Keywords

Introduction

There are about 28,854 ethnic minority (EM) students in Hong Kong (approximately 3.6% of the overall student population) studying in local public sector and directly subsidized schools (hereafter referred to as “local schools”) (Education Bureau (EDB) of the Government of the Hong Kong SAR 2016), and 3,782 non-Chinese residents aged between 0 and 19 in Macau (approximately 1.2% of the overall student population) according to the 2011 Population Census (Direcção dos Serviços de Estatística e Censos de Macau 2011). These EM students are mostly descendants of migrants from India, Pakistan, Nepal, and the Philippines, with a small number of them being Caucasian, Eurasian, or immigrants from other parts of the world. Past research (EDB of the Government of the Hong Kong SAR 2008) indicated that EM students encounter difficulties in learning Chinese as a second language (CSL) , particularly in character writing and reading . Lacking the automatic character recognition ability that characterizes efficient readers of Chinese (Leong et al. 2011), most EM students studying in local schools with compulsory Chinese curricula are facing enormous difficulties in mastering the language (Tsung et al. 2013). Such learning difficulties are often complicated by the huge diversity among EM students in local schools in terms of the number of years spent learning Chinese and the actual level of Chinese language proficiency. While some EM students were born in Hong Kong/Macau and started learning Chinese when they were young, others are new immigrants who only started learning the language upon their arrival in Hong Kong and Macau.

Geographically proximate to each other, both Hong Kong and Macau have been seeking solutions to confront the existing challenges of catering for individual learning needs of EM students. In Hong Kong, for example, streaming is a common practice in schools admitting EM students. EM students from different “streams” often show significant differences among themselves in trajectories of Chinese learning; as an effort to bridge the gaps in CSL learning among EM students, differentiation is being introduced to Chinese instruction in academically and ethnically mixed Chinese language classrooms. Schools in Macau tend to follow a different approach, encouraging EM students to learn Chinese in classrooms together with their ethnic Chinese counterparts, in hopes that they will integrate into the host society upon acquisition of the lingua franca “like a native speaker.”

Given the fact that most teachers in both special administrative regions (SARs) are not trained in teaching Chinese as a second language, many of them adhere to the pedagogies they have been using for teaching native speakers in their classroom with EM students. Such a practice can be problematic as EM students’ learning needs, ways of thinking and trajectory of Chinese development (esp., literacy) can be extremely different from those of their fellow Chinese-speaking counterparts (Leong et al. 2011). As an example of their instructional practice, these teachers tend to spend little time teaching component- and spatial configuration-based orthographic knowledge ; instead, they often go straight to text-based reading and writing practice (Loh et al. 2015a), overlooking the importance of direct instruction to help learners develop their orthographic knowledge (Chang 2009). Tsung et al. (2013) reported that Chinese character learning was considered by the teachers interviewed in Hong Kong as the major obstacle to CSL learning for EM students, who were frustrated and tended to give up quickly. In light of this challenge of Chinese characters for EM students, there is a pressing need to develop new pedagogies, curricula, and materials for teaching them Chinese characters, which could be based on cognitive processes of Chinese character recognition .

To this end, we conducted the present study that examined how EM students visually perceived patterns of Chinese characters. Specifically, it aimed to address the following two research questions. It is hoped that the findings could shed light on EM students’ orthographic knowledge development for the teaching and learning of CSL in Hong Kong and Macau.

  1. 1.

    How do EM students without specific training in component- and spatial configuration-based orthographic knowledge in Chinese visually perceive Chinese characters?

  2. 2.

    How do EM students’ demography and after-school activities relate to their visual perceptions of Chinese characters?

Character Components (Bùjiàn) and Orthographic Knowledge

As pointed out by Liu et al. (2007), Lee and Kalyuga (2011), second language (L2) learning in Chinese is a complex and challenging task, particularly for learners with an alphabetic language background. Given that the Chinese language is logographic without explicit phonetic hints like in alphabetic languages , character recognition is an essential skill that any learner of Chinese needs to have to access word meanings and comprehend texts.

The highly sophisticated orthographic system of the Chinese language often puzzles its L2 learners, including EM students in Hong Kong who mostly speak their heritage language and/or English at home and learn Chinese as a second language through formal instruction in school. Certain Chinese characters (字zì) are pictographs resembling physical objects, with 人 (“human”) as an example, and evolving through time into their current orthographic forms. However, single-unit characters (獨體字) that have a pictographic origin are very small in number; most characters in modern Chinese are made up of multiple, reusable components or bùjiàn (部件)—estimated to be about 540—that fit well into the square space allocated for each character within a text (or 合體字, compound characters ). The majority of compound characters are phonetic-radical compound characters (形聲字) with a component providing partial information about the sound (i.e., phonetic) and another component about the meaning (i.e., semantic radical or simply radical) of the host character (Taylor and Taylor 1995). Competent Chinese language users would be able to guess the meaning or sound of a character based on the hints provided by its components (Zhang 1987).

