Linguistic Innovations and Interactional Features of Casual Online Communication in Japanese

Yukiko Nishimura
Toyo Gakuen University, Chiba, Japan

Editors' Note: If you have difficulty viewing the fonts in this article, you may wish to download and consult a PDF version of the paper which displays the fonts correctly. This PDF document is for reference only. For citation purposes the html document is the official version.

Abstract

This study explores the linguistic and interactional properties of informal asynchronous computer-mediated communication (CMC) in Japanese. Using Bulletin Board Systems (BBS) messages as the primary source of data, the study identifies innovative uses of kanji, other scripts and punctuation, and examines the incorporation of such informal spoken features as final particles. Young Japanese BBS users are found to employ colloquial language online as if conversing offline, and interact appropriately with their fellow participants in their Internet community.

Introduction

At 69.4 million, the number of Internet users in Japan ranks second only to the United States (165.75 million users).¹ Moreover, Japanese speakers constitute 9.7% of the total of 619 million Internet users worldwide, ranking third after English (36.5%) and Chinese (10.9 %).² Not much is known, however, about how Japanese speakers use language online, in sharp contrast with the fairly large amount of research done on English speakers' CMC (Herring, 1996, 2001). These Japanese Internet users, including cell phone users who have access to the Internet, certainly read and post messages and communicate in the Japanese language online, but this has not been well documented so far. This article attempts to describe Japanese speakers' Internet communication, and more specifically, analyzes linguistic and interactional features of informal computer-mediated communication (CMC) in Japanese, using as data messages posted asynchronously on message boards on websites.

Previous studies in English on Internet communication have mainly discussed aspects of CMC that utilize Roman-based scripts; little research has been done on CMC in languages that do not utilize Roman-based characters, such as Japanese. Matsuda (2001) analyses Japanese Web diaries, focusing on the construction of "voice," and discusses one of the four Japanese scripts, katakana, as contributing to "voice" construction, and Matsuda (2002) examines the negotiation of identity and power in e-mail messages written by Japanese teachers of English at colleges and high schools, including analyses of honorific uses. Katsuno and Yano (2002) examine the face marks (kaomoji, the Japanese equivalent of smiley emoticons) used in e-mail and cell phone messages. Sugimoto and Levin (2000) compare how American and Japanese users identify themselves and use emoticons in e-mail messages sent to discussion groups. In addition, there are also some academic as well as semi-academic works in Japanese on CMC: Satake (1995, 2002) gives an account of the styles used by contemporary Japanese youth in general as well as in CMC. Such semi-academic journals as Nihongogaku [Japanese Linguistics] (1993, 1996, 2000, and 2001), and Gendai no Esupuri [Esprit of Today] (1998, 2001) have produced special issues on communication via computers and cell phones. Tanaka (2001) also presents an up-to-date analysis of users' behaviors in exchanging e-mail messages by computer versus cell phone. In Nishimura (to appear), I explain innovative and playful uses of kanji and other scripts on a Japanese BBS website called ni channeru ( 'Channel 2'), where users develop highly specialized, unconventional language use and form a community of practice on the Internet.

Except for the studies by Matsuda (2001), who discusses katakana, and Nishimura (to appear), these existing studies of Japanese CMC, both in English and in Japanese, overlook one of the key elements that characterize Japanese online communication: the vast variety of scripts that the Japanese speakers use innovatively. Through the variety of orthographic choices available to Japanese users, they create rich online discourse that suits their communicative and interactional purposes. Compared with English CMC users, they use creative (and sometimes amusing) orthography and punctuation in similar as well as distinctive ways, supported by Japanese CMC technology. A variety of features in informal spoken Japanese such as sentence-final particles reinforce this tendency. In this way, Japanese Internet users have a high degree of freedom in CMC, and they exploit it to recreate or construct "informal friendly talk" online, which is appropriate to the cultural expectations of their communities.

This study is an attempt to explain how the Japanese orthographic system enriches Japanese CMC in linguistic, interactional and socio-cultural contexts. I begin by explaining the orthographic and punctuation conventions used in written Japanese. Because in CMC the messages are conveyed not by auditory channels but by visual means through keyboarding, some background information on written conventions is necessary. The Japanese writing system is dramatically different from that of Western languages (Suzuki, 1975, 1977; Backhouse, 1984), and thus it is necessary to clarify how each script functions in standard Japanese orthography, in order to understand Japanese CMC.

After identifying briefly the data on which this study is based, I compare Japanese CMC with its close equivalents in English, in order to clarify what aspects of orthographic means in CMC are shared by users of both languages and what other features are distinctive. I mainly use Danet's (2000, p. 17) "common features of digital writing" as a frame of reference, supplemented by studies by several other researchers, in order to enrich the comparison of the specific means of compensating for and adapting to the limitations of CMC in the two languages.

A taxonomy based on English CMC alone is of limited use in accounting for the rich variety of Japanese CMC. Various features specific to informal spoken Japanese are found in Japanese CMC. After considering briefly how informal spoken interactions take place, I examine the way such informal features as sentence-final particles are transferred to Japanese CMC. Finally, by examining closely what Japanese CMC users do, I suggest that users' online interaction is a reflection of communicative styles that occur in casual face-to-face conversation, even though BBS communication is asynchronous. Because the present study is based on a limited set of data, I do not claim that my results necessarily apply to Japanese CMC in general. Rather, they reflect the specific kinds of BBS interactions studied.

Written Japanese: Some Background

Anyone who wishes to study Japanese as foreign language will face a number of obstacles, such as a wide gap between spoken and written styles, a lengthy list of kanji characters to remember in reading and writing, and complicated systems of honorific uses, which are integrated into the grammar (Clancy, 1982). Kanji and honorifics even trouble native Japanese speakers. I begin by describing how Japanese speakers write their language typically using four different scripts, namely, hiragana, katakana, romaji, and kanji.³

Four Scripts in Japanese Orthography

The four scripts used in standard Japanese orthography have the following characteristics. Both hiragana and katakana are kana or syllabary scripts, which use written symbols to represent the sounds of each syllable. Each hiragana symbol (e.g., か for ka) has its corresponding katakana (e.g., カ for ka); though the syllable is pronounced the same, its shape is different in the two syllabaries. This is because hiragana and katakana were created in different ways. Hiragana was originally a simplified form of kanji (Chinese characters) in a cursive form, while some of the katakana symbols were formed by taking some elements of a kanji character (e.g., the radical of カ from kanji, 加 for ka) in an angular form.

Another difference between the two syllabaries lies in their respective functions: hiragana is used to represent inflectional endings such as the past and passive. It is also used to write the postpositional particles that show grammatical relations such as the subject and the object of a sentence. It can also be used to write sentence-final particles, which indicate the speaker/writer's attitude toward the content or the speech situation (this will be discussed in greater detail later). Another use of hiragana is to represent native Japanese ideas and objects for which kanji characters do not exist. Katakana, on the other hand, is used mainly for writing foreign names, loan words,⁴ certain adverbs imitating natural sounds (e.g., ワンワン wan wan 'bow wow' for dog barking ), and so on. Stanlaw (2002) regards katakana as a kind of italic script, used to represent words that are "unique and special" (p. 549), including loan words and onomatopoeic expressions.

Because kana syllabary indicates the sound, it is used to supplement how kanji (see below) is intended to be read, in the form of furigana, which are kana symbols printed or written in smaller font size at the side or on top of kanji characters. When filling out governmental forms, for example, it is customarily requested that important items like names have phonetic transcriptions in kana written at the side. Also kanji supplied with furigana is often seen in books for children.⁵

Romaji is the use of the Roman alphabet. One major function of romaji is to transliterate personal and place names and other Japanese words. There are three systems of Romanization: the Kunrei, Japanese, and Hepburn. In this article, Japanese transliteration is given in slightly modified Hepburn system, which is a closer representation of phonetic orthography than the other two systems.

The use of the Roman alphabet is not limited to transliteration: it is used as eimoji, English words in Japanese writing, especially in the form of acronyms. For example, CD for compact disk, VTR for video tape recorder, and so on, appear abundantly, even in vertical writing, in Japanese newspapers and the like. Though these are originally English words, they are nativized and felt to be part of Japanese vocabulary. Some of them are often used in combination with other Japanese words, such as “OA機器” ooei kiki ‘o(ffice) a(utomation) machinery and tools’ in which オフィス ofisu‘office’and オートメーション ootomeesyon‘automation’are already part of the Japanese vocabulary of loan words. The preference for English acronyms over katakana representation in Japanese writing, it seems, is due to linguistic economy.

