The aim of this paper is to demonstrate that a framework of analysis from spoken discourse analysis can be used successfully to investigatethe interactive written discourse of e-mail.
E-mail is an intermediate form of discourse which has features from both written and spoken language. It is written and asynchronous, but itis also interactive and informal, and participants tend, in spite of the time delay, to refer to the activity they have been engaged in as "conversation"(see for example, Yates and Orlikowski 1993). Accordingly, there has been some work which approaches e-mail discourse from the perspective of spokendiscourse, but this has focused primarily on turn taking and topic maintenance rather than on other levels such as moves and acts. This paper investigateswhether we can extend this work to other levels of analysis, and investigates moves and acts in a sample taken from an e-mail discussion.
Anna-Brita Stenström’s (1994) work is used as the basis for analysis. Although there are other authorities whose levels of analysis differ in some respects from Stenström’s, her framework has two particular features to recommend it: she has developed the framework from the original work by Sinclair and Coulthard (1975), which she has synthesised with work of more recent researchers, and she makes a clear and detailed presentation of the system at the level of moves and acts.
The data for this research are taken from an e-mail discussion list, the Association for Computers and Writing list, ACW-L. The analysis described in this paper forms part of a larger study for which a sample of sixty messages were collected from ACW-L on 14 and 15 May 1997. This paper focuses on twelve messages from the sample. These twelve messages are provided in the Appendix.
An e-mail discussion list was chosen for study because here the entire communication process is carried out by e-mail. In other situations participants might use a mixture of media. For example, in dealing with a colleague one might send an e-mail for a short message, telephone for an urgent matter, or initiate a face to face discussion about a sensitive issue (see Murray1988 for an investigation into "mode and medium switching"). In contrast, for most of the participants in e-mail discussion groups, e-mail will bet he only form of communication which they have with other members of that group. Data are readily available from open discussion lists which provide the non-participating researcher with a valuable opportunity to observe the interaction without influencing it.
Herring (1996) has summarised the ethical issues arising from the use
of such data in research and the need to protect the identities of the
participants in "data from private or semi-private sources". She concludes
that "Messages posted publicly to Usenet and to open-access Listservs are
exempt from this requirement", but she suggests that authors might nevertheless
wish "to mask all participants’ identities" (1996:5-6). The sample for
the current study was taken from an open listserv and is therefore publicly
available, but in the light of the discussions referred to by Herring it
seemed that an ethical approach would be to identify participants by letter,
omitting names and personal e-mail and postal address details. The messages
cover ten main topics (as shown by the subject line of the message header).The
two which form the basis of the analysis in this paper are:
In the data from our discussion list, we find both transactional and interactional communication. A few messages are almost entirely transactional: these tend to be announcements, for jobs, conferences, awards, etc., for example, message 8, an announcement which is the only message on the topic "Professional Development Award". Such messages might begin, as message8 does, with the interactional "FYI" (for your information), or with a brief apology for cross-posting, and then proceed to deliver the transactional information without any further interactional signals. However, such messages formed only a small part of our sample (two messages of the original sample of sixty). Most of the messages are strongly interactional and those messages which do impart information will usually also have a relationship building function.
The purpose of the research is to carry out a preliminary study using a limited set of data from just one discussion group, in order to establish whether the application of techniques of spoken discourse analysis to e-mail discourse is likely to be a promising avenue for future research on a wider scale. Clearly, e-mail is a written discourse and could therefore be analysed as written text, but the focus of this study is on the interactive and dialogic aspects of the discourse, and a discussion of primarily written features is outside the scope of this paper.
In general, researchers who have investigated the discourse structure of computer-mediated communication focus mainly on turn-taking and topic maintenance. Black et al (1983) investigate the sequencing of messages in an asynchronous computer message system which seems to be similar toe-mail and note that one message can initiate several topics and generate responses from several participants, while an individual participant can respond to more than one topic within a single message. Murray (1985) examines turn-taking and topic handling in order to show "the oral/written nature of computer conversation" (1985:211), a recurring theme in discussions about the language of computer-mediated communication. McCarthy, Wrightand Monk (1992:271) in their study on synchronous computer conferencing discuss the problems of "parallel topic development", while Wilkins (1991), describing a system which participants can use either synchronously or asynchronously, shows that participants do not need to negotiate turn-taking, but instead need strategies for topic maintenance.
A major study of the "Dialogue Processes in Computer-Mediated Communication "was carried out by Severinson Eklundh (1986). Using data from the COM computer messaging system, she examines "the structural buildup of dialogues between two individuals" (1986:11). She takes the framework for her investigation from language game theory, which has elements in common with other schools of discourse analysis, for example, she identifies the "question game "which has "prototypical parts or moves" consisting of "Question" and "Answer"(1986:30). She discusses exchange patterns, particularly the issue of whether exchanges are dyads or triads, and she is concerned about the "heavily reduced opportunities for feedback" (1986:11) in computer-mediated communication as compared with other forms of communication. Severinson Eklundh also refers to the "questioning act" (1986:90), and notes that some questions may be preceded by "a motivating or ‘framing’ part", while others may be "followed by one or two statements" (1986:91). However, she does not attempt to identify and label these accompanying elements.
