• 沒有找到結果。

Chapter 3 Methodology

3.2 Data analysis

3.2.3 Glide emergence and stabilization assessment

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

58

regard to the meaning and usage of the adult forms. In addition, onomatopoeia and

conventionalized vocalizations that do not resemble any appropriate adult forms but

are repeatedly used for certain communicative function by the child, referred to as

protoword in Menn’s (1976), were excluded from the present study. The fuzzy sounds

and those integrated with the background noises were also excluded (Sosa and

Stoel-Gammon, 2012).

3.2.3 Glide emergence and stabilization assessment

Considering the normally-developing children, in response to the research

question one, the emergence of the three glides [w, j, ɥ] in terms of various positions,

including word-initial, word-medial, and word-final position, were identified,

irrespective of the target, so long as the sounds were produced phonetically accurately

for the first time. For the [ɥ] sound, only the word-initial and word-medial positions

were noted since the [ɥ] sound is not allowed to appear in the final position. The

emergence of each glide in relation to positions was later compared.

Previous research has indicated high inconsistency in children’s speech

production and fluctuation in the developmental progress (Ferguson and Farwell,

‧ 國

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

59

1975; Zhu, 2002). For the purpose of determining the age of stabilization of the glides

development, the criteria for stabilization should be decided. The present study

adopted the criteria from Zhu (2002):

1. Its accuracy rating in the spontaneous speech sample reached 66.7% level. The

formula for computing Percentage of Consonants Correct (PCC) and Percentage

of Consonants Error (PCE) based on Shriberg & Kwiatkowski (1982) is provided

below.

the number of times of a glide produced correctly

PCC = × 100%

the number of opportunities for the glide in the sample

the number of times of a glide produced incorrectly

PCE = × 100%

the number of opportunities for the glide in the sample

2. In order to minimize the fluctuation and regression of children’s development, its

accuracy rating in all the subsequent speech samples should remain higher than

66.7%. If the child’s accuracy rate reaches 66.7% but drops under 66.7% in the

following months, we do not consider it as stabilized.

The criteria were employed to calculate the accuracy rating across the three glides and

‧ 國

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

60

within the same glide in various positions of the normative data as well as the

phonologically-disordered data. Furthermore, the accuracy rate of the data in each

position was tested with statistical method, two-proportion z-test in order to see

whether the claim that the glides in the initial position are more stable than the other

two positions.

For the purpose of comparing the tendencies and error types of the two groups,

the error types were first identified and were calculated by the formula presented

below:

the number of times of a phonological process used in the given position

× 100%

the total glide errors identified in a given position

The percentage of each phonological process was compared within the same group

and across the two groups after the calculation.

‧ 國

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

61

Chapter 4

Findings and analysis

In this section, the emergence of glides in the Mandarin in relation to various

positions of normally-developing children will be presented in 4.1. Section 4.2

discusses the stabilization of each glide based on the PCC/PCE formula and describe

the order of stabilization of the three glides of the normally-developing children.

Section 4.3 shows the performance of glides from the phonologically-disordered

group. Finally, in section 4.4, the phonological processes involved in the production

of glides in both groups of children, together with the percentages of each process are

reported.

4.1 Emergence of glides in normally-developing group

The speechlike sounds change drastically during the first year. Vowels seem to

predominate the production of the first six months and the sound repertoire expands

considerably between 6 months to 12 months of age across languages (Gleason and

Ratner, 2009). Previous research has discovered that the sounds produced in

children’s late prelinguistic period are highly identical across languages and might

‧ 國

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

62

later be served as the building blocks for the production of words (Stoel-Gammon,

1985).

The first question addressed in the present study is concerned with the three

glides emergence in Taiwan Mandarin in relation to the various positions, including

initial position, medial position (referring to as prenuclear glide), and final position

(referring to as postnuclear glide). The results of the age of emergence of glides from

the two normally-developing children are displayed in Table 4.1 and Table 4.2.