Kanji characters were originally borrowed from Chinese, and each character represents a particular meaning, for example, the native word for 'river' is normally expressed by the character 川. However, this character can be pronounced in at least two completely different ways: the Japanese reading is kawa, and the other reading, modified Chinese reading, is sen. The modified Chinese reading is normally applied to Chinese compound words, e.g., 河　ka + 川　sen = 河川 kasen 'rivers'; When the character 川 is used in isolation, it is normal practice to assign the Japanese pronunciations, which have a single unambiguous meaning. The Chinese readings, in contrast, have many homophones; for example, sen can have a number of different meanings in more than 50 words of Chinese origin, e.g., 千 'thousand', 線 'line', 戦 'battle', 船 'ship' 選 'selection', and so on, which are all pronounced sen.

In writing verbs, Japanese orthography typically uses a combination of a kanji character for the verb stem and hiragana for the inflectional ending. For example, in the Japanese verb warau 'to laugh', the stem, wara is written with the kanji, 笑, which means 'laugh' and the root-form ending, -u is given in hiragana, う. To form the past tense waratta 笑った, the past ending -tta, った, is expressed by hiragana of two different sizes. The smaller symbol, っ, is used to indicate that the consonant is long, and the normal-size symbol, た, represents the past tense ta. Normally the long consonant is phonetically realized by holding the consonant for a moment.

As for the smaller symbols, there are four that are conventionally used, as follows (their katakana equivalents are in parentheses). Examples of how each symbol (bolded and underlined here for clarity) is used in an English loan word are given in (1) through (4) below:

(1) っ (ッ) tsu　as in トップ toppu in transliteration, which means 'top'
(2) ゃ (ャ) ya as in チャンス chyansu 'chance'
(3) ゅ (ュ) yu as in ミュージック myuuzikku 'music'
(4) ょ (ョ) yo as in ショップ shoppu 'shop'

Recent innovations in katakana representation that are widely accepted include small ィ i as in パーティ or paatei 'party.' The smaller symbol innovations are mainly found in loan words in katakana representation, as described by Stanlaw (2002). Japanese online users, however, use these smaller symbols in hiragana innovatively in native words to represent a certain prosodic quality, as we shall see later.

Smith and Schmidt (1996) identify eimoji and kigou in addition to the four scripts in Japanese. Eimoji refers to English words, and kigou are symbols like ☆ and ♪. These two additional types appear frequently in comic books (manga) and magazines for young readers, and advertisements and signboards of various kinds. Japanese CMC messages also have these additional words and symbols. They will be discussed later in this article.

Punctuation Marks and Other Notational Symbols in Written Japanese

In addition to orthography, it is important to understand the standard punctuation system of written Japanese. Japanese standard punctuation marks that have English equivalents are as follows:

the period 。
the comma 、
the question mark ？
the exclamation mark ！
single quotation marks 「　」
double quotation marks 『』⁶
parentheses （　）
the dash ―
brackets ［　］
angle brackets 〈　〉
ellipsis dots ・・・

Though mentioned briefly earlier, it should be noted here that both horizontal and vertical writing are possible.⁷

In addition to these shared punctuation marks, there are unshared marks and symbols.⁸ Below are those used only in Japanese:

emphasis dots or commas, such as in vertical writing and horizontally kirei 'pretty' (Not available in text-only online messages)
middle dots, ・, often used to separate items in a list, such as 月・水・金　getsu・sui・kin　'Mon, Wed, Fri',
the repetition symbol to form a plural, 々, used to represent the same kanji that precedes the symbol, as in 山々 yamayama 'mountains' (cf. 山 yama 'mountain')
lengthening symbol,ー, mainly used in katakana to represent loan words that have long vowels, such as データ, deeta, 'data'

In addition to these shared and unshared marks and symbols, Japanese online users creatively use other symbols and signs that are provided by Japanese word processing software.

Word Processing in Japanese

Unlike handwritten messages, CMC messages are created through keyboarding. To write in Japanese on the computer, users must use so-called romaji-kana-kanji conversion software like Microsoft IME or Justsytem's ATOK, and follow these steps.⁹ An image of a Japanese keyboard is provided below.

Figure 1. Image of the Japanese keyboard

(a) First, the user enters Japanese words, based on their pronunciation, in romaji in the Japanese input mode.¹⁰

(b) Second, the conversion software automatically changes the romaji into the hiragana syllabary, if the combination of romaji letters have corresponding hiragana. (If a single consonant letter, such as "t" or "s", is entered, it is not changed to hiragana, because most of the symbols in the syllabary represent one syllable consisting of a consonant and a vowel, か for "ka"). When the desired word is in hiragana, the user presses "enter" to finalize.

(c) In order to convert hirakana to katakana, the user can press the "katakana key" and then presses "enter" to finalize. Or katakana can be chosen in step (e) below, where the list of choices includes katakana.

(d) In order to use kanji, the user presses the "conversion key" (which is often the space key), so that the system will automatically show his or her most recently used kanji corresponding to the hiragana.

(e) When the most recently used kanji is not the intended character, the user presses the "conversion" (or space) key again to get a list of all the other kanji characters as well as other symbols and characters that have the same pronunciation of the hiragana. The user then chooses the desired character or symbol and presses "enter" to finalize.

Figures 2 and 3 below show lists of kanji candidates that have the sound, sen. There are 58 choices in all, and all the choices cannot be shown in one frame. In Figure 2, the user has selected the character for "thousand," and the symbol chosen in Figure 3 is used in a place name in Hokkaido, meaning "stringed bead."

Figure 2. Choices for "sen"

Figure 3. Additional choices for "sen"

In creating texts through the conversion software, the script choices available to users are not restricted to those given by the romaji-kana-kanji conversion system. For example, the ATOK conversion software also provides moji paretto (a “letter palette”) from which users can select various kinds of symbols, marks and letters, such as scientific and mathematical symbols, the Greek and Russian alphabets, phonetic symbols and so on. The software also includes a system that lets users register their handmade scripts and frequently used expressions, and create a personalized dictionary. With these features, Japanese users have a wide selection of scripts and symbols and can enter even complicated symbols very quickly.

Data: Asynchronous BBS

Having described the resources available to Japanese online users, I turn now to a discussion of Japanese CMC data. The data for this study comes from messages sent asynchronously to Bulletin Board Systems (BBS) that are part of several personal websites featuring popular actors, films, music, and so on. These are typical fan sites where younger people spend a great deal of time reading and posting messages. The data were collected from July to October 2000 and in January 2001. Participants seem to be mostly women in their middle teens to thirties, based on the content of their messages, though there is no way of ascertaining the true identity of the users. One reason for using these asynchronous BBS messages is their accessibility. Messages stored on these sites can be visited and viewed by anyone interested. There is an additional, more substantial reason to examine these BBS messages, which relates to their personal nature. This point is best illustrated when these BBS messages are compared with e-mail messages that have been extensively investigated in the literature.

Both BBS and e-mail messages are exchanged without sharing log-on time, unlike Internet Relay Chat and other synchronous modes of CMC. Herring (2001) points out that synchronicity is one of the major factors that determines the nature of CMC discourse (p. 6). Thus, BBS data and e-mail messages share the common determining factor of asynchronicity.

Important differences, however, exist between the two. Firstly, e-mail messages are typically exchanged between individuals¹¹ and hence are basically private in nature, while BBS messages are sent to and posted on websites and are in principle public. That is, in BBS settings there is an audience and the messages can be read not only by users who post messages, but also by any other interested party, while in one-to-one e-mail, no one other than the recipient reads the message. Who reads the messages, and to whom they can be addressed may influence BBS message contributors and their discourse.

Secondly, there is one more aspect to be considered: whether or not messages are of a personal nature. Thus, both e-mail and BBS messages can further be divided into at least two: personal and non-personal, though this distinction is by no means clear-cut. e-mail messages of a personal nature are located in the computer files of senders and recipients, and nobody "would volunteer his or her ‘in’ and ‘out’ baskets for public scrutiny" (Baron, 2000, p. 248). In contrast, postings to BBSs and other asynchronous systems are readily accessible to researchers.

Gimenez (2000) compares features of e-mail messages in business communication with business letters. Cho (in press) examines work-related e-mails exchanged in a university setting in comparison with memoranda. Both works can be regarded as investigating e-mails of a non-personal nature, although they are private. Matsuda (2002) analyzes e-mail messages sent to a professional discussion group called TESOL Link. These e-mails are also of a non-personal nature. Other researchers, such as Danet (2001), analyze e-mails addressed to them on business-related matters. Though the distinction between personal/non-personal e-mails can be a matter of degree, it is useful to separate them from the viewpoint of this study.

