1. Introduction
2. Methods
2.1 Speech Events
3.1 E-Mail
3.2 Mailing Lists
3.2.1 How Mailing Lists Function
3.2.2 Typology
3.2.3 Accessibility
3.2.4 Direction
3.2.5 Distribution
3.2.6 Orientation
4.1 Announcements - the simplest speech event on mailing lists.
4.3 Discussions
4.3.1 Topic
4.3.2 Quoting
4.3.3 Turn-taking
5.1 Silence
5.2 RFIs Within the Discussion
6. Conclusion
If you find any problems with the files, or the text, or if you want a copy of the paper in a different format, please send me a message.
E-mail is a form of communication whose use is increasing exponentially, as the Internet, and other means of computer-mediated communication (CMC), become more accessible to the general public. Aside from its use in direct interpersonal communication, just as letters, memos, and faxes are used, it is also used for communication among groups that share common interests or goals, through what are known as mailing lists.
Tens of thousands of mailing lists exist, about as many subjects. They give people the opportunity to discuss these subjects using a form of discourse that is relatively new. While this type of CMC is a form of written communication, there are many aspects of e-mail mailing list discourse that are similar to those used in spoken discourse. This has led some researchers to suggest that CMC is a hybrid register.
In this paper, I will discuss how mailing lists function, the different types of mailing lists that exist, and how the type of mailing list can influence the type of discourse that is used on the list. Then I will discuss the different types of speech events that are used on mailing lists. Finally, I will show how those speech events are realized by examining an extended thread from one mailing list.
It will be seen that mailing lists generate three types of speech events, which are the Announcement, the Request for Information, and the Discussion. I will show how these three types of speech events function, and what the differences and similarities are between mailing list speech events and face-to-face communication.
Cyberspace - the world of computer-mediated communication - has no more concrete reality than, say, bellspace, the world of telephone communication. But just as the telephone has spawned its own specific form of discourse, which is recognizable as such, so has electronic communication given rise to new and unique forms and genres of communication.
The computer has become more than just a tool for calculating or processing text, but is now a window open to a world of communication through the Internet. More and more people are getting connected to the Internet and using it as a tool for communication, but, unlike more traditional forms of communication, such as letters or the telephone, they have not grown up with it. The strategies for communicating via the Internet, or other computer networks, are being born and refined right under our eyes. Each new user who embarks on this form of communication needs to know how to effectively communicate, but also needs to know the norms that give such communication the characteristics that allow it to be correctly interpreted by its different users.
This situation is made more complex by its global nature. People from all over the world communicate with computers on a daily basis, and yet on the Internet, no one has an accent. "The rapid growth of the Net has thrown together millions of people from widely divergent backgrounds and levels of computer expertise; [...] we might think of cyberspace as a sprawling urban metropolis with people everywhere." (Sutton, 1995) People whose native languages are varied, find themselves in a forum where the overwhelming language used is English. They need to learn, not only the norms for a form of communication that is new to them, but also how to do it in a second language.
Computer mediated communication (CMC) is a "hybrid register that resembles both speech and writing and yet is neither." (Veselinova & Dry, 1995) Written conventions are seen combined with the characteristics of spoken language, in a dynamic form that is developing into its own register. (Danet, forthcoming 2) Some of the reasons for this are purely technical, and dependent on the nature of the programs used for such communication (the use of subject lines, for example). As Talbott (1995:30) points out, "But since the register is relatively new, "...many email writers are drawing from their experience of both written and spoken communication when composing in this medium." (Daly, 1996)
In one study, Veselinova & Dry (1995) showed that people using one form of CMC "acquired" this register over time, through learning by doing. "In the beginning, LINGUIST subscribers tend to conceive of CMC as written academic communication and use the language appropriate for this. But the longer they stay on the list the more they tend to abandon the formal language and go over to less formal expressions characteristic of spoken language."
In this paper, I will examine one form of CMC, which is the e-mail mailing list. I will show that there are a limited number of speech events on mailing lists, in spite of their subjects, and demonstrate how these speech events function. Then I will examine one long series of messages from one mailing list, showing how these speech events actually occur.