Table 4.1 shows the age of emergence of the Mandarin three glides in terms of

the three positions regardless of the adult target forms of the normally-developing

child WW (0;10-2;5). That is to say, the sounds were noted down as long as the

sounds were accurately produced for the first time.

Table 4.1 Age of emergence of glides

The postnuclear palatal [j] first emerged at 1;0, followed by initial labiovelar [w] at

1;1. Both the initial [j] and prenuclear [j], together with postnuclear [w] appeared at

the age of 1;2. Prenuclear [w] emerged at 1;4. However, the high front rounded [ɥ]

WW 0;10 0;11 1;0 1;1 1;2 1;3 1;4 1;5 1;6 1;7 1;8 1;9 1;10

Initial [w] [j] [ɥ]

Prenuclear [j] [w] [ɥ]

Postnuclear [j] [w]

‧ 國

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

63

did not occur in WW’s system until 1;8 and 1;9, with the prenuclear [ɥ] emerged in

1;8 and initial [ɥ] in 1;9.

Table 4.2 Age of emergence of glides (NN)

Table 4.2 presents the age of emergence of glides of NN. The emergence of

NN’s labiovelar [w] and palatal [j] is earlier than those in WW’s data. The palatal [j]

in three positions occurred at 0;9, along with the initial [w] and postnuclear [w]. The

prenuclear [w] appeared one month later than the other positions of [w].

Unsurprisingly, the labiopalatal [ɥ] emerged later, with the initial [ɥ] occurred at age

1;5, and prenuclear [ɥ] occurred at age 1;8. Since the palatal [j] in all positions,

together with the initial labiovelar [w] and postnulcear [w] emerged at the first month

of the data collection, the exact age of emergence of the palatal [j] and the labiovelar

[w] in the initial and postnulcear position is unknown. We could only claim that the [j]

and [w] both occur at a relatively young age due to the limitation of the data.

If we overlook the positional factors, the emergence order of the three glides in

Mandarin of the two normally-developing children is as follows:

NN 0;9 0;10 0;11 1;0 1;1 1;2 1;3 1;4 1;5 1;6 1;7 1;8

Initial [w] [j] [ɥ]

Prenuclear [j] [w] [ɥ]

Postnuclear [w] [j]

‧ 國

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

64

(c) [j]  [w]  [ɥ] (WW)

(d) [j] / [w]  [ɥ] (NN)

(c) provides the emergence order of the child WW and (d) provides the emergence

order of the child NN. For the child WW, the palatal [j] occurred first in her

production, followed by the labiovelar [w]. The labiopalatal [ɥ] is the last to emerge.

On the other hand, the child NN produced her first [j] and [w] at the same age, and

similar to WW’s data, the labiopalatal [ɥ] is the last to emerge in her system.

4.2 Stabilization of glides in normally-developing group

In this section, we deal with the stabilization of glides in normally-developing

group. Section 4.2.1 displays the data and PCC/PCE throughout 17 months of data

collected in this study. In section 4.2.2, the order of stabilization of glides is listed.

4.2.1 Data and PCC/PCE

The data of normally-developing child, WW and NN, were analyzed through 17

months. WW was observed from age 0;10 to 2;5 and NN was observed from age 0;9

to 2;4. The stabilization of the Mandarin three glides [w, j, ɥ] was examined in terms

of the three positions. Table 4.3 lays out the total tokens of the glide production of the

normally-developing group. The leftmost column includes the correct number of the

sound produced, the error number produced, the total frequency of glides occurred in

the normally-developing group and the PCE/PCC of glides. The three shaded columns

in the middle of the table represent the subtotal of each glide, and the rightmost

column shows the total number of the three glides.

Table 4.3 Glide distribution and PCC/PCE of the normally-developing group

[w] [w] Notes: I, M, and F refer to word-initial, -medial, and –final positions. The glides

appeared in M are prenuclear glides, and those in F are postnuclear glides. For an easier way of presenting them, M and F are used instead.

