The Kanji Code

by Natalie Hamilton

  ABOUT THIS BOOK

  15

  Kanji learning and teaching is not making full use of the phonetic components (Toyoda, 2013). The question is not so much whether phonetics are useful, or which ones are useful, but can they be presented in a way that makes them easy to learn and practical to apply to kanji study?

  I researched some diverse topics as part of the dissertation. An extract of the literature review with some more recent research included follows.

  Learning kanji without adequate support and learning strategies can lead to feelings of helplessness and failure, and giving up

  Usuki defines kanji learning problems as a sense of frustration or helplessness, or anxieties about the difficulty of learning kanji (Usuki, 2000). Respondents in this study reported quickly forgetting kanji even after writing them 50 times, and struggling with the different readings and pronunciation depending on the combinations of characters. A lack of confidence and feelings of failure lead students to consider giving up, and strips them of the enjoyment of learning. Many of the participants spoke of feeling like a failure as they studied hard but continued to get bad results in tests, unable to keep up with the hundreds of kanji they were expected to learn each week. On the other hand, some of the respondents who had persevered reported feelings of enjoyment at the challenge and a sense of accomplishment once they could see that kanji is a system (Usuki, 2000).

  I was one of those students who eventual y gave up studying because of kanji, dropping the subject for the higher school certificate, intimidated by the long list of kanji. So I am intimately acquainted with the feelings Usuki observed.

  Look it up: dictionary issues and kanji

  Getting better at guessing the ON readings provides the practical benefit of being able to look up new words more quickly.

  Using a dictionary is rarely a straightforward process for kanji learners, who often face a ‘chicken and egg’ scenario when it comes to reading unknown kanji characters. That is to say, if they encounter a word written in kanji that they do not understand, they must look it up in a dictionary.

  Yet, if they do not know the pronunciation of the word, they cannot locate it in a standard Japanese-English dictionary, which lists words in their kana order, starting from あ [a] and finishing at ん [n].

  Kanji dictionaries provide a workaround method that does not require knowledge of the reading, but is laborious and requires learners to follow complex steps to locate the kanji. For example, the standard kanji dictionary 16

  ABOUT THIS BOOK

  used by high school students requires users to first work out the kanji’s radical, then count the number of strokes in that radical. They then turn to the radical index at the back of the dictionary, which lists radicals in order of the number of strokes. Once they have located the radical, they scan through a list of kanji containing that radical. When they find the kanji they are looking for, they use the page reference to find the kanji’s entry. This entry is written in Japanese and the definition is likely to be incomprehensible to a Japanese as a Foreign Language (JFL) learner.

  Obviously, this method requires a detailed knowledge of radicals.

  There are some English kanji dictionaries that allow JFL learners to locate kanji by different methods including categorising them by radical type and overall formation, for example, Kanji Learner’s Dictionary (2001). These are certainly easier for JFL learners to use, but the time it takes to look up the meaning of a single kanji means that attempting to read a text made up of around half kanji is a painstaking task for learners. By the time you locate the meaning of the new word, you have lost the flow of the context around it.

  Electronic dictionaries have made the process a little easier, but they can be difficult for beginner learners to use, as they require a learner to draw a kanji character using a stylus in the correct stroke order. If they draw the character too slowly or use the wrong stroke order, the electronic dictionary is unlikely to recognise the character and will instead return a series of seemingly random kanji characters that may be nothing to do with the kanji being looked up. Even now with 13 years of electronic dictionary use under my belt, and a good knowledge of the correct stroke order, I often find I am too slow and the results are far from what I was seeking.

  Online dictionaries have made it faster for learners to look up the pronunciation of unknown characters. Various websites provide hints and learning aids, for example Pop Jisyo (meaning pop-up dictionary) allows users to hover their mouse over an unknown kanji, whereupon a pop-up window is displayed, indicating the meaning and all the ON and kun readings of the kanji character. The Google Translate website allows users to click a speaker icon and hear an automated voice reading a word out loud; it also shows the reading in text. While these are useful learning aids, they are dependent on the learner being connected to the internet, and rather than aiding learners to become self-sufficient in reading kanji, they simply provide an instant answer to the question of a kanji’s reading.

  As learners increase their proficiency, it is a desirable outcome that they will become less reliant on instant learning aids, and more reliant on their own memory and inferring skil s when deducing the readings of unknown kanji in authentic texts such as newspapers and magazines. By learning the phonetics, they will get much better at guessing the readings, and therefore, will be able to look words up more quickly.

  ABOUT THIS BOOK

  17

  They will also be able to take advantage of predictive typing on a keyboard or phone, and start to type in the sound they think it might be, and then scroll through the options that are presented to check if their guess is correct.

  What can we learn from character background learners?

  A 2003 study compared the kanji learning strategies preferred by character background learners (CB) – that is, learners who have been schooled in Chinese characters, for example Chinese speakers – and alphabet background learners (AB) – those schooled using an alphabet-based language (Gamage, 2003). This study provides a rare insight into the impact of orthographic background on learning strategies.

