Chapter 3 Methodology
3.2 Data analysis
3.2.4 Syllable type frequency
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
The variability rate could help us understand how prevalent variability is in children‟s speech. By estimating the variability rate we could know whether a certain
type of syllable is mastered by the children or not.
3.2.4 Syllable type frequency
Evaluating the frequency of different syllable types in children‟s speech could reveal children‟s preference and show the mastery of each syllable types. If the
frequency of a certain syllable type is low, it might be explained that the syllable type is more problematic to children and has not been acquired yet. For example, if the frequency of CGVN (C = consonant, G = glide, V = vowel, N = nasal) was observed to be lower than other syllable types when children were one year old, it would imply that the children have not acquired syllable type CGVN yet and they might use other syllable type to replace it, such as CV or CGV, resulting in production variability and the low frequency of CGVN. So we could predict that syllable types with lower frequency would have higher variability rate.
The frequency of each syllable type children produced would be computed by the formula presents below.
Tokens of a syllable type
Percentage = X100%
Tokens of all syllable type
The numerators would be the tokens of a particular syllable type. The fraction would be presented as a percentage. The frequencies of all syllable types could be
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
ranked into an ordering. The ordering of syllables‟ frequencies could be used to
compare to those in variability rates.
3.2.5 Substitution pattern of syllable type
The multiple forms children produced for a target word would be further analyzed.
Since children‟s productions of words were not mature and stable, children might use
many different phonetic forms to produce a word. The evaluation of substitution pattern could demonstrate which syllable types were more likely to be used to replace immature types, and which syllable types tended to make more error.
Note that there are some syllable omission and syllable contraction patterns in the observation data: 0.6% of the data are syllable omission and 0.4% of the data are syllable contraction. Since the main purpose of this study is to look at substitution types, the patterns involving omission and contraction would not be included.
The substation pattern would be shown in a matrix, in which the row represents the syllable types of target words, and the column represents the actual produced syllable types. The example of the matrix is presented in Table 3.5.
The matrix presents the token and percentage of the substitution between different syllable types. The percentage on the rightmost column represents the frequencies of syllable type which are used to replace others, and the percentage on the bottom row represents which syllable type is replaced by others more often. For
‧
example, in table 3.5, 59% of target syllable types are replaced by CV syllable; CVG syllable are replaced by others more often since CVG syllable has higher percentage than other syllables. Some columns are left blank which mean that children did not use the particular syllable to replace target syllable type.
Table 3.5 The sample matrix of substitution pattern in different syllable types Target
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
47
Chapter 4 Results and Analysis
This section will present the results and overall analysis of the data. The analysis will follow the research questions presented in chapter 1. The overview of the overall data will be presented in section 4.1. Section 4.2 will present the overall pattern of phonological variability pattern of each participant. Section 4.3 will present the results of syllable analysis, including the general syllable type children are likely to produce, the frequency of syllable types in different syllable positions, and the variability rate of different syllable types. Section 4.4 will present the substitution pattern between different syllable types. Last but not least, the relationship between frequencies of syllable types and its variability rates as well as the relationship between frequencies and substitution pattern will be summarized in section 4.5.
4.1 Background information of the data
Two male and four female children were adopted in the study. The observation started at the age of 0;11 to 1;1 in the beginning and ended with 1;10 to 2;0. The children enrolled were all in non-reduplicated babbling stage in which children started to produce different syllable types, and they have already produced their first meaningful word.