In total, 5421 glides were collected from the two children, in which 809 glides were

produced with mistakes and the other 4612 glides were accurately produced. The

distribution of [w], [j], and [ɥ] is 2652, 2615, and 154. The data show that the

labiovelar [w] and labiovelar [j] are distributed almost evenly in children’s early

‧ 國

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

66

utterances; however, the frequency of the labiopalatal [ɥ] is far lower than the other

two glides. The 2652 labiovelar [w] are composed of 335 in the initial position, 641 in

the prenuclear position, and 1676 in the postnuclear position. The 2615 palatal [j] are

composed of 527 initial [j], 1159 prenuclear [j], and 929 in the postnuclear position.

The 154 labiopalatal [ɥ] are composed of 51 in the initial position, and 103 in the

prenuclear position. The PCC of the [w], [j], [ɥ] in I, M, F position is as follows: 95.8%,

80.2%, 80.2%; 95.6%, 92.9%, 81.8%; 90.2%, 44.7%. Two-proportion z-test showed

that the PCC of the three glides in initial position are significantly higher than the

prenuclear and postnuclear positions ([w] I M : Z=6.59, P<.001***; [w] I F: Z=6.31,

P<.001***; [j] I M : Z=2.13, P<.05*; [j] I F : Z=7.49, P<.001***; [ɥ] I M : Z=5.42,

P<.001***).

Table 4.4 and Table 4.5 show the distribution of the three glides and Percentage of

Consonants Correct (PCC) and Percentage of Consonants Error (PCE) of the subgroups

WW and NN.

Table 4.4 Total distribution and PCC/PCE of glides (age 0;10-2;5) of WW 0;10-2;5

initial [w], 252 of which are prenuclear [w], and 597 are postnuclear [w]. The PCC of

initial, prenuclear, and postnuclear [w] are 89.1%, 73.8%, 90.8% respectively. The

965 [j] sounds are composed of 195 [j] in the initial position, 420 in the prenuclear

position, and 350 in the postnuclear position. The PCC of the three positions are

97.9%, 92.9%, and 84.6%. The frequency of [ɥ] is relatively low. Only 59 syllabic

words produced by the child contain the sound, with 13 appear in the initial position

and 46 appear in the prenuclear position. The PCC of the initial [ɥ] is 92.3% and

prenuclear [ɥ] is 43.5%.

Table 4.5 Total distribution and PCC/PCE of glides (age 0;9-2;4) of NN 0;9-2;4

distribution includes 332 in the initial position, 739 in the prenuclear position, and 579

in the postnuclear position. The distribution of the high front rounded [ɥ] includes 38 in

the initial position and 57 in the prenuclear position. To compare the PCC of various

positions within the same phones, we found that both of the children demonstrated a

higher accuracy rate in the initial position.

In order to investigate the developmental progress of the three glides in relation to

positions, the glides are categorized in terms of the positions they appear at each age

and the error production were calculated for the PCE and PCC. WW did not produce

therefore, Table 4.6 to Table 4.22 display the distribution and PCC/PCE of WW from

age 1;1 to age 2;5.

Table 4.6 Distribution and PCC/PCE of glides (age 1;1) WW

1;1 [w] [w]

Table 4.7 Distribution and PCC/PCE of glides (age 1;2) WW 1;2

Table 4.8Distribution and PCC/PCE of glides (age 1;3) WW

Table 4.9 Distribution and PCC/PCE of glides (age 1;4) WW 1;4

From table 4.6 to table 4.9 (1;1-1;4), little data involving glide was uttered by WW. At

age 1;1, only one prenuclear [w] was produced, and it was produced accurately. At

age 1;2, only one postnuclear [j] appeared in the data, and it was accurately produced.

