CALL AND LINGUISTICS 2

Linguistics: Computer-mediated communication and group

journals: expanding the repertoire of participant roles

Zsuzsanna Ittzes Abrams

Abstract

The ability to recognize and adopt a variety of social roles is essential for successful communication. First language research is available on this topic, but very little information is available on participant roles foreign language learners adopt in classroom interactions. The present study investigated the participant roles learners adopted in the two differents writing environments: synchronous computer-mediated-communication (CMC) and pencil-and-paper group journals. The subjects were 46 students of intermediate German, each of whom participated in two CMC and two of group journal assignments over the course of a 15-week semester. The data was coded for participant roles learners adopted. The coding was based on categories established by previous research in discourse analysis, social psychology and sociology in first language research. These categories were then expanded to fit the needs of the language learners participating in this study. The results indicated that while some social roles appeared in both writing contexts (speaker, respondent, scolder, creator of in-group identity), others were only found in CMC (attacker, challenger, supporter and joker). Not only did learners adopt a larger variety of participant roles during CMC than in group jour-nals, these roles were also more interactively negotiated in the CMC environment. # 2001 Elsevier Science Ltd. All rights reserved.

Keywords: Computer-mediated communication; Group journals; Participant roles; Social roles; Learner-to-learner interaction; Classroom research; Written conversation

1. Introduction

Participant roles, which are in many ways culturally guided (D’Andrade and Strauss, 1992) and are definitely interactionally determined through negotiations between or among participants (Harkness et al., 1992; Quinn, 1992; Silverman, 1998) are social personae we adopt in our daily interactions with others. Participant or social roles have been studied by social psychologists (D’Andrade and Strauss, 1992; Smith and Mackie, 2000), sociologists (Goffman, 1974; Larson, 1997) and sociolinguists (Schiffrin, 1994; Tannen, 1989, 1994). These studies, however, are all based on L1 observation; research into the development of participant roles in sec-ond and other languages is not yet available. Yet, the ability to adopt various social roles—for example, supporter, questioner, informant—is necessary for successful interaction with others in any language. It takes skill to be able to use questions alternately to elicit information, to offer sympathy, to joke, to challenge others’ opinions and defend one’s own position, and finding activities which promote the development of this skill in the L2 is an important step in making the L2 learning process more authentic.

Several studies in second language acquisition have already examined the ability of computer-mediated-communication (CMC) to provide learners a forum in which they can produce more language (Kern, 1995; Beauvois, 1998) and more diverse discourse functions (Chun, 1994) than during in-class discussions, and in which they can become the initiators of discourse instead of mere followers of teacher-directed interaction (Cononelos and Oliva, 1993; Beauvois, 1995; Meskill and Ranglova, 2000). The CMC environment has been recognized to be a great equalizer, because of its ability to result in a more evenly distributed amount of participation (e.g. measured in number of words) among interactants than face-to-face discussions, for example (Warschauer, 1996). Such beneficial findings motivated the research focus of the present study, which investigates the interactions of 46 third-semester learners of German in two different writing environments (computer-mediated communication and group journals) to determine what types of social roles learners adopted in each, and whether synchronous CMC (all learners are logged on simul-taneously) might offer opportunities for developing a larger variety of interactional personae—necessary for navigating a larger repertoire of interactional contexts both on- and off-line—than pencil-and-paper group journals.

