Comparisons between Regular and Irregular Phonetic Compounds

Section 4.1 disclosed that in terms of pronunciation identification, a significant interaction of character familiarity and regularity was present in all three grades according to repeated measures two-way ANOVA’s. As for meaning identification, the interaction of character familiarity and regularity was only statistically significant in the sixth grade according to repeated measures four-way ANOVA’s. Consequently, this section goes on to illustrate the simple main effects and main effect.

Starting with the acquisition of pronunciation, Table 4-5 exhibits the performance of each grade on regular and irregular phonetic compounds at different familiarity levels in PT. The second graders’ superior performances on the regular type compared to those on the irregular type were observed at all three levels (F level:

M = .98 > .95, F-1 level: M = .96 > .94, F-2 level: M = .88 > .63). But a significant difference was reported by repeated measures one-way ANOVA only at F-2 level (F(1, 20) = 22.56, p < .001), leaving out F level (F(1, 20) = 2.91, p > .05) and F-1 level (F(1, 20) = .81, p > .05). On the other hand, Grade Four pupils gained higher score on the regular type at F-1 and F-2 levels, but equivalent score at F level (F level: M =

60

1.00 = 1.00, F-1 level: M = .98 > .96, F-2 level: M = .94 > .78). And a significant difference was reported by repeated measures one-way ANOVA only at F-2 level (F(1, 19) = 20.23, p < .001), not at F-1 level (F(1, 19) = 1.31, p > .05). Similarly, the scores of the sixth grade on the regular type surpassed those on the irregular type at F-1 and F-2 levels, but equaled at F level (F level: M = .99 = .99, F-1 level: M = .94 > .80, F-2 level: M = .90 > .32). And a significant difference was reported by repeated measures one-way ANOVA at both F-1 level (F(1, 21) = 12.96, p < .01) and F-2 level (F(1, 21)

= 136.37, p < .001), with the exception of F level (F(1, 21) = 0.00, p > .05). The results proved that compared to the pronunciation of regular phonetic compounds, the pronunciation of irregular phonetic compounds cost the Chinese children more efforts to acquire.. Moreover, with the character familiarity falling, the pronunciation of irregular phonetic compounds became more challenging for the Chinese children to acquire.

Table 4-5 The Performance of Each Grade on Regular and Irregular Phonetic Compounds at Different Familiarity Levels in PT

Grade Familiarity

Turning to the acquisition of meaning, Table 4-6 sets out the performance of each grade on regular and irregular phonetic compounds (at different familiarity levels) in

61

ST-W and ST-S. Since no significant interaction of character familiarity and regularity was reported in Grade 2 or 4, Table 4-6 only manifests the main effect of regularity for the two grades. A surpassing score on the regular type compared to that on the irregular one was obtained by Grade 2, not Grade 4 (Grade 2: M = .88 > .84, Grade 4:

M = .89 < .90). But a significant difference was absent in either grade (Grade 2: F(1, 20) = 3.96, p > .05, Grade 4: F(1, 19) = .89, p > .05). As for Grade 6, in which a significant interaction of character familiarity and regularity was found, the simple main effect is shown in Table 4-6. The sixth graders excelled at the performance on regular phonetic compounds compared to that on the irregular ones at F-1 and F-2 levels, though not at F level (F level: M = .98 = .98, F-1 level: M = .83 > .81, F-2 level: M = .79 > .62). And a significant difference was only found at F-2 level (F(1, 21) = 16.86, p < .01) by repeated measures one-way ANOVA, contrasting with F level (F(1, 21) = .14, p > .05) and F-1 level (F(1, 21) = .28, p > .05). The results indicated that the meaning of irregular phonetic compounds cost the Chinese children more efforts to acquire compared to the meaning of regular phonetic compounds.

Furthermore, the meaning of irregular phonetic compounds turned harder for our Chinese children to master as character familiarity fell.

Table 4-6 The Performance of Each Grade on Regular and Irregular Phonetic Compounds (at Different Familiarity Levels) in ST-W and ST-S

