Gregory A. Bennett Georgetown University* UDK 811.521'42:159.942:004.773 SAY WHAT? THE REPRESENTATION OF LAUGHTER AS A CONTEXTUALIZATION CUE IN ONLINE JAPANESE DISCOURSE INTRODUCTION Over the course of several years of texting and Tweeting while living in Japan, I noticed that many Japanese users utilize orthographic symbols of laughter to convey a tone for, or key, utterances in the digital realm. According to Werry (1996), users can strategically employ symbols to achieve interactional goals throughout digital exchanges. In this paper, a sub-component to a larger study I am conducting with Takafumi Ohyama, I expand upon Werry's (1996) findings by examining representations of laughter as contextualization cues in Japanese digital discourse. I explore how saywhat327,1 a Japanese male college student from Tokyo, uses (^), the kanji character for "laugh" situated between two parentheses, and w, the first letter of the romaji transliteration of that same kanji character, wara, on Twitter to key his utterances as warm/friendly and playful, respectively. At an interactional level, I analyze how he employs these characters to build rapport with other users. I begin by situating this study within literature pertaining to digital discourse as a hybrid mode of communication, focusing specifically on contextualization cues and their written equivalents in online interactions. In order to explore how Japanese users constrain the ways in which they utilize the and w cues in the digital realm, I will explain the symbols in terms of the pragmatic meaning they carry based on prior texts that associate kanji with formality and romaji with informality. I supplement these inferences with insights provided by seven Japanese college students, collected via a survey of media ideologies on and w usage. In these surveys, participants offer that is considered standard and/or formal which conveys a warm/friendly tone and that w is informal and conveys a playful tone. Based on this information, I analyze three examples of Twitter discourse from saywhat327's public account, to demonstrate how he (1) uses to key his utterances as warm/friendly in the formal register of Japanese, (2) uses w to key his utterances as playful in the informal register of Japanese, and (3) uses both variants with informal Japanese to key two separate utterances within the same Tweet as warm/friendly and playful, respectively, based on the semantic content of each utterance. In each of these cases, I argue that saywhat327 builds rapport with his inter- * Author's address: 1437 37th Street NW, Poulton Hall 240, Washington, DC 20057-1051, USA. Email: gab45@georgetown.edu 1 While their accounts are open to public access, I have given pseudo-screennames to all users whose pages I accessed for data in this paper. locutor by keying his utterances via the and w cues. Overall, the different ways in which saywhat327 uses both characters is an indication that variations in stylistic representations of text in digital discourse can, as written contextualization cues, signal different keys for the purpose of achieving interactional goals. LITERATURE REVIEW Over the past two decades, linguists have mused over whether to classify computermediated communication (CMC), or what has more recently been subsumed under the more general term, digital discourse, as speech or writing. This issue became increasingly pertinent in describing the interactional nature of synchronous (instantaneous, one-to-one communication) modes of CMC such as Internet Relay Chat and instant messaging. Werry (1996) addresses this issue, claiming that despite the fact that synchronous CMC exchanges are conducted orthographically, they still exhibit properties that are similar to conversational discourse. Because synchronous CMC consists of real-time, instantaneous exchanges comprised solely of text, he argues, users have developed "a complex set of orthographic strategies designed to compensate for the lack of intonation and paralinguistic cues that interactive written discourse imposes" (Werry 1996: 57) on participants. Werry cites punctuation, capitalization, and nonstandard spellings as three stylistic aspects of text that users can manipulate in order to convey such paralinguistic cues. In turn, he states, these cues can then be utilized to achieve myriad interactional goals such as shifting in and out of different speech registers. The textual variants that Werry describes as being the written equivalent of para-linguistic and prosodic cues in conversational discourse can, therefore, be seen as the digital version of Gumperz' (1982) contextualization