You are on page 1of 24

Reading and Writing: An Interdisciplinary Journal 16: 399422, 2003. 2003 Kluwer Academic Publishers. Printed in the Netherlands.

399

Development of morphological awareness in Chinese and English


YU-MIN KU and RICHARD C. ANDERSON
University of Illinois at Urbana-Champaign, Illinois, USA Abstract. The development of morphological awareness in Chinese and English was investigated in the current study involving 412 Taiwanese and 256 American students in second, fourth, and sixth grades. The results from both Chinese-speaking and English-speaking students indicate that the morphological awareness develops with grade level and is strongly related to reading ability. More procient readers outperformed less procient readers when asked to (1) recognize morphological relationships between words, (2) discriminate word parts having the same or different meanings, (3) select the best interpretations of lowfrequency derivatives and compounds composed of high-frequency parts, and (4) judge the well-formedness of novel derivatives and compounds. Chinese students acquisition of derivational morphology seems to lag behind that of compounding rules, which might reect the nature of Chinese word formation in that there are far fewer derivatives than compounds in Chinese. Key words: Chinese, Morphological awareness, Vocabulary acquisition, Word formation

In daily life, we use or nd others using words that we have never heard before, but we understand each other without further explanation of the novel words. It seems that language users, especially adult native speakers, understand the internal structure of words, and when they need a new word to express an idea, they may invent a word that ts word formation rules. There is evidence from both Chinese and English studies that adults treat complex words analytically. One sort of evidence is that morphological relationships between prime and target words inuence performance in lexical decision and word recognition tasks in both languages (Nagy, Anderson, Schommer, Scott & Stallman, 1989; Taft & Forster, 1976; Taft & Zhu, 1995; Zhang & Peng, 1992). When and how children develop implicit knowledge of the structure of words are fascinating questions, ones of potential importance for understanding childrens reading development. The lions share of reading research has investigated awareness of the phonological structure of words (e.g., Blachman, 2000; Bradley & Bryant, 1983; Goswami, 2000; Ho & Bryant, 1997; Shu, Anderson & Wu, 2000), although previous studies have linked morphological awareness and reading prociency in the early school years (Carlisle & Nomanbhoy, 1993; Casalis & Louis-Alexandre, 2000, Mahony,

400

YU-MIN KU AND RICHARD C. ANDERSON

Singson & Mann, 2000), in the middle school years (Leong, 1989; Tyler & Nagy, 1989), and in high school and college (Mahony, 1994).

Morphological awareness and vocabulary acquisition Scholars have long believed that morphological awareness is important in vocabulary growth (Dale, ORourke & Bamman, 1971; Nagy & Anderson, 1984; Sternberg, 1987; White, Power & White, 1989). For example, Sandra (1994) suggested, morphology could be a powerful device for facilitating the acquisition of polymorphemic vocabulary items and improving the retention of such items (p. 261). Similarly, Carlisle (1995) proposed that morphological awareness might be particularly important because morphological decomposition and problem-solving provide one way to understand and learn the large number of derived words used in the books they read (p. 205). A large proportion of the unfamiliar words children encounter are morphologically complex. Nagy, Osborn, Winsor and OFlahavan (1994) estimated that, among the 10,000 unfamiliar words that an average American fth grader might encounter in reading over the course of a year, about 4,000 would be derivatives of more frequent words. According to Anglins (1993) study of American childrens vocabulary growth between the rst and fth grades, the increase in number of derived words known is over three times greater than the increase in number of root words known by the same children (14,000 derivatives versus 4,000 root words). This high rate of increase in knowledge of derived words presumably reects a process of acquisition that depends on morphological analysis, at least in part. How does childrens morphological awareness relate to vocabulary acquisition? The prevalent belief is that children who are knowledgeable about morphology decompose unfamiliar words into familiar meaningful units prexes, roots, and sufxes and then derive the meanings of the words by combining the units. For example, the prex dis- means not or do the opposite, so when encountering the word disobey for the rst time, children would have an excellent chance of getting the right meaning, not to obey, by using their morphological knowledge. The process of breaking an unfamiliar word into units, and then recombining the units into a meaningful whole, enables children to gure out the meanings of newly encountered words and may enhance memory for these words. Thus, morphological awareness is considered an important factor in childrens rapid vocabulary growth during the elementary school years (Nagy & Anderson, 1984; Tyler & Nagy, 1990; White, Power & White, 1989). Can children, in fact, infer word meanings through morphological analysis? Beginning with Freyd and Baron (1982) several studies have given

DEVELOPMENT OF MORPHOLOGICAL AWARENESS

401

a positive answer to this question. Representative of these studies is one by Wysocki and Jenkins (1987), who studied whether children in grades four, six, and eight could derive word meanings through morphological analysis. Children were rst taught a set of words and then two weeks later were asked to dene words that were either derivationally related or not related to the words they learned in the training session. For example, words like doting, stipulation, and repudiate were taught, and the test included derivationally related words, such as dote, stipulate, and repudiation, as well as unrelated words, such as transgress, abate, and incipient. The results indicated that children in all grades performed somewhat better on words that were derivationally related to previously learned words. This nding conrms the idea that children can derive word meanings through morphological analysis. Childrens ability to decompose unfamiliar words might have been underestimated by Wysocki and Jenkins (1987) because the children may have tried to analyze the morphological structure of the words in the unrelated set, but failed because they lacked knowledge of the base words. The tasks employed by Wysocki and Jenkins (1987) were difcult for elementary school children, because they had to learn new words and then infer the meanings of derivatives of these words. Tyler and Nagy (1989) and White et al. (1989) used simpler tasks. These two studies suggested better understanding of English derivational morphology than previous studies. Children were asked to dene morphologically complex words by Anglin (1993), Lewis and Windsor (1996), and Carlisle (2000). Although delity to adult standards for denitions was not high 20% and 32% for third and fth graders in Carlisle, and 35% for students in grades four through eight in Lewis and Windsor it was evident that students had some ability to use knowledge of word parts to dene words. For example, a child demonstrated his understanding of the agentive sufx -ist in humanist by providing the denition someone who studies human. Can Chinese children infer the meanings of unfamiliar words through morphological analysis? Many scholars (Hatano, Kuhara & Akiyama, 1981; Hoosain, 1992; Shu, Anderson & Zhang, 1995; Tang, 1988) believe that morphological awareness plays an important role in Chinese (and Japanese) reading; however, very little systematic research has been reported in the literature. A Chinese character usually corresponds to a single morpheme and characters are the building blocks of longer, more complex words. According to Hoosain (1992), meanings of the constituents of polymorphemic Chinese words are more manifest than often is the case with constituents of multimorphemic English words (p. 115). Thus, it might be easier for Chinese readers than English readers to encode and retrieve the meanings of polymorphemic words (Hoosain, 1991; Nagy & Anderson,

402

YU-MIN KU AND RICHARD C. ANDERSON

1998). Some evidence for this conjecture comes from Shu, Anderson and Zhangs (1995) cross-cultural study of how Chinese and American children learn unfamiliar words from context. As compared to American children, Chinese children were more likely to learn the meanings of morphologically transparent words than of morphologically opaque words. This implies that Chinese children made more use of morphological analysis to assimilate word meanings. In sum, research on childrens vocabulary acquisition has clearly demonstrated the importance of morphological awareness for learning word meanings. Understanding the morphological structure of words enables children to gure out, encode, and remember the meanings of unfamiliar words. There is some indication that morphological awareness is more important for word acquisition in Chinese than in English.