‧
have clear meaning could be included in the analysis. Tokens without clear meaning or tokens that were acoustically unrecognizable would not be included. There were in a total number of 5868 tokens. However, utterances which consisted of more than two syllables and utterances showed syllable omission and syllable contraction would further be excluded from this study because not all of the participants producing utterances contained more than two syllables with recognizable meaning during the one-year observation, and only two participants produced syllable omission and contraction utterances (there were 0.6% of syllable omission, 0.4% for syllable contraction, and 0.1% for trisyllables in the data). So, after excluding the above types, there were a total number of 5788 tokens which could be analyzed in this study. Table 4.1 provides the information of participants, including gender, age, the duration of observation, and the total tokens participants uttered.Table 4.1 Information of participants Participants Gender Age
range
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
#6 F 1;1-2;0 12 487
Total 5788
As shown in Table 4.1, the most productive children were participants #1, #2, and #3 who produced more than 1100 tokens, while participants #4, #5, and #6 produced the least number of tokens which were under 700 tokens. Participants #1 and #4 were added for counterbalancing the gender differences.
Although participant #3 started the observation from younger age, she produced much more speech tokens than the other five participants. At the beginning of the
observation, participant #3 produced only one speech sound with referential meaning, which was [mama] „mother‟; however, when she was at the age of 1;3, she has
already produced 105 meaningful words in a one hour recording session. It seemed that children‟s productivity and age are not exactly positively correlated.
In the observation data, words participants produced would be further divided into monosyllabic and disyllabic words. Table 4.2 shows the tokens of each word type children produced.
Table 4.2 Tokens of different word types
Monosyllabic words Disyllabic words Total
#1 362 925 1287
#2 424 774 1198
#3 681 982 1663
#4 117 430 547
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
#5 209 397 606
#6 187 300 487
Total 1980 3808 5788
As shown in Table 4.2, all of participants preferred producing disyllabic words than monosyllabic words. There were 3808 tokens of disyllabic words, and only 1980 tokens of monosyllabic words. Each of disyllabic word has two syllables, so the 3808 disyllabic words would have 7616 syllables. Thus, there were in total of 9596 syllabic tokens (1980+3808x2=9596) in the data.
4.2 Phonological variability
In children‟s early production, they tend to produce a target word with many
different phonetic forms. The variability rates of each participant at each age were calculated in order to see whether there were some patterns in their variable production. Table 4.3 shows the number of different target words and number of words that have variation forms of each participant at each age.
Table 4.3 Number of target words and variation words of participants at each age Participants
Age #1 #2 #3 #4 #5 #6
0;11 N/A N/A 1(0) 7(0) 2(0) 1(0)
1;0 N/A 9(1) 8(2) 15(6) 0(0) 1(0)
1;1 28(9) 12(1) 17(5) 18(5) 3(0) 2(0)
1;2 18(7) 14(4) 13(5) 5(2) 17(0) 8(0)
1;3 22(6) 9(2) 34(11) 3(2) 4(1) 3(0)
1;4 27(6) 22(4) 79(20) 11(4) 7(1) 6(1)
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
1;5 28(7) 49(12) 98(26) 7(1) 19(3) 25(6)
1;6 30(8) 77(10) 114(24) 7(0) 18(1) 30(5)
1;7 35(6) 82(22) 87(27) 20(5) 64(5) 27(1)
1;8 37(7) 66(12) 82(23) 21(6) 41(2) 27(4)
1;9 50(9) 76(16) 63(20) 16(4) 73(12) 48(9) 1;10 95(14) 87(15) 74(18) 33(6) 72(16) 38(5)
1;11 72(11) 86(11) N/A N/A N/A N/A
2;0 63(10) N/A N/A N/A N/A
The number of words that had variation forms presents in parentheses. For example, participant #1 produced 28 different words at 1;1, and among which, nine words showed phonological variability, which means participant #1 used more than one phonetic forms to produce these 9 words. If the number in parentheses was zero, it means that children did not produce any variation forms. For example, participants
#3 only produced one word at the beginning of the observation which was [ma55ma55] „mother‟. She produced this word two times during the observation at
0;11, and did not have any variation form. Based on Table 4.3, all of the participants gradually produced more and more amount of different words as they grew older, but the amount of variation form did not gradually increase.