cues. Gumperz offers that as participants interpret contextualization cues, or "constellations of surface features of message form" (Gumperz 1982: 131) such as pitch, intonation, and gesture, they do so based on prior spoken interactions in which they have previously encountered the cue and the tones it signals. These tones that exhibit meaning above the semantic level are what Goffman (1981) calls linguistic "keying" which convey to other participants not to take what is being said for its literal meaning. In the case of synchronous digital discourse, as we saw in Werry's study, the textual constraints of the digital realm cause users to manipulate stylistic aspects of orthography in order to convey contextualization cues as they make sense of what is transmitted throughout their interactions. When Japanese users employ symbols such as and w online to convey linguistic keys, they make sense of which variant to use based on prior contexts in which they have encountered the symbols as speakers do with prosodic and embodied cues in conversation. Here, Becker's (1995) concept of prior text comes into play as users fit old orthographic symbols into new digital contexts. The texts in which participants have previously encountered kanji and romaji scripts provide users a basis upon which they can link with formality and w with informality and thus make judgments about which interactional contexts merit the use of one variant over the other. These associations of formality and informality are rooted in the notion of group relations, whereby the formal register (i.e. kanji script for official reports) is used with members of a speaker's "out-group" — usually someone who is not a close friend or family member — and the informal register (i.e. non-kanji scripts such as romaji for passing notes) is used for showing closeness with those in one's "in-group." As contextualiza-tion cues, and w can thus serve as a means by which users can signal pragmatic meaning based on how the characters look, just as speakers signal pragmatic meaning in conversation based on how their words sound. In this way, Japanese users can utilize and w as contextualization cues to position themselves and others within a participant's in-group or out-group, and thus convey rapport and mitigate interactional distance. Overall, these prior text associations influence the ways in which Japanese users select between and w as contextualization cues in digital discourse. In turn, they comprise user ideologies about how the cues should be used in digital media. Gershon (2010: 3) cites media ideologies as the "set of beliefs about communicative technologies with which users...explain perceived media structure and meaning" and ultimately shape the ways in which they use a particular medium to communicate. Applying this notion to written-out laughter as a contextualization cue in Japanese digital discourse, it holds that users maintain a set of beliefs constraining how laughter should be represented textually based on prior texts of formality, which impacts their conceptualization of the contexts in which they deem it appropriate to employ a particular orthographic representation. In other words, the prior texts that associate kanji with formality and romaji with informality directly influence users' media ideologies about how and in which digital contexts it is appropriate to use and/or w. RESEARCH DESIGN The Archivist: Collecting Twitter Data In order to collect and analyze public exchanges that feature and w usage, I turned to the micro-blogging site Twitter. As Baron (2010) notes, many media that are traditionally considered to be asynchronous (one-to-many communication) mediums can be used synchronously (one-to-one communication) if the exchanges take place in real-time, instantaneous turns (i.e. an e-mail exchange between two users in which both participants send replies immediately upon receiving a transmission). Even though Twitter is considered to be a traditionally asynchronous medium as it is a micro-blogging site, many users direct their Tweets specifically to other users by attaching the @-sign to the respective user's screenname (i.e. saywhat327: @sobasnare Where are you?). These types of Tweets, or "Mentions" as Twitter labels them, appear in a user's Mentions tab whenever another user addresses a Tweet to their account. For this paper, I collected synchronous exchanges that feature (^), and from that group, discovered users who also utilize w. I used The Archivist (http://archivist.visitmix.com/), a free online search engine, to compile a corpus of 243 Tweets taken from 22 different public Twitter