E-mail mailing lists are a genre that is particularly conducive to linguistic investigation - it is very easy to obtain data, no transcription is necessary, and there is no problem with the "observer's paradox", since none of the list members are aware of who the others are, unless they expressly present themselves. "Lurking" on a mailing list is therefore a simple way to observe interaction and behavior. This does not mean that someone doing research on mailing list need necessarily remain silent - the researcher need not present him or herself as a researcher, and may contribute to the discussions, and will be considered just like any other participant.
In my case, I am a member of a number of mailing lists. Some of them for professional needs (LINGUIST, LANTRA), others as support for computer or software problems (MAC-L, NISUS-L), and other out of personal interest (EARLYM-L, Alextech). Having observed, and participated in, a number of mailing lists, I started noticing regularities among the speech events used on lists that may be widely different in their subject matter and their goals. I then started analyzing data from these different lists, in order to establish the patterns that underlie the different speech events used. This analysis has shown that there are three types of speech events that are used in mailing lists. But just like in normal face-to-face conversation, one type of speech event may be "embedded" within another.
The choice of the term "speech event" to describe e-mail mailing list exchanges may be seen to imply that such exchanges are a form of speech, i.e. conversation. A number of researchers have examined this question (Shank, 1993; Veselinova & Dry, 1995; Danet, forthcoming 1, and forthcoming 2; Maynor, 1994) and the general consensus is that of Shank (1993):
"Is Net communication like conversation? Quite a bit. Messages on the Net tend to be informal, to be phrased in conversational form, and can engender a great deal of direct and dyadic interchange. Is Net communication like writing? Absolutely. Messages are written instead of spoken."
E-mail has the characteristics of written texts, being that it "resembles all kinds of everyday notes" (Danet, forthcoming 1). But what sets it apart from other forms of written communication is, curiously enough, an effect of the medium itself - the rapidity with which messages can be transmitted. For, while e-mail is asynchronous, the speed of transmission makes it possible for a number of messages to be transmitted back and forth in a short time. In fact, if two participants in an e-mail exchange are on-line at the same time, it is possible for them to respond to each other's messages with a speed which is only limited by their typing speed, and the length of the messages they send. A number of times I have gotten answers to messages I have sent within minutes.
Therefore, the basic unit of e-mail exchanges is the message, which can be seen as similar to a conversational move (Goffman, 1981:71-72). This also serves as a turn, in the sense that turn-taking in e-mail exchanges is relative to each individual message; messages cannot be interrupted, so only the transition-relevance places (TRPs)in e-mail are at the end (or reception) of a message.
E-mail mailing lists can be seen as communicative organs for small speech communities. They all consist of people who "have [something] in common" (Saville-Troike, 1989:18), and who share the same language, at least sufficiently to discuss the subject at hand. Within these speech communities, there may or may not be a feeling of belonging to a community. In an informal survey done on the LANTRA mailing list, an overwhelming majority of the people who responded said that they felt a "group adherence" and a "feeling of community" towards the list members. One person called it, "An international cyberspace community of translators and interpreters, those who have the equipment and skills to participate in this medium." As Talbott (1995:17) points out, "The online community is real because every medium that passes a 'word' - by which I mean an expressive gesture, an act with an inside - will bear some kind of human community."
E-mail is one of the most widely used forms of CMC. Tens of millions of e-mail messages are exchanged in the world each year, and e-mail is quickly becoming a preferred form of communication for many people, especially those in science and research, who use e-mail to communicate with their peers, anywhere in the world. "For some, it is a way to get in touch and chat with other teens or young adults. For others, it is a way to talk to experts to determine how and why a particular damnable computer or program won't run properly." (Shank, 1993) One of its main attractions is its asynchrony - an e-mail message can be read at any time by its receiver (Negroponte, 1995:167-8). "...private email resembles ordinary letter-writing: the sender composes and sends the message at a time separate from that in which the recipient receives or reads it." (Danet, forthcoming 2) This allows people to carry on extended discussions without the constraint of time requiring them to be present when the other person, or people, are sending messages. This asynchrony enabled the builders of the Internet to establish mailing lists shortly after the network itself was set up.