Chinese radicals, estimated to be about 214 (Chinese National Language Committee 1998), are a centuries-old system for accessing dictionaries that can be dated back to the second century, as detailed in the philological work Shuowen Jiezi (Analytical Dictionary of Characters) by Xu Shen (58–147 AD). The long-established radical system finds its roots in Chinese etymology and semantics. More modern interest in the component system of Chinese characters places an emphasis not only on the phonetic and semantic information or “clues” as displayed in individual characters but also on the structural configurations and composition rules of the components.

Overall, the compositions of Chinese characters can be classified into 15 spatial configurations of structural components/ bùjiàn (Ki et al. 2003; Tse 2000; Wong 2009), namely:

  1. 1.

    single component (square) 日 [sun];

  2. 2.

    single component (triangle) 上 [above];

  3. 3.

    left-right 日 [sun] + 月 [moon] = 明 (bright);

  4. 4.

    upper-lower 小 [small] + 大 [big] = 尖 [sharp];

  5. 5.

    diagonal 夕 [half moon] + 夕 [half moon] = 多 [many];

  6. 6.

    left-middle-right 彳 [walking slowly] + 圭 [jade] + 亍 [stop walking] = 街 [street];

  7. 7.

    top-middle-bottom 亠 [high] + 口 [mouth] + 小 [small] = 京[capital];

  8. 8.

    top (left-right)-down 竹 [bamboo] + 夭 [dog] = 笑 [smile];

  9. 9.

    top-down (left-right) 口 [mouth] + 口 [mouth] + 口 [mouth] = 品 [commodity];

  10. 10.

    left-right (top-down) [water] +  [born] + 丹 [red] = 淸 [clear];

  11. 11.

    surrounding 口 [boundary] + 大 [big] = 因 [reason];

  12. 12.

    outer-inner 木 [wood] + 口 [mouth] = 束 [bundle];

  13. 13.

    half-surrounding (top-down) [rack] + [knot & death] = 罔 [net]

    or (bottom-up)乂 [entangle] + [trap] = 凶 [danger]

    or (left-right) [ruler] + 口 [mouth] = 叵 [improbable];

  14. 14.

    inclusive structure (left-bottom) 走 [walk] +  [join] = 赳 [valiant]

    or (left-top) 尸 [corpse] + 肖 [small] = 屑 [small pieces]

    or (right-top) 戈 [weapon] + 廾 [two hands] = 戒 [get rid of];

  15. 15.

    eudipleural 大 [big] + 人 [human] + 人 [human] = 夾 [hold].

Tse (2002) found that 77 of the most commonly used components constitute 1200 high-frequency Chinese characters. Forty-eight of them are single-component characters themselves (e.g., 人 person), of which 26 are commonly used with such good configuration ability (as deformable components [變形部件] when combined with other components (e.g., 人 and 亻) that they construct one-third of essential Chinese characters. Wong (2009) reported that most of traditional Chinese characters are constructed by two to eight components (with five as the average number). As a result, the memory load for memorizing components is lower than strokes. (The average number of stokes of characters is about 12, and there is no meaningful relationship between the different strokes as opposed to components.) Character learning could thus be possibly enhanced by acquiring these single-component characters with high occurrence and good configuration ability (Tse 2002; Zhang 2012), as the components involved are meaning carriers and thus making better sense to learners.

Previous studies showed that orthographic awareness enhances character learning among Chinese-speaking students (e.g. Shu et al. 2000). This kind of sensitivity has also been found to benefit the learning of characters among second language learners of Chinese, including kindergarteners (e.g., Loh et al. 2013; Tse and Loh 2014), primary school students (Tse et al. 2012), secondary school students (Lee and Kalyuga 2011) as well as adult learners (e.g., Liu et al. 2007). However, L2 studies on learners’ orthographic knowledge are still limited. In addition, these existing studies only addressed a limited number of components (Shu and Anderson 1997) or structures (e.g., Loh et al. 2013; Shu et al. 2003). A complete understanding of L2 learners’ knowledge of diverse spatial configurations of components is yet to be achieved. To this end, the current study examined the perceptions of the 26 most commonly used components with high configuration ability and 15 spatial configurations among EM students learning CSL in Hong Kong and Macau. By prioritizing essential Chinese components and identifying the effect, if any, of sociocultural factors on orthographic knowledge acquisition among secondary EM students, the authors also aimed to propose a number of feasible solutions to the challenges faced by adolescent CSL learners in Hong Kong and Macau in their character learning.

Methods

Participants

One hundred and fifty-one EM students (49 males and 102 females) from a local secondary school in Hong Kong and 62 EM students (34 males and 28 females) from a local secondary school in Macau took part in the study as a closed cohort through convenience sampling. Both schools were EM-dominated mission schools adopting the Hong Kong Chinese language curriculum teaching the Traditional Chinese script and spoken Cantonese. Among them, 141 of them were lower secondary students and 72 of them were upper secondary students (see Table 5.1).

Table 5.1 Distribution of participants with different years of study in school (N = 213)
Full size table

Among the 213 EM students, four major ethnic groups were identified, including Filipinos (N = 80), Pakistanis (N = 66), Nepalese (N = 30), and Indians (N = 21). There were also 14 participants from other ethnic groups (see Table 5.2); most of them were Asian (4 Thais, 2 Myanmarese, 2 overseas Chinese, 1 Korean), with 1 Australian, 1 Portuguese, 1 Irish, 1 Kazakh, and 1 African. Ethnicity information was not available for two students.

