic of
the associative pluralizers, it is possible for the adults to interpret —ne as ‘inherently
associated’ and make the association of the animals of the same species and not make the
association with the animals in the same story. We assume that the participants might
have chosen the alternative interpretation when the contextual evidence was weaker than
the “species-group” interpretation. The results show that many of the adult participants
failed to make an association of the group in the same story. Although we did not expect
their contextual associative interpretation to be 100%, the results of this experiment were
unexpectedly low. Since forcing them to create only an associative reading was not
realistic, this problem is left as a limitation of testing associativity with the TVJT.

We cannot, therefore, make a strong argument about the children’s interpretation
of associative interpretation from this experiment, since adults’ associative interpretation
cannot work as a baseline. In this experiment, therefore, Hypothesis 3 could not be
confirmed.

Furthermore since 4-year-olds and 5-year-olds did not show any differences,

Hypothesis 2 was not confirmed.

50

In sum, this experiment showed that —ne does have an associative meaning, but
that the inherent interpretation of the associative pluralizer is stronger when the

contextual evidence is weaker.

51

CHAPTER 4

CONCLUSION

We have presented three experiments to investigate how children acquire
optional plural morphemes which have semantically complex properties. We looked at
the two morphemes —deul and —ne. —deul is the morpheme that is associated to a more-
than-one interpretation and has been claimed as giving definite or universal quantification
interpretation, (which would require an exhaustive interpretation) or a specific
interpretation (which does not need to be interpreted as exhausting the set in the
discourse). —ne is an associative pluralizer.

First of all, our experimental results provide empirical evidence to say that the
Korean pluralizer —deul does not force an exhaustive reading. This indicates that the
interpretation of —deul does not meet the requirement to have definite interpretation or
universal quantification. Therefore, this finding goes against Park’s (2008) argument
which says Korean —deul involves universal quantification in the course of interpretation.
The existence of a specificity interpretation associated to —deul as claimed by Kim
(2008) was not disproved from our study but we cannot conclude that —deul is necessarily
specific. Further experiments will help to determine this.

Next, we conclude that variable and unreliable input delays the acquisition of the
plural morpheme when we compare with the acquisition of plurality in a language where
invariable and reliable input exists. In English, the morpheme —s which appears
consistently and reliably in the input is comprehended as indicating more than one by
most three-year-olds (Kouider, Halberda, Wood and Careyet, 2006). Chilean Spanish,

which has variability in the input showed incomplete mastery of plural morphology even

52

at around age five and it confirmed a delay of children’s plural acquisition in comparison
with consistent and reliable plurals (Miller and Schmitt 2009, Miller 2007). In Chinese,
plural —men behaves as definite and gives plurality although bare NPs can also have
plural interpretation. In the study of Chinese pluralizer acquisition (Munn, Zhang and
Schmitt, 2009), 5-6-year-old children did not show adult-like interpretation of the plural
morpheme but they showed a plural and singular distinction at that age. In our study of
pluralizers in Korean, first we expected to see children’s development in understanding
the pluralizers through the different behavior between the age groups. But this was not
found since both 4 and 5-year-olds’ interpretations of the pluralizers were at the same
level. Thus Hypothesis 2, whether age difference impacts the ability to use the pluralizers,
could not be confirmed. However, this study showed that Korean children’s
understanding of pluralizers is not yet mastered by 5-year-olds which is later in
comparison with English plural acquisition. This might indicate that the developmental
process for Korean pluralizers is slower than for Chinese but we need further studies in
both languages using the same methodology in order to determine the exact differences.
We also need to confirm if —deul has further properties than a more-than-one
interpretation such as specificity but the children’s results of our study still supports the
hypothesis 1 that the optionality of the morpheme hindered the mastery of the plural
morpheme in comparison with English speaking children.

The last experiment did not go as we expected and could not confirm Hypothesis
3 creating a group of associates which are not identical is harder than pluralzing identical
items. This result was probably due to the two possible interpretations of the associative

group but from the experiment, we can begin to observe how adults identify the

53

associates of the group with associative pluralizers -ne when the two sources of evidence,
contextual and inherent, exist. The results showed that the ‘inherently associated’
interpretation of an associative pluralizer —ne can be more easily realized when there is
no explicit contextual evidence that can force the listeners to make a contextual
association of the individuals.

