Chapter 5 Discussion and conclusion
5.2 Discussion on syllable analysis
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
types more often, the child will produce those types more often. That is, this indicates that child’s syllable patterns may be influenced by his/her caretaker’s linguistic production.
5.2 Discussion on syllable analysis
This study aims to investigate the patterns of overall syllables types and syllable deletions in Taiwan Mandarin-speaking children. Since young children’s productions of words are immature and unstable, they tend to use many different phonetic forms and drop the segments in the outputs. Syllable acquisition and syllable deletion can be attributed to many factors, including markedness factors, frequency effects, perception and production accounts (e.g., Ingram 1986; Echols, 1993; MacNeilage &
Davis, 1993; Stites et al, 2004; Ota, 2006). In this study, the relationship between syllable type frequency and syllable deletion as well as the child-caretaker syllable relationship are carefully examined.
5.2.1 The overall syllable types
For the four children’s speech productions, CV and CVG types have higher frequencies in overall syllable types while VG and VN have lower frequencies. On the one hand, in terms of the syllable types with higher frequency, CV and CVG syllable types are syllables which consist of onset consonants and do not contain consonant clusters. The results show that the frequently used syllable types generally
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
accord with the unmarked syllable criterion proposed by Jakobson (1968). Although CVG ranks as the second most frequently used type and it is not a complete open syllable, this type with the onset consonant and without the consonant cluster is considered to be a simple syllable type for the children. On the other hand, CV and CVG also have the higher frequencies in the caretakers’ syllable productions. The frequency effect can also account for the results (e.g., Ingram, 1988; Chen & Kent, 2005, 2010; Ota, 2006). That is, a target phonological structure produced by a young child is closely related to the input frequency. It is indicated that the ambient languages may influence young children’s syllable structures.
However, VG and VN have the lowest frequency of the four children’s production. They are closed syllables among the twelve syllable types in Taiwan Mandarin and they do not have the onset consonant. Although they have relatively simple structure compared to CGVG or CGVN, they are the two syllable types without onset consonant and are regarded as the difficult and marked syllable types.
Moreover, in terms of markedness theory, the more marked syllable types would be acquired later because young children would display unmarked linguistic structure first (Jakobson, 1968). Accordingly, because VG and VN are the more marked syllable types, young children would acquire these difficult types during the latter course of development (H. Y. Wang, 2014). Nevertheless, CGVG and CGVN
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
consisting of the consonant cluster do not have the lowest frequency of all utterances.
The reason may be due to the fact that the age ranges during the observation. After the age of two, words learned and produced by the children are more complex so that the frequencies of CGVG and CGVN are not the lowest in the data. Thus, syllable types with the lowest frequency may also be consistent with the constraints mentioned above, including the closed syllable, the syllable without onset, and syllable contained consonant cluster which are more marked syllable types.
5.2.2 The syllable deletion types
According to results of syllable deletion in monosyllables, disyllables and multisyllables, CV, CGV and V are the frequently used syllable types among twelve types. CV still has the highest frequency in syllable deletion words, and is then followed by CGV and V in the production of four children. They are the syllables, which include onset consonant and do not contain consonant clusters and codas.
Among syllable types used by children, it is clearly shown that the unmarked syllable CV is the most frequently used type since young children would tend to omit the segments in complex syllable types to become simple syllable types. As a result, the more unmarked syllable types are likely to be used in children’s early immature speech.
In terms of the syllable deletion forms, the children tend to use a coda-dropping
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
strategy to produce reduced words in phonological development (Levitt & Aydelott Utman, 1992; So & Dodd, 1995; Tsay, 2007). The results show that CV is the most frequently used syllable type produced by four children in overall production and syllable deletion productions. The findings are consistent with the study of McCarthy and Prince (1994), who proposed that children’s early words are governed by highly-ranked No-Coda constraints. For instance, complex target forms are easily to be realized as CV depending on the relative ranking of faithfulness constraints on the realization of segments. Hence, it may be predicted that CV syllable type is the most common outputs of syllable errors in children’s speech.