Table 5.2 Overview of participants’ ethnicities (N = 213)
Full size table

Most of the participants (N = 136) were born in their place of residence (i.e., Hong Kong/Macau), with 36 moving to Hong Kong/Macau between 0 and 6 years old, and 41 after six years old (see Table 5.3).

Table 5.3 Age of participants at the time of migration to Hong Kong/Macau (N = 213)
Full size table

Research Instruments

Three tasks, designed by the research team on the basis of the Chinese Language Assessment Tool for Non-Chinese Speaking Students (Education Bureau of the Government of the Hong Kong SAR 2010), were adopted in this study to examine the orthographic awareness of the participants. The tasks aligned with the basic requirements for the first stage in reading in both the Chinese as the first language (CL) primary curriculum (Curriculum Development Council of the Hong Kong SAR 2008) and the CSL general curriculum (Education Bureau of the Government of the Hong Kong SAR 2014), i.e., “ability to recognize and read aloud commonly used characters,” with orthographic knowledge as the prerequisite or an integral part of the curriculum. In addition, a language use survey was administered.

Separation of Chinese Character Components Test

This instrument assessed the participants’ awareness of Chinese character components, more specifically, whether they were able to separate various components from a character. It covered 26 randomly ordered compound characters constituted by the 26 most commonly used components (which were themselves single-component characters) with high configuration ability.

Each participant was required to list the components embedded in their host compound characters so as to break the characters down into individual components. Each compound character (i.e., each item) with all components correctly listed was awarded 1 mark, the full mark of each item. A partial mark was also awarded with one or more but not all components correctly listed. For each item, the partial mark is equal to the number of components identified divided by the total number of components. Regardless of partial or full mark, each Chinese character, or each item, is scored with the range 0–1. The maximum raw score possible for this task was 26. The final scores were reported in decimal form representing a proportion (i.e., total raw score divided by 26).

E.g., 倡 = 亻 (or 人) + 日 + 曰

Constitution of Chinese Character Components Test

This instrument assessed the participants’ ability to use the components commonly used in their curricular materials to construct Chinese characters. Based on seven of the most commonly used single-unit components (i.e., 人, 心, 女, 口, 木, 目 and 子), it required the EM students to compose as many legitimate characters as possible. With these seven components, the theoretical total number of legitimate characters is 39. For each participant, the individual score was the total number of characters composed divided by 39. Each possibly composed character was considered an item. For each item, the item score was the total number of participants who composed that character divided by the total number of participants.

Chinese Character Structure Test

This instrument assessed the participants’ awareness of 15 types of spatial configurations of Chinese characters via 30 multiple-choice-type items. It required the participants to identify the structure of each of the 30 target characters by circling the correct choice among the four given options. The 30 randomly ordered characters were selected from the list of Frequency Statistics of Commonly Used Modern Chinese Characters (Ho and Kwan 2001). Each correct answer would be awarded one mark. The maximum raw score for this test was 30. The final scores were reported in decimal form representing the proportion of correct choices.

(The correct answer is B)

Survey on Participants’ Language Use

The survey consisted of 10 multiple-choice questions . It covered the students’ ethnic and sociolinguistic backgrounds, language domains, as well as their television-viewing, reading, and Chinese language learning habits. It helped us examine the relationship between the students’ after-school activities and their performance on the aforementioned three tasks, with an objective to identify factors affecting CSL literacy development in informal contexts. Up to 5 options were given in each of the multiple-choice questions; the students were asked to choose the option that best described their background or after-school activities.

For example, 你每星期花多少時間温習中文? How much time do you spend every week studying Chinese?

  1. A.

    沒有時間 No time

  2. B.

    1小時以下 Less than one hour

  3. C.

    1至3小時 One to three hours

  4. D.

    3至7小時 Three to seven hours

  5. E.

    其他Others

    (請註明Please specify: _______________)

Data Collection

Multiple one-hour on-site, pencil-and-paper assessment tests took place in the participating schools. The participants were given detailed instructions and allowed to raise queries about the tests before they took them. At least, one research team member was sent to each test site as an invigilator.

Analytical Tools/Methods

The data collected were analyzed both qualitatively and quantitatively. The quantitative analysis was conducted with SPSS (e.g., ANOVA), whereas the qualitative data were examined in the form of typological analysis. The participants’ patterns of deciphering the spatial configurations of Chinese characters were identified by categorizing their answers for Task 1 according to the basic unit(s) they opted for (e.g., components, strokes) and the combinations formed.

Results

Task 1: Separation of Chinese Character Components Test

The mean and standard deviation of the 213 participants’ scores for Task 1 were 0.79 and 0.123, respectively. Cronbach’s alpha was 0.84, indicating a good level of reliability . The results showed that the participants, who had not received systematic, formal training in Chinese character components and their spatial configurations , overall possessed a certain, yet limited, level of knowledge of them. Out of the 213 participants, 122 (or about 57.9%) scored 0.80 or above, among whom 17 (about 8.1% of all 213 participants) earned 0.90 or above, and the scores of the other 105 (about 49.8% of all participants) fell into a range of 0.80–0.89. Only five participants (2.3% of the cohort) earned 0.30 or below.