The data are still limited at age 1;3 and 1;4. The child produced two [w] sounds in the

initial position and two postnuclear glides at age 1;3; the two [w] sounds were

produced. At the age of 1;4, one postnuclear [w] was not accurately produced whereas

the initial [j] was correctly produced. So far, the front rounded glide [ɥ] hasn’t

occurred in the child’s utterance.

The production of the child had an obvious surge since the age of 1;5. The data that

contains glides before age 1;4 are limited to less than four but by the age 1;5, the data

has increased to more than 50.

Table 4.10 Distribution and PCC/PCE of glides (age 1;5) WW

1;5 [w] [w]

Table 4.11 Distribution and PCC/PCE of glides (age 1;6) WW 1;6 Table 4.12 Distribution and PCC/PCE of glides (age 1;7) WW

1;7 [w] [w] Table 4.13 Distribution and PCC/PCE of glides (age 1;8) WW

1;8 [w] [w]

Table 4.14 Distribution and PCC/PCE of glides (age 1;9) WW 1;9

Table 4.15 Distribution and PCC/PCE of glides (age 1;10) WW

1;10 [w] [w]

Table 4.16 Distribution and PCC/PCE of glides (age 1;11) WW 1;11

Table 4.17 Distribution and PCC/PCE of glides (age 2;0) WW

Table 4.18 Distribution and PCC/PCE of glides (age 2;1) WW

Table 4.19 Distribution and PCC/PCE of glides (age 2;2) WW 2;0

Table 4.20 Distribution and PCC/PCE of glides (age 2;3) WW

Table 4.21 Distribution and PCC/PCE of glides (age 2;4) WW

Table 4.22 Distribution and PCC/PCE of glides (age 2;5) WW 2;3

‧ 國

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

76

At age 1;5, the child produced 29 [w] sounds with two errors in the medial position,

with the other positions all accurately produced, and 37 [j] sounds were produced

with 10 errors, nine of which were postnuclear [j] and the other one was the initial

position [j]. The prenuclear [j] was produced without any error. At 1;6, the postnuclear

[j] was stabilized. At 1;7, the first target form containing postnuclear [ɥ] sound

appeared, but was produced inaccurately. Few errors were detected at this age. Aside

from the postnuclear [ɥ], only one postnuclear [w] was inaccurately produced.

Another data explosion occurred at age 1;8, at which the data containing glides

reached above 200. The distribution of the three glides [w, j, ɥ] of this age are 92, 98,

and 12. The error numbers of the production for the prenuclear [w] and postnuclear

[w] are 16 and 1 respectively, and for the prenuclear [j] and postnuclear [j] are 7 and

11 respectively. As for the prenuclear [ɥ], two data were produced with errors. At age

1;9, the initial [ɥ] sound appeared for the first time in the WW’s system and all of them

were accurately produced; however, the 10 [ɥ] sounds in the prenuclear position were

all produced with errors with only one accurately produced. The distribution of the [w]

sounds and [j] sounds are 129 and 109 respectively, with the [w] composed of 7 initial

‧ 國

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

77

[w], 60 prenulear [w], and 62 postnuclear [w] and the [j] composed of 25 initial [j], 45

prenuclear [j], and 39 postnuclear [j]. The distribution of the three glides [w, j, ɥ] at the

age of 1;10 is 83, 67, and 6. The error numbers for the [w, j, ɥ] sounds in each position

are as follows: one out of five for the initial [w], 10 out of 23 for the prenuclear [w], two

out of 55 for the postnuclear [w], zero out of 8 for the initial [j], two out of 28 for

prenuclear [j], 3 out of 31 for postnuclear [j], and three out of six for prenulcear [ɥ]. At

age 2;0, the distribution of the three glides [w, j, ɥ] is 65, 91, 5 respectively. The target

numbers of [w] in initial, medial, and final position are 5, 16, 44, in which 5 errors were

produced in the medial position and 2 errors were produced in the final position. The

target numbers of [j] in initial, medial, and final position are 14, 44, 33, in which one

error was produced in the initial position, one error in the medial position, and two

errors were produced in the final position. The child did not produce any initial [ɥ] at

this age and the error production of the prenuclear [ɥ] is two out of five. From the data

above, we noticed that the child produced the least data in the initial position;

nonetheless, they were produced with the least errors with regard to the three positions.