The messages analyzed in this study are of a personal nature, even though they are publicly posted. Their subjects relate to personal likes and hobbies. Since personal e-mail messages are hard to obtain, there is a strong rationale for studying personal BBS messages in order to examine how "personal subject matters" affect this form of discourse.

The BBS websites studied here also contrast with others examined thus far with respect to users' purpose for posting messages. Collot and Belmore (1996, p. 16), in their analysis of BBS messages, consider users' goals to be requesting and giving information, making announcements, and engaging in discussion. The BBS users in my study, on the other hand, post messages in order to share common interests in particular actors, films, music, and so on and to interact with other users who share the same interests. This goal of entertainment and enjoyment on the part of the users should be taken into consideration when interpreting the messages in the data under investigation.

Table 1 below gives a profile of three representative BBS websites studied here. Although these websites are gone now, similar BBS websites can be easily found elsewhere.¹²

Table 1. Profile of representative BBS

These sites are each managed by a site manager. The manager, who in many cases is the site creator, sets rules and guidelines regarding conduct on the website including the use of appropriate language. The attitudes of the manager might have an influence on participants' message production, especially because the interaction takes place asynchronously, when writers have time to edit and revise their messages to meet the guidelines.

All participants including the site manager use nicknames called handoru neemu ('handle names') when posting their messages. When a new participant appears at the site, the site manager normally asks for a brief self-introduction. After this self-introduction, the newcomer is treated as a member.

The site manager usually responds every time someone posts a message. Thus, we see from Table 1 that the site manager posts messages far more frequently than the other users, and the other participants usually post once or twice, except for the participant who posted 7 times to Idol BBS. On the average, the messages are long and elaborated,¹³ as we shall see later.

Comparison between English and Japanese CMC

In order to clarify what aspects of Japanese CMC are distinctive, I compare messages written by Japanese CMC users with their English counterparts. It would have been best if the comparison had been based on English BBS messages posted on fan sites for entertainment purposes. However, this is not practical since, to my knowledge, there is no study of such sites. Thus, English messages in asynchronous exchanges such as e-mail offer a second best comparison with Japanese CMC messages with asynchronous characteristics.

Since the asynchronous BBS messages under study are digital discourse, Danet's (2001) general characterization of digital discourse provides a useful frame of reference. She identifies nine common features (Figure 1.5, p.17), which are shown in Table 2. To widen the scope of my study further, I supplement Danet's characterization with additional features studied by several other researchers (Cho, in press; Gimenez, 2000; Lan, 2000; Lee, 1996, Werry, 1996), all of whom studied e-mails except for Werry (1996), who analyzes Internet Relay Chat (IRC), and who provides a detailed description of spoken features characterizing chat . As Danet (2001, p. 14) cautions, differences between synchronous IRC and asynchronous BBS should be kept in mind. Thus, I incorporate from Werry's data only those features that are not directly related to the issue of synchronicity.

In Table 2 below, I juxtapose features of Japanese BBS discourse with their English counterparts, mainly based on Danet's classification of nine common features.

Table 2(1). Common features of digital writing based on Danet (2001)

Table 2(2). Common features of digital writing based on Danet (2001)

When we look at Table 2, we notice that most features presented by Danet are shared across the two languages, though the correspondences are not always exact. The features from 1 through 7 are those to compensate for information lacking in this medium, such as paralinguistic and prosodic cues, and the last two are features to speed up typing (Danet, 2001, p. 17). Both English and Japanese CMC users employ similar strategies such as reduplicating letters and symbols for indicating lengthened elements in speech and so on. Speakers of both languages also employ unconventional orthography to represent as accurately as possible the actual sound they wish to simulate, because conventional orthography in both languages does not convey the fine details of actual speech production. Let us now look at each category one by one.

Multiple Punctuation

First, let us consider multiple punctuation, which is shared by the two languages. Punctuation marks that come at the end of a sentence, such as periods, exclamation points, and question marks are multiply used by the users of the two languages. One difference, however, is Japanese users' employment of wavy lines, which are not conventionally used symbols in Japanese writing. The conventional symbol for lengthening is a straight line (see the section "punctuation marks and other notation symbols"). The use of wavy lines reminds viewers of an undulatingly long span of time. The writer uses multiple wavy lines to emphasize visually how long the waiting time is. This practice is reminiscent of the typographic simulation of smoking marihuana on IRC analyzed by Danet and her colleagues (Danet, 2001, chap. 3; Danet et al., 1997, 1998). In the sequence of online interaction analyzed in that study, the main participant typed decreasing numbers of the letter “s,” row by row, to simulate the dissipating smoke.

Eccentric Spelling

Attempts to reproduce actual articulation in typed messages are the major causes for these unconventional spellings. For prolonged utterances letters for the vowel are reduplicated, and words are spelled in what is called "pronunciation spelling" in English, such as "sez" for "says."

The Japanese examples in row 2 of Table 2(1), which use unconventional orthography to indicate colloquial style, need more explanation, since they show the highly expressive nature of some hiragana. As explained above, the smaller-sized っ is conventionally used to represent a long consonant. The inclusion of this symbol in はっじめまして hajjimemasite 'first time to see you' is unconventional because the conventional orthography for this expression is はじめまして hajimemasite. The inclusion of smaller-sized っ gives the viewers a more vivid picture of how this expression might be pronounced; viewers get the impression of a clear and cheerful, high-spirited articulation. The writer's intention in adding this symbol is to convey the speakers' articulation as accurately as possible; in this way the sender can convey a high degree of affect and closeness, as if sharing the same physical space and time of conversation with the viewer. Notice also that this inclusion does not cause any difference in lexical meaning, though in some other words, the presence or absence of this symbol is contrastive in meaning, such as きって（切手）kitte which means 'stamps' versus きて（来て）kite, meaning 'please come' or the gerundive form of the verb, kuru 'come.' The difference brought by the inclusion of っ in the examples here is in the degree of emphasis and is pragmatic in nature.

Another form of unconventional writing in row 2 of Table 2 (1) is the inclusion of smaller-sized symbols, ぁ for a at the end of 終わりましたかぁ owarimashita ka a ‘was it really over’ and ぃ i in 久しぶりぃ hisashi buri i ‘long time no see’. They might look like reduplicated letters, but the reduplication takes place on the vowel, and not on the typed symbols: か ka and あ a are not the same symbols. These smaller-sized scripts signify emphasis in speech articulation, in which the phrase-final vowel is stressed. The final vowel of the emphasized phrase is pronounced somewhat longer and more strongly than the articulation without emphasis. Such inclusions of added vowels normally do not cause a difference in meaning. In conventional writing, these smaller scripts are not used, and the expressions are 終わりましたか owarimashita ka and 久しぶり hisashiburi.

While the conventional writing is ありがとう arigatou 'thank you,' ありがとー arigatoo is used here (the difference is underlined). Use of a lengthing symbol, whether straight or wavy, in place of the hiragana, う u, is often seen not only in CMC but in informal writing as well, where the long vowel [o:] is used in speech. When the gap between pronunciation and conventional writing is highly noticeable, users seem to be more tempted to capture the spoken language in expressing their message, because they feel their voice can be more directly conveyed through such unconventional orthography.

Capital Letters

There are English examples of all capital letters in row 3 of Table 2(1), while in the Japanese example, an English word, HAPPY, is given in capital letters along with Japanese symbols. This is an example of a use of eimoji English word, described above.¹⁴ Since the alphabet is not the major script in the Japanese writing system, English words are seen only sporadically. When they are used, the use itself stands out among the Japanese scripts, and has an attention-catching function. Users' choice of capital letters over lower case probably lies in the fact that capital letters are even more prominent visually. The function of capital letters in entirely alphabetic writing and in the middle of kana and kanji writing, obviously, differs. Thus, the use of capital letters in English CMC to signify shouting does not occur in Japanese CMC, though the function of emphasis seems to be present. The user could employ katakana, ハッピー　happii, following the convention of using katakana for loan words. The use of eimoji here gives viewers some sense of emphasis in this kind of feeling, though the difference between katakana and eimoji would be slight and can be a matter of taste.

Asterisks for Emphasis

I did not find Japanese examples of asterisks used for emphasis. Instead, the example in the table is given in katakana. Normally, yoroshiku is a native expression, and the convention is to write in hiragana よろしく or in kanji 宜しく. The use of kanji gives a rigid or formal feeling, and is not suitable in informal casual interaction. Hiragana would be the most unmarked way of expression. The marked way, katakana is used, because the user seems to attempt to place emphasis on this expression. The function of asterisks for emphasis and its distribution in English CMC may not be the same as that of the marked use of katakana in Japanese CMC. However, since they seem to work for the same purpose, it may be appropriate to regard katakana used in marked ways as some kind of counterpart for the asterisk in English.

