Task 4: Homophonous Naming Task

國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

50

they finished answering, or the star hint of the next trial would show up automatically to inform subjects the coming trial. Their naming response and their reacting time were also recorded in the same way at this phase.

3.1.3.4 Task 4: Homophonous Naming Task

From above visual tasks, we would like to know how the sound similarity and visual representation induced effect on lexical process, which would reflect on the speech errors and the response time. In this section, we recruited a set of homophones as a substitution for color terms. Homophonous Stroop naming test was designed to take homophones instead of color terms to observe whether the lexical process would be similar to that for color terms. With regard to the test material, the target colors were still the same with previous experiments. The colors included red, orange, brown, yellow, green, blue, purple, gray, white, and black again, but the visual stimuli were represented as the corresponding homophones of those colors. The organization and order of stimuli was the same with Stroop naming test. In order to make sure that every aspect kept similar, except for the semantic domain, when we used homophonous characters as visual stimuli, we had to control the frequency of the characters to have them persuasible to replace for the color terms. Even though it is impossible to find the homophones sharing the same frequency with their corresponding colors, we could find a set of homophones whose distribution and correlation of word frequency was similar to those among color terms. The following Table 3-15 is the homophones of color terms and their character frequency in Chinese.

‧

Table 3-15. Word Frequency of Color Terms and Corresponding Homophones

Color Term 紅 棕 橙 黃 綠 藍 紫 灰 白 黑

** indicates there is no data or information

Table 3-15 shows each word frequency for each color term in Chinese. The number indicates the specific times which the word appears in language. The most frequent is white (N=661), and follows red (N=503), black (N=437), yellow (N=367), green (N=200), blue (N=168), gray (N=116), purple (N=30), brown (N=11), and orange (N=11). The words in lower row are the corresponding homophones and their frequency. These are some missing statistics in the homophone candidates. For example, [xwɑŋ35] is one of the case that there is no other characters, except for “黃”, having frequency statistics in the Sinica Corpus. With an ad hot approach, we choose

“皇 (experor)” as a corresponding homophone, because it is the most transparent literally in reading task among all the candidates. Another missing data is the homophone of color white “白 [paj35].” In Chinese, there is not any character which owns the same tonal syllable. To maintain the whole trial in a homophonous condition, we still recruited the term white again in this homophonous naming task.

However, it seems impossible to find a set of homophonous words with absolutely the same frequency. With regard to keep frequency balanced both among the color terms and their homophones, we needed to have these frequency statistics from Sinica Corpus (2005) to pass the examining of correlations test. We had to

3 The frequency data were quoted from Cheng, Huang, Lo & Tsai (2005) “Word List with Accumulated Word Frequency in Sinica Corpus.”

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

52

examine the frequency relation among the homophones we had used to compare with the relation among the color terms, so that we could make sure that frequency bias would induce little influence in naming test. Table 3-16 shows the correlation test between color terms and their homophonous counterparts.

Table 3-16. The Frequency Correlation between Color and Homophone

Descriptive Statistics Mean Std. Deviation N

Color term 175.75 178.65 8

Homophone 8.00 7.48 7

Correlations Color term Homophone Color term Pearson Correlation 1 .961^**

Sig. (2-tailed) .001

Homophone Pearson Correlation .96^** 1

Sig. (2-tailed) .001

**. Correlation is significant at the 0.01 level (2-tailed).

From Table 3-16, the mean scores of the two variables are 175.75 and 8.0 respectively, while the ones of standard deviation are 178.65 and 7.48 respectively.

The Pearson correlation is .96 (p<.05), which reaches to the level of significance. That is to say, the frequency distribution among the color terms correlates the distribution among the homophones positively and significantly (r=.961, p<.05).

The correlation examination shows that the homophone carriers keeps the character frequency balanced with color terms. The trial designs in Stroop’s naming test were repeated in homophonous naming task, except for the homophones in Table 3-15 which were substituted for the original color stimuli, as shown in sampler trials (g) and (h).

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

53

(g)

黑

黑 藍 藍 藍 橙

橙 紫

紫 綠 綠 綠 黑

黑 黃

黃 藍

藍

(h)

嘿

嘿 蘭 蘭 蘭 懲

懲 籽

籽 慮 慮 慮 嘿

嘿 皇

皇 蘭

蘭

The sampler (g) was used in Stroop’s naming test, and (h) is the corresponding trial in the homophonous Stroop’s naming task. All the lexical pairs in (h) were replaced by the homophones in Table 3-15, and the visual colors were arranged in the same order with (g). Of the trial in (g), the color pair “黑藍 [xej55 lan35]”

(black-blue) was replaced with homophonous pair “嘿蘭.” Their visual colors were painted in blue for the former and black for the other respectively. The other pairs orange-purple, green-black, and yellow-blue were replaced in the same manner, as shown in (h). The other phonological trials of this experiment are displayed in appendix 1.

During this phase, there was nothing different from Stroop’s task to name the color of lexical, except for the homophone stimuli. After the gazing at star mark for two seconds, subjects were asked to name the phonological pairs for each trial, and then pressed the response key to have their answers as well as reacting span be recorded. The total response time was limited within 20000 ms, and the next trial would show up automatically.