Language Learning & Technology
Vol. 3, No. 1, July 1999, pp. 75-87


Raffaella Negretti
Università di Verona


Different Internet technologies foster the acquisition of different language skills. In the case of synchronous interaction tools, such as Webchat, the concern is to evaluate whether and how this communication context affects the process of acquiring a second langauge. A collection of Webchat interaction data among English non-native speakers (NNS) and native speakers (NS) is the basis for a microanalytic investigation conducted form a Conversation Analysis (CA) perspective. The main purpose is to discover patterns and conversational strategies used by participatns in this on-line context.

A CA research approach was chosen since it investigates the machinery and the structure of social action in language, avoiding preformulated theoretic categories. This is important since CMC represents a new SLA context, forcing both NS and NNS to produce different structures and strategies. The study analyzes, in particular, whether Webchat implies a reduction of the range in interactional practices, actions performance, sense making, and meaning negotiation, thus affecting the SLA process. Finally, the researcher considers the reliability and validity of this type of qualitative research in this new technological area.

Using some research methodologies taken from CA literature, an analysis of the data focuses first on the overall structure of interaction and sequence organization in connection with the on-line communicaiton setting features. It then passes to turn-taking organization, with attention to recurrent structures and patterns as in openings and closings; turn design (or packaging of actions); expression of parlinguistic features in this on-line context; and some (interlanguage) pragmatic variables. The conclusion resolves the findings and underlines NNS versus NS behaviour, offering hypotheses about SLA through Webchat and synchronous CMC in general, encouraging further investigation.


Interest in Internet technologies for communication and education has recently increased. There are differences in the types of technologies and the types of language skills and communication competencies that each of them requires. Different modes include synchronous and asynchronous interaction e-mail, Webchat, MOOs, IRC-multimedia activities, Web-based reading, and task-oriented activities. Each mode fosters a different kind of linguistic competence, and calls for different skills. In the case of tools for synchronous interaction it is interesting to analyze whether the communication context is linguistically relevant and how it affects the process of acquiring a second language.


This study takes a Conversation Analysis (CA) approach to analyze Webchat interaction among native speakers of English and ESL students. The Web site used for this purpose is Dave Sperling's ESL Webchat (, expressly created in order to put ESL learners in touch with native speakers. The subjects were foreign language university graduates without any previous computer skills who were enrolled in a postgraduate course in Foreign Tourism and Marketing.

Why Conversation Analysis (CA)?

According to Psathas (1995), the basic assumption of CA is that "social actions are meaningful for those who produce them and they have a natural organization that can be discovered and analyzed by close examination. Its interest is in finding the machinery, the rules, the structures that produce that orderliness" (p.2). In particular, CA implies the avoidance of preformulated theoretical categories. According to Psathas (1995), one of the main concepts on which CA is grounded is that "orderliness" is produced by the parties in situ; thus it is situated and occasioned (p. 3). This is very important since the context of interaction--Webchat--constitutes a new and different communicative context which forces both native (NS) and nonnative speakers (NNS) to produce new and different structures and patterns. Furthermore, the adoption of a CA perspective allows the researcher to approach the data without preconceived theories, free to discover, describe, and analyze the conversation and SLA peculiarities in this context, in other words, to study how social actions are organized and locally produced, in the here and now.

Given the present state of SLA research in Internet-based environments and computer mediated communication, a heuristic-inductive approach such as CA is the most useful and fruitful because such a hypothesis-generating method is a good way to begin the study of new interaction/acquisition situations. A qualitative approach can facilitate a preliminary understanding of broad new perspectives that Internet technologies open to SLA and communication. Since it does not establish research questions a priori, any variable of the context may become the focus of investigation even though qualitative research methods are more frequently used in SLA research today. According to Seliger and Shohamy (1989, p.119), this is due either to the fact that it is not easy to apply to classroom learning (the main SLA context) the controls necessary for good experimental research, and to the fact that there has been a growing awareness of distorting effects of the research settings (Tarone,1982). On the other hand, the development of rigorous methods for conducting qualitative research and collecting SLA data allows good results and good reliability and validity:

The ultimate goal of qualitative research is to discover phenomena such as patterns of second language behavior not previously described and to understand these phenomena from the perspective of the participants in the activity. (Seliger & Shohamy, 1989, p.120)

Before explaining why a CA approach was adopted for this investigation of SLA in Webchat, it is important to summarize the main points in CA theory, especially those on which this study is based. CA originated as an approach to the study of the social organization of everyday conduct. The term conversation should not mislead the reader. It includes both informal and institutional talk. Therefore, it is particularly suitable to investigations focusing on the communication that takes place in institutional settings, such as the classroom. It includes verbal and paralinguistic features of talk that play an important role in Webchat.