In sum, the results of this study suggest that 4-to-6-year old children may not
know that the Korean pluralizers are associated to a more-than-one interpretation. Still
much work needs to be done to understand how children arrive at adult-like interpretation

on such semantically complex morphemes.

54

APPENDIX I

Stories used in Experiment 1

1. Story 1 (—deul)

Bada-e gore du-mari-ga suyeong—eul hago-iss-oss-da.
Ocean-LOC whale two-CL-NOM swim-ACC do-PRE-PST-DEC
‘In the ocean, two whales swam.’

Gore han-mari-nun nasgam-eul-garo tto-nass-da.
Whale one-CL-TOP nap-ACC-for leave-PST—DEC
‘One of them went away to take a nap.’

Nameun gore-nun mul-eul ppum-ess-da.

Lefi whale-TOP water-ACC blow-PST-DEC

‘The other one blew water.’

Given sentences

(i) Gore-deul-i mul-eul ppum-ess-da.
Whale-PL-NOM water-ACC blow-PST-DEC
‘The whales blew water.’

(ii) Gore-deul-i suyeong-eul ha—ess-da.
Whale-PL-NOM swim-ACC do-PST-DEC

‘The whales swam.’

Pic 1. Pic 2. Pic 3.

 

2. Story 2 (—deuI)

Apeulika-e kokkili du-mali-ga banana-leul mok-eoss-da.
Afi'ica-LOC elephant two-CL-NOM banana-ACC eat-PST-DEC
‘In Africa, two elephants ate bananas.’

Geulonde kokkili han-mari-ga bae-ga

But elephant one-CL-NOM Stomach-NOM

bullo-so gib-e ga-ass-da.

full-because home-LOC go-PST-DEC
‘But one of them was full , so he went home.’

55

Nameun kokkili-neun mok-i malla-so
Left elephant-TOP throt-NOM thrist-because

mul-eul mass-ess-da.
water-ACC drink-PST-DEC
‘The other one was thirty, so he drank water.’

Given sentences

(i) Kokkili-deul-i banana-leul mok-eoss-da.
Elephant-PL-NOM banana-ACC eat-PST-DEC
‘The elephants ate banana.’

(ii) Kokkili-deul-i mul—eul mass-ess-da.
Elephant-PL-NOM water-ACC drink-PST-DEC
‘The elephants drank water.’

Pic 3..

     

3. Story 3 (-deul)

Sup-e tokki du-mari-ga pul-eul mok-eoss-da.
Woods-LOC rabbit two-CL-NOM grass-ACC eat-PST-DEC
‘In the woods, there were two rabbits ate grass.’

Gurigo tokki han-mari-ga gib-e ga-ass.da.
And rabbit one-CL-NOM home-LOC go-PST—DEC
‘And one of them went home.’

Nameun tokki-noun keun kkoch-eul chaga

Left rabbit-TOP big flower-ACC found
naemse—leul mat-ass-da.

smell-ACC smell-PST-DEC

‘The other one found big flower and smelled the flower.’

Given sentences

(i) Tokki-deul-i pul-eul mok-eoss-da.
Rabbit-PL-NOM grass-ACC eat-PST-DEC
‘The rabbits ate grass.’

56

(ii) Tokki-deul-i kkoch
Rabbit-PL-NOM flower
‘The rabbits smelled flower.’

Pic 2. ..

    

4. Story 4 (—deul)
Madang-e gae du-mari-ga ppyeo-leul
Backyard-LOC dog two-CL-NOM bone-ACC

‘In the backyard, two dogs played with bones.’

Eolmahu gae han-mari-neun bae-ga

After awhile dog one-CL-TOP stomach-NOM
gib-e ga-ass-da.

home-LOC go-PST-DEC

‘Af’ter awhile, one got hungry, so he went away to eat.’
Nameun gae-neun ppyeo-leul
Left dog-TOP bone-ACC

‘The other one buried his bone in the ground.’

Given sentences

(i) Gae-deul-i ppyeo-leul gagigo
Dog-PL-NOM bone-ACC have
‘The dogs played with bones.’

