Does Romaji Help Beginners Learn More Words?

Yoshiko Okuyama 355

CALL Vocabulary Learning in Japanese: Does Romaji Help Beginners Learn More Words?

YOSHIKO OKUYAMA University of Hawaii at Hilo

ABSTRACT This study investigated the effects of using Romanized spellings on beginner- level Japanese vocabulary learning. Sixty-one ﬁrst-semester students at two universities in Arizona were both taught and tested on 40 Japanese content words in a computer-assisted language learning (CALL) program. The primary goal of the study was to examine whether the use of Romaji—Roman alphabetic spellings of Japanese—facilitates Japanese beginners’ learning of the L2 vocabulary. The study also investigated whether certain CALL strategies positively correlate with a greater gain in L2 vocabulary. Vocabulary items were presented to students in both experimental and control groups. The items included Hiragana spellings, colored illustrations for meaning, and audio recordings for pronunciation. Only the experimental group was given the extra assistance of Romaji. The scores of the vocabulary pretests and posttests, the types of online learning strategies and questionnaire responses were collected for statistical analyses. The results of the project indicated that the use of Romaji did not facilitate the beginners’ L2 vocabulary intake. However, the more intensive use of audio recordings was found to be strongly related to a higher number of words recalled, regardless of the presence or absence of Romaji.

KEYWORDS CALL, Vocabulary Learning, Japanese as a Foreign Language (JFL), Romaji Script, CALL Strategies

INTRODUCTION Learning a second language (L2) requires the acquisition of its lexicon. How do American college students learn basic L2 vocabulary in a CALL program? If the vocabulary is written in a nonalphabetic L2 script, such as Japanese, is it more efﬁcient to learn the words with the assistance of more familiar Roman-alphabetic symbols? This experimental study explored these questions in the context of Japa- nese CALL vocabulary learning.

CALICO Journal, 24 (2), p-p 355-379. © 2007 CALICO Journal 356 CALICO Journal, Vol. 24, No. 2 Teachers of less commonly taught languages (LCTLs) are faced with a variety of issues that include lack of pedagogically sound resources and instructional materials (Johnston & Janus, 2003). Although categorically an LCTL, Japanese is in fact the most commonly taught Asian language in the United States. A great challenge awaits learners of this nonalphabetic language, however, because Japanese is ranked as category 4 language (highest) by the US Federal standards in terms of its difﬁculty for American students to acquire. While more college-level course books are being published and software programs being created, many aspects of teaching and learning Japanese still remain to be empirically investigated. One of these underresearched aspects is the effect of nonalphabetic script on word learning.

Japanese Orthography Japanese consists of three types of script: Hiragana, Katakana, and Kanji. The first two are called “syllabaries” because each symbol is a syllabic unit, while Kanji characters are ideographic symbols. Japanese children have to memorize the two syllabaries, with each set made of the basic 46 syllabic units and 61 extensions, before mastering over 2,000 Kanji characters in order to master all three sets of Japanese script. This is no easy task for learners since the difficulty with the syllabaries only worsens when learners come to understand that not all symbols clearly map onto the sound units of the Japanese language (e.g., the same sound ‘e’ in oneesan ‘big sister’ and eego ‘English’ happens to be transcribed with different hiragana symbols). Yet, the perceived difficulty in learning Kana might be relative. According to the orthographic depth hypothesis (Katz & Frost, 1992), it is easier to learn to read words written in a ‘transparent’ script than an ‘opaque’ script. Kana syllabaries are considered a transparent script, a type of orthography in which phoneme-grapheme mappings are highly consistent, and these symbols are processed differently than Kanji characters by native speakers of Japanese (Ellis et al., 2004; Kawaka- mi, Hatta, & Kogure, 2001; Sumiyoshi et al., 2004). English, on the other hand, is called an opaque script, a type of orthography that lacks systematic sound-symbol correspondence. In English, many alphabetic letters are associated with several different sounds (particularly in the case of its five vowels), making the mapping of the letters to the sounds less predictable. Thus, unlike the English alphabet or Kanji characters, “the regularity of the symbol-sound mappings makes hiragana an exceptionally transparent orthography” (Ellis et al., 2004, p. 443).

Romaji versus Japanese Script According to the ‘Standards for Japanese Language Learning’ (National Stan- dards in Foreign Language Education Project, 1999), there is an orthographic bar- rier between English and Japanese: “In order to be able to read Japanese materials written for adult native speakers, students must learn two different syllabic writing systems and approximately 2,000 Chinese characters (kanji), most of which have multiple meanings and readings” (p. 332). The complexity of the Japanese writing Yoshiko Okuyama 357 system poses a great challenge to learners of Japanese as a foreign language (JFL) especially at the beginning level. In the US, Romaji (i.e., Romanized spellings of Japanese text) is commonly used as a starter for JFL beginners. Romaji is not entirely foreign to native speakers of Japanese. The script is used on limited occasions by Japanese native speakers, such as writing their name on a passport or indicating the name of a train station to foreign visitors. However, it is not an integral part of the native orthography and is not mixed with the other scripts within the same text. Romaji does not always transcribe the spoken language in a perfect grapheme-phoneme match. Moreover, both the Hepburn system (a style of Romanization invented by a missionary in 1867) and Kunrei-shiki (a Romanization system adopted by the Japanese government in 1937) have been used in Japan (Hannas, 1997). Two varying ways to transcribe some sounds (e.g., /fu/ vs. /hu/, /zi/ vs. /ji/) cause confusion on the part of readers and writers. Con- trary to Romaji, “grapheme-to-phoneme correspondences are entirely regular” in Hanyu Pinyin (meaning ‘assembling sound’), a phonetic alphabet set of 26 Ro- man letters in 13 letter groups used in addition to the traditional character system in China (Chen, Fu, Iversen, Smith, & Matthews, 2002, pp.1089-1091). Pinyin is taught to school age children in China as a phonetic assistance in learning a set of about 6,000 meaning-based characters. Like Romaji, however, Pinyin is not used as an independent written script nor mixed with the Chinese characters. Thus, it is much easier for (adult) native speakers of Japanese and Chinese to process texts entirely in the traditional orthography rather than in Romaji or Pinyin. Many textbooks for JFL learners published in the US seem to encourage the use of Romaji as an effective pedagogical aid. Books designed for self-study, such as Japanese in 10 Minutes a Day (Kershul, 1992) and Master the Basics: Japanese (Akiyama & Akiyama, 1995), are also written entirely in Romaji. An audiotape- based program, Just Listen ‘n Learn Japanese (Katao & Takada, 1994), is accom- panied with a transcription of the recordings written only in Romaji. By contrast, college-level textbooks for JFL beginners, such as Yookoso (Tohsaku, 1994) and Nakama (Makino, Y. Hatasa, & K. Hatasa, 1998), are mainly written in authentic Japanese orthography. When it comes to CALL software, existing programs lack consistency in spelling Japanese materials. CALL tutorials and commercially available language software vary in the degree of their use of Romaji as opposed to authentic Japanese orthography. For example, NihongoWare 1 presents vocabulary and conversational materials exclusively in Romaji, while TriplePlay Plus! Japanese has these items only in Japanese script. Yet, some programs, such as Robo-Sensei (Nagata, 2004), come in two versions—Romaji only and Japanese script only—for the user to choose from. In the North American context, Romaji is assumed to be an effective learning aid particularly during the initial period of JFL learning. However, when to switch to the authentic script has been controversial among JFL practitioners (Dewey, 2004; Hatasa, 2002). Those who advocate the early introduction of kana and kanji have pedagogical philosophies quite different from those of the proponents of delayed introduction. Divided views on the use of Romaji also exist among JFL 358 CALICO Journal, Vol. 24, No. 2 learners. Even if the kana syllabaries are taught in the early stages of instruction, as in most JFL classes, many students are likely to use Romaji as a quick and easy note-taking device throughout the first year. These students claim that the use of Romaji reduces their language-related anxiety and helps them overcome the challenge of learning Japanese especially when their L1 is English, a Roman alphabetic language. They may adamantly defend their Romaji transcription as a way to absorb a large amount of new words efficiently for quizzes and exams. Others may simply consider this assistive device as a crutch, inhibiting themselves from developing strong reading and writing skills in Japanese. Moreover, JFL learners’ views on Romaji may also be influenced by their teachers’ attitudes toward the script (Dewey, 2004). The question is: does the use of Romaji really ‘facilitate’ JFL beginners’ language learning? Little research has been done to test whether substituting Japanese orthography for Romaji in introductory textbooks or CALL programs is indeed an effective pedagogical tool. In the absence of empirical evidence for the value of Romaji, teachers are left to select courseware or textbooks for beginning-level students based on their experience as an L2 learner or their own instincts. The primary goal of this study is, therefore, to find empirical evidence to address the question of the pedagogical value of Romaji in CALL for JFL beginners.