For item-level analysis, the mean score for each item in Task 1 fell within a range of 0.62–0.94. The items with the largest number of correct responses were “災” (lit. “disaster,” 91.1% correctness out of 213 responses), “初” (lit. “beginning,” 89.2% correctness), “想” (lit. “to think,” 87.3% correctness), “蝴” (lit. one of the characters in the word “butterfly,” 82.6% correctness), and “霜” (lit. “frost,” 82.2% correctness). The items with the smallest number of correct responses included “藥” (lit. “medicine,” 36.6% correctness), “糕” (lit. “cake,” 34.7% correctness), “臉” (lit. “face,” 34.3% correctness), “晴” (lit. “sunny,” 33.3% correctness), and “問” (lit. “to ask,” 25.5% correctness).

There were multiple ways in which the EM students in Hong Kong and Macau visually perceived Chinese characters when asked to break the target characters down to their components. Based on the analysis of the scripts produced by the participants for Task 1, five ways were identified. They included unanalytical agglomeration of strokes , agglomeration of sequenced strokes, arbitrary repetition of (parts of) component, hybrid agglomeration of sequenced strokes and components, and agglomeration of components, which we describe in detail below.

  1. 1.

    Unanalytical agglomeration of strokes (see Fig. 5.1)

    Fig. 5.1
    figure 1

    Chinese character splitting—unanalytical agglomeration of strokes

    Full size image

In the above example, the character “做” (lit. “to do”) was split into eight individual parts, displaying the stroke sequence by adding a new stroke to a set of repeated strokes in the preceding space. Interestingly, it shows close resemblance to how learners of Chinese, including CSL learners, typically practice character writing in their exercises, which do not have a focus on breaking down characters into their constituent components or bùjiàn . Apparently the participant whose pattern of breaking down 做 in the example given did not have a clear concept of character component—he/she made a guesstimate that showed a confusion between stroke and bùjiàn in written Chinese. In other words, the structure of the character was visually perceived by the participant as an unanalytical agglomeration of strokes showing little concept of component constitution of characters.

  1. 2.

    Agglomeration of sequenced strokes (see Fig. 5.2)

    Fig. 5.2
    figure 2

    Chinese character splitting—agglomeration of sequenced strokes

    Full size image

In the above example, the character “災” (lit. “disaster”) was split into seven individual parts, all of which were individual strokes making up the character. Apparently the participant did not have a clear concept of character components. Like the participant who produced the example shown in Fig. 5.2, he/she seemed to be confused between the smallest written unit (i.e., stroke) and the smallest meaningful unit (i.e., bùjiàn or component) in written Chinese. As a result, the participant gave an incorrect answer in which the visual structure of the character was presented simply as an agglomeration of strokes in a sequence he/she was used to writing the character.

  1. 3.

    Arbitrary repetition of (parts of) component (see Fig. 5.3)

    Fig. 5.3
    figure 3

    Chinese character splitting—arbitrary repetition of (parts of) component

    Full size image

In the above example, the character “跟” (lit. “heel”; “to follow”) was split into six individual parts. The participant seemed to display a certain level of awareness of character component with the two components of 跟 correctly produced (the first and second parts in Fig. 5.3). However, it does not seem clear if his/her awareness was intact or he/she might still be confused about what constitutes components, as he/she further split the two components into stoke combinations that are not necessarily meaningful (see, in particular, the fifth part on his/her list). Eventually, the answer was only partially correct in that the visual structure of the character was presented as only an arbitrary repetition of parts of components in a sequence he/she was used to when writing the character.

  1. 4.

    Hybrid agglomeration of sequenced strokes and components (see Fig. 5.4):

    Fig. 5.4
    figure 4

    Character splitting—hybrid agglomeration of strokes and components

    Full size image

In the above example, the character “災” (lit. “disaster”) was split into four individual parts, three of which were individual strokes making up the top half of the character, whereas the remainder was the component “火” (lit. “fire”) making up the bottom half of the character. Like his/her peer who produced in the example shown in Fig. 5.3, this participant seemed to have developed a certain level of awareness of character components giving him/her discernment of “火” from the overall structure of 災 as a correct component, yet his/her orthographic knowledge did not seem to be sufficient for him/her to discern “巛” (archaic; lit. “river”) as another component of the target character. In other words, he/she still seemed to be confused between stroke and component in written Chinese. As a result, his/her answer was only partially correct in that the visual structure of the character was perceived as a hybrid agglomeration of three strokes (in a sequence, he/she might be used to when writing the character) and a component (i.e., 災 having four parts rather than being composed of two components).

  1. 5.

    Agglomeration of components (see Fig. 5.5):

    Fig. 5.5
    figure 5

    Chinese character splitting—agglomeration of components

    Full size image

In the above example, the character “霜” (lit. “frost”) was split into three individual parts, namely “雨” (lit. “rain”), “木” (lit. “wood”) and “目” (lit. “eye”). The participant offered the correct answer by identifying the character as an agglomeration of the three components.

