Professional Documents
Culture Documents
Mandarin
Development
of Indonesian
Immigrants’
Children
A Longitudinal Study in Taiwan
Mandarin Development of Indonesian
Immigrants’ Children
Jenny Yi-Chun Kuo
Mandarin Development
of Indonesian Immigrants’
Children
A Longitudinal Study in Taiwan
123
Jenny Yi-Chun Kuo
National Chiayi University
Chiayi, Taiwan
Taiwan
v
Acknowledgements
I would like to thank the Ministry of Science and Technology for the grants to
conduct the 3-year study and write up the book. I am also grateful for my research
assistants, parents and participants of the study, and valuable comments from Prof.
James Tai and anonymous reviewers from Springer.
vii
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Previous Research on Children of Immigrants in Taiwan . . . . . . . 1
1.3 Mandarin Versus Vietnamese Versus Indonesian . . . . . . . . . . . . . 3
1.4 Research on Immigrants in Other Countries . . . . . . . . . . . . . . . . 4
1.5 Learning Chinese as a Second Language . . . . . . . . . . . . . . . . . . 5
1.6 Language Acquisition Theories . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.7 The Present Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.7.1 The Two-Year-Old Indonesian Girl . . . . . . . . . . . . . . . . 6
1.7.2 The Three-Year-Old Indonesian Boy . . . . . . . . . . . . . . . 7
1.7.3 The Three-Year-Old Indonesian Girl . . . . . . . . . . . . . . . 7
1.7.4 The Four-Year-Old Indonesian Boy . . . . . . . . . . . . . . . . 7
1.7.5 The Four-Year-Old Indonesian Girl . . . . . . . . . . . . . . . . 8
1.7.6 The Five-Year-Old Indonesian Boy . . . . . . . . . . . . . . . . 8
1.7.7 The Five-Year-Old Indonesian Girl . . . . . . . . . . . . . . . . 8
1.7.8 The Six-Year-Old Indonesian Boy . . . . . . . . . . . . . . . . . 9
1.7.9 The Six-Year-Old Indonesian Girl . . . . . . . . . . . . . . . . . 9
1.7.10 The Two-Year-Old Taiwanese Boy . . . . . . . . . . . . . . . . 9
1.7.11 The Two-Year-Old Taiwanese Girl . . . . . . . . . . . . . . . . 10
1.7.12 The Three-Year-Old Taiwanese Boy . . . . . . . . . . . . . . . 10
1.7.13 The Three-Year-Old Taiwanese Girl. . . . . . . . . . . . . . . . 10
1.7.14 The Four-Year-Old Taiwanese Boy . . . . . . . . . . . . . . . . 10
1.7.15 The Four-Year-Old Taiwanese Girl . . . . . . . . . . . . . . . . 10
1.7.16 The Five-Year-Old Taiwanese Boy . . . . . . . . . . . . . . . . 11
1.7.17 The Five-Year-Old Taiwanese Girl . . . . . . . . . . . . . . . . 11
1.7.18 The Six-Year-Old Taiwanese Boy . . . . . . . . . . . . . . . . . 11
1.7.19 The Six-Year-Old Taiwanese Girl . . . . . . . . . . . . . . . . . 11
1.8 Organization of the Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
ix
x Contents
2 Phonological Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.1 General Literature on Phonological Development. . . . . . . . . . . . . 17
2.1.1 Theories of the Acquisition of Phonology . . . . . . . . . . . . 17
2.1.2 Phenomena in Phonological Development. . . . . . . . . . . . 23
2.1.3 Adult role in Phonological Development . . . . . . . . . . . . 24
2.2 Phonology of Mandarin, Taiwan Southern Min,
Vietnamese, and Indonesian . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2.1 Phonology of Mandarin . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2.2 Phonology of Taiwan Southern Min. . . . . . . . . . . . . . . . 26
2.2.3 Phonology of Vietnamese . . . . . . . . . . . . . . . . . . . . . . . 28
2.2.4 Phonology of Indonesian . . . . . . . . . . . . . . . . . . . . . . . 30
2.3 Mandarin Phonological Development by Native Children . . . . . . . 31
2.3.1 Development of Mandarin consonants . . . . . . . . . . . . . . 31
2.3.2 Development and Acquisition of Vowels
and Final Nasals . . . . . . . . . . . . . . . . . . . . . . . . . . ... 32
2.3.3 Development and Acquisition of Tones and Syllables ... 33
2.4 Mandarin Phonological Development by Children
of Vietnamese . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 34
2.4.1 Mandarin Phonological Errors of Vietnamese
Mothers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 34
2.4.2 Mandarin Phonological Development by Children
of Vietnamese Mothers. . . . . . . . . . . . . . . . . . . . . . ... 34
2.5 Mandarin Phonological Development by Children
of Indonesian Mothers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.5.1 The Two-Year-Old Indonesian Girl . . . . . . . . . . . . . . . . 36
2.5.2 The Three-Year-Old Indonesian Boy . . . . . . . . . . . . . . . 41
2.5.3 The Three-Year-Old Indonesian Girl . . . . . . . . . . . . . . . 42
2.5.4 The Four-Year-Old Indonesian Boy . . . . . . . . . . . . . . . . 51
2.5.5 The Four-Year-Old Indonesian Girl . . . . . . . . . . . . . . . . 52
2.5.6 The Five-Year-Old Indonesian Boy . . . . . . . . . . . . . . . . 61
2.5.7 The Five-Year-Old Indonesian Girl . . . . . . . . . . . . . . . . 65
2.5.8 The Six-Year-Old Indonesian Boy . . . . . . . . . . . . . . . . . 69
2.5.9 The Six-Year-Old Indonesian Girl . . . . . . . . . . . . . . . . . 73
2.5.10 The Two-Year-Old Taiwanese Boy . . . . . . . . . . . . . . . . 77
2.5.11 The Two-Year-Old Taiwanese Girl . . . . . . . . . . . . . . . . 82
2.5.12 The Three-Year-Old Taiwanese Boy . . . . . . . . . . . . . . . 84
2.5.13 The Three-Year-Old Taiwanese Girl. . . . . . . . . . . . . . . . 86
2.5.14 The Four-Year-Old Taiwanese Boy . . . . . . . . . . . . . . . . 87
2.5.15 The Four-Year-Old Taiwanese Girl . . . . . . . . . . . . . . . . 92
2.5.16 The Five-Year-Old Taiwanese Boy . . . . . . . . . . . . . . . . 93
2.5.17 The Five-Year-Old Taiwanese Girl . . . . . . . . . . . . . . . . 99
2.5.18 The Six-Year-Old Taiwanese Boy . . . . . . . . . . . . . . . . . 101
2.5.19 The Six-Year-Old Taiwanese Girl . . . . . . . . . . . . . . . . . 103
2.6 Discussion and Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Contents xi
xiii
List of Tables
xv
xvi List of Tables
xix
Chapter 1
Introduction
1.1 Background
Research has found that many children of foreign mothers delayed in their devel-
opment, and language impairment seems to be the major cause (Wu 2004; Lin
2003; Ke 2004; Mo and Lai 2004; Chen and Zhang 2003; Chen and Chen 2003;
Wang 2004).
Vietnamese and Taiwanese mothers in the MLU for age 2–5. The difference
showed up at the age of 6. The MLU for the six-year-old of a Taiwanese mother
reached 6.37, while that of the Vietnamese age-mate remained 3.62. The sentence
patterns used by children of Vietnamese mothers were not fewer than those used by
children of Taiwanese mothers (Kuo 2008).
Children of immigrants from different countries can have different difficulties
when acquiring Mandarin. Although a study of language development of children
of Vietnamese mothers has been conducted (Kuo 2008), it cannot represent the
language development of all immigrant children. Immigrants’ mother tongues can
influence their acquisition of Mandarin and in turn influence their children’s lan-
guage development. Children with Vietnamese mothers were found to have unique
phonological problems and problems common in children with Taiwanese mothers
(Kuo 2008). The present study investigated Mandarin development by children of
27,036 Indonesian mothers, the second largest non-Mandarin-speaking immigrant
group in Taiwan.
Other developed countries have also faced problems of immigrant language learning.
For example, the USA is a country consisting of immigrants of various language and
culture background. The research on immigrant language has focused on the diversity
that the immigrants have brought to the society, the maintenance of immigrants’
mother tongues, and learning the language of the immigrant country (Mckay and
Wong 2000). According to the research in Canada, the use of mother tongue did not
help immigrants learn the language of the immigrant country. The proficiency of the
target language was predicted by the length of language programs and the time of
staying in the immigrant country (Fazio and Stens 1994). The role of parents plays an
important role in the language development of children of immigrants, for example,
parents’ attitude to the language, parents’ interest in the mother tongue and the local
language, and parent–child activities and conversations (Mushi 2002).
Although some of the immigrants in the USA are highly educated, different from
the foreign brides in Taiwan, the Latino workers in the USA have similar back-
ground with workers or foreign brides from Southeast Asia in Taiwan. The situation
of children of foreign brides in Taiwan is similar to children of immigrants in the
USA. Children of immigrants usually communicated in their parents’ mother ton-
gue, which is not English at home. It was hard for them to communicate with people
when they entered nursery school. They were often mistaken as cognitive deficit for
the failure of following the teacher’s instruction. However, they were fluent in the
language of the immigrant country after a few months (Kowalsky et al. 2003).
Why can’t children of immigrants in Taiwan catch up with their peer in the ele-
mentary school? The children of immigrants in the USA entered nursery school and
interacted with English-speaking children at the age of two. Children of immigrants in
Taiwan usually go to kindergarten or elementary school at five or six. Although it has
been claimed that the critical period for language acquisition is from birth to puberty,
the earlier, the better (Elman 1991). Besides, children in the nursery schools in the
USA have a lot of free time to interact with each other, through which language is
developed. The education in Taiwan focuses on collectivism. Students usually listen
to their teacher’s instruction and have little time to interact individually. Due to the
large class size, the teacher has limited time to interact with individual student. In
addition, the children of immigrants in the USA receive the standard forms of parents’
mother tongue. One the other hand, the foreign mothers are usually not allowed to
communicate with their children in their mother tongues. What their children receive
is their limited Mandarin input influenced by their native language, similar to pidgin.
The characteristic of a pidgin is simplified syntax and limited lexicon. But, when it is
acquired as the first language, the lexicon and syntax increase in amount and com-
plexity and become a full-fledged creole. Is the language spoken by children offoreign
mothers a pidgin of their parents’ language? Has it been creolized? Is this the reason
why their language is difficult to understand? This study investigated this special kind
of language. It will not only improve our understanding in creole but also have
implications for language acquisition.
1.5 Learning Chinese as a Second Language 5
Many minority children in Mainland China learning Chinese as a second language are
in a similar situation as children of immigrants in Taiwan. The major difference is that
they use their minority language at home. On the other hand, Indonesian mothers in
Taiwan do not usually speech Indonesian with their children. Children of immigrants in
Taiwan acquire Chinese as their first language. Gulxa et al. (2014) investigated 10
Kazak Chinese bilinguals in their first three years and found their development followed
a regular development process of the two languages: from concrete to abstract and from
simple to complex. There is an L1 transfer of Kazak Syntax to Chinese, while there is no
influence of Chinese on Kazak. They attributed this phenomenon to the predominant
use of Kazak in the families and the drastic linguistic differences between the two
languages. Sun (2010) reported the effect of school environment on minority
preschoolers’ bilingual development. There were bilingual class with all Uygur chil-
dren instructed by a Uygur Chinese bilingual teacher and Chinese Uygur mixed class,
where Chinese and Uygur children were in the same class. The instruction was given in
Chinese, and the environment was decorated in Chinese. The mixed class suppressed
the bilingual class in Chinese pronunciation, vocabulary, and syntax due to better
language input in quality and quantity. The bilingual teachers were less accurate in
Chinese pronunciation and the mixed class used Chinese more with their peer, while the
bilingual class communicate with their peer in Uygur. Input is very important for
language acquisition. The foreign mothers provided even worse Chinese input than
bilingual teachers for their children in Taiwan.
There are three major language acquisition theories: behaviorism, innatism, and
interactionism. Behaviorism treated language learning as habit formation by imitation
and reinforcement (Skinner 1957). Immigrants with limited language proficiency may
not provide a good model for children to imitate and may not be capable of providing
reinforcement. Therefore, behaviorism will predict that children of immigrants may
delay in their development. With poverty of stimulus and lack of negative evidence,
Chomsky (1959) argued that people are equipped with Universal Grammar to acquire
the language in their environment. This is the called innatism. According to innatism,
children of immigrants should acquire Mandarin normally, because they are
immersed in Mandarin environment. Vygotsky (1978) proposed that language is
learned by social interaction, which is supported by Sachs and Johnson’s (1981) study
that hearing children of deaf parents did not acquire English by watching TV. The
importance of child-directed speech is emphasized. Preschool children interact with
their primary caretakers, who were Indonesian mothers in our case. With limited
Mandarin proficiency, the child-directed speech Indonesian mothers provided is
different from that of native mothers. It may limit children’s development due to
6 1 Introduction
The two-year-old girl of an Indonesian mother is neither very extroverted nor intro-
verted. She is the third child in the family and has an eight-year-old sister and a
1.7 The Present Study 7
six-year-old brother. Her father was graduated from a senior high school in Taiwan
and her mother graduated from a junior high school in Indonesia. Her father works in a
factory as a purchase agent. Her mother is an interpreter in the Foreign Spouse Family
Service Center in Chiayi County. In the daytime, this child goes to kindergarten, so her
parents only spend three or four hours with her in the evening one day. When they have
conversation with her, both of them mostly speak Taiwanese to the child. For most of
the time, this child talked in Taiwanese. During the recording, the investigator and the
child played with toys and practiced writing numbers.
The three-year-old girl of an Indonesian mother is very extroverted. She is the third
child in the family and has two older sisters. Her parents were both graduated from
senior high school. Her father works in the hardware store, and her mother is a
housewife and a volunteer at the kindergarten. Her mother has studied Chinese
characters for six months. In the daytime, this child usually stays at home with her
mother or goes to her sister’s kindergarten while her mother works as a volunteer,
so her mother spends all day with her. However, her father spends only five to six
hours with her one day. This child mostly speaks Mandarin with her parents. In the
one-hour recording, the mother and the child played with some toys and were
involved in daily talking.
Taiwan very well and has no problem in communicating with her children. Both the
parents speak Mandarin and Taiwanese to the child. During the recording, the
Indonesian mother and her child played with blocks.
The four-year-old girl and the five-year-old boy of an Indonesian mother were sister
and brother. The girl was outgoing and the boy was introverted. Their mother
adapted to Taiwan very well, and she has stayed in Taiwan for nine years. Before
she came to Taiwan, she did not attend any Mandarin courses. She has attended
Mandarin courses for five years after coming to Taiwan. Therefore, she was not
worried about her interaction with her children. Additionally, both parents were
graduated from junior high school. Her father worked for transportation, and her
mother worked as a servant part-time. The girl spent five hours with her father and
three hours with her mother a day. She usually talked to her parents in Mandarin or
Taiwanese. The child was usually looked after by her aunt, who also talked to her in
Mandarin or Taiwanese. During the recording, her mother and her aunt used blocks
and drew a picture to interact with her.
The five-year-old boy of an Indonesia mother was introverted. His mother has
stayed in Taiwan for nine years, and she adapted very well. She had not taken any
Mandarin courses before coming to Taiwan. She has taken Mandarin courses for
five years after she came to Taiwan. Besides, she was not worried about her
interaction with the children. Both parents were graduated from junior high school.
The father’s job was fruitage and the mother’s was a part-time servant. Moreover,
the child spent five hours getting along with his father and three hours getting along
with his mother a day. Additionally, he usually talked to his parents in Mandarin
and in Taiwanese. Besides, his aunt usually looked after him, and he usually talked
to her in both Mandarin and Taiwanese. His mother and aunt played with blocks
and drew a picture with him during the recording.
life. Now, she adapted to the life in Taiwan very well and has no problem in
communicating with her child. Both the parents speak Mandarin and Taiwanese to
the child. During the recording, the investigator used blocks to play with the child
and also told stories to interact with the child.
The two-year-old boy of Taiwanese parents is outgoing. His parents were graduated
from university, and they were both teachers. His father spent three hours, while his
mother spent seven hours getting along with him a day. He usually talked to his dad
in Mandarin and usually talked to his mom in Taiwanese. He was usually looked
after by her grandmother, who talked to him in Mandarin or Taiwanese. His mother
used the blocks to interact with him during the recording.
10 1 Introduction
The three-year-old girl of Taiwanese parents is not very extroverted or introverted. She is
the second child in the family and has a sister, who is older than her by three years. Both of
her parents were graduated from vocational high schools. Her father works in a food
factory, and her mother is a waitress in a restaurant. In the daytime, the child is mostly taken
care of by her mother. Her father usually spends seven hours a day with her when her
mother goes to work. When they have conversation with her, both of them mostly speak
MandarinChinesetothechild.Formostofthetime,thischildtalksinMandarinChinese.In
the one-hour recording, the mother played with her with house pokers and a toy crocodile
The four-year-old girl of Taiwanese parents was outgoing. Her parents were
graduated from university, and are both teachers. She spent three hours with her
1.7 The Present Study 11
father and seven hours with her mother a day. Additionally, she usually talked to
her dad in Mandarin and talked to her mom in Taiwanese. However, her grand-
mother usually looked after her, and she talked to her grandmother in Mandarin or
Taiwanese. During the recording, her mother used blocks to interact with her.
The five-year-old boy of Taiwanese parents is extroverted. Both his father and
mother work as a staffs in public organizations. His father has a master’s degree and
his mother graduated from university. The child is usually taken care of by his
grandmother and his mother. The parents speak both Mandarin and Taiwanese to
the child, while the grandmother communicates with the child mostly in Taiwanese.
The five-year-old girl with a Taiwanese mother is extroverted. She has a brother.
Her father has a master’s degree and works in industry. Her mother was graduated
from junior college and is a business woman. The child went to the kindergarten
and was usually taken care of by her mother at home. The parents communicated
with the child in Mandarin and Taiwanese.
The six-year-old boy of Taiwanese parents is active and outgoing. He often fights
with other children at school. He has a younger brother. They like to play together.
Both of the parents are elementary school teachers. They lay heavy emphasis on
children’s education. Since the father studied in a graduate program besides work,
the children are mostly taken care of by the mother. The mother teaches first and
second grades in school. She focuses on children’s Mandarin acquisition.
The six-year-old girl of Taiwanese parents is an extroverted girl. Both her father
and mother have senior high school diplomas. Her father runs an iron factory, and
her mother works as a staff in a public organization. The child is usually taken care
of by her grandmother. The parents and grandmother speak both Mandarin and
Taiwanese to the child. During the recording, the child is doing homework and the
mother asked the child some questions.
12 1 Introduction
Table 1.2 summarizes the background information of the participants. The first
column indicates the mother’s nationality, the second column indicates the chil-
dren’s age in the first year, and the third column shows the gender. The fourth
column presents parents’ education. F stands for father and M stands for mother.
Their languages used at home were given in the sixth column. The last column lists
the mothers’ length of stay in Taiwan.
In this chapter, the background of the present study has been provided. The
remaining of the book discusses phonological development in Chap. 2, lexical
development in Chap. 3, and syntactic development in Chap. 4. Finally, Chap. 5
concludes the study.
References
Alves, M. (1999). What’s so Chinese about Vietnamese? The Ninth Annual Meeting of the
Southeast Asian Linguistics Society. Berkeley: University of California.
Chomsky, N. (1959). Review of verbal behavior by B. F. Skinner. Language, 35(1), 26–58.
Elman, J. L. (1991). The importance of starting small. CRL Technical Report 9101, Center for
Research in Language, University of California, San Diego, pp. 3–9.
Fazio, L. L., & Florence, S. (1994). Using multiple regression to predict minority children’s
second language performance. Applied Linguistics, 15(4), 421–441.
Kowalsky, M., Gheiler, D., Galle, L., Goodrich, W., Kuo, J. Y., & Kwon, K. (2003). Learning a
second language in the preschool years. Tampa, USA: The Biennial Meeting of Society of
Research on Child Development.
Kuo, J. Y. (2008). Mandarin acquisition by children of Vietnamese mothers in Taiwan. Taipei:
Crane.
Krashen, S. (1985). The input hypothesis: Issues and implications. London: Longman.
Lust, B. (2006). Child language: Acquisition and growth. New York: Cambridge University Press.
Mckay, S., & Wong, L. (2000). New immigrants in the United States: Reading for second
language educators. Cambridge: Cambridge University Press.
Mushi, S. L. P. (2002). Acquisition of multiple languages among children of immigrant families:
Parent’s role in the home-school language pendulum. Early Child Development and Care, 172
(5), 517–530.
Sachs, J., Bard, B., & Johnson, M. (1981). Language learning with restricted input: Case studies of
two hearing children of deaf parents. Applied Psycholinguistics, 2(1), 33–54.
Scarborough, H. (1989). Index of productive syntax. Applied Psycholinguistics, 11, 1–22.
Skinner, B. F. (1957). Verbal behavior. New York: Appleton-Century-Crofts.
Sneddon, J. N. (2004). The Indonesian language: Its history and role in modern society.
Kensington: UNSW Press.
Sun, M-X., 孫明霞. (2010). 語言環境與少數民族學前兒童雙語習得探析 “Yuyan huanjing yu
shaoshu minzu xueqian ertong shuangyu xide tanxi” [The study of how minority children learn
second languages in preschool education.]. Language and Translation, 2, 64–68.
Vygotsky, L. S. (1978). Mind in society. Cambridge, MA: Harvard University Press.
14 1 Introduction
Chinese
Since both nativist and behaviorist theories cannot deal with regression in sound
production, an appropriate way is to adopt a cognitive problem-solving approach to
the acquisition of phonology. The cognitive problem-solving approach could offer a
better explanation for the regression phenomenon mentioned above. First, nasal-
ization regression could be the children’s overgeneralization in his trial-and-error
articulation attempts. For the second problem, the problem-solving approach tends
to claim that most of the reward is internal instead of the traditional view of
behaviorism (Gleason 2012).
The OT grammar is an input–output mechanism which assumes that each input has
precisely one output. The process of selecting the optimal candidate as the output is
accomplished by the following components:
2.1 General Literature on Phonological Development 19
Eval
Gen C1 >> C2 >> Cn
Candidate a
Candidate b
Input Candidate c
Candidate d Output
Candidate….
Con The set of constraints out of which grammars are constructed (McCarthy and
Prince 1994).
Gen A function defining, for each possible i, the range of candidate linguistic
analyses available to i (McCarthy and Prince 1994).
Eval The function defined by composing all of the constraints in the order in
which they are ranked (McCarthy 2008).
The function Gen (Generator) produces a set of candidates, all of which are
logically possible analyses of the given input. Eval (Evaluator) deals with a system
of ranked constraints which is a formal construction on Con that yields the grammar
of an individual language. Therefore, the mechanism of OT is that Gen with the
given input i generates an infinite set of candidate outputs based on UG. Then, Eval
will assess the candidate outputs in terms of their relative harmony with respect to
the language’s ranking of the constraints (McCarthy and Prince 1994). The
mechanism of OT is schematically presented in Fig. 2.1, which supposes that Con
consists of C1, C2, … to the infinite Cn and that language is defined by the ranking
C1 ≫ C2 ≫ ≫ Cn.
2.1.1.4 OT Tableau
2.1.1.5 Constraints
There are two primary types of constraints. One is the markedness constraints,
which militate against marked structures. The other type consists of the faithfulness
constraints which demand identity between input and output (Gnanadesikan 1996).
In the following sections, the faithfulness constraints and markedness constraints
adopted in the present study will be discussed.
The faithfulness constraints adopted in the present study are shown in the following
and they are elaborated with examples:
a. Ident (aretroflex): The output must be in accord with its input in retroflex.
(Li 2006). For example, given the input /tʂ/, the candidate output [ts] violates
this constraint, while the candidate [ʂ] does not.
b. Ident (as.g): The output has to correspond to its input in aspiration (Li 2006).
For example, given the input /tɕh/, the candidate output [tɕ] violates this con-
straint, while the candidate [tsh] does not.
c. Ident (acont): The output must be equal to the input in continuance (Li 2006).