E-mail mailing lists (also known as Listservs, after the name of the software program used to run many of them) are one of the most widely used genres of CMC. One search engine that indexes mailing lists claims to have 66,692 different mailing lists indexed (http://www.liszt.com , as of 25 Sept., 1996). Their subjects can range from academic (such as the Linguist list, which deals with linguistics), to practical (there are many lists dealing with computers, hardware and software), to self-help groups (Grief-net is one example, which assembles people dealing with grief), to fan clubs (2400 Fulton Street, for fans of the group The Jefferson Airplane), to daily announcements of the latest news (Daily Brief). In spite of the vast difference between the subject matter of different mailing lists, they all tend to exhibit similar modes of interaction, similar speech events. All of these mailing lists can by typologized according to a number of characteristics, which strongly influence the type of discourse which is used in them.
Mailing lists are easy to use, since they require no special software (other than e-mail software), and e-mail can be received on any computer that can be connected to the Internet. They present nothing but text, but in that sterile environment of plain ASCII characters, a great deal of linguistic variety can be observed.
Mailing lists are very different from newsgroups, which are also discussion groups, but which are accessed differently. Anyone can post a message to a newsgroup, while you must be a member of a mailing list to participate. Because of this, Usenet newsgroups have a very high ratio of spam and flaming, compared to the number of messages. (Spam is advertising e-mail, suggesting ways of earning money fast, or ordering magazines cheaply; flaming is the "sudden, often extended flare-ups of anger, profanity and insult." Danet, forthcoming 3)
For the past year, I have been conducting an ethnographic investigation into mailing lists, having been a member of many of them, and using them extensively. I will present a framework for a typology of mailing lists, and I will show how the discourse structure of mailing lists is highly dependent on the characteristics that I will define.
3.2.1 How Mailing Lists Function
Joining (or subscribing to) a mailing list is very easy to do. Once one has found the name of a list that one would like to join, a simple command sent to a computer someplace automatically subscribes the user to the list. For example, to subscribe to the Linguist list, it is necessary to send the following message:
SUBSCRIBE LINGUIST
to
LISTSERV@TAMVM1.TAMU.EDU
Some time later, from a few seconds to a few hours later, the subscriber will receive a message explaining how the list works and what options the person has, as to how they wish to receive their messages. This "welcoming" document also contains instructions for how to unsubscribe, should the person wish to do so.
Once someone has become a member of a list, they will receive all of the messages that are sent to that list. The Listserv software simply takes all of the incoming messages, and sends a copy of each message to each member of the list, at the time the message is received. On some lists this may be a small number of messages, perhaps a few a day, or even a few each week. On other lists, however, the traffic can be much higher; I have been involved in some lists where there were more than 100 messages a day.
There is one alternative to receiving all of these messages separately, for such high-volume lists. For many lists there exists a digest mode, where the subscriber will receive, either: one file per day, containing all of the messages sent during a 24 hour period; or, a message each time there is more than a given number or size of messages waiting (often around 30 kilobytes, which can be anywhere from a few messages to about 20 or 30).
In order to post a message to a list, the user simply needs to address an e-mail message to the address of the list (for the Linguist list, the address is: LINGUIST@TAMVM1.TAMU.EDU), and their message will be sent out to all of the current subscribers. There are some exceptions to this schematic description, which I will discuss later.
Mailing lists, while having varied subjects, all follow a number of similar criteria for acceptance to the list, type of messages allowed etc. These criteria constitute a typology for mailing lists, and, as we shall see, the restrictions placed on a list have a very strong influence on the type of message that is permitted, and therefore on the type of discourse that the list generates.
There are two types of list accessibility: public and private. Private lists are those which may be set up by companies, universities, or even just groups of friends or family, and where new members are not allowed to join. In most cases there is no way to find out about these lists, but on some search engines these private lists can be found. If someone were to try to join such a list, they would either receive some sort of message saying that they do not have permission to do so, or they would simply get no answer.
Within public lists, there are two types: open and closed. Most publicly assessable mailing lists are open, which means that there is no screening of people who sign up. But there are some which are closed, where the list owner reviews the names of new subscribers. This is done for two reasons: to keep out people who may have disturbed the conventions of a given list, or to prevent spam. When subscribing to such a list, you receive a message telling you that "your request for subscription has been forwarded to the list owner," and you then usually receive confirmation within a day or so.
