1 AN ERROR ANALYSIS OF CHINESE CHARACTERS WRITTEN BY BEGINNING LEARNERS OF CHINESE AS A FOREIGN LANGUAGE A RESEARCH PAPER SUBMITTED TO THE GRADUATE SCHOOL IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE MASTER OF ARTS BY SARAH KITTERMAN DR. MEGUMI HAMADA ADVISOR BALL STATE UNIVERSITY MUNCIE, IN JULY, 2012
2 INTRODUCTION Second language acquisition is the process of learning a language other than one s native, or first, language. This includes all that is involved in learning another language: reading, writing, listening, speaking, grammar, etc. Learning to read and write in a second language involves metalinguistic knowledge - the ability for an individual to parse a language into different categories, such as words, parts of speech (nouns, verbs, prepositions, etc.), all the way down to the awareness of phonemes (Koda, 2004) (the sounds that are used in a language to build words). To read and write in any language, the learner must be aware of how the language is parceled into smaller components that are represented by text (Koda, 2004). Mandarin Chinese (here on out referred to as Chinese ) is the language that has the most native speakers and total speakers in the world (Ethnologue, 2012). Enrollment in courses for Chinese as a foreign language has been growing at a rapid pace in the United States over the past few decades, growing 74% between 1980 and 1990 (Abbott & Wilcox, 2009; Arrow, 2004). More recently, it has experienced a 30% growth of enrollment between 2002 and 2006 (Ke & Li, 2011). With such a rapid growth in interest
3 2 and enrollment in Chinese as a second language, research dealing with how methods for learning and teaching Chinese to native speakers of English as well as quantitatively looking at areas of good and poor performance among the same group is surprisingly hard to come by. Students eager to enroll in these courses, however, may not realize the arduous learning process that lies before them, and their teachers may not be as prepared as they could be to teach effectively because of the limited, but growing, number of studies for English speakers of Chinese as a foreign language. Learning to read and write in Chinese as a second language requires different types of metalinguistic awareness than English (Koda, 2004). Chinese characters (the units which make up the writing system of Chinese) cannot be broken down past the syllable level, whereas the English writing system can be deconstructed to individual phonemes (Koda, 2004). Additionally, each Chinese character represents a syllable and a morpheme (the smallest meaningful unit of a word; for instance, cats is made of two morphemes cat and the suffix s that denotes plurality). Going from English, where each letter or group of letters represents an individual sound, to Chinese, where each symbol could represent an entire word, requires changing one s understanding of how spoken language can be represented in text. Unlike English, with its 24 letters that can be combined to form various sounds and words, Chinese has thousands of characters, each with their own unique meanings. This means that to even read a newspaper in Chinese, an individual would have to learn at least 3,000 characters (Arrow, 2004; Ho, Ng, & Ng, 2003; Su, 2010) and an even higher number of combinations of these characters.
4 3 Most studies regarding learning the Chinese writing system deal with reading and the various strategies students use to accomplish this intimidating task (Arrow, 2004; Everson, 1998; Grainger, 2005; Ke & Li, 2011; Kubota & Toyoda, 2001; Mori, 1998; Su, 2010). While much of what these studies have found can be applied to how students tackle the issue of learning to write Chinese characters, research dealing solely with writing characters is relatively scant (Guan, Liu, Chan, Ye, & Perfetti, 2011; Su, 2010). However, it should be kept in mind that the relationship between learning to read and learning to write is close. This can be seen from previous studies of ways that native speakers of languages with alphabets have gone about learning Chinese as well as native speakers of Chinese. For native speakers of Chinese, being able to write novel characters strengthens the ability of the learner to remember it and recognize it later when reading (Chan, Ho, Tsang, & Chung, 2006; Guan, et al., 2011; Tan, Spinks, Eden, Perfetti, & Siok, 2005). Brain scans of native speakers have shown that word recognition while reading also activates areas of the brain associated with writing (Guan, et al., 2011; Siok, Niu, Jin, Perfetti, & Tan, 2008). For non-native learners of Chinese, the combination of reading and writing new characters was more beneficial to remembering characters than simply reading alone (Guan, et al., 2011). The inclusion of pinyin (the standardized method to write Chinese in the Latin alphabet) with the characters helped such learners even more (Guan, et al., 2011). Including stroke-by-stroke instruction of how to write Chinese characters was found to be even more beneficial (Guan, et al., 2011). Other studies have shown that the
5 4 more often a character occurs, the more likely learners will be able to write it correctly (Ke, 1996; McEwen, 2006). Multiple studies have been done on only reading and recognizing Chinese characters; few have been on learning to write Chinese characters. Those that have been done have shown a clear relationship between learning to read and write Chinese characters, but the emphasis of research is still on reading recognition rather than character writing. This study aims to add more to this little-researched aspect of learning Chinese as a foreign language through an error analysis of characters written by first semester college students of Chinese. The significance of this study for the field of second language acquisition is that it looks at something other than English as a second language (which is very well researched) and gives insight into how learners adjust when moving from an alphabet to Chinese characters. The significance to pedagogy is that this study, and later studies, will help those teaching Chinese as a second language teach characters more effectively, further helping students produce and read written Chinese.