For example, given the input /ɕ/, the candidate output [tɕ] violates this con-
straint, while the candidate [s] does not.
d. Ident (aanterior): The segments between input and output share the same value
in anterior (Li 2006). For example, given the input /f/, the candidate output [x]
violates this constraint, while the candidate [p] does not.
e. Ident (astrident): The segments between input and output share the same value
in strident. For example, given the input /ʂ/, the candidate output [t] violates this
constraint, while the candidate [s] does not.
winner is more harmonic than its competing losers (Tesar and Smolensky 2000).
However, Boersma and Hayes (1999) developed a different theory of learning, in
which the constraints can be demoted and promoted. Different from classic OT, the
mechanism of ranking constraints proposed by Boersma and Hayes (1999) is
bidirectional. For example, if there are two incorrectly ranked constraints in a
learner’s ranking, the higher ranked one is demoted and the lower ranked one is
promoted slightly in the full hierarchy of constraints. After learning, the ranking is
reranked and the two constraints are ranked closer, until the ranking is reversed
finally.
In the present study, the classic OT model is adopted to analyze the data since
the one directional mechanism of constraint ranking is simple and clear to observe
which constraint is demoted in the full hierarchy. Besides, by using OT, we can
explain the phonological development of children of Vietnamese and Indonesian
mothers by the constraint rankings and investigate whether there is any similarity or
difference with native children.
In children’s early speech, there is still some babble contained in their speech.
Moreover, during the transition from prespeech to speech, children may produce
words invented by them, defined as ‘protowords.’ Protowords may with or without
adult models, but the appearance of protowords shows that children have begun to
voluntarily control over their vocalizations.
Then, when children start out to learn to pronounce, several phenomena may
appear in the process such as feature changing, cluster reduction, and assimilation.
For children, they learn adult words through trial and error. And, by descriptive
features, it would be more clearly for us to interpret children’s changing of sounds
during the process. So, the often seen patterns in early child phonology include all
initial stops being voiced, using alveolar place of articulation for both alveolar and
velar consonants, and nasal assimilation or assimilating the place of articulation.
Besides the patterns mentioned above, the accuracy of children’s perception of
sounds is also a key point for them to learn speech. For example, some children
learning English may have problem distinguishing [f] from [h] the two sounds are
similar for audition, because which leads to late acquisition of [h] (Velleman 1988).
In addition, in the early period of phonological development, word pronunciation is
affected by the length of the word and stress patterns. For example, children tend to
omit the initial syllable of a multisyllabic word as the syllable is unstressed. So, we
may hear ‘posed’ for ‘supposed.’ This phenomenon may be due to the fact that the
weakly stressed syllables are harder to perceive so that children miss hearing the
whole words.
During the process of learning speech sounds, we can see the differences in
learning strategies among children. Some children would not use a word if they
cannot produce the sounds accurately. But, other children may feel free to rearrange
24 2 Phonological Development
adult words to fit them into their speech. So, some children may attempt one word
at a time until the word is clear to them. On the other hand, some children may try a
whole phrase at once. Nevertheless, as children begin to put words together and
become precise in their articulation, the differences in learning strategies will dis-
appear finally.
By three years of age, the basic features of the adult phonological system are
present in children’s pronunciation and most children can produce all the vowel
sounds and nearly all the consonant sounds. However, the liquids /r/ and /l/ are still
easy to be mistaken even until the age of four or five. Besides, in longer words such
as vacuum cleaner, children may still mispronounce the /v/ sound although utter
accurately in shorter word such as vase. Around the age of seven, children could
pronounce all sounds accurately (Gleason 2012).
Mandarin has twenty-two consonants including three palatals, which were analyzed
as consonant–glide combinations by some scholars (i.e., Cheng 1973; Duanmu
2000). I will follow the tradition in the acquisition literature treating them as single
sound for the ease of transcription and discussion. Table 2.1 shows the twenty-two
phonemes in phonetic symbols.
Mandarin has five vowel phonemes as shown in Table 2.2 (Duanmu 2000).
Mandarin has three glides [j, w, ɥ], as in [ja] ‘tooth,’ [wa] ‘frog,’ and [ɥ] ‘moon,’
which correspond with vowels [i, u, y]. Since the two sets do not contrast with each
other, the vowel symbols were used for transcription. A vowel in Mandarin can
carry four tones with contrastive meanings. Some examples are given in Table 2.3
(Ladefoged and Johnson 2015). The pitch levels are represented by five numerals
with one representing the lowest and five representing the highest.
A Mandarin syllable is made up of an optional onset limited to one consonant
and an optional coda restricted to nasals. Since our focus is acquisition, we will not
get into sophisticated phonological analysis.
Taiwan Southern Min is a Min dialect of Chinese spoken in Taiwan. In the liter-
ature of linguistics, there is a variety of names of Taiwan Southern Min: South Min,
Southern Min, Taiwanese, and Taiwanese Hokkien. Taiwan Southern Min is a tone
language like Mandarin, but it has seven tones to distinguish lexical meanings. In
the following, tones, consonants, and vowels of Taiwan Southern Min will be
introduced.
There are seven tones in Taiwan Southern Min including five long tones: Yinping,
Yinshang, Yinqu, Yangping, and Yangqu and two short tones (entering tones):
Yinru and Yangru. Every tone has two forms: juncture tone and sandhi tone.
Because of the tone sandhi phenomenon in Taiwan Southern Min, every juncture
tone has a matched sandhi tone which appears in context position. According to
Cheng (1997), seven tones of Taiwan Southern Min are shown in Table 2.4.
Entering tones, Yinru and Yangru, only exist in checked syllables, i.e., a syllable
ending with unaspirated stops: /p, t, k, ʔ/, for example, /hap5/ ‘close,’ /lak5/ ‘six,’
and /phaʔ3/ ‘beat.’ The tone sandhi phenomenon in Taiwan Southern Min results in
two tonal forms. For example, given the word ‘short,’ the juncture form is [te53] and
the sandhi form is [te55 te53]. Basically, tone sandhi is applied regressively (Chung
2002; Liu 2009). As the example shown, when the lexical item ‘short’ is alone, it is
a juncture tone [te53] and tone value is high falling; when a lexical item is followed
by another lexical item within a tone group, the tone sandhi rule is applied.
Therefore, in [te55 te53] ‘short,’ tone value of the first one [te] is realized as
high-level tone. Examples of tone sandhi phenomenon of each tone are given in
Table 2.5.
There are six oral vowels and four nasal vowels in Taiwan Southern Min as given in
Table 2.7.
Taiwan Southern Min has oral vowels including front vowels /i, e/, central
vowels /ə, a/, and back vowels /u, ɔ/. Besides, different from Mandarin, Vietnamese,
and Indonesian, Taiwan Southern Min has four nasal vowels /ĩ, ẽ, ɔ̃, ã/, which can
stand alone even they do not occur with nasal consonants as examples shown in
Sect. 2.1 (Chung 2002).
Vietnamese has northern and southern dialects. The northern dialect does not have
retroflex consonants, while the southern one does. Affricates are lacking in southern
Vietnamese, while they exist in northern Vietnamese. The consonants of the two
dialects are given in Table 2.8.
Vietnamese has ten vowels. Below is the vowel chart for the Hanoi variety, a
northern dialect, and other dialects may have different vowel inventories
(Vietnamese language 2007) (Table 2.9).
All vowels are unrounded except the back vowels. Therefore, the high front
rounded vowel [y] in Mandarin is particular challenging for Vietnamese learners of
Mandarin. On the other hand, like Mandarin, all vowels in Vietnamese carry an
inherent tone. Vietnamese tones are not identical with Mandarin tones though.
There are six tones in Vietnamese: high level, low falling, high rising,
dipping-rising, breaking-rising, and low falling constricted. The tone value is listed
in Table 2.10 (Vietnamese Language 2007; Lin 2005).
2.2 Phonology of Mandarin, Taiwan Southern Min, Vietnamese, and Indonesian 29
Indonesian has twenty-two consonants. Twelve of them can occur in both initial
and final positions, and ten of them can only occur initially in the syllable. The
twenty-two consonants are shown in Table 2.11.
As shown in Table 2.11, Indonesian has stops including voiceless–voiced pairs
/p, b/, /t, d/, /k, g/ and one glottal stop /ʔ/, in which /p, t, k, ʔ/ can occur initially and
finally in the syllable, but /b, d, g/ only occur in the initial position. Indonesian
fricatives include /f, v/ and /s, z/ as two pairs of voiceless–voiced fricatives, one
palatal fricative /ʃ/, one velar fricative /x/, and one glottal fricative /h/. /f, s, h/ occur
both initially and finally in the syllable, but /v, z, ʃ, x/ occur only in the initial
position of a syllable. Affricates include voiceless palatal /tʃ/ and the corresponding
voiced palatal /dʒ/ and both of them only occur initially in the syllable. Besides,
Indonesian has nasals /m, n, ɲ, ŋ/, one alveolar trill /r/ and an alveolar lateral /l/.
Palatal nasal /ɲ/ only occurs initially, but /m, n, ŋ, r, l/ occur both initially and finally
in the syllable. In comparison with Vietnamese, the voiced velar stop /g/ and the
voiceless glottal stop /ʔ/ only exist in Indonesian but not in Vietnamese. However,
Vietnamese has one voiceless aspirated alveolar stop /th/ which also exists in
Mandarin but not in Indonesian as well as a voiceless unaspirated palatal stop /c/. In
terms of fricatives, Indonesian has palatal fricative /ʃ/, but Vietnamese does not; on
the other hand, Vietnamese has velar fricative /ɣ/ which is absent in Indonesian. As
for affricates, Indonesian has palatal affricates /tʃ/ and /dʒ/ distinguished by the
feature [+/− voiced]. Mandarin has three pairs of affricates which are voiceless and
distinguished by [+/− aspirated]: alveolar affricates /ts, tsh/, retroflex affricates /tʂ,
tʂh/, and palatal affricates /tɕ, tɕh/. Vietnamese has no affricates. Both Vietnamese
and Indonesian have four nasals /m, n, ɲ, ŋ/ and an alveolar lateral /l/, but alveolar
trill /r/ exists only in Indonesian. In contrast to Mandarin, both Vietnamese and
Indonesian lack retroflex sounds; moreover, the difference between Mandarin,
Vietnamese, and Indonesian is that most of Mandarin consonants are distinguished
by the feature [+/− aspirated], but both consonants of Vietnamese and Indonesian are
distinguished by the feature [+/− voiced] (Macdonald and Darjowidjojo 2007).
There are two semivowels and six vowel phonemes in Indonesian and they are
presented in Table 2.12.
Indonesian has two semivowel phonemes /j/ and /w/ and both of them can occur
initially and finally in the syllable. In comparison with Vietnamese, both
Vietnamese and Indonesian have front vowels /i/ and /e/ and back rounded vowels /
u/ and /o/. Indonesian has a mid-central unrounded vowel phoneme /ʌ/ which
Vietnamese does not have. However, compared with Mandarin, both Vietnamese
and Indonesian lack high front rounded vowel /y/, which is easily replaced by high
front unrounded vowel /i/ if there is no such a sound in the native language.
are still in the period of exploration of sounds. However, they have acquired [p],
[m], [h], and [t]. It seems that the bilabial sounds are acquired quite early, but [n] is
still unstable and sometimes even deleted from pronunciation. Besides, a variety of
substitution sounds are used by the age of twenty months. For example, [l] is often
replaced by [t] and [n].
The second stage is from twenty to thirty months (1; 8–2; 6). In the stage,
children begin the active process of acquisition of a complex set of linguistic rules,
so the development becomes a major part of acquisition in this stage. In this period,
[f], [tɕh], and [tsh] begin to appear but are still unstable. [f] is often substituted by
[p]. And, at this time, [ph] and [th] are substituted by their unaspirated counterparts
although they have been acquired at the preceding stage. Likely, [t] and [n] are still
the most often substitutions for [l]. Besides, all the affricates except [tɕ], the
fricatives, and the retroflex sounds are difficult for children. So, [ɕ] is replaced by
[t], [k], or [tɕ]. Both [ts] and [s] are substituted by [t] or [tɕ]. As for [ʑ], [l] is the
most often substituted sound for it. [tʂ] and [tʂh] are often replaced by [tɕ] and
sometimes by [t]. [ʂ] has many variants such as [t], [k], and [tɕ] and English
interdental fricative [h].
The third stage is from thirty-one to thirty-eight months (2; 7–3; 2). During this
stage, children progress from putting two words together with the well-formed
simple sentences. Moreover, children still continue to add new sounds to his/her
language. In this period, the mutual substitution between [l] and [n] is observed.
Besides, all affricates and fricatives have emerged but are not very stable. The
retroflex sounds are substituted by different ways including the corresponding
affricates or fricatives and some sound approximations.
The fourth stage is from forty to forty-eight months (3; 4–4; 0). In this period,
aspirated stops such as [th] and [ph] can be pronounced, and the substitution of sounds
such as [l], [tɕh], and [ɕ] is dramatically narrowed down. As to the retroflex sounds
and [ts], [tsh], and [s], they do not make any significant progress in this period.
The final stage is from fifty-two to seventy-two (4; 4–6; 0). In this stage,
phonological development progresses considerably in that children begin to
develop his/her perceptual abilities and acquire a number of phonemic contrasts in
speech production. The number of substitution for the retroflex sounds seems to
decrease gradually. But, sometimes [ts] is still used for [tʂ], and [ts] or [tsh] is used
for [tʂh]. The factors for the delay in acquiring the retroflexes might be the influence
of Taiwanese (Hsu 1996). In fact, deretroflexion is quite common among adult
Mandarin speakers in Taiwan (Rau and Li 1994).
Since vowel production involves more invisible articulators, children may need to
adopt trial and error in order to find proper positions of articulators. By the age of
twelve months, children produce [a] first and followed by [i]. Then, by the age of
2.3 Mandarin Phonological Development by Native Children 33
thirteen and fourteen months, children are able to pronounce [u] sound. By the age
of eighteen months, children have good control over all the single vowels except
[y]. That is, they can produce vowels such as [a], [i], [u], [ə], and [o]. Moreover,
from thirteen months to eighteen months, children appear to pronounce the diph-
thongs such as [ai], [au], [ua], [ie], [ui], and [ia] although these diphthongs are not
so stable in this period. Besides, they may be used to substitute for other sounds
which are not acquired. By the age of six, children might still have problem with
[ei] and [ou]. For example, [ei] may lose its glide and pronounced as [e]; [ou] may
be pronounced as [o]. [y] may not be accurately pronounced by children till the age
of six; otherwise, it is substituted by [i] or [ə].
Triphthongs are most difficult for children. Some children even cannot produce
them accurately by the age of six. The first emergent triphthong is [iau] at around
2.6 and followed by [iou] at around 3.2; occasionally, [uai] would appear as early as
one year of age. As for [uei], it is observed very late in language development and is
often replaced by [ue].
Observations show that Chinese final nasals are very often used to substitute
diphthongs without nasal ending. Sometimes the final nasals were dropped. The
first final nasal that appeared was [uan] but was not so stable. It sometimes was
replaced by [an] and [ua]. At around 3.2, [uan] was finally stabilized. [uang] had
three alternatives for the period from 1.8 to 3.2. They are [ua], [ang], and [uang].
[yen] is replaced by [ien]. [uen] emerged around the age of 2.6 and was the only one
used accurately (Hsu 1996).
In the early stage of language development, children may rely on the prosodic
features such as stress and intonation. The declarative utterances rise gradually and
then fall in fundamental frequency. The imperative utterances rise sharply and then
fall. The questions have fundamental frequency contours which rise sharply at the
end of the utterance. Usually, the declarative and imperative intonations are
acquired at the beginning of one-word stage. The question intonation appears
around the twentieth month. Besides, stress patterns are also made by children to
show their emphasis on the intonation.
According to the literature, around one year of age, children appear to use the
level tone and falling tone though not stable. But, before the age of one year, all
rising and falling rising tones are substituted by the level or falling tones. The rising
tone begins to appear shortly after 1.0. The falling rising tone does not emerge till
the age of 1.6. By the age of 2.6, children have acquired the four tones, but they are
sometimes confused between the rising and falling rising tones.
The development of syllable structures can be divided into the following four
stages, beginning with simple structure to more complex one. In the first stage, the
syllable structures appear to be CV, V, VV, and CVC, and CV is the dominant one
in this period. In the second stage, the CVC pattern becomes stabilized and CVV
34 2 Phonological Development
structure is new to appear in the period. Then, the CVV structure is a common one
in this period; there appeared to be nasal syllabic consonants. In the final stage, the
most complex structures appeared are VVV, CVVC, and CVVV.
There is some relationship between syllable structures and tones in children’s
development of speech. In early stage, children deal with simple structures such as
V, CV, and VV and the level, falling, and rising tones are used. The falling rising
tone begins to appear when children construct more complex syllables such as CVV
and VVV. So, it seems that the dipping tone begins to appear after acquiring
complex syllable structures (Hsu 1996).
The two-year-old Indonesian girl has not acquired most of the phonemes. She has
problems with fricatives, affricates, retroflexes, and the high front rounded vowel /
y/. She sometimes substituted labiodental fricative /f/ with high back vowel /u/. For
example, she pronounced /faǹ tsuà n/ ‘sail’ as [uaǹ tsuà n]. She also sometimes
deleted the velar fricative /x/. For example, she pronounced /xua/ ‘flower’ as [ua].
She changed fricatives to affricates as /sā n/ ‘three’ was pronounced as [tsā n]. The
child had problems with the aspirated alveopalatal affricate /tɕh/and sometimes
changed it to the unaspirated one [tɕ], as /tɕhī n tshaí / ‘vegetable’ was pronounced
as [tɕī n tshaí ] or the alveolar one [ts] as /fā n tɕhiè / ‘tomato’ was pronounced as
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 37
[fā n tshiè ]. The alveopalatal fricative was changed to affricates at the same place of
articulation. For example, /tiá u ɕià tɕhy/́ ‘fall’ was pronounced as [tiá u tɕiá tɕhy].
She replaced retroflexes with alveolars like her mother and many adults in Taiwan.
For example, /tʂǔ / ‘cook’ was pronounced as [tsǔ ], /tʂ̄ h/ ‘eat’ was pronounced as
[ts̄ h], and /ʂ/̀ ‘ten’ was pronounced as [s̩ ].
̀ The high front rounded vowel /y/ was
replaced with the unrounded one [i] half of the time; for example, /tɕhy/́ ‘go’ was
pronounced as [tɕhi]́ .
In the second year, aspiration was acquired. All the aspirated stops were pro-
nounced correctly. The unaspirated bilabial stop /p/ was omitted sometimes, e.g., /
pù/ ‘no’ ! [øù].
The child had acquired alveopalatal affricates. The problems with the labiodental
and retroflex still remained. She continued pronouncing the labiodental fricative /f/
as a voiceless velar fricative [x] and replacing retroflex affricates with alveolars,
e.g., /tʂuō tʂ/ ‘table’ ! [tsuō tʂ]. She was able to produce the voiced retroflex
fricative correctly, but not the voiceless one. The child showed regression in the
aspirated alveolar affricate, which was pronounced correctly in the first year.
In the third year, the child still deleted /x/ and deaspirated the aspirated alveopalatal
affricate /tɕh/ occasionally. All the retroflex consonants were still replaced by
alveolars, and the retroflex schwa was pronounced without retroflex. Final velar
nasals were sometimes changed to alveolar nasals or dropped like the final alveolar
nasal sometimes. /u/ was deleted in /uo/ sequence occasionally.
The phonological development of the two-year-old child of an Indonesian
mother is summarized in Table 2.13. The first column lists the target phonemes.
The second column shows the phonetic representations and frequency of occur-
rences in the first year, two-year-old; the third column indicates those in the second
year, three-year-old; the last column presents those in the third year, four-year-old.
The frequency was omitted when there were no phonetic variations.
Let me give an example in OT framework. The Indonesian two-year-old girl
ranked *Retroflex constraint quite high and so replaced the retroflex fricative with
alveolar fricative for three years as shown in Tableau 2.2.
Table 2.13 Acquisition and substitutions of the sounds from the Indonesian girl: 2, 2 + 1, and 2 + 2 years of age
38
Phoneme Phonetic representations, first year Phonetic representations, second year Phonetic representations, third year
Frequency Frequency Frequency
p p p, ø p: 96.87 %, ø: 3.13 % p
ph ph 100 % ph ph
m m m m
f f, u u: 20 % f f
t t t t
th th th th
n n m n
l l l l, n l: 11/12 = 91.67 %, n: 1/12 = 8.33 %
k k k k
kh kh kh kh
x x, ø, u ø: 4 %, u: 2 % x x, ø x: 8/10 = 80 %, ø 2/10 = 20 %
tɕ tɕ tɕ tɕ
tɕh tɕh, tɕ, tsh tɕ: 2 %, tsh: 2 % tɕh, tɕ, ø tɕh: 81.82 %, tɕ: 9.09 %, ø: tɕh, tɕh: 4/5 = 80 %, tɕ: 1/5 = 20 %
9.09 % tɕ
ɕ ɕ, tɕ, tɕh tɕ: 52 %, tɕh: 9 % ɕ, tɕ ɕ: 40 %, tɕ: 60 % ɕ
tʂ ts 100 % ts 100 % ts 100 %
tʂh tsh 100 % tsh 100 % tsh 100 %
ʂ s, ts s: 96 %, ts: 4 % s 100 % ʂ, s, ʂ:1/17 = 5.89 %, s:15/17 = 88.22 %, ts:
ts 1/17=5.89 %
ʐ ʐ s 100 % ʂ: 1/17 = 5.89 %, s: 15/17 = 88.22 %
ts: 1/17 = 5.89 %
ʐ ʐ ʐ l 1/1 = 100 %
ts ts ts, tsh ts: 95 %, tsh: 5 % ts
(continued)
2 Phonological Development
Table 2.13 (continued)
Phoneme Phonetic representations, first year Phonetic representations, second year Phonetic representations, third year
Frequency Frequency Frequency
tsh tsh tsh tsh
s s, ts ts: 33 % s, ts s: 75.86 %, ts: 24.14 % s
ŋ ŋ ŋ ŋ
i i i, ø i: 96.77 %, ø: 3.23 % i
u u u, ø, y u: 81.82 %, ø: 13.64 %, y: u
4.55 %
y y, i i: 50 % y, i y: 90.91 %, i: 9.09 % y
a a a, ai a: 80 %, ai: 20 % a
o o o o
ə ə ə ə
ei ei ei ei
ai ai, a a: 5 % ai ai
e e e e
au au au au
ou ou ou ou
an an, a, aŋ, a: 7 %, aŋ: 7 %, uan: an an
uan 2%
ən ən ən ən
aŋ aŋ aŋ, a, an, aŋ: 80 %, a: 4 %, an: 4 %, en: aŋ, aŋ: 10/12 = 83.33 %, an: 2/12 = 16.67 %
2.5 Mandarin Phonological Development by Children of Indonesian Mothers
en 12 % an
əŋ əŋ əŋ əŋ
ə˞ ə˞ ə 100 % ə 100 %
ia ia ia ia
(continued)
39
Table 2.13 (continued)
40
Phoneme Phonetic representations, first year Phonetic representations, second year Phonetic representations, third year
Frequency Frequency Frequency
ua ua ua ua
ͻŋ ͻŋ ͻŋ ͻŋ
ie ie ie ie
ye ye ye ye
uai uai uai uai
uei uei uei uei
iau iau iau iau
iou iou iou iou
ien ien ien ien
uan uan uan uan
yen yen yen yen
in in in in, i in: 5/6 = 83.33 %, i: 1/6 = 16.67 %
uən uən uən uən
yn yn yn yn
iaŋ iaŋ iaŋ iaŋ
uaŋ uaŋ uaŋ ua ua: 1/1 = 100 %
iŋ iŋ iŋ, in iŋ: 85.71 %, in: 14.29 % iŋ
jͻŋ jͻŋ jͻŋ jͻŋ
uo uo uo uo, o uo: 6/7 = 85.71 %, o: 1/7 = 14.29 %
2 Phonological Development
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 41
Tableau 2.2 OT tableau for the Indonesian girl: 2, 2 + 1, and 2 + 2 years of age
ʂ Ident (astrident) *Retroflex *Affricate *(+cont) *(+anterior)
☞s * *
ʂ *! *
ts *! *
tʂ *! *
In the second year, the voiceless labiodental fricative [f] has been acquired and
pronounced accurately. However, he continued replacing retroflex obstruents with
alveolar obstruents, i.e., /ʂuě i kuǒ / ‘fruit’ ! [suě i kuǒ ]. There were also variants of
stops. For example, the unaspirated voiceless bilabial stop /p/ was sometimes
replaced as the aspirated one [ph] as in /tɕhiaŋ̀ pi/́ ‘wall,’ or the alveolar one [t] as in
/pǎ n tʂ/ ‘board.’ The aspirated voiceless velar stop /kh/ was sometimes replaced
with the alveolar one [th], e.g., /khǎ phień / ‘card’ ! [thǎ phień ]. Some errors in the
vowels occurred in the second year. A coda [n] was added to the vowel /a/, e.g., /tǎ
tɕia/́ ‘fight’ ! [tǎ tɕian]́ or replaced the velar nasal coda, i.e., /tʂhuaŋ̀ /
‘bed’ ! [tʂhuan ̀ ].