In the first place, CA focuses on how individuals in social setting engage in meaningful acts through language and make sense of the world around them. In this view, talk is seen as a social action. This is the reason the terms act or action are frequently used in analyzing a conversation through a CA perspective: language can be used to engage in social actions, and this implies a concern not only for the talk itself, but also for the context in which it takes place. In the case of Webchat, where the context is almost entirely new and unknown to the participants, a CA approach could be helpful in analyzing the different ways in which interlocutors conduct social actions and create meaning through talk. As mentioned earlier, the significant difference between CA and other methodologies concerned with social behavior, such as ethnography or discourse analysis, is that it makes no a priori assumptions. According to Pomertantz and Fehr (1997), "it rejects the use of investigator-stipulated theoretical and conceptual definitions of research questions" (p.66). The questions arise from the data itself and its particularities. In a new communications field, such as Webchat, this free-mindedness helps capture all the peculiarities within the talk and within the context.

Another important characteristic of the CA approach is its focus on details of temporal organization, and the unfolding development of action in interaction. This is extremely helpful in the investigation of Webchat interaction, since as will be shown in greater detail later, the temporal organization of talk in this technological setting is very different from both oral and written talk. It has peculiar modes and structures that call for a redefinition of all strategies for making actions in developing time.

Finally, CA does not try to explain human conduct theoretically, but rather treats rules as situationally invoked standards (Pomerantz & Fehr, 1997, p. 67), by studying social life in situ (Psathas, 1995, p. 1). This approach leads to a deeper insight into the rules and standards of a relatively new environment in which speakers are forced to establish new and alternative strategies for communicating with each other. Regarding the acquisition of a second language through Webchat, the researcher can identify not only what kind of situationally-created rules were adopted by the interlocutors, but also how NNS responded to them, that is, what communicative competencies, devices, strategies, turn-taking, and utterance structures they adopted to communicate effectively in an unusual context.

The main goal of CA is to explain the shared methods speakers use in order to produce and recognize their own and other people's conduct. In this sense, it attaches a great deal of importance to the concept of context, intended both as the temporal organization of utterances that create actions in developing time, and as contingent features such as place, medium, and speakers, but only if these features are shown as relevant to the speakers themselves. Relevance is another important concept in CA theory. Its characteristics are not considered given and worthy of investigation unless the participants explicitly demonstrate through their utterances awareness of these characteristics, that is, unless they are "brought into being by the actions people produce" (Pomerantz & Fehr, 1997, p.70).

It should be clear at this point why the term action is so frequently used in CA. It deals with talk or any interaction as a development of actions which have a social meaning and an inner structure created in situ by the participants themselves in response to contextual contingencies. Sacks (1972) was one of the first to point out that the relationships of individuals in society were visibly demonstrated through their talk, (p.72). Along with other conversation analysts such as Schegloff (1974), Sacks formulated a theory that avoids conceptual theorization in favor of the discovery and analysis of the natural organization of social actions.


In the case of SLA through Webchat, a CA approach does not lead to a generalization about language learning, but rather to the discovery of how non-native speakers produce L2 in this environment: which L2 structures, rules, and practices they adopt or sometimes create in order to effectively communicate in a context that forces them to rearrange their linguistic knowledge.



Webchat was used during a period of one month. However, it was not used systematically due to organizational and structural problems. It was introduced as one of the educational resources in ESL and Internet classes, where the goal was to teach both English and Internet/computer skills using different approaches and technological tools. Classes took place in the computer lab. A total of almost four hours of conversation was collected (see Appendices A, B, C, and D). Students were not monitored during their interaction, but they were asked from time to time to print out their conversations. Sperling's ESL Chat Central was chosen because its explicit purpose is to facilitate contact among native and non-native speakers of English. It requires prior registration via email, adopts a typical frames structure, and allows participants to choose among a variety of options such as number of postings displayed on the screen, insertion of icons in the message, and one-to-one private communication.