2. Review of relevant research

2.1. Participant roles in authentic and pedagogical situations

Proficiency in a language, at least in part, means ‘‘knowing not only the lan-guage code, but also what to say to whom, and how to say it appropriately in any given situation’’ (Saville-Troike, 1989, p. 21). Authentic interactions require that speakers create and interpret context-appropriate meaning (with correct linguistic, sociolinguistic and pragmatic features) as well as negotiate the social roles neces-sary for successful interaction (Halliday, 1970; Hymes, 1972; Goffman, 1974; Canale and Swain, 1980; Kachru and Nelson, 1996; Saville-Troike, 1996). These social roles, which are negotiated interactively, depend on such cultural, personal and situational variables as age, gender, social status, and the purpose and location of the interaction (Pica et al., 1991; Pearson and Lee, 1992; Schiffrin, 1994; Tannen, 1994; Tyler, 1995; Freeman and McElhinny, 1996; Saville-Troike, 1996). Inter-actants negotiate various participant roles dynamically throughout the interaction. These roles—for example that of hearer, speaker, questioner, respondent, informant, supporter, joker, persuader, or controller of interaction—allow partici-pants to maintain the interaction, express social distance or submit to others, or to create in- or out-group identity (Goffman, 1974; Saville-Troike, 1989; Schiffrin, 1994; Smith and Mackie, 2000). Often, the social roles learners can develop are limited due either to the teacher’s or the learners’ interpretations of appropriate classroom behavior (Philips, 1983; Goldstein and Conrad, 1990; Moser, 1992; Strasma and Foster, 1992; Reinertsen and Wells, 1993). According to Hatch (1992), Adler (1993), Larson (1997), and Rampton (1997) the institutional scripts for classroom interactions are still primarily based on teacher–student–teacher turn sequences, in which the instructor directs turn-taking and evaluates most or all of the learner’s output, typically for linguistic appropriateness. These scripts limit learners to adopt only the answering role (respondent), and the role of addressee (listener; Schiffrin, 1994). According to Larson, communicative events, which allow learners to co-author interactions, offer interactional contexts for becoming more ‘‘proficient’’ in directing and determining interaction. Such proficiency is essential for helping learners prepare for (either on-line or face-to-face) authentic interactions.

2.2. Collaborative language learning and CMC

Learner-to-learner interactive language learning tasks can improve the develop-ment of communicative competence because they allow learners to initiate, direct, terminate, and repair interactions (Doughty and Pica, 1986; Nunan, 1990; Chun, 1994; Gass and Varonis, 1994; Warschauer, 1997). One form of interactive learning is group writing. It has been used to teach ESL, train teachers, and promote literacy skills in composition courses (O’Sullivan, 1987; Goldstein and Conrad, 1990; Strasma and Foster, 1992; Reinertsen and Wells, 1993; Arredondo and Rucinski,1994). These researchers found that during interactive learner-to-learner writing, students expressed their ideas more freely and extensively in the L2, nego-tiated meaning more effectively, produced more output and more discourse func-tions, and negotiated a more equalized power-distribution than during face-to-face interactions (O’Sullivan, 1987; Goldstein and Conrad, 1990; Strasma and Foster, 1992; Reinertsen and Wells, 1993; Arredondo and Rucinski, 1994; Leppa¨nen and Kalaja, 1995). In group journals learners reported having noticed and learned from their peers in a low-anxiety environment (Kurtz Allaei and Connor, 1990; Strasma and Foster, 1992). Data also showed that learners aligned themselves in various social group constellations, which helped prepare them for a wider ‘‘array of discourse communities’’ by adopting a variety of participant roles (Johnson and Roen, 1989).

Recent studies found that CMC has similarly equalizing effects socially, in addi-tion to enhancing learners’ linguistic development. In CMC sessions (1) learners fit-ted the discussion to their interests, leaving purely teacher determined territory; (2) participated more evenly; and (3) used a larger variety of discourse functions to express more diverse communicative intents than in teacher-directed discourse (Cononelos and Oliva, 1993; Chun, 1994; Beauvois, 1995, 1998; Kern 1995; Leppa¨-nen and Kalaja, 1995). In addition, Meskill and Ranglova (2000) found that when CMC was used to prepare in-class face-to-face discussions, the ensuing conversation was more fluid, better articulated and learners were participating not only as respondents to but as initiators of questions.

According to the constructivist paradigm of L2 pedagogy (Wolff, 1994; Salomon and Almog, 1998; Ru¨schoff, 1999) learning is a collaboratively constructed process in which learners develop their own web of knowledge through active hypothesis building and testing by doing. From this standpoint, learners are not viewed as passive recipients of ready-made systems of knowledge, transmitted to them by the knower, the instructor. If general knowledge (including the learning of an L2) is enhanced through learner-to-learner collaboration, then the development of inter-actional skills and the creation of discourse (both of which require at least one other participant) must also be enhanced by learner-to-learner collaborative tasks. The equalizing effect of CMC may allow learners to co-create participant roles because it is not left up to the instructor to give the next clue or direct the next step in the social discourse. The learners must negotiate—based on continually incoming information from their interlocutors—how to make their point successfully or how to avoid or repair social missteps. The present study examines whether synchronous CMC is more able to provide an interactional forum for learners to develop such skills than a traditional interactive writing task.