In the third year, the child was already six years of age. He had acquired most
phonemes, but still had problems with retroflexes and pronounced them as alveolars
just like most adult Mandarin speakers in Taiwan. For example, retroflex affricate /
tʂ/ was still replaced by alveolar affricate [ts] as in /tʂə́/ ‘this’; retroflex obstruent /ʂ/
was still replaced by alveolar obstruent [s] as in /ʂí/ ‘be.’ The other problem is
replacing the vowel /u/ with [ou], e.g., /phìŋ kuǒ / ‘apple’ ! [phìŋ koǔ ]. The
phonological development of the three-year-old boy of an Indonesian mother is
summarized in Table 2.14.
Let me illustrate the phonological development of the three-year-old Indonesian
boy in the OT framework. The three-year-old Indonesian boy ranked the marked-
ness constraints *(+s.g), *(+cont), and *Stop over the faithfulness constraint, so the
labiodental fricative became the unaspirated voiceless bilabial stop as shown in
Tableau 2.3. In the second and the third year, *(+s.g) and *(+cont) were demoted
and thus ended with the correct target [f] as in Tableau 2.4.
The three-year-old girl of an Indonesian mother has acquired all the stops and had
mistakes only occasionally. She had some typical Taiwan Mandarin pronunciation.
She replaced labiodental fricative /f/ with the velar one [x], for example, /fɔ́ŋ lì/
‘pineapple’ as [xɔ́ŋ lì]. She also replaced retroflexes with alveolar counterparts, for
example, /tʂə́ kə/ ‘this one’ as [tsə́ kə], /tʂ̄ h/ ‘eat’ as [ts̄ h], /ʂ́/ ‘is’ as [ś], and /ʐə̀n/
‘person’ as [lə̀n]. The child has acquired all the vowels except sometimes they
replaced the rounded high front vowel /y/ with the unrounded one [i]. For example,
/ý mǐ / ‘corn’ was pronounced as [í mǐ ].
Table 2.14 Acquisition and substitutions of the sounds from the Indonesian boy: 3, 3 + 1, and 3 + 2 years of age
Phoneme Phonetic representations, first year Phonetic representations, second year Phonetic representations,
third year
Frequency Frequency Frequency
p p, ph p (93.59 %), ph (6.41 %) p p
ph ph ph ph, m
m once
m m m m
f f, p, ph, x f (21.74 %), p (69.57 %), ph (4.35 %), f f
x (4.35 %)
t t, k, l t (87.18 %), k (1.71 %), l (11.11 %) t t
th th, x, ø th (70 %), x (1.15 %), ø (1.12 %) th th
n n, l n (94.95 %), l (5.05 %) m n
l l, n, k, ø l (97.21 %), n (0.56 %), k (1.12 %), ø l l
(1.12 %)
k k, kh k (99.04 %), kh (0.96 %) k k
h h
k k kh kh
x x, ø x (96.94 %), ø (3.06 %) x x
tɕ tɕ, tɕh, ts, th tɕ (77.27 %), tɕh (19.09 %), ts (2.73 %), th tɕ, ø, t tɕ (81.82 %), ø (4.54 %), t (13.64 %) tɕ
(0.91 %)
tɕh tɕh tɕh tɕh
h h
ɕ ɕ, ɕ, s, k , tɕ ɕ (88.68 %), ɕ (1.89 %), s (1.89 %), k ɕ, tɕ ɕ (81.82 %), tɕ (18.18 %) ɕ
(1.89 %), tɕ (5.66 %)
2.5 Mandarin Phonological Development by Children of Indonesian Mothers
tʂ ts, tɕ ts (99.34 %), tɕ (0.66 %) tʂ, ts, tɕ, tʂ (4.08 %), ts (85.72 %), tɕ (4.08 %), th ts 100 %
th , t (4.08 %), t (2.04 %)
tʂh tsh, tɕh tsh (97.18 %) tʂh, tsh tʂh (9.09 %), tsh (90.91 %) tsh 100 %
tɕh (2.82 %)
ʂ s, ɕ s (97.26 %), ɕ (2.74 %) s 100 % s, ɕ
43
(continued)
Table 2.14 (continued)
44
Phoneme Phonetic representations, first year Phonetic representations, second year Phonetic representations,
third year
Frequency Frequency Frequency
s (97.83 %), ɕ
(2.17 %)
ʐ l 100 % ʐ l 100 %
ts ts, tɕ ts (4.74 %), tɕ (5.26 %) ts, tʂ, k ts (87.5 %), tʂ (8.33 %), k (4.17 %) ts
tsh tsh tsh tsh, tsh (92.59 %), ø
ø (7.41 %)
s s s s
ŋ ŋ ŋ ŋ
i i i i
u u u u
y y, i y (59.1 %), i (40.91 %) y y
a a, ia, au a (96.97 %), ia (1.82 %), au (1.21 %) a a
o o o o
ə ə, i, ei, o ə (99.25 %), i (0.30 %), ei (0.15 %), o ə ə
(0.30 %)
ei ei, ie ei (93.65 %), ie (6.35 %) ei ei
ai ai, a, ia ai (68.18 %), a (29.09 %), ia (2.73 %) ai, a ai (66.67 %), a (33.33 %) ai
e e e e
au au, a, iau au (71.95 %), a (26.83 %), iau (1.22 %) au au
ou ou ou ou
an an, a an (90 %), a (10 %) an an, uan
uan once
(continued)
2 Phonological Development
Table 2.14 (continued)
Phoneme Phonetic representations, first year Phonetic representations, second year Phonetic representations,
third year
Frequency Frequency Frequency
ən ən ən, in ən (77.78 %), in (22.22 %) ən
aŋ aŋ aŋ aŋ
əŋ əŋ, an əŋ (93.33 %), an (6.67 %) ən 100 % əŋ
ə˞ ə˞ ə 100 % ə 100 %
ia ia ia ia
ua ua ua ua
ɔŋ ɔŋ ɔŋ ɔŋ
ie ie, e, i, ei, iau ie (24 %), e (40 %), i (12 %), ei (20 %), ie, e ie (56.25 %), e (43.75 %) ie
iau (4 %)
ye ye ye ye
uai uai, ua uai (88.64 %), ua (11.36 %) uai uai
uei uei, u uei (97.87 %), u (2.13 %) uei uei
iau iau, ia iau (86.13 %), ia (13.87 %) iau iau
iou iou, ə, ou, iͻ iou (90.82 %), ə (2.04 %), ou (3.06 %), iou iou
iͻ (4.08 %)
ien ien, en, iŋ, ei, ien (65.08 %), en (25.40 %), iŋ (3.17 %), ien, en ien (73.68 %) ien
i ei (4.76 %), i (1.59 %) en (26.32 %)
uan uan, an, ua, uan (73.44 %), an (1.56 %), ua (12.5 %), uan uan
2.5 Mandarin Phonological Development by Children of Indonesian Mothers
Phoneme Phonetic representations, first year Phonetic representations, second year Phonetic representations,
third year
Frequency Frequency Frequency
iaŋ iaŋ iaŋ iaŋ
uaŋ uaŋ, ͻŋ uaŋ (42.86 %), ͻŋ (57.14 %) uaŋ uaŋ, uaŋ (86.67 %), ͻŋ
ͻŋ (13.33 %)
iŋ iŋ iŋ, in iŋ (28.57 %), in (71.43 %) iŋ
jɔŋ jɔŋ jɔŋ jɔŋ
uo uo, o, ou uo (71.31 %), o (26.23 %), ou (2.46 %) uo, o, u uo (80 %), o (14.29 %), u (5.71 %) uo, o uo (98.91 %), o
(1.09 %)
2 Phonological Development
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 47
Tableau 2.4 OT tableau for the Indonesian boy: 3 + 1 and 3 + 2 years of age
/f/ Ident (aanterior) *Stop *(+s.g) *(+cont) Ident (astrident)
☞f * *
p *! *
ph *! * *
x *! * *
In the second year, the phonetic variations decreased in both types and tokens. The
replacement of labiodental fricative /f/ with the velar one [x] has disappeared. The
replacement of retroflexes with alveolars still happens occasionally. The alveopa-
latal fricative was still pronounced as affricates most of the time. The alveolar
fricative was also pronounced as affricates sometimes. The vowel /u/ was dropped
occasionally. She still replaced the rounded high front vowel /y/ with the unrounded
one [i] occasionally. /a/ was replaced with [ai] occasionally. She dropped nasals in
codas or changed the place of articulation occasionally, which was not observed in
the first year.
In the third year, the only problems remained were retroflexes and nasals in coda.
Retroflexes were replaced with alveolars most of the time. The velar nasals in codas
were still replaced with alveolar nasals when following /i/ but not other vowels
sometimes. The frequency is lower than that in the second year. The phonetic
representations of the three-year-old girl are summarized in Table 2.15.
48
Table 2.15 Acquisition and substitutions of the sounds from the Indonesian girl: 3, 3 + 1, and 3 + 2 years of age
Phoneme Phonetic representations, first year Phonetic representations, second Phonetic representations, third year
year
Frequency Frequency Frequency
p p p p
ph ph, x, f, th x: 8 %, f: 8 %, th: ph, p ph: 77.78 %, p: ph
8% 22.22 %
m m m, l m: 87.5 %, l: 12.5 % m
f f, x x: 33 % f f
t t, p, k, ts p: 4 %, k: 2 %, ts: t t
2%
th th th th
n n m n
l l l, n l: 98 %, n: 2 % l
k k k k, ø ø: once
once
kh kh kh kh
x x x x
tɕ tɕ, t t: 2 % tɕ tɕ
tɕh tɕh, ɕ ɕ: 6 % tɕh tɕh
ɕ ɕ ɕ ɕ
tʂ tʂ, ts ts: 94 % ts, tɕ ts: 93.33 %, tɕ: 6.67 % ts 53/53 = 100 %
tʂh tʂh, tsh, kh tsh: 81 %, kh: 2 % tsh 100 % tsh 8/8 = 100 %
ʂ ʂ, s, tsh, kh, x, k, s: 66 %, tsh: 2 %, kh: ʂ, s ʂ: 25 %, s: 75 % s 43/43 = 100 %
th 18 %, x: 5 %, k:
5 %,
th: 2 %
ʐ l 100 % ʐ ʐ, l ʐ: 2/4 = 50 %, l: 2/4 = 50 %
2 Phonological Development
(continued)
Table 2.15 (continued)
Phoneme Phonetic representations, first year Phonetic representations, second Phonetic representations, third year
year
Frequency Frequency Frequency
ts ts, k, tʂ k: 10 %, tʂ: 10 % ts, tʂ ts: 90.91 %, tʂ: 9.09 % ts, tʂ, tɕ tʂ once
tɕ once
tsh tsh, kh kh: 6 % tsh tsh
s s, ʂ, kh ʂ: 9 %, kh: 27 % s, ʂ s: 97.92 %, ʂ: 2.08 % s, ʂ ʂ twice
ŋ ŋ ŋ ŋ
i i i i
u u u, ø u: 94.34 %, ø: 5.66 % u
y y, i i: 48 % y y
a a, au au: 3 % a a
o o o o
ə ə ə ə
ei ei ei ei
ai ai ai ai
e e e e
au au au au, iau iau twice
ou ou ou ou
an an, ən, uan ən: 10 %, uan: 15 % an an
ən ən ien 100 % ən
2.5 Mandarin Phonological Development by Children of Indonesian Mothers
(continued)
Table 2.15 (continued)
50
Phoneme Phonetic representations, first year Phonetic representations, second Phonetic representations, third year
year
Frequency Frequency Frequency
ua ua ua ua
ͻŋ ͻŋ ͻŋ ͻŋ
ie ie ie ie
ye ye ye ye
uai uai uai uai
uei uei uei uei
iau iau iau iau
iou iou iou iou
ien ien ien ien
uan uan uan uan
yen yen yen yen
in in in in
uən uən uən uən
yn yn yn yn
iaŋ iaŋ iaŋ iaŋ
uaŋ uaŋ uaŋ uaŋ
iŋ iŋ iŋ, in iŋ: 69.23 %, in: iŋ, in, iŋ: 5/8 = 62.5 %, in: 2/8 = 25 %, i:
30.77 % in 1/8 = 12.5 %
jɔŋ jɔŋ jɔŋ jɔŋ
uo uo uo uo
2 Phonological Development
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 51
The Indonesian four-year-old boy had problems with stops, fricatives, laterals,
nasals, alveopalatals, and retroflexes. The child replaced the unaspirated voiceless
bilabial stop /p/ with the voiceless labiodental fricative [f] as /pań fǎ / ‘solution’ was
pronounced as [fan ́ fǎ ]. On the other hand, he also changed fricatives to stops. For
example, /xē i sə/́ ‘black’ was pronounced as [xē i tə]́ and /sɔ̄ ŋ ʂǔ / ‘squirrel’ was
pronounced as [tɔ̄ ŋ tǔ ]. The aspirated voiceless alveolar stop /th/ was replaced by
alveolar affricates [tsh]. For example, /thú/ ‘throw up’ was pronounced as [tshu]́ . The
alveolar lateral /l/ and alveolar nasal /n/ were used interchangeably as /naǹ ʂə̄ ŋ/
‘male’ was pronounced as [laŋ̀ tə̄ ŋ] and /laǹ tɕhioù / ‘basketball’ was pronounced as
h
[naǹ tɕ iou ̀ ]. Alveolar nasals and velar nasals in codas were mixed. For example, /
kuan ́ tɕȳ n/ ‘champion’ was pronounced as [kuaŋ́ tɕȳ n] and /tshəŋ̀ kɔ̄ ŋ/ ‘succeed’
was pronounced as [tʂhəǹ kɔ̄ ŋ]. The alveopalatal fricative /ɕ/ was changed to the
palatal fricative [ҫ] or alveolar fricative [s]. For example, /ɕiǎ ŋ taú / ‘think of’ was
pronounced as [ҫiǎ ŋ taú ] and /tʂhɔŋ̀ ɕī n/ ‘restart’ was pronounced as [tshɔŋ̀ sī n]. The
retroflexes were replaced by their alveolar counterparts. For example, /khǎ tʂu/́
‘stuck’ was pronounced as [khǎ tsu]́ , /tʂhū an/ ‘wear’ was pronounced as [tshū an],
and /ʂā y/̀ ‘shark’ was pronounced as [sā y]̀ . The voiced retroflex fricative was
replaced by the alveolar lateral [l] all the time as /ʐəǹ tɕiā / ‘I’ was pronounced as
̀ tɕiā ]. The child sometimes changed alveolar affricates to retroflex or alveo-
[lən
palatal ones. For example, /thú ts̝/ ‘rabbit’ was pronounced as [thú tʂ] and /ts̩ /́
‘rabbit’ was pronounced as [tɕhi]́ . The child did not really have problems with any
manner or place of articulation, but was confused about when to use which one. All
the vowels were correct. His mother also replaced retroflexes and final velar nasals
with alveolars as she pronounced /tʂǎ u/ ‘find’ as [tsǎ u] and /nań ʂə̄ ŋ/ ‘male’ as [naǹ
sən]. The mother’s errors were more systematic.
In the second year, the child produced the labiodental fricative /f/ correctly, but still
have problems with the voiceless retroflex fricative, which was replaced with an
alveolar one [s], e.g., /ʂū / ‘book’ ! [sū ]. The other problem was the dropping of
velar nasal coda, i.e., /ɔŋ/ ! [ɔ] as in /xɔŋ̀ lɔ̀ puo/ ‘carrot.’ Moreover, the child still
replaced retroflex affricate /tʂ/ with alveolar affricate /ts/ as in /tʂuō tʂ/ ‘table.’
52 2 Phonological Development
In the third year, the child was already seven years of age. He had acquired all the
phonemes and pronounced them almost correctly. But, sometimes, the child may
replace the unaspirated voiceless alveolar stop /t/ with the aspirated one [th], e.g.,
/tɔŋ́ u/́ ‘animal’ ! [thɔŋ́ u]́ . Besides, the retroflex affricate /tʂ/ was sometimes
replaced with alveolar affricate /ts/, i.e., /tʂə́ kə/ ‘this’ ! [tsə́ kə]. The phonological
development of the four-year-old boy of an Indonesian mother is summarized in
Table 2.16.
Let me illustrate with an example of the four-year-old boy’s phonological
development in the OT framework. *(+cont) was high-ranked in the first year, so
the voiceless bilabial stop was changed to voiceless labiodental fricative as in
Tableau 2.5.
In the second and the third year, *(+cont) was demoted and ended with the target
[p] as in Tableau 2.6.
The Indonesian four-year-old girl has acquired most of the phonemes and just had
problems with retroflexes and high front rounded vowels like many
Mandarin-speaking adults in Taiwan. She replaced retroflexes with alveolars. For
example, /tʂǔ / ‘cook’ was pronounced as [tsǔ ]. /tʂhū / ‘out’ was pronounced as
[tshū ]. /ʂ̄ / ‘teacher’ was pronounced as [s̩̄ ]. /ʐàn/ was pronounced as [làn]. The high
front rounded vowel was pronounced without lip rounding, as /tɕhý/ ‘go’ was
pronounced as [tɕhí] occasionally.
Table 2.16 Acquisition and substitutions of the sounds from the Indonesian boy: 4, 4 + 1, and 4 + 2 years of age
Phoneme Phonetic representations, first year Phonetic representations, second Phonetic representations, third year
year
Frequency Frequency Frequency
p p, f p: 2/6 = 33.3 %, f: 4/6 = 66.6 % p p
ph ph ph ph
m m m, l m: 5/6 = 83.33 %, m
l: 1/6 = 16.67 %
f f f f, p p: once
t t t t
th th, tsh, th: 5/12 = 41.6 %, tsh: 6/12 = 50 %, tʃ: 1/12 = 8.3 % th, th: 6/7 = 85.71 %, th
tʃ tsh tsh: 1/7 = 14.29 %
n n, l n: ½ = 50 % m n
l: 1/2 = 50 %
l n, l l: 11/13 = 84.6 % l l
n: 2/13 = 15.3 %
k k k k
kh kh kh kh
x x x x
tɕ tɕ tɕ tɕ, k k twice
tɕh tɕh tɕh tɕh
ɕ ɕ, s, ts ɕ: 4/8 = 50 %, s: 3/8 = 37.5 %, ts: 1/8 = 12.5 % ɕ ɕ, k k once
tʂ ts, t ts: 14/17 = 82.3 %, t: 3/17 = 17.7 % tʂ, tʂ: 1/4 = 25 %, t: tʂ, tʂ: 4/144 = 2.78 %, ts:
2.5 Mandarin Phonological Development by Children of Indonesian Mothers
Phoneme Phonetic representations, first year Phonetic representations, second Phonetic representations, third year
year
Frequency Frequency Frequency
ʂ s, ts, s: 3/31 = 9.6 %, ts: 1/31 = 3.2 %, tʂh: 23/31 = 74.1 %, ʂ, s ʂ: 3/4 = 75 %, s: ʂ, s s: 7/78 = 8.97 %, s: 71/78 = 91.03 %
tʂh, t, ɕ t: 3/31 = 9.6 %, ɕ: 1/31 = 3.2 % 1/4 = 25 %
ʐ l 1/1 = 100 % ʐ l, n l: 16/17 = 94.12 %, n: 1/17 = 5.88 %
ts ts, tʂ, t ts: 1/3 = 33.3 %, tʂ: 1/3 = 33.3 %, t: 1/3 = 33.3 % tʂ 7/7 = 100 % ts, l: 35/38 = 92.11 %, n: 3/38 = 7.89 %
tʂ
tsh tɕh, tʂh tɕh: ½ = 50 %, tʂh: 1/2 = 50 % tsh, tsh: 4/5 = 80 %, tʂh: tsh, tsh: 8/9 = 88.89 %, tʂh: 1/9 = 11.11 %
tʂh 1/5 = 20 % tʂh
s s, t, ɕ s: 4/19 = 21.05 %, t: 13/19 = 68.42 %, ɕ: s, ʂ s: 7/12 = 58.33 %, ʂ: s, ʂ s: 28/30 = 93.33 %, ʂ: 2/30 = 6.67 %
2/19 = 10.52 % 5/12 = 41.67 %
ŋ ŋ ŋ ŋ
i i i i
u u u u
y y y y
a a a a
o o o o
ə ə ə ə
ei ei ei ei
ai ai ai ai
e e e e
au au au au
ou ou ou ou
an an, aŋ an: 3/11 = 27.27 %, aŋ: 8/11 = 72.72 % an an
ən ən ən ən
(continued)
2 Phonological Development
Table 2.16 (continued)
Phoneme Phonetic representations, first year Phonetic representations, second Phonetic representations, third year
year
Frequency Frequency Frequency
aŋ aŋ aŋ aŋ
əŋ əŋ, ən əŋ: 3/6 = 50 %, ən: 3/6 = 50 % əŋ əŋ, əŋ: 2/7 = 28.57 %, ən: 5/7 = 71.43 %
ən
ə˞ ə˞ ə˞ ə 19/19 = 100 %
ia ia ia ia
ua ua ua ua
ɔŋ ɔŋ ɔŋ ɔŋ
ie ie ie ie
ye ye ye ye
uai uai uai uai
uei uei uei uei
iau iau iau iau
iou iou iou iou
ien ien ien ien
uan uan uan uan
yen yen yen yen
in in in in
uən uən uən uən
2.5 Mandarin Phonological Development by Children of Indonesian Mothers
yn yn yn yn
iaŋ iaŋ iaŋ iaŋ
uaŋ uaŋ uaŋ uaŋ
(continued)
55
Table 2.16 (continued)
56
Phoneme Phonetic representations, first year Phonetic representations, second Phonetic representations, third year
year
Frequency Frequency Frequency
iŋ in 3/3 = 100 % iŋ iŋ
jɔŋ jɔŋ jɔŋ jɔŋ
uo uo uo uo
2 Phonological Development
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 57
Her mother also replaced retroflexes with alveolars and pronounced the high
front rounded vowel without lip rounding. In addition, she replaced alveolar nasals
with alveolar laterals, changed the place of articulation of final nasals, simplified
complex vowels, and pronounced the first tone as the fourth tone. For example, the
mother pronounced /nień / ‘read’ as [lień ]. She changed the final alveolar nasal to a
velar one, as /uən ́ / ‘ask’ was pronounced as [uəŋ́ ]. She tended to drop the first vowel
in a sequence; for example, /tɕhień mień / ‘front’ was pronounced as [tɕhieǹ meń ]
and /xǔ o tɕien
́ / ‘rocket’ was pronounced as [xǒ tɕień ]. She pronounced the first tone
as the fourth tone as /xiā ŋ kū / ‘mushroom’ was pronounced as [xiaŋ́ kū ]. These
errors were not observed in the child.
In the second year, the girl still replaced retroflexes with alveolars all the time. In
addition, she had problems with place of articulation of nasals in codas. She tended
to replace velar nasals with alveolar nasals next to /i/, which could be caused by
assimilation. Nasals following other vowels were pronounced correctly.
In the third year, retroflexes were still replaced with alveolars over 90 % of the
time. Velar nasals were replaced with alveolar nasals next to /i/ with even higher
frequency than in the second year. The acquisition and substitutions of the sounds
of the four-year-old girl are summarized in Table 2.17.