L2 Vocabulary Learning across Different Nonalphabetic Orthographies The main issue of second language acquisition discussed in this study is college- level students’ L2 word learning in a nonalphabetic language. L2 vocabulary gain plays a crucial role especially in beginning SLA (Ellis, 1995). Although the topic of vocabulary is no longer undervalued in SLA research due to a rapid increase in L2 vocabulary research, little is available regarding how L2 learners process, store, and retrieve words written in nonalphabetic script. Because research on L2 word recognition skills tends to be conducted on the learners of alphabetic languages, some issues specific to the lexical acquisition of nonalphabetic languages are yet to be investigated. As mentioned above, a major challenge to the beginners in Japanese is to learn new vocabulary in the orthographically different script. In fact, for L2 learners of any nonalphabetic language, the orthographic form is one key element affecting lexical learning: “The form of items is more likely to influence difficulty than meaning, because there is much more shared knowledge of meaning between two distinct languages than there is shared form” (Nation, 2001, p. 29). Learning new words by itself is a complex process and requires learners to access the semantic representation of the new word while simultaneously making sound-spelling correspondence. Naturally, the process becomes even more demanding if the L2 script is completely different from that of L1. “For learners whose first language is not related to the second language, the learning burden will be heavy” (Nation, 2001, p. 24) because the “learning burden of the written form of words will be strongly affected by first and second language parallels” (p. 45). L2 orthography also plays a significant role even beyond word recognition. L2 learners’ acquisi- Yoshiko Okuyama 359 tion of reading skills is dependent upon their “ability to identify the printed (i.e., orthographic) form of a word or lexical item in order to activate the word’s meaning, structural/syntactic information, and other pragmatic or world knowledge as- sociations” (Fender, 2003, p. 291). The orthographic distance between L1 and L2 creates a cognitive overload in deciphering L2 lexical items or processing L2 reading materials (Akamatsu, 1999; Fender, 2003; Laufer, 1997; Koda, 1997, Tan et al., 2003), whereas the similarity in L1 and L2 spelling patterns facilitates word recognition (Muljani, Koda, & Moates, 1998). The orthographic mismatch may make L2 word learning challeng- ing for beginners but not impossible. For example, Wong, Perfetti, and Liu (2003) found that native speakers of English were able to develop sensitivity toward structural complexity and compositional relationship of Chinese radicals at the early stage of L2 learning. According to their lexical processing model of Chinese characters, understanding Chinese characters involves three interlinked constitu- ents: orthographic, phonological, and semantic. Since one-to-one grapheme-phoneme mappings are unavailable in the logographic script of Chinese, L2 learners’ visual-orthographic processing turned out to be the most critical element. In this model, learners of Chinese must first work on stroke analysis, seeking information in the orthographic unit of the character, and then access the phonological unit of the character as well as its semantic unit. Yet, first-year learners of Chinese have not made a strong connection between the orthographical and phonological units. Chung (2002, 2003) also looked at alphabetic learners of Chinese, shedding light on how to reduce cognitive overload derived from L1-L2 script discrepancy. When it comes to alphabetic learners of Japanese, research has been done primarily on American students’ processing of Kanji characters (e.g., Matsunaga, 1995, 2001) rather than Kana. As mentioned before, Japanese syllabaries and kanji characters require different types of processing. Therefore, drawing directly on the findings of research in logographic script does not suffice in understanding how L2 learners process and store words written in syllabic script. The current study fills a critical need by investigating whether Romaji, a phonetic assistance consisting of visually familiar Roman alphabet letters, helps JFL beginners overcome the burden of learning words in Hiragana, a nonalphabetic orthography.

CALL Strategies This study also investigated what strategies students were likely to use when learning L2 words in multimedia software. Gathering useful information on learners’ strategies has always been a challenge to researchers. Survey studies on vocabulary learning strategies (e.g., Kudo, 1999) provide an insightful account on learners’ self-reported strategy use, but there is often a discrepancy between what strategies L2 learners report having used and what they actually used in language- learning problem solving. To obtain a holistic view of L2 learner behaviors, we need to incorporate a method of recording the students’ actual strategy use. Here, CALL research technologies come in handy: the computer can be pro- 360 CALICO Journal, Vol. 24, No. 2 grammed to compile inventories of behavioral patterns followed by users of language software. The use of hypermedia presents intensive interaction between user and computer and executes a variety of tasks, allowing for a rich recording of online learning strategies in an unobtrusive way. Recently, there has been a surge of interest in CALL strategies and CALL user tracking technologies (Al- Seghayer, 2001; Ashworth, 1996; Collentine, 2000; Hwu, 2003; Hegelheimer & Tower, 2004; Laufer & Hill, 2000; Liou, 2000; Vincent & Hah, 1996). For example, Ashworth’s (1996) CALL program, The Observer, was designed to store L2 learners’ keyboard activities (e.g., mouse-clicks and cursor move- ments) in computer files. He suggested the possibility of observing other online actions, such as transcriptions of user input in computer-mediated conversational exchanges or frequency counts of accessing online dictionaries and thesauruses in a reading comprehension task. Liou (2000) also emphasized the advantage of using computers’ recording abilities to collect learner data. Hwu (2003) used WebCT’s student-tracking system to collect learner data during CALL activities. Other SLA studies have also incorporated user behavior tracking technologies as a data collection methodology or have documented strategies employed by CALL users for L2 word learning (Collentine, 2000; Laufer & Hill, 2000; Al-Seghayer, 2001), showing great promise for conducting L2 vocabulary research in this elec- tronic medium. CALL is thought to have great potential in increasing the amount of L2 input and improving the relatively low L2 achievement by learners of Asian languages (McMeniman, 1997). Although Chinese, Japanese, Hebrew and other nonalphabetic LCTLs have traditionally suffered from a shortage of orthographically well designed CALL programs due to problems in “displaying ideographs or right- to-left writing” (Ariew, 1991, p. 34), the development of Unicode has improved CALL programs’ capability of handling foreign fonts (Corda & Van Der Stel, 2004). However, it is still unknown how efficiently CALL can assist learners of non-Western, nonalphabetic languages. For instance, what sort of CALL strategies do JFL students use in a self-paced computer environment? What CALL strategies facilitate the learning of this orthographically complex difficult language, especially with respect to trying to absorb as many new words as possible? The current study attempts to demonstrate the feasibility of applying computer technology to document CALL strategies employed in learning nonalphabetic vocabulary materials. The findings from the study will also add insight on the relationship between L2 script and vocabulary learning in a CALL environment, providing useful implications to JFL teachers as well as software developers.