This paper takes as its framework for analysis Stenström’s (1994) model of discourse analysis. This was based on the 1975 model of Sinclair and Coulthard, which was developed to analyze the discourse of classroom interaction. Stenström refined this through a "synthesis" of earlier work from a range of sources, and has applied the resulting model to the "unpredictability of natural conversation" (1994:xii). Stenström’s model is based on five levels of discourse: the transaction, the exchange, the turn, the move, and the act, and she provides tables of the full set of moves and acts which she has identified in spoken conversation (see below).
Moves:
Stenström defines a move as "the verbal action which carries the
conversation forward" (1994:36). In effect, the move is a unit which has
a function relating to the progression of the conversation (in contrast
to the act, which has a function relating to the speaker’s communicative
intention). A move consists of one or more acts, so, for example, a speaker
who is seeking information would make an [Initiate] move which could take
the form of a <question act. The moves which Stenström has identified
are given in Table 1 below.
| [Summons] | calls thelistener’s attention |
| [Focus] | introducesthe [initiate] |
| [Initiate] | opens theexchange |
| [Repair] | holds upthe exchange |
| [Response] | continuesor terminates the exchange |
| [Re-open] | delays thetermination of the exchange |
| [Follow-up] | terminatesthe exchange |
| [Backchannel] | signals thelistener’s attention |
Table 1: Moves (Stenström1994:36)
Acts:
An act is defined by Stenströmas signalling "what the speaker wishes to communicate" (1994:38) or "signalling the speaker’s intention" (1994:221). An act is the smallest element in spoken discourse: one or more acts combine to make a move, just as one or more moves combine to make a turn. Acts are essentially what the speaker is doing in the conversation, regardless of the informational content of the words themselves (for example, it would be possible for a speaker tosay "I agree" in such a tone that it would be clear to any hearers that this was in fact a <disagree). Stenström identifies three classes of act: primary, secondary and complementary.
Acts are a useful descriptive category in spoken discourse, where the interaction between speaker and hearer is important and the interactional function can be more significant than the informational content of the discourse. Even brief conversations can showa variety of acts. Acts are a less useful category of analysis for certain types of written discourse — expository text, for example, could consist mainly of <inform. However, we would predict that acts are a useful category of analysis for any discourse which is interactional, whether it is spoken or written. For example, one might expect a letter to contain a variety of acts, such as <greetings, <informs, <requests,<invitations, <questions, <answers, etc. Since e-mail discussion lists are interactional and perceived as "conversations" we would expect to find a wide range of acts present. We therefore need to determine whether we can find a range of acts in the e-mail data which is similar to the range identified by Stenström, and whether the acts in the data come from all three of the classes (primary, secondary and complementary acts) identified by Stenström.
Primary acts:
Stenström defines a primary act as "the only obligatory act in a move" (1994:224). Stenström’slist of primary acts (1994:39-40) is given in Table 2 below.
| <accept | agrees to a <request, <suggest, etc. |
| <acknowledge | signals receipt of information |
| <agree | signals agreement with what was just said |
| <alert | calls the addressee’s attention |
| <answer | Responds to a <question/<request |
| <apology | Expresses regret |
| <call-off | prompts a conversational closing |
| <check | asks for clarification |
| <closer | ends a conversational closing |
| <confirm | Responds to a request for confirmation |
| <disagree | Expresses disagreement |
| <evaluate | judges the value of what the previous speaker said |
| <greeting | greets somebody or bids farewell |
| <inform | Provides information |
| <invite | asks if somebody ‘would like to do X’ |
| <object | signals a different opinion |
| <offer | Presents something for acceptance/rejection |
| <opine | gives one’s personal opinion |
| <query | Expresses attitude and strong surprise |
| <question | asks for information, confirmation, clarification |
| <react | Expresses attitude and strong feelings |
| <reject | Disagrees to a <request, <suggest, etc. |
| <reply | Responds to a <statement |
| <request | asks somebody to do something |
| <smoother | Responds to an <apology |
| <statement | informs or expresses opinion |
| <suggest | puts forward an idea or a plan |
| <thanks | Expresses gratitude |
Table 2: Primary acts (Stenström 1994:39-40)
Secondary acts:
A secondary act "accompanies a primary act" (Stenström 1994:225). Secondary acts add emphasis or further information to the primary act (and would account for the preceding and following statements which Severinson Eklundh found accompanying questions in her data from the COM system — see above). Stenström’s list of secondary acts is given in Table 3 below
| <clue | follows a primary act and gives a hint |
| <emphasizer | Underlines what was said in the primary act |
| <expand | gives complementary information |
| <justify | defends what was said in the primary act |
| <metacomment | Comments on current talk |
| <precursor | Precedes a primary act and gives information |
| <preface | Introduces a primary act |
Table 3: Secondary acts(Stenström 1994:44)
Complementary acts:
Stenström states that a complementary act "accompanies primary/secondary
act" (1994:222). Complementary acts are primarily interactional, and are
low in informational content ("well", "erm", "you know", etc.). They are
therefore the category that one would least expect to find in writing.
