• 沒有找到結果。

Chapter 2 Literature Review

3.4 Discussion

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

3.4 Discussion

There was no significant difference on the proportion of looks between the tonal

competitor and the unrelated distractor. The mismatch of the first segments may

override the possible effect of the tone in the activation stage. Therefore, subjects

would nearly consider the tonal competitor as the unrelated distractors.

When could tone be accessed during the spoken character recognition? The

divergent time of the curves between target and the segmental competitor could be

extrapolated as the time point when listeners could distinguish the tonal difference of

segmental competitor from target. The result showed that the divergent time of the

segmental competitor and the target was at about 601-700-msec. This divergent time

was earlier than the offset of the target (at about 703msec), which suggests that tone

may begin to affect the spoken character processing before subjects catch the whole

acoustic information.

While the results suggest that tonal information could affect the spoken character

processing before the end of the auditory stream, however, the divergent time of the

curves between the tonal competitor and target was the same among four tones. If

tonal information had impact on the processing, fixation proportions of tonal

competitor should have significant differences in different time bin among four tones.

Different tonal characteristics could not affect listeners’ decision on choosing the

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

target. This result could imply that the tonal information has weak influence on the

spoken character processing.

Additionally, weak tonal effect could be the reason why the curves of segmental

competitor diverged from the target in later time, comparing to the time when the

curves of target diverging from unrelated distractor. Owing to the totally different

syllable structure between target and unrelated distractor, the time when listeners are

able to distinguish two characters could consider to be the earliest time listeners could

identify the tonal and segmental information. It was worth suspecting that tonal

information could affect spoken character processing in an earlier time.

The weak and late tonal effect might be because of the great segmental disparity

between the tonal competitor and target. As the auditory stimuli unfold, the initial

segments of the character of tonal competitor make participants consider it different

from the target and thus fixate less on it in an early time. The great segmental

differences of tonal competitor might be the reason why tone could not take effect

during spoken character processing.

The last issue was the characteristics of tone in processing spoken character. There

was a latest time of divergence between the segmental competitor and the target in

Tone 3, in which of Lai and Zhang (2008) that the last isolation point in a gating

paradigm is in Tone 3. However, in the present study, the order when the curves of

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

competitor and target diverged in Tone 1, Tone 2, and Tone 4 were inconsistent with

the order of Isolation Point in Lai and Zhang (2008).

To sum up, the result of experiment one showed that tone would constrain before

the unfolding of the whole syllable information. However, the totally different

segmental structure between the target and the tonal competitor caused listeners

barely consider the tonal competitor as unrelated distractor. Experiment two examined

the earlier time point when tone would affect lexical processing when the initial

segments were the same.

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

53

Experiment Two

In Experiment 2, the design was similar as experiment one. However, the targets

shared the same initial cohort (two segments) with two types of competitors. One type

of the cohort competitors shared the initial cohort and has the same tone with the

target (cohort-tone competitor, CTC), the other shared the cohort but the tone is

different from targets (cohort-only competitor, COC). For the unrelated distractor,

similar as experiment one, the tonal and segmental structure was totally different from

the target and the competitors.

The time point of divergence between the curves of target and the competitors, and

between the competitors and the unrelated was examined. Thus, if tone affected the

processing in early phase of spoken character recognition, the time point of

divergence between the curves target and the competitor would be earlier in

cohort-only competitor than in cohort-tone competitor. In addition, the proportion of

looks to cohort-only competitor would be lower than that of the cohort-tone

competitor. On the contrary, if the tone information did not affect the spoken character

processing in early time, the curve of the cohort-only competitor would be similar as

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

the cohort-tone competitor. Table 6 illustrates the predictions for the accounts whether

tone is accessed in initial character in Experiment 2.

Table 6. Predictions of two accounts for initial tonal processing in Experiment 2

Accounts Predictions

Fixation Proportion Divergent time between TAR and competitors 1. Tone is

accessed early CTC > COC COC is earlier than CTC 2. Tone is

accessed late CTC = COC Comparable time point

4.1 Method

4.1.1Participants

Thirty-two participants, including 21 females and 11 males were recruited through

on-line sign-up sheets and paid to participate in the experiment. Their mean age was

21.7 years old, ranging from 19 to 25 years old. All participants had normal or

correct-to-normal vision and were native speakers of Mandarin Chinese.