METHODOLOGY The main purpose of the study was to investigate the effects of using Romaji on English-speaking college students’ learning of beginning-level Japanese words. The research question was whether or not the availability of Romaji signiﬁcantly facilitated learners’ short-term learning of such words when using a CALL program. Yoshiko Okuyama 361 The primary independent variable of the study was the type of orthography used to present the Japanese vocabulary items (i.e., Japanese vocabulary instruction in Hiragana only or together with Romaji). The dependent variable was L2 learners’ immediate vocabulary increase in Japanese, measured by both kana and sound recognition tests in the same CALL environment. Because the research purported to test the impact of the independent variable (i.e., script type) on the dependent variable (i.e., L2 vocabulary learning), an experimental design was selected. The students in the control group learned new words in a ‘Hiragana-only’ version of the CALL program, while the students in the experimental group learned the same words in a ‘Romaji-plus’ version of the program. It was hypothesized that the use of Romaji would result in better attainment of the newly learned nonalphabetic (Japanese) words by English-speaking students.

Subjects The target population for this experiment was defined as English-speaking learners of entry level Japanese who were enrolled in American universities. A sample of students from first-semester Japanese courses was thought to be representative of the identified population. Sixty-one students of first-semester Japanese in two research universities were randomly assigned to either the experimental or control group. The control group (n = 31) was made of 18 students from Arizona State University (ASU) and 13 students from Northern Arizona University (NAU), while the experimental group (n = 30) consisted of 17 students from ASU and 13 students from NAU. The average age of the students was 21.7 years. Although there was a gap in the mean ages of the students at the two universities (M = 20.8 for ASU; M = 23.0 for NAU), an unpaired t test showed that the discrepancy in age distribution was not statistically significant when the alpha level was set at .01: t (59) = -2.101, p = .0399. A frequency distribution of each of the other demographic characteristics was also made between the groups by university. There were more male students (23 in ASU, 16 in NAU) than female students (12 in ASU, 10 in NAU). More than half of each group (n = 20 for both groups) had no experience with the Japanese language prior to the semester. Sixteen ASU students identified themselves as ‘fluent’ in a language other than English, while only 6 NAU students identified themselves as ‘fluent’ in another language. However, Each group had the same number (n = 5) of international students (i.e., nonnative speakers of English). The demographic comparisons of the two school groups indicated that, whereas NAU students were relatively older, ASU students had slightly more linguistically en- riched backgrounds. Despite these few differences, the overall background characteristics of ASU and NAU students were similar. Thus, it was appropriate to treat the two school groups as one and to draw conclusions about the population. The students at both institutions used the same textbook, Yookoso! in their classes.1 Having little or no previous knowledge of Japanese, the students at both ASU and NAU were introduced to the hiragana and katakana syllabaries in class within the first few weeks of the semester. They were instructed to use the vo- 362 CALICO Journal, Vol. 24, No. 2 cabulary CALL program to learn new Japanese words (i.e., words other than those presented in the textbook) as a supplemental learning task.

Pilot Phase of the Project Early prototypes of the CALL program were piloted for feasibility and effectiveness in a series of sessions with 20 volunteer subjects. Feedback from the volunteer subjects enabled the researcher to address problematic areas and strengthen the program’s design and functionality, including the use of English translations to reinforce learners’ accurate interpretation of visual and audio input and the addition of situational contexts (e.g., dialogues) to enrich the vocabulary learning.

Hardware and Software The instructional materials used in the vocabulary CALL program were an adaptation of a commercially available CD-ROM software program, Learn to Speak Japanese (1994), a ‘self-paced’ language program for beginning-level learners of Japanese. Four of 20 lessons in Learn to Speak Japanese, four lessons were selected and adapted for this study.2 The existing words of the four lessons and their accompanying illustrations and audio recordings were entirely replaced with new words and color drawings as well as new recordings by two different Japanese speakers (a male and a female). The modified lesson materials contained a total of 40 words that were not included in the first-semester Japanese class (for more information, see the Tasks section below). The CALL program, designed for Macintosh,3 consisted of three modules: (a) a preview module on how to use the CALL program, (b) a lesson module containing the exercises, and (c) a testing module to administer tests on vocabulary recall. It is important to emphasize that the lesson module was designed not only to provide L2 vocabulary lessons to the students, but also to gather ‘learner data’ on each student (i.e., data on online learning behaviors such as how long the student spent on each lesson). Similarly, the testing module was programmed to execute the test batteries and then to record an individual student’s test scores and other information (e.g., reaction time to test items). The method of computer-mediated data collection was chosen because of its ability to discretely and objectively capture language-learning data and language- testing data and to reduce internal threats to validity such as teacher experience during the learning session (Chapelle and Jamieson, 1991). Specifically, in this study, the program randomized the test items in the testing module. Students were tested on the same words but in different orders, thereby greatly reducing subject sensitivity toward individual items that could have developed in the lesson module. This method also randomly distributed words of varying difficulty (e.g., word length ranging from two-syllable to five-syllable words) throughout the test. Thus, computer technology helped not only to control the effect of extraneous variables but also to secure accuracy and consistency in measurement. The study was able to pool a large amount of data without data collection errors. Yoshiko Okuyama 363 Tasks The aim of instruction in this experiment was to expose the students to 40 Japa- nese lexical items in two versions of a CALL program (with Romaji and without Romaji) in order to test which version can teach L2 vocabulary lessons more effectively. All the words were concrete nouns in Japanese and were of appropriate length and complexity for the student’s proficiency level. Words containing consonant-glide combinations (i.e., hiragana syllables such as kya, hyi, ryo) and geminates (i.e., double consonants as in sakka) were avoided except for one item (sentakki ‘refrigerator’). The selected words were relatively high frequency words, and the range of the number of syllables varied from two to five. All words were hiragana type words that fit each lesson’s vocabulary theme and which had clear English equivalents. Because the students lived in Arizona, familiar creatures of desert life (e.g., scorpion and hummingbird) were also included. The Japanese instructors at each school were also consulted on the final corpus of vocabulary items in order to eliminate any words of potential familiarity to the students. For preliminary vocabulary assessment, 10 words were drawn randomly from the 40 for each of two pretests. All 40 words used in the lessons and the tests are listed in Table 1. Table 1 List of Japanese Vocabulary Items Lesson 1 Structural Emphasis: Likes and dislikes Vocabulary theme: (… は、すきですか。) Desert animals 1. しか shika ‘deer’ 2. へび hebi ‘snake’ 3. とかげ tokage ‘lizard’ 4. うずら uzura ‘quail’ 5. さそり sasori ‘scorpion’ 6. やまねこ yamaneko ‘wildcat’ 7. こうもり koomori ‘bat’ 8. ふくろう fukuroo ‘owl’ 9. はちどり hachidori ‘hummingbird’ 10. あらいぐま araiguma ‘raccoon’ Average word length: 3.4 syllables Lesson 2 Structural emphasis: Describing things Vocabulary theme: (... は、すごい/きれい/とおいです。) The physical world/the cosmos 1. しま shima ‘island’ 2. たき taki ‘waterfall’ 3. うみ umi ‘sea’ 4. もり mori ‘forest’ 5. かざん kazan ‘volcano’ 364 CALICO Journal, Vol. 24, No. 2

