Chapter 4 Results and Analysis
4.5 Relationship between frequency, variability rate, and substitution pattern
4.5.2 Substitution pattern and frequency
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
syllable type in children‟s speech and it also had lower variability rates which only
accounted for 7.1 percent of utterances in disyllabic tokens. CGV ranked as second place in frequency, and also showed low variability rates.
Although CVG was frequently used in children‟s disyllabic word production, it also showed relative high variability rates in disyllabic words. The data of participants showed that when they produced words with CVG, participants often dropped the final consonant, and the same phenomenon happened in words contained CVN. So the variability rates for CVG and CVN were always higher. The results are consistent with earlier findings which suggested that coda dropping is very common in phonological acquisition (Tsay, 2007; So & Dodd, 1995).
VN and CGVG had the highest variability rates which accounted for 47.5%
and 26% respectively. Not surprisingly, these two syllable types were less frequently used syllable types in the data.
To sum up, variability rates were in some extent influenced by frequency.
Syllable types with higher frequency presents lower variability rates. The more the children produce a certain kind of syllable type, the lower the variability rate is.
4.5.2 Substitution pattern and frequency
The relationship between syllable type‟s substitution pattern and frequency is also of interest. Based on Table 4.13, children replaced the syllable types they were
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
not mastered with CV the most, followed by CGV, GV, and V. The overall frequencies of syllable types among all the 9596 tokens was presented in Table 4.15
Table 4.15 Percentages of syllable types in all syllabic tokens Syllable type Tokens Percentages
CV 3975 41.4%
CVG 1639 17.1%
CGV 1510 15.7%
CVN 526 5.4%
GVG 315 3.3%
V 497 5.2%
GV 481 5%
CGVG 253 2.6%
CGVN 232 2.4%
VG 51 0.5%
GVN 66 0.7%
VN 61 0.6%
Total 9596
Based on Table 4.15, CV was the most frequently used syllable type with the usage of 41.1%, followed by CVG, CGV, CVN, V, and GV. The results show that children do not necessarily replace syllable types they are not mastered yet with more frequent ones. For example, CVG and CVN had higher frequency than CGV and V, but the percentage of substitution of CVG only reached 3.9% and CVN only reached
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
0.9%. Clearly, CVG and CVN were not frequently used in replacing. As mentioned in section 4.5.1, children tended to drop the coda consonant when producing immature syllable types. That is, they used syllable types without coda consonant to replace others more frequently, so even CVG and CVN had higher frequency, they were not frequently used to replace immature syllable type.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
83
Chapter 5 Discussion 5.1 Summary of the findings
In this study, the phonological variability of word production by six Mandarin-speaking children aged 0;11 to 2;0 is observed. Two aspects are included:
the phonological variability of words and the syllable types composed the words. We examined the frequencies, the variability rates, the substitution pattern of each syllable type, and calculated the phonological variability pattern of each participant.
There were a total number of 5788 words and 9596 syllabic tokens analyzed in the study. The results and findings are summarized below:
Table 5.1 Ranking of syllable types in different measure (1) Syllable type frequency CV > CVG > CGV
Syllable combination frequency CV-CV > CGV-CGV> CVG-CVG (2) Syllable type variability rate Monosyllable: VG>CGVN>CGVG
Disyllable: VG, VN, CGVN, CGVG (3) Substitution pattern CV>CGV>GV>V
(1) Variability pattern: although each participant showed somewhat different pattern of variability, there were still some similarities. Firstly, the overall variability
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
patterns of six participants all showed peaks and valleys. Second, the variability rates gradually decreased as the participants grew older, and among six participants, participants #1 had the most regular decreasing pattern. Third, every participant had a noticeable variability peak at different age. The variability peaks correspond with the time when these participants entered a new stage of phonological development.
(2) Syllable type frequency: in both monosyllabic and disyllabic words, CV had the greatest frequency of occurrence, followed by CVG and CGV. CVN and GV ranked as fourth place in monosyllabic and disyllabic words respectively. For the least frequent syllable types, VG, GVN, and VN showed the lowest frequency in bout monosyllabic and disyllabic words. The frequency of GVG was relatively low, accounted for only 0.07% in first syllable, but reached 4.9% in second syllable. The production of GVG was mostly limited to second syllable, resulting in its lower frequency in first syllable. The results of syllable combinations showed that combinations presented higher frequency are those with consonant-vowel reduplicated forms.