Table 2.17 Acquisition and substitutions of the sounds from the Indonesian girl: 4, 4 + 1, and 4 + 2 years of age
58
Phoneme Phonetic representations, first year Phonetic representations, second Phonetic representations, third year
year
Frequency Frequency Frequency
p p p p
ph ph ph ph
m m m m
f f f f, xw f: 13/16 = 81.25 %, xw: 3/16 = 18.75 %
t t t t
th th th th
n n, m n: 98.24 %, m: 1.75 % m n
l l l l
k k k k
kh kh kh kh
x x x x
tɕ tɕ tɕ tɕ
tɕh tɕh tɕh tɕh
ɕ ɕ ɕ ɕ
tʂ ts 100 % ts 100 % tʂ, ts tʂ: 8/134 = 5.97 %, ts: 126/134 = 94.03 %
tʂh tʂh 100 % tsh 100 % tʂh, tʂh: 2/31 = 6.45 %, tsh: 29/31 = 93.55 %
tsh
ʂ s 100 % s 100 % ʂ, s ʂ: 22/341 = 6.45 %, s: 319/341 = 93.55 %
ʐ l 100 % ʐ ʐ, l ʐ: 24/52 = 46.15 %, l: 28/52 = 53.85 %
ts ts ts ts
tsh tsh tsh tsh
s s s s
(continued)
2 Phonological Development
Table 2.17 (continued)
Phoneme Phonetic representations, first year Phonetic representations, second Phonetic representations, third year
year
Frequency Frequency Frequency
ŋ ŋ ŋ ŋ
i i i i
u u u u
y y, i y: 93.75 %, i: 6.25 % y y
a a a a
o o o o
ə ə ə ə
ei ei ei ei
ai ai ai ai
e e e e
au au, aŋ au: 98.33 %, aŋ: 1.66 % au au
ou ou ou ou
an an, aŋ an: 96.77 %, aŋ: 3.22 % an an
ən ən ən ən
aŋ aŋ aŋ aŋ
əŋ əŋ əŋ əŋ
ə˞ ə 100 % ə˞ ə˞, ə ə˞: 3/24 = 12.5 %, ə: 21/24 = 87.5 %
ia ia ia ia
2.5 Mandarin Phonological Development by Children of Indonesian Mothers
ua ua ua ua
ɔŋ ɔŋ, au, ɔŋ: 88.88 %, au: 5.55 %, ͻ: ɔŋ ɔŋ
5.55 %
(continued)
59
Table 2.17 (continued)
60
Phoneme Phonetic representations, first year Phonetic representations, second Phonetic representations, third year
year
Frequency Frequency Frequency
ie ie ie ie
ye ye ye ye
uai uai uai uai
uei uei uei uei
iau iau, iau: 96.87 %, tɕiau: 3.12 % iau iau
tɕiau
iou iou iou iou
ien ien ien ien
uan uan uan uan
yen yen yen yen
in in in in
uən uən uən uən
yn yn yn yn
iaŋ iaŋ iaŋ iaŋ
uaŋ uaŋ uaŋ uaŋ
iŋ iŋ iŋ, in iŋ: 57.14 %, in: 42.86 % iŋ, in iŋ: 14/50 = 28 %, in: 36/50 = 72 %
jɔŋ jɔŋ jɔŋ jɔŋ
uo uo, o uo: 99.31 %, o: 0.68 % uo uo
2 Phonological Development
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 61
The Indonesian five-year-old boy had problems with labiodental fricatives, retro-
flexes, final nasals, and complex vowels. He replaced the labiodental fricative /f/
with [xu]; for example, [fē i tɕī ] ‘airplane’ was pronounced as [xuē i tɕī ]. He replaced
all the retroflexes with alveolars as /tʂə́ iaŋ́ / ‘this way’ as [tsə́ iaŋ́ ], /tʂ̄ h ‘eat’ as [ts̄ h],
/ʂ/́ ‘is’ as [s̩ ],
́ and /ʐəǹ / ‘person’ as [ləǹ ]. One of the vowels in a sequence was
omitted sometimes. For example, /kǒ u/ ‘dog’ was pronounced as [kǒ ], /mień /
‘noodles’ as [men ́ ], /tʂ ū aŋ/ ‘bed’ as [ts ā ŋ], and /uǒ / ‘I’ as [ǒ ]. The velar nasal /ŋ
h h
following /i/ was changed to the alveolar nasal [n] as /tɕhī ŋ/ ‘clean’ was pronounced
as [tɕhī n]. Many of these errors were also observed in the mother (Table 2.21).
His mother also replaced retroflexes with alveolars, changed the place of artic-
ulation of final nasals, pronounced the high front rounded vowel without lip
rounding, and simplified complex vowels. In addition, she replaced the alveolar
nasal with the alveolar lateral, the first tone with the fourth tone. For example, the
mother pronounced /nień / ‘read’ as [lień ]. She changed the final alveolar nasal to a
velar one, as /uən ́ / ‘ask’ was pronounced as [uəŋ́ ]. She tended to drop the first vowel
in a sequence; for example, /tɕhieǹ mié n/ ‘front’ was pronounced as [tɕhieǹ meń ]
and /xǔ o tɕien ́ / ‘rocket’ as [xǒ tɕień ]. She pronounced the first tone as the fourth
tone as /xiaŋ́ kū / ‘mushroom’ was pronounced as [xiaŋ́ kū ] like the stress in a stress
language like her L1, Indonesian. The tone errors were not observed in the child.
In the second year, the alveolar variants [s, ts, tsh] of retroflex obstruents /ʂ, t ʂ, tʂh/
coexisted with retroflexes [ʂ, tʂ, tʂh] and alveopalatals [ɕ, ʨ] except the voiceless
aspirated retroflex affricate /tʂh/ and the voiced retroflex fricative /ʐ/ were consis-
tently replaced by the alveolars [tsh] and [z].
In the third year, almost all the phonemes were acquired except that the aspirated
voiceless retroflex affricate /tʂh/ was still replaced by the unaspirated one [tʂ], e.g., /
tʂhī / ‘eat’ ![tʂī ]. The phonological development of the five-year-old boy of an
Indonesian mother is summarized in Table 2.18.
Table 2.18 Acquisition and substitutions of the sounds from the Indonesian boy: 5, 5 + 1, and 5 + 2 years of age
62
Phoneme Phonetic representations, first year Phonetic representations, Phonetic representations, third year
second year
Frequency Frequency Frequency
p p p
ph ph ph ph
m m m m
f f, x f: 91.66 % f f
x: 8.33 %
t t t t
th th th th
n n m n
l l l l
k k k k
kh kh kh kh
x x x x
tɕ tɕ tɕ tɕ
tɕh tɕh tɕh tɕh
ɕ ɕ ɕ ɕ, tɕ ɕ: 96.77 %, tɕ: 3.23 %
tʂ tʂ 100 % tʂ, ts tʂ: 25 %, ts: 75 % tʂ, ts tʂ: 26.53 %, tʂ: 73.47 %
tʂh tʂh 100 % tsh 100 % tʂh, tsh tʂh: 25.81 %, tsh: 74.19 %
ʂ s 100 % s 100 % ʂ, s ʂ: 16.38 %, s: 83.62 %
ʐ l 100 % ʐ ʐ, l ʐ: 36.36 %, l: 63.64 %
ts ts ts ts
tsh tsh tsh tsh
s s s s
(continued)
2 Phonological Development
Table 2.18 (continued)
Phoneme Phonetic representations, first year Phonetic representations, Phonetic representations, third year
second year
Frequency Frequency Frequency
ŋ ŋ ŋ ŋ
i i i i
u u, ə u: 96.96 %, ə: 3.03 % u u
y y y y
a a a a
o o o o
ə ə ə ə
ei ei ei ei
ai ai ai ai
e e e e
au au au, a au: 75 %, a: 25 % au
ou ou, o ou: 93.54 %, o: 6.45 % ou ou
an an an an
ən ən ən ən
aŋ aŋ aŋ aŋ
əŋ əŋ əŋ əŋ
ə˞ ə 100 % ə˞ ə˞, ə ə˞: 5.26 %, ə: 94.74 %
ia ia ia ia
2.5 Mandarin Phonological Development by Children of Indonesian Mothers
ua ua ua ua
ɔŋ ɔŋ ɔŋ ɔŋ
ie ie ie ie
ye ye ye ye
(continued)
63
Table 2.18 (continued)
64
Phoneme Phonetic representations, first year Phonetic representations, Phonetic representations, third year
second year
Frequency Frequency Frequency
uai uai uai uai
uei uei uei uei
iau iau iau iau
iou iou iou iou
ien ien, en ien: 97.43 %, en: 2.56 % ien ien
uan uan uan uan
yen yen yen yen
in in in in
uən uən uən uən
yn yn yn yn
iaŋ iaŋ iaŋ iaŋ
uaŋ uaŋ, aŋ uaŋ: 83.33 %, aŋ: 16.66 % uaŋ uaŋ
iŋ iŋ 100 % iŋ iŋ, in iŋ: 27.5 %, in: 72.5 %
jɔŋ jɔŋ jɔŋ jɔŋ
uo uo, o uo: 95.23 %, o: 4.76 % uo uo
2 Phonological Development
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 65
The Indonesian five-year-old girl had problems with alveolar laterals, retroflexes,
final nasals, and high front rounded vowels. She changed the alveolar lateral /l/ to
nasal [n]; for example, /lə̌ ŋ/ ‘cold’ was pronounced as [nə̌ ŋ]. All the retroflexes
were replaced by their alveolar counterparts. For example, she pronounced / kɔ̄ ŋ ʂǔ /
’ princess’ was pronounced as [kɔ̄ ŋ tsǔ ], [ʂhū laì ] ‘come out’ as [tshū laì ], and /ʂueí
́ / ‘sleep’ as [sueí tɕiaú ]. There were also aspiration and palatal assimilation. She
tɕiau
pronounced the unaspirated voiceless alveopalatal affricate as the aspirated one as /
́ tɕ y/́ ‘enter’ was pronounced as [tɕ iń tɕ y]́ . The final nasals tended to be
h h h
tɕin
dropped; for example, the child pronounced /ɕiǎ u/ ‘small’ as [ɕiǎ ], /uaŋ́ tɕi/́ ‘forget’
as [uá tɕi]́ , /phaŋ́ / ‘fat’ as [pha]́ , /khaú tɕiń / ‘get close’ as [khaú tɕi]́ , and /pì kɔ̌ ŋ/ ‘nose
hole’ as [pì kɔ̌ ]. The high front rounded vowels were sometimes pronounced
without lip rounding; for example, /tɕhy/́ ‘go’ was pronounced as [tɕi]́ .
In the second year, the alveolar lateral /l/ was pronounced correctly. The retroflexes
were still replaced by alveolars with lower frequency. She had no more problems
with aspiration and palatal assimilation. The velar final nasals were sometimes
changed to alveolar nasals.
In the third year, the girl has acquired all the phonemes except that retroflexes were
consistently replaced by alveolars. The retroflex mid-central lax vowel was also
pronounced without retroflex (Table 2.19).
Let us put /tʂ/ in the OT framework. The five-year-old Indonesian girl
high-ranked *Retroflex and so turned it to an alveolar affricate and remained for
three years as shown in Tableau 2.7.
Table 2.19 Acquisition and substitutions of the sounds from the Indonesian girl: 5, 5 + 1, and 5 + 2 years of age
66
Phoneme Phonetic representations, first year Phonetic representations, second year Phonetic representations, third year
Frequency Frequency Frequency
p p p p
ph ph ph ph
m m m m
f f f f
t t t t
th th th th
n n n n
l l, n l: 81.82 %, n: 18.18 % l l
k k k k
kh kh kh kh
x x x x
tɕ tɕ, tɕh tɕ: 77.27 %, tɕh: 22.73 % tɕ tɕ, ts tɕ: 70.83 %, ts: 29.17 %
tɕh tɕh tɕh tɕh
ɕ ɕ ɕ ɕ
tʂ ts 100 % tʂ, ts tʂ: 13.64 %, ts: 86.36 % ts 100 %
tʂh tsh 100 % tsh 100 % tsh 100 %
ʂ s 100 % s 100 % s 100 %
ʐ l 100 % ʐ l 100 %
ts ts ts ts
tsh tsh tsh tsh
s s s s
ŋ ŋ ŋ ŋ
i i i i
(continued)
2 Phonological Development
Table 2.19 (continued)
Phoneme Phonetic representations, first year Phonetic representations, second year Phonetic representations, third year
Frequency Frequency Frequency
u u u u
y y, i y: 88.89 %, i: 11.11 % y y
a a a a
o o o o
ə ə ə ə
ei ei ei ei
ai ai ai ai
e e e e
au au au au
ou ou ou ou
an an an, a an: 95 %, a: 5 % an
ən ən ən ən ən
aŋ aŋ, an, a aŋ: 36.36 %, an: 45.45 %, a: 18.18 % aŋ aŋ
əŋ əŋ 100 % əŋ əŋ
ə˞ ə˞ ə 100 % ə 100 %
ia ia ia ia
ua ua ua ua
ie ie ie ie
ye ye ye ye
2.5 Mandarin Phonological Development by Children of Indonesian Mothers
Phoneme Phonetic representations, first year Phonetic representations, second year Phonetic representations, third year
Frequency Frequency Frequency
ien ien ien ien
uan uan uan uan
yen yen yen yen
in in in in
uən uən uən uən
yn yn yn yn
iaŋ iaŋ iaŋ iaŋ
uaŋ uaŋ uaŋ uaŋ
iŋ iŋ, in iŋ: 20 %, in: 80 % iŋ, in iŋ: 54.55 %, in: 45.45 % iŋ
ͻŋ ͻŋ, ͻ ͻ once ͻŋ ͻŋ
jͻŋ jͻŋ jͻŋ jͻŋ
uo uo uo uo
2 Phonological Development
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 69
The six-year-old Indonesian boy still had phonetic errors in aspirations, fricatives,
retroflexes, and final nasals. He occasionally deaspirated the aspirated velar stops
and alveopalatal affricates. He replaced the aspirated voiceless velar stop with the
unaspirated one; for example, /khǎ ū / ‘test’ was pronounced as [kǎ ū ] and /tɕhio iə̌ n/
‘earthworm’ was pronounced as [tɕī o iə̌ n]. He replaced the alveopalatal fricative
with an affricate or an alveolar stop. For example, /ɕī pī an/ ‘riverbank’ was pro-
nounced as [tɕī pī an], /ɕíe ɕíe/ ‘thank’ as [tɕhíe tɕhíe], and /ɕǐ au kǒ / ‘dog’ as [tíau
kǒ ]. The retroflexes were replaced by their alveolar counterparts. For example, /tʂ̄
táu/ ‘know’ was pronounced as [ts̩̄ táu], /tʂhàŋ tɕǐ n lú/ ‘giraffe’ as [tshàŋ tɕǐ n lu], /ʂ́
thò/ ‘stone’ as [s̩ ̀ thò], and /ʐə̀n/ ‘person’ as [lə̀n]. He mixed the place of articulation
of final nasals; for example, /sán kā i/ ‘spread’ was pronounced as [sáŋ kā i] and /tháŋ
ʂā ŋ/ ‘burn’ was pronounced as [thaŋ́ ʂaŋ́ ]. The substitution of retroflexes with
alveolars was also observed in his mother, but not other errors.
In the second year, the child continued to replace retroflex affricates with alveolar
affricates and replaced a labiodental fricative with a velar fricative followed by a
labiovelar glide. For example, /f/ was replaced by [xu] as in /fā n tɕhie/̀ ‘tomato.’ The
voiced retroflex fricative was pronounced correctly, but the voiceless one /ʂ/ was
still replaced by an alveolar one [s], e.g., /ʂueǐ kuǒ / ‘fruit’ ! [sueǐ kuǒ ]. In addi-
tion, the child replaced the voiced bilabial nasal stop /m/ with the lateral liquid [l],
e.g., [mień pei] ‘quilt’ ! [lień pei]́ . All the vowels and finals have been acquired
except the rounded high front vowel was still pronounced without lip rounding.
In the third year, when the child was eight years of age, all the phonemes were
pronounced correctly, except the vowel /u/ was dropped in /uo/ sequence, e.g., /phiŋ̀
kuǒ / ‘apple’ ! [phiŋ̀ kǒ ]. It was correct in the first two years. The phonological
development of the six-year-old child of an Indonesian mother is summarized in
Table 2.20.
Table 2.20 Acquisition and substitutions of the sounds from the Indonesian boy: 6, 6 + 1, and 6 + 2 years of age
70
Phoneme Phonetic representations, first year Phonetic representations, second year Phonetic representations, third
year
Frequency Frequency Frequency
p p p p
ph ph ph ph
m m m m
f f f f
t t t, th t (88.89 %), th (11.11 %) t
th th th th
n n m n
l l l l
k k k k
kh kh, k k (16 %) kh kh
x x x x
tɕ tɕ tɕ tɕ
tɕh tɕh, tɕ tɕ (10 %) tɕh tɕh
ɕ ɕ, tɕ, t tɕ (50 %), t (10 %) ɕ ɕ
tʂ ts 100 % ts 100 % ts 100 %
tʂh tsh, s tsh (96 %), s (4 %) tʂh, tsh tsh (80 %) tah 100 %
ʂ s, ts, tsh s (50 %), ts (40 %), tsh (10 %) s 100 % s 100 %
ʐ l 100 % ʐ, l ʐ (33.33), l (66.67 %) l, z l (87.5 %), z (12.5 %)
ts ts ts ts
tsh tsh tsh tsh
s s s s
ŋ ŋ ŋ ŋ
(continued)
2 Phonological Development
Table 2.20 (continued)
Phoneme Phonetic representations, first year Phonetic representations, second year Phonetic representations, third
year
Frequency Frequency Frequency
i i, ei ei (6 %) i, ø, y i (82.61 %), ø (15.22 %), y (2.17 %) i
u u u, ø u (68.18 %), ø (31.82 %) u
y y y y
a a a a
o o o o
ə ə ə ə
ei ei ei ei
ai ai ai ai
e e e e
au au au au
ou ou ou ou
an an, aŋ aŋ (18 %) an an
ən ən ən ən
aŋ aŋ, an an (7 %) aŋ aŋ
əŋ əŋ əŋ, ən əŋ (66.67 %), ən (33.33 %) əŋ
ə˞ ə˞ ə˞ ə˞
ia ia ia ia
ua ua ua ua
2.5 Mandarin Phonological Development by Children of Indonesian Mothers
ͻŋ ͻŋ ͻŋ ͻŋ
ie ie ie ie
ye ye ye ye
uai uai uai uai
(continued)
71
Table 2.20 (continued)
72
Phoneme Phonetic representations, first year Phonetic representations, second year Phonetic representations, third
year
Frequency Frequency Frequency
uei uei uei uei
iau iau iau iau
iou iou iou iou
ien ien ien ien
uan uan uan uan
yen yen yen yen
in in in in
uən uən uən uən
yn yn yn yn
iaŋ iaŋ iaŋ iaŋ
uaŋ uaŋ uaŋ uaŋ
iŋ iŋ iŋ, in iŋ (85.71 %), in (14.29 %) iŋ
ͻŋ ͻŋ ͻŋ ͻŋ
jͻŋ jͻŋ jͻŋ jͻŋ
uo uo uo uo
2 Phonological Development
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 73
The six-year-old Indonesian girl has acquired most phonemes in Mandarin except
alveopalatals, retroflexes, and final nasals. The child sometimes changed alveopa-
latal affricates to alveolar ones; for example, /tɕhyà n pu/́ ‘whole’ was pronounced as
[tshɔŋ̀ pu]́ . She mixed retroflexes and alveolars. She replaced retroflexes with
alveolars most of the time. For example, /tʂ̄ taú / ‘know’ was pronounced as [ts̄ taú ],
/piá n tʂhəŋ̀ / ‘become’ was pronounced as [piá n tshəŋ̀ ], /niá n ʂū / ‘study’ was pro-
nounced as [niá n su], and /ʐəǹ / ‘person’ was pronounced as [ləǹ ]. At other times,
alveolars were replaced by retroflexes. For example, /tś tɕǐ / ‘self’ was pronounced
as [tʂ́ tɕǐ ] and̄ /lan
̀ sə/́ ‘blue’ was pronounced as [laǹ ʂə]́ . She occasionally changed
the place of final nasals or even dropped it. For example, /tɕhyà n pu/́ ‘whole’ was
pronounced as [tshɔŋ̀ pu]́ , /xuɑŋ̀ sə/́ ‘yellow’ was pronounced as [xuà n sə]́ , and /sə̄ n
lin/ ‘forest’ was pronounced as [sə̄ n li]. The child did not seem to have difficulty
with particular place articulation, but was confused about when to use which one.
Only the replacement of retroflexes with alveolars was observed in the mother.
In the second year, the mistakes with low frequency in the first year disappeared but
some new errors were observed. There were substitutions of the bilabial nasal /m/
with the high back vowel /u/ and the substitution of the alveolar lateral /l/ with the
alveolar nasal [n]. The final velar nasal /ö/ was replaced by the alveolar nasal [n]
with higher frequency (30 %). The final alveolar nasal, which was dropped occa-
sionally in the first year, was pronounced correctly most of the time while replaced
by the velar nasal 25 % of the time. The retroflex mid central vowel /Ä/ was
pronounced without retroflexation. She replaced aspirated retroflex affricates and
fricatives with alveolars 100 % of the time. The unaspirated retroflex affricates / tÏ/
co-exist with the alveolar and alveolpalatal variants.
In the third year, the girl has acquired all the phonemes except replacing retroflexes
with alveolars consistently. The aspirated voiceless retroflex affricates /tÏh/ were
pronounced correctly 75 % of the time though. The acquisition and substitutions of
the sounds by the six-year-old Indonesian girl for three years were summarized in
Table 2.21.
The six-year-old Indonesian boy and girl both high-ranked *Retroflex and
changed the retroflex fricative to the alveolar fricative throughout the three years as
shown in Tableau 2.8.
Table 2.21 Acquisition and substitutions of the sounds from the Indonesian girl: 6, 6 + 1, and 6 + 2 years of age
74
Phoneme Phonetic representations, first Phonetic representations, second year Phonetic representations, third
year year
Frequency Frequency Frequency
p p, f f (2 %) p p
ph ph ph ph
m m m, u m (87.5 %), u (12.5 %) m
f f f f f
t t, ø ø (2 %) t t
th th th th
n n, l l (5 %) m n
l l l, n l (67.74 %), n (32.26 %) l
k k, t t (8 %) k k
kh kh kh kh
x x, f f (3 %) x x
tɕ tɕ, ɕ ɕ (3 %) tɕ tɕ
tɕh tɕh, tsh tsh (13 %) tɕh tɕh
ɕ ɕ ɕ ɕ
tʂ tʂ, ts ts (63 %) tʂ, ts, tɕ tʂ (8.33 %), ts (79.17 %), tɕ (12.5 %) ts 100 %
tʂh tʂh, tsh tsh (45 %) tsh 100 % tsh, ts tʂh (75 %), ts (25 %)
ʂ ʂ, s s (87 %) s 100 % s 100 %
ʐ ʐ, l l (80 %) n 100 % l 100 %
ts ts, tʂ tʂ (16 %) ts ts
tsh tsh tsh tsh
s s, ʂ ʂ (15 %) s, ʂ, ts s (88.24 %), ʂ (5.88 %), ts (5.88 %) s, ʂ ʂ once
ŋ ŋ ŋ ŋ
(continued)
2 Phonological Development
Table 2.21 (continued)
Phoneme Phonetic representations, first Phonetic representations, second year Phonetic representations, third
year year
Frequency Frequency Frequency
i i, y, ø y (4 %), ø (2 %) i i
u u u u
y y y y
a a a a
o o o o
ə ə ə ə
ei ei ei ei
ai ai ai, ei ai (96.3 %), ei (3.7 %) ai
e e e e
au au au, a au (94.44 %), a (5.56 %) au
ou ou ou ou
an an, ͻŋ ͻŋ (4 %) an an
ən ən ən ən
aŋ aŋ, an an (3 %) aŋ, an aŋ (70 %), an (30 %) aŋ
əŋ əŋ əŋ əŋ
ə˞ ə˞ ə 100 % ə˞
ia ia ia ia
ua ua ua ua
2.5 Mandarin Phonological Development by Children of Indonesian Mothers
ɔŋ ɔŋ ɔŋ ɔŋ
ie ie ie ie
ye ye ye ye
uai uai uai uai
(continued)
75
Table 2.21 (continued)
76
Phoneme Phonetic representations, first Phonetic representations, second year Phonetic representations, third
year year
Frequency Frequency Frequency
uei uei uei uei
iau iau iau iau
iou iou iou iou
ien ien ien ien
uan uan uan uan
yen yen yen yen
in in, i i (11 %) in, iŋ in (75 %), iŋ (25 %) in
uən uən uən uən
yn yn yn yn
iaŋ iaŋ iaŋ iaŋ
uaŋ uaŋ uaŋ uaŋ
iŋ iŋ iŋ, in iŋ (83.33 %), in (16.67 %) iŋ
jɔŋ jɔŋ jɔŋ jɔŋ
uo uo uo uo
2 Phonological Development
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 77
Tableau 2.8 OT tableau for the Indonesian children: 6, 6 + 1, and 6 + 2 years of age
ʂ Ident (astrident) *Retroflex *Affricate *(+cont) *(+anterior)
☞s * *
ʂ *! *
The Taiwanese two-year-old boy has acquired most of the phonemes and just had
mistakes occasionally (under 10 %). The only frequent error is the replacement of
retroflexes with alveolars. For example, he pronounced /tʂáu ɕiáŋ tɕī / ‘camera’ as
[tśau ɕiáŋ tɕi], /tʂhùan/ ‘boat’ as [tshùan], /ʂū / ‘book’ as [sū ], and /tɕī tɕhí ʐə̀n/
‘robot’ as [tɕī tɕhí lə̀n]. He also deretroflexed the retroflex vowel; for example, /ʂ̀ ə́˞/
‘twelve’ was pronounced as [s̩ ̀ ə́]. He dropped final nasals or changed the place of
articulation occasionally. He also simplified vowel sequence occasionally, but it
was not systematic in terms of which one was dropped. For example, /kǔ o tʂ̄ / ‘juice’
was pronounced as [kǒ ts̩̄ ] and /nà kúo lài/ ‘bring’ was pronounced as [nà kú lài].