Moreover, the data of the [ɥ] sound were the least among the three glides.

with possible meanings at a relatively young age but the number of tokens is low.

Table 4.23 Distribution and PCC/PCE of glides (age 0;9) NN 0;9 Table 4.24 Distribution and PCC/PCE of glides (age 0;10) NN

0;10 [w] [w]

Table 4.25 Distribution and PCC/PCE of glides (age 0;11) NN 0;11 Table 4.26 Distribution and PCC/PCE of glides (age 1;0) NN

1;0 [w] [w] Table 4.27 Distribution and PCC/PCE of glides (age 1;1) NN

1;1 [w] [w]

Table 4.28 Distribution and PCC/PCE of glides (age 1;2) NN 1;2 Table 4.29 Distribution and PCC/PCE of glides (age 1;3) NN

Table 4.30 Distribution and PCC/PCE of glides (age 1;4) NN 1;4

Table 4.31 Distribution and PCC/PCE of glides (age 1;5) NN 1;5 Table 4.32 Distribution and PCC/PCE of glides (age 1;6) NN

1;6 Table 4.33 Distribution and PCC/PCE of glides (age 1;7) NN

1;7 [w] [w]

Table 4.34 Distribution and PCC/PCE of glides (age 1;8) NN 1;8 Table 4.35 Distribution and PCC/PCE of glides (age 1;9) NN

1;9 [w] [w]

Table 4.36 Distribution and PCC/PCE of glides (age 1;10) NN

1;10 [w] [w]

Table 4.37 Distribution and PCC/PCE of glides (age 1;11) NN 1;11 Table 4.38 Distribution and PCC/PCE of glides (age 2;0) NN

Table 4.39 Distribution and PCC/PCE of glides (age 2;1) NN 2;0

Table 4.40 Distribution and PCC/PCE of glides (age 2;2) NN

Table 4.41 Distribution and PCC/PCE of glides (age 2;3) NN

Table 4.42 Distribution and PCC/PCE of glides (age 2;4) NN

2;2 [w] [w]

‧ 國

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

85

Similar to WW’s data, the glide production of NN’s data had an evident surge at

age 1;5. The total production of glides is 104, 63 of which are [w], 38 of which are [j],

and three of which are [ɥ]. The 63 labiovelar [w] are composed of two in the initial

position, 10 in the prenuclear position, 51 in the postnuclear position, and the 38 palatal

[j] sounds are composed of 14 in the initial position, 19 in the prenuclear position, and

five in the postnuclear position. The three rounded palatal [ɥ] are composed of one in

the initial position and two in the postnuclear position. Another burst of data occurred

at age 1;11, when the total production of glides reached 347. The distribution of the

three glides [w, j, ɥ] is 166, 172, 9. Table 4.38 to Table 4.42 reports the performance of

glides from 2;0 to 2;4. The PCC of each glide in the three positions reaches the criteria

of stabilization except for the prenuclear palatal [j] at 2;1. Moreover, we found that in

all the data observed, the production of the labiovelar [w] and the palatal [j] are similar

to each other while the number of labiopalatal [ɥ] is much lower than the other two

glides. This was found in both WW and NN’s data.