6 LITERATURE REVIEW Background Research on Second Language Reading and Writing With the growing interest among the student populace in learning Chinese as a foreign language, researchers have been turning their attention to finding how native and non-native speakers learn Chinese. More studies are being done, more textbooks are being published for Chinese foreign language instruction, and more societies for people teaching or learning Chinese are being established across the United States, much of these being funded by the governments of the United States, China, and Taiwan (Ke & Li, 2011). For this study, research done with reading (especially word recognition) and writing (of characters, not composition) will be considered. The two fields of research are related in that they use the same linguistic information and much of the same cognitive functions. The linguistic information involved is a spoken language and its assigned writing system. The shared cognitive processing is in the transfer of graphic information to the spoken language for reading and vice versa for writing. Studies have also shown that writing and reading are very intertwined in Chinese among native and
7 6 non-native speakers, as discussed above (Chan, et al., 2006; Guan, et al., 2011; Ke, 1996; McEwen, 2006; Siok, et al., 2008; Su, 2010; Tan, et al., 2005). Learning to read and write a language is not like learning to listen or speak a language; the latter is a natural process that is innate to the human species, but the former requires years of dissection and instruction to master (Schau, 2000). There are believed to be two levels of components to reading higher level processing and lower level processing (Schau, 2000). Higher level processing entails content knowledge (where the reader already knows about the topic he or she is reading) or contextual assistance (where the reader uses the information from the text itself to gain more information about specific words or phrases) (Schau, 2000). Lower level processing entails gaining meaning from the bits and pieces that make up words and phrases; that is, letters and words are the primary source of getting meaningful information from a text (Schau, 2000). For example, when reading a children s book, most adults would employ higher level processing as there would be no new words and the stories are probably already familiar to them. Reading something more difficult, like the inner workings of a potential quantum computer, would be exceedingly difficult for the layperson who would need to rely on having to sound out individual words that were novel; that is, lower level processing would be more important here. Having these two levels work together involves using higher level processing for most of what is being read, but using lower level processing to understand new or unexpected material in the text (Schau, 2000).
8 7 Learning to read often works from lower processing skills up (Schau, 2000); that is, learning what the smallest meaningful written symbol that is relevant to spoken language is and building larger pieces of the language from there. This lower level processing is key to being a good reader; being able to automatically process this visual information leaves room for cognitive processing of the larger picture (Schau, 2000). The less work an individual needs to do with these lower level processes, the easier it is to read a given text and think about the big picture of what the text describes. Writing works in the same direction. People learn their native language s writing system from the smallest meaningful written symbols (i.e. letters in an alphabet), work their way up to small words, then larger words, then sentences, etc. While it may not be as dramatic as with reading, high level processing skills can take over writing after the lower level processes have become automated; this can be seen in students writing class notes while looking at the professor and not their notebook or being able to spell most words without thinking about it and instead focusing on what words or sentences will be written next. While learning to speak and listen in a second language may be more difficult than learning to speak or listen in one s native language, it is believed that learning to read and write in a second language is even more arduous (Schau, 2000). When learning to read and write in one s native language, there is already a phonological foundation on which to build the written language; this foundation rarely exists for individuals learning to read and write in a second language (Guan, et al., 2011). Additionally, although learning to read and write in a second language is difficult for second language learners who are skilled readers in their first language, it is even more difficult for individuals
9 8 who are poor readers and writers in their native language (Arrow, 2004). This is because learners of a second language will use their prior knowledge of how sounds and orthography (the writing system for a language) are connected (Guan, et al., 2011; Su, 2010). If this prior knowledge is lacking in any way, it will be detrimental to their understanding of the connection in a second language. While higher level processes are at work regardless of language (that is, the reader will attempt to use previous knowledge and context to help tackle a difficult text whether it is in their native language or in a second language), lower level processes impede successful reading if the orthography of a second language is different from the first language (Schau, 2000; Su, 2010). If the orthography is different between the two languages, the learner would have to find new ways to process the second orthography (Chung, 2008; Kubota, 2005; Schau, 2000; Su, 2010). For example, native speakers of languages with an alphabetic writing system process morphosyllabic or logographic (a writing system that is only meaning-based and has no relation to sound) symbols differently than native speakers with a morphosyllabic or logographic writing system (Kubota, 2005). The above information pertains to writing Chinese characters, and not just reading them, because both levels of processing reading (lower and higher) and the relation between the two deal with writing as well. For example, for the higher level processing, at times, collocations may help individuals remember characters he or she may not have remembered if the characters were alone. For the lower level processing, simply
10 9 remembering one radical (a piece that makes up a character) within a character may be enough to recall the rest of the character. This is because visual processing of a written language depends upon the type of orthography that is used (Schau, 2000). The Orthographic Depth Hypothesis (Katz & Frost, 1992; Su, 2010) makes this clearer. There are different levels of depth into which an orthographic system can be categorized. The closer that each written symbol is to the phoneme it represents, and the more transparent this relationship is, the shallower it is said to be (Katz & Frost, 1992; Schau, 2000; Su, 2010). Thus, languages like Spanish or Italian are considered to be shallow orthographies. English is considered a deep orthography because, while its alphabet sometimes follows the phonology (the mental representation of sounds) of the language, it often diverges (take the pronunciation of the vowel /a/ in the words have and save, for example). Different orthographic depths require different reading and writing strategies (Su, 2010); for instance, deeper languages use more visual strategies while shallower languages use more phonological strategies to remember how words are written (Su, 2010). Chinese has an even deeper orthography than English (Guan, et al., 2011). Due to the nature of Chinese characters (which will be discussed below), there are minimal clues to how a given character sounds and each character represents a morpheme and possibly an entire word. It can be very difficult to discern how an unknown character sounds or what it may mean (Cheung, Chan, & Chong, 2007; Schau, 2000; Shu & Anderson, 1997; Su, 2010).