The mother also replaced retroflexes with alveolars and deretroflexed retroflex
vowels. She also changed the final velar nasal following /i/ to the alveolar nasal.
In the second year, the boy has improved a lot in pronunciation. The only problem
remained was retroflex and final nasal. He still replaced retroflexes with alveolars
and deretroflexed the retroflex vowel. In addition, he changed the final velar nasal
following /i/ to the alveolar nasal.
The third year is similar to the second year. The errors in retroflex and final nasal
seemed to have fossilized. He still replaced retroflexes with alveolars and dere-
troflexed the retroflex vowel and still changed the final velar nasal following /i/ to
the alveolar nasal. The phonological development of the two-year-old Taiwanese
boy is summarized in Table 2.22.
Table 2.22 Acquisition and substitutions of the sounds from the Taiwanese boy: 2, 2 + 1, and 2 + 2 years of age
78
Phoneme Phonetic representations, first year Phonetic representations, second Phonetic representations, third year
year
Frequency Frequency Frequency
p p, f, m p (94.25 %), f (4.59 %), m (1.14 %) p p
ph ph, p ph (88.88 %), p (11.11 %) ph ph
m m, p m (98.87 %), p (1.12 %) m m
f f f f
t t, l t (99.1 %), l (0.89 %) t t, ø ø once
th th, tɕh th (97.91 %), tɕh (2.08 %) th th, ø ø once
n n, l n (97.61 %), l (2.38 %) m n, l n (97.18 %)
l (2.82 %)
l l, ʐ l (97.29 %), ʐ (2.7 %) l l, ø ø once
k k, ø k (99.15 %), k k
ø (0.84 %)
kh kh, x kh (97.5 %), kh kh
x (2.5 %)
x x, ø x (99.23 %) x x, ø ø once
ø (0.76 %)
tɕ tɕ tɕ tɕ
tɕh tɕh tɕh tɕh
ɕ ɕ ɕ ɕ
tʂ tʂ, ts tʂ (41.98 %) tʂ, ts tʂ (6.25 %), ts (93.75 %) tʂ, ts tʂ (50.43 %), ts (49.57 %)
ts (58.01 %)
tʂh tʂh, tsh tʂh (35.71 %), tsh (64.28 %) tʂh, tʂh (14.29 %), tsh tʂh, tsh, tʂh (66.66 %), tsh (16.67 %), ts
tsh (85.71 %) ts (16.67 %)
(continued)
2 Phonological Development
Table 2.22 (continued)
Phoneme Phonetic representations, first year Phonetic representations, second Phonetic representations, third year
year
Frequency Frequency Frequency
ʂ ʂ, s, ts, ʂ (63.46 %), s (35.25 %), ts (0.64 %), ɕ s 100 % ʂ, s ʂ (55.38 %), s (44.62 %)
ɕ (0.64 %)
ʐ ʐ, l ʐ (21.05 %), l (78.94 %) l 100 % ʐ, l ʐ (88.89 %), l (11.11 %)
ts ts ts ts
tsh tsh tsh tsh
s s, f s (97.95 %), f (2.04 %) s s, ʂ ʂ once
ŋ ŋ ŋ ŋ
i i i i
u u, ø u (98.43 %), ø (1.56 %) u u
y y y y
a a a a
o o o o
ə ə, o, əŋ ə (96.41 %), o (3.4 %), əŋ (0.3 %) ə, o ə (95 %), o (5 %) ə
ei ei, i ei (96.96 %), i (3.03 %) ei ei
ai ai ai ai
e e e e
au au au au
ou ou ou ou
2.5 Mandarin Phonological Development by Children of Indonesian Mothers
In the first year, the child has acquired most phonemes in Mandarin while having
problems in producing alveolar nasals, alveopalatal fricatives, and retroflex affri-
cates. She tended to replace an alveolar nasal /n/ with an alveolar lateral [l] and
replace an alveopalatal fricative /ɕ/ with an unaspirated alveopalatal affricate [tɕ].
She pronounced retroflex affricates as alveolar affricates, which were also observed
in her parents’ speech. The father even pronounced the retroflex fricative /ʂ/ as an
alveolar fricative [s], e.g., /ʂi/́ ‘is’ ! [si]́ . The child had problems with vowel /o/
and /u/. She pronounced /o/ as [ej] and /u/ as [u], [ej] [i], or [ou] inconsistently, e.g.,
/tʂə́ kə̄ / ‘this’![tsé i kə], /khəì /̀ ‘can’ ! [khoǔ ǐ ]. Adults did not have vowel
problems.
In the second year, the alveolar nasal /n/ and an alveopalatal fricative /ɕ/ have been
acquired. However, the retroflex fricative /ʂ/ was still replaced by an alveolar
fricative [s], for example, /mǎ liŋ̀ ʂǔ / ‘potato’ ! [mǎ liŋ̀ sǔ ]. Retroflex affricates
were not observed in the data of the second year. A new error has been observed,
that is, the replacement of aspirated alveopalatal affricates with unaspirated ones
sometimes, e.g., /tɕhī ɛn pǐ / ‘pencil’ ! [tɕī ɛn pǐ ]. All the vowels have been
acquired. The child has acquired most of the phonemes except the retroflexes,
whose correct forms were not available in the input.
The child did not make any further progress in the third year. All the errors have
seemed to be fossilized. The retroflexes, fricatives, and affricates were replaced with
alveolar ones for example, /ʂaŋ́ miań / ‘on the top’ ! [saŋ́ miań ]. Aspirated
alveopalatal affricates were still replaced by unaspirated ones sometimes.
The phonological development from two-year-old Taiwanese girl for three years
is summarized in Table 2.23.
Putting in OT framework, like the Indonesian two-year-old girl, the Taiwanese
two-year-old boy also high-ranked * retroflex and so changed the retroflex affricate
to alveolar affricate in the first two years as shown in Tableau 2.9. In the third year,
the alveolar (49.57 %) and retroflex affricates (50.43 %) coexist. Those two kinds
of rankings coexist as shown in Tableau 2.9 and 2.10.
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 83
Table 2.23 Acquisition and substitutions of the sounds from the Taiwanese girl: 2, 2 + 1, and
2 + 2 years of age
Phoneme Phonetic representations, first year Phonetic Phonetic
representations, representations,
second year third year
Frequency Frequency Frequency
p p p p
ph ph ph ph
m m m m
f f (50 %), x (25 %), u (25 %) f f
x, u
t t t t
th th, ɕ th th
n l n n
l l, ø ø (9 %) l l
k k k k
kh kh kh kh
x x, t t (8 %) x c
tɕ tɕ tɕ tɕ
tɕh tɕh tɕh, tɕ (67 %) tɕh
tɕ
ɕ ɕ ɕ ɕ
tʂ ts, tɕ ts (14 %) ts ts (14 %) tʂ, ts (54 %)
ts
tʂh tʂh
ʂ ʂ s s (100 %) ʂ, s s (76 %)
ʑ ʑ
ts ts ts, tɕ (50 %) ts
tɕ
tsh tsh
s s s, ɕ ɕ (73 %) s
i i, u u (13 %) i i
u u u u
y y y
a a a a
o ei o o
ɤ ɤ, ei, i, ou, ei (6 %), ou (6 %), a ɤ ɤ
a (6 %)
e e
ai ai ai ai
ei ei ei ei
au au au au
ou ou, əŋ
(continued)
84 2 Phonological Development
The three-year-old Taiwanese boy had acquired all the vowels and finals in
Mandarin while had problems with some initial consonants. He had problems with
manner of articulations. Aspiration and frication seem to confuse him. He turned
aspirated voiceless alveolar stops /th/ to aspirated voiceless alveolar affricates /tsh/
while turned an unaspirated voiceless alveolar affricate /ts/ to an unaspirated
voiceless alveolar stop [t], for example, /tsaí laì / ‘again’ as [taí laì ]. He replaced an
alveopalatal fricative /ɕ/ with an unaspirated alveopalatal affricate [tɕ], for example,
/ɕǐ / ‘happy’ ! [tɕǐ ]. He pronounced the retroflex fricatives as alveolar fricatives or
stops and retroflex affricates as alveopalatal affricates, for example, /ʂi/́ ‘copula BE’
as [ɕi], /ʂou taú / ‘receive’ as [tou taú ], /tʂuē i tʂə/ ‘chase’ as [tsuē i tsə], and /tʂə̄ n tʂə̄ ŋ/
‘really’ as [tə̄ n təŋ́ ]. His mother replaced retroflexes with alveolars. No other errors
were observed in his mother’s speech.
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 85
In the second year, the child got rid of problems in aspiration and frication. He
pronounced aspirated stops and unaspirated affricates correctly. The alveopalatal
fricative /ɕ/ was also pronounced correctly. However, he still had problems with
retroflexes. Alveolar affricates coexist with the correct forms. The voiceless retro-
flex fricative /ʂ/ was consistently replaced by the alveolar one [s], e.g., /ʂu/̄
‘book’ ! [sū ], while the voiced one /ʐ/ was replaced by the lateral liquid [l], for
example, /ʐɤ́ kǒ u/ ‘hot dog’ ! [lɤ́ kǒ u].
In the third year, the child had acquired most phonemes except the typical Taiwan
Mandarin problems. They are retroflexes and the rounded high front vowel. He still
replaced the voiceless retroflex affricate /tʂ/ with voiceless alveolar affricate [ts] and
replaced the voiceless retroflex fricative /ʂ/ with the alveolar fricative [s], for
example, /ʂí/ ‘is’ as [sí]. The rounded high front vowel /y/ was pronounced correctly
in the first two years, but was replaced with the unrounded one [i] in the third year,
e.g., /ɕiǎ u ɕyɔ̀ŋ/ ‘baby bear’ ! [ɕiǎ u ɕiɔ̀ŋ]. The target forms that the child acquired
seem to be Taiwanese-interfered Mandarin. A summary of the child’s phonemes
and phonetic variants in three years is given in Table 2.24.
Table 2.24 Acquisition and substitutions of the sounds from the Taiwanese boy: 3, 3 + 1, and
3 + 2 years of age
Phoneme Phonetic representations, Phonetic representations, Phonetic representations,
first year second year third year
Frequency Frequency Frequency
p p p p
ph ph ph ph
m m m m
f f f f
t t t t
th tsh th th
n n n n
l n l l
k k k k
kh kh kh kh
x x x, f x
tɕ tɕ tɕ tɕ
tɕh tɕh tɕh tɕh
ɕ ɕ, tɕ, ʂ ʂ (29 %) ɕ ɕ
tʂ tʂ, tɕ, t, tʂ (23 %), tɕ tʂ, ts, s ts (97 %), s tʂ (50 %), ts
ts (3 %), (2 %) (50 %)
t (31 %), ts
(43 %)
(continued)
86 2 Phonological Development
In the second year, the only problems left were retroflexes and omission of /u/ in
/uo/ sequence. As in the first year, she still replaced retroflexes with alveolars. Other
errors have disappeared.
In the third year, the girl consistently replaced retroflexes with alveolars. They seem
to have fossilized. The aspirated retroflex affricate /tʂh/ was pronounced as alveolar
affricate with or without aspiration. The phonemes and phonetic variations of the
Taiwanese three-year-old girl for three years are summarized in Table 2.25.
Both Taiwanese three-year-old boy and girl high-ranked *Retroflex and thus
pronounced retroflex fricative as alveolar fricative throughout three years as in
Tableau 2.11.
In the second year, the omission of final nasals and voiceless velar fricatives was no
longer observed. Neither was nasal assimilation. The voiced retroflex fricative /ʐ/
was acquired. The only problem left was the replacement of voiceless retroflex
fricatives and affricates—/ʂ/, /tʂ/, and /tʂh/ with alveolar ones—[s], [ts], and [tsh],
e.g., /ʂā fā / ‘sofa’ ! [sā fā ], /tʂǐ / ‘paper’ ! [tsǐ ], and /tʂhǐ / ‘ruler’ ! [tshǐ ].
Table 2.25 Acquisition and substitutions of the sounds from the Taiwanese girl: 3, 3 + 1, and 3 + 2 years of age
88
Phoneme Phonetic representations, first year Phonetic representations, second year Phonetic representations, third year
Frequency Frequency Frequency
p p p p
ph ph, kh kh (10 %) ph ph
m m m m
f f f f
t t t t
th th, x x (5 %) th th
n n m n
l l l l
k k k k
kh kh kh kh
x x x x
tɕ tɕ tɕ tɕ
tɕh tɕh tɕh tɕh
ɕ ɕ ɕ ɕ
tʂ ts 100 % ts 100 % ts 100 %
tʂh tsh 100 % tʂh, tsh tʂh (20 %) tsh (80 %) tʂh, tsh tsh (57.14 %), ts (42.86 %)
ʂ s 100 % s 100 % s 100 %
ʐ l, n l (85 %), n (25 %) ʐ, l ʐ (33.33 %), l (66.67 %) l 100 %
ts ts ts ts
tsh tsh tsh tsh
s s s s
ŋ ŋ ŋ ŋ
i i, u, ian u (5 %), ian (5 %) i i
(continued)
2 Phonological Development
Table 2.25 (continued)
Phoneme Phonetic representations, first year Phonetic representations, second year Phonetic representations, third year
Frequency Frequency Frequency
u u, ø ø (10 %) u, ø u (69.23 %), ø (30.77 %) u
y y y y
a a a a
o o, ͻŋ ɔŋ (2 %) o o
ə ə ə ə
ei ei ei ei
ai ai, an an (6 %) ai ai
e e e e
au au au au
ou ou ou ou
an an, ai ai (3 %) an an
ən ən ən ən
aŋ aŋ aŋ aŋ
əŋ əŋ əŋ əŋ
ə˞ ə˞ ə 100 % ə˞
ia ia ia ia
ua ua ua ua
ɔŋ ɔŋ ɔŋ ɔŋ
ie ie ie ie
2.5 Mandarin Phonological Development by Children of Indonesian Mothers
ye ye ye ye
uai uai uai uai
uei uei uei uei
iau iau iau iau
(continued)
89
Table 2.25 (continued)
90
Phoneme Phonetic representations, first year Phonetic representations, second year Phonetic representations, third year
Frequency Frequency Frequency
iou iou iou iou
ien ien ien ien
uan uan uan uan
yen yen yen yen
in in in in
uən uən uən uən
yn yn yn yn
iaŋ iaŋ iaŋ iaŋ
uaŋ uaŋ uaŋ uaŋ
iŋ iŋ iŋ iŋ
jɔŋ jɔŋ jɔŋ jɔŋ
uo uo uo uo
2 Phonological Development
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 91
In the third year, the child acquired the aspirated retroflex affricate /tʂh/, but still
replaced unaspirated voiceless retroflex affricates—/tʂ/ and the voiceless retroflex
fricative /ʂ/ with alveolar ones—[ts] and [s], e.g., /tʂə́ tʂā ŋ/ ‘this piece’ as [tsə́ tsā n]
and /ʂi/́ ‘is’ as [si]́ . A new error was observed. That is the replacement of final velar
nasal by alveolar nasal. The child’s phonological development for three years is
summarized in Table 2.26.
Table 2.26 Acquisition and substitutions of the sounds from the Taiwanese boy: 4, 4 + 1, and
4 + 2 years of age
Phoneme Phonetic representations, first year Phonetic Phonetic
representations, representations,
second year third year
Frequency Frequency Frequency
p p, ph p (86 %) p p
ph (14 %)
ph ph, tɕh ph (89 %) ph ph
tɕh (11 %)
m m m m
f f f f
t t, ts t t
th th, kh, th (56 %) th th
ph, t kh (27 %)
ph (9 %)
t (9 %)
n ŋ n n
l l, n l (87 %) l l
n (13 %)
k k, th th (9 %) k k
kh kh, th kh kh
x x, ø x (73 %) x x
ø (27 %)
tɕ tɕ, ɕ tɕ (92 %) tɕ tɕ
ɕ (8 %)
tɕh tɕh, tɕ, tɕ (25 %) tɕh tɕh
ts ts (13 %)
ɕ ɕ ɕ ɕ
tʂ tʂ, ts, t, tʂ (71 %), ts (12 %), t ts tʂ, tʂ (52 %),
th, ʂ (12 %), th (2 %), ʂ (2 %) (100 %) ts ts (48 %)
(continued)
92 2 Phonological Development
The Taiwanese four-year-old girl has acquired most of the phonemes in Mandarin
and just replaced retroflexes with alveolars and changed final velar nasals to
alveolar nasals. She replaced retroflexes with alveolars as /tʂə́ kə/ ‘this one’ was
pronounced as [tsə́ kə], [tʂ̄ h] ‘eat’ as [ts̩̄ h], and /ʂ/ ‘is’ as [s̩ ].
́ He changed the velar
nasal following high front vowel to the alveolar nasal; for example, /ly̌ ɕiŋ̀ / ‘travel’
was pronounced as [ly̌ ɕiǹ ]. The same errors were observed in the mother.
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 93
Tableau 2.12 OT tableau for the Taiwanese children: 4, 4 + 1, and 4 + 2 years of age
ʂ Ident (astrident) *Retroflex *(+anterior) Ident (aanterior)
ʂ *!
☞s * *
In the third year, the correct target of retroflex coexisted with alveolar variations
with low frequency. The retroflex mid-central lax vowel /ə˞/ was consistently
pronounced without retroflex [ə], which was pronounced correctly in the last two
years. Final velar nasal next to /i/ was replaced by alveolar nasal with higher
frequency 62.96 % (Table 2.27). The child’s phonological development for three
years is summarized in Table 2.27.
The Taiwanese four-year-old boy and girl also high-ranked * retroflex and so
changed the retroflex fricative to alveolar fricative as shown in Tableau 2.12. The
high ranking of *retroflex by both Indonesian and Taiwanese children proves it to
be a universal constraint.
The five-year-old Taiwanese boy had difficulties with some places of articulation.
He mixed alveolars and retroflexes, changed alveolar nasal to lateral, and changed
final velar nasal to alveolar. He used retroflexes and alveolars interchangeably. He
could pronounce both, but was not clear when to use which one. Retroflexes were
changed to alveolars in the following examples. /tʂə́ lǐ / ‘here’ was pronounced as
[tsə́ lǐ ]. /saí tʂhə̄ / ‘racing car’ was pronounced as [saí tsə̄ ]. /ʂǔ ei kǔ o/ ‘fruit’ was
pronounced as [sǔ ei kǔ o], and /ʐəǹ / ‘person’ was pronounced as [ləǹ ]. On the other
hand, alveolars were changed to retroflexes in the following examples. /tshəŋ̀ /
‘level’ was pronounced as [tʂhəǹ ], /tš sə/́ ‘purple’ as [tʂ̌ sə]́ , and /sā n/ ‘three’ as
[ʂā n]. He changed the alveolar nasal to a lateral as /naǹ / ‘difficult’ was pronounced
̀ ]. Velar nasals in coda were changed to alveolar nasals; for example, /t ī ŋ/
h
as [lan
‘listen’ was pronounced as [t ī n] and /tɕiŋ́ ts̩ / ‘mirror’ was pronounced as [tɕiń ts̩ ].
h
Table 2.27 Acquisition and substitutions of the sounds from the Taiwanese girl: 4, 4 + 1, and 4 + 2 years of age
94
Phoneme Phonetic representations, first year Phonetic representations, second Phonetic representations, third year
year
Frequency Frequency Frequency
p p p p
ph ph ph ph
m m, p, ø m (99.20 %), p (0.39 %), ø (0.39 %) m m
f f, x f (90 %), x (10 %) f f
t t t t
th th, ø th (96.77 %), ø (3.22 %) th th
n n m n, l l once
l l l l
k k k k
kh kh kh kh
x x x x
tɕ tɕ, ø tɕ (98.24 %), ø (1.75 %) tɕ tɕ
tɕh tɕh tɕh tɕh
ɕ ɕ, tɕh ɕ (98.93 %), tɕh (1.06 %) ɕ ɕ
tʂ tʂ, ts, tɕ tʂ (18.23 %), ts (80.50 %), tɕ (1.25 %) ts 100 % tʂ, ts tʂ (6.85 %), ts (93.15 %)
tʂh tʂh, tsh, x tʂh (15.38 %), tsh (80.76 %), x (3.84 %) tsh 100 % tʂh, tsh tʂh (7.14 %), tsh (92.86 %)
ʂ ʂ, s, ɕ ʂ (21.66 %), s (77.22 %), ɕ (1.11 %) s 100 % ʂ, s ʂ (3.1 %), s (96.90 %)
ʐ ʐ, l ʐ (33.33 %), l (66.66 %) l 100 % ʐ, l ʐ (15.38 %), l (84.62 %)
ts ts ts ts
tsh tsh tsh tsh
s s s s
(continued)
2 Phonological Development
Table 2.27 (continued)
Phoneme Phonetic representations, first year Phonetic representations, second Phonetic representations, third year
year
Frequency Frequency Frequency
ŋ ŋ ŋ ŋ
i i i i
u u, iau, ien u (97.72 %), iau (1.13 %), ien (1.13 %) u u
y y y y
a a a a
o o o o
ə ə ə ə
ei ei ei ei
ai ai, uei ai (97.05 %), uei (2.94 %) ai ai
e e e e
au au au au
ou ou ou ou
an an, iaŋ an (97.61 %), iaŋ (2.38 %) an an
ən ən ən ən
aŋ aŋ aŋ aŋ
əŋ əŋ əŋ əŋ
ə˞ ə˞ ə˞ ə 100 %
ia ia ia ia
2.5 Mandarin Phonological Development by Children of Indonesian Mothers
ua ua ua ua
ͻŋ ͻŋ ͻŋ ͻŋ
ie ie ie ie
(continued)
95
Table 2.27 (continued)
96
Phoneme Phonetic representations, first year Phonetic representations, second Phonetic representations, third year
year
Frequency Frequency Frequency
ye ye ye ye
uai uai uai uai
uei uei uei uei
iau iau iau iau
iou iou iou iou
ien ien ien ien
uan uan uan uan
yen yen yen yen
in in in in
uən uən uən uən
yn yn yn yn
iaŋ iaŋ iaŋ iaŋ
uaŋ uaŋ uaŋ uaŋ
iŋ iŋ, in iŋ (3.22 %), in (96.77 %) iŋ, in iŋ (57.14 %), in (42.86 %) iŋ, in iŋ (37.04 %), in (62.96 %)
jͻŋ jͻŋ jͻŋ jͻŋ
uo uo, o uo (99.15 %), o (0.84 %) uo uo
2 Phonological Development
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 97
His mother also replaced retroflexes with alveolars, but not the other way
around. For example, /tʂə̄ n tə/ ‘real’ was pronounced as [tsə̄ n tə]. The change of
velar nasals in coda to alveolar nasals was also observed in the mother. For
example, /ɕī en ʂə̄ ŋ/ ‘husband’ was pronounced as [ɕī en ʂə̄ n].