The ESL students were eight graduates in Foreign Languages and Literature from Università Cattolica del Sacro Cuore. Their proficiency level was advanced in reading and writing, and intermediate in speaking and listening. Although these students had studied English for some time, they experienced many difficulties in managing their linguistic knowledge for communicative purposes. Their lack of practice in listening and their unfamiliarity with the most common structures and expressions of oral English were evident from the beginning of classes. The use of Webchat under the supervision of a tutor improved their proficiency so that at the end of the two months they could handle a conversation with native speakers quite easily, and were excited about their progress. The great motivation and involvement demonstrated by the students prompted the present research .

In addition to students from Catholic University, native and nonnative students from other parts of the world also joined the chat. Unfortunately, the only way of inferring which of these other interactors were native or non-native speakers of English was through the information they provided about themselves, such as place of birth, country of residence, education, studies, and explicit reference to their proficiency level. There was a total of 36 participants. Of those who had posted messages to the chat, 17 were NS of English connecting mainly from the US and Australia, and 19 were NNS connecting from Brazil, Germany, Greece, Italy, Japan, Korea, Taiwan, and Spain. Interactions were recorded over four days, for a total of 2 hours and 58 minutes of conversation which included 392 turns (see Appendices A, B, C, and D). Each posting was time-stamped at the moment it was sent in order to facilitate a precise reconstruction of time sequence and enhance the reliability of the data.


The analysis was conducted using some of the tools described by Heritage (1997) with regard to CA in institutional talk. His analysis focuses on how institutional realities are evoked, transformed, and even manipulated during interaction. The assumption is that through interaction, the institutional setting imperatives are evidenced and made real to the participants. This approach can be very useful since Webchat is a conversational setting whose restrictions on the range of linguistic and strategic choices available might affect the interlocutors' speech. In this case, the goal was language learning since interaction took place during EFL classes. From the point of view of SLA, it is important to consider whether students were aware of these restrictions since one of the differences between acquisition and learning is awareness of the process. As will be shown later, NNS showed great awareness of the rules and restrictions imposed by the technological setting. They were less aware of the institutional setting with regard to communication goals and the topics of the informal talk. This is the reason why the present study deals with SLA in Webchat and not with second language learning.

The approach taken by Heritage (1997) to CA microanalysis was especially useful for this study, in particular the analysis of the overall structure of interaction, turn-taking organization, and lexical choice. In Webchat, all these features impose heavy restrictions on the range of structures, sequences, and selection of terms speakers are accustomed to in oral interaction. At the same time, these restrictions allow the creation of new standards for communication which take advantage of the technological features of Webchat.

The present study also employs the tools of CA analysis developed by Pomerantz and Fehr (1997). These focus mainly on the actions performed by the participants in a sequence that is context itself. The study focuses, in particular, on the relation between the context of Webchat and the way participants chose to package actions, that is, their selection of terms, turn design, and linguistic structures in order to express certain meanings. It is important to keep in mind that this packaging process is mostly non-deliberate. Timing proved to be a revealing feature as well. In Webchat, timing presents a double challenge since it is largely independent of the participants' choices. Turns cannot overlap since they are displayed in a vertical sequence, and speakers don't have a chance to negotiate when to start, finish, or give a turn as they would be able to do in face-to-face interaction. Onscreen interaction develops non-sequentially as a parallel structure in which different strands of conversation alternate. This complexity of sequencing and timing is peculiar to the context and, as will be demonstrated, is competently managed by the interlocutors. NNS demonstrated not only awareness of these peculiarities of Webchat, but also an increasing competence in exploiting these structures. In terms of SLA, it could be said that they had acquired certain communication skills. This question should be investigated further in order to find out whether these skills are limited by, strictly linked to, and influenced by the technological context, or are of a more general conversational type.

Some General Questions

Some general questions raised by Pomerantz and Fehr (1997) led to the analysis of this conversational environment in order to discover its mechanisms and SLA characteristics:

    1. Does Webchat imply a reduction or modification of the range of interactional practices that participants would use in a "normal" interaction? And if so, how is this manifested in the talk? Does the context of Webchat affect the way speakers "package actions" (p. 72). According to Heritage (1997), context has an important role in shaping the structure of the interaction, in the sense that participants create meanings and use strategies of sense-making through a sequence of actions in talk which constitute the context itself. Thus context is built, invoked, and managed during the interaction.
    2. How do the speakers negotiate meanings and create a sequential context in a technological setting that is different from oral communication? In particular, what typical structures, recurrent patterns, and conversational mechanisms are used by native and nonnative speakers in this setting?