Stenström’s list of complementary acts is given in Table 4 below.
| <appealer | invites feedback |
| <booster | Assesses what the speaker himself says |
| <empathizer | ‘involves’ the listener |
| <filler | fills a gap in the discourse |
| <frame | marks a boundary in the discourse |
| <hedge | helps avoiding commitment |
| <monitor | helps putting something right |
| <staller | plays for time |
| <starter | helps getting started |
| <uptake | accepts what was said and leads on |
Table 4: Complementary acts (Stenström 1994:46)
In e-mail discussions we are dealing with multi-party interaction, usually with a very large number of participants. This is in direct contrast to most studies in spoken discourse analysis which focus on two-party dialogue or on group interaction among a small number of participants. Stenström’s examples likewise are mostly two-party dialogue or very small group interaction. She does include one example of a discussion among eight committee members, but says (1994:187) "with so many people involved, a description of the entire interaction in terms of five neat discourse levels will certainly cause problems." The sample in this study includes contributions from thirty participants in sixty turns, and so we would not expect the whole of our data to fit neatly into Stenström’s categories. We would expect to find that we need to adapther system, extending the definition of some moves and acts, and creating new ones where no suitable classification exists. Nevertheless, as this paper will show, Stenström’s work does provide a very useful framework for approaching the analysis of e-mail discourse.
This paper will now examine some sequences from the data in order to illustrate how Stenström’s framework can be applied to this type of discourse. It will consider three sequences :two relatively straightforward exchanges from the topic "Testing, testing", and a complex turn from the topic , "Questions about basic writers andcomputers".
In topic "Testing, testing", J has just joined the list, and begins the topic by sending his first message to the group (message 4). His behaviour here is unusual: normally when new participants join a list they simply begin to contribute to the discussion when they are ready, possibly prefacing their first remark with something like "I am new to this list and I would like to ask...". J, however, is announcing his presence and seeking some kind of welcome, which he duly receives. In terms of Stenström’s framework, the progress of the topic is unremarkable, except for the fact that the turns on this topic are interspersed with turns on other topics. "Testing, testing" is in fact a brief topic— J seems to be satisfied with the responses he receives, and by message19 of the data he is contributing on another topic, giving his final message on "Testing, testing" in message 20.
There were nine messages on this topic in our sample, one shared [Initiate], four [Responses] and four [Follow-ups],giving four instances of the pattern: [Initiate] - [Response] - [Follow-up].
The three messages below representone of these instances.
Message 4
Is this mike working?
J...
Message 11
J... wrote:
Is this mike working?
J...
yes, J..., it is.
w...
Message 13
On Wed, 14 May 1997, W.. wrote:
J... wrote:
Is this mike working?
J...
yes, J... it is.
w...
Bit of an echo, huh? Thank you, W...- J...
The turns in these messages are brief, and the main moves are straightforward: message 4 [Initiate], message 11[Response], message 13 [Follow-up].
In terms of acts, the [Initiate]in message 4 is a simple move consisting of one act, a <question, "Is this mike working?"
Message 11 includes a quotation from the previous message. The software will have inserted the quotation, but the sender will have chosen to insert a quotation and decided how much of the previous message to quote. When all or part of a message is quoted at the start of a turn, as it is here, it has the effect of a focusing move, although this particular type of [Focus] is characteristic of e-mail and not of spoken conversation. Message 11 then consists of two moves: a [Focus] consisting of "J... wrote" together with the quotation, and a[Response] "yes, J..., it is." In fact, Stenström uses [Focus] solelyas a move which "introduces the [initiate]" (1994:36). But here we have a move which is introducing a [Response]. This move is functioning in very much the same way as Stenström’s [Focus], and it seems reasonableto extend [Focus] to include moves which introduce a [Response] or [Follow-up]or [Repair]. This kind of move is common in e-mail because of the timedelay — it is often necessary or helpful to link a message to a previous message, whereas in spoken discourse turns follow each other sequentially. One could, however, find a [Focus] being used to introduce a [Response]in spoken discourse if an aside sequence comes between an [Initiate] and the [Response].