Variability rate differed considerably both across children and across ages, ranging from 0% (participant #4 at 1; 6) to 45.8% (participant #3 at 1; 2). At the level of the individual word, variability rate could range from 0% to 100%, so a high overall variability rate did not indicate instability in the child‟s production of all
‧
variable. Table 4.4 presents two words with different level of variability rate produced by participant #1. The marker of tone was omitted in the table because the child produced all of the variation forms with the same tone, which is the level tone.Table 4.4 Production of [xwa55] and [tʂhɤ55tʂhɤ55] of participant #1 at 1; 3 Target word IPA transcription Number of Occurrences [xwa55]
Target word IPA transcription Number of Occurrences [tʂhɤ55tʂhɤ55] tɕytɕy 2
Variability 12 forms 12/40x100=30%
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Based on Table 4.4, participant #1 used two different phonetic forms to produce the word [xwa55] „flower‟. The variability rate of [xwa55] „flower‟ reached only
16.7% at the age of 1;3. However, participant #1 used 12 different phonetic forms to produce the word [tʂhɤ55tʂhɤ55] „car‟ also at his 1;3. The variability rate was up to 30%.
The following section will introduce the patterns of overall phonological variability of each participant.
4.2.1 Overall variability pattern
Individual patterns of overall variability are shown for each child at each age in Figure 4.1.
‧
Figure 4.1 Overall variability patterns of each participant
0
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Based on Figure 4.1, although each participant exhibited a somewhat unique pattern of variability, some generalizations were evident. First of all, the overall variability patterns of six participants all showed peaks and valley. None of them showed a decreasing or increasing linearity in their variability patterns. Secondly, the variability rates of most of the participants gradually decreased as they grew older and older. Thirdly, although every participant showed the various forms at different age, they all had a noticeable variability peak. For participants #1, #3, and #4, the variability peaks appeared at 1;0 to 1;2, which was relatively early compared to other participants.
As for differences, firstly, these six participants could be divided into three
groups: participant #1 had slightly different behavior comparing to other participants.
Participants #2 and #3 were “early talkers”. They were productive and uttered their
first meaningful words at earlier age, which was one-year-old for participant #2, and 11-month-old for participant #3. Participants #4, #5, and #6 had much more similar developmental process.
The variability peak of participant #1 appeared at the very beginning of the observation, and the variability rates for participant #1 exhibited a much more regular decreasing pattern than others. By looking carefully into the observational data, we found that participant #1 tended to produce the same target words at every age. For
‧
presents the tokens and variability rates of these three words.
Table 4.5 Number of forms/tokens and variability rates for words of participant #1
1;1 1;2 1;4 1;6 1;9 tɕhjow35] „ball‟, [tʂhɤ55 tʂhɤ55] „car‟, and [kɑw55 kɑw55] „building blocks‟ produced by participant #1 at different age. Different forms referred to the use of more than one phonetic form to produce a word, and if the child used only one phonetic form to produce the word, the number of different forms is 0. The variability rates for these three words gradually decreased as participants #1 grew older, and there were no variable forms for these three words at 1;9, that is, participant #1 used one form consistently to produce the words, resulting in a lower overall variability rates.
For participants #2 and #3, they reached their variability peaks at their 1;7 and 1;2 respectively. It was also the time when these two participants started to produce
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
two-word utterances, such as [aj51 ni21] „love you‟ and [pɑw21pɑw21 na35] „baby takes (something)‟. The onset time of two-word utterances of participants #2 and #3 occurred earlier than previous study suggested by Stoel-Gammon (2006). Although participants #2 has produced two-word utterances at 1;6, almost all of her productions were two-word utterances and sometimes even three-word phrases when she was at 1;7, in which the variability peak appeared,. Participant #3 started to produce two-word utterances as early as she was 1;2. Although she only produced one two-word utterance, which was [ma21ma35 tswɔ35] „mother sits‟, she gradually produced more and more two-word utterances.