In the second year, the aspirated voiceless bilabial stop was pronounced without
aspiration. The frequency of the replacement of retroflexes with alveolars
decreased. The voiced retroflex fricative /ʐ/ was pronounced correctly. The retroflex
mid-central vowel was pronounced without retroflex all the time. Final velar nasals
were changed to alveolar nasals sometimes.
In the third year, the aspiration problem disappeared. The only problem left is
retroflex. Retroflex affricates and fricatives were replaced with alveolar counter-
parts. Voiced retroflex fricative /ʐ/ was replaced by [l] again as in the first year most
of the time. The phonetic variations of the Taiwanese five-year-old boy for three
years are summarized in Table 2.28.
Table 2.28 Acquisition and substitutions of the sounds from the Taiwanese boy: 5, 5 + 1, and
5 + 2 years of age
Phoneme Phonetic representations, Phonetic Phonetic
first year representations, representations,
second year third year
Frequency Frequency Frequency
p p p p
ph ph p 100 % ph
m m m m
f f f f
t t t t
th th th th
n n n n
l l l l
k k k k
kh kh kh kh
x x x x
tɕ tɕ tɕ tɕ
tɕh tɕh tɕh tɕh
ɕ ɕ ɕ ɕ
(continued)
98 2 Phonological Development
The child can pronounce all the phonemes in Mandarin quite accurately except
retroflexes, which were sometimes replaced by alveolar counterparts. Voiced ret-
roflex fricatives /ʐ/ were also replaced by the lateral [l] sometimes, e.g., /xuá i ʐəǹ /
‘bad person’ as [xuá i ləǹ ]. The mother also replaced retroflexes with alveolars, but
never changed the manner of articulation.
In the second year, the child continued to replace retroflexes with alveolars. The
voiceless retroflex fricative /ʂ/ was consistently replaced with the voiceless alveolar
fricative [s]; for example, /ʂǔ ei kǔ o/ ‘fruit’ was pronounced as [sǔ ei kǔ o] and the
voiced retroflex fricative /ʐ/ was consistently replaced by the alveolar lateral /l/,
̀ / ‘human being’ ! [ləǹ ]. In addition to the aspirated voiceless alveolar
e.g., /ʑən
affricate [ts] used to replace the aspirated voiceless retroflex affricate /tʂh/ in the first
100 2 Phonological Development
year, two other phonetic variants—/tɕh/ and /th/—showed up in the second year,
e.g., /tʂhuɑŋ̀ / ‘bed’ as [tshuɑŋ̀ ]. The child might have noticed his inaccurate pro-
nunciation and was trying other variants to approach the target.
All the phonemes were acquired in the third year except the unaspirated retroflex
affricate /tʂ/, which was replaced with the alveolar affricate [ts] sometimes. The
child’s phonological development for three years is summarized in Table 2.29.
Let us put /tʂ/ in OT framework. The five-year-old Taiwanese children
high-ranked * Retroflex and so changed retroflex affricate to alveolar affricates
throughout three years as in Tableau 2.13.
Table 2.29 Acquisition and substitutions of the sounds from the Taiwanese girl: 5, 5 + 1, and
5 + 2 years of age
Phoneme Phonetic representations, Phonetic Phonetic representations,
first year representations, third year
second year
Frequency Frequency Frequency
p p p p
ph ph ph ph
m m m m
f f f f
t t, th t (80 %), th (20 %) t t
th th th th
n n n n
l l l l
k k k k
kh kh kh kh
x x x x
tɕ tɕ tɕ tɕ
tɕh tɕh tɕh tɕh
ɕ ɕ ɕ ɕ
tʂ ts, s ts (86 %), s (14 %) tʂ, ts, tɕ ts (93 %) tʂ, ts tʂ (46 %), ts (54 %)
tʂh tsh tsh (100 %) tsh, th tsh (50 %), tʂh
th (50 %)
ʂ ʂ, s ʂ (2 %), s (98 %) ʂ, ɕ s (100 %)
ʑ l l (100 %) l l (100 %)
ts ts, tɕ ts (10 %), tɕ (90 %) ts, tʂ ts
tsh tsh tsh
s s s, ʂ s
(continued)
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 101
Table 2.29 (continued)
Phoneme Phonetic representations, Phonetic Phonetic representations,
first year representations, third year
second year
Frequency Frequency Frequency
i i i i
u u u u
y y y y
a a a a
o o o o
ɤ ɤ ɤ
e e e e
ai ai ai ai
ei ei ei ei
au au au au
ou uo ou ou
an an an an
ən ən ən ən
aŋ aŋ aŋ aŋ
əŋ əŋ əŋ əŋ
ɤ.ɪ ɤ.ɪ ɤ.ɪ
The child has acquired all the phonemes in Mandarin. The only errors sometimes
occurred were in retroflex consonants and final velar nasals. He sometimes pro-
nounced retroflex consonants as alveolar consonants just like his mother, e.g., /lǎ u
ʂī / ‘teacher’ as [lǎ u sī ]. This is actually quite common in Mandarin in Taiwan. Also,
he replaced the final velar nasal with an alveolar nasal, e.g., /thaŋ/ ‘soup’ ! [than].
In the second year, the child pronounced the aspirated retroflex affricate /tʂh/ and
voiced retroflex fricative /ʐ/ correctly, but still replaced unaspirated retroflex
affricate /tʂ/ and voiceless retroflex fricative /ʂ/ with their alveolar counterparts—
[ts] and [s] sometimes, e.g., /tʂǐ / ‘paper’ as [tsǐ ] and /xaì ʂǐ / ‘still’ as [xaí sǐ ]. The
final velar nasal was produced accurately.
102 2 Phonological Development
No progress was made in the third year. The child consistently replaced the
unaspirated retroflex affricate /tʂ/ with the alveolar affricate [ts]. As for the voiceless
retroflex fricative /ʂ/, the correct phonetic representation coexisted with the alveolar
counterparts—[s]. A summary of the acquisition of sounds by the six-year-old boy
is presented in Table 2.30.
Table 2.30 Acquisition and substitutions of the sounds from the Taiwanese boy: 6, 6 + 1, and
6 + 2 years of age
Phoneme Phonetic representations, first year Phonetic Phonetic
representations, representations,
second year third year
Frequency Frequency Frequency
p p p p
ph ph ph ph
m m m m
f f f f
t t, th th (20 %) t t
th th th th
n n, l l (95 %) b b
l l l l
k k k k
kh kh kh kh
x x x x
tɕ tɕ tɕ tɕ
tɕh tɕh tɕh tɕh
ɕ ɕ ɕ ɕ
tʂ tʂ, ts ts (94 %) tʂ, ts ts (88 %) tʂ, ts ts (41 %)
tʂh tʂh, tsh, tɕh tah (68 %), tɕh (18 %) tʂh, ts ts (20 %) tʂh
ʂ ʂ, s s (97 %) ʂ, s s (33 %) ʂ, s s (50 %)
ʑ ʐ, z z (81 %) ʑ
ts ts ts ts
tsh tʂh tʂh (100 %) tsh
s s s s
i i i i
u u u u
y y y y
a a, ai ai (25 %) a a
o o o o
ɤ ɤ ɤ
(continued)
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 103
The Taiwanese six-year-old girl mixed alveolar laterals with alveolar nasals,
replaced retroflexes with alveolars, changed velar nasals in coda to alveolars, and
simplified vowel sequence. The child used alveolar laterals and nasals inter-
changeably. For example, /lǎ n tuó / ‘lazy’ was pronounced as [nǎ n tuó ] and /nioù
nǎ i/ ‘milk’ was pronounced as [liòu nǎ i]. She replaced retroflexes with alveolars.
For example, she pronounced /tʂ̩̌ / ‘paper’ as [tš̩ ], /xǔ o tʂhə̄ / ‘train’ as [xǔ o tshə̄ ], /ʂū /
‘book’ as [sū ], and /ʐə̀n/ ‘person’ as [lə̀n]. The velar nasal was changed to an
alveolar nasal as /mìŋ thiē n/ ‘tomorrow’ was pronounced as [mìn thiē n]. The /i/ in /
ie/ sequence was dropped sometimes as /ná pī en/ ‘there’ was pronounced as [ná
pēn].
Most of the child’s errors were also observed in the mother such as mixing
alveolar laterals with nasals, replacing retroflexes with alveolars, and changing
velar nasals to alveolar nasals. In addition, the mother sometimes pronounced the
voiceless labiodental fricative as a velar fricative. For example, /ī fù/ ‘clothes’ was
pronounced as [ī xù].
In the second year, the exchange of laterals and nasals disappeared. However, she
still replaced retroflexes with alveolars with lower frequency. The velar nasal was
changed to an alveolar nasal when followed by /i/ with lower frequency.
104 2 Phonological Development
In the third year, the retroflexes were replaced by alveolars consistently. The
mid-central retroflex vowel was pronounced without retroflex all the time. The
problems with final nasals disappeared. The phonological development seems to
have stabilized. The acquisition and phonetic variations of the Taiwanese
six-year-old girl are summarized in Table 2.31.
Let us see the acquisition of retroflex in the OT framework. The six-year-old
Taiwanese girl high-ranked the faithfulness constraint ident (aretroflex) and
changed alveolar affricate to retroflex affricate. That could be due to overcorrection
when acquiring retroflexes. She demoted it and reached the target in the following
two years as shown in Tableau 2.14 and 2.15.
Table 2.31 Acquisition and substitutions of the sounds from the Taiwanese girl: 6, 6 + 1, and
6 + 2 years of age
Phoneme Phonetic representations, first year Phonetic Phonetic
representations, second representations,
year third year
Frequency Frequency Frequency
p p p p
ph ph ph ph
m m m m
f f f f
t t t t
th th th th
n n, l n (98 %), l (2 %) m n
l l, n l (99 %), n (1 %) l l
k k k k
kh kh kh kh
x x x x
tɕ tɕ tɕ, tɕ (92.86 %), tɕh tɕ
tɕh (7.14 %)
tɕh tɕh tɕh tɕh
ɕ ɕ ɕ ɕ
tʂ tʂ, tʂ (3 %), ts (97 %) tʂ, tʂ (22.22 %), ts ts 100 %
ts ts (77.78 %)
tʂh tʂh, tʂh (8 %), tsh (92 %) tʂh, tʂh (66.67 %), tsh 100 %
tsh tsh tsh (33.33 %)
ʂ ʂ, s ʂ (1.1 %), s (98.9 %) ʂ, s ʂ (50 %), s s 100 %
(50 %)
ʐ l, n ʐ (6.4 %), l (35.4 %), n ʐ l, n l (91.49 %), n
(41.9 %), z (16.1 %) (8.51 %)
(continued)
2.5 Mandarin Phonological Development by Children of Indonesian Mothers 105
Tableau 2.15 OT tableau for tsh Ident (as.g) Ident (aretroflex) *(+anterior)
the Taiwanese children: 6 and
6 + 2 years of age tsh *
☞ tʂh *!
Children of Indonesian immigrants have unique and common difficulties with the
children of Vietnamese or Taiwanese mothers. The phonetic variants of children of
Indonesian and Taiwanese mothers are summarized in Table 2.32. The first column
lists the target phonemes in Mandarin. The phonetic variants observed in the
children of Indonesian were listed in the second column and those produced by the
children of Taiwanese were listed in the last column. Omission is indicated by Ø.
Like the children of Vietnamese and Taiwanese, the children of Indonesian have
difficulties in retroflexes, aspirations, and nasals. They tended to replace retroflexes
with alveolars just like many adult Mandarin speakers in Taiwan. For example, the
three-year-old Indonesian boy pronounced /tʂə́/ ‘this’ as [tsə́], which is also
observed in Huang (2011) and Hsu (1996). *Retroflex outranked ident (aretroflex).
This could be due to the universal markedness of retroflex or the children’s input.
Retroflexes usually do not preserve in adult Mandarin in Taiwan. That could just be
their target rather than an error.
Children of Indonesian mothers failed to discriminate aspiration. For example,
unaspirated bilabial stop /p/ and unaspirated velar stop /k/ were replaced by their
aspirated counterparts. Deaspirations were also observed. Fuzzy discrimination of
aspiration was also found in the children of Taiwanese mothers. Hsu (1996) also
found the substitutions of aspirated stops with unaspirated stops in the children of
Taiwanese parents.
2.6 Discussion and Summary 107
Table 2.32 Comparison of phonetic variation between children of Indonesian and Taiwanese
mothers
Phoneme Phonetic variations
Indonesian Taiwanese
p p, ph, f, ø, m p, ph, f, m
ph ph, p, x, f, th ph, p, kh, tɕh
m m, l, u m, p, ø
f f, u, p, ph, w, x f, u, x
t t, k, l, th, ø t, l, thø, ts
th th, x, ø, tsh, tʃ th, tɕh, ø, ɕ, x, kh, ph, t
n n, l, m n, l, ŋ
l l, n, k n, l, ŋ
k k, kh, ø, t k, ø, th
kh kh kh, x, th
x x, ø, u, f x, ø, t
tɕ tɕ, tɕh, t, ɕ, ts, th, ø tɕ, ɕ, ø, tɕh
tɕh tɕh, tɕ, tsh, ɕ, ø tɕh, tɕ, ts
ɕ ɕ, tɕ, tɕh, ɕ, s, kh, ts, k, t ɕ, ts, ʂ, tɕh
tʂ ts, tɕ, t tʂ, ts, tɕ, t, th, t, x
tʂh tʂh, tsh, tɕh, kh, th, s tʂh, tsh, tɕh, kh, ts, th, t, x
ʂ ʂ, s, ts, ɕs, tsh, kh, x, k, th, tʂ, t, ɕ ʂ, s, ts, ɕ, t
ʐ ʐ, l ʐ, l, n, z
ts ts, tɕ, k, tʂ, t ts, tʂ, tɕ, t, kh
tsh tsh, kh, tɕh, tʂh, tʂ, k, ø tsh, tʂh, kh
s s, ts, ʂ, kh, t, ɕ s, f, ʂ, ɕ
ŋ ŋ ŋ
All the children have problems with nasals. Younger children tended to drop
nasals in coda, while older children changed the place of articulation. The
two-year-old Indonesian girl pronounced /tan/ ‘egg’ as [ta]. The four-year-old
Indonesian boy pronounced /tshū an/ ‘wear’ as [tshū aŋ] ‘window.’ The six-year-old
Indonesian boy pronounced /ʂā ŋ/ ‘hurt’ as [san] ‘three.’ Changing the place of final
nasals was also observed in Indonesian mothers. Children of Indonesian were often
confused by aspiration like their mothers. For example, the three-year-old
Indonesian boy said /tɕin/ ‘enter’ as [tɕhin]. Children in both groups were accu-
rate in tones, regardless of the Indonesian mothers’ tone errors. The Indonesian
mother of the four-year-old girl and five-year-old boy pronounced the high-level
tone as the falling tone, for example, /ɕiā ŋ kū / ‘mushroom’ as [ɕiáŋ kū ].
The unique problems observed in the children of Indonesian include simplifi-
cation of complex vowels and preference for affricates and palatals. Younger
children of Indonesian mothers tended to simplify complex vowels, which have not
been observed in the children of Taiwanese or Vietnamese mothers. For example,
the Indonesian two-year-old girl pronounced /tiau/ ‘lose’ as [ta]. Only one vowel
108 2 Phonological Development
was left in the three-vowel sequence. The Indonesian boy pronounced /tuan/
‘broken’ as [tan]. Only one vowel was left in the two-vowel sequence. /a/ tended to
the vowel preserved.
The challenging alveopalatal affricates for the children of Vietnamese and
Taiwanese mothers (Kuo 2008) do not seem to be a problem for the children of
Indonesian mothers. They preferred affricates to fricatives. The two-year-old
Indonesian girl pronounced the alveolar fricative /s/ as the alveolar affricate [ts], for
example, /sə/ ‘color’ as [tsə]. The six-year-old Indonesian boy pronounced the
alveopalatal fricative /ɕ/ as alveopalatal affricate [tɕ]; for example, /ɕi/ ‘river’ was
pronounced as [tɕi].
Children of Indonesian mothers changed alveopalatals to palatals, which were
not observed in the children of Taiwanese or Vietnamese mothers. The
three-year-old and four-year-old Indonesian boys changed alveopalatal fricatives to
palatal fricatives. For example, the three-year-old Indonesian boy pronounced /tɔ̄ ŋ
ɕī / ‘thing’ as [tɔ̄ ŋ ҫī ]. The four-year-old Indonesian boy pronounced /ɕiaŋ tau/
‘think’ as [ҫiaŋ tau]. These kinds of errors were also observed in their Indonesian
mothers. For example, the mother of Indonesian three-year-old boy pronounced /tś
tɕǐ / ‘self’ as [tɕí tɕǐ ].
The results of the present study show that the Mandarin phonological devel-
opment by the children of immigrant mothers might be influenced by their mothers’
Mandarin (L2), because they made similar errors. Since the same errors were
observed among the children of mothers of different nationalities, they cannot be
attributed to the mothers’ L1 alone. Despite the similar exposure to Taiwan
Mandarin as native children, children of immigrants made unique errors. This
indicates the importance of primary interlocutors, their mothers. Children of
immigrants have received input mediated by some universal constraints, whose
rankings are dynamic in the process of development. Innatism and interactionism
may not be mutually exclusive as the proposal for ‘innately guided learning’ (Gould
and Marler 1987; Jusczyk and Bertoncini 1988; Marler 1991).
References
Chinese
Chung, Rong Fu. 鐘榮富. (2002).台語的語音基礎 “Taiyu de Yuyin Jichu” (The nature of
Taiwanese). Taipei: The Crane Publishing Company.
Lin, Ling-Ying. 林玲英. (2005). 越籍配偶漢語語音偏誤現象初探 “Yueji peiou hanyu yuyin
pianwu xianxiang chutan” (The study of Vietnamese foreign spouses’ phonetic errors in
learning Mandarin Chinese). National Conference on Linguistics. Hsinchu: National
Chiao-Tung University.
Chapter 3
Lexical Development
Vocabulary is crucial not only because words allow speakers to express themselves,
but also because of the strong association between vocabulary and reading. If
students do not have sufficient vocabulary to understand the texts at school, they
cannot learn the content-area material for academic success.
Very young children understand the pragmatic intent of adults’ utterances before
they can understand the words themselves. This earlier comprehension is at the
emotional, social, and contextual levels. Only very slowly do children come to
understand the use of words independent of context (Gleason 2012). In this chapter,
I describe the theories for lexical development, phenomena in lexical development
in general and in Mandarin. Then, I will discuss the present study on Mandarin
lexical development of children of Indonesian and Taiwanese mothers in Taiwan.
There are two major theories of lexical development: learning theory and develop-
mental theories. According to learning theories, children learn the meanings of their
first words by associative learning. That is, by repeated exposure to a word (e.g.,
hearing doggie) paired with a particular experience (e.g., seeing the family dog), the
child makes connections between the word and meanings. Exclusively replying on
associative learning would be slow and result in many errors. However, children’s
learning of words is rapid, predictable, and remarkably accurate (Gleason 2012).
In contrast to the behavioral model, developmental theories consider semantic
development within the wider context of the child’s unfolding social, cognitive, and
linguistic skills (Gleason 2012). Clark (1993) claimed that by the time children start
to learn a language, they have developed a set of ontological categories including
objects, actions, events, relations, states, and properties. They learn words to match
these categories. In order to perform this mapping, children rely on their rudi-
mentary understanding of speakers’ attention and intentions and how these relate to
what is likely to be communicated (Tomasello 1995). Other lexical principles have
been suggested to assist children’s word learning task, for example, whole object
principle (Markman and Wachtel 1988), principle of mutual exclusivity (Markman
1987; Markman and Wachtel 1988) and principle of contrast (Clark 1987).
However, these lexical principles can be overridden by linguistic and world
knowledge (e.g., Akhtar 2002; Hall 1994).
Children as young as 18 months old start to make word-referent mapping after only
a few exposure to a new word, often also without explicit instruction (Houston-Price
et al. 2005). By early in their second year, most children have begun to produce
words themselves. They begin with words related to what is intellectually and
socially most meaningful to them (Anglin 1995). Also, early words tend to share
phonetic features. They are shorter and occur frequently in speech (Storkel 2004).
Children’s early productive vocabulary consists of words from a variety of
grammatical classes, while common nouns constitute nearly 40 % of the average
English—speaking child’s first 50 words (Bates et al. 1994). Among nouns, those
that are easiest to distinguish from the surroundings, such as animate beings or
things that move, are the earliest learned (Gentner 1999). This nouns bias has also
been found in several other languages such as Spanish, Dutch, French, Hebrew,
Italian, despite striking structural differences across languages (Bornstein and Cote
2004). However, some research shows less marked noun preference in Mandarin or
Korean (Tardif et al. 1999; Choi and Gopnik 1995). Several explanations about
noun bias have been proposed. One hypothesis is that children’s vocabulary reflects
the input directed to them. Studies have shown that labeling of objects dominates
the adult speech to English-speaking children (Goldfield 1993). An alternative
explanation is that nouns are linguistically simpler than verbs. In addition, the
concepts represented by nouns are clear and concrete (Gentner 1983, 1988). The
first proposal can account for differences across languages in acquisition, while the
second will predict universality in language acquisition. In my study, I will
investigate whether this noun bias exists in children acquiring Mandarin in Taiwan.
Even before children begin producing words themselves, adults’ labeling and
gazing behaviors serve to focus children’s attention on objects. As children acquire
language, adults’ speech to them incorporate increasingly rich information about
3.1 General Literature on Lexical Development 113
the categories they are acquiring. For example, parents explain that ‘Bats live in
caves.’ (Gelman et al. 1998).
The labels adults provide for children seem to be based on children’s own ways
of categorizing the world (Golinkoff et al. 1995), which corresponds to so-called
basic-level categories. Basic level is the most general level at which objects are
similar because of their forms, functions, component parts (Poulin-Dubois 1995), or
motions. Mother’s speech has also been shown to have an effect on the ways that
children come to understand the use vocabulary relating to their own inner states
(Beeghly et al. 1986; Tingley et al. 1994).
In addition to the special vocabulary directed to children, some characteristics of
child-directed speech may facilitate semantic development. Input language, espe-
cially when young children begin to understand and use words, is more clearly and
slowly enunciated and is characterized by exaggerated intonation and clear pauses
between utterances (Sachs et al. 1976). In addition, the words being focused on tend
to be placed in sentence-final position with especially marked pitch and stress
(Fernald and Mazzie 1991). We wonder whether Indonesian mothers in Taiwan
with limited Mandarin proficiency are able to provide this kind of input to facilitate
their children’s Mandarin acquisition or not.
Most of the input of children is from their primary caretakers, their mothers in our
study. Immigrant mothers are second language learners of Mandarin, whose
Mandarin is likely to be interfered by their first language like most second language
learners. Therefore, it is helpful to understand Indonesian vocabulary composition
when studying children of Indonesian mothers’ lexical development. A discussion
of Vietnamese is helpful for comparing with the previous study on children of
Vietnamese mothers (Kuo 2008).
As much as 60 % of Vietnamese vocabulary is derived from Chinese, especially
words relating to science and politics due to a thousand years of Chinese domi-
nation. Many compound words are Sino-Vietnamese, consisting of native
Vietnamese words combined with Chinese borrowings. A native Vietnamese word
can be distinguished from a Chinese borrowing by reduplication, which indicates
intensity. A native Vietnamese word can be reduplicated and the meaning does not
change when the tone is shifted. As a result of French colonization, Vietnamese also
has words borrowed from French. Recently many words are borrowed from
English, for example, TV (Alves 1999). Like Chinese, Vietnamese also has a
classifier system to indicate classes of nouns. Among the common categorization
bases are shape, animacy, material, and size. The classifier cái is used for most
inanimate objects, which can extend to all other classifiers.