The [Focus] in message 11 begins, "J... wrote". This is a secondary act which "precedes a primary act and gives information" (1994:44), and therefore could reasonably be classed as a <precursor, even though Stenström normally takes <precursors as preceding <questions. Moreover, it simply has an introductory rather than a face-saving effect, which suggests that it should be classed as a <precursor rather than a <preface or a <metacomment (see Stenström 1994:86).
The primary act in the [Focus] move of message 11 is the <quotation. Here we have had to create a new class of act not identified by Stenström, but this seems to be justified because quotations are a distinctive feature of e-mail and are used very frequently. It is possible to have quotations in spoken discourse, but in that situation they would probably be carrying new information as in the following fabricated example: "I was just putting the dinner out when Thomas rushed in and said ‘I’ve forgotten my tuba’ ". Here, " ‘I’ve forgotten my tuba’ " simply forms part of the whole <inform, whereas the <quotations in e-mail have a reference rather than an informational function.
The second move in message 11 is a [Response]. This is a complex move consisting of the primary act <answer, "Yes", and the secondary act <expand, "it is", together with the complementary act of <nomination, "John". The act of <nomination is not included in Stenström’s set of acts, but it is identified by Sinclair and Coulthard (1975). In the classroom situation which they describe, nomination is used primarily in an [Initiate] move by the teacher to select the next speaker. In e-mail discussions it is used mainly as part of a [Response], where it serves both an informational function (indicating a reply to a specific participant) and an interactional function, reinforcing the relationship by the positive politeness strategy (Brown and Levinson 1987) of first name use. In spoken conversation we can use <nomination in a similar way as part of a [Response] such as "Yes, Alex".
Message 13 also consists of two moves, in this case a [Focus] and a [Follow-up]. Again, the [Focus] consists of a <precursor and a <quotation. Then we have the [Follow-up] consisting of a primary act <statement, a complementary act <empathizer (in the form of the tag "huh"), a primary act of <thanks, "Thank you", and a complementary act of <nomination, "W...". This move, then, consists of four acts, but two of these are primary acts. When Stenström is discussing complex moves, she states that they consist of "an obligatory primary act and an optional secondary and/or complementary act" (1994:37).This implies that a move cannot have more than one primary act. But [Responses]which consist of an <answer followed by <thanks are common in spoken conversation ("Yes, thanks"). Stenström also states that "Moves consisting of more than three different types of act are rare" (1994:38). In e-mail where an individual turn may carry the interaction further than is usual in spoken discourse, they are not unusual, but more research is needed to investigate the frequency and composition of such moves.
All three messages end with the first name of the sender. This is a convention from written rather than spoken discourse, but this use of the sender’s name is important because some e-mail software does not display the header information. Ending with the name, therefore, serves to identify the participant and does not really fall inside the framework of acts.
So far the exchange and the moves and acts within it correspond very well to the familiar patterns of spoken discourse, and even the variations are in keeping with the general framework. However, we also find in this exchange a feature which could not be explained by spoken discourse analysis. As we noted above, quotation is used frequently in e-mail, but e-mail software also makes it easy to quote all or part of a previous message which itself contains a quotation. In Message 13, J uses this facility to create the [Focus], in this case creating a double embedding by quoting the whole of W’s message which itself included the quotation of J’s [Initiate]. This double embedding creates a textual effect of reiteration, emphasised by the visual effect of the double angle brackets. J, who has already created the microphone metaphor by using the subject line "Testing, testing" and by asking in message 4, "Is this mike working?", uses this visual effect to extend his metaphor, likening it to an echo.
The second sequence to be examined is another exchange from the same topic, "Testing, testing":
Message 4
Is this mike working?
J...
Message 16
Is this mike working?
It seems to be feeding back ; --K...
Message 20
On Thu, 15 May 1997, K... wrote:
Is this mike working?
It seems to be feeding back ;--K...
Ouch. That one hurt my ears.
In this exchange, message 4 is the same [Initiate] as before. Message 16 has a [Focus] and a [Response]. The[Focus ] is a simple move consisting only of a <quotation. The [Response] has a primary act, <answer which includes a pun, "feeding back". This pun is followed by a smiley ; . The smiley represents a winking face, and, functioning like a real wink in face to face conversation, is intended to show that the previous remark should be taken humorously. Although it is a graphic device, it is functioning in the text as a comment on the previous statement and is therefore acting in the same way as the secondary act <metacomment.
Finally, message 20 also has two moves. There is a [Focus] consisting of a <precursor and a <quotation. This is followed by a [Follow up] which begins with a primary act <react followed by a secondary act <justify , and these together extend the microphone metaphor.
This topic consisted almost entirely of short turns. The average turn in the data is much longer than those in the examples given above (excluding header and signature details, the average length of turn for the sixty messages of the data was 179 words). As an example of a longer turn, we will examine the first message of the topic "Questions about basic writers and computers".