(3) Syllable type variability rate: the results showed that the syllable types that had higher variability rates in monosyllabic words were VG, CGVN, and CGVG. In disyllabic words, VN and CGVN presented higher variability rates. CV and V
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
were the last two in the ranking of variability rates in both monosyllabic and disyllabic words.
(4) Substitution pattern: the ranking of syllable types that were most frequently used to replace other types was CV>CGV>GV>V, and the ranking of syllable types that were most frequently replaced by other types was CVG>CGVG>CVN. The findings suggest that the strategy participants used in replacing others was to
simplify the syllable by only dropping the coda consonant.
5.2 Discussion on variability pattern
Ferguson and Farwell (1975) proposed the ‘whole-word system of phonological representation.’ which claimed that children are able to be aware of relatively large
phonological units, such as syllables, at early stages of phonological development.
When the amount of vocabulary in children’s lexicon grows, phonemic representation
develops since there is a need to discriminate the speech sounds efficiently in production and perception. Children’s productions of each sound and each word then
become more and more stable. As a result, children’s phonological development is a process from whole-word representation to phonemic representation. Variability can serve as evidence since the lack of segmental phonemic representation detail in child's underlying phonological representation would result in greater degree in variability.
In this study, the overall variability patterns of each participant were examined.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Each participant in the study presented a certain amount of phonological variability during the observation. However, the variation was not unlimited. In fact, it appeared to be quite principled. As mentioned above, the variability patterns of participants showed peaks and valleys, and every participant had a noticeable variability peak.
Although the variability peaks of all the participants appeared at different age, their productions exhibited great differences after the variability peaks. Participants #2, #3, and #4 started to produce two-word utterances and participants #5 and #6 produced much more different words after the variability peaks appeared.
The results may correspond to the study of Sosa and Stoel-Gammon (2006) which proposed that the increase of phonological variability at the onset of two-word combination speech reflects a reorganization of the linguistic system, from one-word stage to two-word stage. Since the phonological system are unstable, variability rates then increase. But according to Vihman (1996), the system would once again stabilize and variability would then decrease. The process may continue until children were 7 years old. Noted that the variability rates of participants were mostly high at early stage during the observation, the reason might due to the fact that the development of neuromotor control for speech is not mature yet. Children at this stage are biologically limited by the neuromotor control ability, so they cannot produce a word with a same form consistently as well as accurately (A. Smith & Goffman, 1998; Walsh& Smith,
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
2002).
To sum up, examining the overall variability patterns shows us the process of
phonological development. At early stage of development, since the speech motor control is not mature, the variability rate is high. When children’s lexicon starts to
grow, their phonemic representation development then becomes mature, resulting in a more stable production of each word. However, when linguistic reorganization happens, variability rate increases again. This study did not show the once-again stabilized pattern of variability.
5.3 Discussion on syllable type analysis
The study of overall variability pattern aims to investigate the developing process of children’s phonological system. Since children’s early lexicon
representation is the whole-word, the syllable types which composed the variable repeated words in children’s speech are also of interest. Phonological variability can be attributed to many factors including the immature of speech neuromotor control, phonological complexity, and phonological overload (Holm et al., 2007; Macrae, 2013; McLeod & Hewett, 2008; Sosa & Stoel-Gammon, 2006 & 2012). In this study, the relationship between syllable type frequency and variability rate as well as the substitution pattern were examined.
The findings in this study can be explained by the markedness theory
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
presented by Jakobson (1941/68) which proposed that children would acquire the unmarked form of language first, and only later acquire the more marked ones. Those phonetic forms most commonly found cross-linguistically are considered to be
unmarked, and would therefore be the first to be acquired. Those who less frequently attested in the world’s languages are considered to be marked, and would be acquired
later. Jakobson’s markedness theory of syllable can be summarized as below:
(1) Open syllable is more unmarked than closed syllable
(2) Syllable with onset is more unmarked than those without onset (3) Syllable contained consonant cluster is marked.