Childrens acquisition of morphological knowledge In one of the rst studies, Berko (1958) assessed the ability of American children, ranging in age from ve and one half to seven, to produce plural and past tense inections. Children saw a picture card and heard an oral text missing a target word. They were asked to supply the missing word. For example, a child was shown a card with a bird-like animal and then two bird-like animals. At the same time, the child heard the experimenter say, This is a wug. Now there is another one. There are two of them. There are . The child was supposed to say wugs. Berkos results indicated two that children in the preschool years have already acquired some knowledge of regular inections. Analyses of young childrens speech have provided further evidence supporting the notion that children have acquired basic rules of English inection and are able to use the rules productively and spontaneously at a relatively young age (Kuczaj, 1977; Marcus, Pinker, Ullman, Hollander, Rosen & Xu, 1992). For example, children may overgeneralize the rules for creating the past tense, producing words like goed, and eated, and for creating plurals, producing words such as mans and foots. The productive use of compounding rules has been observed in spontaneous speech among children as young as three years of age. Clark (1981) claimed, Children, like adults, innovate in order to ll lexical gaps (p. 307). In other words, children may create new words to express intended meanings when they lack well-established terms in their lexicons. For example, children may coin compound words like x-man for a car mechanic, garden-man for a gardener, and car-smoke for exhaust. Children understand the basic relationship between the component terms in modier-head compounds at an early

DEVELOPMENT OF MORPHOLOGICAL AWARENESS

403

age, for instance, that doghouse refers to a type of house whereas housedog refers to a type of dog (Clark, 1981; Clark, Gelman & Lane, 1985). However, there has been little research on the development of awareness of compounds over the elementary school years. Research with adults who speak various languages, including English and Chinese, suggests that transparent compounds are processed in terms of constituent morphemes whereas opaque compounds are processed as a whole (e.g., Dominiek, 1990; Wang, Peng, Guan & Kuang, 1999). It is uncertain when school children begin to see transparent compounds in terms of constituents. In one early study, Silvestri and Silvestri (1977) obtained results which they interpreted to mean that kindergarteners represent familiar transparent compound words as unanalyzed wholes whereas by fourth grade children understand how the meanings of components contribute to the meaning of an entire compound. School English contains morphologically complex words with a wide range of semantic transparency. According to Nagy and Anderson (1984), many complex words in school English have meanings that are totally predictable (p. 310) from constituents (e.g., senselesssenselessly, wash washcloth). At the other extreme are words in which there is no discernible semantic connection (p. 311) between a constituent and the whole word, at least for a modern-day child as opposed to an historical linguist (e.g., x prex, groovegroovy). Nagy and Anderson (1984) found that the frequent words that predominate in the early grades are less transparent than the less frequent words introduced in higher grades, which might condition the age at which children acquire insights into morphology. Childrens acquisition of derivational knowledge is the most studied aspect of morphological knowledge among school-age children. In her pioneering study, Berko (1958) reported that preschoolers and rst graders were unable to produce proper derived forms. For example, when the experimenter asked children what they would call a man whose job is to zib, about 90% of children failed to use the agentive afx -er to produce zibber. However, Carlisle and Nomanbhoy (1993) found that when a derived form preserved the pronunciation of its base word, the correct response rate was 40.9%, whereas when the derived form required phonological changes from its base word, the rate was 11.2%. One possible explanation for the different results in these studies is that Berko (1958) used nonsense base words whereas Carlisle and Nomanbhoy (1993) used real words that children might have known. Tyler and Nagy (1989) conducted a comprehensive investigation of fourth-, sixth-, and eighth-grade American students knowledge of different aspects of derivational sufxes. They found that students at fourth grade level already had the knowledge of the relational aspect of the sufx and were

404

YU-MIN KU AND RICHARD C. ANDERSON

able to capitalize on the relationship between the derivatives and the base. Childrens knowledge of the syntactic contribution of the sufx increased with grade level. Similarly, childrens knowledge of distributional constraints on the use of the sufx also increased with grade but was acquired later than relational and syntactic knowledge. Derwing and Baker (1979) employed the comes from task for example, having people judge whether knowledge comes from know to investigate the inuence of phonetic and semantic features on the ability to recognize morphological relationships. Results indicated that the ability to recognize real morphological relationships and reject false ones increases with age. Younger subjects tended to accept false relationships based on obvious similarity in a single feature (catkitty). Older subjects were able to recognize nonobvious relationships that are taught in school and also better able than younger subjects to recognize relationships when word parts differed in pronunciation but had the same spellings (breakbreakfast). Carlisle (1988) and Leong (1989) categorized relationships between a derived word and its base word into four types, depending on whether the derived word involves an orthographic or phonological change from its base. The four types are: (1) no change (e.g., carecareful), (2) orthographic change (e.g., beginbeginner), (3) phonological change (e.g., electricelectricity), and (4) both orthographic and phonological change (e.g., deepdepth). A consistent nding is that students are most likely to produce derived forms when no change in orthographic or phonological form is required (Carlisle, 1988; 1995; Fowler & Liberman, 1995; Leong, 1989). Students make more errors in producing derived forms that require phonological or both orthographic and phonological change. Carlisle (1988) found that 82% of childrens errors involved words that required a phonological change to make the derived forms, and most of the errors involved simply retaining the phonological form of the base word. This nding conrms Tyler and Nagys (1989) thesis; Children will tend to acquire strategies which make as few changes as possible when forming a new word from an old one (p. 665). The smaller the change in the structure of the derived words, the easier it is to acquire these words. The inuence of phonological or orthographic change from a base to its derived form was also investigated by Mahony (1994) and Mahony, Singson and Mann (2000), but the task in these studies was to recognize the morphological relatedness of pairs of words. Good readers and older students correctly identied more related word pairs than poor readers or younger students. However, Mahony and her colleagues concluded that the degree of phonological distortion in the derived form does not inuence students judgment of morphological relatedness. It seems that phonological and orthographic