On the other hand, the vast majority of Indonesian words, however, come from
the root lexical stock of its Austronesian (incl. Old Malay) heritage. Indonesian has
borrowed from many other languages, such as Sanskrit, Arabic, Persian,
114 3 Lexical Development
Portuguese, Dutch, and Chinese. There are about 750 Sanskrit loan words in
modern Indonesian, 1000 Arabic loans, some of Persian and Hebrew origin, so 125
words of Portuguese origin and around 10,000 loan words from Dutch. The latter
also comprises many words from other European languages, which came via Dutch,
the so-called International Vocabulary. The Chinese loan words are usually con-
cerned with cuisine, trade, or often just exclusively things Chinese. There is a
considerable Chinese presence in the whole of Southeast Asia (Sneddon 2004).
In the present study, we collected one-hour spontaneous speech samples from two-,
three-, four-, five-, six-year-old children of Indonesian mothers and Taiwanese
mothers and designed elicitation tasks for classifiers, which were lack in
3.4 Mandarin Lexical Development by Children of Indonesian Mothers 117
50
0
2 3 4 5 6
Age
The nouns in the spontaneous speech of Indonesian children were more than those
of Taiwanese children. On average, Indonesian children have acquired 74.44 nouns
and Taiwanese children have acquired 69.6 nouns. The nouns are divided into
semantic categories such as animals, commodities, food, body parts, fruit, vehicle,
toys, clothing, furniture, stationary, color, plant, personal names, kinship, and place
names. The largest portion of Indonesian children’s nouns consists of animals
(12.30 %) while that of Taiwanese children consists of food (9.75 %). The details
are shown in Table 3.1. The first column indicates the semantic category in bold
and an example below. The first row indicates the nationality of the participant’s
mother. The percentage shows the proposition of the semantic category among the
118 3 Lexical Development
children’s nouns. The last row gives the mean number of nouns produced by each
group.
The following lists errors in noun usages found in Indonesian children’s spon-
taneous speech. Most of their errors involve superordinate categories as in Example
1 and relational terms as in Example 2. Example 1–2 are from the eight-year-old
Indoneisan girl, who was confused about the superordinate categories animal and
plant and relational terms husband and wife.
1. *CHI: 還有 花 也 是(*) 動物
%pho: xài iŏu xuā iĕ s̩ ́ tɔ́ŋ ú
%eng: The flowers are animals, too.
*INV: 花…花 算 是(*)….植物 啦!
%pho: xuā xuā suán s̩ ́tʂ̀ ú la !
%eng: Flowers…are plants.
*CHI: 動物
%pho: tɔŋ́ ú
%eng: Animal.
*INV: 植物 啦! 動物 是(*) 會 動 的 啊! 對 不 對**會 講 話 的 啊
%pho: tʂú la ! tɔ́ŋú s̩ ́ uéi tɔ́ŋ tə tuéi pù tuéi xuéi tɕiăŋ xuá tə
%eng: They are plants. The animals can move, right? They can speak.
*CHI: 動物**不會**
%pho: tɔ́ŋ ú pù xuéi
%eng: No, animals cannot.
2. *CHI: 還有 紅太狼, 紅太狼 是(*) 灰太狼 的 老公[老婆]
%pho: xài iŏu xɔŋ̀ thái làŋ xɔŋ̀ thái làŋ s̩ ́ xuēi thái làŋ tə lău kɔ̄ ŋ
%eng: Another character is Wolnie (Red Wolf). Wolnie is Wolffy’s husband [wife].
Examples 3–5 are from the Indonesian 4-year-old boy. He was confused with
nouns sharing some features. In Example 3, he was confused between cucumbers
and eggplants, which are both long.
3. *CHI: 小[*]黃瓜
%pho: tɕ[*]iău xuaŋ̀ kuā
%eng: Cucumber.
*INV: 不是[*],這[*]是[*]茄子
%pho: pù ś̝[*] ts[*]ə́ ś̝[*] tɕʰiè ts̟
%eng: No. This is an eggplant.
In Example 4, he took light for fire because they are both bright.
4. *INV: 沒有啦,那不是[*]火啦
%pho: mèi iŏu la ná pù ś̝[*] xuŏ la
%eng: Ni, that is not fire.
3.4 Mandarin Lexical Development by Children of Indonesian Mothers 121
*CHI: 那個…電[*]燈[*]而[*]已喔?
%pho: ná kə teń [*] tən[*] ə[̀ *] ĭ o
%eng: That….is only a light.
Example 5 involves the overextension of the vegetable suffix cai to eggplant.
5. *INV: 那這[*]是[*]什[*]麼?
%pho: ná ts[*]ə́ ṣ[́ *] s[*]ə mə
%eng: What is this?
*CHI: 高麗[*]菜….嗯….
%pho: kāu léi[*] tshái ən
%eng: Cabbage….well…
*INV: 這[*]是[*]茄子阿
%pho: ts[*]ə́ ṣ́[*] tɕhiè tṣ a
%eng: This is an eggplant.
*CHI: 茄[*]子菜[*]
%pho: tɕhè[*] tṣ tshái
%eng: Eggplant.
In Example 6, he was confused between clothes and pants.
6. *INV: 好,那這[*]是[*]什[*]麼你知[*]道嗎?
%pho: xău ná ts[*]ə́ ṣ[́ *] s[*]ə mə n ĭ tṣ[*] táu ma
%eng: Ok, do you know what this is?
*CHI: 衣[*]服
%pho: í[*] fù
%eng: Clothing.
*INV: 這[*]是[*]褲子
%pho: ts[*]ə́ ṣ[́ *] khú tṣ
%eng: They are pants.
*CHI: 褲子…阿這[*]個…
%pho: khú tṣ a ts[*]ə́ kə
%eng: Pants…this…
On the other hand, most of the errors made by children of Taiwanese parents are
due to Taiwanese transfer and overextension. Example 7 was from the
three-year-old boy, who transferred the word ‘money’ from Taiwanese. The word /
kho/ usually follows a numeral in Taiwanese, but the child reduplicated it.
Examples 8–10 are from the six-year-old Taiwanese boy. Example 8 is also caused
by Taiwanese transfer.
7. 摳摳
%pho: kho kho
%eng: money
122 3 Lexical Development
8. Child: 擦子[橡皮擦]
%pho: tshā ts1[ɕiáŋ phі̀ tshā]
%eng: eraser
Examples 9 and 10 were due to overextension. In Example 9, the boy coined the
terms xiaowangguo ‘small kingdom’ for the kingdom with dwarfs. In Example 10,
he overextended the rain umbrella for the sun umbrella.
9. Child: 是小王國
%pho: sí ɕiaǔ uàŋ kuò
%eng: It is a small kingdom.
Mother: 小人國,不是小王國
%pho: ɕiaǔ zəǹ kuò, pù sí ɕiaǔ uàŋ kuò
%eng: It is a country with dwarfs, not a small kingdom.
10. Child: 這張媽怎麼沒有撐雨傘,熱死了
%pho: tsə́ tsāŋ mā tsə̌ n mo meі̀ tshəŋ y̌ sǎ n, z̨ə́ s1̌ lə
%eng: Why didn’t you use an umbrella? It was hot.
Mother: 因為要照相所以我沒有撐雨傘
%pho: і̄ n ueі́ iaú tsaú ɕiaŋ́ suǒ і̌ uǒ meі̀ iǒ u tshəŋ y̌ sǎ n
%eng: Because I will take the picture, I don’t use the umbrella.
Mother: 這叫陽傘,那個,撐太陽,擋太陽的叫陽傘
%pho: tsə́ tɕiaú iàŋ sǎ n, ná kə tshəŋ ́ thaі́ iàŋ, tǎ ŋ thaі́ iàŋ tə ɕiaú iàŋ sǎ n
%eng: This is called a sun umbrella. It is used to block the sunshine.
The verbs acquired by Indonesian children (mean = 48.44) were fewer than those
by Taiwanese children (mean = 61.6). The verbs are divided into weather verbs,
stative verbs, change of state verbs, body function verbs, motion verbs, position
verbs, action verbs, cognition verbs, perception verbs, emotion verbs, communi-
cation verbs, and social interaction verbs. Both groups produced mostly action
verbs, which constitute 47.24 % of the verbs by Indonesian children and 43.82 %
of Taiwanese children. The details are shown in Table 3.2. The first column indi-
cates the semantic category and an example below. The first row indicates the
nationality of the participant’s mother and the number of participants. The per-
centage shows the proposition of the semantic category among the children’s verbs.
The last row gives the mean number of verbs produced by each group.
Most of the errors of verb use were phonological errors, which have been
discussed in the last chapter, so will not be discussed again here. An example from
the Taiwanese 2-year-olds was given in Example 11. Most of the semantic errors
were due to Taiwanese interference. For example, the Taiwanese verb 惜惜 was
used by both the three- and six-year-olds.
3.4 Mandarin Lexical Development by Children of Indonesian Mothers 123
Indonesian children (mean = 18.22) also acquired fewer adjectives than Taiwanese
children (mean = 22.70). According to Blackwell (2005), we divided the adjectives
into color, evaluative, age, human propensity, physical property, and quantity.
Indonesian children acquired most evaluative adjectives (31.21 %) and Taiwanese
children acquired most physical properties (32.83 %). Table 3.3 presents a summary
of adjectives produced by the two groups. The first row indicates the nationality of
124 3 Lexical Development
the participant’s mother and the number of participants. The percentage shows the
proposition of the semantic category among the children’s adjectives. The last row
gives the mean number of adjectives produced by each group.
Few classifiers were produced in the spontaneous speech. The first classifier use
appeared at age two for both groups. The two-year-old Indonesian girl produced only
one classifier—ge, while the two-year-old Taiwanese girl produced four classifiers—
ge, zhang, shao, and ben. Indonesian children on average produced 2 classifiers and
Taiwanese children produced 2.7 classifiers. The four-year-old Taiwanese boy pro-
duced most types of classifiers—zhang, ge, tiao, ben zhong, and ba. The
three-year-old Indonesian boy produced most tokens, but 31 out of the 34 were ge.
A list of classifiers occurred in the speech sample was given in Table 3.4. The
first column lists the classifiers. The first row lists the mothers’ nationality. The
numbers indicate the mean tokens of classifiers observed. The last row gives the
mean number of classifiers produced by each group.
Since classifiers did not occur a lot in spontaneous speech, we conducted a classifier
elicitation task (Kuo 1999) to elicit classifier production of the same group of children
in the second year. Therefore, the participants were one year older than in the spon-
taneous speech. The age of the participants was three, four, five, six and seven.
3.4 Mandarin Lexical Development by Children of Indonesian Mothers 125
Twelve pictures were specially designed to elicit the production of shape clas-
sifiers in Mandarin Chinese with two or more of the target objects present in the
pictures. The target objects were the referents of the collocating nouns of the target
classifiers. The target objects appeared with things that they often occur with in a
natural context. The presence of two or more target objects creates an obligatory
context for the use of classifiers (Kuo 1999).
Eight categories are identified as relevant cognitive categories in Chinese clas-
sifiers—material, shape, consistency, size, location, arrangement, quanta, and
attributes of parts (Tai 1994). The first seven categories were from Allan (1977)
based on the use of classifiers in more than fifty languages. The material category
has three subcategories: animacy, inanimacy, and abstract and verbal nouns. Shape
is divided into saliently one-dimensional, saliently two-dimensional, and saliently
three-dimensional. Consistency consists of three subtypes: flexible, hard or rigid,
and non-discrete. Size means big or small. Location refers to the place where the
object exists. Arrangement refers to the relation among members that a collective
noun denotes. Quanta refers to the function of quantification. Attributes of parts are
proposed as a basis for human categorization by Tversky and Hemenway (1984).
Shape classifiers were chosen because they are semantically more productive in a
sense that they can classify a larger number of objects on semantic bases (Loke
1996). In addition, shape is determined by visual perception. Also, shape is easier to
manipulate than other classifications.
The participants were tested individually at their homes. They were asked to
describe pictures provided by the researcher including naming and counting the
items in the pictures. After the instructions of the task, the experimenter demon-
strated by orally describing a picture. To make sure that the participant understood
the procedure, one picture was given as a warm-up. The participant was given
immediate feedback. Assistance without providing the classifiers was allowed. The
total procedures took about 15 min for each child. The process was tape-recorded
and transcribed.
126 3 Lexical Development
The results show that the children of Taiwanese mothers did a bit better than
children of Indonesian mothers. The mean accuracy rate for the Indonesian group is
25.43 %, and the mean accuracy rate for the Taiwanese group is 32.17 %. Their scores
for each classifier are given in Table 3.5. 1D stands for saliently one-dimensional, 2D
for saliently two-dimensional, and 3D for saliently three-dimensional.
Both groups did best for ge, with Indonesian even higher than Taiwanese.
Indonesian children used ge almost across the board, which may lead to the high
accuracy rate for ge. Here is an example of classifier errors. Example 12 was
produced by the four-year-old Indonesian boy, who used jiao ‘feet’ for pants and
then changed to ge failing to utter the conventional tiao.
(12). *INV: 有多少[*]長[*]褲呢?
%pho: iŏu tuō s[*]ău tsh[*]àŋ khú nə
%eng: How many trousers are there?
*CHI: 兩個腳…腳
%pho: liăŋ kə tɕiău tɕiău
%eng: Two…feet..feet.
*INV: 這[*]是[*]腳沒錯…可是[*]這[*]是[*]長[*]褲阿,對不對?有多少[*]
長[*]褲?
%pho: ts[*]ə́ ṣ́[*] tɕiău mèi tshuó khə̆ ṣ́[*] ts[*]ə́ ṣ́[*] tsh[*]àŋ khú a tuéi pù
tuéi iŏu tuō s[*]ău tsh[*]àŋ khú
%eng: Feet….these are trousers, right? How many trousers are there?
*CHI: 一個長[*]褲
%pho: i kə tsh[*]aŋ̀ khú
%eng: one ge trousers
A classifier comprehension task (Sera et al. 2013) was also given in the second
year. Four classifiers tiao 條, ge 個, zhi 枝, and zhi 隻 were chosen for the classifier
3.4 Mandarin Lexical Development by Children of Indonesian Mothers 127
comprehension task because they are acquired early and have high agreement
among adults (Batteen et al. 2007). The task consisted of 18 unique trials, 3 shape
and 3 kind trials, for each classifier (ge, zhi, and tiao). All shape trials consisted of
three line drawings of non-living things (e.g., an apple, a crayon, and a scarf). All
kind trials consisted of three drawings that depicted living things (e.g., a girl, a bird,
and a snake). Structuring the trials in this fashion avoided potential problems
associated with the homophones of zhi. A total of nine shape and nine kind trials
were presented to each participant, with each different classifier being queried three
times. So, for example, a child would be presented with drawings depicting a ball, a
pencil, and a belt three times. One time the child would be asked to ‘Give me one
ge,’ the other time they would be asked to ‘Give me one zhi,’ and another time they
would be asked to ‘Give me one tiao.’ The pictures depicting each item were
slightly different for each query.
The Indonesian children on average answered 63.25 % correctly, which is lower
than the mean of children with Taiwanese parents (75.08 %). Taiwanese
3-year-olds’ responses were only reliably better than chance on the shape trials of
zhi (t(15) = 4.4, p < 0.001) and the kind trials of tiao (t(15) = 2.5, p < 0.05). The
responses of all other age groups were well-above chance on all trial types. (Sera
et al. 2013). On the other hand, Indonesian children did not perform above chance
till seven years old. There were also large individual differences. The five-year-old
girl did better than the six-year-old girl. The results of individual Indonesian par-
ticipant’s performance in classifier comprehension task are given in Table 3.6. The
first column indicates the target classifiers. The top row lists the participants’ age
and gender. In order to trace the data from the first year, +1 was used to indicate the
second-year data. For example, 2 + 1♀ means the three-year-old girl.
Children’s acquisition of classifiers is guided by both cognitive principles and
frequency. Shape classifiers were first acquired by children and then were followed
by classifiers of high frequency such as jian for clothing and nouns that are
semantically vague such as thing. The first cognitive basis they acquired is shape. In
addition to the overextension of the general classifier ge, children’s classifier errors
did not violate shape principles. They used a classifier for nouns denoting objects of
the same shape but different material.
References
Akhtar, N. (2002). Relevance and early word learning. Journal of Child Language, 29, 677–686.
Allan, K. (1977). Classifiers. Language, 53, 285–311.
Alves, M. (1999). What’s so Chinese about Vietnamese? The Ninth Annual Meeting of the
Southeast Asian Linguistics Society. University of California, Berkeley.
Anglin, J. (1995). Classifying the world through language: Functional relevance, cultural
significance, and category name learning. International Journal of Intercultural Relations, 19,
161–181.
Bates, E., Marchman, V., Thal, D., Fenson, L., Dale, P., Reznick, S., et al. (1994). Developmental
and stylistic variation in the composition of early vocabulary. Journal of Child Language, 21,
85–123.
Batteen, C., Davis, N., Fuller, J., Kuo, J. Y.-C., Li, D., Sera, M., et al. (2007). Classifier use by
adult Mandarin speakers. The 5th Conference of the European Association of Chinese
Linguistics. Leipzig, Germany. September 4–7.
Beeghly, M., Bretherton, I., & Mervis, C. (1986). Mothers’ internal state language to toddlers: The
socialization of psychological understanding. British Journal of Developmental Psychology, 4,
247–260.
Bornstein, M., & Cote, L. (2004). Cross-linguistic analysis of vocabulary in young children:
Spanish, Dutch, French, Hebrew, Italian, Korean, and American English. Child Development,
75, 1115–1139.
References 129
Chien, Y.-C., Lust, B., & Chiang, C.-P. (2003). Chinese children’s comprehension of
count-classifiers and mass-classifiers. Journal of East Asian Linguistics, 12, 91–120.
Choi, S., & Gopnik, A. (1995). Early acquisition of verbs in Korean: A cross-linguistic study.
Journal of child language, 22, 497–529
Clark, E. (1987). The principle of contrast: A constraint on language acquisition. In B.
MacWhinney (Ed.), Mechanisms of language acquisition. Hillsdale, NJ: Erlbaum.
Clark, E. (1993). The lexicon in acquisition. Cambridge, UK: Cambridge University Press.
Erbaugh, M. S. (1986). Taking stock: The development of Chinese noun classifiers historically and
in young children. In C. Craig (Ed.), Noun classes and categorization (pp. 399–436).
Amsterdam: John Benjamins Publishing Company.
Fang, F. (1985). An experiment on the use of classifiers by 4-to-6-Year-Olds. Acta Psychologica
Sinica, 17, 384–392.
Fernald, A., & Mazzie, C. (1991). Prosody and focus in speech to infants and adults.
Developmental Psychology, 27, 209–221.
Gelman, S., Coley, J., Rosengran, K., Hartman, E., & Pappas, A. (1998). Beyond labeling: The
role of maternal input in the acquisition of richly structured categories. Monographs of the
Society for Research in Child Development, 63(253), 1.
Gentner, D. (1981). Some interesting differences between verbs and nouns. Cognition and Brain
Theory, 4(2), 161–178.
Gentner, D. (1983). Nouns and verbs. Symposium Presented at the Meeting of the New England
Child Language Association. Medford, MA: Tufts University.
Gentner, D. (1988). Cognitive determinism: Object reference and relational reference. Paper
presented at the Boston University Child Language Conference, Boston, MA.
Gentner, D. (1999). Individuality and early word meaning. Paper presented at the VIIIth
International Congress for the Study of Child Language, San Sebastian, Spain.
Gleason, J. B., & Ratner, N. B. (2012). The development of Language (8th ed.). Cambridge:
Pearson.
Goldfield, B. (1993). Noun bias in maternal speech to one-year-olds. Journal of Child Language,
20, 85–99.
Golinkoff, R., Shuff-Bailey, M., Olguin, R., & Ruan, W. (1995). Young children extend novel
words at the basic level: Evidence for the principle of categorical scope. Developmental
Psychology, 31, 494–507.
Hall, D. (1994). How mothers teach basic-level and situation-restricted count nouns. Journal of
Child Language, 21, 391–414.
Houston-Price, C., Plunkett, K., & Harris, P. (2005). “Word learning wizardry” at 1;6. Journal of
Child Language, 32, 175–189.
Hsu, J. (1996). A study of the stages of development and acquisition of Mandarin Chinese by
children in Taiwan. Taipei: Crane.
Hu, Q. (1993). The acquisition of Chinese classifiers by young Mandarin-speaking children. Ph.D.
Thesis, Boston University.
Huang, Y. V. (2007). The use of count and mass classifiers in Chinese preschoolers. MA thesis,
National Taiwan Normal University, Taipei.
Jiang, Y.-W. (2000). Basic-level effects in Chinese lexicon. MA thesis, National Tsing Hua
University, Hsinchu.
Kuo, J. Y.-C. (1999). Strategies for learning classifiers. In Proceedings of the 1999 Second
Language Research Forum, pp. 424–442.
Kuo, J. Y.-C. (2008). Mandarin acquisition by children of Vietnamese mothers in Taiwan. Taipei:
Crane.
Kuo, J. Y.-C. (2015). Lexical development by children of Thai immigrants in Taiwan. Studies in
English Language and Literature, 35, 1–12.
Loke, K. K. (1996). Norms and realities of Mandarin shape classifiers. Journal of Chinese
Language Teachers Association, 31(2), 1–22.
130 3 Lexical Development
Markman, E. (1987). How children constrain the possible meanings of words. In U. Neisser (Ed.),
Concepts and conceptual development: Ecological and intellectual factors in categorization.
Cambridge, UK: Cambridge University Press.
Markman, E., & Wachtel, G. (1988). Children’s use of mutual exclusivity to constrain the
meanings of words. Cognitive Psychology, 20, 121–157.
Myers, J. (2000). Rules vs. analogy in Mandarin classifier selection. Language and Linguistics, 1
(2), 187–209.
Poulin-Dubois, D. (1995). Object parts and the acquisition of the meaning of names. In K. Nelson
& Z. Reger (Eds.), Children’s language (Vol. 8). Hillsdale, NJ: Erlbaum.
Sachs, J., Brown, R., & Salerno, R. (1976). Adults’ speech to children. In W. won Raffler Engel &
Y. Lebrun (Eds.), Baby talk and infant speech. Lisse, The Netherlands: Swets and Zeitlinger.
Sera, M. D., Johnson, K., & Kuo, J. Y.-C. (2013). Classifiers augment and maintain shape-based
categorization in Mandarin speakers. Language and Cognition, 5(1), 1–23.
Sneddon, J. N. (2004). The Indonesian Language: Its history and role in modern society. Sydney:
UNSW Press.
Storkel, H. (2004). Do children acquire dense neighborhoods? An investigation of similarity
neighborhoods in lexical acquisition. Applied Psycholinguistics, 25, 201–222.
Tai, J. (1994). Chinese classifier systems and human categorization. In M. Chen & T. Ovid (Eds.),
In honor of professor William S-Y. Wang: Interdisciplinary studies on language and language
change (pp. 479–494). Taipei: Pyramid.
Tardif, T. (1996). Nouns are not always learned before verbs: Evidence from Mandarin speaker’s
early vocabulary. Developmental Psychology, 32, 492–504.
Tardif, T., Gelman, S., & Xu, F. (1999). Putting the “noun bias” in context: A comparison of
English and Mandarin. Child Development, 70, 620–635.
Tingley, E., Gleason, J., & Hooshyar, N. (1994). Mothers’ lexicon of internal state words in speech
to children at mealtime. Journal of Communication Disorders, 27, 135–155.
Tomasello, M. (1995). Pragmatic contexts for early verb learning. In M. Tomasello & W.
Merriman (Eds.), Beyond names for things: Young children’s acquisition of verbs. Hillsdale,
NJ: Erlbaum.
Tsay, S., Ke, H., Tsai, J., Lin, H., & Kuo, Y. 蔡素娟、柯華葳、蔡志賢、林惠玲、郭怡君.
(2005). 新移民女性及其子女的發展與適應研究-新移民女性之子女的語言發展研究 I
“Xinyimin nuxing ji qi zinu de fazhan yu shiying yanjiu—xinyimin nuxing zhi zinu de
yuyan fazhan yanjiu” [Language development by children of new female immigrants]. NSC
report. (NSC 93-2420-H-194-003-KD)
Tversky, B., & Hemenway, K. (1984). Objects, parts and categories. Journal of Experimental
Psychology, 113, 169–193.