This paper explores whether the context characteristics in Webchat are perceived as relevant and real by the speakers. It examines their utterances in order to see how this relevance influences the acquisition of a second language.

Reliability and Validity Issues

According to Peräkylä (1997, p. 206) reliability and validity present the serious challenge for CA research. In order to satisfy the requirements of reliability and validity the transcript must include as much information as possible about the conversation, such as prosodic features, events, intonation, and all other relevant features of the interaction.

Providing a reliable transcript was easy in the case of this study since it was possible to record the conversations exactly as they developed on the screen and as they were seen by the interactors. The fact that there were no visual or aural cues to infer meaning was both a challenge and an advantage. On the one hand, data important for the analysis of verbal interaction was missing. On the other hand, the participants didnÕt have access to these channels of communication.

As far as validity is concerned, Peräkylä stated that since CA is mainly concerned with a "naturalistic" description and representation, there is less space for the researcher's own interpretations and thus a greater possibility to attain validity (1997, p. 212). Seliger and Shohamy (1989, p.104) recommend the following criteria to control for validity in heuristic research:

    1. representativeness, or the degree to which the observed data represent normal behavior (similar to Peräkylä's naturalistic description);
    2. retrievability, or accessibility of electronically stored data for future research; and
    3. confirmability, or an attempt on the part of the researcher to support each assertion or description with many instances from the data.

As far as representativeness is concerned, the researcher was not present to observe the interaction since the participants were required to print out the interaction themselves. Unfortunately, at times they became so involved in the conversation that they forgot to do so. This resulted in a loss of portions of the conversation, and at the same time, boosted the reliability of the data because the transcript represented a very spontaneous and naturalistic interaction. As Tarone (1983) and Ellis (1985) point out, such spontaneous speech, or vernacular style, is particularly important when analyzing SLA processes because it is the only style in which the normal pattern of language is not disturbed as the speakers do not resort to strategies typical of less naturalistic and more controlled situations. Not much needs to be added as far as retrievability is concerned since the transcripts were obtained by mechanically printing out the computer screens, which makes the data easily available in the original form for further investigation.



Overall Structure of Interaction and Sequencing

One of the main differences between Webchat and face-to-face interaction is that the communication in Webchat is disrupted and discontinuous since many different topic strands and interactions can be carried out simultaneously, often by the same participant. A participant can receive multiple responses to different previous turns and use the same turn to simultaneously post several messages contributing to different strands. In this sense, Webchat turns are not sequential and do not adhere to the time pattern of adjacent pairs and sequences as described by Schegloff (1968, 1979). In analyzing phone conversations, he found some recurrent sequences of adjacent turns, such as in greetings, openings, and closings. He discovered that what the speakers do during their turn is related to what their interlocutors did in the turn that immediately preceded it. For example, in what he called summons-answer , there is an initial turn in which a speaker calls for his or her interlocutorÕs attention, followed by a turn in which the interlocutor indicates comprehension and ability to respond sequences (Schegloff,1968). It is interesting to point out that while such adjacent pairs were frequently found in phone conversations, they were quite infrequent in Webchat. Most of the time, the response to a turn was delayed and many adjacent pairs were intermingled temporally, which disrupted the flow typical of oral talk.

This temporal characteristic of Webchat requires a great deal of motivation and commitment on the part of the speakers, since the expectation of a relevant reply to each turn is often unfulfilled or delayed. Thus, L2 learners face a double challenge: they must acquire the typical structures and sequences of the foreign language and then adapt them to this new communicative setting with its contingent standards and rules.

As mentioned earlier, the overall structure of the interaction in Webchat is split into many topic strands. In many exchanges, it exhibits a parallel architecture, with speakers simultaneously contributing to different strands of the conversation. This kind of structure can be seen on Day 1, Moves 36-96 featuring simultaneously (a) the public interaction between Honk and Carmilla who are engaged in exchanging personal information; (b) Mariposita and Nightsky's conversation; (c) Naghib's incursions with his soliloquy in an attempt to attract attention and establish contact; (d) Leonito's opening and greeting reply by Mariposita and Fede; and (e) the beginning of a conversation between Arosio and Powpow in which the latter expresses his confidence in being able to handle the medium of communication.