Task 2: Constitution of Chinese Character Components Test

Thirty-nine legitimate characters were created by the participants using the seven components provided in the task. These characters are listed below in a descending order of occurrence (more specifically, the proportion of the 213 participants who produced each character) (see Table 5.4).

Table 5.4 List of legitimate Chinese characters created with the given components by the participants (in descending order)
Full size table

The ten most frequently occurring characters were “想” (lit. “want,” “think”), “好” (lit. “good”), “如” (lit. “if”), “相” (lit. “inspect,” “together”), “林” (lit. “forest”), “森” (lit. “forest”), “仔” (lit. “son,” “small,” “young”), “休” (lit. “stop,” “rest”), “品” (lit. “taste,” “quality”) and “李” (lit. “plum,” and a common family name too). Among them, six carried the component “木” (lit. “wood”) and three carried the component “子” (lit. “child”), with left-right (6 characters) and top-down (4 characters) spatial configurations (3 characters with a combination of both) as the most frequently occurring ones. All of the aforementioned characters were commonly used in both Hong Kong and Macau and were relatively familiar to the participants (e.g., 87.3% participants answered the Task 1 question on “想” correctly).

The ten least frequently occurring characters were “惢” (lit. “suspicious”), “恷” (lit. “violent”), “众” (lit. “crowd”), “囚” (lit. “to imprison”; “the imprisoned”), “杺” (lit. “a type of tree with yellow heartwood”), “从” (lit. “to follow,” “belonged to”), “囝” (lit. “child”), “昍” (lit. “bright”), “孨” (lit. “weak,” “cowardly”), and “喿” (lit. “chirping of birds”). Among them, only “众”, “囚”, “从”, and “囝” are considered commonly used characters, with “众” and “从” as simplified equivalents of “眾” and “從”, respectively. Three of them carried the component “心” (lit. “heart”); another three carried the character/component “木” (lit. “wood”), with left-right and top-down combined (5 characters), left-right (3 characters), and surrounding (2 characters) as the most frequently occurring spatial configurations . Given that all of the aforementioned characters were rarely used in both Hong Kong and Macau (and thus unlikely taught in school curriculum), it is highly likely that they were guesses made by the participants using their orthographic knowledge of spatial configurations .

Task 3: Chinese Character Structure Test

The mean and standard deviation of the 213 participants’ performance on the Character Structure Test were 0.65 and 0.129, respectively. Cronbach’s alpha was 0.623, indicating an acceptable level of reliability . Among the 213 participants, 28 of them (13.1%) earned 0.80 or above, whereas 23 of them (12.3%) earned lower than 0.50. Most participants fell into the 0.50–0.77 range (76.1%). Only 4 participants scored 0.30 or below, making up a mere 1.9% of the cohort. These findings seem to suggest that the participants as a group possessed a certain level of awareness in spatial configurations of character components despite a lack of formal training in orthographic structure of characters.

Our item-level analysis revealed that the five target characters (out of 30) with the largest number of correct responses from the 213 participants were “做” (lit. “to do”) (left-right; 93.9% correctness); “的” (lit. “something/somebody’s”) (left-right; 91.5% correctness); “和” (lit. “and,” “peace”) (left-right; 91.5% correctness); “晶” (lit. “crystal”) (top-down and left-right; 88.7% correctness); and “月” (lit. “moon”) (single component; rectangle; 86.4% correctness). The five characters with the smallest number of correct responses were “我” (lit. “I”, “me”) (single component; square; 39.0% correctness); “問” (lit. “ask”) (half-surrounding and top-down; 34.7% correctness); “時” (lit. “time”) (left-right; 28.6% correctness); “為” (lit. “for”) (diagonal; 19.7% correctness); and “喉” (lit. “throat”) (left-right; 15.0% correctness).

Relationship of Participants’ Demography and After-School Activities with Orthographic Knowledge

As shown in Table 5.5, among the four major ethnic groups (i.e., the Filipinos, Pakistanis, Nepalese and Indians), based on the sample means, the Filipinos appeared to score the highest whereas the Indians the lowest for Task 1. For Task 2, the Pakistanis appeared to score the highest whereas the lowest-scoring group was the Filipinos. The Filipinos appeared to score the highest for Task 3, with the Indians as the lowest-scoring group (see Table 5.5). Results of the ANOVA for each task, however, showed that ethnicity had a significant effect on the learners’ performance on Task 2 only (F = 5.679, p < 0.001); no significant differences among the four ethnic groups were observed for Tasks 1 and 3.

Table 5.5 Mean scores of Tasks 1–3 according to ethnic groupings
Full size table

For all three of the tasks, ANOVA revealed no significant effect of the age at which the participants arrived in Hong Kong or Macau (see Table 5.6).

Table 5.6 Mean scores of Tasks 1–3 according to age of arrival in Hong Kong or Macau (N = 213)
Full size table

We also analyzed if the participants’ television (TV)-viewing habits would have any effect on their performance on Tasks 1 to 3. As shown in Table 5.7, the mean scores of the participants who reported watching TV after school appeared to outperform those who did not on all three tasks. Further t tests, however, showed that participants who have more TV viewing performed significantly better in Task 2 only, and there was no significant effect of TV viewing on the performance on Tasks 1 and 3.