(ii) Gae-deul-i ppyeo-leul ttang-e
Dog-PL-NOM bone-ACC ground-LOC
‘The dogs buried the bone.’

naemse-leul
smell-ACC

mat-ass-da.
smell-PST-DEC

 

gagi go nol-ass-da.
have play-PST—DEC
gopaso

felt hungry

ttang-e mud-ess—da.
ground-LOC bury-PST-DEC
nol-ass-da.

play-PST-DEC

mud-ess-da.
bury-PST-DEC

 

57

5. Story 5 (-ne)

Yeomso du-mari-ga gib-e sal-ass-da.
Goat two-CL-NOM home-LOC live-PST—DEC
‘Two goats are living in the house.’

Eoneunal yeonso—neun dengsan-eul ga-ass-da.

One day goat-TOP mountain climbing-ACC go-PST-DEC

‘One day they went out to climb the mountain.’

Yeomso han-mari-neun pigonhae-so gib-e ga-ass-da.
Goat one-CL-TOP tired-because home-LOC go-PST-DEC

‘One goat came back home because he is tired.’

Given sentences

(i) Yeomso-ne-neun dengsan-eul ga-ass-da.
Goat-PL-TOP mountain climbing-ACC go-PST-DEC
‘The Goats went out to climb the mountain.’
(ii) Yeomso-ne-neun gib-e ga-ass-da.
Goat-PL-TOP home-LOC go-PST-DEC
‘The goats came back home.’
Pic 1. Pic 2. . Pic . ...

"War-Pr: -‘ ‘ - $7339 ~.'-.1"'~".’= : .

 

     

6. Story 6 (—ne)

Paendo du-mari-ga mudae-eso norae-leul ha-ess-da.
Panda two—CL-NOM stage-LOC sing-ACC do-PST-DEC
‘Two pandas were singing in a band.’

Keun paendo-neun gita-ga burogyeo-so
Big panda-TOP guitar-NOM broken-because

sac gita-leul sa-ro naga-ass-da.
new guitar-ACC buy-to go out-PST-DEC
‘Big panda broke his guitar so he went out to buy a new one.’

Gageun paendo-neun honga gita-leul chye-oss-da.

Little panda-TOP alone guitar-ACC play-PST-DEC
‘Little panda played guitar alone.’

58

Given sentences

(i) Paendo-ne-neun norae-leul ha-ess-da.
Panda-TOP sing-ACC do-PST-DEC
‘The pandas sang a song.’

(ii) Paendo-ne-neun gita-leul chye-oss-da.
Panda-TOP guitar-ACC play-PST-DEC
‘The pandas played the guitar.’

     

7. Story 7 (-ne)
Gamgali du-mari-ga haneul-eul ppaleugae nalgo-iss-oss-da.
Dragonfly two-CL-NOM sky-ACC fast fly-PRE-PST-DEC

‘Two dragonflies are flying fast in the sky.’

Agi gamgali-neun shuil-yeogo gureum-e ang-ass-da.
Baby dragonfly-TOP take a rest-to cloud-LOC sit-PST-DEC
‘Baby dragonfly sat on the cloud to take a rest.’

Namneun gamgali-neun namugagi-e ang-ass-da.
Lefi dragonfly—TOP tree-LOC sit-PST-DEC
‘The other dragonfly hanging on the branch.’

Given sentences

(i) Gamgali-ne-neun ppaleugae nalgo-iss-oss-da.
Dragonfly-PL-TOP fast fly-PRE-PST-DEC
‘The dragonflies flew fast.’

(ii) Gamgali-ne-neun gureum-e ang-ass-da.
Dragonfly-PL-TOP cloud-LOC sit-PST-DEC

‘The dragonflies set on the cloud.’

Pic 2. Pic 3.

   

59

8. Story 8 (—ne)

Mulgogi du-mari-ga hon gib—e sal-ass-da.
Fish two—CL—NOM old house-LOC live-PST—DEC
‘Two fishes live in an old house.’

Mulgogi han-mari-neun mun-eul gochi-lyeogo

Fish one-CL-NOM door-ACC , fix-to

miyeok-eul gagyeowa-ass-da.

seaweed-ACC bring-PST-DEC
‘One fish brought seaweed to fix the door.

Nameun mulgogi-neun byeok-eul gochi-lyeogo
dol-eul gagyeowa-ass-da.

Stone-ACC bring-PST-DEC

‘The other fish brought a stone to fix a wall.

Given sentences

(i) Mulgogi-ne-neun hon gib-e sal-ass-da.
Fish-PL-TOP old house-LOC live-PST-DEC
‘The fishes were living in the old house.’