4.1.2 Material

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

4.1.2.1 Stimuli

There were 140 monosyllabic Chinese characters in the experimental stimuli,

including 28 target characters, 28 cohort-tone competitor character, 28 cohort-only

competitor characters, and 56 unrelated characters. The segmental structure of these

stimuli comprised 70 CVC, and 70 CGVC. The stimuli was consisted of 36 characters

with Tone 1, 35 characters with Tone 2, 35 characters with Tone 3, and 34 characters

with Tone 4. The features of target and competitors were controlled as follows. First

of all, the frequency of these characters were controlled in a range between 7~200, as

computed from the CKIP Electronic Dictionary (The CKIP Electronic Dictionary is

an electronic lexicon for Mandarin Chinese containing 88,000 entries). There was no

significant difference in character frequency across the target and the cohort-tone and

cohort-only competitors (F (2, 81) = 2.105, p=.128). Secondly, the average stroke of

these characters were controlled in a range between 5~20. There was no significant

difference of the stroke across the target and two types of competitors were balanced.

(F (2, 81) = 0.122, p=.885). Lastly, the average number of homophone was under 10.

There was no significant difference in the average number of homophone across the

target and the two types of competitors (F (2, 81) = 1.172, p=.315).

Table 7. Means and SDs of character frequency, strokes, and homophone number for target, cohort-tone competitor and cohort-only competitor

Frequency Stroke Homophone

Note. TAR: target; CTC: cohort-tone competitor; COC: cohort-only competitor; UR:

unrelated distractor

4.1.2.2 Recording

The target and the competitors were recorded by a 25-year-old female Chinese

native speaker through the Audio-technica MB 4k/c cardioid condenser microphone.

The recording data was digitalized at a sampling rate of 44100 Hz, 16 bits through the

software Praat. The mean durations of target, cohort-tone competitor, and cohort-only

competitor characters were 745.3 msec, 741.5 msec, and 755.2 msec, respectively.

4.1.2.3 Auditory Stimuli Pretest

To ensure that the tone of the auditory stimuli was clear for the subject to recognize,

a survey was conducted. Four females and one male whose average age was 26.4

years old took part in the survey. The participants listened to all of the auditory stimuli

and then ticked off the tone (Tone 1 to Tone 4) and typed the Chinese character they

considered to be. The result showed that average of the accuracy of each experimental

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

stimulus was 99%.

4.1.3 Design

Two conditions were manipulated in this experiment. There were two syllable

structures CVC and CGVC in the present study. The initial cohort in CVC syllable

structure was CV, and the initial cohort in CGVC syllable structure was CG. The

cohort-tone competitors shared cohort structure and tone with the target. The

cohort-only competitors shared only the cohort structure with the target. For each

target, the two types of competitors shared the same rhyme. For example, a stimuli set

included a target / tang1 / ‘soup’, a cohort-tone competitor / tai1 / ‘fetus’, a

cohort-only competitor / tai4 / ‘peaceful’, and two unrelated distractors (the segmental

and tone were different from target: / pʰow3 / ‘to cut open’, and / xən2 / ‘scar’). An

experimental trial comprised one target, one competitor, which was either cohort-tone

competitor or cohort-only competitor, and two unrelated distractors. The entire

experiment consisted of 62 trials, including 56 experimental trials, 4 filler trials and 2

practices. The filler trials and practice trials were not included for analysis. The

experimental trials were mixed and randomly distributed into four lists. In each of the

lists, the number of each condition was equal and the conditions were

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

counterbalanced across subjects. There were two blocks of 32 trials, of which the first

two trials were fillers. The relationship between the target and the competitor in the

first block was exchanged in the second block.

4.1.4 Layout of visual stimuli

The layout of visual stimuli was same as Experiment 1.

4.1.5 Apparatus & procedure

The apparatus and procedure follows that of Experiment 1.

4.2 Data analysis

All the analyses and measures were the same as in Experiment 1.

4.3 Result

The mean of reaction time and correct hits on matching the acoustic target character

to the character in visual display were computed for each participant. Mean reaction

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

time for response was 1315.3 msec (SD = 275.3) and mean accuracy rate was 99 %

(Max = 100%, Min = 98%).