11 10 Another factor at play in second language acquisition is that of first language transfer. Language transfer is where an individual learning another language takes his or her knowledge of his or her previously learned language(s) and applies it to the second language (Winford, 2003). This process can be conscious or subconscious. The positive result of language transfer allows the learner to more easily acquire rules and words of the second language, most often when the two languages have commonalities. The negative result of language transfer, or interference, is when the learner s acquisition of a second language is impeded due to the second language being different from the first (Brown, 2007). It is often expected that these transfers, whether negative or positive, are the result of the learner using his or her knowledge of the native language to face challenges in acquiring the second language (Jarvis & Odlin, 2000; Winford, 2003). While English is not considered a shallow language according to the orthographic depth hypothesis (Katz & Frost, 1992), it does conform to some phonological rules. With this in mind, individuals learning Chinese may expect the same type of relationship between the spoken and written language that English has, which could end in a misunderstanding of how writing (and reading) Chinese is supposed to work. Even if the learner is able to grasp that each character represents a syllable, the knowledge that every syllable in Chinese has its own morphemic meaning may take a considerable amount of time to comprehend due to negative transfer from English, the morphemes of which regularly consist of two or more syllables.
12 11 The Chinese Writing System Chinese is the oldest continually written orthography in existence, spanning a history of over 4,000 years (Arrow, 2004). Beginning as pictographs, or pictures, of what they were representing, Chinese characters evolved over time to what we have today (Arrow, 2004). While it is easy to discern the pictographic roots of some characters (i.e. 女 (nǚ) for woman ), most characters have become quite abstract and the reader must rely more on pieces that make up a character, called radicals, to discern their meanings. While Chinese is often described as logographic, this term is incorrectly applied (Su, 2010). Only around 10% of characters are strictly semantic (meaning-based) (Ho, Ng, et al., 2003), which is why the term morphosyllabic, meaning the orthography represents morphemes and syllables, is a term which is growing in popularity to describe the Chinese writing system (Guan, et al., 2011; Su, 2010). Unlike English, Chinese is a tonal language, which means that the tone in which a word is said can change the meaning of the word. For instance, 妈, pronounced mā, has a high steady tone and means mother. 麻, pronounced má, has a rising tone and is a generic term for hemp or flax. 马, pronounced mǎ, has a falling and then rising tone and means horse. 骂, pronounced mà, has a falling tone and means to scold or to abuse. 吗, pronounced ma, has no tone and is used at the end of a question as a kind of verbal question mark. Another difference from English is the number of allowed syllable structures. English has a very complicated syllable structure system while Chinese has a very simple
13 12 one. The only allowed syllable structures in Chinese are V (vowel), VV (diphthong), CV (consonant-vowel), CVV (consonant-diphthong), VC (vowel-consonant), and CVC (consonant-vowel-consonant) (Su, 2010). The last of these structures can only be ended by the consonants n, ŋ, and ɻ. Furthermore, each syllable is a morpheme or a word. This leads to numerous homophones present in Chinese, even with the various tones (there are only around 1,200 unique syllables in Chinese with tones incorporated) (Ho, 1989; Koda, 2004). Such homophones can be confusing to native speakers and learners of Chinese, both in listening and writing. Each character in Chinese represents a syllable as well as a morpheme (Ho, Ng, et al., 2003; Koda, 2004; P. D. Liu, Chung, McBride-Chang, & Tong, 2010; Shu & Anderson, 1997; Su, 2010). Because of this, the ability for an individual to be fluent in reading and writing Chinese requires that he or she knows over 3,000 characters (Arrow, 2004; Ho, Ng, et al., 2003; Koda, 2004; Su, 2010). As with any writing system, this requires the individual to know the pronunciation, the shape (what it looks like or how it is written), and the meaning (Allen, 1992; Arrow, 2004; Chung, 2008) of each character. Each Chinese character is composed of one or more radicals (Allen, 1992; Koda, 2004), which are the smallest meaningful unit in a Chinese character (Ho, Ng, et al., 2003; Shen & Ke, 2007; Shu & Anderson, 1997; Su, 2010). Characters that are only made of one radical, or simple characters (Koda, 2004; Su, 2010), make up around 10%-20% of all Chinese characters (Everson, 1998; Ho, Ng, et al., 2003; P. D. Liu, et al., 2010). Characters that have two or more characters are called compound characters; around