  The participants were beginner level Japanese language students at the undergraduate level in Australia. The strategies used most often were determined to be the ones that the students perceived to be the most helpful. The kanji learning strategies were separated into three categories:

  ‘shape (visual strategies), meaning (semantic strategies) and pronunciation (phonological strategies)’ (Gamage, 2003), noting that recent kanji textbooks emphasise the importance of contextual strategies.

  Significantly, AB learners reported using ‘repeated writing’ strategies significantly more often than CB learners did, while CB learners made better use of the phonological aspects of kanji (Gamage, 2003). CB learners were significantly more likely to use phonetic strategies such as ‘grouping kanji with similar pronunciations and phonetic [components]’ and remembering both the ON reading and kun reading at the same time. In contrast, AB

  learners tended to use visual strategies including ‘picture association to kanji’.

  Gamage notes that while AB learners prefer to use cognitively less demanding strategies, CB learners tend to favour strategies that look at the internal structure of a kanji, such as grouping them by radical or stroke count. Gamage concludes that since CB learners are, to a certain extent, already familiar with the shape and meaning of kanji, they are likely to be able to bypass rote-learning strategies and instead, transfer their prior knowledge of kanji due to the wide exposure in their native language.

  Gamage recommends that cognitive processing strategies that are oriented to a learner’s orthographic background be employed.

  Rather than taking a fatalist view and saying, ‘It’s easier for Chinese students’, and AB learners just have to learn by rote, a more positive approach is to ask the question: ‘What can we learn from how CB learners approach learning Japanese kanji?’ If Chinese learners group kanji with similar pronunciations and phonetic compone
nts, this method should be equal y effective for non-Chinese learners, if they are taught the phonetic components and their associated pronunciations.

  18

  ABOUT THIS BOOK

  Phonetic components – an unknown quantity and hard to teach

  Although the radicals and their role as indicators of kanji meaning are general y well known, the role of phonetic components is not easy to define (Miwa, 2012). This is partly because the roles of the two types of components are not mutual y exclusive. For example, it is possible for the same component to play the role of the phonetic component in one kanji, and the semantic component in another (Toyoda, 2000).

  Furthermore, while semantic components are mostly on the left and phonetic components are mostly on the right, this is only a rule of thumb and there are many exceptions (Flores d’Arcais, Saito & Kawakami, 1995; Toyoda, 2000). For example, the character for ‘knife’ 刀 [katana] is used as a semantic component in about 80 complex characters and as a phonetic (dao) in around 12 (Flores d’Arcais et al., 1995).

  Inexplicably, Japanese students are not introduced to the phonetic components in a structured form, and Japanese children learn kanji readings without these clues (Townshend, 2011). They have only recently become a topic of focus in Japan, with books released in 2012. The idea was definitely in the air while I was writing my dissertation.

  Kanji is phonetic and reading it isn’t that different from reading alphabets like English

  Until recently, it was commonly believed that as a logographic script, kanji, and indeed Chinese, were only read for meaning, not sound. Most textbooks describe hiragana and katakana as the phonetic scripts, as if kanji was completely different.

  Experimental evidence in 1991 supported the theory that reading Chinese characters involves phonological recoding not dissimilar to the process used when reading alphabetic scripts (Lam, Perfetti & Bel , 1991). This suggests that individual kanji characters have visual components which provide clear clues to their phonetic readings, and that native speakers and readers of logographic languages are equipped with knowledge of these phonological clues, whether consciously or not. Flores d’Arcais et al. (1995) concluded that readers of logographic scripts work out the pronunciation of unknown kanji characters, either via phonological information (post-lexical y) or through the phonological information associated with the character. Furthermore, they state that the existence of phonetic radicals allows readers to follow a grapheme-to-phoneme route (Flores d’Arcais et al., 1995).

  Bernhardt (1991) theorised that subvocalisation of some sort might be occurring during second language (L2) reading in Japanese or Chinese.

  Subvocalisation refers to the process of sounding words out in one’s head while reading. In other words, JFL learners ‘ hear when they read’

  ABOUT THIS BOOK

  19

  (Bernhardt, 1991, p. 76). This strengthens the case against the idea that kanji are not a phonetic script. The results of a study by Matsunaga &

  Crosby (1997) indicated that native Japanese participants perceive both sounds and meaning during the process of reading kanji, and the authors concluded that phonological coding is involved in kanji processing. If this were the case, it would be useful for JFL learners to be instructed on the phonological encoding inherent in kanji reading.

  Sounding it out helps with inferring

  Subvocalisation accounts for the preference of CB learners towards using phonetic strategies. One study by Kondo-Brown (2006) investigated how non-native learners of Japanese with English as their first language infer unknown kanji words in authentic texts. The participants were all advanced level Japanese language students in Hawaii, although some students were recorded as having higher proficiency than others. Some students had Japanese heritage, meaning that Japanese-speaking parents in an English country had raised them, while others were from a whol y English background, a factor that was taken into account in the discussion.