5.3.1 Syllable type frequency
In this study, syllable types with higher frequency were CV, GV, CGV, and CVG, while VG, GVN and VN had lower frequency in monosyllabic words as well as in disyllable words. Among the syllable types with higher frequency, firstly, CV and GV are open syllable and they are also the only two open syllables among the 12 syllable types in Mandarin. Secondly, four of these syllable types are syllables contained onset consonants. And thirdly, three of these syllable types do not contain consonant cluster. To sum up, syllable types with higher frequency accord with the unmarked syllable criterion proposed by Jakobson, so it could be generalized that syllable types with higher frequency in this study are more unmarked syllable types.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
As for syllable types with the lowest frequency in this study, all of them are closed syllable and two of them have no onset consonant. The frequency of syllable types contained consonant cluster such as CGVG and CGVN, although do not had the lowest frequency, compare to other syllable types their frequencies are still relatively low. So, again, the results might also be consistent with the constraints mentioned above, which stated that closed syllable, syllable without onset, and syllable contained consonant cluster are more marked syllable types.
5.3.2 Substitution pattern
As for substitution pattern, the ranking of syllable types that were most frequently used to replace other types was CV, followed by CGV, GV and V. Noted that the frequencies of these four syllable types were higher than other types; however, syllable types with high frequency did not necessarily used frequently in replacing others. For example, CVG had high frequency, even higher than CGV, but the percentage of CVG used in replacing others only reached 3.9%. The reason may due to the fact that coda-dropping is very common in phonological acquisition (Tsay, 2007; So & Dodd, 1995), and it was also a commonly used strategy by children in this study when replacing other types. For example, CVN was most frequently replaced by CV; CGVG was most frequently replaced by CGV. So, the strategy participants used when replacing immature syllable types was to simplify the syllable by only dropping
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
the coda consonant. That is, syllable types that were most frequently used in replacing others are all open syllables. Furthermore, they are also syllables contained onset
consonants. The results accord with McCarthy and Prince (1994), who proposed that children’s early productions were governed by highly-ranked No-Coda constraints,
which predicts that CV syllable types appear to be the most common output of syllable errors.
Syllable types that were frequently replaced by others were CVG, CGVG, and CVN. All of these syllable types also presents similar features: they are all closed syllables, and one of them contains consonant cluster. So, the findings of substitution pattern may indicate that children have the tendency to replace a more marked syllable type with a more unmarked one.
5.3.3 Syllable type variability
Studies have shown that phonological variability is most likely to occur when one or more aspect of the word is unstable in child’s phonological system; that is, phonological elements were presented in a child’s speech, but not yet mastered (Holm
et al., 2007; Leonard et al., 1982; McLeod & Hewett, 2008). In this study, syllable
types that had higher variability rates in monosyllabic words were VG, CGVN, and CGVG. In disyllabic words, VN and CGVN presented higher variability rates.CV and V were the last two in the ranking of variability rates in both monosyllabic and
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
disyllabic words.
Syllable types presented higher variability rates are more marked syllables. All of them are closed syllables; VG and VN are syllables without onset; CGVG and CGVN are syllables contained consonant clusters. The results are in support of the study of Leonard et al. (1982). They suggested that words with higher variability rates are most often those which have more advanced forms, sounds, or word shapes. Among the syllable types with high variability rates, two of them are the most maximal syllable in Mandarin: CGVG and CGVN. So, it is much more difficult for children to master, resulting in high variability rates.
As for VN and VG, although they have relatively simple structure compared to CGVG or CGVN, they are the only two syllable types without onset consonant in Taiwan Mandarin, which are the more marked syllable types. So, it would acquired later by children. As proposed by Macrae (2013), later-acquired sounds are produced less accurately and with more variability than earlier-acquired sounds.
Comparing the syllable type frequency and variability rate, the results show that syllable types with higher frequency would present lower variability rates, and
syllable types with lower frequency would show higher variability rates. For example, CV was the most frequently used syllable type in children’s speech and it presented
the lowest variability rates in both monosyllabic and disyllabic words. The results
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
accord with Sosa and Stoel-Gammon (2012). They suggested that high frequency facilitates both word recognition and word production, that is, the relationship
between syllable type frequency and variability rate may respond to the rule that
‘practice makes perfect’, so the more the children produce a certain kind of syllable
type, the lower the variability rate of that type is.