14 13 80%-90% of Chinese characters fall into this category (Everson, 1998; Ho, 1989; Ho, Ng, et al., 2003; Koda, 2004; Su, 2010). These radicals are classified into two categories: semantic and phonetic (Allen, 1992; Ho, 1989; Koda, 2004; Shu & Anderson, 1997; Su, 2010). Semantic radicals generally give the meaning or semantic category of a character and are most often found on the left or top of a character (Allen, 1992; Ho, Ng, et al., 2003; Koda, 2004; P. D. Liu, et al., 2010; Shu & Anderson, 1997; Su, 2010). There are around 200 semantic radicals in Chinese (Ho, Ng, et al., 2003; Su, 2010). Phonetic radicals give phonological clues to how a character is pronounced and are often found on the right or bottom of characters (Allen, 1992; Ho, Ng, et al., 2003; Su, 2010). There are 800 to 1,200 phonetic radicals in Chinese (Ho, Ng, et al., 2003; Koda, 2004; Su, 2010). Semantic radicals tend to be more transparent in their meanings than phonetic radicals are in their phonological clues (Ho, Ng, et al., 2003; Su, 2010). A reader may know either kind of radical directly or through knowing their functions from other characters (Ho, Ng, et al., 2003; Ho, Wong, & Chan, 1999). Even with such inference, the accuracy of a prediction of a character s sound based on its phonetic radical is only around 40% (Ho, 1989; Ho, Ng, et al., 2003; Koda, 2004), which drops even lower, to 24%, if tone is taken into consideration (Everson, 1998; Ho, Ng, et al., 2003). Moreover, the more frequently a character occurs, the less likely it is that the phonetic radical will correspond with the sound of the character or that the semantic radical will be related to the character s semantic category (Everson, 1998). Some radicals take multiple shapes that are only found in certain parts of a character (for example, 心 (xīn), the heart radical,
15 14 which is semantic, is only found on the bottom of a character, while its other form 忄 is only found on the left side of a character) (Allen, 1992). Additionally, many phonetic radicals can stand alone as a complete character, but semantic radicals usually cannot (Su, 2010). These radicals are further divided into strokes, which is a line, curve, or angle that is written without lifting the pen or brush from the paper (Ho, 1989). There are eight or 24 basic strokes that make up radicals, depending on how they are being counted (on the upper end of the number of strokes, some of these can be further divided into smaller stroke combinations, leading to the smaller count) (Guan, et al., 2011; Su, 2010). Both strokes and radicals have very strict rules regarding the direction and order in which they are written (Allen, 1992; Ho, Ng, et al., 2003; P. D. Liu, et al., 2010). The order for radicals is usually left to right and top to bottom (Allen, 1992). Levels of Orthographic Awareness for Those Learning the Chinese Writing System As learners go through their courses or levels of self-teaching, they go through three levels of orthographic awareness (Ke & Li, 2011; Su, 2010). The first level is precomponent processing. During this stage, entire characters are added to the learner s lexicon without a lot of processing (Allen, 1992); the learner will likely be able to notice that there are radicals, but with such a small vocabulary and a large number of radicals, they do not have the means to notice patterns of what these radicals may mean, sound
16 15 like, or certain positional constraints, all of which are very important for remembering characters (Ho, Ng, et al., 2003; Ke & Li, 2011; Shen & Ke, 2007; Su, 2010). Thus, written errors found early in the first level or orthographic awareness involve writing general shapes that look similar to the target character. Those more advanced in this level may be able to start discerning what semantic category a new character belongs to based on its semantic radical, but would not be able to guess a new character s pronunciation (Allen, 1992; Ke & Li, 2011; Shu & Anderson, 1997; Su, 2010). They also start to be able to recognize illegal radicals and legal radicals in illegal positions when more advanced in the first level of orthographic awareness (Su, 2010; Wang, Liu, & Perfetti, 2004; Wang, Perfetti, & Liu, 2003). At this point, written errors would be expected to be with only one radical within a compound character. During this level, characters learned near the beginning and end of a course will more likely be remembered than those taught in the middle of the course (Ke & Li, 2011). The fewer strokes a character has, the more likely it will be written correctly, and the errors that learners in this level make will not be the same types of errors that native speakers of Chinese make (Ke & Li, 2011). While many learners rapidly learn Chinese in their first year or two of study, it still takes most learners knowledge of over 3,000 characters (about three years of classroom study) to continue to the next stage according to some research (Ke & Li, 2011; Shen & Ke, 2007). The second stage of orthographic awareness is the component processing stage (Ke & Li, 2011; Su, 2010). At this point, the learner is able to apply their understanding of how Chinese characters are constructed to identify and name known characters and