Chapter 4
Syntactic Development
Three major theories have been proposed for syntactic development: semantic
bootstrapping, syntactic bootstrapping (Valian 1986, 1991; Radford 1990), and
distributional learning (Maratsos 1988). The first two focus on innate mechanism.
Semantic bootstrapping matches syntactic structures to innate ontological cate-
gories. Syntactic bootstrapping assumes the innate universal grammar. These two
do not predict differences between children of native parents and immigrant parents.
However, distributional learning predicts delay in the development with limited
input in children of immigrants. Many problems have been found in semantic
bootstrapping (Mills 1986; Karmiloff-Smith 1979; Pérez-Pereira 1991; Levy 1988;
Bowerman 1990). On the other hand, distributional learning is supported in several
studies (Elman 1990; Finch and Chater 1992; Plunkett and Marchman 1991). The
goal of this chapter is to examine whether there is a difference in the syntactic
development between children of Taiwanese and Indonesian mothers. It will have
implication for language acquisition theories.
In the present study, syntactic development was measured by length (mean
length of utterance), sentence structures (Index of Productive Syntax), and a
question elicitation task. Since these measurements were first developed with
English-speaking children, I will first review this literature followed by literature on
Mandarin syntactic development.
After Stage II, the MLU goes from 2.5 to 3.0 when the different types of sentences
such as negation, questions, and passive sentences are developed gradually by
young children since they attempt to acquire the sentence structures from what they
hear in the daily life in order to express themselves. According to Bassano and
Mendes-Maillochon (1994), children developed their different sentence types based
on different intonation patterns, which closely matched those used by adults to mark
different sentence modalities.
4.1.2.1 Negation
4.1.2.2 Questions
There are three ways of forming questions in English, rising intonation with a
statement, yes/no questions, and wh-questions. Children at earlier stage seem to use
rising intonation to express questions (Klima and Bellugi 1966). Yes/no question
appears much later about in Stage III, because speakers need to invert the subject
and the auxiliary verb. Wh-question is used for obtaining more important infor-
mation instead of a yes or no answer, so speakers need to place wh-words such as
what, who, where, and when in the initial position to form a question in addition to
subject auxiliary inversion, for example, What should we do now? or When does the
show start? In the first period, children usually omit the auxiliary when asking wh-
questions, e.g., What that? or Where Mommy go? Next, they start to insert the
auxiliary into the structure of questions but do not reach the full mastery of
expression of questions since they sometimes do not invert the subject and the
auxiliary like Where he is going? The wh-questions can become more complicated
when they are formed from complex sentences which may undergo long-distance
movement across sentences. However, children are able to incorporate all of the
syntactic rules necessary to produce well-formed wh-questions with highly complex
rules (De Villier et al. 1990).
Klima and Bellugi (1966) claimed that children can produce correct forms of
yes/no questions earlier than produce wh-questions since there is only one rule in
yes/no question in which the subject of the sentence and auxiliary verb should be
switched and wh-questions involve two rules to complete the sentence patterns: the
placement of wh-words and inversion of subject and auxiliary.
134 4 Syntactic Development
Wootten et al. (1979) explored the order of acquisition of the various wh-
questions by children and found that the sentence patterns beginning with wh-words
such as what, where, and who were acquired first, but questions about when, how,
and why were used later. Also, the same order was found in the comprehension
study conducted by Winzemer (1980) since the concepts of how, when, and why are
about manner, time, and causality, which are more abstract.
During the preschool years, Stage IV (MLU = 3.0–3.5), more complex gram-
matical structures begin to be produced by preschoolers such as passive con-
struction, coordinations, and relative clauses; however, their acquisition of certain
sentence structures does not reach the full mastery until some years later.
4.1.2.3 Passive
The structure of passive is for focusing on the recipient of an action, for example,
‘John was hit by a ball.’ which is a full passive sentence, which focuses on the
person, John. The other passive form without an agent specified in the sentence is
called truncated passive, e.g., ‘The vase is broken.’ Not surprisingly, passives
rarely occur in children’s spontaneous speech. Thus, if researchers would like to
investigate the development of structure of passive, they have to develop a task to
elicit these passive patterns in an experiment. In addition, because the order of the
agent and the object is reversed in passives in English, this particular construction
can reveal a great deal about how children acquire word order rules that play a
major role in English syntax.
Horgan (1978) elicited English-speaking children’s production of passives and
discovered that the truncated passives were produced more than full passives by
children in the earlier period. In addition, she found that the subjects in full passives
produced by young children are animate, while those in truncated passives are
inanimate subjects. Thus, Horgan concluded that full and truncated passives were
acquired separately by children.
Bever (1970) proposed that young children around three or four generalize the
order of words in English passives based on word order strategy in which English
use noun–verb–noun sequences in the active voice meaning agent–action–object.
Thus, when they hear a passive sentence, they ignore the copula and the word, by
and infer the meaning of the passive noun–verb–noun sequence to be active.
However, other studies which explored children learning languages other than
English showed that this strategy is not universal. It was indicated that passive
construction was acquired much earlier by non-Indo-European language-speaking
children than English-speaking children. Children in southern Africa used passives
in daily conversation by age two since they mostly marked the topic of a sentence in
the conversation (Pye 1988; Suzman 1987).
4.1 General Literature on Syntactic Development 135
4.1.2.4 Coordination
During Stage IV, young children attempt to combine two or more sentences to form
the expression of complex or compound proposition by simply using and. There are
two main forms of coordination: sentential coordinations, in which two or more
whole sentences are conjoined by means of conjunctions, e.g., ‘Mary is singing and
Mary is dancing,’ and phrasal coordinations, in which only phrases in the sentences
are combined, e.g., ‘Mary is singing and dancing.’ Thus, some researchers were
interested in the order of these two coordinations in the development of language of
children. Bloom et al. (1980) studied four children longitudinally and found that for
three of them, both forms appeared in their speech in the same period, but one child
produced phrasal coordination before sentential coordination. In addition, they
discovered that the development of coordination was influenced by semantic fac-
tors. They found that their subjects used conjunction and to express a variety of
meanings, and these meanings were developed in a fixed order. First, they devel-
oped the meaning of additive, which showed no dependency relation between the
conjoined clauses, e.g., ‘Mary is singing and Carrie is dancing.’ Few months later,
they began developing coordination which showed temporal relations, e.g., ‘She
goes home and takes a shower.’ After that, they used and to encode causal relations,
e.g., ‘Go straight on and you’ll see the church.’ Therefore, we can see that coor-
dinations encode a variety of meanings and children use conjunction word, and, in a
semantically limited way and develop in a fixed order.
In Stage IV, embedded sentences such as relative clauses begin to appear in young
children’s speech, but the full mastery of these constructions is not reached until
Stage V. In Bloom’s (1980) study, after development of coordination, relativization
began to occur in young children’s speech, but scarcely. Therefore, Hambruger and
Crain (1982) and Tager-Flusberg (1982) designed an elicitation task for relative
clauses and found that object relatives were produced first by children at age four.
For example, ‘John met a girl who wears a pink dress.’ The grammatical structure
which adds a clause at the end of a sentence is much easier than that which is
interrupted by a clause in the middle of a sentence.
Hsu (1996) proposed the stages of the syntactical development in Mandarin similar
to the one in English proposed by Brown (1973). In Stage I, young children around
1;0–1;6 produced one-word utterances that were usually composed of nouns, for
example, 車車 che che ‘a car,’ and since young children in this stage did not have
enough syntactic knowledge and vocabulary, they began to use the same word with
different intonation contours to convey their expression. In addition, Hsu general-
ized that the participants in his study tried to select the word conveying the
important information, e.g., 帽 mao for maozi ‘a hat.’ In Stage II, simple sentences
began to appear in children’s speech, e.g., 超人給你 chaoren gei ni ‘superman give
you’ ‘(I) give you the superman.’ During Stage III, Hsu found that children at age
two started to produce embedded sentences in the speech, e.g., 這是小阿姨綁的
zhe si xiao a yi bang de ‘little aunt tie de’ ‘This is tied by my little aunt,’ in which
de is a relativization marker. After two years at Stage IV, they began to produce
4.2 Mandarin Syntactic Development by Native Children 137
longer sentences. Hsu divided compound sentences used in Stage IV into two kinds.
First, children around 2;7–4;2 produced compound sentences without any con-
junctions, e.g., 你給我,我就給你 ni gei wo, wo jiu gei ni ‘(If) you give me (one
thing), I’ll give you (another thing).’ The second kind of compound sentences with
conjunctions appeared after age four, e.g., 如果你認識我弟弟,就行了 Ru kuo ni
ren shi wo di di, jiu xing le ‘If you know my younger brother, that’s fine.’
Comparing with the development of English-speaking children, we can see that the
syntactic development of Chinese-speaking children is faster since Chinese does
not have grammatical morphemes like English. Therefore, after Stage I, Chinese
children directly developed simple sentences instead of grammatical morphemes
like English-speaking children. In addition, negations and questions do not involve
any inversion of subjects and auxiliaries in Mandarin, so children around two
attempt to produce embedded clauses; however, for English-speaking children,
embedded sentences appeared at age three. Thus, we can see that Chinese children
develop syntactic constructions much earlier than English children do. Since
embedded sentences appear earlier in children’s speech, some researchers continued
to explore the mastery of complex sentences, which is an important developmental
milestone in language acquisition (Chang 1984; Cheung 2006).
Chang (1984) explored the comprehension of different types of relative clauses
by children in grade 1, 2, 4, and 6. There are four types of relative sentences
depending on the grammatical function of the head noun in the main sentence and
relative clause: SS, SO, OS, and OO. S stands for subject and O for object. She
found that SS and SO were much easier to be produced by children than OS and
OO since in Chinese, SS and SO contain no interruption in the main clause, while
the main clauses of OS and OO are interrupted.
Cheung (2006) studied longitudinally the order of development of different types
of complex sentences by four children around two to four. In his study, relative
clauses, serial verb constructions, descriptive clauses, pivotal constructions, and
subject/object complement sentences were investigated. He found that generally
serial verb constructions and descriptive clause appeared earliest among other types
of complex constructions by children at age two. Several months later, children
began to produce pivotal constructions and relative clauses. After the development
of relative clauses, subject/object complements were developed much later and
object complement appeared in children speech earlier than subject complements.
children of Vietnamese and Taiwanese mothers in the MLU for age 2–5. The
difference showed up at the age of 6. The MLU for the six-year-old of Taiwanese
mother reached 6.37, while the Vietnamese age-mate remained 3.62 as shown in
Table 4.1.
In addition, Scarborough’s (1989) Index of Productive Syntax was adapted with
particular structures in Mandarin to measure children’s syntactic growth. It was
found that the sentence patterns used by children of Vietnamese mothers were not
fewer than those used by children of Taiwanese mothers as shown in Table 4.2.
However, different error patterns revealed. Both groups of children omitted or
misused function words. The younger children tended to omit function words. The
older children tended to misuse function words. Errors in complement selection
were found in the two-year-old of the Taiwanese mother, who used the intransitive
verb buxing ‘cannot’ as a transitive verb. The word order errors were observed only
in children of Vietnamese mothers as in Example 1 from a six-year-old child of a
Vietnamese mother. The letter stands for the mother’s nationality, and the number
stands for the child’s age. For example, V6 stands for a six-year-old of a
Vietnamese mother.
(1) V6: *媽媽 買 電池 給 我 明天
%pho: mama mai tiàn t́ʂh gě i uǒ miə́ŋ tian
%Pinyin: mama mai dianchi gei wo mingtian
%eng gloss: mother buy battery give me tomorrow
%eng: Mother will buy me batteries tomorrow.
We collected one-hour spontaneous speech samples from two, three, four, five, and
six-year-old children of Indonesian mothers and Taiwanese mothers. The partici-
pants were the same as in the study of phonological development in Chap. 2 and
lexical development in Chap. 3. We visited the participants at their homes and
recorded their conversions, transcribed with IPA, and then analyzed the syntax by
length, sentence structures, and syntactic errors. A question elicitation task was
given in the third year for questions are rare in spontaneous speech.
The mean length of utterance was used to measure the length of utterances. The
Indonesian group seemed to delay in the syntactic development. The mean MLU
(M = 2.95) was smaller than the Taiwanese group (M = 3.61), and such difference
showed up as early as age 2 as shown in Table 4.3.
The sentence types were measured by Mandarin Index of Productive Syntax (Lin
2009), which adapted Scarborough’s (1989) Index of Productive Syntax with
particular structures in Mandarin. The ba construction, verb reduplication, aspect
markers (i.e., le, zheng, zhe), and sentence final particles observed in children’s
speech were added. It was found that the sentence patterns used by the Indonesian
group (M = 34.70) were fewer than those used by the Taiwanese group
(M = 39.10) as shown in Table 4.4.
Table 4.4 Mandarin Index of Productive Syntax by children of Indonesian and Taiwanese
Age (years) Indonesian Taiwanese
Gender Male Female Male Female
2 22 23 43 18
3 45 42 28 44
4 19 42 38 42
5 38 40 47 39
6 32 44 46 46
Mean 31.20 38.20 40.40 37.80
Mean 34.70 39.10
1
With kind permission from Crane Publishing: Mandarin Acquisition by Children of Vietnamese
Mothers in Taiwan (2008), pp. 99, 103, by Jenny Yichun Kuo, Examples (2, 6, 15).
142 4 Syntactic Development
(10) INV: 你 功課 寫完 了 沒?
%pho: nǐ kƱ̄ ŋ khə̀ ɕiě uań lə meí
%Pinyin: ni gongke xie wan le mei
%eng gloss: You class write-Asp Asp not
%eng: Do you have done your homework?
I5: *去 學校 寫 [在 學校 寫完 了]
%pho: tɕʰỳ ɕyé ɕiaù ɕiě [tsai ɕyé ɕiaù ɕiě uań lə]
%Pinyin: qu xuexiao xie zai xuexiao xiewan le
%eng gloss: *go school write [at school write-Asp Asp]
%eng: I have finished it in school.
Indonesian children also misused the ba construction as in Example (11).
(11) I6: *把 我 割 [割到 我]
%pho: pǎ uɔ̌ kə̄ [kə̄ taù uɔ̌ ]
%Pinyin: ba wo ge [gedao wo]
%eng gloss: *ba I cut [cut me]
%eng: I got a cut.
Although Taiwanese children omitted function words, they never omitted con-
tent words. Indonesian children omitted content words as in Example (12).
A five-year-old Indonesian child omitted you ‘have.’
(12) I5: 我 [有] 很多 鉛筆。
%pho: uɔ̌ iǒ u xə̌ n tuɔ̄ tɕʰiɛ̄ n pǐ
%Pinyin: wo you henduo qianbi
%eng gloss: I [have] many pencil
%eng: I have a lot of penciles.
4.4.2.1 Participants
The participants were the same with those whose natural speech sample were
collected. There were nine children of Indonesian mothers and ten children of
Taiwanese mothers. It was two years after the collection of natural speech sample,
so the participants were two years older. The age of the children was four, five, six,
seven, and eight.
4.4.2.2 Materials
The same set of four colored pictures with minor differences used in Kuo’s (2008)
study was used. There were four or five animals holding different objects in each
picture.
4.4.2.3 Procedures
The participants were tested individually at their homes. The children were given
four pictures, and their mother had of one of them. The children were required to
ask questions to find out which of the four pictures their mothers was holding.
When the child found out, the mother took another picture. The game continued till
the mother had all the pictures. The total procedures took about 15 min for each
child. The process was tape recorded and transcribed.
The mean accuracy rate of questions by children with Indonesian mother is a little
bit lower than their age-mates with Taiwanese mothers. Accuracy is defined by the
tokens of correct use out of the total use of questions. The mean of Indonesian
children is 77.39 %, while that of Taiwanese children is 79.73 %. The error types
will be discussed later. The accuracy rate of individual children’s questions is listed
in Table 4.5. NA stands for not available. The child refused to participate in the
task.
It is unusual the older children did not ask better questions. The older children
were able to reach the communicative goal with fragments, that is, the limitation of
the task. Taiwanese children produced both A-not-A and particle questions as in
(12–14). They used question particles most of the time.
(12) T4: 小豬 頭上 有沒有 蘋果?
%pho: ɕiaǔ tʂū thouʂ̀ aŋ́ iǒ u meì iǒ u phiŋk̀ uǒ
%Pinyin: xiaozhu toushang youmeiyou pinguo
%eng gloss: Little pig head above A-not-A apple
%eng: Does the pig have an apple?
144 4 Syntactic Development
References
Bassano, D., & Mendes-Maillochon, I. (1994). Early grammatical and prosodic marking of
utterance modality in French: A longitudinal case study. Journal of Child Language, 21, 649–
675.
Bellugi, U. (1967). The acquisition of negation (Unpublished doctoral dissertation). Cambridge,
MA: Harvard University.
Bever, T. G. (1970). The cognitive basis for linguistics structure. In J. R. Hayes (Ed.), Cognition
and the development of language. New York: Wiley.
Bloom, L., Lahey, J., Hood, Lifter, K., & Fiess, K. (1980). Complex sentences: Acquisition of
syntactic connectives and the semantic relations they encode. Journal of Child Language, 7,
235–261.
Bowerman, M. (1990). Mapping thematic roles onto syntactic functions: Are children helped by
innate linking rules? Linguistics 28, 1253–1289.
Brown, R. (1973). A first language: The early stages. Cambridge.
Brown, R., & Fraser, C. (1963). The acquisition of syntax. In C. N. Cofer & B. Musgrave (Eds.),
Verbal behavior and learning: Problems and processes. New York: McGraw-Hill.
Chang, H. (1984). The comprehension of complex Chinese sentences by children: Relative
clauses. Chinese Journal of Psychology, 26, 57–66.
Cheung, H. (2006). The uses of complex sentences in Mandarin-speaking children. NSC report.
148 4 Syntactic Development
de Villiers, P. A., & de Villiers, J. G. (1979). Early language. Massachusetts: Harvard University
Press.
de Villiers, J. G., Reoper, T., & Vainikka, A. (1990). The acquisition of long distance rules. In
L. Frazier & J. G. de Villiers (Eds.), Language processing and acquisition. Dordrecht, the
Netherlands: Kluwer Academic.
Elman, J. L. (1990). Finding structure in time. Cognitive Science, 14, 179–211.
Finch, S., & Chater, N. (1992). Bootstrapping syntactic categories. In Proceedings of the 14th
Annual Conference of the Cognitive Science Society.
Gleason, J. B., & Ratner, N. B. (2012). The development of language (8th ed.). Cambridge:
Pearson.
Gleitman, L. R., & Wanner, E. (1982). Language acquisition: The state of the art. In E. Wanner &
L. R. Gleitman (Eds.), Language acquisition: The state of art (pp. 3–48). Cambridge, MA:
Harvard University Press.
Hambruger, H., & Crain, S. (1982). Relative acquisition. In S. A. Kuczaj (Ed.), Language
development, Vol. 1: Syntax and semantics. Hillsdale, NJ: Erlbaum.
Horgan, D. (1978). The development of the full passive. Journal of Child Language, 5, 65–80.
Hsu, J. (1996). A study of the stages of development and acquisition of Mandarin Chinese by
children in Taiwan. Taipei: Crane.
Karmiloff-Smith, A. (1979). A functional approach to child language: A study of determiners and
reference. Cambridge: Cambridge University Press.
Klima, E., & Bellugi, U. (1966). Syntactic regularities in the speech of children. In J. Lyonsand &
R. Wales (Eds.), Psycholinguistic papers. Edinburgh: Edinburgh University Press.
Kuo, J. Y. (2008). Mandarin acquisition by children of Vietnamese Mothers in Taiwan. Taipei:
Crane.
Levy, Y. (1988). The nature of early language: Evidence from the development of Hebrew
morphology. In Y. Levy, I. Schlesinger, & M. D. S. Braine (Eds.), Categories and processes in
language acquisition (pp. 73–98). Hillsdale, NJ: Erlbaum.
Lin, H-Y. (2009). Toward an index of productive syntax in Mandarin. MA thesis. National Chiayi
University.
Maratsos, M. (1988). The acquisition of formal word classes. In Y. Levy, I. M. Schlesinger & M.
Braine (Eds.), Categories and processes in language acquisition (pp. 31–44). Hillsdale, NJ:
Erlbaum.
Mills, A. E. (1986). The acquisition of gender: A study of English and German. Berlin:
Spinger-Verlag.
Pérez-Pereira, M. (1991). The acquisition of gender: What Spanish children tell us. Journal of
Child Language, 18, 571–590.
Plunkett, K., & Marchman, V. (1991). U-shaped learning and frequency effects in a multi-layered
perception: Implications for child language acquisition. Cognition, 38, 43–102.
Pye, C. (1988). Precocious passives (and antipasives) in Quiche Mayan. Paper presented at the
Child Language Research Forum, Stanford, CA.
Radford, A. (1990). Syntactic theory and the acquisition of English syntax. Oxford: Basil
Blackwell.
Scarborough, H. (1989). Index of productive syntax. Applied Psycholinguistics, 11, 1–22.
Suzman, S. (1987). Passives and prototypes in Zulu children’s speech. African Studies, 46,
241–254.
Tager-Flusberg, H. (1982). The development of relative clauses in child speech. Child Language
Development, 18, 104–111.
Valian, V. (1986). Syntactic categories in the speech of young children. Developmental
Psychology, 22, 562–579.
References 149
Valian, V. (1991). Syntactic subject in the early speech of American and Italian children.
Cognition, 40, 21–81.
Winzemer, J. A. (1980). A lexical expectation model for children’s comprehension o wh-questions.
Paper presented at the Fifth Annual Boston University Conference on Language Development.
Wootten, J., Merkin, S., Hood, L., & Bloom, L. (1979). Wh-questions: Linguistic evidence to
explain the sequence of acquisition. Paper presented at the biennial meeting of the Society for
Research in Child Development, San Francisco.
Chinese
The syntactic development of children of Indonesian mothers was slower than that
of children of Taiwanese parents in length, sentence types, and questions accuracy.
They have shorter mean length of utterances, lower Mandarin Index of Productive
Syntax, and lower accuracy rate in questions. The mean MLU (M = 3.02) was
smaller than the Taiwanese group (M = 3.61) and such difference showed up as
early as age 2. The sentence patterns used by the Indonesian group (M = 36.11)
were also fewer than those used by the Taiwanese group (M = 39.10). The mean
accuracy rate of questions by children with Indonesian mother is a little bit lower
than their age-mates with Taiwanese mothers. The mean of Indonesian children is
77.39 % while that of Taiwanese children is 79.73 %. Over all, the syntactic
development of Indonesian children is slower than native Taiwanese children.
The results are different from children of Vietnamese mothers, whose MLU
shows little difference from children of Taiwanese mothers in the MLU till age six
with even more sentence types (Kuo 2008).
5.4 Limitations 153
5.4 Limitations
There are two limitations of the study. First, the children in the two groups did not
match exact age in months. The age difference could have caused differences for
younger groups. Second, due to the small sample size, the results may not be
representative for all the children of immigrants. I was unable to control all the
factors that affect language acquisition such as the children’s IQ, siblings, and
socioeconomic status. Although IQ was not measured, none of them was reported
abnormal. Individual differences were not the focus of the present study. Families
with Indonesian mothers tend to have lower socioeconomic status than native
Taiwanese families. I have done my best to match the status of Taiwanese families
with them.
5.5 Implications
References
Gentner, D. (1981). Some interesting differences between verbs and nouns. Cognition and Brain
Theory, 4(2), 161–178.
Kuo, J. Y. (2008). Mandarin acquisition by children of Vietnamese mothers in Taiwan. Taipei:
Crane.
Tardif, T. (1996). Nouns are not always learned before verbs: Evidence from Mandarin speaker’s
early vocabulary. Developmental Psychology, 32, 492–504.
Tardif, T., Gelman, S., & Xu, F. (1999). Putting the “noun bias” in context: A comparison of
English and Mandarin. Child Development, 70, 620–635.
Yip, V., & Matthews, S. (2010). The acquisition of Chinese in bilingual and multilingual contexts.
International Journal of Bilingualism, 14(1), 127–146.