Figure 10 plots the fixation proportions of target, competitor, and unrelated

distractor for every millisecond from 200 msec, the time when visual display showing

to the target acoustic onset 0 msec until 1000 msec after acoustic target onset in

cohort-tone and cohort-only conditions. Starting from about 301msec, both of the two

competitors attracted more fixations than the unrelated distractors. As expected, the

time when the curve of target diverged from cohort-only competitor was earlier than

when the target curve diverged from cohort-tone competitor curve. In addition, as

illustrated in Figure 11, the fixation proportions of cohort-tone competitor was higher

than that of cohort-only competitor.

Figure 12 and Figure 13 plots the fixation proportions of target, competitor, and

unrelated distractor in CVC and CGVC syllable structure respectively for every

millisecond from 200 msec, the time when visual display showing to the target

acoustic onset 0 msec until 1000 msec after acoustic target onset in cohort-tone and

cohort-only conditions. For the syllable structure of CVC, the time when the curve of

target diverged from cohort-only competitor was earlier than when the target curve

diverged from cohort-tone competitor curve. However, as for CGVC, the time when

the curve of target diverged from cohort-only competitor was similar when the target

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

curve diverged from cohort-tone competitor curve.

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

Figure 10. Fixation proportions to targets, competitors, and unrelated distractors for trials with cohort-tone or cohort-only competitors in Experiment 2. The x-axis shows time in milliseconds from visual display onset, 200 msec before target acoustic onset, for the 1200 msec period.

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

Figure 11. Fixation proportions to cohort-tone/cohort-only competitors across two experimental conditions in Experiment 2. The x-axis shows time in milliseconds from the display onset, for 1200 msec.

Figure 12. Fixation proportions to targets, competitors, and unrelated distractors in CVC syllable structure for trials with cohort-tone or cohort-only competitors in Experiment 2. The x-axis shows time in milliseconds from visual display onset, 200 msec before target acoustic onset, for the 1200 msec period.

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

Figure 13. Fixation proportions to targets, competitors, and unrelated distractors in CGVC syllable structure for trials with cohort-tone or cohort-only competitors in Experiment 2. The x-axis shows time in milliseconds from visual display onset, 200 msec before target acoustic onset, for the 1200 msec period.

Figure 14. Fixation proportions to cohort-tone/cohort-only competitors in CVC and CGVC syllable structure across two experimental conditions in Experiment 2. The x-axis shows time in milliseconds from the display onset, for 1200 msec.

We performed the analyses of variance (ANOVAs) by participants (F1) and items

(F2) for the fixation proportions in time bins of 100 msec, starting from the onset of

acoustic target for 1000 msec in cohort-tone and cohort-only condition (Table 8).

Additionally, we performed the same analyses as above-mentioned in CVC and

CGVC respectively (Table 9 and Table 10). As Table 8 illustrated, in both of the two

conditions, the differences among target, competitor, and distractor were significant

during 301 msec to 1000 msec. Table 9 showed that the differences among target,

competitor, and distractor were significant during 401 msec to 1000 msec in both

conditions. Table 10 showed that the differences among target, competitor, and

distractor were significant during 401 msec to 1000 msec in cohort-tone condition but

during 301 msec to 1000ms in cohort-only condition.

Table 8. Analyses of variance by participant and item comparing mean fixation proportions to tonal and segmental competitors with those of the target and unrelated distractors from 1 msec to 1000 msec after acoustic target onset in Experiment 2

Condition Test 1-100 101-200 201-300 301-400 401-500 501-600 601-700 701-800 801-900 901-1000

F 1(2,62) 0.66 0.90 0.91 5.70 30.72 81.58 116.20 249.00 364.60 609.10

p 0.521 0.414 0.410 0.005 0.000 0.000 0.000 0.000 0.000 0.000

F 2(2,54) 0.45 0.54 0.57 3.33 19.41 77.86 162.20 183.70 321.10 485.70

p 0.642 0.584 0.569 0.043 0.000 0.000 0.000 0.000 0.000 0.000

F 1(2,62) 0.36 0.62 0.36 4.32 28.56 66.35 193.50 390.90 599.20 1055.00

p 0.701 0.541 0.702 0.018 0.000 0.000 0.000 0.000 0.000 0.000

F 2(2,54) 0.24 0.96 0.49 6.05 34.81 63.29 180.50 327.20 741.50 1758.00

p 0.790 0.390 0.614 0.004 0.000 0.000 0.000 0.000 0.000 0.000 All

Time bin (ms)