However, for least frequent syllable types of syllable deletion forms, the frequencies of GVG, GVN, VG and VN have the low frequencies among all the deleted syllable types. The frequencies of GVG and GVN account for less than 2% of the total, being only 0.7% and 1.1% respectively. These less frequently used syllable types are closed syllables and VG and VN and thus have no onset consonant, so these syllable types appear less frequently then open syllables. These results may due to the fact that Mandarin severely limits the possible coda consonants. Some studies of syllable omission in Mandarin have claimed that open syllables are more likely to undergo phonetic reduction as opposed to closed syllables (Cheng & Xu, 2009;
Burchfield & Bradlow, 2014). As a result, the findings are in accordance with
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
previous studies, indicating that languages favoring coda-less syllables may show overall higher rates of phonetic deletion
5.2.3 The syllable deletion patterns
In terms of the pattern of syllable deletion, the four children were likely to drop the segments in the syllable-final position than syllable-initial and syllable-middle position. On the one hand, the deletion of the glide [w] in syllable-final position has
the highest percentage of all the deletion categories, and is then followed by the nasal [n] and [ŋ] in syllable-final position whereas the percentage of the prenuclear glide [j]
and prenuclear glide [w] have the least frequencies of all categories. This deletion pattern is consistent with previous studies, demonstrating that young children show limited ability at final consonant production in syllable-final position (Fee & Ingram, 1982; Demuth & Johnson, 2003). Demuth and Johnson (2003) also proposed that because young children’s phonological system is unstable and immature, they are inclined to drop word-final segments and produce the targets as CV words or truncated words. On the other hand, it seemed that the glide [j] in the syllable-final position does not explained by the coda-dropping strategy because the percentage of postnuclear glide [j] has the second last frequency of all categories. But, the reason may be due to the fact that the young children seldom produced the target words which contained the postnuclear glide [j]; hence, the syllable deletion of postnuclear
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
glide [j] occurred less frequently in the analysis.
Nevertheless, when carefully examining each child’s deletion patterns, we can observe that child #1 tended to reduce the segments in the syllable-initial position rather than in the syllable-final position. The deletion of initial consonant had the highest percentage of her all reduction categories, and was then followed by final nasals and the deletion of prenuclear glide [w]. The initial consonant deletion is also in accordance with the studies of Zhu and Dodd (2000) and Tsay (2006), who claimed that young children are likely to omit the syllable-initial consonant in their early phonological acquisition. Young children are expected to delete some components such as syllable-initial consonant at the syllable level. The syllable-initial consonant deletion is very common in young children’s data (Zhu & Dodd, 2000).
In general, although there are individual differences in terms of dropping the segments in positions, young children still have undergone the syllable deletion in their immature productions. Thus, the findings of syllable deletion pattern may indicate that children tend to reduce the consonants and glides in every syllable position.
5.2.4 The syllable relationship between children and caretakers
Based on the results of children’s overall syllable production pattern, CV is obviously the most frequently used syllable type, and is then followed by CVG. For
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
caretakers’ syllable patterns, CV, CVG and CVN have higher frequencies in the results. According to the statistical findings, the syllable types of the four child-caretaker pairs are positively correlated. The two highest ranking of frequencies of syllable types were CV and CVG in the productions of pair #13, pair #2 and pair #3.
VG and VN had the least frequencies in the production of pair #1, pair #2 and pair #3.
Nevertheless, pair #4 presented slightly different patterns from the other pairs. Child
#4 used CV and CGV more frequently but used GVN and CGVN less frequently;
caretaker #4 produced CV and CVG more frequently but VN and VG appeared less frequently. One possible reason for there being slightly different syllable patterns may lies in the individual differences (Levelt et al., 2000). The children may exhibit individual differences in the productions, preferring the different complexity of the syllable types in the syllable-initial or the syllable-final.
Nevertheless, with regard to the syllable patterns of the four children-caretakers, the results are significantly correlated. In general, our findings are congruent with those previous studies (e.g., de Boysson-Bardies & Vihman, 1991; Chen & Kent, 2005; Zamuner et al., 2005; Ota, 2006). The ambient language can be observed from children’s final consonant production in their early stages (de Boysson-Bardies &
Vihman, 1991). Moreover, the finding of Chen and Kent (2005) indicate that young
3 In order to be succinct, the following child-caretaker relationship would be marked as a “pair” (e.g., pair #1 referring to child #1 and caretaker #1).
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
children’s syllable productions are relevant to their caretakers. Children’s phonological productions may be influenced by the overall productivity and the phonetic content of the ambient language. Furthermore, it can be reasoned that a young child produces a target phonological structure which is strongly related to the input frequency (Ota, 2006). To conclude, the findings lend some credence to the accounts that children’s syllable patterns have largely been influenced by their linguistic environment (i.e. caretakers’ speech).