Participants #4, #5, and #6 were less productive in our study groups. They did not consider as fast talkers. Participant #4 was productive at the beginning of the observation but the variability rates were also high. The variability peak appeared at the very beginning of the observation. However, the variability rates gradually decreased and even reached 0% at the age of 1;6. At the age of 1;7, the variability rates once again increased when participant #4 entered two-word stage.
The production of participants #5 and #6 were very small and the variability rates were 0% at the beginning. The low variability rates at their early stage did not mean that their productions were stable; instead, the reason was due to the small amount of utterances, as shown in Table 4.3 presented earlier. Participants #5 and #6
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
produced only 0 to 3 words at the beginning. The variability rates were low because they barely uttered more than 2 words in the one-hour observation. The variability peaks appeared at their 1;3 and 1;4 respectively. Although these two participants did not utter any two- word utterance at the end of the observation, they apparently produced much more different words after they went through the variability peaks.
The number of words participant #5 produced increased to 19 words and participant
#6 increased to 25 words.
To sum up, the findings generally confirm the study of Stoel-Gammon and Sosa (2006) which showed that an increase in variability might correspond to instability in the phonological system when it undergoes reorganization. In this study, it is the time when the onset of two-word utterances and the increased amount of words appeared.
4.3 Syllable analysis
In this section, the tokens of different syllable types in different syllable position were calculated in order to see the general syllable types produced by children. The frequency and variability rate of each syllable type would be discussed in this section.
The frequencies of syllable types would inform which syllable type is more frequently used and which type is less used. The variability rates of each syllable type would show which syllable type is not stable.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
4.3.1 Frequency of syllable type
It is mentioned that the tokens collected in this study included monosyllabic words and disyllabic words. To see whether different types of tokens affect the development in syllable acquisition, monosyllables and disyllables would be analyzed separately. During the observation, we found that not all of the participants could produce utterances longer than two syllables, so the tokens analyzed here were all one-to-two syllabic utterances. The frequency of monosyllabic words and disyllabic words will be presented in section 4.3.1.1 and 4.3.1.2 respectively.
4.3.1.1 Monosyllabic words
Table 4.6 and Figure 4.2 provide the tokens of each syllable type in monosyllabic words. The bar graph in Figure 4.2 was used to compare the differences among frequencies of syllable types.
Table 4.6 Tokens and percentages of syllable types in monosyllabic words
V CV VG VN GV CVG CVN GVG GVN CGV CGVG CGVN
Tokens 93 667 27 25 75 549 127 108 26 189 65 29
Percentage 4.7% 33.7% 1.4% 1.3% 3.8% 27.7% 6.4% 5.5% 1.5% 9.5% 3.3% 1.5%
Total tokens 1980
‧
Figure 4.2 Percentages of syllable types in monosyllabic words
The results are statistically significant (χ211,0.001=153.9, p<.001). Based on Figure 4.2, CV was the most frequently used syllable type in monosyllabic words, accounted for 33.7 percent of the total 1980 syllables. CVG and CGV ranked as the second and third place, with the usage of 27.7 percent and 9.5 percent of utterances respectively.
CVN ranked as the fourth place, with the usage of 6.4 percent. As for the least frequent syllable types, VN, GVN, VG had percentage less than 2 percent. Clearly, the findings indicate that children prefer to use CV syllable the most. The frequency ranking according to Figure 4.2 would be CV>CVG>CGV>CVN.