6. かせい kasee ‘Mars’ 7. どせい dosee ‘Saturn’ 8. たいよう taiyoo ‘Sun’ 9. せきどう sekidoo ‘equator’ 10. ながれぼし nagareboshi ‘shooting Average word length: star’ 3.0 syllables Lesson 3 Structural emphasis: This and that Vocabulary theme: (これ/あれ/それは何ですか。) Ocean life 1. ます masu ‘trout’ 2. かに kani ‘crab’ 3. かい kai ‘seashell’ 4. さめ same ‘shark’ 5. かめ kame ‘turtle’ 6. ひとで hitode ‘starfish’ 7. くじら kujira ‘whale’ 8. うなぎ unagi ‘eel’ 9. かもめ kamome ‘seagull’ 10. とびうお tobiuo ‘flying fish’ Average word length: 2.6 syllables Lesson 4 Structural emphasis: Where is it? Vocabulary theme: (… は、どこにありますか。) Household items 1. はさみ hasami ‘scissors’ 2. かびん kabin ‘vase’ 3. かがみ kagami ‘mirror’ 4. でんち denchi ‘battery’ 5. そうじき soojiki ‘vacuum cleaner’ 6. でんたく dentaku ‘calculator’ 7. ごみばこ gomibako ‘trash can’ 8. ふうとう fuutoo ‘envelope’ 9. れいぞうこ reezooko ‘refrigerator’ 10. せんたくき sentakuki ‘washing Average word length: machine’ 3.8 syllables

Each word was presented with a color drawing and an English translation. The use of the English gloss was necessary to ensure the clarity of meaning of the L2 word, which may not always be immediately evident in a drawing alone. There were also buttons for the pronunciation of a single word and for a conversation containing the relevant vocabulary item. The audio recordings of single-word pronunciations as well as dialogues were made in a male voice for some and in Yoshiko Okuyama 365 a female voice for others. Both speakers were native speakers of standard To- kyo Japanese. Each dialogue was a dyadic conversation scripted to present the keyword in a sentence structure familiar to the ﬁrst-semester Japanese learners. For example, in lessons 1 and 2, Speaker A asked a yes/no question (e.g., X wa suki desuka? ‘Do you like X?’ or X wa tooi desuka? ‘Is X far away?’), to which Speaker B responded positively or negatively. In lessons 3 and 4, Speaker A asked an interrogative question (e.g., Sore wa nan desuka? ‘What is that?’ or X wa doko desuka? ‘Where is X?’) to which Speaker B supplied pertinent answers. These sentence structures were also in accordance with the grammar instruction that both the ASU and NAU students had already been provided in their regular classroom lessons. The spellings of the vocabulary and dialogues were presented in Hiragana because of the students’ proﬁciency level. The Romaji spellings of the 40 words (not the dialogues) were included in the version of the CALL program used by the students in the experimental group. Figure 1 shows the structure of the vocabulary displays in the lesson module. Figure 1 Vocabulary Display in the Lesson Module

Button Button Button Drawing with Navigation buttons to hear dialogue to review for grammar button for and see its hiragana explanation single-word written script syllabary pronunciation

In the lesson module, the following types of information were collected: (a) total learning time (i.e., the time the subject spent per lesson and on all the lessons), (b) audio access (i.e., the number of clicks made on the drawing to hear the Japa- nese pronunciation of the word), (c) kana access (i.e., the number of clicks made on Kana to review the syllabary of the 46 basic hiragana symbols and to learn some tips on how to read sets of symbols as words), (d) grammar access (i.e., the number of clicks made on grammar button to access grammatical explanation), and (e) dialogue access (i.e., the number of clicks made on dialogue button to hear the dialogue and have it displayed in Japanese text). 366 CALICO Journal, Vol. 24, No. 2 The testing module was made up of two posttests. The purpose of the posttests was to measure students’ short-term vocabulary learning. The immediate recall of the instructed L2 vocabulary was evaluated in terms of L2 sound recognition and of L2 script recognition. The ﬁrst posttest, “Sound Games,” consisted of four pages of sound recognitions on all the 40 words and measured students’ ability to identify the illustration (for meaning) that corresponded to an L2 sound cue. The second posttest, “Kana Games,” presented four pages of kana script recognition test on the same 40 words and assessed students’ ability to select the correct illustration based on the hiragana spelling shown in the upper left-hand corner of the display (see Figure 2). All the test directions were written in English.

Figure 2 Sample Kana Game

In the testing module, the program collected learner information (e.g., subject’s name, response items, and score of each vocabulary test automatically calculated by the computer) and stored the data in the individual subject ﬁle.

Procedures The students were first given two pen-and-paper pretests, each of which contained 10 Japanese words. The first pretest was a sound recognition test, and the second pretest was a written hiragana recognition test. The pretests were administered to ensure that there was no significant difference in Japanese vocabulary knowledge between the experimental and control groups as well as between the two university groups. After a brief orientation session in which the features of the program were presented and explained, the students started the vocabulary lessons at their own computer stations. They were told that the main instructional objective was to learn new Japanese words using the multimedia software. They were allowed to navigate from one lesson to another at their own pace. After completing all of the lesson modules, the students took the posttest in the testing module. The Yoshiko Okuyama 367 students completed both the lessons and tests in one sitting at the laboratory. After exiting the program, the students were asked to fill out a short questionnaire as the last procedure. The main purpose of the exit questionnaire was to collect various information on the learners. The first portion of the questionnaire was designed to obtain subject characteristics (e.g., gender, age, major, native language, prior exposure to Japanese and length of experience in learning Japanese, and prior foreign language learning experience). Questions related to the student’s demographic background were adapted from Grace’s (1995) CALL experimental study on L2 learners of French. Other parts were drawn from a study of JFL learners by Okamura (1995) for comparative purposes.

RESULTS Group Differences The results of the analysis of the subject characteristics from the questionnaire showed that both gender and age were evenly distributed between the groups and the universities. The only substantial difference was the amount of previous Japa- nese exposure. In the control group (n = 31), 25 students had had some experience learning Japanese (e.g., as a language course requirement in high school) prior to taking the first-semester Japanese course in college. In the experimental group (n = 30), only half of the students had had some prior exposure to the language. However, in spite of the discrepancy in their previous Japanese experience, the groups did not differ in their knowledge of the Japanese vocabulary items. The results from a t test confirmed that the group-based difference in pretest scores was not statistically significant at the .01 level. Table 2 shows the means and standard deviations of the pretest scores by group and pretest type.

Table 2 Means and Standard Deviations of Pretest Scores by Pretest Type and by Group Pretest Group M SD Prestest 1 Cont (n = 31) 1.000 1.033 Exp (n = 30) 1.233 1.223 Pretest 2 Cont (n = 31) 1.613 1.407 Exp (n = 30) 1.067 1.015

The very low scores in both pretests indicate that the students’ knowledge of Japanese words was minimal prior to the CALL instruction and was evenly spread across the groups.