accounts. Of these users, I chose to focus on saywhat327 because he employs both and w in his Tweets. The examples I analyze here are taken from his public Twitter account and are comprised of both asynchronous and synchronous interactions that he conducted with three different users: sobasnare, ykt42, and kels81. Media Ideologies Survey Data In order to gain a preliminary sense of how Japanese users perceive and w usage in digital discourse, I administered a survey of media ideologies asking participants to comment on hypothetical and actual Twitter discourse that featured both contextualization cues. I asked seven college-aged Japanese students to answer essaystyle questions that targeted their media ideologies about how and to what effect and w are used in online discourse. Of these seven participants, all commented on usage as follows: standard, friendly, respectful, formal, soft, and warm, and five students claimed that w is playful, affectionate, friendly, casual, light-hearted, and for use between close friends. These results are in line with the prior text associations between kanji usage in formal written discourse with people in one's out-group, and romaji usage in informal written discourse with people in one's in-group. It is based on these media ideologies about how and to what effect both variants are used in digital discourse that I analyze three exchanges taken from saywhat327's public Twitter account. ANALYSIS The User: saywhat327 Based on saywhat327's Twitter account profile, he is a male college student attending a major university in the Tokyo region of Japan. He is a member of the Folk Song Society on campus and is chief of the percussion squad. Most of his Tweets pertain to music and the recording industry. At the time of my data collection he had produced a total of 1,191 Tweets, was following 142 other Twitter users, and had 125 followers. saywhat327 invokes both the formal and informal register when posting updates and directing Tweets at other users. In both types of Tweets, however, he employs and w to textually represent laughter. Overview of Examples I have selected three Twitter exchanges in which saywhat327 was a participant and used and/or w to represent laughter. In the first example, saywhat327 uses the variant to represent laughter together with formal Japanese as he builds rapport with and shows respect to another musician, sobasnare. In the second, say-what327 represents laughter using the w variant with informal Japanese during a synchronous exchange of Tweets with ykt42 as he builds rapport through playful criticism about ykt42's restaurant. In my final example, saywhat327 uses both and w to represent laughter in the same Tweet to key the utterance that directly pre- cedes each variant as warm/friendly and playful, respectively, based on the utterance's semantic content, achieving rapport with user kels81 as he replies to her Tweet. For each of these examples, I will base my analysis on the media ideologies that constrain and w-usage in Japanese digital discourse as expressed by the participants in my survey: that is standard for representing laughter to key utterances as warm/friendly and can be used in the formal register, and that w signals a playful key and can be used in the informal register. Each example Tweet is presented as follows: the original Japanese text followed by a word-by-word gloss and English translation. Example 1: Using in the Formal Register The following example consists of an exchange in which saywhat327 is responding to a Tweet that sobasnare, a user whom he follows on Twitter, posted. Per the information listed in his account profile, sobasnare is a male resident of Tokyo who plays percussion in a local band and follows saywhat327 on Twitter. Below, say-what327 Tweets a public link to a YouTube video of Yokose, a percussionist whose work he enjoys. sobasnare responds by uttering a comment about the video and referencing the lyrics to Yokose's song, "Storm". saywhat327 then replies to sobasnare, tweeting about how he once saw Yokose performing live at an event in which he played the beat to a famous contemporary children's song called "The Maru Maru Mori Mori Dance": (0l)[4:lBpm] saywha(327: youtube. com/watch? v=93Vwym... tit31 ¿A/ lite*" youtube. com/watch ?v=93 Vwym.., Yokose name-marker to a ¡treat extent (slang). (02) [7:43pm] sobasnare: youtube.com/watch?v=93Vwym,., That's absolutely Yokose. ^¡b»!