3. Methods

3.1. Subjects

Forty-six students (24 women, 22 men) from four sections of third-semester Ger-man¹ at the University of Arizona participated in this study. In this course, students met four times a week to practise the four skills and culture; the main course objec-tive was preparation for authentic interactions with other speakers of German. Each subject participated in both CMC sessions and in both group journal writing events. Both tasks were intended to help learners brain-storm ideas, synthesize information, exchange relevant experiences, support and counter arguments related to the main themes covered during the semester. These written-conversations were meant to help learners prepare for longer reading and writing assignments and in-class discussions. Although not the focus of this study, it was anticipated that learners would be able to transfer the social skills of offering support, challenging, joking, informing and questioning to other interactional contexts, such as face-to-face conversations or on-line discussions with other speakers of German.

3.2. Interactive writing tasks: written conversation

3.2.1. Group journals

Traditionally, learners had written group journals to exchange ideas with their peers in German on topics relevant to the course syllabus. During the semester this research study was conducted, students completed two group journal assignments, the first one during weeks 4–6, the second during weeks 11–13. Each assignment spanned the course of three weeks. Students in each section of third-semester German were randomly placed by their instructor into groups of three or four (depending on class size—18–23 students). Learners wrote anonymously using code names. Each group had a spiral-bound notebook at the reserve desk of the main library;² these could be checked out for two hours (however, students reported writing an average of only 23 min for each entry). Prior to each group journal assignment, learners received a list of five topics based on material in the course syllabus. Learners were asked to write a minimum of 60 words, in at least two non-consecutive entries per assignment.

3.2.2. Synchronous computer mediated communication

Two synchronous CMC activities, collected from two 50-min class periods, pro-vided the rest of the data. After receiving instructions for spelling,³ learners logged on with a code name. The five discussion topics were listed on the left hand side of the students’ screen. Learners could click on any of the five topics they wanted to discuss, stated their opinions about these discussion topics, and responded to other learners’ input in relation to these questions in German. Comments for each ques-tion were stored in a different ‘‘envelope,’’ and only those comments were visible at any given time that pertained to the topic the student selected.

3.3. Discussion topics

For both interactive writing tasks (CMC and group journal), topics were selected based on suggestions made by the instructors and the learners according to the themes covered during the semester. The overlying topics were the same for both the CMC and the group journal tasks, but each pair of prompt questions focused on a slightly different aspect of the overlying topic to avoid redundancy, boredom and

² Learners’ suggested writing the group journal as an in-class activity, but this would have reduced the writing time per student, and would have resulted in some learners waiting while others finished their entries. Another suggestion was to hand out the group journal notebooks to a member of each group instead of keeping them at the library, and when each learner was finished writing an entry, s/he could pass the notebook on to the next group member. This method, however, had been impractical in the past: some learners dropped the class while in possession of the group journal, others did not respect due dates, thus reducing the available time for other group members for writing their entries, or left them no time to write an entry at all.

³ Since the keyboards were not equipped for the German writing system, learners were told at the beginning of each CMC session how to compensate for the Umlaut letters and the letter ‘‘ß;’’ this infor-mation remained on the board throughout the CMC sessions to prevent problems due to orthographic mismatches. Learners indicated no difficulties with this writing system.

possible practice effects (Ruth and Murphy, 1988). Some examples of the matched-question-pair design can be seen in Table 1.

Learners were encouraged to pursue other relevant lines of discussions in both the group journal and the CMC tasks. They were allowed to use dictionaries for all writing tasks. None of the instructors participated in these interactions.

3.4. Data analysis

Learners’ output from both writing contexts was coded for the participant roles based on categories suggested by Goffman (1974), Larson (1995, 1997), Schiffrin (1994), and Smith and Mackie (2000). The data were triangulated a week after the original analysis; the researcher re-analyzed large sets of data to ensure systemacity in coding; the intrarater-reliability was 0.94. In addition, several learners and instructors were asked after coding to reconfirm the researcher’s interpretation of the participant roles; these respondents were asked what they thought the learner was trying to accomplish with particular questions or comments. The results of this reconfirmation supported the researcher’s coding criteria.