Cohort-Tone

Cohort-Only

Note: F 1 = 32 participants, F 2 = 28 items

Table 9. Analyses of variance by participant and item comparing mean fixation proportions to tonal and segmental competitors with those of the target and unrelated distractors in CVC syllable structure from 1 msec to 1000 msec after acoustic target onset in Experiment 2

Table 10. Analyses of variance by participant and item comparing mean fixation proportions to tonal and segmental competitors with those of the target and unrelated distractors in CGVC syllable structure from 1 msec to 1000 msec after acoustic target onset in Experiment 2

The mean fixation proportions to target, competitor, and the unrelated distractors

and standard errors were shown in Figure 15. For each time bin, a one-way ANOVA

was performed for trials with cohort-tone or cohort-only competitors and their targets

and distractors, followed by the post-hoc comparisons (Table 11). The mean fixations

Condition Test 1-100 101-200 201-300 301-400 401-500 501-600 601-700 701-800 801-900 901-1000 F 1(2,62) 0.06 0.68 0.92 2.67 11.32 31.64 -0.39 119.40 236.70 454.90

p 0.938 0.512 0.403 0.077 0.000 0.000 1.000 0.000 0.000 0.000 F 2(2,54) 0.04 0.54 0.69 2.08 7.72 37.54 71.36 81.06 227.80 659.10

p 0.958 0.590 0.509 0.145 0.002 0.000 0.000 0.000 0.000 0.000 F 1(2,62) 0.08 1.30 1.40 0.85 10.59 45.30 89.00 140.90 277.00 699.60

p 0.925 0.281 0.253 0.431 0.000 0.000 0.000 0.000 0.000 0.000 F 2(2,54) 0.12 2.17 2.91 2.07 19.71 30.83 73.82 117.30 323.00 702.90

p 0.892 0.134 0.072 0.147 0.000 0.000 0.000 0.000 0.000 0.000 CVC

Time bin (ms)

Cohort-Tone

Cohort-Only

Note: F 1 = 32 participants, F 2 = 28 items

Condition Test 1-100 101-200 201-300 301-400 401-500 501-600 601-700 701-800 801-900 901-1000 F 1(2,62) 1.66 0.47 0.45 2.18 11.81 27.89 50.15 120.60 184.30 327.80

p 0.198 0.629 0.638 0.121 0.000 0.000 0.000 0.000 0.000 0.000 F 2(2,54) 1.35 0.23 0.35 1.28 14.02 39.77 102.20 109.20 118.50 138.20

p 0.276 0.796 0.711 0.294 0.000 0.000 0.000 0.000 0.000 0.000 F 1(2,62) 0.67 0.08 0.30 3.16 13.32 24.41 85.99 281.90 504.10 673.70

p 0.517 0.925 0.742 0.049 0.000 0.000 0.000 0.000 0.000 0.000 F 2(2,54) 0.41 0.12 0.52 3.88 16.23 31.47 105.30 238.00 462.10 1172.00

p 0.665 0.888 0.603 0.034 0.000 0.000 0.000 0.000 0.000 0.000 CGVC

Time bin (ms)

Cohort-Tone

Cohort-Only

Note: F 1 = 32 participants, F 2 = 28 items

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

proportions to cohort-tone competitor and target was significantly different by

participants and by item analyses from 501-1000-msec. The mean fixations

proportions to cohort-only competitor and target was significantly different by

participants and by item analyses from 401-1000-msec. When the cohort-only

competitor diverged from the target was earlier than when the cohort-tone competitor

diverged from the target.

The mean fixations proportions to cohort-tone competitor and unrelated distracter

was significantly different by participants analyses from 301-1000-msec and by item

analyses from 401-1000-msec. The mean fixations proportions to cohort-only

competitor and unrelated distracter was significantly different by participants and by

item analyses from 301-1000-msec. When the unrelated distractors diverged from two

types of competitors was on a consistent time bin by participants analysis.