Written-out Laughter

The category of written-out laughter can be extended to include other vocalizations and non-linguistic sound/noise like music. Because interactions are not face-to-face, several kinds of extralinguistic information are missing in CMC. In order to supply as much missing information as possible, users in the two languages attempt to express various non-linguistic sounds visually by keyboarding in similar ways. One difference here is the use of musical notes in Japanese examples to show explicitly that expressions belong to music or have some musical characteristics. The use of musical notes and other non-linguistic symbols appears at the end of Table 2(2), under "Multiple/Non-linguistic symbols" and will be discussed later in the section "other features".

The Japanese examples for auditory information (sound/noise) can be grouped as onomatopoetic or mimetic words, such as kyaa, uumu, chan chan, and een, which are also abundant in informal spoken conversation. The inclusion of such vocabulary in CMC evokes a higher degree of involvement and sense of participation in "conversation", because such vocabulary encodes high affect (Baba, 2001).

Description of Actions

Three different notations for "grin" are given in theEnglish examples: asterisks enclosing the verb, the verb in full spelling in angle brackets, and the initial letter alone of the verb in angle brackets. In the Japanese examples, parentheses are used to differentiate certain expressions from the main text itself. One type of action description is to give the verb stem in kanji alone enclosed by parentheses. As explained earlier, when used as a verb, kanji are not normally used in isolation; they are used with hiragana for inflectional endings, tense, and so on. Yet even when the characters are used without any accompanying kana, their meaning is clearly understood by readers with basic literacy in Japanese, because the characters are ideographs and convey meanings. Such use of isolated kanji in parentheses is typical in Japanese CMC and is very frequently found not only in BBSs, but also in informal e-mail messages (even in a two-party chat session in my limited experience of synchronous CMC), because senders can communicate their feelings, actions, and so on by means of kanji instantaneously and easily. Recipients can visually get the message at a glance as well.

Additional examples of isolated kanji in parentheses include the following:

(涙）namida “tears”
(謎）nazo “puzzlement”
(照）tere “embarrassment”
(惚) hore “being attracted”
(歓喜）kanki “joy”
(恥）haji “shame”
(爆）baku “explosion, burst”
(汗）ase “perspiration”
(死）shi “death”
(苦笑）kusyou “bittersweet laughter”
(感涙）kanrui “tears of gratitude”
(殴！）naguru “punch !”

This way of expressing feeling is rather unconventional and not very often used in other communication modes except for personal casual letters, as reported by Kataoka (1979, p. 121).

Not only kanji in parentheses but also other types of expressions indicate actions, gestures, and mental states. Parentheses are used to separate this kind of comment from the messages themselves. In row 5 of Table 2(1), the message writer explains the situation using a somewhat metaphorical expression, 'holding sweat in the palm.'

"Smiley" Icons

One major difference between smiley icons used in English CMC and those in Japanese is that the English ones are to be read sideways, while the Japanese ones are right-side up, as illustrated in row 7 of Table 2(2). Smileys rendered in Roman typographic characters consist of colons, parentheses and other typographic characters. Because the kinds of symbols available in Japanese are more varied, including non-linguistic ones, there seems to be more variety in face marks―Japanese emoticons.¹⁵ In addition to those in Table 2(2), other examples of face marks in messages are:

(5) 復活おめでと～♪良かったね（*＾▽＾*）
hukkatu omedeto～♪yokatta ne （*＾▽＾*）
'Congratulations on your comeback♪[as if singing]That was good(*＾▽＾*）"
This face mark represents the mouth wide open, laughing loudly and cheerfully, with asterisks used to indicate (rosy) cheeks.

(6) ヨロシクお願いしますっ！m（。＾＿＾。）m
yoroshiku [in katakana] onegai shimasu ? [glottal stop] emoticon for yoroshiku 'Please remember me.'
Yoroshiku is an expression very often used to keep good relations with others in Japanese conversation. This particular face mark is listed in an online kaomoji jiten 'Face Mark Dictionary' under yoroshiku, which can be visited at http://www.diana.dti.ne.jp/~s-tuzi/.

(7) 許してくださいませ　§^｡^§
yurushite kudasai mase
'Please forgive me [for any inconvenience that may arise].'
The symbol § for legal or bureaucratic “section” is used to represent hair. This face mark represents a somewhat quiet, obedient attitude.

These examples illustrate that variety in face marks is enhanced by variety in scripts used in Japanese CMC. Some of them seem to be used so frequently that they are "lexicalized" in dictionaries, many of which are available online. When message writers use them, they may create these face marks manually, or they may copy and paste already made marks. If these face marks are entered manually, it might take some time, but since they appear in asynchronous messages, users can take time creating messages with such colorful frills.

Abbreviations

Abbreviations appear at least in two different types of processes: One is acronyms like LOL for laughing out loud, and the other is rebus writing. As for acronyms, which are encoded by stringing together the initial letters of words in phrases in English, what the acronyms mean may not always be understood by users, unless they are already familiar with the language used in CMC. Though the Japanese examples for abbreviations may not be the exact counterpart of the English examples, there is certainly a process of shortening from full phrases. The whole message for the new year's greeting in Japanese is a long one, and the first two syllables from each of the four phrases that make up the whole greeting are used to form this abbreviated greeting. This expression is not limited to CMC, but is in circulation especially among young people.

Regarding rebus writing, which is the use of a single symbol to represent a word or syllable with a similar sound, the English example seems clear, while the Japanese example may not be; the process by which the sequence of numbers “4649” in the Japanese example can be rebus writing needs explanation. In Japanese the number 4 is pronounced yon or shi, 6, roku, and 9 ku or kyuu. The first number, 4, is read yo from yon, the second number, 6 as ro from roku, the third number, again 4, is this time read shi, and the final number 9 is read ku. When all are combined, we get yoroshiku. To replace certain words with numbers and the other way around are not very unusual practices in general, especially where word play is concerned. In the case of Japanese abbreviations, the replacement of words with numbers, it seems, is more motivated by word play and play with writing than by a desire to speed up typing.

Abbreviations and the next category, use of all lower case (Table 2(2)), are for the purpose of speeding up typing, and hence they matter more in synchronous CMC than in the asynchronous mode. In general, to be able to type messages in a shorter time is preferred in both modes in CMC, and there is a tendency for abbreviations to occur, perhaps more often in synchronous than asynchronous modes. Since the Japanese data are from asynchronous CMC, it is not absolutely necessary for users to type quickly. Still, similar processes seem to be operating. The examples in this category shorten a long phrase and play with words.

All Lower Case

As in the case of capital letters, since the alphabet is not used as often as other major scripts, I did not find examples typed in all lower case in the Japanese data. One possible phenomenon similar to this would be to write everything in hiragana or katakana, without taking the trouble of converting to kanji. However, there are a great many homophones in Japanese and it would be extremely difficult for readers to understand the intended meaning without the use of kanji, even with the help of context. This would thus be an unlikely counterpart to writing in all lower case. Another possible candidate for the equivalent process to all lower case would be to send misconverted kanji as it is, before or without correcting it, because all lower case writing has errors that need to be corrected, from a prescriptive perspective, and the misconverted kanji are also errors that should be corrected as well. However, misconverted kanji could cause misunderstanding, while all lower case would not, and these processes cannot be considered exactly the same. Since this last category concerns synchronous modes where typing speed matters, this topic will not be discussed further in this article.

Other Features

In the cells at the bottom of Table 2(2), I have included a few examples that are not specifically discussed by Danet. One example is from Cho (2000), in which she shows an example of multiple uses of the dollar symbol, which is very expressive, emphasizing the money visually represented by the dollar symbol. The Japanese examples include the use of non-linguistic symbols like stars and musical notes. The message accompanied by the symbols conveys encouragement; the effect of the ☆★☆ symbols is to cheer up someone in a light-hearted manner. The use of musical notes is similar. Since the musical note denotes music and singing, the message writer feels somewhat merry, as if she is going to sing. If these symbols were not used, but conventional punctuation marks were used instead, the cheerful feeling/atmosphere would be lost. Though these two kinds of examples could have been discussed under multiple punctuation, the dollar symbol and ☆ are not punctuation marks, strictly speaking.

To sum up, we have seen in Tables 2(1) and (2) that CMC users in the two languages employ similar methods to overcome technical shortcomings. They make use of various strategies to supply information that cannot otherwise be conveyed through this medium. CMC users in both languages devise various innovative and unconventional ways of writing in order to supply prosodic, paralinguistic, phonological, and auditory information with language-specific means.