Table 5.7 Mean scores of Tasks 1–3 as related to participants’ television-viewing habits (N = 213)
Full size table

Table 5.8 further shows the mean scores of those TV-watching students (168 in total) as related to the language of the TV programs they watched (i.e., if they watched TV programs in Chinese/with Chinese subtitles or programs not in Chinese). The sample means in Table 5.8 appeared to suggest that watching Chinese TV programs or TV programs with Chinese subtitles helped to improve the participants’ ability of using given character components to construct Chinese characters (i.e., Task 2), and there did not appear to be any effect on their ability of separating the components for characters (Task 1) or identifying the spatial configurations of characters (Task 3) (see Table 5.8). Further t tests confirmed that the participants who watched Chinese TV programs performed significantly better in Task 2 only.

Table 5.8 Mean scores of Tasks 1–3 as related to participants’ choice of television programs (N = 166)
Full size table

Finally, ANOVA revealed that the participants’ time spent on studying Chinese outside of class did not have a significant effect on their performance on any one of the three tasks. However, based on the sample means presented in Table 5.9, it appeared that there was some positive effect on their Chinese orthographic knowledge at or beyond the 3-h threshold in a week. Those participants who reported having spent three to 7 hours per week studying Chinese appeared to earn the highest mean scores for Tasks 1 and 2, namely 0.801 (Task 1) and 0.213 (Task 2), respectively. No such tendency, however, was found for Task 3, in which the “More than 7 hours” group appeared to earn the highest mean score (i.e., 0.683).

Table 5.9 Mean scores of Tasks 1–3 as related to time spent on studying Chinese (N = 213)
Full size table

Discussion

Participants’ Task Performance

To answer the first research question, despite the lack of systemic, formal training in component-related orthographic knowledge , there was a relatively high percentage of the participants earning 0.8 or above (57.9%) for Task 1 (i.e., Separation of Character Components). We argue that this fairly good performance among the participants might be attributed to their knowledge of Chinese (semantic) radicals , which teachers from both schools claimed as an essential part of their Chinese language (CL) curricula. The students’ prior training in Chinese radicals might have played a role in enhancing their orthographic awareness with respect to character components in general. At the item level, the fact that they carry common components with simple spatial configurations such as top-down, left-right, and left-middle-right seemed reasonable for the largest number of correct responses to occur for characters such as “災”, “初”, “想”, “蝴” and “霜”. Likewise, the less common components and relatively sophisticated, multi-tiered spatial configurations of the characters such as “藥”, “糕”, “臉”, “晴”, and “問” might explain the finding that they received the smallest number of correct responses.

As for Task 2 (Constitution of Chinese Character Components Test), the results showed that the participants were able to create at least a few characters from the list of common components with which the students should be rather familiar. The top ten characters produced by the students (see Table 5.4) were actually basic vocabulary items in CSL beginners’ curricula; their frequent appearance seemed to indicate that the participants’ mental lexicon was accessed during the test. Among the 10 characters, six were of a left-right structure and four a top-down structure (three actually involved a combination of both), which seems to suggest that it was easier for the participants to master these two spatial configurations of components.

One should note that obsolete or rarely used characters also came up on the list of legitimate characters produced by the participants in Task 2. Their frequency of appearance, however, was very low (see Table 5.4). Given that these characters are either rare characters seldom used in daily life or simplified Chinese characters , they were supposed to be totally new to the participants. It seemed evident that they were educated guesses of those who produced them. On the other hand, the guesses might constitute a good piece of evidence that supports the active use of orthographic knowledge among those students. In other words, the fact that these characters are legitimate ones suggests the active engagement of some EM students, albeit small in number, with the components they were familiar with as well as their knowledge of common spatial configurations of components of Chinese characters (e.g., left-right and top-down or a combination of them).

Results of Task 3 (Chinese Character Structure Test) showed that the participants as a whole group seemed to possess a considerable level of structural awareness of Chinese orthography, which tends to support some of the evidence obtained from their performance on the other two tasks discussed earlier. Although knowledge of spatial configurations was not formally taught in either school, the students overall displayed a considerable level of ability (with 87.8% of them earning a score of 0.50 or above) to discern different spatial configurations.

On the other hand, it is interesting to note that all the items that most participants got correct were highly frequent characters that were supposed to have been learnt by the participants at the elementary level with relatively simple and easily recognizable spatial configurations. The five items that most participants failed to get correct, on the contrary, appeared to have ambiguous character components and spatial configurations (e.g., “我”, “問”, “時”, “為”, and “喉”). Thus, the participants’ prior character knowledge and the complexity of spatial configurations of character components seemed to be a major contributing factor of their achievement in Task 3.

Influence of Demography and After-School Activities

To answer the second research question, with reference to the sample mean scores (see Table 5.5), it appeared that the Filipinos scored the highest in Task 1 and the Indians the lowest; the Pakistanis scored the highest and the Filipinos the lowest in Task 2; and the Filipinos scored the highest in Task 3, and the Indians the lowest. However, ANOVA revealed that ethnicity actually did not have any significant effect on Tasks 1 and 3 as opposed to Task 2. Because of a lack of information about the specific formal and informal Chinese learning experiences of these ethnic groups, it was unclear to us why such patterns existed.