Table 1

Examples of parallel discussion topics

Topic Questions in CMI Questions in GJ

Friendship What is the difference, in your opinion,

between a friend and an acquaintance?

Family relationships Should it be allowed for parents with

different skin colors to adopt children?

Why or why not?

Sport and Money Last year there was no baseball for

several months because of the baseball

strike. Many players felt that they were

not getting paid enough. Some people

agreed, while others felt that players

already get paid too much. Which side

of the argument do you support? Why?

Gender In may European countries men and

women can work equally in the

military, for the police or the fire

department. Do men and women

have equal rights in the USA in these

departments? Do you think both

genders can/should have the same

responsibilities and duties?

What is the worst thing a friend could do to you without alienating you for ever? Has this ever happened to you?

What are the most important qualities of a good parent? Why do you consider these personal traits most important?

Many people are angry that athletes get paid so much while the general public has to work very hard and still gets paid relatively little. What is your opinion?

When you are at a party, do you think women should be allowed to approach men or not? Does it depend on the context? What would you do?

4. Findings

Several participant roles were found in both writing contexts: speaker, scolder, respondent, commenter, informant, questioner, creator of in-group identity and knower. Others were found only in the CMC sessions: attacker, challenger, supporter, and joker. Table 2 illustrates the distribution of the different participant roles during the two sessions of each writing context.

4.1. Participant roles found in both writing contexts

4.1.1. Speaker, scolder, respondent and commenter

A typical interaction in both writing tasks began with a subject’s comment as ‘‘speaker’’ when an interactant made a basic statement to initiate a discussion by presenting his or her opinion on a subject matter, eliciting various responses:⁴

(CH) Ich finde homosexuelle Ehen sehr schlecht. Die Ehe sollt zwischen einen Mann und eine Frau sein.

/I think same-sex marriages are wrong. Marriage should be between a man and a woman./

Table 2

Participant roles—frequency of occurrences in each writing session^a

Participant roles		Computer conferencing						Group journals
		Session 1			Session 2			Session 1		Session 2
Present in both	Speaker	55	(23%/5%)		62	(26%/7%)		78	(50%/9%)	85	(52%/8%)
writing contexts	Respondent	23	(10%/2%)		31	(13%/3%)		26	(17%/3%)	23	(14%/2%)
	Scolder	7	(3%/06%)		4	(2%/04%)		1 (06%/01%)			0 (0%/0%)
	Commenter	14		(6%/1%)	16 (7%/2%)				0 (0%/0%)	4	(2%/04%)
	Informant	15		(6%/1%)	20 (8%/2%)			47	(30%/6%)	38	(23%/4%)
	Questioner	22		(9%/2%)	25	(10%/3%)			0 (0%/0%)	5	(3%/05%)
	Creator of	26	(10%/2%)		26	(11%/3%)		3	(2%/04%)	7	(4%/07%)
	in-group identity
	Knower	5	(2%/04%)		7	(3%/07%)		2	(1%/02%)	3	(2%/03%)
Found only	Attacker		0	(0%/0%)	2 (08%/02%)			0 (0% / 0%)		0 (0% / 0%)
in CMI	Challenger	18		(8%/2%)	14 (6%/1%)				0 (0%/0%)		0 (0%/0%)
	Supporter	6	(3%/05%)		4	(2%/04%)			0 (0%/0%)		0 (0%/0%)
	Joker	49	(20%/4%)		32	(13%/3%)			0 (0%/0%)		0 (0%/0%)
Total No. of PRs				240		243			157		165
Total No. of idea units				1120		934			853		1015

^a The first percentages indicate frequency of participant roles distributed over the total number of participant roles within each writing event. The second percentages indicate the frequency of participant roles distributed over the total number of idea-units learners produced in each writing event.

⁴ The learners’ output is presented in the original, unadulterated form; no corrections of grammar or vocabulary were made.