Overall L2 CAL-based Vocabulary Learning Table 3 presents a summary of the pretest-posttest results for both groups combined. 368 CALICO Journal, Vol. 24, No. 2 Table 3 Means and Standard Deviations of Pretest and Posttest Scores for Combined Groups by Test Type Test M SD Prestest 1 1.115 1.127 Pretest 2 1.344 1.250 Posttest 1 21.262 9.588 Posttest 2 18.738 11.275

While the mean scores of the pretests were very low, those of the posttests indicate a sizable increase in the students’ vocabulary knowledge after the CALL instruction: students correctly identiﬁed approximately half of the 40 items on each posttest (21.3 [53%] in posttest 1 and 18.7 [47%] in posttest 2). This general outcome provides evidence that students learned an appreciable amount of vocabulary as the result of using the CALL program.

Research Questions and Hypotheses The main research question in this study was whether or not Romaji facilitated Japanese beginners’ short-term learning of hiragana words in CALL. The primary research hypothesis was

There is a systematic relationship between the presence or absence of Romaji assistance and students’ gain in knowledge of Japanese vocabulary.

Romaji and Japanese Vocabulary Learning To investigate the effect of Romaji, two-tailed independent t tests were performed between the control and experimental groups on the two posttests. Generally speaking, a one-tailed t test is more powerful. However, the two-tailed procedure was used because both the possible positive and negative effects of Romaji on vocabulary learning had to be considered. The results of the t tests did not show a signiﬁcant difference in either posttest 1 or posttest 2 (t = .262, df = 59, p = .794 and t = .364, df = 59, p = .717, respectively). Another variable measured was reaction time (RT) to the vocabulary items in the posttest, here measured in computer ticks (60 ticks = 1 sec). The experimental group’s mean RT was slightly lower than that of the control group, 183 ticks versus 230 ticks, for posttest 1 and virtually the same for posttest 2. Independent t tests did not show a signiﬁcant difference for either posttest.

CALL Strategies and Japanese Vocabulary Learning As mentioned earlier, this study also examined the students’ use of several learning strategies. These strategies were labeled as “CALL strategies” because they inform us of how students approached the L2 learning tasks in a CALL environ- Yoshiko Okuyama 369 ment. The program recorded the students’ use of the CALL strategies in individual computer ﬁles. The following acronyms are used in reporting the analyses of the use of these strategies:

1. TL = total learning time spent on the CALL program, 2. AA (Audio Access) = clicking the audio recording button for each new L2 word, 3. KA (Kana Access) = clicking the kana tutor button to review the hiragana syllabary, 4. GA (Grammar Access) = clicking the Grammar Help button to learn about the sentence structures used in the program, and 5. DA (Dialogue Access) = clicking the dialogue button to listen to a dyadic conversation containing the target word.

Students’ use of each strategy (other than TL) was measured by the number of clicks made on the relevant button. Table 4 presents descriptive statistics of students’ use of all CALL strategies.

Table 4 Means and Standard Deviations of Use of CALL Strategies Strategy M SD TL 1339.508 631.627 AA 188.410 182.836 KA 1.918 4.961 GA 2.721 2.788 DA 71.311 37.649

The unit of measure for TL was 1 sec; the average TL (1,340 in the table) is equal to approximately 22 minutes. The standard deviation is very large (631.627), indicating a wide disparity in the length of time individual students spent in the program. Among the other strategies, the mean score of AA was found to be the highest (M = 188.410 clicks). The second most frequently used strategy was DA (M = 71.311 clicks). The least used functions were KA (M = 1.918 clicks) and GA (M = 2.721 clicks).

To examine whether the students in the experimental group studied the CALL vocabulary differently than those in the control group, t tests were performed on the use of the strategies. Table 5 displays the results of the t tests. There were no signiﬁcant differences between the groups in the use of any of the strategies. 370 CALICO Journal, Vol. 24, No. 2 Table 5 Between-group Comparisons of the Use of CALL Strategies Strategy Mean diff. df t p TL 34.452 59 0.211 .833 AA 72.629 59 1.570 .122 KA 1.347 59 1.062 .293 GA -0.680 59 -0.951 .345 DA -9.226 59 -0.956 .343

To determine whether students’ use of CALL strategies were related to their vocabulary learning, the control and experimental groups were combined and cor- relations between their use of strategies and posttest scores were computed (see Table 6).

Table 6 Correlation Coefﬁcients of Use of CALL Strategies and Posttest Scores TL AA KA GA DA Posttest 1 .309 .499 -.079 -.078 .056 Posttest 2 .285 .513 -.143 -.133 -.048

The figures in Table 6 show that using AA (sound) was highly correlated with both posttests: .499 for Posttest 1 and .513 for Posttest 2. A stepwise regression analysis confirmed that the frequent use of audio access was indeed a strong predictor of vocabulary learning. To further investigate the extent to which this strategy alone contributed to vocabulary learning, a linear regression was run on AA and the two posttests. In this analysis, AA was used as the predictor variable and the posttest scores as the criterion variable. For posttest 1 (sound recognition), the coefficient of determination (r2) was .249, meaning that approximately 25% of the variance in posttest 1 was explained by the variance in AA (Audio Access). For posttest 2 (kana recognition), the coefficient of determination was .264, meaning that about 26% of the variance in posttest 2 was explained by the variance in AA. Based on the results this analysis, AA is a clearly important predictor of vocabulary learning. To summarize this section, the results of statistical analyses in this study did not provide empirical support for the beneficial role of Romaji in learning new hiragana words within a short instructional period. Instead, the results showed a systematic relationship between the use of L2 audio and L2 vocabulary learning.