Message 1
Hi.
We're thinking of developing interactive exercise software for basic writing
students, and we're trying to figure out whether to do a network version
and, if so, what that should look like. Could you take a few minutes to
answer some questions?
For your basic writing students whouse computers ...
1. Do most of your students use Macs or Windows machines?
2. Do they use standalones (personal or school-owned), or are they on a network?
3. If they're on a network, what kind of a network is it? (Novell?
Appleshare? something else?)
4. If you have a network, do you use it to run applications or more as a
way to share files?
5. Would you like a networked version of interactive exercise software? If
so, what network capabilities or features would you like to see?
You can respond to me off-list, if you'd prefer. Thanks in advance for your
help.
Sincerely,
D...
This message opens the first topic, and consists of three moves, a [Summons], a [Focus] and an [Initiate].The [Summons] is a primary act of <greeting, "Hi". The [Focus] consists of a secondary act <preface, "We're thinking of developing interactive exercise software for basic writing students, and we're trying to figure out whether to do a network version and, if so, what that should look like", and a primary act <request, "Could you take a few minutes to answer some questions?"
The [Initiate] is much more complex .It consists of a secondary act <precursor, "For your basic writing students who use computers"; five primary act <questions each initiating a sub-topic within the main topic; a primary act <suggest, "You can respond to me off-list, if you'd prefer"; a primary act <thanks, "Thanks in advance for your help"; and finally a primary act of <greeting, "Sincerely". However, if we look more closely at the five questions, we can see that, apart from the first which is simply a <question, theyare made up of more than one act. Question 2 has a secondary act of <expand, "personal or school-owned"; question 3 has a <precursor, "If they’re on a network", and an <expand, "Novell? Appleshare? something else?"; question 4 has a <precursor, "If you have a network"; and question five is in fact two questions, the second of which is introduced by a <precursor, "If so". We therefore have a highly complex [Initiate] consisting of nine primary acts and six secondary acts.
This kind of complexity is rare in spoken language. Although the framework of moves and acts provides us with a basis for analysing this message, the patterns of acts within the moves, particularly within the [Initiate], are different from those that we find in spoken conversation. The sender is taking the interaction much further in one turn than would be normal in a face to face interaction, where one would expect to find a turn by another participant after each of the primary acts. The fact that the message can be analysed in terms of moves and acts shows that it is composed of elements common to interactive discourse, but this accumulation of primary acts within one turn depends on the fact that the message is in fact written. If D had been carrying out face to face interviews, she would have asked her questions one at a time, getting a response to each before going on to the next. The medium of e-mail both enables her and encourages her to present all of her questions simultaneously. If she had asked her five questions consecutively in spoken conversation she would have been placing an unreasonable cognitive load on the hearer— many hearers would have forgotten at least some of the questions by the time they had completed one or two answers. In this message, the sender is relying on the fact that readers can move at their own pace through a written document, keeping the text in front of them while they compose their replies, and reading it section by section, or referring back to anything they have forgotten. Furthermore, if D had asked the questions one at a time using e-mail, waiting for a response to each before going on to the next, the time delay would have been very tedious, and a strong disincentive to all participants involved.
However, there are problems with carrying the discourse so far within one turn. Having sent her message, D receives a response in message 3 from R, who clearly objects to D’s request and uses parody to demonstrate his objection to what she is asking. R takes this up again in message 12. R has a remarkably strong negative reaction to D’s message, apparently disapproving of the nature of her request. But D has used a <preface in her [Focus] which in spoken discourse might have prevented this. As Stenström shows (1994:86), "<prefaces have a face-saving effect in that they prepare B for what is going to happen next". So in face to face conversation, R could have intervened at this point and declared that he felt the approach was inappropriate. In practice, however, R could not respond to the <preface at the appropriate stage in the interaction because it was sent with the rest of the message. It is possible that this led to a stronger negative reaction than would have occurred if he could have responded at an earlier stage. Perhaps one factor which contributes to flaming is this accumulation of primary acts within one turn.
Of the eight moves which Stenström identifies, we can find six in the e-mail data:
[Summons] message 1, "Hi".
[Focus] message 1, "We’re thinking of developing interactive exercise software for basic writing students, and we’re trying to figure out whether to do a network version and, if so, what it should look like."
[Initiate] message 4, "Is this mike working?" This is a simple [Initiate]. [Initiates] can be much more complex, as in message 1 where the sender asks five questions.
[Repair] message 6, "Are you trying to tell me that my query was inappropriate?" D had requested information in message 1, and this was parodied by R’s [Response] in message 3. D is now checking her understanding of R’s [Response].
[Response] message 11, "yes, J...,it is." This is a direct [Response] to a specific message (4) and a specific sender, J. Other [Responses] relate to the discussion so far, rather than an individual message, and some messages are addressed to the group rather than to an individual participant.