The age at which the participants arrived in Hong Kong or Macau did not have any statistically significant effect on their mean scores of the three tasks. Based on the mean scores, it seemed that the participants who immigrated to Hong Kong/Macau between “0 and 3 years old” scored the highest in Task 2; the participants who immigrated to these two places between “0 and 6 years old” scored highest in Task 3; and the participants who immigrated between “3 and 6 years old” scored the highest in Task 1. However, no between-group differences can be found from the ANOVA results. It suggests that in these places, age of arrival might not be a good indicator of immigrant students’ actual exposure to the target language or general language proficiency in Chinese and, consequently, did not show any effect on actual performance on the three tasks. The development of Chinese orthographic knowledge or overall Chinese proficiency might well be the result of classroom instruction in Hong Kong and Macau. Such a speculation also seems to agree with the findings discussed below that after-school or informal experiences with Chinese to a larger extent did not have any significant effect on the participants’ measured orthographic competencies in this study.

With regard to the participants’ TV-viewing habits, the t tests indicated that those who watched TV programs, especially those with Chinese subtitles, performed significantly better in Task 2, but not Tasks 1 and 3. In other words, watching TV programs with Chinese subtitles helped to improve the participants’ ability to use components to construct Chinese characters, but not breaking whole characters down into components and identifying spatial configurations. This finding might be explained by the different nature of the tasks. More specifically, Task 2 was based on orthographic sensitivity and mental lexicon, both of which could be possibly enhanced via TV viewing, whereas Tasks 1 and 3 seemed to require systematic input on orthographic knowledge, which is not possible to deliver in the form of self-learning in informal contexts through TV captions.

Time spent studying Chinese outside of class did not have a significantly positive effect on student performance, even though based on the sample mean scores (see Table 5.9), those who spent 3 hours or more studying Chinese outside of class appeared to have greater performance than their counterparts in Tasks 1 and 2, and those who spent “more than 7 hours” studying Chinese outside of class appeared to perform the best across all of the three tasks. The fact that no systematic training on Chinese orthographic knowledge had been provided to the cohort before the assessments might explain why the time spent studying Chinese outside of class did not show any statistical significance related to mean scores. It also seems to suggest that language input from their self-study, as opposed to classroom instructional experience, might not be directly related to the students’ orthographic knowledge being tested in the three tasks.

Implications for CSL Teaching and Learning

Traditionally, Chinese character teaching tends to place a lot of emphasis on “roots” (i.e., radicals) based on a classification system that used to be indispensable for dictionary use, “stroke order” based on a set of ancient calligraphic principles, as well as the etymology of single-component characters . Radicals are, by no means, equivalents of components in a strict sense, as the indexicality of radicals lies in the etymology that serves lexicographical purposes, whereas that of components lies in the structural functionality that serves pedagogical purposes. The traditional pedagogy, which lacks explicit attention to spatial configurations of the components in multi-component characters , however, fails to address the needs of CSL learners, particularly EM students with diverse sociolinguistic backgrounds, who often lack the level of exposure to print and cultural input that their native-speaking peers commonly have. Another reason that the radical-oriented traditional pedagogy of character teaching might not be as productive to EM students is that certain radicals in the Traditional Chinese script can be highly complicated in form (extreme examples include “龍” [lit. “dragon”] and “龜” [lit. “tortoise”]), which are not easily identified or deciphered by CSL learners. Thus, teaching the skills related to spatial configurations of character components (such as those featured in the present study) seems a comparatively sustainable approach to the learning and teaching of Chinese characters among CSL learners.

Preliminary findings from Loh, Mak, and Tam (2015b) indicated that students with low Chinese proficiency possess a number of characteristics: (1) they cannot discriminate between the structures and components of Chinese characters and only rely on their intuition for random speculation; (2) due to their poor knowledge of character structures and components, they tend to mix up components with similar visual characteristics, while using wrong components to construct wrong characters; and (3) they see Chinese characters as individual images, while selecting parts of a character they like most and making free or even random associations in an attempt to analyze the orthographic structure . On the contrary, students with good Chinese language proficiency possess a different set of characteristics: (1) they are able to discern between various character components and even identify the phonetic and the (semantic) radical components in a character; (2) they tend to be able to make use of their knowledge of phonetic and radical components to construct new characters; and (3) they are able to accurately identify spatial configuration of Chinese characters.

The findings of the current study seem to echo the aforementioned research. The high-scoring participants in Task 1 could discern between different components and even show a good concept of Chinese orthographic structures beyond strokes and stroke sequences without formal, systematic training in Chinese components. Task 2 was the most difficult one among the three, yet the high-scoring participants were able to apply their prior knowledge (including mental lexicon and their vague concept of spatial configurations) to create characters, which was well-supported by the list of archaic, yet legitimate, characters in the answers which might not be understood by the participants themselves. The same applies to Task 3 in which high-scoring participants did not receive any formal training in Chinese spatial configurations prior to the study, and there appears to be a widespread use of common sense and visual matching when tackling the easily recognizable items.