Mailing lists come in two flavors: moderated and unmoderated. The Linguist list is an example of a moderated list. All messages are reviewed by a moderator, for pertinence, spam, and professional relevance, before they are accepted for distribution. In the case of the Linguist list, messages are grouped and then sent out. For other lists, the messages may be sent out individually.
Moderated lists have advantages and disadvantages, as well as important factors which constrain the discourse structure of messages, and the type of subjects that are dealt with. (Gruber, 1996) While they can eliminate all superfluous messages (spam, me-too messages, where someone will quote a long message and simply add the words "I agree." to the end of it), they can also strongly repress the list members' right to initiate topic, or to carry on a thread.
Moderated lists may have two types of expression: free or controlled. While the choice of posting a message by the moderator may be considered to be a way of controlling expression, it is more a yes or no choice. Some mailing lists go much further, and have strict guidelines as to what type of message are considered acceptable. One example is the TESL-L list, which is very successful, with more than 20,000 members. In order to have a message accepted for posting on the list, it must pass a series of 18 restrictions, which are communicated to new members when they sign up. While the owners claim that these restrictions are necessary in order to have a list which functions smoothly with the high number of participants, it is obvious that speech is constricted.
Unmoderated lists, on the other hand, are places where anything goes. Some of them can be quite disciplined, and there may be a discipline enforced by the list community itself, but others may contain a large number of messages which are off-topic, that is, not related to the discussion topic of the list. It is also relatively easy for spam to get through, although I have noticed that most people tend to try to ignore this as much as possible. One list that I have participated in for a long time is LANTRA-L. This is a mailing list for translators and interpreters, and it generates over 100 messages a day. There are often messages which are off-topic, but, as I mentioned previously, the list members tend to react when that happens, and remind the senders of such messages that they should refrain from doing so. Sometimes flame wars will break out as a result of this, but there are rarely as vehement as those seen in newsgroups.
There are two ways that a mailing list may be distributed, one-to-many, or many-to-many. It is obvious that in one-to-many mailing lists, the discourse is created at one end of the chain, and the subscribers have no real input into the list. This kind of list can be a presentation of news, announcements of new products, or just a regular mailing of articles or opinions by one person. One example:
Welcome to the Red Rock Eater News Service. This message contains some notes about RRE. Please consider hanging onto it for future reference.
As a subscriber to RRE, you will receive periodic messages from the
editor. The normal traffic on RRE is perhaps five messages per week.
...
The RRE News Service is a filter, not a discussion list. It is rigged
so that only its editor can send messages to it. If you have a message
or file that might be of broad interest, feel free to send it along to
rre-maintainers@weber.ucsd.edu and we'll check it out.
This kind of list, while technically a mailing list, does not offer any interactive communication between the members.
Many-to-many lists make up the vast majority of mailing lists. These are the ones where people may discuss the diverse subjects mentioned above, and in most cases they are relatively free and open. Members usually send either request or comments to the list, and can respond to other messages, either by answering questions, or responding to others' comments. In many cases, these exchanges take on the form of conversations, where a number of people may respond to one another, and a subject may be discussed for a period of time stretching from a day or so, to a number of weeks.
Within the many-to-many lists, three types of messages may be seen. The first is information exchange. Very common on computer lists, this consists of one person sending a question to the list, either seeking a solution to a problem, or advice.
Another type of message is the announcement. One member, or perhaps a company, has something to announce: it may be a product, a call for papers, the date and time of a meeting or information about a new web site. This sort of message does not usually require a response, but they sometimes do receive responses. People may either respond requesting further information, or there may be a discussion that develops about the information given. (Discussions can develop from any of the different types of messages.)
The third type of message is the simple discussion. One list member initiates a new topic, and others respond. These discussions may be short and simple, or they may go on for long periods of time, often for weeks. The topic may change over time, and will eventually die out.
The three types of messages mentioned above, while distinct, often exist simultaneously on any given mailing list. An active list contains a number of conversations, or threads, at any given time, and these threads may or may not interest all of the list members, who have the choice of participating in them or not. Most mailing lists I have observed have this tripartite structure. These three types of messages correspond to the three types of speech events seen on mailing lists, which will be discussed in part 4 of this paper.