(CT) Re #19: Es ist ein unbeku¨mmert Antwort. /re #19: That is an inconsiderate answer./

(JK) CH, ich kann sich nicht verstehen. Lieber ist wichtig. . .das ist alles. Ich denke das, Gott Lieber haben muss, nicht Hass.

/CH, I can’t understand you. Love is important. . .that is everything. I think that God must have love, not hate./

(MN) Re #19: Kennst du irgendjemand, wer ist homosexuell? /Re #19: Do you know anybody who is homosexual?/

(CH) Re MN: Ja, ich kenne homosexuellen Leute. Ich hasse nicht homosexuellen Leute. Homosexuellen Leute sind nett und freundlich. Aber sie sollen nicht heiraten.

/Re MN: Yes, I know homosexual people. I don’t hate homosexual people. Homosexual people are nice and friendly. But, they should not marry./

The speaker, in the first CMC session, states his opinion about same-sex marriages without including a question to involve the other interactants. In response, CT takes the role of scolder. He disagrees with the speaker and makes a disapproving meta-comment about the value of CH’s statement. The third person takes on the role of commenter by commenting on the speaker’s opinion. As the speaker did not ask a question, JK’s reaction to it does not reflect the role of respondent (i.e. in responding to a question).

The role performed by the fourth interlocutor can be seen as either that of respon-dent or challenger; he either intended to pose a question in response to the speaker’s claims or he challenged the speaker’s opinion by his question, implying that it was not an informed comment. The fifth contribution, by CH, however, suggests that MN’s question was interpreted as a challenge by the original speaker, who reacted to it defensively. After recognizing that his communicative intent was misunderstood, CH attempts to repair the social relationship with the other participants.

4.1.2. Informant and questioner

As mentioned earlier, the typical pattern of interaction began with a speaker’s opin-ion. In many cases, the next interactant provided new, additional factual information to the subject, performing the role of informant (i.e. expanding knowledge). This can be seen in the following example, taken from the first Group journal assignment.

(RW) Ich lese die Journal von [EC]. Ich glaube daß sie ist richtig aber falsch mit einige Dinges. . .Meine Familien, zum beispiel, ist jetzt gut aber in den letzten paar Jahren, Leben war nicht sehr gut. . .

/I am reading EC’s journal. I think she’s right about some things but wrong about others. My family, for example, is fine now but in the last few years, life was not very good./

This learner began his journal entry by acknowledging the previous statement regarding the decline of quality of families (i.e. respondent), and then provided information about his family to underscore his message, taking on the role of infor-mant in the second half of his journal entry.

Besides providing new information, interlocutors often took on the role of ques-tioner in response to a previous contribution. In this role, learners posed a question pertaining to a previous comment to clarify its content, to indicate an interest in the continuation of the discussion or to request directly further information from another interlocutor.

(BB) [CS] hat ein interessant Idee. . .Aber ich verwundere, ‘‘Wie alt sind die Kinder?’’

/[CS] has an interesting idea. . .But I wonder, ‘‘how old are the children?’’/

4.1.3. Creator of in-group identity and knower

For communicative purposes, creator of in-group identity and knower are very similar. However, there are subtle differences between them. Learners take the role of creator of in-group identity to establish the fact that they belong to an exclusive group of individuals who share some aspect of life that is not shared by people in general. The following examples are taken from the second and the first Group journal assignments respectively.

(DC) Ich bin einen Republican so habe ich viele Diskutieren u¨ber Dole und Pra¨-sident Clinton.

/I’m a Republican, so I have a lot to say about Dole and President Clinton/

Here, DC identifies himself as being a Republican, providing the other inter-locutors with a framework with which they can differentiate between his sarcastic and serious statements. A knower also establishes group membership, but explicitly states knowing a fact the other participants do not know. This role is different from that of an informant, who presents factual information about things and ideas without expressing personal affiliation with a special group of people. In the fol-lowing example taken from the second group journal, DG establishes her role as knower (of parenting) immediately at the beginning of her journal entry and expands this role by providing the others with factual information based on her personal experience.

(DG) Wir haben zwei To¨chter (neun und zehn Jahre alt). . .Mit guten Eltern du nie wißt daß Zeit ist ein Opfer. Manchmal Zeit ist ein Opfer. Meine Tochter hat gesagt, ‘‘Mutter ist nie selbstsu¨chtig (selfish).’’ Glaubst du das nicht! Aber sie kann nicht sehen wie mu¨de ich bin.