DEVELOPMENT OF MORPHOLOGICAL AWARENESS

405

variability affect production (Carlisle, 1988; Leong, 1989) but not recognition of morphologically complex forms (Mahony, 1994; Mahony et al., 2000). Chinese childrens development of morphological awareness through the elementary school years has not been studied until recently. Li, Anderson, Nagy and Zhang (2001) developed several tasks to examine the morphological awareness of rst and fourth graders from the Peoples Republic of China. In one task employed in the rst grade, students were introduced to a pair of homophonic characters in the context of familiar words. Then the experimenter orally presented another two-character word and asked students to choose which written character was appropriate for use in this word. For example, the experimenter showed the homophone characters /xin1/ (heart) and /xin1/ (new) and explained that the character can be used in words such as /xin1zang4/ (heart) while the charcan be used in words such as /xin1nian2/ (new year). Then acter the experimenter presented another two-character word familiar from oral language but unfamiliar in written form, /xin1wen2/ (news). Students were to write this new word. Another task supposed to select the character employed by Li and her colleagues (2001) with rst graders involved orally presenting three two-character words that shared a syllable. In two of the words, the syllable represented the same morpheme. Students had to indicate which word contained a character that was different from the other two. /hong2cha2/ (black tea), /lu4cha2/ For example, students heard /jian3cha2/ (to examine). All three words contain the (green tea), and syllable /cha2/, but only the rst two represent the same morpheme. Li et al. (2001) found that good readers outperformed poor readers on tasks assessing morphological awareness in both the rst grade and the fourth grade. Wang (1999) used some of the same tasks with Taiwanese rst and second graders and also found that performance on the tasks was related to reading prociency.

Goals of the present study Research in developmental psycholinguistics has provided evidence that morphological awareness plays an important role in English-speaking childrens vocabulary and reading acquisition. However, it is not yet wellestablished whether morphological awareness is also important for children learning Chinese, a very different language with a very different writing system. The present study is designed to investigate whether morphological awareness contributes to the vocabulary acquisition and reading prociency of Chinese children as well as American children.

406

YU-MIN KU AND RICHARD C. ANDERSON

Table 1. Types of morphemes and examples in Chinese and English. Morpheme type Root word Bound root Inectional afx Derivational afx Chinese /shan1/ (mountain) /gou3/ (dog) /fang2/ (house) /zhuo1/ (desk) /le/ verbal aspect /men/ plural /wu2/ (not) /hua4/ verbalizing English Book Hand Anti- (against, opposite) -logy (study) -ed (past tense) -s (plural) -er (agentive) -ly (adverb)

Chinese is different from English in terms of orthographic symbols and grapheme, morpheme, and phoneme relations. However, the word formation rules of Chinese are similar to those of English. Morphemes in Chinese and in English can be classied in terms of the same basic elements. In each language, in addition to free morphemes or root words, there are three types of bound morphemes: (1) inectional sufxes, which are grammatical morphemes that change a root words aspect, tense, number, case, etc.; (2) derivational afxes, which usually change the part of speech of roots; and (3) bound roots, which must combine with derivational afxes or other roots to form words. Table 1 lists examples of the different types of morphemes in Chinese and English. In each language, through afxation and compounding, three types of words are formed: inected words, derived words, and compounds. Despite the similarities in word formation, there are some differences between Chinese and English morphology. First, the primary way of forming words in Chinese is to combine roots. Second, although there are far fewer inectional and derivational afxes in Chinese than in English, there are numerous bound roots in Chinese. Unlike most bound roots in English (e.g., anti- and tele-), Chinese bound roots are less positionally restricted (Packard, 2000). Third, because approximately 89% of Chinese characters represent unique morphemes, characters usually provide the reader with visually distinct and reliable cues for decomposing polymorphemic words. This study addresses the question of whether, because of language-specic differences in morphology, Chinese- and English-speaking children follow different courses in developing morphological knowledge. In other words, the current study attempts to determine, not only whether there are common elements in the acquisition of Chinese and English morphological knowledge,

DEVELOPMENT OF MORPHOLOGICAL AWARENESS

407

but in addition whether there are certain aspects of morphological knowledge that are important for one language but not for the other. Comparable Chinese and English tasks were developed with the goal of equating the difculty of each task in the two languages. To control for familiarity, Chinese and English words were equated in frequency of usage. To control for degree of semantic transparency and degree of morphological relatedness of words, native speakers of Chinese and English rated the words used in the tasks. Because, unlike English, word formation in Chinese seldom involves changing phonological or orthographic form, tasks contained only word pairs from the two languages that share the same orthographic and phonological form. In the second grade classrooms in the two countries, materials were read aloud by the teacher; this was intended to minimize the possibility that performance would reect variation in ability to read the words rather than variation in morphological knowledge. In summary, comparable tasks were constructed to assess Chinese- and English-speaking childrens awareness of morphology. The general hypothesis was that among both groups of children morphological awareness would increase with language experience and correlate highly with vocabulary knowledge and reading prociency. One specic hypothesis was that morphological awareness would be more strongly related to reading prociency among Chinese-speaking than English-speaking children. The basis for this hypothesis is that, unlike English, in Chinese [a] word formation hardly ever involves phonological or orthographic change and [b] most Chinese characters represent only one morpheme. Another specic hypothesis was knowledge of derivational morphology would develop more slowly among Chinese-speaking children. The basis for this hypothesis is that Chinese derivational afxes are less productive than English derivational afxes.

Method Participants Participants were 282 American and 436 Taiwanese second, fourth, and sixth grade students. The Taiwanese students came from public schools in Taichung City and Dayuan, the American students from public schools in Danville, Mahomet, and Fisher, Illinois. In each grade, four classes were included; two classes were selected from a school or schools in which most of the students were from working-class families, while the other two were selected from another school in which most of the students were from middleclass families. Upon the request of one participating American sixth-grade teacher, two more of her classes were also included in the study.