(ii) Mulgogi-ne-neun miyeok-eul gagyeowa-ass-da.
Fish-PL-TOP seaweed-ACC bring-PST-DEC

‘The fishes brought seaweed.’

Pic 2. Pic 3.

    

60

  

APPENDIX II

Stories used in Experiment 2

1. Story 1
Ako dasos—mari-ga gang-eso suyeong-eul ha-ess-da.
Alligator five-CL-NOM river-LOC swim-ACC do-PST-DEC
‘5 alligators were swimming together in a river.’
Gamsihu ako ne-mari-ga tongnamu-eso swi-oss-da.
After a while alligator four-CL-NOM log-LOC rest-PST-DEC
‘After a while, four alligators rested on a log.’
Ako han-mari-neun gyesok suyeong-eul ha-ess-da.
Alligator one-CL-NOM keep swim-ACC do-PST-DEC
‘One alligator kept swimming.’
Given sentences
(i) Ako-deul-i tongnamu-eso swi-oss-da.

Alligator log-LOC rest-PST-DEC

‘The alligators rested on a log’
(ii) Ako-deul-i modu tongnamu-eso swi-oss-da.

Alligator all log-LOC rest-PST-DEC

‘The alligator all rested on a log’

 

 

2. Story 2

Yeonmos-e oli dasos-mali-ga suyeong—eul ha-ess-da.
Pond-LOC duck five-CL-NOM swim-ACC do-PST-DEC
‘In the pond, there were five ducks swimming.’

Onenal bi-ga wa-seo oli ne-mali-neun
One day rain-NOM come-because duck four-CL-TOP
usan-eul ss-oss-da.

umbrella-ACC put-PST-DEC

‘One afternoon, it started raining. 4 ducks used an umbrella.’

61

Georonde oli han-mali-neun bi-ga goa-seo
But duck one-CL-TOP rain-NOM come-because

usan-eul sseu—jian-ass-da.
umbrella-ACC put-NEG-PST-DEC

‘But 1 duck likes rains, so he didn’t use an umbrella.’

Given sentences

(i) Oli-deul-i usan-eul ss-oss-da.
Duck-PL-NOM umbrella-ACC put-PST-DEC
‘The Ducks used umbrellas.’

(ii) Oli-deul-i modu usan-eul ss-oss-da.
Duck-PL-NOM all umbrella-ACC put-PST-DEC

‘The ducks all used umbrellas.’

Pic 1.

   

3. Story 3

Gongwon-e nabi dasos-mali-ga nalgo iss-oss-da.
Garden-LOC butterfly five-CL-NOM fly exist-PST-DEC
‘In the garden, there were 5 butterflies were flying.’

Geomsim-e bae-ga gopa-seo nabi ne-mari-neun

Noon-At stomach-NOM hungry-because butterfly four-CL-TOP
kkoch-eso kkul-eul mog-oss-da.

flower-LOC honey-ACC eat-PST-DEC
‘At noon, they got hungry so 4 butterflies were on flowers to suck nectars.’

Georonde nabi han-mali-neun kkoch-i

But butterfly one-CL-TOP flower-NOM
obso-seo aiseukeulim-eul mog—oss-da.

not exist-because ice cream-ACC eat-PST-DEC

‘But 1 butterfly couldn’t find a flower so she ate ice cream.’

Given sentences

(i) Nabi-deul-i kkoch-eso kkul-eul mog-oss-da.
Butterfly-PL-NOM flower-LOC honey-ACC eat-PST-DEC
‘The butterflies suck nectar from flowers.’

62

(ii) Nabi-deul-i modu kkoch-eso kkul-eul
Butterfly-PL—NOM all flower-LOC honey-ACC

‘The butterflies all suck nectar from flowers.’

Pic 1.