4.2.2 Order of stabilization of glides

The criteria for determining whether the glides are stabilized or not was adopted

‧ 國

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

86

from Zhu (2002), in which the accuracy rating of the spontaneous sample is

considered stabilized when it reached 66.7%. Besides, for the avoidance of the

fluctuation and regression of the developmental progress, the accuracy rate in all the

subsequent speech samples should not drop under 66.7%. In table 4.5, WW only

produced one prenuclear glide [w]. Although the PCC is 100%, we do not consider it

to be stabilized since the PCC dropped under 66.7% at the age of 1;5, 1;8, and 1;10;

as a result, the stabilization of prenuclear [w] is 1;11. The first production of initial [w]

occurred at age 1;3. The PCC is 100% and has remained above 66.7% since then, so

1;3 is regarded as the age of stabilization of initial [w]. The postnuclear [w] is

stabilized at the age of 1;5. Following the same way of analysis, the initial [j] and

prenuclear [j] are stabilized by the time the first production with possible meaning

emerged, which are 1;4 and 1;5 respectively. The postnuclear [j] has undergone

regression and is not stabilized until 1;6. For the late-emerging sound [ɥ], WW did not

produce any syllabic word containing it with possible meaning until 1;9; however, all

six syllabic words are produced without any mistakes, and the PCC is 100%; hence, 1;9

is thought of as the age of stabilization, while the prenuclear [ɥ] is not stabilized within

‧ 國

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

87

the 17-month-data. The PCC of 2;1 is 33.3%, below the criteria of stabilization. WW

did not produce any form containing prenuclear [ɥ] at 2;2 to 2;5, therefore, the

performance of the prenuclear [ɥ] remains unclear. The shaded three columns

demonstrated the distribution and accuracy rate of each glide without distinguishing the

positions. The [j] sound is stabilized at the age of 1;4 and [w] sound at 1;5. The [ɥ]

sound is stabilized at age 2;2. In other words, the order of the stabilization of the three

glides is [j]  [w] [ɥ]. Palatal [j] is the first to stabilize, followed by the labiovelar

[w], and the labiopalatal [ɥ].

On the other hand, the child NN produced the target forms containing glides at the

earliest data collected at age 0;9. Without the consideration of positions, the three

glides [w], [j], [ɥ] are stabilized at 1;11, 1;8, and 1;11. The stabilization of the

labiovelar [w] in the initial position, prenuclear position, and postnuclear position are at

age 1;5, 1;9, 1;11, and the stabilization of the palatal [j] in the initial position,

prenuclear position and prenuclear position are at age 1;0, 1;5, 2;2. As for the less

produced labiopalatal [ɥ], the stabilization of the initial position and the prenuclear

position is at age 1;9 and 2;0.

‧ 國

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

88

The developmental process of the Mandarin three glides [w, j. ɥ] throughout 17

months of the two normally-developing children WW and NN are presented in Figure

5.1 and Figure 5.2. The vertical axis represents the PCC and the horizontal axis

represents the age. The blue line, orange line, and red line represent [w], [j] and [ɥ]

respectively.

Figure 5.1 illustrates the stability trend of the Mandarin three glides in WW’s

data from age 0;10 to 2;5, from which we found that the labiovelar [w] and palatal [j]

demonstrate higher stability than the labiopalatal [ɥ]. Moreover, the palatal [j] seems

to be even more stable than the labiovelar [w] since the orange line is above the blue

line throughout the developmental process.

‧ 國

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

89

Figure 5.1 Glide development of WW

Figure 5.2 Glide development of NN

‧ 國

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

90

Figure 5.2 shows the glide the stability trend of the Mandarin three glides in

NN’s data from age 0;9 to 2;4. Similar to WW’s stability trend, the palatal [j]

performs higher stability than labiovelar [w] and labiopalatal [ɥ]. Nonetheless, to

compare Figure 5.1 and Figure 5.2, we discovered that NN’s data underwent greater

fluctuation.

The order of the stabilization of the three glides can be drawn from the analysis

above, Table 4.43 and Table 4.44 present the age of stabilization of each glide in terms

above, Table 4.43 and Table 4.44 present the age of stabilization of each glide in terms