Due to the complexity of these multiple exchanges a new participants does not always succeed in joining the on-going talk. For instance, Rufus enters the chat in Move 71 and tries to connect in Moves 76 and 80. But since no one pays attention to his attempts to contact other participants, he soon loses interest and leaves the chat in Move 88.

This already complex structure is further complicated by the fact that in this Webchat, participants can interact in a one-to-one mode, that is, exchange messages without the other speakers seeing them displayed on the screen. This mode of interaction is usually chosen when interactors perceive the need for more intimate communication, especially when exchanging personal information about themselves. Moreover, participants exploit this feature when they feel that the topic under discussion is particularly important and when they do not want to be interrupted by other speakers' postings. By excluding other interlocutors, they can make their exchange flow more smoothly. From a pragmatic point of view, this strategy reduces the distance between the two interlocutors who are chatting one-to-one, creating a sense of confidentiality that is also reflected in the choice of a more informal register and the use of icons.

It is important to keep in mind that NNS use this mechanism of Webchat interaction quite frequently and almost always with the appropriate language features. The influence of the context is quite evident since NNS, even when interacting one-to-one, do it in the L2 rather than L1. This shows that they perceive the setting-situated English-only rule as very strong and real, in a way they probably wouldn't if they were in a foreign country and had to engage in oral face-to-face communication.


An interesting example of this strategy, and also an instance of how complex the conversational structure may become, is demonstrated by the analysis of the data from Day 2, Moves 110-122, where many one-to-one exchanges cross the shared flow of conversation, creating parallel and interfering strands. In order to retrace the organization at this point, it is necessary to use a parallel numeration of pages (e.g., 26, 27, 27a, 27b, 27c) and of moves (e.g., 114, 114a, 114b, 114c). The pattern can be seen in a one-to-one chart. The left column shows the precise time a turn was sent and its sequential position. The top row lists all participants selected for frequent use of the one-to-one form of interaction. It is clear that interaction sequences in Webchat intermingle and overlap, in the sense that what appears on the screen in the time line does not coincide with the unique sequence of turns typical of oral talk. In Webchat, conversational pairs are disrupted and the response to a turn may be displayed after turns are posted by other participants. This becomes more complex when many one-to-one conversations take place simultaneously, so that almost every participant sees on the screen not only the general conversation but also his or her private conversation sequence.

Another important aspect of Webchat is the organization of the sequence of actions, rather than of turns. According to Pomerantz and Fehr (1997, p.69), it is context itself that is relevant from a CA perspective. Although several conversational strands usually take place in parallel, one can find an occasional uninterrupted sequence of actions with two or more participants contributing to turns in a discontinuous fashion. This is usually indicative of their interest in the topic. For instance, on Day 2, Arosio and Tomty start exchanging personal information, getting to know each other, and posting it to the public room. Their conversation attracts the attention of other participants, for instance Adelaide, who become involved in the sequence as well (Moves 3, 4, 6, 7, 9, 10, 12, 17, 20, 21, 23 in the one-to-one chart).

This shows that in Webchat, there can be a single sequence of actions that serves as a general context, if the topic is of interest to all participants. However, multiple parallel sequences with different interaction contexts are more common.

Turn-Taking in Webchat

The organization of turn-taking in Webchat is highly complex and is governed by specific patterns and rules. Since there is no smooth sequential order, interlocutors are forced to manage turn-taking and turn-giving in ways that are different from oral talk.

In studying turn-taking in oral interaction, Sacks, Schegloff, and Jefferson (1974) found that there are some systematic features through which interactors organize their turns. For instance, transitions occur quite smoothly and at relevant points in the talk and overlap is brief and preferably avoided. In general, the turn-taking system is both context sensitive and context free in the sense that it is influenced by the context of the interaction, including the immediately preceding talk, restrictions on the channels of communication, topic, speakers, and time.

The present data show these features are also maintained in Webchat, albeit in a different way. First, the organization of turns in Webchat is highly context-sensitive. Turns are very rarely displayed sequentially, and interlocutors are forced to mentally follow the logical sequence of the different strands of interaction, relying on the name of the speakers and the content of their turns. There are many examples of disrupted sequences.


The transcript of Day 1 shows that Zorba and Nightsky's conversation follows a pattern with long time delays, with the interactors participating in other interactions. For instance, Zorba addresses Nightsky in Turn 19, then the sequence continues in Turn 22 (where Nightsky, replying, uses a very characteristic Webchat strategy of internal turn organization), then continues again at Turn 27, 30, 34. At the same time, other ongoing sequences of interaction are crossing cross each other, disrupting the pattern of turn-organization and making it difficult to understand who is responding to whom. However, NNS demonstrate their awareness of the disrupted sequences and ability to handle the strategies of cross-posting and turn-delaying.