 

4. Story 4
Namgeuk-e paengguin dasos-mali-ga iss-oss-da.
Antarctica-LDC penguin five-CL-NOM EX-PST-DEC
‘In Antarctica, there were 5 penguins.’
Oneu gyeoul paengguin ne-mali-ga al-eul pum-oss-da.
One winter penguin four-CL-NOM egg-ACC warm-PST-DEC
‘In one winter, 4 penguins kept their eggs warm.’
Geuronde paengguin han-mali-nun sagwa-leul pum-oss-da.
But penguin one-CL-TOP apple-ACC warm-PST-DEC
‘But 1 penguin didn’t have an egg, so she warmed an apple.’
Given sentences
(i) Paengguin-deul-i al-eul pum-oss-da.
Penguin-PL-NOM egg-ACC warm-PST-DEC
‘The penguins warmed the eggs.’
(ii) Panguin-deul-I modu al-eul pum-oss-da.
Penguin-PL-NOM all egg-ACC warm-PST-DEC
‘The penguins all warmed the eggs.’
Pic 2

Pic 1.
‘5‘

Pic 2.

mog-oss-da.
eat-PST-DEC

 

PI" is» 3 Wfi

I
J

.....- ..._.__.- ._._.. .L

63

APPENDIX 111

Stories used in Experiment 3

1. Story 1

Gom du-mali-wa mal se-mali-ga nonggang-e iss-oss-da.
Bear two-CL-and horse three-CL-NOM farm-LOC EX-PST-DEC
‘2 bears and 3 ponies were in the farm.’

Modu chinhan chingu—da.

All close friend-DEC

‘They all are close friends.’

Gom han-mali-wa mal se-mali-neun sule-leul dang-geoss-da.
Bear one-CL-and horse three-CL-NOM cart-ACC drag-PST-DEC
‘1 bear and 3 ponies dragged a cart.

Nameun gom han-mali-neun sule-leul mil-eoss-da.
Left bear one-CL-NOM cart-ACC push-PST-DEC
‘The other bear pushed a cart.’

Given sentences

(i) Gom-ne-ga sule-leul dang-geoss-da.
Bear-PL-NOM cart-ACC drag-PST-DEC
‘Bear-ney dragged a cart’

(ii) Gom du-mali-ga sule-leul dang-geoss-da.
Bear two-CL-NOM cart-ACC drag-PST-DEC
‘2 bear (bare noun) dragged a cart.’

Pic 1. Pic 2.
i
r .3 v .1
i . ‘h 3
r ;l
- :Q-smmrsw " '5‘

 

2. Story 2
Yang du-mali-wa doaegi se-mali-ga nonggang-e iss—oss-da.
Sheep two-CL-and pig three-CL-NOM farm-LOC EX-PST-DEC

‘Two sheep and three pigs were in the farm.’

Modu chinhan chingu—da.

All close friend-DEC
‘They all are close friends.’
Yang han-mali-wa doaegi se-mali-neun pungson-eul gap-ass-da.

Sheep one-CL-and pig three-CL-NOM ballon-ACC held-PST-DEC
‘One day, one sheep and three pigs held a balloon to play.’

Nameun yang han-mali-neun kkoch-eul gap-ass-da.
Left sheep one-CL-NOM flower-ACC held-PST-DEC
‘The other sheep held a flower.’

Given sentences
(i) Yang-ne-ga pungson-eul gap-ass-da.
Sheep-PL-NOM ballon-ACC held-PST-DEC

‘The sheeps held a ballon.’
(ii) Yang du-mali-ga pungson-eul gap-ass-da.
Sheep two-CL-NOM ballon—ACC held-PST-DEC
‘2 sheep held a ballon.’
Pic 1.

   

3. Story 3

Se du-mali-wa dolgole se-mali-ga bada-eso
Bird two-CL-and dolphin three-CL—NOM ocean-LOC
gompeu-leul ha-ess-da.

jump-ACC do—PST-DEC

‘Two birds and three dolphins were jumping in the ocean.’

Modu chinhan chingu-da.

All close friend-DEC

‘They all are close friends.’

Geuronde se han-mali-wa dolgole se-mali-neun

But bird one-CL-and dolphin three-CL-TOP
gompeu-leul momchu-go suyeong-eul ha-ess-da.

jump-ACC stop-COMP swim-ACC do-PST-DEC

‘But one bird and three dolphins stopped jumping and they swam.’

65

Namneun se han-mali-neun
Lefi bird one-CL-TOP
‘The other bird rested on a tree.’

Given sentences

(i) Se-ne-ga suyeong-eul
Bird-PL-NOM swim-ACC
‘The birds were swimming.’

(ii) Se du-mali-ga
Bird two-CL-NOM
‘2 bird were swimming.’