The analysis presented in this section has been based on Danet's common features on English CMC. However, this classification cannot account for the rich variety of Japanese CMC. In fact, various features specific to informal spoken Japanese are also found in Japanese CMC. I have already pointed to Japanese CMC users' extensive use of colloquial vocabulary. Another important feature is expressions of familiarity and politeness. In the next section, I examine the way informal spoken features showing politeness and familiarity are transferred to Japanese CMC.

Online Japanese Discourse as "Conversation"

Features of Informal Spoken Japanese

As shown in the discussion above, computer-mediated discourse shares a number of features with informal spoken language. In order to understand what informal spoken features appear in Japanese CMC, it is useful to examine characterizations of informal face-to-face conversation in Japanese. Maynard (1989, pp. 23-44) presents a detailed description of Japanese conversational language, which is characterized by: (1) fragmentation of talk, (2) final particles, (3) fillers, (4) ellipsis, (5) postposing, (6) verb morphology (plain versus polite), (7) sentence-final forms, (8) insertion of metacommunicational remarks, (7) propositional twisting, (8) questions as conversational elicitors, and (9) rhythmic ensemble.

Among these features, the fragmentation of talk is based on pauses within utterances; since no auditory information is communicated in BBS exchanges, this characteristic will not be discussed. However, it should be pointed out that BBS users show a tendency to create fragmentary messages through innovative punctuation marks (e.g., ♪) between chunks of words and phrases, as well as the use of short messages that do not form a full sentence syntactically. Similarly, the feature of rhythmic ensemble will not be considered either, since it concerns the tempo of each syllable and non-verbal behavior like head movement (e.g., nodding), which are also not available in online interaction.

Final Particles

Among the remaining seven characterizations, I mainly focus on two areas, particles and verb morphology, or, more specifically, the polite and plain styles of verb endings. Particles can be categorized into two groups: grammatical particles and particles that express the speaker's attitude. The first type encodes the grammatical structure of the sentence, and is not of concern here. The second type can be further subcategorized into two: particles in sentence initial positions that draw interactants' attention, such as e and aa, and particles that ensure rapport, which normally appear at the end of phrases, clauses and sentences. This category of particles will be the focus in this article.

According to Maynard (1989), the particles of rapport, sometimes referred to as sentence-final particles, include ne(e), sa(a), no, ka na and yo na, yo, ka, wa, and ze. A brief summary of the functions and uses of the particles (except for the combined particles, ka na and yo na) is given in Table 3 below. These particles have been studied by scholars such as Barke (2001), Cook (1990, 1992), McGloin (1990), and Uyeno (1971), to name a few. Maynard observes, "in general frequent insertion of particles encourages rapport between the conversation partners and achieves a closer monitoring of the partners' feelings" (p. 28).

Table 3. Summary of final particles

In my CMC data, there are also a number of particles in the website messages. The most widely used particle of rapport in the messages is ne(e) and the next is yo. All the participants (N=23 in the representative three BBSs) used at least one particle once in their messages. Example 8 illustrates how these particles are used in actual messages. In (8) and subsequent examples, each of the messages is given in the original first, then transliteration, and finally, in English translation.

(8) Final particles bolded¹⁶

1 Re: そうなの？投稿者：GMTNでちゅ 2000/10/08(Sun) 13:51:30
2 あ、リンク貼らせて頂きましたｖｖ逃がさんよ？
3 最近、踊るから離れてるGMTNだけど仲良くしてね。
4 只今GMTNの心は富樫でいっぱいいっぱい。可愛いよね？
5 ホワイトアウトのサイトになっちゃおうかなあ？（死）

1 Re: sou na no ? Sender: GMTN dechu Date and time
2 A, rinku harasete itadaki mashita vv nigasan yo ?
3 Saikin, Odoru kara hanareteru GMTN da ke do nakayoku shite ne.
4 Tada ima GMTN no kokoro wa Togashi de ippai ippai. Kawaii yo ne ?
5 Howaito auto no saito ni nattchaou ka naa ? (shi)

1 Response: Title: Is that so? Sender: GMTN Date & Time
2 Oh, you let me put a link [to your website], vv I won't let you get away, OK? [vv replaces the heart symbol]
3. Recently, I'm not really into Odoru [title of film], but let's stay in touch, okay?
4. Now my heart is full, full of Togashi [character in film]. He is cute, isn't he?
5. Shall I make my site a Whiteout site, do you think?' [kanji for death]

(8) was posted in response to an earlier message posted by the manager of the website, who said that she had been hospitalized. vv in line 2 of (8) is a representation of a heart symbol, but can also function as a kind of punctuation mark discussed earlier. In the five lines of this message we see a particle of rapport in every line: "no" in line 1, "yo" in 2, "ne" in 3, "yo ne" in 4, and "naa" in 5. This indicates how widely used these particles are. The user 'talks' to the previous message sender, using multiple rapport-providing final particles. A particle, no, even occurs in the title of the message (line 1). As seen in the translations, this type of particle functions to elicit or monitor the partner's reaction on the topic discussed; in the BBS setting, these reactions will be elicited at a later time, when the message is read. Also, since the messages in a BBS can be read by other users who may have the same interests, the writer may also be addressing the readers as well in this affect-encoding manner.

Polite versus Plain Styles

We have seen how final particles are used in many lines, and how the highly affective tone of the messages is conveyed. Let us now turn to levels of politeness. I compare conversational and CMC use of sentence-ending forms, specifically the polite and plain verb forms. There are at least two levels of politeness and formality in Japanese conversational discourse: polite/formal and plain/informal. These levels are distinguished by verb endings typically known as desu/masu polite style and da/(r)u plain style. Maynard (1989) states there are few polite/formal uses of sentence endings in her data set of Japanese conversation, and that when they are used, there are reasons, such as quoting other people who might be the speaker's superior. Utterances directly addressed to the conversational partner in her data are all in non-polite forms (p. 37-38). In general, people in a superior position, such as teachers as opposed to students, are more likely to receive respect in Japanese traditional culture, and hence are addressed with more polite language. In contrast, in conversation among peers the most unmarked form is the plain da/(r)u ending. Maynard's data are consistent with cultural expectation of conversational behavior among people with equal status.

The BBS message writers, however, show different behavior in their use of polite and plain verb endings. In my data about two-thirds of the verbs have polite/formal forms, and plain forms, when they are used, are rather scattered. Normally one user is consistent with respect to which verb ending is used in a single message. This phenomenon can be interpreted by taking the BBS setting into account. Since the users do not share physical space, they do not know other users, even though they share common interests and hope to interact with their fellow users. Thus it would be more appropriate to address their fellow users with polite forms. Interlocutors may be their senior/superior in age. Users who are not very conscious of relative age or status differences between themselves and others may implicitly assume that others are their equals; they would be more likely to use plain address forms. Also, since most users are likely to be women, and women tend to use polite language more often than men, this may explain why more cases of polite endings appear in the data. For a discussion on politeness and women's use of Japanese, see Ide and Yoshida (1999), for example. Another factor that might contribute to the polite use of language is the site manager's guidelines. As explained above, she expects her contributors to use polite language on her website. Users' compliance with her request results in these polite forms.

There are, however, cases where a writer mixes these two styles in one message, either with a few informal forms among predominantly formal forms, or a few formal forms among mostly informal endings. On mixing polite desu/masu and plain da/suru forms, Maynard (2002) summarizes the views of scholars who have discussed mixing polite desu/masu and plain da/suru forms. She cites Haga's (1962) remark, "in a discourse where da-endings dominate with occurrences of sporadic masu/desu endings, the latter functions (1) to mark formality, (2) express humor and sarcasm, (3) insert personal comment, and to (4) mark vocatives directly addressing the partner " (Haga, 1962, pp. 77-83, cited in Maynard, 2002, p. 278). Example (8) given above shows this situation. The endings are predominantly informal, and the only occurrence of the formal masu ending is in line 2. The underlined part, harasete itadaki mashita in line 2 is also a noteworthy expression, in that it has a humble form in Japanese honorific language. Literally translated, line 2 would be "I humbly received your letting me put a link…" Such humble expressions are more consistent with the formal verb itadakimashita than plain itadaita.

This use of a formal ending with honorific vocabulary is a marked use in an otherwise informal, friendly style. The message writer seems, at least at the beginning of her message, to make the message sound polite, because the subject matter concerns the addressee's website, and this politeness conveys a feeling of respect for the addressee.