DISCUSSION Effects of Romaji It was hypothesized in this project that the use of Romaji would help English- speaking learners acquire Japanese vocabulary because of the similarity of the Yoshiko Okuyama 371 symbols used in Romaji and the Roman alphabet. However, it was found that the experimental group who learned the Japanese content words with Romaji in the CALL program did not score higher in the posttests than the control group who learned the same vocabulary without Romaji assistance. This finding was consistent with the results from both the sound and script posttests. The opportunity to view Japanese vocabulary in Romaji had no effect on the number of words correctly identified by the students in the first-semester Japanese classes. JFL learners’ attitudes toward Romaji has been raised as a significant variable in a survey study (Dewey, 2004). Is it possible that the results in the current study were influenced by the subject group’s preference for Japanese script? In the exit questionnaires, the students were asked whether the presence or absence of Romaji would have helped them remember more Japanese words in the CALL program. The responses from students in both groups were almost evenly divided: 25 students favored the Romaji assistance in the CALL program and 23 students preferred the Japanese script only in the program. (The rest of the students were undecided on the issue.) Of greater importance, these orthographic preferences were evenly distributed between the control and experimental groups. Thus, the beginning-level Japanese learners’ personal beliefs about Romaji were unlikely to have affected the overall L2 vocabulary outcome. Is it possible that the students in the experimental group did not pay attention to Romaji because of their solid familiarity with the hiragana syllabary? As many JFL practitioners can attest, it is highly unlikely that students in first-semester Japanese can fully master the Hiragana syllabary within the first few weeks of instruction. Their knowledge of the syllabary tends to be shaky until the end of the semester. The reaction time data of the study showed that those who viewed the vocabulary items with Romaji during the CALL lessons were slightly faster in matching L2 audio cues with correct meanings. This finding may suggest that the students in the experimental group made use of Romaji to some extent. The difference in reaction time between the two groups may imply that Romaji assisted the experimental group in speeding up the activation of lexical memory, but it should be remembered that the difference was not statistically significant. Contrary to the pedagogical assumptions supporting the use of Romaji, such orthographic assistance did not have an impact on beginning-level CALL-based vocabulary learning in Japanese. Hatasa’s (2002) study on the classroom use of Romaji provided evidence that confirms this finding. Students who were taught with the prolonged use of Romaji in class did not perform significantly better on their midterm and final exams than those who had an early introduction of authentic Japanese orthography. No effects of the use of Romaji were found either on their short-term or long-term development of Japanese proficiency at the introductory level in Hatasa’s study. Therefore, once Hiragana script has been introduced, JFL beginners are ready to handle the “learning burden” of L2 written forms (Nation, 2001) in acquiring basic Japanese words. What are possible explanations for the lack of effects of Romaji assistance, then? One explanation may come from the orthographic depth hypothesis (Katz & Frost, 1992) and the correlation between orthographic transparency and the 372 CALICO Journal, Vol. 24, No. 2 ease in developing L1 word recognition skills (Ehri, 1999; Ellis & Hooper, 2001; Fender, 2003; Seymour, Aro, & Erskine, 2003; Ellis et al., 2004; Kim, Davis, Burnham, & Luksaneeyanawin, 2004). For example, in an experimental study (Ellis et al., 2004), Japanese elementary school children rarely made errors in reading the target words in the transparent script of Hiragana but were not so successful with the opaque script of Kanji. Similarly, Greek children were more successful at reading the words due to the transparent script of Greek with highly regular grapheme-to-phoneme mappings than English-speaking children in the same grades whose orthography is far less transparent than the Greek writing system. Ellis et al. concluded that “it is much harder to learn to read aloud in orthographically opaque scripts” (p.455) and that “self-teaching might be more difficult in orthographically opaque scripts than in transparent ones” (p.456). Kim et al. (2004) examined the visual sensitivity of Thai readers and Korean Hangul readers. Thai is written in 72 alphabetic symbols of some visual complexity (e.g., many similar looking symbols), yet has high regularity between symbol and sound. Hangul, on the other hand, is a script of 24 alphabetic symbols that has a clear one-to-one match with Korean phonology and presents phonologically similar phonemes, such as /n/, /d/, and /t/, with visually similar graphemes. They found that the nature of Thai orthography demands higher visual sensitivity in processing words and that the lack of such sensitivity negatively affects readers of Thai, but not Hangul readers. In a similar study, Kim and Davis (2004) also found that visual processing problems did not result in poor reading in Hangul because the Korean orthography is visually transparent. Drawing from these findings, one can speculate that L2 vocabulary is more easily learned in a transparent script (e.g., Japanese Hiragana and Korean Hangul) than in an opaque script (e.g., Japa- nese Kanji and the English alphabet) and that it is more likely for L2 learners to become self-sufficient in decoding new words in transparent orthographies early in the learning process. Once Hiragana is introduced to JFL learners, reverting to the Roman alphabet does not appear to bring pedagogical merit to their beginning-level vocabulary learning.

CALL Strategies In the second set of analyses, the frequency of using the sound button to access audio recordings was strongly correlated with higher scores in both the sound and script recognition posttests. Hegelheimer and Tower (2004) also found a close relationship between achievement and repetitive use of L2 audio. On the other hand, the strategy of spending more time in the CALL program was found to only marginally influence L2 vocabulary retention. In other words, time on task had little impact on the number of words recalled. This was also similar to what Hegelheimer and Tower (2004) discovered. The students who frequently used the dialogue recordings stayed in the program for a longer period of time mainly because the auditory presentation of each dialogue took longer than that of a single word. Thus, although those who listened to the dialogues took longer to complete the CALL lessons, the length of CALL learning time per se was not a strong Yoshiko Okuyama 373 predictor of vocabulary learning. The other tools available in the CALL software, such as the kana tutor and grammar review, were rarely utilized by the students (the mean number of mouse clicks on the kana button was 1.9 and on the grammar button 2.7) and had no impact on their vocabulary learning. Why is accessing L2 sound so important in retaining newly learned words in short-term memory? Baddeley (2000) distinguished between two domains of working memory: verbal and visual-spatial. It has been widely accepted that word recognition is primarily a phonological process in Hiragana, Hangul, or Pinyin (Simpson & Kang, 1994; Kawakami et al., 2001; Kim & Davis, 2002; Chen et al., 2002; Ellis et al., 2004) but more of a visual-spatial process in Chinese (Tavas- soli, 1999; Sugishita & Omura, 2001; Ho, Chan, Tsang, & Lee, 2002; Chen et al., 2002; Flaherty, 2003). For example, a neurolinguistic study (Chen et al., 2002) provided fMRI images of different brain activities involved in native speakers’ processing Chinese characters and its phonetic counterpart, Pinyin. If Hiragana also requires a high degree of phonological processing in reading words, English-speaking JFL beginners need not struggle as much due to their prior experience with another phonological script, the English alphabet. Further- more, the symbol-sound mapping of the Hiragana syllabary is much more consistent than that of English orthography. Thus, adding another phonetic script of high regularity, Romaji, to the CALL program was probably redundant and did not enable the students to store any more words than their short-term memory could hold. From a pedagogical point of view, it might be more beneficial for JFL beginners to develop the solid knowledge of all the Hiragana symbols and adequate word recognition skills based on that knowledge. If Hiragana itself was a sufficient device for the students to learn new words on their own in this brief CALL instruction, then mastering the syllabary within the first year of Japanese learning should be of high priority. If the use of a phonetic notation device enhances the learning of Kanji (a logograph that demands more visual-spatial processing), then having Hiragana might be sine qua non for learning more L2 words via Kanji characters in the following years. The current study demonstrated the important role of L2 sound in CALL vocabulary learning. The intensive use of audio recordings was linked to better lexical learning not only in the recognition of L2 sound form (Posttest 1) but also in the recognition of L2 orthographic form (Posttest 2). If the use of audio recordings has a substantial impact on L2 vocabulary retention, focus on L2 phonological form needs to be recognized as an important CALL learning strategy. CALL users should be encouraged to access the software’s audio input frequently if the goal of an L2 task is to learn new words. Whether students prefer Romaji assistance or not, as long as they utilize L2 audio intensively, they are likely to learn new vocabulary more effectively in CALL-based materials.

Limitations of the Study The implications of this study should be viewed in the light of some limitations. First, the implication that introductory-level students benefit not from reading in 374 CALICO Journal, Vol. 24, No. 2 Romaji but from listening to L2 sound applies only to the short-term learning of L2 vocabulary. Second, the study focused on students’ receptive knowledge of L2 vocabulary acquired in a CALL environment and did not address how students could develop productive knowledge of L2 vocabulary. Third, the main purpose of this study was to examine the effects of Romaji assistance for JFL beginners in CALL. The students represented in the study were first-semester Japanese students who had become acquainted with the hiragana syllabary within the first few weeks of instruction. Thus, the findings should not be interpreted as pedagogical implications for those who have no familiarity with Japanese orthography. Last of all, because of the correlational nature of much of the analysis in the study, the results of the study do not offer evidence in support of a cause-effect relationship between any variables.