 

4. Story 4
Ttokki du-mali-wa dalamgwi
Rabbit two-CL-and squirrel

namu-eso swi-oss-da.
tree-LOC rest-PST-DEC
ha-ess-da.

do-PST-DEC

suyeong-eul ha—ess-da.

swim-ACC do-PST-DEC

 

se-mali-ga sup-e iss-oss-da.
three-CL-NOM wood-LOC EX-PST-DEC

‘Two rabbits and three squirrels were in the words.

Modu chinhan chingu-da.

All close friend-DEC

‘They all are close friends.’

Oneunal Ttokki han-maIi-wa dalamgwi se-mali-ga

One day rabbit one-CL-and squirrel three-CL-NOM
ttangkong-eul mog-oss-da.

acom-ACC eat-PST-DEC

‘One day, one rabbit and three squirrels ate an acorn.

Ttokki han-mali-neun danggeun-eul mog—oss-da.
Rabbit one-CL-TOP carrot-ACC eat-PST-DEC

‘The other rabbit ate a carrot.

Given sentences

(i) Ttokki-ne-ga ttangkong-eul
Rabbit-PL-NOM acom-ACC
‘The rabbits ate acorn.’

mog-oss-da.
eat-PST-DEC

66

(ii) Ttokki du-mali-ga ttangkong-eul mog—oss-da.
Rabbit two-CL-NOM acorn-ACC eat-PST—DEC
‘2 rabbits ate acorn.’

Pic 2.

      

67

APPENDIX IV

Filler sentences used in Experiment 1

1. Story 1

(i) Gore-ga bada—e iss-ess-da.
Whale-NOM ocean-LOC EX-PST-DEC
‘The whales were in the ocean.’

(ii) Gore du-mari-ga naggam-eul ga-ro ga-ass-da.
Whale two-CL-NOM nap-ACC sleep-to go-PST-DEC
‘The two whales went to take a nap.’

2. Story 2

(i) Kokkili-ga apeulika—e iss-eoss-da.
Elephant-NOM africa-LOC EX-PST-DEC
‘The elephants were in Africa.’

(ii) Kokkili du-mari-ga sala-gess-da.
Elephant two-CL-NOM disappear-PST-DEC
‘The two elephants disappeared.’

3. Story 3

(i) Tokki han-mali-ga gib-e ga-ass-da.
Rabbit one-CL-NOM home-LOC go-PST-DEC
‘One rabbit went home.’

(ii) Tokki-ga bada—e iss—eoss-da.
Rabbit-NOM ocean-LOC EX—PST—DEC
‘The rabbits were in the ocean.’

4. Story 4

(i) Gae han-mali-ga gib-e ga-ass-da.
Dog one-CL-NOM home-LOC go-PST-DEC
‘One dog went home.’

(ii) Gae-ga gib-an-e iss-ess-da.
DogNOM house-inside-LOC EX-PST-DEC

‘The dogs were in the house.’

68

5. Story 5

(i) Yeomso-ga gib-e
Goat-NOM house-LOC
‘The Goats were living in a house.’
(ii) Yeomso du-mari-neun
Goat two-CL-NOM
‘The goats were tired.’
6. Story 6
(i) Paendo-ga mudae-e
Panda-NOM Stage-LOC
‘The pandas were on the stage.’
(ii) Paendo du-mali-ga
Panda two-CL-NOM
‘The two pandas went out to buy a guitar.’
7. Story 7
(i) Agi gamgali-neun
Baby dragonfly-TOP
‘The baby dragonfly sat on the clouds.’
(ii) Gamgali du-mali-neun
Dragonfly two-CL-TOP
‘The two dragonflies liked the clouds.’
8. Story 8
(i) Mulgogi-neun hon
F ish-PL-TOP old
‘The fishes were living in the old house.’
(ii) Mulgogi du-mali-neun
Fish two-CL-TOP
‘The two fishes fixed the wall.’

69

sal-ass-da.
live-PST-DEC

pi gon-haess-da.
tired-PST-DEC

iss-oss—da.

