Chapter 3 Methodology
3.2 Data Analysis
3.2.5 The pattern of syllable deletion
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
higher. The frequency of each type of syllable deletion would be analyzed by the formula below:
Percentage = Tokens of a deleted syllable type
Tokens of all deleted syllable type X 100%
The numerators were the tokens of a particular reduced syllable type. The fraction was presented as a percentage. The frequency of child’s syllable reduction could reveal their syllable tendencies in the development.
3.2.5 The pattern of syllable deletion
Since the children’s productions of words were not mature and stable, they might delete the segments in the syllables to produce a word. The following examples showed the different categories of deletions in the analysis. There were six categories
in the analyses, including deletion of initial consonant, final nasal [n]/[ŋ], prenuclear glide [j], prenuclear glide [w], postnuclear glide [j] and postnuclear glide [w].
If the children produced GVN-CVG syllable structure in [wən35 tʰaj55] tire, rather than CGVN-CVG in [lwən35 tʰaj55], they omitted the initial consonant. If the children produced the GV-GV in [ja21 ja35] goat, rather than GVN-GVN in [jɑŋ21
jɑŋ35], they deleted the final nasals. Another example showed that the children produced CV in [pʰa21] run, instead of CVG in [pʰaw21], showing that the
postnuclear glide was omitted.
Furthermore, many studies have showed that children in early speech tended to
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
drop the segments in the final-syllable positions (Roark & Demuth, 2000; Demuth &
Johnson, 2003) in terms of hierarchical constraints “No-Coda” (e.g., Pater, 1997;
Levelt et al., 2000). For example, CVC target forms were easily to be realized as CV
depending on the relative ranking of faithfulness constraints on the realization of segments. However, some researchers proposed that children might drop the syllable-initial consonant in the early speech (e.g., Zhu & Dodd, 2000; Tsay, 2006).
For example, the child would produce V-CV syllable structure in [ɤ35 ma21]
hippopotamus, instead of CV-CV in [xɤ35 ma21]. This child deleted the initial consonant in the production. Hence, the deletion patterns could show a tendency of children’s production patterns.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
54
Chapter 4
Result and analysis
In this section we will examine the results of the analysis. This analysis follows the research questions of this study as laid out in Chapter 1. Section 4.1 presents an overview of the overall data. Section 4.2 demonstrates the results of syllable analyses based on different syllable numbers, including both the frequent overall syllable types and the frequent syllable deletion in the production. Different patterns of syllable deletion are presented in section 4.3. Finally, section 4.4 clearly presents the child-caretaker’s output syllable relationship.
4.1 Data Background Information
Four children and four caretakers were enrolled in the analysis. The observation started at the age of 0;10 to 1;1 and ended at the age of 2;3 to 2;6. The children’s speech data was collected from the non-reduplicated babbling stage when the children started to produce different syllable types as well as their first meaningful words.
As noted in Chapter 3, on the one hand, the analysis of this study only included sounds that have clear meaning and can be inferred from their respective contexts or from adults’ repetition. Tokens without clear meanings or tokens that were acoustically unrecognizable are excluded. In total the data contains 16515 syllabic tokens produced by children and 4586 tokens produced by caretakers, as shown in
‧
Table 4.1. Table 4.1 lists of the relevant information concerning four children, including: gender, age, duration of observation and total tokens of words that the four participants uttered.
Table 4.1 Background information regarding the children Participant Gender Age
In table 4.1, the duration of observation continued for 18 months, starting from the babbling stage and continuing to the telegraphic speech stage. Child #3 was the most productive of all the children while child #4 was the least productive child. Child #1 produced 3352 types of tokens and child #2 produced 5041 tokens. Child #3 produced 5390 tokens while child #4 only produced 2732 tokens. There were 16515 syllabic tokens of utterances suitable for analysis in this study.
In addition, the following Table 4.2 displays relevant information regarding the four caretakers, including gender, age, duration of observation and total syllabic tokens.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table 4.2 Background information regarding the caretakers
Participant Gender Relation Syllabic tokens Caretaker #1 Female Mother of child #1 1496
Caretaker #2 Female Mother of child #2 1071 Caretaker #3 Female Mother of child #3 973 Caretaker #4 Female Mother of child #4 1046
Total 4586
The researcher transcribed a random sample of the four caretakers’ recordings, accounting for 10% of all recordings. As shown in Table 4.2, Caretaker #1 was the most productive mother of all the caretakers, producing 1496 syllabic tokens.