17 16 guess the meanings of and sound of simple, transparent novel characters (Ke & Li, 2011; Shen & Ke, 2007; Su, 2010). Many characters have found their ways into the learner s long-term memory, especially characters and radicals that are transparent in meaning or sound and which frequently occur (Ke & Li, 2011; Shu & Anderson, 1997; Su, 2010). At this stage, learners are able to notice and correct errors in others writings and his or her own (Kubota, 2005). Reaching the second stage may take anywhere from six months to two years according to some research (Su, 2010). Written errors at this stage would likely include new, complex, and/or rare characters missing strokes. The third stage of orthographic awareness is the automatic component processing stage (Ke & Li, 2011; Su, 2010). At this point in their learning, students have a nativelike awareness of the Chinese writing system and are able to recognize and produce characters based on frequent components, as well as radicals that do not strictly belong to phonetic or semantic radicals (Ke & Li, 2011). They are able to make good guesses on the meanings and pronunciations of unfamiliar characters, have good judgment on the legality of pseudocharacters (characters that have follow legal placement of radicals but are not real characters), and most errors made in writing characters are phonological, which is also the most common type of error made by native speakers (Ke & Li, 2011).. Learning to Read and Write in Chinese As stated before, reading and writing are related due to their shared characteristics. Firstly, they are unnatural; that is, speaking and listening to a language is innate to human
18 17 beings, but reading and writing must be consciously learned. Secondly, they share the same cognitive processes, as they both deal with the transfer of information between the spoken language and its writing system. Reading involves this information goes from the page to the spoken language, while writing goes from the spoken language to the page. They both involve the abstract process of transforming aural information into a physical representation that others can understand. Due to their related nature and scarce research into learning to write Chinese as a second language, studies involving both learning to read and write in Chinese will be considered. These studies deal with learners of Chinese as a second language unless otherwise stated. Chinese as a foreign language instructors are taking various studies on how learners acquire Chinese into consideration. For instance, by using knowledge gained from studies on how people learn a second language, more foreign language classrooms are shifting from being teacher-centered to learner-centered (Arrow, 2004). A common way of teaching Chinese characters is to divide new characters into their radical components and showing how they fit together (Kubota, 2005). Indeed, teaching radicals to students has been found to be beneficial, as it is important for word and character recognition (Ho, Ng, et al., 2003; Shu & Anderson, 1997). While semantic processing is the dominant process in reading Chinese characters (Schau, 2000; Su, 2010), phonological processing (connecting the written symbols on a page to the phonology, or sounds, of a language) still plays an important role in recognizing and understanding characters (Schau, 2000). Phonological processing is key because it accesses short-term memory, which is necessary for remembering what was
19 18 read earlier in a sentence to be able to piece the whole sentence together or to help remember a new character (Schau, 2000). Studies with English learners of Japanese as a second language regarding Japanese Kanji (which are or are based on Chinese characters) have shown that participants remembered pseudocharacters better if there was a phonetic radical (Mori, 1998; Schau, 2000). However, most of the focus from teachers is on the semantic radical, leaving learners to discover the sounds of phonetic radicals on their own (Everson, 1998; Su, 2010). This is exceedingly difficult for students not only because phonetic radicals are less transparent than semantic radicals, but also because many characters taught early on in a Chinese language classroom have irregular phonetic radicals (Su, 2010), making it even more challenging to find any kind of phonetic patterns. While teachers are becoming more aware of how people learn a second language successfully, learners are still mostly unaware of what may be more beneficial to them (Arrow, 2004). This often leads to students becoming frustrated with the language and dropping out of courses (Grainger, 2005; Su, 2010) and quickly forgetting what they had learned (Grainger, 2005). There are many things that students should be aware of to help them in their journey to fluency in a second language. For instance, it has been found that good language learners use more learning strategies than poor language learners (Arrow, 2004; Butler, 2011). Learning strategies that learners use to make predictions about the sounds and meanings of unknown words are through intralingual inference (using knowledge from the second language), interlingual inference (applying knowledge from one
20 19 language to another), and through extralingual inference (applying knowledge from the outside world) (Arrow, 2004). Ways that Chinese as a second language learners use to remember Chinese characters can be quite varied, and many will develop or use ways they find most useful to them (Su, 2010). This is certainly tied to how each student approaches Chinese and what they find the most difficult about characters. Many students use flashcards with the character, meaning, and pinyin (the way Chinese is transcribed in the Roman alphabet) or among other methods of memorization (Arrow, 2004; Everson, 1998; Grainger, 2005), most likely due to students finding the shape of a character the most difficult aspect (Arrow, 2004). Other methods for remembering how to write a character were repetition of writing a given character multiple times (Arrow, 2004; Everson, 1998; Grainger, 2005; Ke & Li, 2011; Kubota & Toyoda, 2001), creating mnemonic devices about how the character looked (usually this is done by using the radicals in a character; for example, to remember the character 好 (hǎo) good, one can divide the radicals and say that a woman ( 女 ) with her child ( 子 ) is a good thing) (Everson, 1998; Ke & Li, 2011; Kubota & Toyoda, 2001), or by associating a new character with characters already known (Grainger, 2005). Another method learners use builds on what many classrooms do, which is dividing the character into its radicals and remembering them and how they fit together (Grainger, 2005; Ke & Li, 2011; Kubota & Toyoda, 2001; Su, 2010), although students tend to prefer rote memorization of characters over this method after a year of study (Ke & Li, 2011). In a study by Kubota and Toyoda (2001), it was found that students who