EX-PST-DEC

gita-leul
guitar-ACC

gureum-e
cloud-LOC

gureum-eul
cloud-ACC

gib-e

house-LOC

byeok-eul
wall-ACC

sa-ro ga-ass-da.
buy-to go-PST-DEC

ang-ass-da.
sit-PST—DEC

goa-haess-da.
like-PST-DEC

sal-ass-da.

live-PST-DEC

goch-yess.da.
fix-PST-DEC

APPENDIX V

Filler sentences used in Experiment 2

1. Story 1

(i) Ako-neun gang-e
Alligator-TOP river-LOC
‘The alligators were in the river.’

(ii) Ako-neun pigon-heass-da.
Alligator-TOP tired-PST-DEC
‘The alligator were tired.’

2. Story 2

(a) Oli-ga yeonmos-e
Duck-PL-NOM pond-LOC
‘The ducks were in the pond.’

(b) Oli yeososs-mali-ga
Duck six-CL—NOM
‘The ducks were total of six.’

3. Story 3

(i) Nabi-neun bae—ga
Butterfly—TOP stomach-NOM
‘The butterflies were hungry.’

(ii) Nabi-neun bang-e
Butterfly-TOP house-LOC
‘The butterflies were inside of the house.’

4. Story 4

(i) Paengguin-eun Nam geuk-e
Penguin-TOP antarctica-LOC
‘The penguins were living in Antarctica.’

(ii) Panguin yeososs-mali-ga
Penguin six-CL-NOM

‘The penguins were total of six.’

70

iss-oss-da.
EX-PST—DEC

iss-oss-da.
EX-PST-DEC

iss-oss-da.
EX-PST-DEC

gopa-ass-da.
houngry-PST-DEC

iss-oss-da.
EX-PST-DEC

sal-ass-da.

live-PST-DEC

iss-oss-da.
EX—PST-DEC

APPENDIX VI

Filler sentences used in Experiment 3

1. Story 1

(i) Gom-gwa mal-eun nonggang-e iss—oss-da.
Bear-and horse-TOP farm-LOC EX-PST-DEC
‘The bears and horses were in the farm.’

(ii) Mal-eun chingu-ga ob-oss-da
Horse-TOP friend-NOM NonEX--PST-DEC
‘The horse did not have friends.’

2. Story 2

(i) Yang-gwa doaegi-neun chingu-yeoss-da.
Sheep-and pig-TOP friend-PST-DEC
‘The sheep and pigs were friends.’

(ii) Yang-gwa doaegi-neun uli-e iss-oss-da.
Sheep-and pig-TOP hut-LOC EX-PST-DEC
‘The sheep and pigs were inside of the hut.’

3. Story 3

(i) Se-wa dolgole-ga gompeu-leul ha-ess-da.
Bird-and dolphin-NOM jump-ACC do-PST-DEC
‘The birds and the dolphins jumped.’

(ii) Se—wa dolgole-ga san-e iss-oss-da.
Bird-and dolphin-NOM mountain-LOC EX-PST-DEC
‘The birds and the dolphins were in the mountain.’

4. Story 4

(i) Ttokki-wa dalamgwi-neun sup-e iss—oss-da.
Rabbit-and squirrel-TOP mountain-LOC EX—PST—DEC
‘The rabbits and the squirrels were in the mountain.’

(ii) Ttokki-neun chingu-ga ob-oss-da

Rabbit-TOP friend-NOM NonEX--PST-DEC
‘Rabbit did not have friends.’

71

REFERENCES

An, Y. (2007). Korean tul and English all. LSO Working Papers in Linguistics 7:
Proceedings of WIGL. 1-16.

Baek, M. (2002). A study on Korean plural senses. Discourse and Recognition, Seoul,
South Korea, 9 (2), 59—78.

Bamer, D., and Snedeker, J. (2005). Quantity judgments and individuation: Evidence that
mass nouns count. Cognition, 97, 41-66.

Borer, H. (2005). Structuring sense. Vol 1. In Name Only. Oxford: Oxford University
Press.

Brown, R. (1973). A first language: The Early Stages. Cambridge, MA: Harvard
University Press.

Bybee, J. (1985). Morphology: A Study of the Relation between Meaning and F orm. John
Benjamins Publishing Co.

Chierchia, G. (1998). Reference to kinds across languages. Natural Language Semantics,
6, 339-405.

Chung, S. (2000). On reference to kinds in Indonesian. Natural Language Semantics, 8,
157-171.

Corbett, G. (2000). Number. Cambridge: Cambridge University Press.