«^ «WtiiM With1} I RT Woww! I also nostalgic! Thank you! ST-OR-M object marker} RT @saywhat327 bit.ly/HA6aNb HM 3 ^ • @saywhat327 bit.ly/HABaNb Yokose name-marker to a great extent (slang). (03) [8:13pm] saywhaG27: Woww! Now I'm all nostalgic! Thanks! The S-T-O-R-M/ RT @saywhat327 bit.ly/HA6aMb That's absolutely Yokose. @sobasnare t O XE o -e-r ! ? reinforced maple (English import wordI is catted thing to be (format) question-marker?! @ sobasnare Yeah, that's Yokose for you! Next, 1 had the pleasure of seeing him tap out the beat to the Maru Maru Mori Mori dance! (31) Is the red Canopus drum set the reinforced maple one!? In line 1, saywhat327 gives an evaluation of the YouTube video link he includes in his non-Mention Tweet using slang Japanese, positioning himself as a member of his followers' in-group. sobasnare ratifies saywhat327's positive evaluation in line 2, emphasizing it with exclamation points: Woww! Now I'm all nostalgic!, and aligns himself with saywhat327 by employing the informal register as well. The combination of sobasnare's positive evaluation in the informal register and his use of exclamation points at the end of each utterance shows rapport with saywhat327 at an interactional level. saywhat327 responds to sobasnare in line 3, shifting into the formal register with to be (formal), for Yeah, that's Yokose for you! and to see-CAUS to receive (formal)-PAST for I had the pleasure of seeing him. His use of exclamation points at the end of each utterance mirrors the form of sobasnare's previous utterances, continuing the rapport-building process that began in line 2. In shifting from informal Japanese into the formal register, however, saywhat327 creates distance with sobasnare despite his use of exclamation points.2 Because sobasnare initiated the rapport-building process when he responded to saywhat327's original Tweet with informal Japanese, saywhat327 runs the risk of offending sobasnare by distancing himself through formal language, even if his intent is to convey respect and continue building rapport. To mitigate this, saywhat327 keys Next, I had the pleasure of seeing him tap out the beat to the Maru Maru Mori Mori dance! as warm and friendly by using the standard, formal to represent laughter. In this case, functions as a contextualization cue that draws upon the users' knowledge of the prior text in which kanji characters are associated with formality and the symbol ^ means laughter. Because he has shifted into the formal register, saywhat327 is constrained by this prior text to use the variant instead of w to signal the warm and friendly key in a digital setting. In keying his utterances as warm and friendly by using as a contextualization cue, saywhat327 de-emphasizes the distance created by his shift into the formal register as he simultaneously shows sobasnare respect and continues to build rapport. However, because he is using the formal register, saywhat327 is constrained by the prior text that associates w with informality. Thus, he does not use it here. Example 2: Using w in the Informal Register When saywhat327 uses the informal register, however, w becomes permissible as a contextualization cue. The example below consists of an excerpt of a longer 2 It is certainly possible that sociolinguistic factors such as age and/or social status of the adressee, sobasnare, could influence saywhat327's shift into the formal register here, though it is difficult to speak to the certainty of this potential influence given the limited ethnographic information I could obtain via each user's Twitter profiles. The nature of their online accounts provides information about each respective user in terms of their interests and their professions. However, I cannot deduce either user's specific age, only that they are within the realm of 18-22 as university undergraduate students. Lastly, I know nothing more about their relationship to each other, only that at the time of this study they were following each other's account on Twitter. These ethnographic limitations notwithstanding, it is surely worth noting that sociolinguistic factors such as age and/or social status are certainly probable influences that constrain saywhat327's use of the formal register. exchange between saywhat327 and his friend, ykt42, which is conducted using informal Japanese. According to his Twitter account profile, ykt42 is a male college student who plays the bass and shares saywhat327's interest in music. ykt42 works at a tempura restaurant in Tokyo and follows saywhat327 on Twitter, who follows him in return. saywhat327's interactions with ykt42 are synchronous and largely informal. Prior to this excerpt, saywhat327 responds to an asynchronous Tweet that ykt42 posted about a ramen (thin noodle soup) dish he enjoyed. Over the course of five subsequent synchronous exchanges, the