21 20 divided characters into their radical components did a better job on a short-term memory writing task than those who used repetitious writing of a character. For those who found remembering the sound of a character to be difficult, reading texts aloud, using sound correspondences with their native language (i.e. one participant remembered 东 (dōng), east, by the sun rising in the east and an alarm clock goes dong around sunrise) (Everson, 1998; Ke & Li, 2011; Kubota & Toyoda, 2001), and color coding tones was thought to be beneficial to them (Arrow, 2004). Focusing on remembering phonetic radicals may be beneficial to those who find remembering the sounds of characters difficult, as it has been found that those with an alphabetic native language already, and often involuntarily, search for phonetic clues in Chinese characters (Kubota & Toyoda, 2001; Mori, 1998). As for remembering the meanings of characters, students found self-made vocabulary dictionaries and saying their native language equivalent while looking at the character to be helpful (Arrow, 2004). While some participants of Arrow s (2004) study claimed that reading authentic materials was helpful, other papers (Grainger, 2005) have shown that using authentic materials for learning Chinese is exceedingly difficult. The reason for this is that Chinese is classified as a category four language, which means that it takes three times as long to achieve second language fluency than learning a category one language like Spanish or French (Grainger, 2005; Su, 2010).
22 21 Character Errors in Learning Chinese Errors are very common in learning any second language, but errors by those learning to write Chinese characters are different than those by learners of syllabic or alphabetic languages. According to Kubota (2005), common errors in writing include: confusion with characters that are morphologically similar to other characters, misshapen radicals (Ishida, 2000), strokes that are written in an incorrect direction, strokes that are too long or too short, confusion with homonymous characters (Ishida, 2000; Su, 2010), and missing one or more characters in a multi-character word. Among language learners with an alphabetic or syllabic native language orthography, the most common errors are writing strokes in the wrong direction and with stroke length (Kubota, 2005). Other studies have found that using incorrect radicals or omitting strokes was another common error among learners of Chinese as a foreign language (Hatta, Kawakami, & Tamaoka, 1998; Su, 2010). Such errors come about because learners of Chinese as a second language have a higher cognitive load than learners of, for example, an alphabetic language. All aspects of a character - sound, meaning, and shape - must be learned immediately and remembered for the long term (Chung, 2008; Guan, et al., 2011; Perfetti, Liu, & Tan, 2005; Taft, Zhu, & Peng, 1999; Wang, et al., 2003). This results from the typical way that characters are presented in classrooms and textbooks. Multiple characters are presented at once, each with what they look like, what they mean, and how they sound (Chung, 2008; Everson, 1998). With all of these presented at the same time, learners
23 22 often involuntarily sacrifice two of the attributes of a character to only remember one due to such a high cognitive load (Chung, 2008). Ways to decrease some of this cognitive load have been found by some students learning Chinese. For example, students who practice characters in the context of other characters, whether it be to make words or within a sentence, were shown to be more effective at writing characters correctly than practicing characters individually (Ke & Li, 2011). Also, the more often a character occurs in Chinese, the less likely errors will be made in writing (Ke, 1996; Ke & Li, 2011). It should be kept in mind, however, that one s fluency in speaking or listening in Chinese will not help him or her significantly in writing or reading Chinese due to the nature of the characters. Part of the reason for this is that phonetic radicals are more numerous than semantic radicals and are generally less transparent than semantic radicals (Ke & Li, 2011). Another important factor is that learning Chinese characters can be so arduous for native speakers of English because of the orthographic differences between the two languages. Studies have shown that learners of a second language with a morphosyllabic or logographic writing system have a harder time learning the script if they come from a native language with an alphabetic orthography rather than a native language of a logographic or morphosyllabic orthography (Chikamatsu, 1996; Grainger, 2005; Hatta, et al., 1998; Hatta, Kawakami, & Tamaoka, 2002; Schau, 2000). In other words, if someone whose native language is English is learning Chinese, it would be more difficult and take more time to learn the script than for a native speaker of Japanese.