408

YU-MIN KU AND RICHARD C. ANDERSON

Twenty-six American and twenty-four Taiwanese students were eliminated because they were absent during one experimental session. This left a total of 668 participants, 412 Taiwanese students (131 in second grade, 145 in fourth grade, and 136 in sixth grade) and 256 American students (65 in second grade, 62 in fourth grade, and 129 in sixth grade). Tasks Six tests were developed to assess different aspects of morphological awareness, vocabulary, and reading prociency. The tests were the Recognize Morphemes Test, the Discriminate Morphemes Test, the Judge Pseudowords Test, the Select Interpretations Test, the Select Vocabulary Test, and the Reading Comprehension Test. Recognize Morphemes Test. A morpheme recognition task was constructed to test childrens knowledge of the morphological relationships between pairs of words. Students saw 20 pairs of words followed by yes and no. For each pair, students were asked to judge whether the second word comes from the rst one, that is, whether the meanings of the two words are related. For example, English-speaking children were asked to indicate whether the word teacher comes from the word teach. Chinese-speaking children were asked /shu1jia4/ to indicate whether the meaning of the two-character word /shu1/ (bookshelf) is related to meaning of the rst component character (book). All the words were familiar to the children from oral language. To make the English and Chinese word pairs comparable, the complex word in each English word pair had no phonological or spelling change from the root word. Skilled readers of Chinese or English rated the words in the tests for morphological relatedness. These ratings were used to equate the items in the Chinese and English tests. Discriminate Morphemes Test. The morpheme discrimination test was designed to determine whether children understand that a word part may have different meanings in different complex words. The test lists 20 groups of words. Each group consists of three words that share a part. In two words, the common part has about the same meaning. The task for the children is to circle the odd word, the one in which the common part has a different meaning. For an English example, among the words classroom, bedroom, and mushroom, the meaning of room in the rst two words means a division of a building, with its own walls, oor and ceiling; but the room in mushroom means something else. For a Chinese example, among the words /shang1pin3/ (merchandise), /shang1dian4/ (shop; store), and /shang1liang2/ (to consult), the last one is odd because the meaning of the

DEVELOPMENT OF MORPHOLOGICAL AWARENESS

409

character is not business or merchant. Words used in the tests were familiar to students from oral language. Second-grade teachers rated whether their students were familiar with the words meanings. These ratings were used to equate the Chinese and English tests. Select Interpretations Test. The purpose of this task was to examine whether children could apply their knowledge of the morphology of compounds and derivatives to select proper interpretations of 16 low-frequency derived and compound words that contained high-frequency base words. The English words were chosen from Francis and Ku era (1982) and the Chinese words c from the Institute of Information Science Academia Sinica (1993). Only words that occurred fewer than 5 times per million were included in the tests. The task for the subjects was to select the proper interpretation of each word from among four choices. Here is an example in English: rebuild: (1) to build a house with bricks, (2) a man whose job is to build houses, : (3) a tall building, and (4) to build again. An example in Chinese is, , (2) , (3) , (1) . To select the meaning of the word rebuild correctly, and (4) English-speaking students have to understand the meaning of the prex re-. Depending solely on the knowledge of the base word build, without recognizing the meaning added by the prex re- would not be enough to select the correct interpretation from the alternatives. Similarly, in the Chinese example, (marksman), in order to select the proper interpretation, Chinesespeaking students have to know that is an agentive sufx and that it denotes a person who has certain skills or abilities. Judge Pseudowords Test. A checklist test was constructed to measure the childrens ability to apply word formation rules to novel vocabulary items. The items were intermixed with the general vocabulary items from the Select Vocabulary test described later. The task for the subjects was simply to check whether or not they knew the meanings of the items. The items were possible and impossible derivatives and compounds. Possible derivatives were formed observing distributional constraints (e.g., -ly attaches to adjectives but not to verbs) whereas impossible derivatives violated distributional constraints. Possible compounds conformed to word formation rules and all of them were semantically plausible. Impossible compounds were words that were semantically implausible. High-frequency stems were used to form pseudo-derivatives and pseudo-compounds, so that responses would depend on awareness of word formation rules rather than knowledge of the constituent stems.

410

YU-MIN KU AND RICHARD C. ANDERSON

The English test consisted of 20 possible pseudowords and 20 impossible pseudowords. There were two types of possible items. One type was a possible compound of two words, but not an existing word in the language, such as cowhouse; the other type was a possible derivative but not an existing one, such as heartful. The impossible items included ill-formed compounds, such as mansmall, and ill-formed derivatives, such as muchable. The Chinese test also contained 20 possible pseudowords and 20 impossible pseudowords. The possible items were of two types. One was a possible compound of two characters but not an existing word, such as ; the other was a possible derivative but not an existing one, such as . The impossible items included ill-formed compounds, such as , . The scoring formula for the Judge and ill-formed derivatives, such as Pseudowords Test was proportion yes for possible items minus proportion yes for impossible items, divided by one minus proportion yes for impossible items. Select Vocabulary Test. This was a wide range, general vocabulary test in the checklist format (Anderson & Freebody, 1983). The students simply indicated whether they knew the meanings of the items. As already stated, the items were intermixed with the items from the Judge Pseudowords Test. The English version of the Select Vocabulary Test contained 100 words chosen from a word frequency book (Francis & Ku era, 1982). Some were c very high-frequency words, which children as young as second graders should know, such as strong, simple, and student, and some were lowfrequency words, which may not be known by sixth graders, such as concede, fuzzy, and seam. The range of word frequency was from 5 to 202 per million. To control for guessing, the test included 20 pronounceable nonwords. These nonwords were constructed from pronounceable nonsense syllables, such as derg, and did not contain any identiable morphemes. The Chinese test consisted of 100 general words selected from a word frequency book (Institute of Information Science Academia Sinica, 1993). , , These words included some very high-frequency words, such as , and some low-frequency words, , , and . The and range of word frequency was from 3 to 200 per million. To control for guessing, 20 two-character nonwords, each containing one non-character, were constructed. A non-character was composed of two legal components but was not an existing combination in Chinese, such as . Scores were corrected for guessing using the high threshold formula proposed by Anderson and Freebody (1983). The formula for estimating the number of words that a child knows is the proportion of hits on words minus the proportion of false alarms on nonwords, divided by one minus the

DEVELOPMENT OF MORPHOLOGICAL AWARENESS

411

proportion of false alarms on nonwords. A hit means that a child checks a real word yes, whereas a false alarm means that a child checks a nonword yes. Reading Comprehension Test. The subjects reading prociency was measured with grade-appropriate reading comprehension tests. Different tests with similar formats and contents were administered to Chinese-speaking and English-speaking students. The English test was the reading comprehension subtest in the Metropolitan Achievement Tests, Seventh Edition. The Chinese test was compiled from several reading comprehension tests available in Taiwan; seven passages were selected for second graders and nine for fourth and sixth graders; children had to answer ve multiple-choice questions after reading each passage. The difculty of the passages and the questions that accompany the passages increased with grade level. A z score was computed for each student, standardized within grade level and language group. Procedure There were two separate test sessions on different days for all Chinesespeaking students and fourth-grade and sixth-grade English-speaking students. On the rst day, the Reading Comprehension Test was administered. The second day was for the Recognize Morphemes Test, Discriminate Morphemes Test, Select Interpretations Test, Judge Pseudowords Test, and the Select Vocabulary Test. For English-speaking second-grade classes, the four morphological tests were divided into two test sessions because class sessions were only 25 minutes in length. Each test was administered as a group test to the whole class. The classroom teachers read the directions and explained sample items. Because children in the second grade might not be able to read some of the words, teachers in the second-grade classes were asked to read each word aloud in the Recognize Morphemes, Discriminate Morphemes, and Select Interpretations tests. Fourth-grade students were encouraged to ask for help from their teachers if they had trouble reading any of the words.