In order to see whether there are individual differences between participants, the frequencies of syllable types produced by each participant were also calculated, as
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table 4.7 Percentages of syllable types of each participant in monosyllables
V VG VN CV CVG CVN CGV GV GVG GVN CGVG CGVN
#1 6.9% 0.5% 1.3% 41.1% 24% 5.5% 7.2% 3.5% 4.4% 1.4% 1.9% 1.9%
#2 8% 1.6% 0.7% 31.6% 29.4% 3.7% 8.5% 6.3% 5.8% 1.4% 1.6% 0.9%
#3 2.4% 1.6% 2.5% 32% 24.4% 9.8% 10.3% 3.5% 5% 1.3% 4.8% 2.2%
#4 3.4% 3.4% 0% 47.8% 19.7% 0.8% 10.3% 0.8% 7.7% 0% 6% 0%
#5 4.3% 1.4% 0% 16.3% 41.6% 7.7% 12.4% 3.3% 8.1% 0.9% 3.3% 0.5%
#6 2.1% 0% 0% 40.1% 32.6% 3.7% 10.2% 1.6% 3.7% 2.1% 2.1% 1.1%
Based on Table 4.7, half of the participants produced all the syllable types in Mandarin. Participants #4, #5, and #6 produced only 9, 11, and 10 syllable types respectively at the end of the observation.
For the most frequently used syllable type, participant #5 presented slightly different pattern comparing to other participants. CVG was the most frequently used syllable types by participant #5, accounted for 41.6% of the utterances, and then followed by CV and CGV, accounted for 16.3% and 12.4% of the utterances. The least frequent syllable type produced by participant #5 in monosyllables was VN, CGVG, and GVN. The other participants showed similar pattern: CV was the most frequently used syllable type, followed by CVG and CGV.
To sum up, CV, CVG, and CGV were used more frequently than other syllable types in monosyllables of all participants. As for the least frequent syllable type, although every participant showed slightly difference, VG, VN, GVN, CGVN had
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
lower frequency in all participants‟ production.
4.3.1.2 Disyllabic words
In this section, syllable types in different syllable position of disyllabic words and the preferred syllable combination were analyzed. The syllable-initial position in the disyllabic words is referred to as 1st syllable and the syllable-final position is represented as 2nd syllable in Table 4.8.
Table 4.8 and Figure 4.3 present the tokens and frequencies of syllable types in disyllabic words. The results in first syllable are statistically significant (χ211,0.001=318.4, p<.001). The results in second syllable are also statistically significant (χ211,0.001=212.2, p<.001).
Table 4.8 Tokens and percentages of syllable types in disyllabic words
1st syllable V CV VG VN GV CVG CVN GVG GVN CGV CGVG CGVN
Tokens 255 2122 11 21 182 473 154 3 19 448 70 50
Percentages 6.7% 55.7% 0.3% 0.6% 4.8% 12.4% 4% 0.07
%
0.5% 11.8% 1.8% 1.3%
Total tokens 3808
2nd syllable V CV VG VN GV CVG CVN GVG GVN CGV CGVG CGVN
Tokens 134 1760 13 15 224 499 173 185 21 513 118 153
Percentages 3.5% 46.2% 0.3% 0.4% 5.8% 13.1% 4.5% 4.9% 0.8% 13.5% 3.1% 4%
Total tokens 3808
‧
Figure 4.3 Percentages of syllable types in disyllabic words
In Figure 4.3, the bars on the left named 1st syllable show all syllable types‟
frequencies of the first syllable in disyllabic words, and those named 2nd syllable in the right represent the frequencies of the second syllable in disyllabic words.
Frequency of each syllable type would be calculated separately as different categories, but they were put together in the same graph for easier comparison.
Figure 4.3 shows that CV, CVG, CGV had the highest frequency in both first syllable and second syllable in disyllabic words, among which CV was the most frequently used syllable types, accounted for 55.7 and 46.2 percent of the production in first and second syllable respectively. The frequency of CV was three times higher than other syllable types. The frequencies of syllable types in disyllabic words were similar to the results in monosyllabic words, in which CV, CVG, and CGV were also the most frequently used syllable types. The ranking of frequencies of syllable types is
6.7%
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
CV> CVG, CGV> GV, V in disyllabic words.
For the least frequent syllable types, VG, VN, and GVN had frequency less than
For the least frequent syllable types, VG, VN, and GVN had frequency less than