Moreover, it is important to keep in mind that what is displayed on the screen does not necessarily follow the logical sequence of interaction between two participants: sometimes the turns overlap because one of the them sends a new posting without waiting for a response from the other.

Basic Sequences

In discussing openings and closings, the first feature of Webchat that comes to mind is the great amount of interaction devoted by participants to these social functions (Pomerantz & Fehr, 1997, 66) Although the technological environment of Webchat causes a great deal of disruption in the interaction sequences, some basic sequences can be traced to reconnecting turns. Sacks (1979) identified some basic units of speech such as pairs and adjacent pairs which have functions similar to the social actions of Schegloff (1968, 1979). Openings and closings, in particular, follow what he termed summons-answer sequences in which the first opening turn calls for attention and response from the interlocutor; then the third and fourth turns carry on this opening/connecting/introducing function. Webchat contains these basic sequences too, but they display special features imposed by the technological setting either in the way meanings are conveyed, or due to temporal and spatial restrictions.

Openings In opening sequences, participants introduce themselves and make themselves visible by establishing contact with the rest of the room. In observing identification and recognition sequences in telephone conversation openings, Schegloff' (1979) found that interlocutors usually prefer a less explicit strategy, without overtly calling for attention or asking for a reply by the other participants.

Here are some examples. On Day 1, in Move 68, Powpow opens the interaction by greeting everyone in a friendly and informal way. When he does not receive any feedback, he resorts to a more explicit way of calling for attention by using icons and symbols that arouse curiosity in the other participants (see Move 73). In Move 79 the opening sequence is closed by ArosioÕs reply. At the same time, Rufus does not succeed in getting acknowledged despite two opening turns that call for attention more and more insistently (see Moves 71, 76, and 80).

Schegloff (1968) noted that self-identification usually takes place not in the first turn of the opening sequence but in the second one, or in response to a question by one of the interlocutors. In Webchat, identification is more of a self-introduction aimed at having oneÕs presence acknowledged by the other participants. For example, on Day 2 in Move 40, Twitty enters the conversation (first turn of the opening sequence) and receives an immediate reply in Moves 41 and 45 which represent a very rare case of spatially adjacent pairs. Then, in Move 47, the opening sequence goes on with recognition and self-identification by Twitty.

Thus, Schegloff's findings hold true for Webchat since identification (in Webchat, public acknowledgment) is worked out in interactional sequences.


Greetings Speakers perform the social action of greeting in two main ways: They either post different messages to each person in the room, or more frequently greet the entire room using more general structures and lexical items. This strategy is a less personal way of packaging the action of greeting, but it is also linguistically, a more economical way of performing the action. The latter strategy is very typical in the Webchat context, where it is important to convey a maximum of information in the shortest way possible.

Closings In a fairly typical pre-closing/closing sequence, the pre-closing statement usually solicits farewells from the other speakers, while the closing statement calls for a definitive goodbye from the participant who is about to leave. In many cases, pre-closings present some kind of justification or reason why the speaker is leaving, as if he or she were trying to avoid conveying a negative feeling such as boredom or disinterest to the room (Ping in move 136 and "747-400" in move 147). Sometimes, pre-closing and closing take place in the same turn, as in ZorbaÕs Move 31 on Day 1.

In Webchat openings, participants typically use explicit strategies that convey their attitudes such as inserting quotation marks or little icons into the text. This is a very compact and economical way of conveying multiple meanings. It also demonstrates the usersÕ awareness of various Webchat features and possibilities. Therefore, the answer to the earlier question about whether the packaging of actions is influenced and linked to the Webchat setting, is the affirmative. Participants use the formulas and formats typical of, and possible only in, this technological setting.