/We have two daughters (9 and 10 years old). . .Good parents don’t think that time is a sacrifice. Sometimes time is a sacrifice. My daughter said: ‘‘Mother is never selfish.’’ Don’t believe that! But she cannot / should not see how tired I am./

4.2. Participant roles present only in CMC

4.2.1. Attacker and challenger

The first of these participant roles, attacker, was performed by learners when they made personal, insulting comments directed at other interactants. In this example, DC attacks the previous learner, who—during a discussion on good gift ideas— stated that he wanted a prostitute.

(DC) Ist deine Mutter ein Teil davon! /Your mother is one of them!/

In other instances, learners took on the role of challenger, in which a learner explicitly disagrees with the content of another interactant’s statement. Earlier, ‘‘RE’’ expressed the opinion that children with biracial parents would have a diffi-cult time fitting in. TM disagrees with RE’s opinion:

(TM) Quatsch! Was meinst du, dass Kind wird mehr Probleme wenn deine Eltern eine verschiedene Farbe ist? Wenn die Eltern gute Eltern sind, dann die Hautfar-ben sind nicht so wichtig!

/Nonsense! What do you mean the child will have more problems when the parents are biracial? If the parents are good parents, the skin color is not that important!/

Such comments elicited excited commentary from the original writers and other interactants. Both DC and TM tried to re-direct their statements to ensure that communication would not be irreparably damaged. DC said afterwards that he was only kidding, and TM paraphrased his statement without the ‘‘Nonsense!’’ com-ment, explaining further what he had meant in order to soothe any ruffled feathers he may have caused. Several students who visited German chat-rooms reported that often native-speaker interlocutors aggressively rejected and ridiculed them during e-exchanges. Thus, developing skills to recognize, deflect or counter attacks and chal-lenges, defend or explain one’s own opinions is important if the learners wish to partake successfully in authentic German Internet chat rooms.

4.2.2. Supporter

This role is performed when a learner responds to the emotional content of another’s comment, providing encouragement or sympathy, as illustrated by the following two examples:

(CG) Herr Kappellmeister, du sollst sehr priviligiert fu¨llen. /Mr. Kappellmeister, you should feel privileged./

(DG) re #22 Es tut mir leid; es ist traurig, wenn du hast Heimweh und vermisst deiner Familie.

/I am sorry; it is sad when you are homesick and miss your family./

4.2.3. Joker

On several occasions during CMC, students took control of the interaction by interjecting a teasing question, a lighthearted statement, a vulgarity or an off-topic comment. In the role of joker, learners attempted to change the flow of interaction, the focus or the mood of the discussion. In both CMC sessions, in each of the four sections of German 201, and for each of the discussion topics, the flow of interaction was ‘‘disrupted’’ several times by an interjection that attempted to direct the conversation to a different, less serious topic. The following is an example from the first CMC task:

(DH) Ich weiss nicht, ich habe viele Bekannten, aber nur ein bisschen Freunde. Mr. Rogers war mein nachbar.

/I don’t know, I have many acquaintances, but only a few friends. Mr. Rogers was my neighbor/

This comment was made at the beginning of the discussion about friendships and was not a response to any of the previous comments. DH made two more attempts to make this discussion more light-hearted and to create camaraderie with the other interlocutors.

5. Discussion and conclusion

The main purpose of this study was to examine whether the CMC context pro-moted the adoption of a wider array of participant roles—resembling more or pre-paring learners better for authentic interactive situations—than more traditional group journals. As stated earlier, it is essential for learners to develop the skills to co-construct the discourse of their interactions, to take on actively a large variety of participant roles, so they can be successful participants and authors of what happens in the classroom, as opposed to being merely reactors to the instructor’s direction. Classroom discussions that are more evenly distributed, and offer learners better opportunities to practise speaking with different social voices will prepare them better to partake in more realistic interactions outside the classroom as well, where one-sided, responsive conversations are not routine. As the data revealed, while both CMC and group journal elicited more than the traditional roles of listener and respondent (Hatch, 1992; Larson, 1997; Rampton, 1997), in CMC learners used more diverse participant roles than in group journals. The following section will explore possible causes for these differences.