Grade Familiarity

62

4.4.2 Discussion

Section 4.4.1 provides a comparison of the acquisition of regular and irregular phonetic compounds, both in terms of their pronunciation and the in terms of their meaning. Regarding the acquisition of pronunciation, the first crucial finding was that irregular phonetic compounds imposed greater challenge for our Chinese children compared to the regular ones. Table 4-5 describes significantly superior performances on the regular type in all three grades. The pronunciations of regular phonetic compounds were more accessible due to the fact that their phonetic radicals had the same pronunciations as the whole characters. But the phonetic radicals of irregular phonetic compounds were pronounced differently from the whole characters, impeding the recognition of the pronunciations of irregular characters. This finding implied that our children possessed phonetic radical awareness (Ho et al., 2003; Shu et al., 2000), and that this awareness benefited the acquisition of character pronunciation. The second graders already knew that Chinese characters are decomposable, and that the constituent phonetic radicals bear phonetic cues to the pronunciation of the whole characters that contain them. They made good use of the phonetic radicals, which elevated their performance on regular phonetic compounds with available phonetic cues. Irregular phonetic compounds, without such facilitation, proved to be challenging unsurprisingly. Lee, Huang, Kuo, Tsai, and Tzeng (2010) further found that adult readers continued to use phonetic radicals to infer the pronunciations of Chinese pseudocharacters. The practicality of phonetic radicals seems to be far-reaching and can be observed even in adulthood. The current finding also verified Zhou and Marslen-Wilson’s (1999) model of Chinese character recognition. Regular phonetic compounds with helpful phonetic radicals were acquired prior to the irregular ones since phonetic radicals could activate their phonological representations and helped retrieve the pronunciation of whole

63

characters.

Moreover, the existence of regularity effect actually depended on character familiarity. The second crucial finding stated that the pronunciation of irregular phonetic compounds turned more difficult for our participants to acquire as character familiarity decreased. As presented in Table 4-5, the regular type received significantly higher scores at F-1 and F-2 levels only, but not at F level. Because the participants could easily retrieve the pronunciations of all familiar characters, regular and irregular, no significant difference was expected between the two types of characters at F level. In contrast, the availability of pronunciation clues from the phonetic radicals became particularly indispensable in the phonological recognition of unfamiliar characters, which the children had little knowledge of. The participants’

performance on regular and that on irregular phonetic compounds would thus differ to a great degree. Such interaction of character regularity and familiarity was also recorded in Shu et al. (2000).

This interaction of character regularity and familiarity indicated that phonetic radical awareness only significantly assisted with the pronunciation acquisition of unfamiliar characters. Our participants had acquired the pronunciations of all familiar characters. Therefore, they might readily retrieve the pronunciations of both regular and irregular phonetic compounds without relying on radical analysis. Alternatively, other than general word knowledge, they might also call on phonetic radicals to help with the phonological access of regular characters. Irregular characters were without such privilege but could still be pronounced courtesy of the general knowledge the children had for familiar characters. Whichever the case, the performance on familiar regular and irregular phonetic compounds would be equally good. Nevertheless, unfamiliar characters told a very different story. Since our participants were ignorant of the pronunciations of unfamiliar characters, they performed poorly on irregular

64

characters. But they could still infer the pronunciations of regular characters from reliable phonetic radicals even though they lacked general knowledge of the characters. Therefore, significantly poorer performance on unfamiliar irregular phonetic compounds compared to that on unfamiliar regular ones was expected since the latter received facilitation from phonetic radical awareness.

A scrutiny of the scores of the three groups revealed that the second, fourth and sixth graders all had phonetic radical awareness in phonological access as mentioned in Table 4-5. The three grades performed similarly at F and F-2 levels but differently at F-1 level, at which a significant regularity effect in phonological access was only found in the sixth grade, not in the second or fourth grade. Grade 2 pupils could recognize the pronunciation of irregular characters as well as that of regular ones probably on account of their additional general word knowledge gained from outside reading. Their stronger parental support might help them learn to pronounce the irregular characters at F-1 level, hence neutralizing the regularity effect. Grade 4 pupils, however, seemed to be familiar with certain test items only. Although all target characters at F-1 level fell under the same frequency range, two items, “手帕”

(shǒupà, ‘handkerchief’) and “秩序” (zhìxù, ‘order’), tended to occur more often at school compared to other items. Elementary school teachers in Taiwan often check whether students bring handkerchiefs and tissue paper with them, and ask students to keep order in class. Therefore, it is natural that children know how to pronounce the two characters albeit their classification at F-1 level. In brief, at F-1 level, all three grades successfully employed their phonetic radical awareness to infer the pronunciation of regular characters, but only the second grade had supplementary general word knowledge of the pronunciation of irregular characters while the fourth grade merely knew particular irregular characters, which elevated their scores.

Due to the fact that a significant regularity effect was found exclusively at F-2

65

level for the second graders, only an examination of characters at this level could ensure the presence of phonetic radical awareness in the second grade. The participants had acquired this awareness, and knew how to take advantage of it when pronouncing familiar and especially unfamiliar characters. Progressing to the fourth and then the sixth grades, they based on previous achievement and kept building their phonetic radical awareness with more known characters and phonetic radicals. The heightened awareness might in return bestow more privileges on the recognition of unfamiliar characters. It is possible, then, to develop phonetic radical awareness in phonological access with more reading stimulus.

This pattern of development is in line with the previous research. Shu et al. (2000) reported phonetic radical awareness in the second graders, and Ho et al. (2003) found that the first graders already knew the position, function and sound values of phonetic radicals. The present study also spotted phonetic radical awareness in the second graders, who possessed knowledge of the position, function and sound values of phonetic radicals, and who could analyze both familiar and unfamiliar characters with this knowledge. Phonetic radical awareness continued to mature in the fourth and sixth graders as more characters were mastered.