Figure 15. Mean fixation proportions to targets, competitors, and unrelated distractors for cohort-tone and cohort-only conditions in the 1,000 msec period following

acoustic target onset in Experiment 2. Each data point represents the average of fixation proportions across participants in the time bin of 100 msec and the error bars show the standard error of the data.2

2 The calculation of mean fixation proportions in each time bin included the area of target, competitor, unrelated distractors, cross , and the remaining areas outside the above-mentioned areas in the visual display

Fixation Proportion (Mean +/- 1 SE)

Cohort-Tone Competitor

Fixation Proportion (Mean +/- 1 SE)

Cohort-Only Competitor

TAR COC UR

Table 11. Analyses of variance by participant and item comparing mean fixation proportions to competitors with those of the target and unrelated distractors from 1 msec to 1000 msec after acoustic target onset in Experiment 2

The mean fixations proportions to cohort-only competitor and target was

significantly different by participants analyses from 401-500-msec in Tone

1[401-500-msec: z1 = 3.32, p = .003]. The mean fixations proportions to cohort-only

competitor and target was significantly different by participants analyses from

501-600-msec in Tone 4[501-600-msec: z1 = 4.21, p < .001]. Lastly, the mean

fixations proportions to cohort-only competitor and target was significantly different

by participants analyses from 601-700-msec in Tone 2 and Tone 3

[601-700-msec :Tone 2: z1 = 5.47, p < .001; Tone 3: z1 = 6.60, p < .001 ].

The result indicated that in Tone 1, the divergent time between the curve of

cohort-only competitor and target was significantly earlier than Tone 4, which was

also earlier than Tone 2 and Tone 3. Figure 16 plots the fixation proportion over time

Condition Test 1-100 101-200 201-300 301-400 401-500 501-600 601-700 701-800 801-900 901-1000 TAR - CTC z1 1.13 1.30 1.36 < 1 < 1 6.07 10.40 18.23 23.24 30.35

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

in two conditions among four tones.

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

Figure 16. Mean fixation proportions to targets, competitors, and unrelated distractors for cohort-tone and cohort-only conditions in the 1,000 msec period following

acoustic target onset in Experiment 2. Each data point represents the average of fixation proportions across participants in the time bin of 100 msec and the error bars show the standard error of the data. Fixation proportions to target, competitors and unrelated distractors over time for 4 tones are shown respectively.

The mean fixation proportions to target, competitor, the unrelated distractors and

standard errors were shown in Figure 17 for CVC syllable structure and Figure 18 for

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

CGVC syllable structure. A one-way ANOVA was performed for trials with

cohort-tone or cohort-only competitors and their targets and distractors for each time

bin, followed by the post-hoc comparisons (CVC for Table 12 and CGVC for Table

13). In CVC, the mean fixations proportions to cohort-tone competitor and target was

significantly different by participants and by item analyses from 501-1000-ms. The

mean fixations proportions to cohort-only competitor and target was marginally

significant different by participants analyses and significantly different by item

analyses from 401-1000-ms. In CGVC, the mean fixations proportions to cohort-tone

competitor and target was significantly different by participants and by item analyses

from 501-1000-ms. The mean fixations proportions to cohort-only competitor and

target was also significantly different by participants and by item analyses from

501-1000-ms. Table 14 showed the time when mean fixation proportion had

significant difference between the competitor and its target and unrelated distractors

by participants in two conditions of Experiment 2.

立 政 治 大 學

N a tio na

l C h engchi U ni ve rs it y

Figure 17. Mean fixation proportions to targets, competitors, and unrelated distractors in CVC syllable structure for cohort-tone and cohort-only conditions in the 1,000 msec period following acoustic target onset in Experiment 2. Each data point represents the average of fixation proportions across participants in the time bin of 100 msec and the error bars show the standard error of the data.

Figure 18. Mean fixation proportions to targets, competitors, and unrelated distractors in CGVC syllable structure for cohort-tone and cohort-only conditions in the 1,000 msec period following acoustic target onset in Experiment 2. Each data point represents the average of fixation proportions across participants in the time bin of

Figure 18. Mean fixation proportions to targets, competitors, and unrelated distractors in CGVC syllable structure for cohort-tone and cohort-only conditions in the 1,000 msec period following acoustic target onset in Experiment 2. Each data point represents the average of fixation proportions across participants in the time bin of