[Follow-up] message 20, "Ouch. That one hurt my ears."
[Re-open] There are no examples of[Re-opens] in the data.
[Backchannel] The function of a [Backchannel]is to show the hearer’s attention and to encourage the speaker to continue. There are no [Backchannels] in the data and we would not expect to find them in e-mail discussions because the other participants cannot see what an individual contributor has written until the turn is complete. Severinson Eklundh (1986) also comments on this absence of backchannelling. Thus, just as senders cannot be interrupted by other participants before they have completed their turn, they likewise cannot be encouraged to continue by backchannelling.
The moves identified by Stenström cover most of the moves found in the data, though there are occasions where Stenström’s framework cannot fully account for what is taking place in the data. For example, in message 3 we have a situation where the same piece of text is fulfilling two functions: it is responding to the first message by parody, and, because it parodies the form of D’s message, it takes the form of an [Initiate], which D duly responds to. This double function is a complex one, but it does have some resemblance to the Response/Initiate move identified by Coulthard and Brazil and exemplified here in P’s turn(1992:71):
T: Can anyone tell me what this means?
P: Does it mean ‘danger men at work’
T: Yes
(The Response/Initiate move is also accepted and used by Francis and Hunston 1992).
>Message 12 reveals an additional problem. This message is quite complex and includes a [Response] to D’s[Repair], a [Re-initiate] asking for a public reply to R’s spoof [Initiate],and a [Response] to D’s [Initiate].
Stenström does not have a [Re-initiate]move, and her [Re-open] does not account for this type of move: it is in fact more like a pre-closing move, as her example, below, demonstrates(1994:123):
A ...would twelve o’clock be OKAY [initiate]
B: LOVELY [respond]
A: RIGHT [re-open]
B: YES
The [Re-initiate] move was identified by Sinclair and Coulthard (1975) who found many [Re-initiates] in their classroom data. Possibly there were no occurrences in Stenström’s data, but this kind of move can occur in ordinary conversation, as shown by Francis and Hunston (1992).
However, although we do find some moves which cannot be accounted for within Stenström’s system of analysis, we find that in general Stenström’s framework provides a useful basis for approaching moves in the e-mail data, and those moves which cannot be accounted for within this framework are nevertheless amenable to analysis by a modest adaptation of the system.
Stenström lists twenty eight types of primary act, of which twenty-three are found in the data and five are not.
There are no examples of <acknowledge. Stenström describes the <acknowledge as "the most economical way of responding" (1994:110), and it takes forms such as "all right", "quite" , etc. In e-mail discussions participants do not need to respond unless they wish to contribute to the topic, and so there is no need for a minimal response of this kind.
There are no occurrences of <call-offs or <closers. This is not surprising: discussions can continue without a break for years until group members stop participating or until the listowner decides to close down the list. <Call-offs prompt a closing, and <closers finalise a closing in conversation: during the life of a discussion list these would not normally be needed.
There are no <rejects. In e-mail discussions, <rejections tend to be handled by absence of response. This is particularly true of requests made to the group as a whole. When no individual has been selected to respond, members of the group feel free to ignore the request. This happens in the case of message 9: this message initiates a new topic, "Signature image", but there are no replies on this topic. (However, replies could have been sent directly to this one person, initiating a separate private discussion. This can happen when participants feel that the material is not relevant to the whole group.) An exception to the use of non-response to reject a request is when participants object to the request as inappropriate to the list. In such a situation there will often be an overt objection. In the data there is an example of this particular situation being handled by parody: R (message 3) objects to D’s request for information (message 1) by making a parallel request for information from D which he clearly expects to be considered inappropriate by D.
There are no examples of <smoothers. The apologies which occur in the data receive no response.
The primary acts identified by Stenström account for most of the primary acts in the data, although it is occasionally necessary to define a new category. The most prominent example is the need for an act of <quotation, as described above.
In the e-mail data there were examples of all the secondary acts identified by Stenström except <clues. Senders are producing asynchronous written discourse where the other participants are physically distant. Perhaps for this reason they tend to be more explicit than in spoken discourse, where one can assess from the reaction of the hearer whether or not the utterance has been understood. Thus in this respect e-mail discourse resembles written discourse — spoken discourse has a tendency to be "vague" (Crystal 1995:291).