In light of the above studies, and the different learning needs of L1 and L2 learners of the Chinese language, specific curricula and teaching materials that enable systematic acquisition of Chinese orthographic knowledge based on L2 learners’ level of proficiency and stage of learning would be highly desirable. These findings also suggest that formal training in character components and their spatial configurations could show great potential to reinforce L2 learners’ orthographic knowledge and would equip EM students, and learners in general, with the skills and awareness they need for learning to read and write in their target language. In what follows, we propose a few specific recommendations for teaching Chinese orthographic knowledge to EM learners or CSL learners, who typically do not receive any formal training in the discernment and application of character components and their spatial configurations:

Teaching Chinese Character Components in a Systematic and Contextualized Manner

As indicated earlier in this chapter, the Chinese radical system was originally devised for L1 learners to study etymology. Radicals include not only components but also whole characters, which can easily confuse CSL learners as they do not have an equal Chinese culture input as their L1 counterparts. Teaching CSL learners systematically about character components and spatial configurations instead would help learners acquire advanced orthographic knowledge at large, which is essential for achieving higher levels of Chinese proficiency. Such advanced orthographic knowledge includes the relationship between orthographic form, sound, and meaning, as well as the proportion, symmetry, and size of a component as constituting parts of Chinese characters. Based on the hypothesis of comprehensible input (Krashen 1982) and variation theory (Marton and Booth 1997; Marton et al. 2004), Tse et al. (2007) proposed the Integrative Perceptual Approach to Teaching Chinese Characters , which highlights the importance of systematic training on Chinese orthographic knowledge (inclusive of character components and spatial configurations), which is essential for the automatization of Chinese character recognition and a prerequisite for the development of Chinese reading ability. Such an approach would be a good starting point for relating the teaching of Chinese characters’ components to students’ mental lexicon and everyday life along with their development of all four skills in CSL learning.

Enhancing Students’ Orthographic Awareness with Exercises Based on the Components that Have Been Covered in Their CL Lessons

Most EM CSL learners are L1 speakers of alphabetic languages. Their knowledge components, such as vocabulary, are built up step-by-step. Component-related exercises could strengthen students’ character writing skills; and exercises with component constitution games, such as component cards, flash cards, and access to mental lexicon (Loh et al. 2015; Tse 2000) could also elevate their levels of motivation and confidence in learning Chinese through peer collaborative learning and communicative language teaching in the CSL classroom.

Teaching Students About Spatial Configurations of Chinese Characters with Special Reference to Component Position, Symmetry, and Proportion

While teaching Chinese components to EM CSL learners, it would be important to start introducing the two levels of spatial configurations specific to Chinese characters (Kao and Chen 2012), at an early stage to enhance their understanding of the orthographic concepts for the Chinese language. The two levels of spatial configurations are namely (1) the relationship among strokes in a component or single character (e.g., “士” vs. “土” with different lengths of strokes signifying different things) and (2) the relationship among components in a character (e.g., the ways in which different forms of the same component “人” are used for different positions in a character). With reference to the findings of the current study, particularly the most frequently occurring characters in Task 2, we have identified certain common components (e.g., “木” and “子”) and spatial configurations (i.e., left-right and top-down) more easily recognized by the participants. Teachers may consider introducing the concept of spatial configurations in Chinese orthography for CSL beginners with the aforementioned configurations and components to facilitate better learning for their students, such as teaching the left-right configurations with common components “木”, “女”, and “子” (as in “林” and “好”) instead of starting with the less easily recognized ones (e.g., half-surrounding and top-down) with more complicated components (e.g., “門”, as in “問”).

Conclusion

In this chapter, we reported a study on the orthographic knowledge among EM students at various levels in Hong Kong and Macau who learned Chinese as a second language. The participants overall performed better on separating characters into their constituent components (Task 1) and identifying the spatial configurations of characters (Task 3), which tapped their passive knowledge of Chinese orthography, than on producing characters with given character components (Task 2), which required more active knowledge of applications of character components. The current findings on the patterns in EM adolescent CSL learners’ perception of Chinese characters also echoed with our previous research on Chinese reading strategies used by learners with different proficiency levels (Loh et al. 2015). Without the ability to discern among different levels of Chinese orthographic structures, it would be difficult for low-level or beginning CSL learners to develop effective reading strategies, which might cause them to resort to random association of images with specific (parts of) characters.

This study sheds light on curriculum development and pedagogies for cultural and academically diverse EM CSL learners, especially with respect to effective approaches to teaching orthographic knowledge for learners with different proficiency levels in CSL classrooms. The results on the students’ orthographic knowledge project their basic competency required for reading in the current CL and CSL curricula, and pave the way for an ongoing study on the relationship between orthographic knowledge and students’ performance in Chinese learning. Teaching of character components and their spatial configurations, as revealed in this chapter, would be an essential scaffold in EM CSL curricula for reinforcing learners’ orthographic awareness before extensive text-based reading and writing practice. In this spirit, the Integrative Perceptual Approach for Teaching Chinese Characters is recommended for beginner- and intermediate-level CSL teaching.