Turn Design. In this study, speakers used two recurrent internal turn-taking mechanisms to manage the complex interaction in Webchat. The first was a strategy of turn-giving. It is related to speaker selection (one of the turn-allocation techniques described by Sacks, Schegloff, & Jefferson, 1974) The techniques of current speaker selects next and self-selection minimize overlap in oral interaction, but in Webchat, where overlap is not possible, they have the function of reconnecting the sequence of turns, by making explicit who is addressing whom. Although the technique of speaker selection is not common in face-to-face interaction, the present data include hundreds of turns in which the interlocutor's name is explicitly stated, so that both the room and the interlocutor know to which interaction strand the participant is posting (see Moves 3, 5, 7, 103, 104, 105). This confirms Psathas' (1995, p. 38) observation that the system is self-organizing and produced in situ by the interaction parties according to the needs and restrictions of the setting

This strategy is often combined with another internal turn-taking features unique to Webchat which are frequently used by participants to address different interlocutors in the same turn, as in Powpow (Moves 16, 22, 30). It should be noted that this strategy is appropriately used not only by NS but also by NNS.

This internal turn-taking organization not only conforms to Grice's (1975) principle of least effort but is also highly functional from a social point of view. It is not only a practical and convenient way of organizing a conversation, but it is also a technique for combining two different social actions in the same turn in a way that would be impossible in face-to-face interaction. For example, Nightsky uses it on Day 1 in Move 30 in order to perform the social action of closing ("Hi Tis"--mistakenly using "hi" instead of "bye") and to respond to Zorba in a way that further enhances their intimacy and familiarity through the use of icons. Like the speaker selection strategy, the combining of several functions in one turn was also exploited by the NNS from the very start.

Cohesion An important social action typical of Webchat is the creation of cohesion and connection among the participants. Speakers often post messages to the entire room instead of addressing a particular interlocutor. This strategy is used by participants mainly in openings and closings, but sometimes also to serve the social functions of attracting attention and establishing contact as in Powpow's turn 73, Mariposita's turn 78, and Rufus' turn 80. Many moves are devoted to what can be called the fatic function of speech, that is, verifying the connection with the interlocutor(s), establishing communication, attracting attention, and showing desire to interact.


Paralinguistic Features

In face-to-face interaction, interlocutors use paralinguistic features such as intonation, intonation, gestures, body posture, and eye contact, as a communicative strategy. However, Webchat restricts the range of available paralinguistic devices. Many moves in the present data demonstrate the use of alternative devices to convey the semantic load usually carried by paralinguistic features in oral speech.

Among the most frequently used devices are capital letters, emoticons, onomatopoeia, punctuation, little icons, and embedded pictures available from the Webchat menu. Both NS and NNS frequently used the last feature which is unique to Webchat.

Uppercase Letters and Emoticons Uppercase letters are used to indicate loudness of speech. They serve as a visual substitute for an aural clue to attract the attention of other speakers. For instance, on Day 1, in Move 9 Carmilla first uses lowercase and then shifts to uppercase letters to signal a switch to a different interlocutor when she was using the double-addressee turn pattern described earlier.

Emoticons are used to substitute for visual cues such as facial expressions and eye contact. They normally convey a positive attitude or give a particular shade of meaning to the content of the message, such as irony or amusement. There are many instances of the use of emoticons in the present data. For instance, on Day 1 in Moves 5 and 11, Powpow attempts to create a sense of familiarity and participation by using " :) ."

It is interesting to note that NNS relied exclusively on the uppercase and never used emoticons. By contrast, emoticons were very popular with NS. This difference in choice of communicative strategies was probably due also to the fact that Italian-speaking EFL learners had scant acquaintance with e-mail and CMC, whereas the English-speaking participants came from countries where network technologies for communication are fairly common.

Onomatopoeia Onomatopoetic devices were exploited in the Webchat as well. Onomatopoeia represents aural cues that typically present in face-to-face interaction, and are traditionally used in writing to convey feelings and shades of meaning. The present data contain many instances of these communicative devices (e.g., Day 1, Move 16, "Oh hey!"--to attract attention; Move 34, "Beurk! Ha!!!!!!!"--no comment; Moves 58 and 69,"Oh!" and "Ah!"--interjections; Move 104, "Hiiiiiiii,"--long vowel; Moves 142, 159, and 174, "zzz. . .zzz. . ."--snoring sound).

Punctuation The Webchat data show extensive use of punctuation such as question and exclamation marks, dots, and commas by the participants. They used punctuation for multiple purposes, ranging from indicating prosody and intonation contours, to semantic shades and implicatures. In particular, exclamation marks, whose function is to indicate pitch and intonation, were used more frequently than in normal writing. Their use was mainly associated with assertions, expressions of surprise, and happiness--remarks which in other contexts of interaction would be conveyed through strategies of sense-making such as eye contact, gestures, and facial expressions. For instance, On Day 1 in Move 16, Yolanda Kuo used seven exclamation marks in an utterance consisting of just a few words. She used them for two reasons: a) to express surprise at Carmilla's statement, and b) to alert everyone that she was about to leave the chat.