24 23 At all levels of fluency, it appears that there are several important factors to keep in mind for students who are learning Chinese as a second language. First is that successful learners use several learning strategies (Arrow, 2004; Butler, 2011). Second is that radical awareness leads to better reading and better written characters than remembering characters as a whole (Ho, Ng, et al., 2003; Shu & Anderson, 1997; Su, 2010). Third, it should be remembered that Chinese takes a very long time and a lot of practice to become fluent, especially compared to other languages (Chung, 2008; Grainger, 2005; Ke, Wen, & Kotenbeutel, 2001; Schau, 2000; Su, 2010), so if the Chinese class seems to be moving more slowly than a Spanish class, there is no need for concern or to drop the course out of frustration. In summary, it is expected that beginning learners of Chinese as a foreign language will make several character errors because of their low level of orthographic awareness (Allen, 1992; Su, 2010) and due to a high cognitive load given to them based on the presentation of novel characters from their textbooks (Chung, 2008; Y. Liu, Yao, et al., 2009a). Learners errors are predominately found within a character rather than through the confusion with other characters, as they do not have a large enough lexicon to confuse characters based on meaning or homophones (Allen, 1992; Ho, Ng, et al., 2003; Ke & Li, 2011; Shen & Ke, 2007; Su, 2010). Learners are also expected to perform poorly in stroke order because, at this level in their orthographic awareness, characters are remembered as a whole or as their radicals, but not through the smallest constituents (Allen, 1992; Ho, Ng, et al., 2003; Ke & Li, 2011; Shen & Ke, 2007; Su, 2010).
25 RESEARCH QUESTIONS Most studies done with Chinese as a foreign language learners has been done on character recognition in reading; little to no research has been done on how students learn and how students produce Chinese characters. For my study, I have the following questions about first semester students of Chinese as a foreign language, the answers to which I hope will shine more light on this inadequately researched area of second language acquisition: 1) Will learners at such an early stage of orthographic awareness make errors predominately with semantic radicals, phonetic radicals, or will there be no difference? It could be assumed that semantic radicals would have fewer errors because they are more transparent and fewer in number than phonetic radicals (Everson, 1998; Ho, Ng, et al., 2003; Ke & Li, 2011; Su, 2010). This question, however, has not been answered for Chinese as a foreign language students in any studies that I am aware of. The answer to this question would
26 25 not only provide insight into this scantly researched area, but could potentially improve teaching methods for teachers of Chinese as a foreign language. 2) Does incorrect and/or varied stroke order have any impact on producing semantic radicals and phonetic radicals? Again, this does not seem to be well researched for Chinese as a foreign language learners and the answer could have vast implications for teaching if incorrect and/or varied stroke order leads to poorly formed or incorrect radicals of either category. 3) Does incorrect and/or varied stroke order have any impact on producing characters overall? Once more, there seems to be little to no research regarding this for Chinese as a foreign language learners. The implications for teaching are even more important here; if the reason that characters are written incorrectly is to the students inability to use correct stroke order, more class time would surely need to be dedicated to teaching correct stroke order.
27 METHOD Participants An was sent to the professor of CH 101 at Ball State University at the end of the Fall 2011 semester requesting she ask for participants who were native speakers of English for my study. Five participants took part in this study. All five are Ball State University students who were at the end of their first semester of Mandarin (CH 101) and were taught in the same class by the same teacher. This level was selected not only to elicit more errors (as beginning students of any foreign language will make more errors than advanced students), but also to see how beginning learners of Chinese go about writing a completely foreign orthography after only four months of study. All participants are native speakers of English. Native speakers of English were wanted for participants because they would have no native speaker intuitions about the nature of Chinese characters. Four participants were female and one was male. Random numbers were assigned to each participant. Participants 34, 57, 82, and 20 had been studying Mandarin
28 27 (writing, reading, speaking, and listening) for four months, although participant 20 did have some experience with Mandarin at seven years old, but she said it was very minimal and she did not remember any of it prior to restarting learning it four months prior to this study. Participant 82 considered Spanish to be a second language and Participant 20 considered herself to be fluent in French and at an intermediate level in German. Participant 5 was excluded from this study due to being both left-handed and dyslexic both would very interesting topics to be looked into for future research. Materials Materials used in this study were the courses three books: the textbook, workbook, and character workbook from the third edition of the series Integrated Chinese 中文听说读写 : Level 1 Part 1 (Y. Liu, Yao, et al., 2009a, 2009b). Specifically, Lesson 5 reading comprehension exercise B found on page 85 of the workbook was used for one of the writing tasks. A digital camcorder was used as well as a tripod. SPSS 19 was used for correlation analysis. Analysis of stroke order was done with mdbg.net, chineseetymology.org, and the character workbook from the course (Y. Liu, Yao, Bi, Ge, et al., 2009)
29 28 Tasks Two writing tasks were presented to the participants to complete. The first task was a free writing exercise requesting information about their best friend to be written in Chinese with additional questions to help participants think about what details they could write about. The free writing task was given in order to allow participants the opportunity to write characters in which they were confident and to possibly avoid those in which they were not confident. The second exercise was a translation exercise that was from the last chapter completed by the class in their class s exercise workbook (Y. Liu, Yao, et al., 2009b). The passage, Lesson 5 reading comprehension exercise B found on page 85, was in Chinese, which I translated into English for the participants to translate back into Chinese. This translation task was given so as to persuade participants to attempt to write characters they may not have felt comfortable in writing, thus it would be more likely to elicit errors. Procedures Seven signatures were acquired with five people coming to participate by the end of the semester. Participants met with me individually in a quiet room on Ball State University s campus with few distractions. Participants were given consent forms explaining the process of data collection when they came to participate. Participants then read through and signed a consent form to the study. Then, questions were asked about
30 29 their native languages and how long they had been studying Chinese. I then gave a short oral explanation of what they were to do for the study. Participants were then given two writing exercises, each marked with their participant number which was randomly assigned by a random number generator ( with the minimum number being one and the maximum number being 100. The two writing exercises were created and selected in order to elicit long responses. A digital camcorder was placed on a tripod which was placed on the opposite side of the dominant hand of the participant and above his/her shoulder in order to record stroke order. It was angled in such a way so as not to capture the faces of the participants. No sound from the video was taken into consideration so as to ensure anonymity. At the beginning of each video, a notecard containing the participant s random number was shown to ensure data from each participant could be matched. After the participants were finished with their exercises, the video recording was transferred to a password protected laptop and deleted from the camcorder. The participants language histories and writing samples were scanned and stored on the same password protected laptop with the hard copies kept in a locked desk.