5.4 Concluding remarks
The purpose of this thesis is to discuss the issue concerning phonological variability of children acquiring Taiwan Mandarin. Two aspects are included: the overall variability pattern, and the syllable types that composed the variable phonetic forms of words. The overall variability pattern, the syllable type frequency, variability rates, and the substitution pattern of immature syllable types were examined. From the analysis and discussion, the patterns of variability displayed by these children observed here indicate that variability pattern is influenced by the development of speech neuromotor control, the phonemic representation, and also the linguistic reorganization process. At early stage in phonological development, since the speech motor control is not mature, the variability rate is high. When children’s lexicon starts to grow, their phonemic representation development then becomes mature. However, when linguistic reorganization happens, and in this case, the onset of two-word combination and the increased amount of different word, variability rate increases
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
again. Presumably, the system would once again stabilize and variability would then decrease. The process may continue until children were 7 years old. This study didn’t
show the once-again stabilized pattern of variability.
Concerning the syllable type frequency, variability rates, and substitution pattern, the results would be in support of the markedness theory of syllable proposed by Jakobson (1941/68). The comparison showed that syllable types with higher frequency as well as syllable types used frequently in replacing others are more unmarked ones. Syllable types with lower frequency and syllable types that were replaced by others frequently are more marked ones. Furthermore, syllable types contained more complex structure, that is, a more marked syllable, would present high variability rates. It would be beneficial for future work to trace patterns of phonological variability for more participants, for longer duration, and carefully examine the
development of both variability and accuracy of individual words over time. Furthermore, comparing the syllable types used in motherese and children might also provide different view in the syllable acquisition in Taiwan Mandarin.
All in all, the findings in this thesis provide evidence of the process of phonological
variability in Taiwan Mandarin
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
References
Anderson, J. (2007). Phonological neighborhood and word frequency effects in the
stuttered disfluencies of children who stutter. Journal of Speech, Language, and
Hearing Research, 50, 229–247.
Anderson, J. (2008). Age of acquisition and repetition priming effects on picture
naming of children who do and do not stutter. Journal of Fluency Disorders, 33,
135–155.
Chao, Y-R. (1968). A Grammar of Spoken Chinese. Berkeley. University of
California Press.
Cheng, C-C. (1973). A Synchronic Phonology of Mandarin Chinese. The Hague:
Mouton.
Chen, L-M. (2005). Development of Vocalic production in a Mandarin learning infant:
a longitudinal case study. Providence Forum: Language and Humanities, 1(1),
169-192.
Davis, S., & Hammond, M. (1995). On the status of onglides in American English.
Phonology,12,159-182.
Demuth, K., & Johnson, M. (2003). Truncation to subminimal words in early French.
Canadian Journal of Linguistics, 48, 211-241
Dodd, B., Leahy, J., & Hambly, G. (1989). Phonological disorders in children:
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
underlying cognitive deficits. British Journal of Developmental Psychology, 7, 55-71.
Duanmu. S. (1990). A Formal Study of Syllable, Tone, Stress, and Domain in Chinese Languages. Cambridge: MIT dissertation.
Duanmu, S. (2007). The Phonology of Standard Chinese. New York: Oxford University Press.
Duanmu, S. (2010). Syllable Structure: The Limits of Variation.Paperback edition.
Oxford University Press.
Ferguson, C. A., & Farwell, C. B. (1975). Words and sounds in early language acquisition. Language, 51(2), 419–439
Gierut, J., Morrisette, M., & Champion, A. (1999). Lexical constraints in phonological acquisition. Journal of Child Language, 26, 261–294.
Gleason, J. B., & Ratner, N. B. (2009). The development of language. Boston, Mass:
Pearson Press.
Green, J. R., Moore, C. A., & Reilly, K. J. (2002). The sequential development of jaw and lip control for speech. Journal of Speech Language and Hearing Research, 45(1), 66–79.
van Geert, P., & van Dijk, M. (2002). Focus on variability: New tools to study intra-individual variability in developmental data. Infant Behavior &
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Development, 25, 340–374.
Holm, A., Crosbie, S., & Dodd, B. (2007). Differentiation normal variability from
Holm, A., Crosbie, S., & Dodd, B. (2007). Differentiation normal variability from