Complementary acts are primarily interactional. In the data there were examples of all types of complementary act except <staller. What is surprising is not that this one is missing, but that we do have examples of the others. It is interesting to consider the reasons for the existence of complementary acts in e-mail discussions. Are the participants deliberately imitating spoken discourse, deliberately seeking maximum informality and emphasising the interactional aspects of the discourse? McCarthy notes that "style in English has drifted in general from the literate to the more oral between the eighteenth and twentieth centuries" (1993:171-2) and suggests that "many texts composed and transmitted in the written medium and meant to be read ‘mark’ themselves on the surface of the text as echoing that which might be spoken and heard." (1993:172).This is particularly noticeable in popular texts. McCarthy is studyingthe ‘fanzine’ and he believes that the move towards spoken discourse features is "partly a result of modern technology (the word-processor) liberating the authors from the constraints posed by editors and publishers, pushing the fanzine genre evermore towards the spoken mode and away from the written modes adopted by the more staid institutional sources previously available to fan readers" (1993:172). However, the participants in ACW-L could be said to be those same "staid institutional sources". The lack of constraints certainly apply — there are no publishers, editors, or secretaries to make the text conform to the conventions of written discourse — but is this really a rebellion against established norms or simply a change of style appropriate to a situation which is both informal and interactional?
In the data, although there are a few absences (discussed above) we can find examples of most of the acts identified by Stenström. The wide variety of acts found in this sample shows strong similarity with spoken discourse and is in sharp contrast to the limited range of acts found in most types of written text.
As a tool for the analysis of e-mail discussions, the framework of spoken discourse analysis is remarkably successful. E-mail discussions are a combination of the transactional and the interactional. They include some messages, for example announcements, which are almost entirely transactional. These use the structure and language of written discourse, and for this type of message Stenström’s framework is not particularly useful. However, most of the messages in the data combine a transactional with an interactional function, and a few are almost entirely interactional. For these messages which are partly or entirely interactional, Stenström’s framework succeeds very well and enables us to make useful observations about the structure of e-mail discourse on the levels of moves and acts.
pHaving established that we can use Stenström’s framework to analyse the discourse of e-mail discussions, what does the framework tell us about this new discourse? First it demonstrates, in a striking way, that a significant element of such discussions is the interactive function. Although this discussion group was set up with the "idea of formally generating and providing information to be equally shared by both commercial and academic users" (message 39), the messages actually posted to the discussion are far less transactional than this original aim suggests. Instead, the mixture of discourse in this discussion group resembles the mixture that one might find in a common-room, with some personal conversations taking place, and some more public discussions addressed to a wider circle, alongside some formal notices on a notice board.
Using Stenström’s framework also helps us to identify elements which resemble elements of spoken discourse, but which serve a different function. For example, we can investigate the occurrence of spoken discourse markers and their function in a discussion which is communicated in writing.
Finally, using the framework we can identify those elements which do not normally occur in spoken discourse. Some of these elements are features from written discourse, for example the form which is provided with message 8 (the Professional Development Award announcement) and which participants are invited to complete. Other differences from spoken discourse are a direct result of what is and is not possible using current e-mail software. For example, quotations can easily be incorporated without retyping, so they are used frequently to maintain the topic or to show which particular participant or which part of a message is being replied to.
pWhat we find, then, is a fascinating and complex discourse. It is composed of the transactional and the interactional, of the public and the private. It exhibits many of the features of spoken discourse, but it also uses features of written discourse, and features which are not normally found outside e-mail, all working together to forma recognisable whole.
Black, S. D./ Levin, J. A./Mehan,H./Quinn, C. N. (1983): "Real and non-real time interaction: Unraveling multiple threads of discourse". Discourse Processes, Vol 6: 59-75.
Brown, P./Levinson, S. (1987): Politeness:Some universals in language usage. Cambridge: Cambridge UniversityPress.
Brown, G./Yule, G. (1983): DiscourseAnalysis. Cambridge: Cambridge University Press.
Cheepen, C. /Monaghan, J. (1990):Spoken English: A Practical Guide. London: Pinter Publishers.
Coulthard, M./Brazil,B. (1992): "ExchangeStructure". In: Coulthard, M. (ed.): Advances in Spoken Discourse Analysis London, Routledge: 50-78.
Crystal, D. (1995): The CambridgeEncyclopedia of the English Language. Cambridge: Cambridge University Press.
Francis, G./Hunston, S. (1992): "AnalysingEveryday Conversatio’." In: Coulthard, M. (ed.): Advances in Spoken Discourse Analysis London, Routledge: 123-161.
Herring, .S. C (1996): "Introduction". In: Herring, S. C. (ed.): Computer-Mediated Communication: Linguistic, Social and Cross-Cultural Perspectives. Amsterdam, John Benjamins:1-10.
McCarthy, J. C./Wright, P. C./Monk,A. F. (1992): "Coherence in text-based electronic conferencing: coupling text and context". Journal of Language and Social Psychology 11/4:267-277.
McCarthy, M. (1993): "Spoken discourse markers in written text". In: Sinclair J.M./Hoey, M./ Fox, G. (eds.): Techniques of Description: Spoken and Written Discourse London, Routledge:170-182.
Murray, D. E. (1985): "Composition as Conversation: The Computer Terminal as Medium of Communication". In: Odell, L./Goswami, D. (eds.): Writing in Nonacademic Settings. New York, The Guilford Press: 203-227.