Crain and Thornton. (1998). Investigations in Universal Grammar: a guide to
experiments on the acquisition of syntax and semantics. Cambridge, MA: MIT Press.

Dressler, W. (1989). Prototypical differences between inflection and derivation.

Zeitschrift fu"r Phonetik, Sprachwissenschaft und Kommunikationsforschung, 42,
3—10.

Ferenz, K. S., and Prasada, S. (2002). Singular or plural? Childrenfs knowledge of the
factors that determine the appropriate form of the count nouns. Journal of Child
Language, 29, 49—70.

Gordon, Peter. (1996). The truth-value judgment task. In Methods for Assessing

Children’s Syntax. (Ed.), By Dana McDaniel, Cecil McKee, and Helen Cairns, 211-
232. Cambridge, MA: MIT Press.

72

Im, H. (2000). The plural marker —tul and eventuality. Aysanhakpo 24, Seoul, South
Korea.

Joh, Y. (2009). Plurality and Distributivity. Dissertation Abstracts International, A: The
Humanities and Social Sciences, 69, 09, 3527.

Kang, B. (1994). Plurality and other semantic aspects of common nouns in Korean.
Journal of East Asian Linguistics, 3, 1—24.

Kim, C. (2005). Order and Meaning: Numeral Classifiers and Specificity in Korean. In
Proceedings of the 24th West Coast Conference on Formal Linguistics, (Ed.), John
Alderete et al., 218-226. Somerville, MA: Cascadilla Proceedings Project.

Kim, C. (2005). The Korean Plural Marker tul and Its Implication, Dissertation Abstracts
International, A: The Humanities and Social Sciences, 65, 12, June, 4543-A.

Kim, J. (2008). The Parameterization of Plural Markings in Classifier Languages.
Kansas Working Papers in Linguistics, 30, 159.

Kouider, 8., Justin H., Justin W and Susan C. (2006). Acquisition of English number

marking: T he singular—plural distinction. Language Learning and Development, 2
(1), 1—25.

Kwak, E. (2003). Interpretation of plural noun phrases in Korean. Linguistics, Seoul,
South Korea, 35, 3—38.

Longobardi, G. (1994). Reference and proper names: A theory of N movement in syntax
and logical form. Linguistic Inquiry, 25(4), 609-665.

Longobardi, G. (1996). The Syntax of N-rising: A Minimalist Theory. OTS Working
Papers, Research Institute for Language and Speech, Utrecht University.

Madigan, S., Yamada M and Peng A. (2008). Inclusive Reference in Korean: Implications
for a Theory of Plurality. Proceedings of Harvard ISOKL 2007.

Miller, K and Cristina S. (2009). Variable vs. Consistent Input: Comprehension of
Plural Morphology and Verbal Agreement in Children. Merging Features. Oxford
Press.

Miller, K. (2007). Variable input and the acquisition of plurality in two varieties of
Spanish. Ph D. Dissertation, Michigan State University.

Moravcsik, E. (2003). A semantic analysis of associative plurals. Studies in Language, 27
(3), 469-503.

73

Munn, A., Xiaofei, Z and Cristina, S. (2009). Acquisition of plurality in a language
without plurality’“. Merging Features. Oxford Press.

National Academy of the Korean Language. Seoul: Ministry of Culture & Tourism.
(2000). The Revised Romanization of Korean.

Nakanishi, K and Ritter, E. (2008). Plurality in Languages without a count-mass
distinction. Mass/Count Workshop, University of Toronto.

Park, S. (2008). Plural marking in classifier languages: a case study of so-called plural
marking -tul in Korean. Toronto Working Papers in Linguistics, 28, 281-295.

Stowell, T. (1989). Subjects, Specifiers and X-bar theory. In M. R. Baltin and A. S.
Kroch (Eds), Alternative Conceptions in Phrase Structure, 232-262.

Vassilieva, M. (2008). A syntactic analysis of nominal and pronominal associative plurals.
Preceedings of the 315‘ Annual Penn Linguistics Colloquium, 14 (1), 26

Yang, C. (2005). On productivity. Yearbook of Language Variation, 5, 333-370.

Zapf, Jennifer A., and Smith, Linda B. (2009). Knowing more than one can say: The early
regular plural*. J. Child Language, 36, 1145-1155.

74

MICHIGAN STATE UNIVERSITY LIBRARIES

3 1293 03063 7759