Results The basic results of the study are shown in Table 2. The table contains the mean proportion correct following adjustment for guessing on the measures of morphological awareness and vocabulary knowledge for Chinese-speaking and English-speaking students in second, fourth, and sixth grades. Reading prociency is not shown in Table 2, because different measures were used in different grades and the comparison across grades would not be meaningful.

412

YU-MIN KU AND RICHARD C. ANDERSON

Table 2. Means and standard deviations for American and Taiwanese students by measure and grade level Measures Grade level Second Recognize Morphemes Discriminate Morphemes Select Interpretations Judge Pseudowords Select Vocabulary American Taiwanese American Taiwanese American Taiwanese American Taiwanese American Taiwanese 0.60 (0.25) 0.51 (0.26) 0.58 (0.25) 0.55 (0.25) 0.44 (0.21) 0.40 (0.26) 0.25 (0.27) 0.15 (0.26) 0.27 (0.27) 0.23 (0.40) Fourth 0.80 (0.24) 0.68 (0.26) 0.69 (0.21) 0.71 (0.20) 0.56 (0.21) 0.62 (0.24) 0.37 (0.26) 0.38 (0.37) 0.60 (0.21) 0.71 (0.21) Sixth 0.84 (0.19) 0.84 (0.20) 0.81 (0.12) 0.81 (0.12) 0.71 (0.19) 0.80 (0.18) 0.41 (0.22) 0.43 (0.29) 0.80 (0.14) 0.90 (0.12)

Note: Scores have been adjusted for guessing; standard deviations are in parentheses.

Table 3. Correlations between measures of morphological awareness and vocabulary knowledge. Measure 1. 2. 3. 4. 5. Recognize Morphemes Discriminate Morphemes Select Interpretations Judge Pseudowords Select Vocabulary 1 0.55 0.51 0.23 0.51 2 0.49 0.68 0.29 0.56 3 0.42 0.53 0.38 0.68 4 0.30 0.35 0.37 0.43 5 0.47 0.57 0.63 0.46

Note: Correlations for Chinese-speaking children appear below the diagonal; those for English-speaking children above the diagonal. All correlations are statistically signicant at the 0.01 level.

Analyses of variance veried that performance improved as a function of grade on every measure in both language groups (all Fs > 25, all Ps < 0.001). The correlations between the four tests of morphological awareness and vocabulary are given in Table 3. Correlations for Chinese-speaking children appear below the diagonal; those for English-speaking children above the diagonal. The table shows that the measures had moderate to high intercorrelations. The Discriminate Morphemes and Select Interpretations tests have higher correlations with vocabulary knowledge than do the Recognize Morphemes and Judge Pseudowords tests. The pattern of correlations for the two language groups is similar. The correlations are a little higher

DEVELOPMENT OF MORPHOLOGICAL AWARENESS

413

Table 4. Performance of American and Taiwanese students on compounds and derivatives by grade and measure Grade Measure Discriminate morphemes Select interpretations Compound Derivative American Second Fourth Sixth Compound Derivative Judge pseudowords Compound Derivative

0.59 (0.25) 0.56 (0.34) 0.69 (0.22) 0.69 (0.25) 0.83 (0.13) 0.78 (0.19)

0.42 (0.26) 0.48 (0.23) 0.34 (0.35) 0.15 (0.28) 0.52 (0.30) 0.60 (0.24) 0.44 (0.35) 0.30 (0.30) 0.68 (0.22) 0.65 (0.25) 0.43 (0.29) 0.39 (0.27)

Taiwanese Second 0.60 (0.25) 0.42 (0.35) Fourth 0.75 (0.22) 0.62 (0.26) Sixth 0.85 (0.13) 0.71 (0.20)

0.41 (0.27) 0.39 (0.34) 0.32 (0.36) 0.00 (0.37) 0.58 (0.50) 0.56 (0.39) 0.52 (0.50) 0.25 (0.34) 0.79 (0.19) 0.82 (0.25) 0.53 (0.35) 0.34 (0.29)

Note: Scores have been adjusted for guessing; standard deviations are in parentheses.

for the Chinese-speaking students, except in instances involving the Judge Pseudowords Test. Three of the four morphological measures included both derivatives and compounds, although it should be cautioned that the measures were not designed to equate derivatives and compounds on extraneous factors that could inuence performance. Of the 16 low-frequency words used in the Select Interpretation Test, 6 were derivatives and 10 were compounds. In the Discriminate Morphemes Test, 6 out of 20 items involved distinguishing derivatives from pseudo-derivatives (e.g., teacher, farmer, and shoulder). The Judge Pseudowords Test contained 20 derivatives and 20 compounds. Table 4 presents the means and standard deviations of students performance on derivatives and compounds in the three tests that contained both types of words. Analyses of variance with grade level as the between-subjects factor and word type (derivative vs. compound) as the within-subjects factor conrmed a signicant effect of grade on each measure in both language groups (all Ps < 0.01). With regard to type of complex word, the general trend was for Chinesespeaking children to score higher on compounds while English-speaking children scored higher on derivatives, but there were inconsistencies among the three tests. Chinese-speaking students had signicantly higher scores on compound words than on derived words from the Judge Pseudowords Test (F(1,409) = 170.90, P < 0.001) and the interaction between word type and grade was also signicant F(2,409) = 3.71, P < 0.05). From Table 4, it is