Murray, D. E. (1988): "The context of oral and written language: a framework for mode and medium switching". Language in Society 17: 351-373.
Severinson Eklundh, K (1986): "Dialogue processes in computer-mediated communication: A study of letters in the COM system". Linköping Studies in Arts and Science 6, University of Linköping.
Sinclair J McH/Coulthard R. M. (1975): Towards and Analysis of Discourse. Oxford: Oxford University Press.
Stenström, A.-B. (1994): An Introduction to Spoken Interaction. London: Longman.
Wilkins, H.(1991): "Computer Talk: Long-Distance Conversations by Computer". Written Communication 8 /1: 56-78.
Yates, J.A./ Orlikowski, W. (1993):Knee-jerk Anti-LOOPism and other E-mail Phenomena: Oral, Written, and Electronic Patterns in Computer-Mediated Communication. Working Paper, Cambridge, Mass. Massachusetts Institute of Technology, WP#3578-93 June1993: 2-20
The sample was taken from ACW-L on14 and 15 May 1997. The e-mail address of this list is acw-l@ttacs6.ttu.edu
This appendix contains twelve messages from the data. The date and subject lines of the headers are shown, but the remaining header details (Reply-To, Originator, Sender, Precedence, From, To, Status) have been removed. Any references in the data to participants’ names and contact details have been replaced by ellipses together with an indication in square brackets of what has been removed.
Message 1
Date: Wed, 14 May 1997 12:27:49 -0500
Subject: Questions about basic writers
and computers
Hi.
We're thinking of developing interactive
exercise software for basic writing
students, and we're trying to figure
out whether to do a network version
and, if so, what that should look
like. Could you take a few minutes to
answer some questions?
For your basic writing students whouse
computers ...
1. Do most of your students use Mac
sor Windows machines?
2. Do they use standalones (personal
or school-owned), or are they on a network?
3. If they're on a network, what
kind of a network is it? (Novell?
Appleshare? something else?)
4. If you have a network, do you
use it to run applications or more as a
way to share files?
5. Would you like a networked version
of interactive exercise software? If
so, what network capabilities or
features would you like to see?
You can respond to me off-list, if
you'd prefer. Thanks in advance for your
help.
Sincerely,
D...[First name]
****************************************
...[Full name, role, address, e-mail
address, phone, fax, www]
"Sooner or later, everything gets
done. Or not."
****************************************
Message 3
Date: Wed, 14 May 1997 14:49:43 -0500
Subject: Re: Questions about basic
writers and computers
Hi D...[First name]
I am thinking of starting a publishing
company, geared toward the
educational market, though general
fiction, cookbooks, self-help manuals
and technical software may also be
developed. We are trying to get as
much information as possible to make
our startup a success, so would you
mind taking a few minutes to answer
the following questions?
For the educational market ...
1.Do most of your sales come through
catalogues? Direct sales or
institutional?
2.Do desk copies to instructors really
garner student sales?
3.What percentage of bedford's revenues
come from educational textbooks?
4.What percentage of bedford's revenues
come from general fiction?
5. What were your 1996 pre-tax revenues?
Was this an increase/decrease
over 1995?
You can respond to me off-list, if
you prefer. Thanks in advance for
your help.
Sincerely,
R...[First name]
*******************************************************************
...[Full name, role, address, e-mail
address]
Message 4
Date: Wed, 14 May 1997 15:32:29 -0500
Subject: Testing, testing.
Is this mike working?
J...[First name, e-mail address]
Message 6
Date: Wed, 14 May 1997 16:18:13 -0500
Subject: Re: Questions about basic
writers and computers
R...[First name]
Huh?
I'll respond if you like, but, gosh,
I strongly suspect you're spoofing me
here ...
Are you trying to tell me that my
query was inappropriate? I'm not
promoting anything, I'm just trying
to find out more about what people do.
If you'd like me to tell you how
I work on books together, I'd be happy to
share that with you.
D...[First name]
At 02:49 PM 5/14/97 -0500, you wrote:
Hi D...[First name]
I am thinking of starting a publishing
company, geared toward the
educational market, though general
fiction, cookbooks, self-help manuals
and technical software may also be
developed. We are trying to get as
much information as possible to make
our startup a success, so would you
mind taking a few minutes to answer
the following questions?
For the educational market ...
1.Do most of your sales come through
catalogues? Direct sales or
institutional?
2.Do desk copies to instructors really
garner student sales?
3.What percentage of bedford's revenues
come from educational textbooks?
4.What percentage of bedford's revenues
come from general fiction?
5. What were your 1996 pre-tax revenues?
Was this an increase/decrease
over 1995?
You can respond to me off-list, if
you prefer. Thanks in advance for
your help.
Sincerely,