By contrast, in example (9) and the response to it in example (10), the discourse is dominated by formal/polite masu/desu endings with a few scattered plain forms. Observe (9) and (10) below:

(9) Polite forms bolded

1 ハジメマシテ。投稿者：HNT A 2000/09/30(Sat) 20:41:54
2 こんばんは。
3 BBSでは、初めまして、です。
4 初めて書きこむには、なんだかとっても勇気がいりました。
5 こんなに緊張するものだとは・・・（苦笑）
6 ・・・「THE　MOVIE」やるんですか？？！
7 知りませんでしたよ～～！！
8 ぎゃ～～、うれしいいいいいい！！どうしましょ？！
9 っていうか、予告みたい！！
10 はああ・・・楽しみ♪どうしましょ！！

1 Title: hajimemashite. Sender: HNT A Date and time
2 Kon ban wa.
3 BBS de wa, hajime mashite, desu.
4 Hajimete kakikomu ni wa, nan daka totte mo yuuki ga irimashita
5 Kon na ni kinchou suru mono da to wa ・・・ (kusyou)
6 ・・・ "THE MOVIE" yaru n desu ka ??!
7 shirimasen deshita yo ～～!!
8 gya ～～, ureshiiiiiii !! dou shimasho?!
9 tte iu ka, yokoku mitai !!
10 ha a a ・・・tanoshimi♪dou shimasho?!

1 Title First time [to post] Sender: HNT A Date and time
2. Good evening.
3. On BBS this is my first time.
4. In order to post [my message] for the first time, somehow a lot of courage was necessary
5. such a stressful thing (I hadn't expected that) … [kanji for wry smile]
6. Is "THE MOVIE" going to be shown??!
7. I didn't know that, you know ～～!!
8. Wow ～～ veeeeeery happyyyyyyyy!! What shall I do?!
9. You know, I want to see the preview!!
10. Haaa・・・　can't wait♪What shall I do?

(10) Polite forms bolded and honorific forms underlined¹⁷

1 Re: ハジメマシテ。投稿者：MSSG K ＠管理人 2000/09/30(Sat) 23:24:31
2 わっvvv(*^_^*)
3 いらっしゃいませ～♪いつもメールではお世話に…（笑）
4 そっ、そんなに緊張なされましたかっ！！！！(*>_<*)
5 別にとって食いはしませんから (あたりまえだ…）
6 フフフ～♪７日やりますね～。
7 標準で録画されますか？（笑）じっくり！まったり！（謎）見ましょう ね～(-_-)予告は見れましたか？(*^_^*)
8 そうそう、あとメールの件ですが（笑）了解しやしたっ！！
9 ニヤリ、楽しみに待ってて下さいねvvv
10 （またメールで改めて連絡しまっすぅ～♪）

1 Title Re: Hajime mashite Sender: MSSG K @Manager Date and time
2 Wa [symbol to represent long consonant] vvv(*^_^*)
3 Irasshyaimase ～♪itsu mo meeru de wa o sewa ni…（kanji for laugher）
4 So [long consonant] sonna ni kinchyou nasare mashita ka [long consonant] ！！！！ (*>_<*)
5 betsu ni totte kui wa shimasen kara (atarimae da…)
6 Huhuhu ～♪nanoka yarimasu ne～
7 Hyoujun de rokuga saremasuka ? [kanji for laugher] jikkuri ! mattari ! [kanji for puzzlement] mimasyou ne～(-_-) yokoku wa miremashi ta ka？(*^_^*)
8 Sou sou, ato meeru no ken desu ga [kanji for laugher] ryoukai siyashi [variant of mashi]ta [long consonant] !!
9 Niyari, tanoshimi ni mattete kudasai ne vvv
10 (mata meeru de aratame te renraku shimassuu [emphasized form of masu] ～♪)

1 Title Re: First time [to post] Sender: MSSG K @Manager Date and time
2 Oh Wow vvv(*^_^*)
3 Welcome ～♪Thank you (for your honorable labor in) e-mail as always … [kanji for laughter]
4 [Stuttering]Were, were you that nervous [honorific]! ! ! (*>_<*)
5 [No one] will take you and eat you up. (Of course not...[plain form])
6 Huhuhu [laughter] ～♪on the 7th they will show it, right～
7 Are you going to record [honorific] in standard [playing mode] [kanji for laughter] Let's watch [it]
leisurely! appreciatively [kanji for puzzlement] ～(-_-) Were you able to see the preview? (*^_^*)
8 Oh by the way, as for the other topic, e-mail, [kanji for laughter], I understand !!
9 Grinning, please be waiting [for it] eagerly vvv
10 (I'll contact you on another occasion by e-mail～♪）

In (9), the writer posts a message for the first time, and she explains how nervous she was when doing so, all in polite formal language. When the topic shifts to the movie of her interest, she changes the style to plain informal endings, which conveys an excited tone here. The use of informal forms in the latter half of (9) can be interpreted as signaling her increased excitement.

According to Haga (1962, p. 74, cited in Maynard, 2002, p. 278), such informal use surrounded by formal style "expresses an interpersonal familiarity and closeness with the partner." Formal desu/masu endings create distance between the speaker and the hearer, and informal da forms can shorten such psychological distance. Maynard (2002, p.280) presents her own analysis of this mixing of styles, and concludes that in a desu-masu dominant discourse, da is selected, "(1) when the speaker takes a perspective internal to the narrative setting and immediately responds within that framework, (2) when the speaker presents background information semantically subordinate within the discourse structure, and (3) when the speaker finds the partner close enough and the speaker uses a style similar to the style in which he or she self-addresses." This analysis fits Example (10). The message writer feels very close to the addressee. The remark after "no one will take you and eat you up" has the plain da-form, in atarimae da and is not very important semantically; it could be an afterthought and hence subordinate to the main discourse. The use of parentheses also supports this analysis, because parentheses are generally used when the content is only marginally relevant to the main topic.

Finally, there is a new element in my data, which has, to my knowledge, not been accounted for in the literature. It appears in line 5 in Example (10). After an ordinary (and yet very playful) sentence betsu ni totte kui wa shimasen kara '[No one] will take you and eat you up,' this CMC user adds the remark atarimae da 'Of course not' in parentheses. Though this addition is extremely obvious and might sound odd, it is not strange, especially to those who are familiar with popular Japanese comic dialogues.¹⁸

In most Japanese comic dialogues, one actor poses as a man of common sense, while the other behaves like a trickster. Every time the man of common sense states something ordinary and unwitty, the trickster reacts with a witty and sometimes sarcastic reply.¹⁹ In the example in line 5, the writer apparently mimics this comic dialogue all by herself. The main part is a statement by a man of common sense, and this statement is commented on by a trickster in parentheses. What is extraordinary here is that one writer plays a double role: a man of common sense and a trickster. In a sense, the writer creates dramatic discourse or a stage performance.

This is not so strange if one takes account of the situation in which this CMC takes place. Interaction on BBS websites can be viewed by other users who visit the site. The writer is well aware of such viewers. Moreover, this BBS website is for entertainment, so that the writer, the creator of the site, expects to present a kind of humorous dramatization that can be enjoyed by visitors. Frequent uses of the kanji character for laughter and various amusing emoticons and innovative punctuation marks such as the musical note ♪　also exemplify the writer's intention to make messages enjoyable for the addressee and other online viewers as well.

This also explains the usage of the da-form in atarimae da. The da here is extremely noticeable, both because it has an abrupt sound (Maynard, 2002, p. 279), and because it is surrounded by so many polite expressions, including three honorific forms: o in o sewa (line 3), nasare in kinntyou nasaremasita (line 4), and sare in rokuga saremasuka (line 7). Such a great gap between very polite language and a short sudden plain form creates a humorous effect, which also contributes to the entertainment purpose of the website.

Concluding Remarks

This study has shed some light on the communicative behavior of young Japanese BBS users. We have seen that they employ a variety of flexible options to address readers who have varying degrees of familiarity and involvement. Their use of language largely resembles that of informal spoken Japanese with genuinely creative orthography.

We have noticed that users employ such informal spoken features as final particles in order to interact with other users online as if they were in face-to-face conversation. We have also seen greater use of polite forms than plain forms. Some users may be very involved and assume a highly interactive stance by using more plain forms, while others who may not be as involved but are more concerned with avoiding a rude impression may convey social distance with polite masu forms. If they are quite familiar with each other, they may engage in message exchange without masu/desu polite forms even in online interaction. When the participants are new or do not have a sufficient sense of familiarity to use plain forms, they may use more masu/desu polite forms than da/suru forms. Thus, the varying degrees of style mixture in CMC may be related to the extent to which the users feel involvement and closeness to other users.