414

YU-MIN KU AND RICHARD C. ANDERSON

clear that difference between performance on compounds and derivatives was greater for younger students. On the Discriminate Morphemes Test, Chinesespeaking students in all three grades performed better on compounds than derivatives, F (1,409) = 131.25, P < 0.01, however, on the Select Interpretations Tests performance was similar on the two types of words. Like the Chinese-speaking children, the English-speaking children had higher scores on the compounds than the derivatives from the Judge Pseudowords Test (F(1,253) = 30.38, P < 0.001). The English-speaking students performed equally well on compounds and derivatives from the Discriminate Morphemes Test, whereas on the Select Interpretation Test, they had signicantly higher scores on derivatives than on compounds, F(1,253) = 19.35, P < 0.01). A principal component analysis of the four morphological measures was performed and, then, rst principal component scores were computed to serve as an overall indicator of students morphological awareness. The rst principal component is the best linear combination of the four specic measures, best in the sense that it accounts for the maximum amount of covariance among the measures. Using a principal component score as the overall indicator of morphological awareness minimizes the inuence of extraneous features of the individual measures, such as peculiar task demands, idiosyncrasies of particular items, and performance oors or ceilings. Thus, the rst principal component score is as pure an indicator of morphological awareness as the data afford. The total variance accounted for by the rst principal component was 59% and 56% for Chinese-speaking and English-speaking students, respectively. The Discriminate Morphemes and Select Interpretation Tests were the most inuential measures in the formation of the component scores for both language groups. The loadings for the two language groups are virtually identical; thus, it seems acceptable to combine the data of Chinese-speaking and English-speaking students to form a common principal component. The total variance accounted for by the rst principal component from the pooled data was 58%. Component scores based on the pooled data provide a common metric to compare the importance of morphological awareness in vocabulary knowledge and reading comprehension for the two language groups. An analysis of variance of pooled component scores indicated a signicant grade effect, F(2,662) = 161.72, P < 0.001, which conrms once again that older students are more aware of the underlying principles that govern the formation of morphologically-complex words. The overall difference between Chinese-speaking and English-speaking students was not signicant (F < 1, P = 0.33); however, there was a signicant interaction of grade and language, F(2,662) = 4.09, P < 0.05. Figure 1 depicts the development of morphological awareness across grade levels for Chinese-speaking

DEVELOPMENT OF MORPHOLOGICAL AWARENESS

415

Figure 1. Morphological awareness as a function of grade for American and Taiwanese students. Table 5. Correlations of morphological awareness with vocabulary and reading for Taiwanese and American students at three grade levels. Measure Grade level Second Taiwanese Vocabulary Reading Vocabulary Reading 0.51 0.73 0.52 0.67 Fourth 0.56 0.73 0.62 0.60 Sixth 0.60 0.63 0.54 0.63

American

Note: All correlations are statistically signicant at the 0.01 level.

and English-speaking students. As Figure 1 suggests, the interaction was due to the difference between the morphological awareness of Chinesespeaking and English-speaking second graders, F(1,194) = 5.02, P < 0.05. Chinese-speaking sixth graders were slightly but not signicantly ahead of English-speaking sixth graders. Table 5 shows the correlation of students morphological awareness (as represented by rst principal component scores) with vocabulary and reading. Among both Chinese-speaking and English-speaking students, morpholo-

416

YU-MIN KU AND RICHARD C. ANDERSON

Table 6. Regression analyses predicting reading prociency from morphological awareness after entering vocabulary knowledge American R2 change Second grade Step 1. Vocabulary Step 2. MA Fourth grade Step 1. Vocabulary Step 2. MA Sixth grade Step 1. Vocabulary Step 2. MA R2 change F Taiwanese R2 change R2 change F

0.393 0.559

0.393 0.166

40.82 23.39 51.32 6.25 44.84 40.19

0.230 0.551

0.230 0.321

38.60 91.39 64.45 82.46 139.74 20.51

0.461 0.513

0.461 0.052

0.311 0.564

0.311 0.253

0.261 0.440

0.261 0.179

0.510 0.576

0.510 0.065

MA morphological awareness. P < 0.05; P < 0.01.

gical awareness is highly related to vocabulary knowledge and reading prociency. The correlations between morphological awareness and vocabulary knowledge are similar for Chinese-speaking and English-speaking students, whereas the correlations between morphological awareness and reading prociency are slightly higher for Chinese-speaking students. Hierarchical regression analyses were performed to explore the extent to which morphological awareness makes a contribution to childrens reading development beyond sheer amount of vocabulary knowledge. Table 6 indicates the additional variance in reading prociency attributable to morphological awareness when it is entered after vocabulary knowledge. In each grade morphological awareness makes a signicant independent contribution to the reading prociency of both Chinese-speaking and English-speaking students. What is noteworthy is the substantial contribution of morphological awareness to the reading development of Chinese-speaking second and fourth graders. The additional variance accounted for was 32% and 25%.

Discussion The current study provides strong evidence that both Chinese-speaking and English-speaking students morphological awareness develops with increasing experience with language and that awareness of word structure is highly

DEVELOPMENT OF MORPHOLOGICAL AWARENESS

417

related to vocabulary and reading acquisition. The study shows that the development of morphological awareness among English- and Chinese-speaking children is similar, despite the large differences in their languages and writing systems, and the differences in culture and traditions of education in Taiwan and mid-America. The nding that childrens morphological awareness develops with grade is consistent with previous research (Carlisle, 1988; 1995; Tyler & Nagy, 1989; Leong, 1989; Singson et al., 2000; Wysocki & Jenkins, 1987), but most of this research investigated English-speaking childrens knowledge of English derivational morphology. The current study shows that Chinese-speaking childrens knowledge of Chinese morphology also develops with grade level. The current study goes beyond previous studies in either English or Chinese by exploring childrens acquisition of compounding rules as well as derivational morphology. The Taiwanese children probably would have outperformed the American children if the full range of compounds and derivatives in Chinese and English had been examined. As we have stressed, English-speaking children have trouble when the pronunciation or spelling of morphemes shifts from word to word (e.g. Carlisle, 1988), but such words were excluded in the present study because in Chinese there are seldom shifts in spoken or written forms of morphemes and we were attempting to equate the tasks in the two languages. Chinese-speaking children displayed somewhat more awareness of the morphology of compounds than English-speaking children, especially in the fourth and sixth grades. In contrast, Chinese-speaking children displayed less awareness of derivational morphology; this was especially true of second graders, marginally true of fourth graders, while the difference between Chinese-speaking and English-speaking children had largely disappeared among sixth graders. These patterns are readily understandable considering the structure of the two languages. Compounds are more common and probably more transparent in Chinese than in English. Derivational afxes in Chinese are less productive than derivational afxes in English. A contributing factor may be differences in instruction about word formation in Taiwan and America. Interviews with two American second-grade teachers in one participating elementary school revealed that second graders have been introduced to derivational principles governing high-frequency and productive afxes like un-, re-, -less, and -er. As a matter of fact, being able to use word analysis (analyzing complex words into root words, inections, derivational afxes) to identify words is part of the state second grade English-Language Arts Standards in Illinois. In contrast, Taiwanese teachers told us that Chinese instruction focuses on learning to pronounce and write each character correctly rather than on morphological analysis of words.