Suggestions for Future Studies Milton and Meara (1995) estimated that advanced learners of English as a Second Language (ESL) possibly acquire about 2,500 words per year. If this estimate is valid for any foreign language, successful vocabulary learning represents a very important pedagogical agenda. We need to continue our investigation on how L2 words are learned through different tasks, at different levels, as well as for different effects (e.g., short-term vs. long-term retention). If L2 word recognition is critical in developing L2 reading ability at later stages, research on Japanese word recognition will help provide insight for L2 reading experts. Furthermore, because the distance between L1 and L2 orthographic forms can burden L2 vocabulary learning, more studies need to be conducted on lexical acquisition in a nonalphabetic language. Although the results of this study did not show a measurable outcome for the use of Romaji, one can investigate long-term effects of using Romaji by conducting subsequent experiments and following the same students over a semester or a year. The long-term effects of Romaji need to be investigated because many JFL textbooks and CALL materials still use Romaji, and many JFL students continue to use it as a quick and easy note-taking device even after their learning materials have completely shifted to the authentic orthography. Individual differences in English speakers’ predisposition for Hiragana symbol learning (e.g., visual memory capacity) also need to be examined in a CALL environment. Due to the rapid growth of distance learning courses in American higher education, more and more language courses are offered online. However, the display of non-English fonts continues to pose challenges to course designers/instructors as well as students. For instance, WebCT, one of the most popularly adopted instructional delivery systems in the US, cannot be easily applied to develop a Japanese language course (e.g., the Java-supported chat of WebCT 4.1 does not allow users to type in Japanese, resulting in some confusion and typos derived from Romaji input). Continuing research on the interaction between learner outcomes and L2 script use will help bring about a better understanding of how students can effectively use CALL instructional materials in Asian languages. Yoshiko Okuyama 375 CONCLUSION The primary goal of this study was to explore the role of L2 orthography in computer-assisted Japanese vocabulary learning. What this study revealed was strong evidence for the advantage of using, not Romaji assistance, but rather L2 audio recordings. The insights from the study are perhaps most applicable to future development of CALL software or for CALL vocabulary research. Nowadays, there are many commercially available CALL programs for JFL learners either to supplement their classroom learning or to learn the language on their own. With clear guidelines for language software use, teachers and students of foreign languages are better able to make pedagogically wise decisions. For the adequate incorporation of CALL materials into a classroom curriculum, SLA researchers need to explore empirically what works best in virtual learning environments.

NOTES

1 The Yookoso! textbook is designed to cover all the four language skills: listening, speaking, reading, and writing. Although the book presented all the regular chapters primarily in Japanese orthography, the preliminary chapter is written all in Romaji as a transitional phase. 2 The Learning Company kindly provided a sample copy of the original Learn to Speak Japanese software and granted permission for its adaptation and use in this study. 3 The HyperCard program was used to program all the three modules. Each module was created as a stack of cards, and the three stacks were linked in sequence for easy navigation.