The above findings come from Japanese fan sites. Obviously, if we were to look at other kinds of websites, different phenomena might be observed. Users in the sites studied seem to aim at being friendly with each other because of shared interests. It would be hard to find friendly interaction in websites where conflicting opinions and interests coexist. It would be intriguing to investigate how users on other websites employ language, and how they interact. It would be of great interest to see whether culturally-constrained Japanese communication patterns are maintained, or whether novel behavior would be found. This type of research could make a theoretical contribution to our knowledge of Japanese online communication in sociocultural context.

Studies of Internet communication owe a great deal to previous studies on the nature of, and differences between, spoken and written language (e.g., Chafe, 1982; Ochs, 1979; Tannen 1982). Research about communication on the Internet strengthens the claim that various genres of communication (regardless of medium) are located along a continuum, and that there is no clear dichotomy between speech and writing (e.g., Baron, 1984; Collot & Belmore, 1996; Yates, 1996). CMC may have an intermediate position between spoken and written language, sharing certain features with both, while displaying at the same time other features unique to online genres. The investigation of online discourse deepens our understanding of communicative behavior generally, in that we encounter types of communicative events that have not previously been observable for study. At the same time, this study shows the importance of attending to the characteristics of the specific language involved, and to the sociocultural context of its use online, for an understanding of emergent patterns of online communication.

Acknowledgments

I am grateful to Patricia M. Clancy and Mary Bucholtz for their valuable comments on earlier versions of this article. However, any errors that may remain are my responsibility.

Footnotes

1. As of December, 2002, Soumusyou Tyuusin Doukou Chousa [Survey of information and communications in

Japan

] March 7, 2003.

2. As of September, 2002, Global Internet Statistics (by language).

3. Strictly speaking, there is one more kind, which is suuji, numeral script. Numbers are normally given in Arabic and Chinese numeral characters. As a convention, Arabic numerals are used in horizontal writing and Chinese numerals in vertical writing, though vertical writing sometimes includes Arabic numerals.

4. In Nishimura (to appear) I discuss unconventional kanji uses online, in which loan words of English origin are represented by kanji.

5. Furigana symbols are attached to all the kanji characters used in some comic books (not necessarily for young readers) to supplement how each kanji character is read. In some instances furigana is used even for English words to indicate how the English word is read.

6. Single/double quotation marks function in Japanese in the opposite way from English conventions, in that 『　』 is used to mark a quote within a quote, while 「　」　is for simple quotes.

7. Horizontal writing is transferred from Western languages, and native writing is vertical, in which lines are read/written from right to left. In online communication, lines go only horizontally, except in documents created with some word processing software that allows vertical writing, such as Microsoft Word.

8. Unshared punctuation marks used only in English include colons and semicolons. Though in Japanese colons can be used in time expressions, e.g., 10:00, just like English, semicolons are not found. This may explain Japanese writers' relative difficulty in using these marks when writing English prose.

9. Readers interested in Japanese word processing might like to visit:
http://www.microsoft.com/windows/ie/downloads/recommended/ime/default.asp,
http://www.justsystem.co.jp/atok/atok16w/
Further, the following website explains how to enter Japanese on the computer: http://www.seanspot.com/jwrite/MS-IME/UseIME/jwrite-ime-enter-entertext.htm. The site below offers a primitive but usable virtual Japanese keyboard, and a minimal experience of what it is like to select a syllable or word in Roman characters and see it transformed into hiragana and katagana: http://www.mojio.net/japan/jkwe.html.

10. It is also possible to enter kana symbols directly through keyboarding without the process of romaji-kana conversion, though users need to know where each kana symbol is assigned on the keyboard. See Figure 1. Image of the Japanese keyboard.

11. For characterization of e-mail in comparison with other CMC modes, see Baron (1998) and Danet (2001), for example. The focus in Baron's (1998) study on e-mail is "one-to-one dialogue with an identified interlocutor" (p.142).

12. These fan sites are very easy to find through search engines such as Google. Just type the name of a popular actor, search, and then a substantial number of them will appear. I am indebted to Keiko Nishimura for drawing my attention to these sites.

13. As to the length of messages in Table 1 above, the number of lines in one message is entered. To illustrate how it is determined, example (8) in the section on "final particles" is regarded as a message of four lines. The first line is not included in the message because it has the subject/title and information on sender and time. This information is not considered to be the message itself, though the subject/title conveys some informational content in certain cases. Similarly, Example (9) is a 9-line message. Example (10), in which Line 7 exceeds a normal screen width, is counted as a 10-line message. The length in terms of number of lines is to measure relative length roughly, and is not intended to be used as a strict measure.

14. Other examples of Eimoji, English words, include, for instance, A子, which can be used to refer to a person, eiko. In this case, the alphabet A is used in combination with a kanji character, 子, pronounced ko 'child', which is often used at the end of a girl's name. The motivation behind the use of A子 is to mask the real identity of an individual, in place of the individual's authentic name written either in kanji or kana.

15. For details on emoticons, see Katsuno and Yano (2002).

16. Senders' names in the examples are changed to pseudonyms hereafter.

17. In example (10), there are four face marks, three of them different. Although the exact meanings of these are difficult to ascertain, they can roughly be considered to mean:
(*^_^*) in line 2 and 7 shows some kind of shyness, according to an online face mark dictionary at http://kao.popkmart.ne.jp/tereru.html. The asterisks referring to the cheeks (-_-) in line 7 look mysterious, since there is a kanji for mystery in the same line.

18. Danet (2001) makes a strong case for CMC as stage performance. Atarimaeda as stage performance supports her claim. BBS’s, along with several other CMC modes, are suitable for the display of performance because of the presence of an audience, the fellow participants on the site.

19. This type of comic stage performance is referred to as manzai in Japanese popular culture. See Bensky (1998) for an introduction to manzai.

References

Baba, J. (2001). Pragmatic function of Japanese mimesis in emotive discourse. Southern Japan Seminar, Atlanta. Retrieved March 10, 2003 from http://web.aall.ufl.edu/SJS/Baba.pdf.

Backhouse, A. E. (1984). Aspects of the graphological structure of Japanese. Visible Language, 18(3), 219-228.

Barke, A. J. (2001). Inconspicuous gender marking: The case of Japanese sentence-final particles, yo and ne. In K. Horie and S. Sato (Eds.), Cognitive-functional linguistics in an East Asian context (pp. 235-255). Tokyo: Kuroshio.

Baron, N. S. (1984). Computer mediated communication as a force in language change. Visible Language, 18(2), 118-141.

Baron, N. S. (1998). Letters by phone or speech by other means: The linguistics of e-mail. Language and communication, 18, 133-170.

Baron, N. S. (2000). Alphabet to e-mail: How written English evolved and where it's heading. New York and London: Routledge

Bensky, X. B. (1998). Commodified comedians and mediatized manzai: Osakan comic duos and their audience. Asian Studies on the Pacific Coast Conference Papers. Retrieved July 14 2003 from http://mcel.pacificu.edu/aspac/papers/scholars/bensky/bensky.htm

Chafe, W. L. (1982). Integration and involvement in speaking, writing, and oral literature. In D. Tannen (Ed.), Spoken and written language: Exploring orality and literacy. Advances in Discourse Processes Series Vol. IX (pp. 35-53). Norwood, NJ: Ablex.

Cho, Natasha. (In press). Linguistic features of electronic mail in the workplace: A comparison with memoranda. In S. Herring (Ed.), Computer-Mediated Conversation. Cresskill, NJ: Hampton Press.

Clancy, P. M. (1982). Written and spoken style in Japanese narratives. In D. Tannen (Ed.), Spoken and written language: Exploring orality and literacy. Advances in Discourse Processes Series Vol. IX. (pp. 55-76). Norwood, NJ: Ablex.

Collot, M. & Belmore, N. (1996). Electronic language: A new variety of English. In S. C. Herring (Ed.), Computer mediated communication: Linguistic, social and cross-cultural perspectives (pp. 13-28). Amsterdam and Philadelphia: John Benjamins.

Cook, H. M. (1990). An indexical account of the Japanese sentence-final particle no. Discourse Processes, 13, 401-439.

Cook, H. M. (1992). Meanings of non-referential indexes: A case study of the Japanese sentence-final particle ne. Text, 4, 507-539.

Danet, B. (2001). Cyberpl@y: communicating online. Oxford and New York: Berg. Companion website: http://atar.mscc.huji.ac.il/~msdanet/cyberpl@y/ (best viewed in Internet Explorer).

