Chapter 3 Methodology
3.2 Data analysis
3.2.3 Glide emergence and stabilization assessment
國
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