Caretaker #2 produced 1071 syllabic tokens and caretaker #3 produced 973 syllabic tokens. Caretaker #4 produced 1046 syllabic tokens. There were 4586 syllable tokens of utterances analyzed in the study.
4.2 Frequency of syllable types
The tokens of overall different syllable types in different syllable numbers were calculated in order to see the general syllable types produced by children. The tokens collected in this section included the calculations of overall syllable types and deleted syllable types.
In order to answer the research questions, the frequencies of syllable types were examined in order to indicate which syllable type was more frequently used and which type was less used. The frequency of overall syllable types is shown in section 4.2.1, and the frequency of syllable deletion is examined in section 4.2.2.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
4.2.1 Frequency of overall syllable types
This section demonstrates the overall frequency of syllable types among the four children. It is noted that analysis of syllable tokens includes monosyllabic words, disyllabic words and multisyllabic words. To discuss whether different types of tokens influence the syllable development in acquisition, monosyllables, disyllables and multisyllables are analyzed accordingly. Because all the children could produce utterances longer than disyllable words when they were at two years old, there are three categories of their syllable utterances. In terms of disyllables and mulitsyllables, the tokens of syllable types in different syllable positions would not be separately analyzed. Based on the analyses of syllable frequencies, we can see the general and preferred syllable patterns occurring in children’s speech. The frequency of monosyllabic words, disyllabic words and multisyllabic words is shown in section 4.2.1.1, section 4.2.1.2 and section 4.2.1.3 respectively.
4.2.1.1 Overall monosyllabic words
In terms of statistical analysis of overall monosyllables, since the assumption of homogeneity of variance was not met for the data, F(11, 36) = 6.16, p<.05, we used an alternative Welch’s adjusted F ratio and reported it as Welch’s F(11, 13.90) = 15.65, p<.05. The analysis using the Welch test shows that there were significant differences among the twelve syllable types in monosyllabic words. Moreover, the
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
result of the follow-up with Duncan’s multiple range test is shown in Table 4.3.
Table 4.3 Duncan's Multiple Range Test for overall monosyllables
Duncan grouping Mean N Monosyllables
A 157.00 4 CV
B 64.75 4 CVG
C 40.25 4 CGV
C 36.25 4 GV
D 31.00 4 CVN
D 28.00 4 GVG
D 26.00 4 CGVG
D 22.25 4 V
D 15.75 4 CGVN
D 7.25 4 GVN
D 5.50 4 VG
D 5.50 4 VN
According to the above Table 4.3, group A (i.e. CV) is significantly different (p<.05) from the B, C and D groups. Group B (i.e. CVG) is also significantly different (p<.05) from the other syllable types. Hence, we may ascertain that CV and CVG are superior to the other syllable types, and statistically significantly different.
In addition, the tokens and percentages of each syllable type in monosyllabic words are presented in the following Table 4.4 and Figure 4.1. The bar graph in Figure 4.1 was used to clearly compare the differences among frequencies of syllable types.
‧
Table 4.4 Tokens and percentages of overall syllable types in monosyllabic words
Type V CV VG GV VN CVG CVN CGV GVG GVN CGVG CGVN
Figure 4.1 Percentages of overall syllable types in monosyllabic words
Based on Table 4.3 and Figure 4.1, CV is the most frequently used syllable type in monosyllabic words, accounting for 35.7% of the total 1758 syllables. CVG and CGV rank second and third place, with a usage of 14.7% and 9.2% of utterances respectively. GV ranks fourth place, with a usage of 8.2%. However, as for the least frequently used syllable types, VG, VN and GVN are less than 2% of all utterances.
The ranking of frequencies of syllable types is thus CV > CVG > CGV > GV. The
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
results obviously show that the children preferred CV in monosyllables the most due to the fact that CV has the highest frequency.
Moreover, in order to discuss whether there are individual differences among children, the following Table 4.5 displays the frequencies of overall monosyllable types produced by each child.