The participant roles attacker, challenger, supporter and joker—which may endanger one’s or others’ sense of self vis-a-vis his/her interlocutors (Schiffrin, 1994)—appeared only in the context of CMC tasks. Some possible reasons why these roles were performed only during CMC are (1) the immediacy of CMC, (2) CMC’s perceived fleeting nature, and (3) anonymity resulting in ‘‘flaming’’and increased risk-taking.

First, the immediacy of CMC might have a liberating effect on the learners. The attacker, challenger, orjoker might feel more comfortable committing a face-threatening act because there is the possibility of immediate repair—the remediation or a reneging of the attack—if one’s attack is taken more seriously than it was intended. Such misreadings of social/linguistic intent group journals might be the result of lack of pragmatic or linguistic proficiency. However, by negotiating the social roles, learners could repair many instances of miscommunication even with limited linguistic abilities. The group journals, on the other hand, took a longer time to co-compose, and would have allowed the attack or challenge to linger without being explained, negotiated and modified, therefore, these more face-threatening roles were avoided.

The fact that the role of the supporter was present only in the CMC context might also be due to the issue of immediacy. When someone expressed sadness or related a personal experience in the group journal tasks, other learners did not react in a personal manner. They did not address each other by name, they used very few second person pronouns, and the topics remained more general with very few per-sonal anecdotes.

The role of joker was also only found in the CMC context, probably also due to the immediacy of CMC interactions, and the notion of shared physical space. In group journals, there was no ‘‘immediate context’’ for the interaction, and the quick turn-taking pattern required for joking could not take place during an assignment lasting three weeks.

The second explanation for the presence of these more aggressive roles could be the seemingly fleeting nature of written responses on the computer screen. Due to the length of learners’ output in the CMC tasks (approximately 68 words per entry), each utterance moved swiftly from the bottom of the screen to the area of the dialog box that could only be reached by a scroller. Therefore, the attack was soon ‘‘invi-sible,’’ and as such, was not considered as great a transgression of a sense of self as it would have been were it more permanently visible like in the group journals.

Thirdly, group journal tasks were written in smaller groups, in which group dynamics may have reflected a more intimate interactional environment. Even though learners did not know their group mates’ names, perhaps fear of recognition of one’s handwriting or the perceived intimacy of a smaller group prevented the use of extreme emotions. CMC has been shown to encourage negative, often aggressive comments known as ‘‘flaming’’ (Sproull and Kiesler, 1991) due to a lack of fear of recognition, and the ability to act out other personae than one’s own, since there is no personal consequence to making social and interactional transgressions hidden behind the screen.

These findings indicate that CMC was more able than group journals to elicit a larger variety of participant roles which they collaboratively negotiated. Instead of self-determining the participant roles, learners probed, reneged, and adjusted their social roles in cooperation with the other interactants. In both instances of CMC, learners took on the roles of speaker, informant and knower 31% (CMC1) and 37% (CMC2) of the time. These roles are fairly non-interactive. In other words you can speak, inform and know without having other people reacting or especially inter-acting with you. The remaining 69 and 63% of the time, learners adopted roles that require social collaboration, such as respondent, scolder, commenter, questioner, creator of in-group identity, attacker, challenger, supporter and joker. In complete contrast to this, the group journal context elicited the not necessarily interactive roles of speaker, informant and knower 81% (group journal 1) and 77% (group journal 2) of the time, while the requisitely interactive roles were adopted only 20 and 23% of the time.

Wildner-Bassett (1990) and Harder (1980) report that L2 learners often feel that they are inadequate social entities because they cannot express the same range and

variety of emotions and opinions in the L2 as in their native language, due to lim-itations in linguistic ability and the education context (perceived or actual tradi-tional teacher-learner roles). While it still remains to be seen whether CMC is able to prepare learners for oral communication, this research suggests that CMC—at least in the written modality—can help learners (1) adopt a wider range of participant roles, and (2) develop the skills to view interaction as a process in which they con-stantly have to negotiate collaboratively both their own and other interactants’ participant roles.