topic eventually shifts to tendon, an item on the menu at the restaurant where ykt42 works. In the exchange below, saywhat327 is complaining about how expensive it is to order tendon, a rice bowl with tempura and sauce on top, according to his preferences: (01) [11:17am] saywhat327: gykt42 Ov«< t T S^fcl^ h> Itttd @ykl42 Donburi NEG set meal by to eat-want emphatic marker to be but &KU\Z III® j*tt ft16 i3— @saywhat32 7 small bovsl to be in a state (informal) because emphatic marker-ELONG t6+ 1 5 0 p3 x wti ttia (j jt> but + 150 money-counter topic-marker expensive emphatic marker ( J U.') (fl£saywhat327 Bccausc it's in a small bowl, you know- but yeah. +150 yen is expensive, isn't it C J JO (03) [11:21am] sa>-what327: @ykt4210*3«?} ttffi V ft^l^ hj %ykt42 Crunchy POSS state of being as to eat-want emphatic marker it tt£" 150 n tf. . . B'J CO Ss [C to be but 150 money-counter however.,. Separate POSS device location-marker @ykt42 1 wranna eat the tempura when they're cnjnchy, but 150 yen... I wonder if it's bad of me to ask for them on a separate dish and stuff ww (04) [11:25am] ykt42: @saywhat327 SO C I IZ $say*'hat327 Separate POSS dish direction-marker contrastive topic-marker an/an it to serve-POT-NEG emphatic marker to be emphatic marker-ELONG »o ~c m^-c ig r «^a ¿Jtrt tempura sauce without as to ask salt by to eat among other things! Laugh ^saywhat327 They can't really serve it on a separate dish, you know-At least ask for the tempura without sauce and use salt or somethingf In line 1, saywhat327 issues a question criticizing the service charge that ykt42's restaurant imposes upon customers when they request that the tempura be served in a separate dish from the rice bowl with but why's it so expensive? in the informal register. Here, saywhat327 does not cue his utterance with any representation of laughter. Despite the fact that he issued the criticism in the form of a question, it can still be considered mildly offensive to publicly criticize ykt42's restaurant on Twitter where other users can view the interaction. Given saywhat327's criticism, ykt42 is caught in an interactional bind that strains his loyalties to saywhat327, a friend, and his employer, the company of which the restaurant where he works is a subsidiary, and of which he is a representative. ykt42 first aligns himself with his employer, giving a reason for the service charge in line 2: Because it's in a small bowl, you know - using informal Japanese. While this utterance serves to explain the reason for the service charge and defends his employer, ykt42 softens his disalignment with saywhat327 by utilizing the elongated sound marker . After doing so, he realigns himself with saywhat327 by affirming saywhat327's original complaint with a repetition that echoes his sentiments: but yeah, +150 yen is expensive, isn't it (Z^). In using the emoticon (/ff), ykt42 solidifies the alignment move he makes in agreeing with saywhat327's complaint and seems to rebuild rapport with his friend. Thus, ykt42 creates an environment in which saywhat327 can continue his criticism of the restaurant's service charge. As we can see in line 3 above, saywhat327 begins his reply to ykt42's Tweet with another complaint in the informal register: I wanna eat the tempura when it's crunchy, but 150 yen.... Again, while his use of ellipses (...) softens his complaint, issuing such a criticism as but 150 yen in a public Twitter domain can still be considered offensive and possibly insulting toward ykt42's employer. To mitigate this, however, saywhat327 issues an alternative to the service charge in the form of a question, followed by laughter which he represents using w: I wonder if it's bad for me to ask for them on a separate dish and stuff ww. Based on the prior text that associates romaji with informality, it follows that say-what327 uses w with an utterance in the informal register. At an interactional level, the w symbol in this example serves as a cue that signals the playful key, which con-textualizes saywhat327's utterances as non-serious criticism and serves to continue building rapport with ykt42. In using the w symbol twice, saywhat327 emphasizes this key, mitigating the distancing effect of his criticism and suggested change to ykt42's restaurant's fee policy. In this case, while saywhat327 could have used the variant as a contextualization cue because it is considered standard, it would not signal the playful key to ykt42. He would only be keying his suggestion as friendly or warm and thus his criticism could be interpreted as friendly, yet legitimate dissatisfaction. Doing so would not serve to mitigate the