  The results indicated that students were better at inferring the meanings of unknown characters when they were able to process the words phonological y, that is, correctly pronounce the kanji words. ‘The more comprehensive a [student’s] control of the phonological properties of a word, the greater the ability to inference’ (Kondo-Brown, 2006, p. 144).

  This means that if you see a kanji compound that you don’t know, but you are able to work out the pronunciation, you have a better chance of guessing the meaning.

  Since students of advanced Japanese are expected to learn vocabulary from context while reading authentic texts (Kondo-Brown, 2006), the ability to correctly pronounce unknown kanji characters should therefore be a valuable skill for JFL learners.

  Learning the phonetics and other visual clues to the ON reading therefore puts students in a better position to be able to guess the meaning of new words, and gives them a sense of confidence and control over their learning, countering the feelings of helplessness documented by Usuki.

  The Chinese ON readings have more in common with modern Chinese readings than you may realise Students from a Chinese background tend to be better at learning Japanese.

  This is partly because of the similarity in the readings and meanings, but also because they have already been trained in how to interpret the components of kanji.

  20

  ABOUT THIS BOOK

  The Chinese character script hanzi and kanji share the same etymological origin as well as the same semantic core (Lieh-Ting, 2008). In a study on two-character kanji compound words between Japanese and Chinese, Hong (2005) analysed 17,049 kanji compounds based on the similarities between Chinese and Japanese orthography and meaning, and found that 10,189 or 59.76% of all kanji compounds had either the same or overlapping orthography and meaning.

  As well as the similarities in meaning, there are also regular and systematic correspondences between the phonetic elements of modern hanzi and kanji ON reading, which allow L1 Chinese learners to quickly establish phonetic associations, assisting in learning and inferring the pronunciation of unfamiliar kanji words (Lieh-Ting, 2008). Kanji words that originate from hanzi and are pronounced in the ON reading give Chinese learners of Japanese a sense of familiarity, which allows them to memorise the kanji much more easily through both visual representations (shapes) as well as the approximation of the sounds based on their Chinese counterparts (Lieh-Ting, 2008). For example, the kanji compound 安全 refers to ‘safety’

  in both Japanese and Chinese, and its pronunciation is [anzen] in Japanese and [ānquán] in Chinese, making it very easy for Chinese learners to memorise both the meaning and the sounds based on their first language (L1) knowledge (Lieh-Ting, 2008).

  Wang (2004) analysed vowel correspondences between hanzi and the ON reading of 2,136 kanji characters from the Joyo Kanji to find thirteen correspondence rules of Sino-Japanese vowels. Wan (1999) described the Sino-Japanese phonological correspondence between ancient hanzi and modern hanzi as an assonance relationship. Phonological similarities serve as an auxiliary tool for L1 Chinese learners to learn kanji through sound-spelling correspondences. A number of studies have suggested that this correspondence could be used to assist L1 Chinese learners in inferring Japanese kanji pronunciation (Lieh-Ting, 2008). Lieh-Ting proposed an instructional method to take advantage of these similarities, in which kanji were introduced to students in phonetic groups, through a web-based Sino-Japanese phonological correspondence system. The effectiveness of this method was tested by assigning two groups of participants kanji compound homework over a five-week period. One group was given access to the web-based system, while the other group were instructed to use hard-copy dictionaries to look up the meaning and pronunciation of kanji compounds.

  The results of the five-week study were empirical y examined to test the effect of the proposed method on learners’ strategy use. The results showed that the participants who use Lieh-Ting’s instructional method went from relying mainly on visual-based strategies to using both visual and phonological-based strategies. Furt
hermore, the participants who used Lieh-Ting’s method performed significantly better than those who did not ABOUT THIS BOOK

  21

  (Lieh-Ting, 2008). While being careful to acknowledge that negative interference from hanzi knowledge sometimes results in problems, and hanzi and kanji are not identical, Lieh-Ting (2008) concludes that the benefits for Chinese background learners may still be greater than such interference problems and that knowledge of hanzi may allow L1 Chinese learners to gain an edge over alphabetical background learners.

  These methods are useful to learners who know Chinese; a similar approach of teaching the phonetic clues in kanji could help students without prior knowledge of Chinese.

  The natural advantage visualisers have in learning kanji In recent years, the idea of learning styles has gained momentum and been used in teacher training. Learning style refers to an individual’s natural and preferred method of absorbing and retaining new information and skil s (Reid, 1995) and to their overall approach to the business of learning (Ehrman, 1996). Cognitive styles are defined as an individual’s ‘preferred and habitual modes of perceiving, remembering, organising, processing and representing information’ (Dörnyei, 2005, p. 129) and relate more to style than to ability.

  The Verbal-Imagery Style dimension divides individuals into those who are outgoing and tend to represent information while thinking aloud, and those who are more inward and tend to think in mental pictures (Dörnyei, 2005).

  Dörnyei states that verbalisers are better at working with verbal information, while imagers, also referred to in the literature as visualisers, are superior at working with visual or spatial information (Dörnyei, 2005). Verbalisers, tend to prefer a stimulating environment, whereas imagers are more content with a static environment.