REFERENCES Akamatsu, N. (1999). The effects of first language orthographic features on word recognition processing in English as a second language. Reading and Writing: An Inter- disciplinary Journal, 11 (3), 381-403. Akiyama, N., & Akiyama, C. (1995). Master the basics: Japanese. Hauppauge, NY: Bar- rons Educational Series. Al-Seghayer, K. (2001). The effect of multimedia annotation modes on L2 vocabulary acquisition: A comparative study. Language Learning & Technology, 5 (1), 202- 232. Retrieved October 30, 2006, from http://llt.msu.edu/vol5num1/AlSeghayer/ default.html Ashworth, D. (1996). Hypermedia and CALL. In M. Pennington (Ed.), The power of CALL (pp. 79-94). Houston, TX: Athelstan. Ariew, R. (1991). Effective strategies for implementing language training technologies. Applied Language Learning, 2 (2), 31-44. Baddeley, A. D. (2000). The episodic buffer: A new component of working memory? Trends in Cognitive Sciences, 4 (4), 417-423. 376 CALICO Journal, Vol. 24, No. 2 Chapelle, C., & Jamieson, J. (1991). Internal and external validity issues in research on CALL effectiveness. In P. Dunkel (Ed.), Computer-assisted language learning and testing: Research issues and practice. (pp. 38-59). New York: Newbury House. Chen, Y., Fu, S., Iversen, S. D., Smith, S. M., & Matthews, P. M. (2002). Testing for dual brain processing routes in reading: A direct contrast of Chinese characters and Pinyin reading using fMRI. Journal of Cognitive Neuroscience, 14 (7), 1088- 1098. Chung, K. K. H. (2002). Effective use of Hanyu Pinyin and English translations as extra stimulus prompts on learning of Chinese characters. Educational Psychology, 22 (2), 149-164. Chung, K. K. H. (2003). Effects of Pinyin and first language words in learning of Chinese characters as a second language. Journal of Behavioral Education, 12 (3), 207- 223. Collentine, J. (2000). Insights into the construction of grammatical knowledge provided by user-behavior tracking technologies. Language Learning & Technology, 3 (2), 44-57. Retrieved October 30, 2006, from http://llt.msu.edu/vol3num2/col lentine/index.html Corda, A., & Van Der Stel, M. (2004). Web-based CALL for Arabic: Constraints and challenges. CALICO Journal, 21 (3), 485-495. Dewey, D. P. (2004). Connections between teacher and student attitudes regarding script choice in first-year Japanese language classrooms. Foreign Language Annals, 37 (4), 567-579. Ehri, L. C. (1999). Phrases of development in learning to read words. In J. Oakhill & R. Beard (Eds.), Reading development and teaching of reading: A psychological perspective (pp.79-108). Oxford: Blackwell. Ellis, R. (1995). Modified oral input and the acquisition of word. Applied Linguistics, 16 (4), 409-41. Ellis, N. C., & Hooper, A. M. (2001). It is easier to learn to read in Welsh than in English: Effects of orthographic transparency demonstrated using frequency-matched cross-linguistic reading tests. Applied Psycholinguistics, 22 (4), 571-599. Ellis, N. C., Natsume, M., Stavropoulou, K., Hoxhallari, L., Van Daal, V. H. P., Polyzoe, N., Tsipa, M., & Petalas, M. (2004). The effect of orthographic depth on learning to read alphabetic, syllabic, and logographic scripts. Reading Research Quar- terly, 39 (4), 438-468. Fender, M. (2003). English word recognition and word integration skills of native Ara- bic- and Japanese-speaking learners of English as a second language. Applied Psycholinguistics, 24 (2), 289-315. Flaherty, M. (2003). Sign language and Chinese characters on visual-spatial memory: A literature review. Perceptual and Motor Skills, 97 (4), 797-802. Grace, A. C. (1995). Beginning CALL vocabulary learning: Translation in context. Unpub- lished doctoral dissertation, The University of Arizona, Tucson, Arizona. Hannas, W. C. (1997). Asia’s orthographic dilemma. Honolulu, HI: University of Hawaii Press. Yoshiko Okuyama 377 Hatasa, Y. A. (2002). The effects of differential timing in the introduction of Japanese syllabaries on early second language development in Japanese. The Modern Lan- guage Journal, 86 (3), 349-367. Hegelheimer, V., & Tower, D. (2004). Using CALL in the classroom: Analyzing student interactions in an authentic classroom. System, 32 (2), 185-205. Ho, C. S.-H., Chan, D. W.-O., Tsang, S.-M., & Lee, S.-H. (2002). The cognitive profile and multiple-deficit hypothesis in Chinese developmental dyslexia. Developmental Psychology, 38 (4), 543-553. Hwu, F. (2003). Learners’ behaviors in computer-based input activities elicited through tracking technologies. Computer Assisted Language Learning, 16 (1), 5-29. Johnston, B., & Janus, L. (2003). Teacher professional development for the less commonly taught languages. Minnesota University, Minneapolis. Center for Advanced Re- search on Language Acquisition. (ERIC Document Reproduction Service No. ED479299) Kato, H., & Takada, N. (1994). Just listen ‘n learn Japanese: Beginning through intermedi- ate. New York: McGraw-Hill/Contemporary Books. Katz, L., & Frost., R. (1992). Reading in different orthographies: The orthographic depth hypothesis. In R. Frost & L. Katz (Eds.), Orthography, phonology, morphology, and meaning (pp. 67-84). Amsterdam: New Holland. Kawakami, A., Hatta, T., & Kogure, T. (2001). Differential cognitive processing of Kanji and Kana words: Do orthographic and semantic codes function in parallel in word matching tasks? Perceptual and Motor Skills, 93 (4), 719-726. Kershul, K. (1992). Japanese in 10 minutes a day. Seattle, WA: Bilingual Books Inc. Kim, J., Davis, C., Burnham, D., & Luksaneeyanawin, S. (2004). The effect of script on poor readers’ sensitivity to dynamic visual stimuli. Brain and Language, 91 (1), 326-335. Kim, J., & Davis, C. (2002). Using Korean to investigate phonological priming effects without the influence of orthography. Language & Cognitive Processes, 17 (6), 569-592. Koda, K. (1997). Orthographic knowledge in L2 lexical processing: A cross-linguistic perspective. In J. Coady & T. Huckin (Eds.), Second language vocabulary acquisition: A rationale for pedagogy (pp. 35-52). New York: Cambridge University Press. Kudo, Y. (1999). L2 Vocabulary Learning Strategies. NFLRC NetWork #14, 1-46. Learn to speak Japanese [Computer software]. (1994). Knoxville, TN: HyperGlot Soft- ware Company. Laufer, B. (1997). What’s in a word that makes it hard or easy: some intralexical factors that affect the learning of words. In N. Schmitt & M. McCarthy (Eds.), Vocabu- lary description, acquisition and pedagogy (pp.140-155). New York: Cambridge University Press. Laufer, B. & Hill, M. (2000). What lexical information do L2 learners select in a CALL dictionary and how does it affect word retention? Language Learning & Technol- ogy, 3 (2), 58-76. Retrieved October 30, 2006, from http://llt.msu.edu/vol3num2/ laufer-hill/index.html 378 CALICO Journal, Vol. 24, No. 2 Liou, H. (2000). Assessing learner strategies using computers: New insights and limitations. Computer Assisted Language Learning, 13 (1), 65-78. Matsunaga, S. (1995). The role of phonological coding in reading kanji: A research report and some pedagogical implications (Technical Report #6). Honolulu, HI: Uni- versity of Hawaii Second Language Teaching and Curriculum Center. Matsunaga, S. (2001). Subvocalization in reading kanji: Can Japanese texts be compre- hended without it? In H. Nara (Ed.), Advances in Japanese language pedagogy (pp. 30-46). Columbus, OH: The National East Asian Languages Resource Cen- ter. Makino, S., Hatasa, Y. A., Hatasa, K. (1998). Nakama 1: Japanese communication, culture, context. New York: Houghton Mifflin Company. McMeniman., M. (1997). Introduction. Setting the Australian context: The teaching and learning of languages and technology. In M. McMeniman & N. Viviano (Eds.), The role of technology in the learning of Asian languages: A report on the Griffith University National Priority Reserve Fund Project (pp. 5-10). Melbourne, Aus- tralia: Language Australia. Milton, J., & Meara, P. M. (1995). How periods abroad affect vocabulary growth in a foreign language. Issues in Teaching Languages, 107-108, 17-34. Muljani, D., Koda, K., & Moates, D. (1998). The development of word recognition in a second language. Applied Psycholinguistics, 19 (1), 99-113. Nagata, N. (2004). Robo-Sensei: Personal Japanese tutor [Computer software]. Boston: Cheng & Tsui Company. National Standards in Foreign Language Education Project (1999). Standards for foreign language learning in the 21st century (2nd ed.). Lawrence, KS: Allen Press. Nation, I. S. P. (2001). Learning vocabulary in another language. New York: Cambridge University Press. NihongoWare 1: An interactive approach to learning Business Japanese [computer software]. (1992). Webster, NY: Ariadne Language Link Co. Okamura, Y. (1995). Students’ and nonteachers’ perception of elementary learners’ spoken Japanese. The Modern Language Journal, 79 (1), 29-40. Seymour, P. H. K., Aro, M., & Erskine, J. M., (2003). Foundation literacy acquisition in European orthographies. British Journal of Psychology, 94 (2), 143-174. Simpson, G. B., & Kang, H. (1994). The flexible use of phonological information in word recognition in Korean. Journal of Memory and Language, 33 (3), 319-331. Singleton, D. (1999). Exploring the second language mental lexicon. New York: Cam- bridge University Press. Sugishita, M., & Omura, K. (2001). Learning Chinese characters may improve visual recall. Perceptual and Motor Skills, 93 (3), 579-594. Sumiyoshi, C., Sumiyoshi, T., Matsui, M., Nohara, S., Yamashita, I., Kurachi, M., & Niwa, S. (2004). Effect of orthography on the verbal fluency performance in schizophrenia: Examination using Japanese patients. Schizophrenia Research, 69 (1), 15-22. Yoshiko Okuyama 379 Tan, L. H., Spinks, J. A., Feng, C., Siok, W. T., Perfetti, C. A., Xiong, J., Fox, P. T., & Gao, J. (2003). Neural systems of second language reading are shaped by native language. Human Brain Mapping, 18 (1), 158-166. Tavassoli, N. T. (1999). Temporal and associative memory in Chinese and English. Journal of Consumer Research, 26 (2), 170-187. Tohsaku, Y. (1994). Yookoso!: An invitation to Japanese. New York: McGraw-Hill. TriplePlay Plus! Japanese [computer software]. (1994-1997). Syracuse, NY: Syracuse Language Systems. Vincent, E., & Hah, M. (1996). Strategies employed by users of a Japanese computer assisted language learning (CALL) program. Australian Journal of Educational Technology, 12 (10), 25-34. Wong, M., Perfetti, C. A., & Liu, Y. (2003). Alphabetic readers quickly acquire orthographic structure in learning to read Chinese. Scientific Studies of Reading, 7 (2), 183-208.

AUTHOR’S BIODATA Yoshiko Okuyama is Assistant Professor in the Department of Languages at the University of Hawaii at Hilo (UHH). While working at UHH, she completed her dissertation and earned her Ph.D. from the University of Arizona in 2000. She currently teaches courses in Japanese, introductory linguistics, psycholinguistics, and second language acquisition theory and has served as UHH Language Lab Coordinator.

AUTHOR’S ADDRESS Yoshiko Okuyama, Ph.D. Department of Languages The University of Hawaii at Hilo PO Box 6917 Hilo, HI 96720 Phone: 808 982 9871 Fax: 808 974 7736 (Attn: Yoshiko Okuyama) E-mail: [email protected] 380 CALICO Journal, Vol. 24, No. 2