distancing effect of his criticism, and thus, saywhat327 would not maintain the rapport-building process. In using the w variant, however, saywhat327 signals the playful key to indicate that the interaction is non-serious, which allows ykt42 to engage him in an exchange of playful criticism, as seen in ykt42's response in line 4: At least ask for the tempura on the side and use salt or something! Here, ykt42 issues a response criticism, cueing it with^3 and the rapport-building process continues. Altogether, while formality seems to be a constraint that impacts the appearance of either the or w variant in saywhat327's Tweets they both seem to mitigate distance and build rapport as contextualization cues that key utterances as warm/friendly and playful, respectively. Example 3: Comparing with w in the Informal Register In most of saywhat327's Tweets, (%) and w appear to be constrained by register formality. In example 1, we saw that he used (%) with formal Japanese and in example 2, he employed the w variant with an informal utterance. In this example, however, we will see that saywhat327 uses both variants to key two informal utterances within one Tweet. Below, saywhat327 is responding to an asynchronous Tweet that kels81 posted for public viewing. kels81 is a female Japanese college student who say-what327 follows on Twitter, plays guitar in a local band in Tokyo, and works in a music instrument shop as a packing assistant. The following example begins with kels81's asynchronous Tweet in which she issues a public, general invitation to other users for a Skype chat and a request for music to a girl named Narumi: (01) (5:12am] kelsSl ig« (02) [5:34am] saywhat327: -ï X Skype UiU-; kelsSl ! Sparc time person Skype to do-VOL-ELONG (formal) S kcls8l to be (formal)! iurics^ S-A- sm (*ov.v)o And S'arumi name marker-ELONG soundsource please give me (*ov.v)o Attention all bored people: let's Skypc-JTm kelsSl! And Narumi- please give me a sound source (*ov.v)o m TS. Kfh- (5E) v Playing by car? 3 Though I have not explored this variant in my paper because saywhat327 does not use it, in the media ideologies survey results, % is considered by some to be a contextualization cue that has a different keying effect than (%) and w. The answers provided in the survey pertaining to this variant were staggered in terms of describing how and to what effect %is used. It is worth noting that ykt42 does not use either the (%) or w variant, so while the % variant may or may not necessarily be associated with signaling the playful key for all users, I argue that the prior text associations of % representing laughter serve to maintain a basic key of friendliness here, which allows ykt42 to continue with playful criticism, given the content of his utterance. In line 1, kels81 addresses two audiences in her Tweet: her group of Twitter followers such as saywhat327 (Attention all bored people), and a girl named Narumi whom she addressed by her first name rather than her Twitter screenname (And Narumi-). She issues a different request to each audience using formal Japanese — she invites her group of Twitter followers to join her in a Skype chat (let's Skype-i>) and requests that Narumi provide her with music (please give me a sound source (*ov.v)o). Because saywhat327 is a follower of kels81, he falls within the group of people that kels81 addresses in her invitation to Skype chat and thus, as a recipient of the invitation, issues a reply in which he rejects her offer using informal Japanese: I'm free, but I'm at work now (^). In responding to kels81's invitation written in the formal register with informal Japanese, saywhat327 positions himself as a member of kels81's in-group. In spite of that, however, the content of his utterance is an implicit denial of kels81's original request to chat on Skype - he acknowledges that while he is indeed a member of the group of all bored people in stating I'm free, he follows this with the contrastive marker but to preface his reason for rejecting her offer: I'm at work now without explicitly writing a rejection such as No, I can't. This softens his rejection by issuing a statement that foregrounds his current state of affairs rather than highlighting the fact that he is not accepting her invitation. Yet, in rejecting her offer, he still distances himself from kels81. To mitigate this, say-what327 keys his utterance as warm/friendly with the standard (^). Though it stands to reason that saywhat327 would use w to match the informal register, based on the survey data that deem the standard form of representing laughter online, we can account for saywhat327's use