418

YU-MIN KU AND RICHARD C. ANDERSON

This study conrms strong relationships of morphological awareness to vocabulary knowledge and reading comprehension, which supports the notion that childrens insights into the internal structure of words play an important role in reading development. Similar ndings have been reported in other recent studies. For example, Mahony et al. (2000) found a correlation of 0.40 between English-speaking elementary school students vocabulary and morphological awareness, assessed with a test similar to the Recognize Morpheme Test. In the current study the correlation between these two measures was 0.47. Carlisle (2000) used two production tasks that required third and fth graders to produce correct derived forms and provide denitions for morphologically complex words. She found correlations of performance on these two tasks with vocabulary and reading comprehension were about 0.40 and 0.20 for third graders and 0.60 and 0.66 for fth graders. The correlations in the three grades examined in the present study ranged from 0.52 to 0.62 for vocabulary and from 0.60 to 0.67 for reading comprehension. The lower correlations Carlisle obtained for third graders may be attributable to the additional demands of production tasks she used, which may be especially difcult for children of this age. Two previous studies have examined the relation of Chinese childrens morphological awareness to reading achievement and vocabulary knowledge. Wang (1999) found correlations of 0.59 and 0.45 between morphological awareness and reading and vocabulary, respectively, in a sample of rst and second graders. The gures obtained by Li et al. (2001) were 0.26 and 0.59 for rst graders and 0.56 and 0.47 for fourth graders. The current study found higher correlations between morphological awareness and reading achievement (r = 0.630.73) and similar correlations between morphological awareness and vocabulary knowledge (r = 0.510.60). Correlations with reading achievement might have been higher in the present study because the tests used in this study were more extensive, with at least 7 passages and 35 questions, whereas the other two studies used sentence reading tests or short passage reading tests. Another possible explanation is that the measures in Wang (1999) and Li et al. (2001) were mostly designed to assess the childrens awareness at the level of single characters, namely, the insight that the radical in a semantic-phonetic compound character conveys a clue to its meaning, while the measures in this study focused on the childrens understanding of the structure of words composed of two or more characters. Theoretically, it would seem that morphological awareness is directly related to vocabulary acquisition but only indirectly related to reading prociency. Thus, we had expected higher correlations between morphological awareness and vocabulary knowledge than between morphological awareness and reading level. The data were not consistent with this expectation. Perhaps

DEVELOPMENT OF MORPHOLOGICAL AWARENESS

419

the reason is that the reading achievement measures we employed were more reliable than the measures of vocabulary knowledge. Chinese characters usually represent only one morpheme and there are far fewer phonological and orthographic shifts in the formation of complex words in Chinese than in English. At the same time, Chinese has many more homophones than English, so the Chinese child who is not making good use of morphological information will frequently be confused. Therefore, because of the characteristics of Chinese and English, we had expected morphological awareness to be more strongly related to reading level among Chinese-speaking children than among English-speaking children. The data were weakly consistent with this hypothesis. The robust correlations obtained in the present study between morphological awareness and reading level could not have been due to hidden variation in vocabulary knowledge. In the rst place, we constructed tests using only high-frequency roots and afxes. Every child should know these word parts or, at least, has had innumerable opportunities to learn them. In the second place, morphological awareness accounts for substantial amounts of variance in reading level after variance attributable to vocabulary is removed (see Table 6). Thus, the following conclusion is warranted: Children who are good readers for their age are aware of the information in word parts, able to decompose complex words into informative parts, and able to use information in word parts to estimate the meanings of unfamiliar complex words. That similar ndings were obtained with languages as different as Chinese and English supports the conjecture that morphological awareness is universally important in learning to read.

Acknowledgement The research reported in this paper was supported in part by a grant from the Spencer Foundation.

References
Anderson, R.C. & Freebody, P. (1983). Reading comprehension and the assessment and acquisition of word knowledge. In B. Hutton (Ed.), Advances in reading / language research, a research annual (pp. 231256). Greenwich, Connecticut: JAI Press. Anglin, J.M. (1993). Vocabulary development: A morphological analysis. Monographs of the Society for Research in Child Development, 58 (10. Serial No. 238). Chicago, Illinois: The University of Chicago Press. Berko, J. (1958). The childs learning of English morphology. Word, 14, 150177.

420

YU-MIN KU AND RICHARD C. ANDERSON

Blachman, B.A. (2000). Phonological awareness. In M.L. Kamil, P.B. Mosenthal, P.D. Pearson & R. Barr (Eds.), Handbook of reading research (Vol. III, pp. 483502). Mahwah, New Jersey: Erlbaum. Bradley, L. & Bryant, P.E. (1983). Categorising sounds and learning to read: A casual connection. Nature, 310, 419421. Carlisle, J.F. (1988). Knowledge of derivational morphology and spelling ability in fourth, sixth, and eighth graders. Applied Psycholinguistics, 9, 247266. Carlisle, J.F. (1995). Morphological awareness and early reading achievement. In L.B. Feldman (Ed.), Morphological aspects of language processing (pp. 189209). Hillsdale, New Jersey: Erlbaum. Carlisle, J.F. (2000). Awareness of the structure and meaning of morphologically complex words: Impact on reading. Reading and Writing: An Interdisciplinary Journal, 12, 169 190. Carlisle, J.F. & Nomanbhoy, D.M. (1993). Phonological and morphological awareness in rst graders. Applied Psycholinguistics, 14, 177195. Casalis, S. & Louis-Alexandre, M.-F. (2000). Morphological analysis, phonological analysis and learning to read French: a longitudinal study. Reading and Writing: An Interdisciplinary Journal, 12, 303335. Clark, E.V. (1981). Lexical innovations: How children learn to create new words. In W. Deutsch (Ed.), The childs construction of language (pp. 299328). London: Academic Press. Clark, E.V., Gelman, S.A. & Lane, N.M. (1985). Compound nouns and category structure in young children. Child Development, 56, 8494. Dale, E., ORourke, J. & Bamman, H.A. (1971). Techniques of teaching vocabulary. Palo Alto, California: Field Educational Publications. Derwing, B. & Baker, W. (1979). Recent research on the acquisition of English morphology. In P. Fletcher & M. Garman (Eds.), Language acquisition (pp. 209223). New York: Cambridge University Press. Dominiek, S. (1990). On the representation and processing of compound words: Automatic access to component morphemes does not occur. Quarterly Journal of Experimental Psychology, 42, 529567. Fowler, A.E. & Liberman, I.Y. (1995). The role of phonology and orthography in morphological awareness. In L.B. Feldman (Ed.), Morphological aspects of language processing (pp. 157188). Hillsdale, New Jersey: Erlbaum. Francis, W. & Ku era, H. (1982). Frequency analysis of English usage. Boston, Massachuc setts: Houghton Mifin. Freyd, P. & Baron, J. (1982). Individual differences in acquisition of derivational morphology. Journal of Verbal Learning and Verbal Behavior, 21, 282295. Goswami, U. (2000). Phonological and lexical processes. In M.L. Kamil, P.B. Mosenthal, P.D. Pearson & R. Barr (Eds.), Handbook of reading research (vol. III, pp. 252267). Mahwah, New Jersey: Erlbaum. Hatano, G., Kuhara, K. & Akiyama, M. (1981). Kanji help readers of Japanese infer the meaning of unfamiliar words. The Quarterly Newsletter of the Laboratory of Comparative Human Cognition, 3, 3033. Ho, C.S.-H. & Bryant, P. (1997). Learning to read Chinese beyond the logographic phase. Reading Research Quarterly, 32, 279289. Hoosain, R. (1991). Psycholinguistic implications for linguistic relativity: A case study of Chinese. Hillsdale, New Jersey: Erlbaum.