With regard to the acquisition of meaning, the first key finding is that irregular phonetic compounds were acquired later than regular phonetic compounds. Table 4-6 demonstrates that the score on the regular type significantly surpassed that on the irregular type in Grade 6. Regular phonetic compounds, with reliable phonetic radicals, were more easily pronounced. And once pronounced, their meanings might also be activated due to the sound-meaning links formed in oral language. On the contrary, the unreliable phonetic radicals in irregular phonetic compounds provided no easy access to pronunciations or meanings, resulting in the inferior performance in meaning retrieval. This finding entailed that phonetic radical awareness (Ho et al.,

66

2003; Shu et al., 2000) was present in our Chinese children, and that this awareness could enhance the acquisition of character meaning. The sixth graders were aware of the structure and embedding phonological clues of Chinese characters. And they could use these clues to access not only the pronunciation but also the meaning of characters.

Regular characters equipped with reputable phonetic radicals were thus acquired prior to the irregular ones. This finding could also be explained by the model of Chinese character recognition proposed by Zhou and Marslen-Wilson (1999). Since the phonological representations of phonetic radicals were activated and supported the identification of character pronunciation, regular phonetic compounds with phonetically related phonetic radicals were identified more easily.

The role of phonological activation in the semantic retrieval of Chinese characters has also been discussed in previous research. Guo, Peng, and Liu (2005) found that when processing a homophone of a target character, the meaning of both the homophone and the target would be activated. Their explanation was that the phonological codes, or the pronunciations of characters, were automatically activated in familiar character recognition of children. This shared pronunciation of the homophone and the target then further activated the meaning of both characters. The access of meaning through pronunciation thus proved to be automatic in character recognition. In the present study, the pronunciations of regular phonetic compounds, familiar or unfamiliar, were easily identified through dependable phonetic radicals.

Once identified, the meanings of regular phonetic compounds could also be acquired.

And the absence of this merit in irregular characters increased their difficulty level.

An additional consideration of character familiarity revealed its impact on the regularity effect, and led to the second key finding, which was that the meaning of irregular phonetic compounds became even less accessible to our children compared to that of regular ones when character familiarity fell. As displayed in Table 4-6, the

67

participants’ significantly better performance on the regular type was observed only at F-2 levels, but not at F level. Since the children knew the meaning of all characters, regular or irregular, at F level, facilitation from phonetic radicals would not wield profound influence. At F-2 level, however, available phonetic radicals turned crucial in the meaning inference of unfamiliar characters, of which they lacked knowledge.

Without the advantage, the participants’ performance on irregular characters thus lagged behind. This interaction of character regularity and familiarity resembled that in Shu et al. (2000).

The interaction of character regularity and familiarity showed that phonetic radical awareness only significantly facilitated the meaning acquisition of unfamiliar characters. With familiar characters, our participants already knew the meaning of both regular and irregular characters and needed little help from radical analysis. They might still derive the meaning of regular characters from phonetic radicals, but they did equally well on irregular characters in light of their general word knowledge. In contrast, with unfamiliar characters, the participants witnessed the virtue of phonetic radical awareness when they successfully identified the meaning of regular characters but not the meaning of irregular ones. The lack of general word knowledge enabled the effect of phonetic radical awareness to emerge and to bring benefit for unfamiliar regular characters selectively.

A discrepancy among the three grades also existed. According to Table 4-6, only the sixth graders exhibited phonetic radical awareness in semantic access. As our pupils learned more characters and phonetic radicals with grades, they kept cultivating and intensifying their phonetic radical awareness. And when they reached the sixth grade, they could finally systematically capitalize on phonetic radicals to retrieve not only the pronunciation but also the meaning of characters. Therefore, a more developed phonetic radical awareness seemed to be a prerequisite for it to facilitate

68

meaning access. The second and fourth graders, with their still developing phonetic radical awareness, could not benefit from this facilitation.

However, grade six students only showed phonetic radical awareness in semantic access at F-2 level, not at F-1 level. Their good performance at F-1 level on regular characters could be due to their affirmed phonetic radical awareness, and that on irregular characters could be due to their semantic radical awareness. An examination of their answers on irregular characters revealed that they made considerably fewer mistakes on irregular and also transparent characters than that on irregular and also opaque characters. Therefore, they seemed to employ semantic radical awareness to recognize the meaning of irregular characters, and proved to be successful in elevating the score.

Previous researchers predominantly focused on the development of phonetic radical awareness in phonological access, and recorded such awareness in children as young as the first grade (Ho et al., 2003; Shu et al., 2000). The current study further reported the impact of phonetic radical awareness in semantic access, and concluded that only the sixth graders consistently demonstrated and benefited from such awareness.