of the variant. In using the standard variant to key his utterance as warm/friendly, he reduces the distancing effect of his rejection and shows rapport with kels81 at the interactional level. saywhat327 follows his rejection with a criticism of her activity at work: If you're just packing guitars, I'd say you're goofing off, huh w. While he also produces this utterance in the informal register, his criticism (I'd say you're goofing off, huh) has a stronger distancing effect at the interactional level than his rejection I'm free, but I'm at work now in the previous utterance. His criticism, therefore, counteracts the rapport building process that he began with previously. To minimize this effect, saywhat327 employs w, keying his criticism as playful thusly reinstating the rapportbuilding process. While he could have used the variant again to maintain continuity across both utterances, his switch to the w variant indicates not only that for saywhat327, the two variants can be used in the informal register but also that they are indeed separate contextualization cues used to key utterances that differ by semantic content. In this example, we can gain a sense of how saywhat327 distinguishes between variants: he uses to key his rejection of kels81's offer as warm/friendly, yet somehow it is not the appropriate cue with which he may key his criticism. The difference between both utterances in saywhat327's Tweet is that the second (I'd say you're goofing off, huh) has a stronger distancing effect than the first (I'm free, but I'm at work now). Thus, we can infer that the warm/friendly key signaled by is not enough to mitigate the distancing effect of a criticism, as it would not give any indication to kels81 that he is criticizing her playfully. Rather, in using the variant, he could signal to kels81 that while he is criticizing her in a warm/friendly way he is still expressing legitimate disapproval which would ultimately lead to creating distance between both users. Therefore, in using the w variant saywhat327 mitigates the stronger distancing effect of his criticism by keying I'd say you're goofing off, huh as playful such that he may foreground a tone of non-seriousness rather than conveying friendly, yet legitimate criticism. Altogether, this is entirely dependent on the notion that w is inherently different from in its function; namely, that it signals the playful key rather than the warm/friendly key. saywhat327's use of both variants in the same register yet in different semantic contexts (rejection and criticism, respectively) indicates that they are indeed different contextualization cues that achieve more than merely representing laughter in text. Rather, both variants are distinct forms that cue different keys - warm/friendly and playful, whereby the neutrality of warmness/friendli-ness is associated with achieving rapport with out-group members and the teasing nature of play is restricted to the arena of in-group camaraderie - and saywhat327 utilizes these differences to appropriately contextualize particular utterances such that he may achieve rapport with the other user. CONCLUSION I have shown that saywhat327 uses in a formal context to signal the warm/friendly key while showing respect, w in an informal context to signal the playful key throughout an exchange of criticisms, and both variants within the same Tweet written in the informal register to signal the warm/friendly key and the playful key for the rejection and criticism that directly precedes each respective variant. Ultimately, while he uses both variants to achieve rapport with the other user, he employs them with particular utterances to signal the appropriate key for contextu-alizing their semantic content. Additional analysis is needed to uncover the particularities of how other users with differing media ideologies utilize and employ these variants in their discourse. Certainly, however, the common thread between say-what327 and other users who harbor different media ideologies about how to represent laughter in digital discourse lies in the fact that they utilize these different representations of laughter in the first place. That fact alone speaks to the idea that users feel compelled to convey contextualization cues in the digital realm to achieve different interactional goals, and do so by manipulating the stylistic aspect of the text that constitutes their exchanges. Whether via changing orthography as we have seen with and w, or opting to use a period-mark or even abstaining from doing so at the end of an utterance, for example, these are conscious decisions that users make in the