DEVELOPMENT OF MORPHOLOGICAL AWARENESS

421

Hoosain, R. (1992). Psychological reality of the word in Chinese. In H.C. Chen & O.J.L. Tzeng (Eds.), Language processing in Chinese (pp. 111131). Amsterdam: NorthHolland. Institute of Information Science Academia Sinica (1993). Corpus-based frequency count of words in journal Chinese: Corpus-based research series No. 2 (Technical Rep. No. 93-02). Taipei, Taiwan. Kuczaj, S.A. (1977). The acquisition of regular and irregular past tense forms. Journal of Verbal Learning and Verbal Behavior, 16, 589600. Leong, C.K. (1989). The effects of morphological structure on reading prociency-a developmental study. Reading and Writing: An interdisciplinary Journal, 1, 357379. Lewis, D.J. & Windsor, J. (1996). Childrens analysis of derivational sufx meanings. Journal of Speech and Hearing Research, 39, 209216. Li, W., Anderson, R.C., Nagy, W.E. & Zhang, H. (2001). Facets of metalinguistic awareness that contribute to Chinese literacy. In W. Li, J.S. Gaffney & J.L. Packard (Eds.), Chinese childrens reading acquisition: Theoretical and pedagogical issues (pp. 87106). Boston, Massachusetts: Kluwer Academic Publishers. Mahony, D. (1994). Using sensitivity to word structure to explain variance in high school and college level reading ability. Reading and Writing: An Interdisciplinary Journal, 6, 1944. Mahony, D., Singson, M. & Mann, V. (2000). Reading ability and sensitivity to morphological relations. Reading and Writing: An Interdisciplinary Journal, 12, 191218. Marcus, G.F., Pinker, S., Ullman, M., Hollander, M., Rosen, T.J. & Xu, F. (1992). Overregularization in language acquisition. Monographs of the Society for Research in Child Development, 57 (4, Serial No. 228). Chicago, Illinois: The University of Chicago Press. Nagy, W.E. & Anderson, R.C. (1984). How many words in printed school English? Reading Research Quarterly, 19, 304330. Nagy, W.E. & Anderson, R.C. (1998). Metalinguistic awareness and the acquisition of literacy in different languages. In D. Wagner, R. Venezky & B. Street (Eds.), Literacy: An international handbook (pp. 155160). Boulder, Colorado: Westview Press. Nagy, W.E., Anderson, R.C., Schommer, M., Scott, J.A. & Stallman, A.C. (1989). Morphological families in the internal lexicon. Reading Research Quarterly, 24, 262282. Nagy, W.E., Osborn, J., Winsor, P. & OFlahavan, J. (1994). Structural analysis: Some guidelines for instruction. In F. Lehr and J. Osborn (Eds.), Reading, language, and literacy: Instruction for the twenty-rst century (pp. 4558). Hillsdale, New Jersey: Erlbaum. Packard, J.L. (2000). The morphology of Chinese: A linguistic and cognitive approach. Cambridge, UK: Cambridge University Press. Sandra, D. (1994). The morphology of the mental lexicon: Internal word structure viewed from a psycholinguistic perspective. Language and Cognitive Processes, 9, 227269. Shu, H., Anderson, R.C. & Zhang, H. (1995). Incidental learning of word meanings while reading: A Chinese and American cross-culture study. Reading Research Quarterly, 30, 7695. Shu, H., Anderson, R.C. & Wu, N. (2000). Phonetic awareness: Knowledge of orthographyphonology relationship in the character acquisition of Chinese children. Journal of Educational Psychology, 92, 5662. Silvestri, S. & Silvestri, R. (1977). A developmental analysis of the acquisition of compound words. Language, Speech, and Hearing Services in the Schools, 8, 217221. Singson, M., Mahony, D. & Mann, V. (2000). The relation between reading ability and morphological skills: Evidence from derivational sufxes. Reading and Writing: An Interdisciplinary Journal, 12, 219252.

422

YU-MIN KU AND RICHARD C. ANDERSON

Sternberg, R.J. (1987). Most vocabulary is learned from context. In M.G. McKeown & M.E. Curtis (Eds.), The nature of vocabulary acquisition (pp. 89105). Hillsdale, New Jersey: Erlbaum. Taft, M. & Forster, K.I. (1976). Lexical storage and retrieval of polymorphemic and polysyllabic words. Journal of Verbal Learning and Verbal Behavior, 15, 607620. Taft, M. & Zhu, X. (1995). The representation of bound morphemes in the lexicon: A Chinese study. In L.B. Feldman (Ed.), Morphological aspects of language processing (pp. 293 315). Hillsdale, New Jersey: Erlbaum. Tang, T.-C. (1988). Studies on Chinese morphology and syntax. Taipei, Taiwan: Student Book Co. Tyler, A. & Nagy, W.E. (1989). The acquisition of English derivational morphology. Journal of Memory and Language, 28, 649667. Tyler, A. & Nagy, W.E. (1990). Use of derivational morphology during reading. Cognition, 36, 1734. Wang, C.-C. (1999). Learning to read Chinese: The role of phonological awareness and morphological awareness. Unpublished doctoral dissertation, University of Illinois at Urbana-Champaign, Illinois. Wang, C., Peng, D., Guan, L. & Kuang, P. (1999). The role of surface frequencies, cumulative morpheme frequencies, and semantic transparencies in the processing of compound words. Acta Psychologica Sinica, 31, 257265. White, T.G., Power, M.A. & White, S. (1989). Morphological analysis: Implications for teaching and understanding vocabulary growth. Reading Research Quarterly, 1989, 24, 283304. Wysocki, K. & Jenkins, J.R. (1987). Deriving word meanings through morphological generalization. Reading Research Quarterly, 22, 6681. Zhang, B. & Peng, D. (1992). Decomposed storage in the Chinese lexicon. In H.C. Chen & O.J.L. Tzeng (Eds.), Language processing in Chinese (pp. 131149). Amsterdam: NorthHolland. Address for correspondence: Richard C. Anderson, Center for the Study of Reading, 158 Childrens Research Center, 51 Gerty Drive, Champaign, IL 61820, USA Phone: +1-217-333-4437; E-mail: csrrca@uiuc.edu

You might also like