Ellipsis points (. . .) are used in the Webchat as pause markers, sometimes to signal the intention of switching to a new topic, and sometimes to signal the speaker's intention to yield the floor to someone else (see Day 1, Moves 1, 5, 20, 122, 127, and 135).


It is important to point out that NS resorted to the use of onomatopoetic devices quite frequently, whereas NNS never used them. On the other hand, NNS did use punctuation as a communicative strategy. In particular, NNS used many exclamation marks, probably due to the fact that they were less competent in handling other devices such as onomatopoeia and emoticons which represented L2 colloquial expressions with which they were not familiar.

It is obvious that both NS and NNS showed awareness of the limitations of the Webchat environment. They compensated for the inability to use paralinguistic features that carry semantic load in normal face-to-face communication by resorting to a variety of economical devices, such as the ones described above, instead of resorting to lengthier messages containing the more elaborate language of explanations and descriptions.


Participants clearly demonstrate the context relevance of Webchat in their interactions. There are many ways in which participants deal with restrictions on sense-making procedures imposed by the Webchat environment. These restrictions are translated into peculiar ways of performing actions during the interaction. In particular, the overall structure of turn-taking and the sequencing of actions were influenced by medium which forced interactors to use special ways of packaging actions, expressing paralinguistic meanings, conveying their identities and roles, and making lexical choices.

This preliminary study attempted to describe a variety of ways in which these features are utilized as real and relevant in Webchat. In the future, it would be interesting to assess the impact of synchronous communication such as Webchat on L2 pragmatic competence, for example, in the use of opening and closing formulas, the understanding of implicatures, and appropriate choice of politeness strategies.


Raffaella Negretti is a research collaborator at Verona University's Language Center. She focusses on language research, learning, and teaching through the aid of computer and on-line technology. She graduated from the University of Verona (languages and literature) in 1996, specializing in linguistics and langauge learning. She has collaborated with Università Cattolica di Milano, teaching Internet research for linguistics and English through computer support, as well as Web development for educational purposes. Her main interests are Discourse analysis, Conversation Analysis, and pragmatics, applied to second language acquisition and learning via Internet.



Ellis, R. (1985). Understanding second language acquisition, Oxford: Oxford University Press.

Grice, H.P. (1975). Logic and Conversation. In P. Cole & J. Morgan (Eds.), Syntax and Semantics 3: Speech Acts. New York: Academic Press.

Heritage, J. (1997). Conversation analysis in institutional talk: Analysing data. In D.Silvermann. (Ed.), Qualitative research (pp. 161-181). London: Sage.

Peräkylä, A. (1997). Reliability and validity in research based on transcripts. In D. Silvermann (Ed.), Qualitative research (pp. 201-220). London: Sage.


Pomerantz, A., & Fehr, B.J. (1997). Conversation analysis: An approach to the study of social action and sense-making practices. In T. Van Dijk (Ed.), Discourse as a social interaction (pp. 1-37). Thousand Oaks, CA: Sage.

Psathas, G. (1995). Conversation analysis - The study of talk-in-interaction. Thousand Oaks, CA: Sage.

Sacks, H. (1972). Lectures, unpublished mimeosi, 1964-68: University of California, Los Angeles. Irvine, CA: University of California.

Sacks, H., Schegloff, E., & Jefferson, G. (1974). A simple systematics for the organization of turn-taking for conversation. Language, 53, 361-82.

Schegloff, E. (1968). Sequencing in conversational openings. The American Anthropologist, 70 (6), 1075-1095.

Schegloff, E. (1979). Identification and recognition in telephone conversation openings. In G. Psathas (Ed.), Everyday language: Studies of ethnomethodology (pp. 56-65), New York: Irvington.

Seliger, H.W., & Shohamy, E. (1989). Second language research methods. Oxford: Oxford University Press.

Tarone, E. (1982). Systematicity and attention in interlanguage. Language Learning, 32, 69-84.

Tarone, E. (1983) On the variability of interlanguage systems, Applied Linguistics, 4 (2), 143-63.


Appendix A

Appendix B

Appendix C

Appendix D

About LLT | Subscribe | Information for Contributors | Masthead | Archives