31 ANALYSIS AND RESULTS Data was coded by me through the use of the writing task and video of each participant writing the exercises. Characters that were erased were not considered, as it was determined by the participants to be mistakes. Analysis was done to determine if the characters they wrote were correctly written at the entire character, semantic radical, phonetic radical, stroke order within the semantic radical, and stroke order within the phonetic radical levels. The time that the participant began and finished writing each character was also recorded. Stroke order was determined through a recommended textbook for their course (Y. Liu, Yao, Bi, Ge, et al., 2009) and through mdbg.net, an online Chinese dictionary. Semantic radicals and phonetic radicals were determined through chineseetymology.org, an online etymology of the history of Chinese characters. A character was considered incorrect if it was missing a radical, used an incorrect radical, or a radical was not written correctly (i.e. too many strokes). A radical was considered incorrect if it was the wrong radical or was not written correctly. Stroke order was considered incorrect if a stroke was written in the wrong direction, the order of
32 31 strokes was incorrect, if one stroke was separated into two or more strokes, or if two or more strokes were combined into one stroke.
33 32 Table 1 Scores on Both Tests and Combined Free Writing Correct Radicals (%) Partic. Correct Strokes (%) Avg. Time Per Char. (sec) No. of Chars. Correct Chars. (%) Sem. Phon. Sem. Phon. Time : : : : Avg : Translation Partic. Correct Radicals (%) Correct Strokes (%) Avg. Time Per Char. (sec) No. of Chars. Correct Chars. (%) Sem. Phon. Sem. Phon. Time : : : : Avg : Combined Partic. Correct Radicals (%) Correct Strokes (%) Avg. Time Per Char. (sec) No. of Chars. Correct Chars. (%) Sem. Phon. Sem. Phon. Time : : : : Avg :
34 33 Table 1 shows a comparison of the number of characters, the percentages of correct characters, radicals, and strokes, times, and average times per character for the free writing task, the translation task, and of both tasks combined. Table 2 Number of Radicals Participant Semantic Radicals Phonetic Radicals Free Writing Total Translation Total Combined Total Table 2 shows the number of semantic radicals and phonetic radicals per participant and combined for the free writing task, the translation task, and the two tasks combined.
35 % Correct No. of Characters 34 Figure 1.1 Number of Characters No. of Translation Chars. No. of Free Chars Participant Figure 1.1 gives a graph of how many characters each participant wrote for the free writing exercise and the translation exercise, as well as the total number of characters. Figure 1.2 Radicals Participant Semantic Free Phonetic Free Semantic Translation Phonetic Translation Figure 1.2 gives a graph of the percentage of correct of radicals (on the surface) each participant wrote for the free writing exercise and for the translation exercise.
36 % Correct % Correct 35 Figure 1.3 Stroke Order Semantic Free Phonetic Free Semantic Translation Phonetic Translation Participant Figure 1.3 gives the percentage of correct stroke order for each participant for the free writing task and for the translation task. Figure 1.4 Combined Scores Participant Correct Chars. Semantic Radical Phonetic Radical Semantic Stroke Phonetic Stroke
37 36 Figure 1.4 gives a graph for all of the percentages of correct characters, radicals, and stroke orders for the free writing task and translation task combined. Figure 1.5 Number of Radicals Total Participant Free Semantic Free Phonetic Translation Semantic Translation Phonetic Total Semantic Total Phonetic Figure 1.5 shows the total numbers of semantic radicals and phonetic radicals used by each participant in each writing task and with both tasks combined, and for all participants in each writing task and both tasks combined. Table 3 through Table 6 (found in Appendix A) present the data from each participant. Each table gives every character that was written by the participant. For the categories of correctness, a 1 represents that the character, radical, or stroke order was correct and a 0 represents that it was not. The free writing task is presented first and the translation task second. At the end of each of these tasks, total numbers of characters and radicals are presented as well as the number of correct characters and radicals written by