Danet, B., Ruedenberg, L., & Rosenbaum-Tamari, Y. (1997). “Hmmm...where's that smoke coming from?" Writing, play and performance on Internet Relay Chat. Journal of Computer-mediated Communication 2 (4). Retrieved November 12, 2003 from http://jcmc.indiana.edu/vol2/issue4/

Danet, B., Ruedenberg, L., & Rosenbaum-Tamari, Y. (1998). ""Hmmm...where's that smoke coming from?" Writing, play and performance on Internet Relay Chat." In F. Sudweeks, M. McLaughlin, & S. Rafaeli, (Eds.), Network and netplay: Virtual groups on the Internet (pp. 47-85). Cambridge, MA: AAAI/MIT Press.

Gendai no esupuri. [Esprit of today]. (1998). Vol. 370. Special Issue on "Intaanetto Shakai [Internet Society]." Tokyo: Shibunndo.

Gimenez, J.C. (2000). Business e-mail communication: Some emerging tendencies in register. English for Specific Purposes, 19, 237-251.

Global Internet Statistics (by langauge) (2002). Retrieved March 8, 2003, from http://www.glreach.com/globstats/

Haga, Y. (1962). Kokugo Hyougen Kyooshitsu. [Classrooms for/lessons in expressions in Japanese]. Tokyo: Tookyoodoo.

Herring, S. C. (Ed.) (1996). Computer-mediated communication: Linguistic, social and cross-cultural perspectives. Amsterdam and Philadelphia: John Benjamins.

Herring, S. C. (2001). Computer-mediated discourse. In D.Tannen, D. Schiffrin, & H. Hamilton (Eds.), Handbook of discourse analysis (pp. 612-634). Oxford: Blackwell.

Ide, S., & Yoshida, M. (1999). Sociolinguistics: Honorifics and gender differences. In N. Tsujimura (ed.), The handbook of Japanese linguistics (pp.444-480). Malden, Mass. and Oxford, UK: Blackwell.

Kataoka, K. (1997). Affect and letter-writing: Unconventional conventions in casual writing by young Japanese women. Language in Society, 26, 103-136.

Katsuno, H. & Yano, C. R. (2002). Face to face: On-line subjectivity in contemporary Japan. Asian Studies Review, 26, 205-232.

Lan, L. (2000). e-mail: A challenge to standard English? English Today, 16, 23-29, 55.

Lee, J. Y. (1996). Charting the codes of cyberspace: A rhetoric of electronic mail. In L. Strate, R. Jacobson, & S. B. Gibson (Eds.), Communication and cyberspace: Social interaction in an electronic environment (pp. 275-296). Cresskill, NJ: Hampton Press.

Matsuda, P. K. (2001). Voice in Japanese written discourse: Implications for second language writing. Journal of Second Language Writing, 10, 35-53.

Matsuda, P. K. (2002). Negotiation of identity and power in a Japanese online discourse community. Computers and Composition, 19, 39-55.

Maynard, S. K. (1989). Japanese conversation: Self-contextualization through structure and interactional management. Norwood, N.J. : Ablex.

Maynard, S. K. (2002). Linguistic emotivity : Centrality of place, the topic-comment dynamic, and an ideology of pathos in Japanese discourse. Amsterdam and Philadelphia: John Benjamins.

McGloin, N. H. (1990). Sex difference and sentence-final particles. In S. Ide & N. H. McGloin (Eds.), Aspects of Japanese women's language (pp. 23-41). Tokyo: Kuroshio .

Nihongogaku. [Japanese Linguistics]. (1993). Special Issue on “Pasokon tsuushin” [Communication by personal computers.] Tokyo: Meiji Shoin, December.

Nihongogaku. [Japanese Linguistics]. (1996). Special Issue on “Denshi shakai no komyunikeisyon” [Communication in electronic society.] Tokyo: Meiji Shoin, November.

Nihongogaku. [Japanese Linguistics]. (2000). Special Issue on “Keitai to gengo koudou, higengo koudou” [Cell phones and linguistic and non-linguistic behavior.] Tokyo: Meiji Shoin, October.

Nihongogaku. [Japanese Linguistics]. (2001). Special Issue on “Keitai meeru” [Cell phone (e)mail.] Tokyo: Meiji Shoin, September.

Nishimura, Y. (2001). Kakareru hanashi kotoba―intaanetto jyou ni mirareru atarashii komyunikeeshon sutairu [Written ‘Spoken’ Discourse: Novel communication style found on the Internet]. In Toyo Gakuen University and Toyo Women's College Association for Studies on Language (Ed.), Shiriizu kotoba no supekutoru Taiwa [Dialogue: Series in Spectrum of Languages] (pp. 27-57). Tokyo: Liber Press.

Nishimura, Y. (to appear). Establishing a community of practice on the Internet: Linguistic behavior of online Japanese communication. Proceedings of the 29th Annual Meeting of the Berkeley Linguistics Society.

Ochs, E. (1979). Planned and unplanned discourse. In T.Givon (Ed.), Discourse and syntax (pp. 51-80). New York: Academic Press.

Satake, H. (1995). Aratana buntai wo mosaku suru. [Exploring new styles.] Gengo [Language], 24(1), 52-59.

Satake, H. (2002). Hugou no mondai [Issues on signs and symbols]. In Y. Hida & T. Sato (Eds.), Gendai Nihongo Kouza [Contemporary Japanese Linguistics Series] . Vol 6 (pp.104-126). Tokyo: Meiji Shoin.

Smith, J., & Schmidt, D. (1996). Variability in written Japanese: Towards a sociolinguistics of script choice. Visible Language, 30(1), 47-71.

Souomyshou [Ministry of Public Management, Home Affairs, Posts and Telecommunications, Japan.] (2003). Tsuusin Doukou Chousa. [Survey of information and communications in Japan] Retrieved March 8, 2003 from http://www.johotsusintokei.soumu.go.jp/statistics/data/030307_1.pdf

Stanlaw, J. (2002). "Hire" or "Fire"? Taking AD-vantage of innovations in the Japanese syllabary system. Language Sciences, 24(5-6), 537-574.

Sugimoto, T. & Levin, J. (2000). Multiple literacies and multimedia: A comparison of Japanese and American uses of the Internet. In G. Hawisher & C. Self. (Eds.), Global literacies and the World-Wide Web. (pp. 133-153). London: Routledge.

Suzuki, T. (1975). On the twofold phonetic realization of basic concepts: In defense of Chinese characters in Japanese. In Peng (Ed.), Language in Japanese society: Current issues in sociolinguistics (pp. 193-214) Tokyo: University of Tokyo Press.

Suzuki, T. (1977). Writing is not language, or is it? Journal of Pragmatics, 1, 407-420.

Tanaka, Y. (2001). Keitai denwa to denshi meeru no hyougen. [Expressions in messages via cell phones and e-mail on the computer]. In Y. Hida & T. Sato (Eds.), Gendai Nihongo Koza [Contemporary Japanese Language Series]. Vol.2 Hyougen [Expressions.] ( pp. 98-127) Tokyo: Meiji Shyoin.

Tannen, D. (Ed.) (1982). Spoken and written language: Exploring orality and literacy. Advances in Discourse Processes Series Vol. IX. Norwood, NJ: Ablex.

Uyeno, T. (1971) A study of Japanese modality: A performative analysis of sentence particles. Unpublished Ph.D. dissertation, University of Michigan.

Werry, C. C. (1996). Linguistic and interactional features of Internet Relay Chat. In S. C. Herring (Ed.), Computer-mediated communication: Linguistic, social and cross-cultural perspectives (pp. 47-63). Amsterdam and Philadelphia: John Benjamins.

Yates, S. J. (1996). Oral and written linguistic aspects of computer-conferencing. In S. C. Herring (Ed.), Computer-mediated communication: Linguistic, social and cross-cultural perspectives (pp. 29-46). Amsterdam and Philadelphia: John Benjamins.

About the Author

Yukiko Nishimura is Professor in the Faculty of Humanities at Toyo Gakuen University, Chiba, Japan. She completed her undergraduate education at Japan Women's University, Tokyo, and did graduate work at the International Christian University, Tokyo and at the University of California, Berkeley. She also taught Japanese at Yale University and most recently was a Visiting Researcher (2002-2003) in the Department of Linguistics at the University of California, Santa Barbara. Previous work includes contrastive studies in English-Japanese discourse, issues involving Japanese learners of English in Japan, especially on writing, and English-language learners in the State of California. Her continuing research interests in contrastive analyses of English and Japanese, specifically in the areas of spoken and written discourse, have led as a natural extension to the study of Internet discourse in the two languages.
Address: Faculty of Humanities, Toyo Gakuen University 1660 Hiregasaki, Nagareyama-shi, Chiba, 270-0161 Japan. Telephone: +81-471-50-3001; Fax: +81-471-50-3006.