Table 4.5 Percentages of overall syllable types of each child in monosyllabic words
V CV VG GV VN CVG CVN CGV GVG GVN CGVG CGVN
Child #1 7.8% 25.0% 0.4% 7.6% 1.6% 18.1% 7.4% 8.0% 8.0% 1.3% 9.2% 5.6%
Child #2 2.8% 37.2% 1.2% 12.8% 1.2% 13.4% 7.5% 8.9% 4.5% 2.8% 4.7% 3.0%
Child #3 4.9% 38.0% 1.0% 5.5% 1.4% 14.4% 8.6% 9.2% 7.2% 1.6% 4.1% 4.3%
Child #4 5.0% 45.2% 3.0% 6.6% 0.7% 12.5% 3.3% 11.2% 5.6% 0.3% 6.3% 0.3%
As shown in Table 4.5, for the most frequently used syllable type, CV has the highest frequency of occurrences among the four children’s monosyllabic words, and is followed by CVG. However, as for the third place, child #2 presented slightly different pattern compared to that of the other children. GV ranks as the third most frequently used syllable type with a usage of 12.8%. The other children exhibited similar patterns: CV was the most frequently used type, and is then followed by CVG and CGV. Briefly, CV, CVG and CGV were used more frequently than other syllable types of the children’s monosyllables. For the least frequently used syllable types, although there were slightly different syllable patterns, VG, VN and GVN had lower frequencies in the production of the children.
‧
4.2.1.2 Overall disyllabic words
Based on the statistical analysis of overall disyllables, because the assumption of homogeneity of variance was not met for the data, F(11, 36) = 11.80, p<.05, we used Welch’s adjusted F ratio and reported it as Welch’s F(11, 13.77) = 24.75, p<.05. This shows that the results of children’s overall disyllables are significant. The analysis using the Welch test shows that there were significant differences among the twelve syllable types in disyllabic words. Moreover, the results of the follow-up in Duncan’s multiple range test is shown in Table 4.6
Table 4.6 Duncan's Multiple Range Test for overall disyllables
Duncan grouping Mean N Disyllables
A 957.50 4 CV
According to Table 4.3 as seen above, group A (i.e. CV) is significantly different (p<.05) from the B and C groups. Group B (i.e. CVG) was also significantly different
‧
(p<.05) from the other syllable types. It can thus be obviously shown that CV and CVG were superior to the other syllable types, and statistically significantly different.
The tokens and percentages of each syllable type in disyllabic words are displayed in the following Table 4.7 and Figure 4.2. The bar graph in Figure 4.2 clearly compares the differences among syllable type frequencies.
Table 4.7 Tokens and percentages of overall syllable types in disyllabic words
Type V CV VG GV VN CVG CVN CGV GVG GVN CGVG CGVN
Figure 4.2 Percentages of overall syllable types in disyllabic words
As shown in Table 4.7 and Figure 4.2, we see that CV is the most frequently used syllable type in disyllabic words, accounting for 46.7%, and is then followed by CVG.
5.9%
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
The frequencies of syllable types in disyllabic words were similar to the results in monosyllabic words, where CV, CVG, and CGV are also the most frequently used syllable types. However, as for the least frequently used syllable types, VG and VN, they make up less than 1% of disyllabic words. The frequency ranking is thus CV >
CVG > CGV.
In addition, in order to determine whether there were individual differences among the children, the following Table 4.8 revealed the frequencies of overall disyllable types produced by each child.
Table 4.8 Percentages of overall syllable types of each child in disyllabic words
V CV VG GV VN CVG CVN CGV GVG GVN CGVG CGVN
Child #1 7.9% 36.6% 1.3% 3.8% 1.9% 12.2% 9.6% 6.3% 4.4% 2.4% 6.6% 6.9%
Child #2 5.0% 45.9% 0.4% 6.1% 1.1% 13.0% 6.5% 9.9% 2.9% 1.2% 3.6% 4.4%
Child #3 4.1% 53.6% 0.5% 2.3% 0.3% 13.4% 5.4% 7.3% 3.9% 1.3% 3.4% 4.4%
Child #4 8.6% 47.6% 1.1% 3.8% 0.5% 11.0% 1.5% 14.6% 4.4% 0.1% 5.9% 1.0%
Based on Table 4.8, the most frequently used syllable type, CV, had the highest frequency of occurrences of disyllabic words. Child #2 and child #3 show similar patterns: CV was the most frequently used type, followed by CVG and CGV.
However, child #1 and child #4 show slightly different syllable patterns and order.
Child #1 preferred CVG and CVN, with the usage of 12.2% and 9.6% respectively.
Child #4 instead preferred CGV and CVG, which accounted for 14.6% and 11.0%
respectively.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
In a nutshell, CV, CVG and CGV had higher frequencies than other syllable types in the disyllables of all children. For the least frequently used syllable types, although there were slightly different patterns, VG, VN and GVN had lower frequencies in the sample of the four children’s disyllabic words.