digital realm, and for many, regardless of their differing media ideologies, such decisions leave a lasting impact on the interaction itself. Acknowledgements No gratitude could possibly express my appreciation toward Takafumi Ohyama, the colleague with whom I have collaborated on this project since its inception. This paper reflects only one part to a much larger mosaic of findings that he and I have crafted in partnership. I am endlessly grateful toward Dr. Ana Zwitter Vitez for her flexibility, support, and guidance as editor while I grappled with the revisions process. Many additional thanks to Deborah Tannen, Anna Marie Trester, Sarah McKeown, Alicia Martinez, and the anonymous reviewers from this edition of Linguistica for their invaluable comments on earlier drafts of this paper. I take responsibility for any inadequacies that remain. References Baron, Naomi (2010) "Discourse structures in instant messaging: The case of utterance breaks." In: S. Herring (ed), Language@Internet 7, 1-32, 10 May 2012. Becker, Alton L. (1995) Beyond Translation: Essays Toward a Modern Philology. Ann Arbor: University of Michegan Press. Gershon, Ilana (2010) The Breakup 2.0: Disconnecting over New Media. Ithaca/London: Cornell University Press. Goffman, Erving (1981) Forms of Talk. Philadelphia: University of Pennsylvania Press. Gumperz, John (1982) Discourse Strategies. Cambridge: Cambridge University Press. Werry, Christopher (1996) "Linguistic and interactional features of Internet Relay Chat." In: S. Herring (ed), Computer-Mediated Communication: Linguistic, Social and Cross-Cultural Perspectives. Amsterdam and Philadelphia: John Benjamins, 47-63. Gloss Conventions POSS = Possessive marker NOM = Nominalizer NEG = Negation PAST = Past tense PROG = Progressive tense POT = Potential tense PASS = Passive tense ELONG = Elongated sound English import word = A borrowed word from English uttered with Japanese pronunciation Abstract SAY WHAT? THE REPRESENTATION OF LAUGHTER AS A CONTEXTUALIZATION CUE IN ONLINE JAPANESE DISCOURSE In this paper, I compare two variants for representing laughter in synchronous Japanese Twitter discourse: the kanji symbol meaning "laughter" situated between two parentheses, and w, the first letter of the romaji transliteration of that same word wara. I argue that they are digital equivalents of what Gumperz (1982) calls contextualization cues and that users employ each variant to convey a pragmatic tone, or what Goffman (1981) refers to as linguistic keys. To demonstrate, I analyze the public Tweets of saywhat327, a male Japanese university student. I find that saywhat327 (1) uses to key his utterances as warm/friendly in the formal register of Japanese, (2) uses w to key his utterances as playful in the informal register of Japanese, and (3) uses both variants with informal Japanese to key two separate utterances within the same Tweet as warm/friendly and playful, respectively, based on the semantic content of each utterance. In each of these cases, I argue that @saywhat327 builds rapport with his interlocutor by keying his utterances via the ('%) and w cues. Keywords: digital discourse, contextualization cues, discourse analysis, interactional sociolin-guistics, orthographic variation. Povzetek SAY WHAT? REPREZENTACIJA SMEHA KOT KLJUČ KONTEKSTUALIZACIJE V JAPONSKEM SPLETNEM DISKURZU V prispevku primerjam dva načina reprezentacije smeha v sodobnem japonskem diskurzu na Twitterju: med dvema oklepajema uporabljena pismenka zapisa kanji, ki pomeni "smeh", in w, ki je v latiničnem zapisu romaji prva črka besede wara. Menim, da ta dva načina zapisovanja predstavljata digitalna ekvivalenta ključev kontekstualizacije (Gumperz 1982) in da z izbiro enega od možnih načinov zapisa uporabnik doseže želeni pragmatični ton, kar Goffman (1981) poimenuje tudi jezikovni ključ. Tezo dokazujem z analizo javnih sporočil na omrežju Twitter uporabnika saywhat327, japonskega študenta moškega spola. Analiza pokaže, da saywhat327 uporablja (1) pismenko za kodiranje tistih izrekov, ki predstavljajo topel oziroma prijazen ton znotraj uradnega jezikovnega registra v japonščini, (2) črko w za kodiranje igrivih izrekov v neformalnem registru in (3) obe varianti pri neformalnem diskurzu, kadar želi v dveh ločenih izjavah istega tweeta ustvariti najprej topel, nato igriv ton, s čimer tvori končno semantično podobo vsake od izjav. Na podlagi zgornje analize menim, da zapisa in w kodirata odnos, ki ga @ saywhat327 pri svojih objavah vzpostavi z naslovnikom. Ključne besede: elektronski diskurz, ključi kontekstualizacije, analiza diskurza, interakcijska sociolingvistika, ortografska variacija.