4.2.1.3 Overall multisyllabic words
In terms of the statistical analysis of multisyllables, the assumption of homogeneity of variance was not met for the data, F(11, 36) = 22.15, p<.05, so we used Welch’s adjusted F ratio and reported it as Welch’s F(11, 13.72) = 18.73, p<.05.
It is indicated that the results of children’s overall multisyllables are significant. The analysis using the Welch test shows that there were significant differences among the twelve syllable types in disyllabic words. Results of the follow-up using Duncan’s multiple range test are shown in Table 4.9.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table 4.9 Duncan's Multiple Range Test for overall multisyllables
Duncan grouping Mean N Multisyllables
A 771.00 4 CV
B 200.25 4 CVG
C 123.50 4 CVN
C 115.25 4 CGV
C 107.50 4 V
C 91.50 4 CGVN
C 66.00 4 CGVG
C 59.25 4 GV
C 52.75 4 GVG
C 29.50 4 GVN
C 14.25 4 VG
C 10.50 4 VN
According to Table 4.9, group A (i.e. CV) is significantly different (p<.05) from the B and C groups. Group B (i.e. CVG) is also significantly different (p<.05) from the other syllable types. Hence, we can see that CV and CVG are superior to the other syllable types, and statistically significantly different.
The tokens and percentages of each syllable type in disyllabic words are displayed in the following Table 4.7 and Figure 4.3. The bar graph in Figure 4.3 clearly compares the differences among syllable type frequencies in multisyllabic words.
‧
Table 4.7 Tokens and percentages of overall syllable types in multisyllabic words
Type V CV VG GV VN CVG CVN CGV GVG GVN CGVG CGVN
Figure 4.3 Percentages of overall syllable types in multisyllabic words
In Table 4.7 and Figure 4.3, it can be clearly seen that CV is the most frequently used syllable type in multisyllabic words, accounting for 47.0% of the total 6565 syllables, and followed by CVG. The frequency of CV is three times higher than other syllable types. The frequencies of syllable types in multisyllabic words are slightly different from the results presenting monosyllabic words and disyllabic words. However, for the least frequently used syllable types, VG and VN have the lowest frequencies of all multisyllables. The frequency ranking is thus CV > CVG > CVN > GGV.
6.5%
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Furthermore, in order to see whether there were individual differences among children, the frequency of overall multisyllable types produced by each child is shown in the following Table 4.10.
Table 4.10 Percentages of overall syllable types of each child in multisyllabic words
V CV VG GV VN CVG CVN CGV GVG GVN CGVG CGVN
Child #1 9.0% 38.4% 1.1% 3.1% 1.1% 14.4% 9.7% 3.8% 4.0% 2.0% 3.7% 9.6%
Child #2 4.3% 44.5% 0.5% 4.2% 0.5% 13.2% 9.5% 8.7% 2.9% 1.6% 3.6% 6.3%
Child #3 6.1% 50.9% 0.7% 2.4% 0.3% 11.7% 7.1% 6.2% 2.8% 2.7% 4.0% 5.1%
Child #4 9.4% 53.2% 1.6% 5.2% 1.0% 8.9% 2.0% 8.6% 3.8% 0.1% 5.3% 0.8%
Based on Table 4.10, the most frequently used syllable type, CV, has the highest frequency of occurrences of multisyllabic words. Child #4 showed slightly different syllable patterns and order from the other children. He preferred V and CVG, with a usage of 9.4% and 8.9% respectively. The least frequent syllable types produced by child #4 in multisyllables were GVN, CGVN, VN and VG. The other children showed similar patterns: CV is the most frequently used type, followed by CVG and CVN.
In brief, CV and CVG have higher frequencies than other syllable types in multisyllables of all children. For the least frequently used syllable types, although there were slightly different patterns, VG and VN have lower frequencies among the four children’s multisyllabic words. Furthermore, CV is the most frequently used syllable type in the young children’s speech production. The unmarked syllable CV was obviously the most frequently used type for four children.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
4.2.2 Frequency of syllable deletion
The results in the previous section 4.2.1 show the rankings of the frequently used syllable types used by the four children. In order to evaluate the frequency of deleted syllables, we chose the syllables which had all gone through syllable deletion.
Meaningful words that did not have reduced syllable forms are not included. Besides, since the form of syllable deletion is to omit the segments in the syllable structure, and CGVG and CGVN are two most complex syllable structures in Mandarin, these two types do not occur in the output of syllable deletion forms. As previously noted, the analysis of syllable tokens includes monosyllabic words, disyllabic words and multisyllabic words. To discuss whether different types of tokens influence the children’s syllable deletion patterns, monosyllables, disyllables and multisyllables are analyzed. The frequency of monosyllabic words, disyllabic words and multisyllabic words is shown in section 4.2.2.1, section 4.2.2.2 and section 4.2.2.3 respectively.
4.2.2.1 Deleted monosyllabic words
In terms of statistical analysis of deleted monosyllables, because the assumption of homogeneity of variance was not met for the data, F(9, 30) = 19.82, p<.05, we used Welch’s adjusted F ratio and reported it as Welch’s F(9, 11.92) = 8.68, p<.05. The analysis using the Welch test shows that there are significant differences among the twelve syllable types in deleted monosyllabic words. Additionally, results of the
‧
follow-up with Duncan’s multiple range test are shown in Table 4.11.
Table 4.11 Duncan's Multiple Range Test for deleted monosyllables
Duncan grouping Mean N Deleted
According to the above Table 4.11, group A (i.e. CV) is significantly different (p<.05) from B group. Nevertheless, the other eleven syllable types which are under the same group B show no significant effect. We can tell that CV is superior to the other syllable types, and statistically significantly different while the other syllable types are not significantly different.
The tokens and percentages of each syllable type in monosyllabic words are presented in the following Table 4.12 and Figure 4.4. The bar graph in Figure 4.4 clearly compares the differences among frequencies of deleted syllable types.
‧
Table 4.12 Tokens and percentages of deleted syllable types in monosyllabic words
Type V CV VG GV VN CVG CVN CGV GVG GVN CGVG CGVN
Figure 4.4 Percentages of deleted syllable types in monosyllabic words
Based on the data contained in Table 4.12 and Figure 4.4, the most frequently used syllable type, CV, still has the highest frequency of syllable deleted forms in monosyllabic words, reaching 46.0%. CGV and GV rank second and third, with a usage of 18.8% and 10.9% respectively. As for the least frequently used syllable types, GVG and GVN, they make up than 2% of all monosyllables. The frequency ranking is thus CV > CGV > GV.
It is clearly shown that CV is the most frequently used syllable type in the children’s deleted syllable types and overall syllable types; however, the second place
6.5%
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
and the third place of each ranking are different. That is, the frequency ranking is CV
> CGV > GV for syllable deletion while the ranking is CV > CVG > CGV for overall syllable production.
In order to examine whether there were individual differences among the four children, the frequencies of syllable deletion types in monosyllabic words produced by each child are shown in the following Table 4.13.
Table 4.13 Percentages of syllable deletion types of each child in monosyllabic words
V CV VG GV VN CVG CVN CGV GVG GVN CGVG CGVN
Child #1 16.7% 25.0% 4.2% 12.5% 10.4% 10.4% 4.2% 8.3% 4.2% 4.2% 0.0% 0.0%
Child #2 5.9% 52.9% 4.2% 10.1% 0.8% 4.2% 2.5% 18.5% 0.0% 0.8% 0.0% 0.0%
Child #3 5.0% 50.0% 0.0% 5.0% 0.0% 20.0% 5.0% 15.0% 0.0% 0.0% 0.0% 0.0%
Child #4 2.2% 47.2% 2.2% 12.4% 1.1% 5.6% 2.2% 25.8% 1.1% 0.0% 0.0% 0.0%
According to Table 4.13, CV is definitely the most frequently used type of all children’s deleted forms in monosyllables. Child #1 and child #3 show different syllable patterns compared to the other two children. Firstly, the CV frequencies of child #1 are comparatively lower than those of other children’s, only reaching 25.0%.
Besides, child #1 preferred V and GV, with a usage of 16.7% and 12.5%. For the least frequently used type, in addition to GVG and GVN, VG and CVN still have the lower frequencies of her syllable deletion pattern. Secondly, child #3 preferred the usages of CVG and CGV, reaching 20.0% and 15.0% respectively. He did not produce VG, VN,
Besides, child #1 preferred V and GV, with a usage of 16.7% and 12.5%. For the least frequently used type, in addition to GVG and GVN, VG and CVN still have the lower frequencies of her syllable deletion pattern. Secondly, child #3 preferred the usages of CVG and CGV, reaching 20.0% and 15.0% respectively. He did not produce VG, VN,