The final characteristic for typologizing mailing lists is the list orientation. There are two, which are process-oriented, and product-oriented. This distinction, while simple, may not always be clear, since the two can overlap. A process-oriented list is one where the discussion itself takes precedence over any tangible results of the information that is exchanged. Fan-club lists are a good example of this, as are self-help group type lists. In many ways, these lists tend to contain much more phatic communication, and much less actual "information." People communicate here to communicate, rather than to accomplish goals.
Product-oriented lists, however, tend to be such as the computer list where one has a problem. Too much discussion is frowned upon, and often criticized by one or more list member as being "off-topic." Messages can be more terse, and often contain little or no superfluous comments. Another example is a mailing list I am on for beta-testing a word-processing program, where the goal is to carry out the testing, and exchange information which is important to it. There is no topic initiation that is off-topic.
Here is a summary of the above typology:
I Accessibility
A Private
B Public
1 Open
2 Closed
II Direction
A Unmoderated
B Moderated
1 Controlled
2 Free
III Distribution
A One to many
B Many to many
IV Orientation
A Product
B Process
Posting a message to a mailing list can be considered like "...a group of persons who are sitting in a dark cave: anybody [sic] knows that there are some others and that they might (but most not) respond to one's own utterance." (Gruber, 1996) The poster does not know if anyone will "hear" him, or whether or not they will answer. "... list-subscribers have to decide collectively which contribution is relevant to the context of a certain discussion (list), and contributors who want to deliver a relevant contribution have to anticipate which message will be regarded as relevant by the others." (Gruber, 1996) Sometimes no one will answer; this silence, like silence in spoken interaction, can have communicative value. Here it means that no one judges the post interesting or important enough to respond, or, in the case of a request for information, no one has an answer; "...gaining the focus of the group's attention or the "conversational floor" depends entirely on the extent to which other participants acknowledge and respond to one's postings." (Herring, 1993)
However, one or more people may choose to answer, which, in essence, "validates" the post as being worth of recognition. One usually does not address a particular person (although this can happen); one addresses the group, in hopes that there will be a response. As Goffman (1981:3) points out; "When a word is spoken, all those who happen to be in perceptual range of the event will have some sort of participation status relative to it." Whether or not they choose to exercise their participation status is up to them.
Once a thread (a topic) has been initiated, responses tend to be linked to previous messages by quoting text posted by another list member, and, often, responding to that person, but in public. The person could, just as easily, choose to respond by private e-mail (and perhaps many do - list members would not know), but given the fact that the original message was public, it seems to make sense that the response also be public.
There can be four possible types of message posted to a mailing list:
"- initiating messages which successfully stimulate a new discussion
- initiating messages which fail to stimulate further discussion
- continuing messages which cause further discussion
- continuing messages which are "dead ends." (Gruber, 1996)
Each of these messages involves three kinds of participants: an addresser, the group of addressees, and one or more responders. In most mailing lists it seems to be the norm that only a small percentage of the list members actually participate, while the rest "lurk", or read without contributing. These lurkers are still participants who are "ratified participants but are not specifically addressed by the speaker" (Goffman, 1981: 9-10), but do not need to contribute. In fact, most mailing lists would be overwhelmed if every member contributed regularly. One list I am on, LANTRA, has usually from 600 to 900 members, and the traffic is often above 150 messages a day. Were this traffic to be increased six- or eightfold, most members would unsubscribe. Of course, this would reduce the traffic, and it is possible that many mailing lists are self-regulating in this way.
I therefore define an exchange as one message being sent to the mailing list, and its members, and one or more list members responding to it. But what if there is no response, or at least no overt response? Should we accept, as Gumperz (1982:1) says, that "Only when a move has elicited a response can we say communication is taking place."?
As we will see, examples can be found where a message, not having received an answer, is reposted. In this case the addresser seems to have received a response, in the lack of response. This silence may also be a simple question of a face-preserving strategy (Brown and Levinson, 1987) (Goffman, 1967). I would like to suggest that this too be considered a speech event, in the same way, as in face-to-face communication, (Saville-Troike, 1989:146-8) an unanswered comment can have communicative value.
There are three types of speech events found on mailing lists. They are Announcements, Requests for Information, and Discussions.
4.1 Announcements - the simplest speech event on mailing lists.
The simplest form of speech event on mailing lists is the announcement. Announcements may be of many types. There is the "official" announcement, of a conference, or a call for papers:
Date: Tue, 30 Jul 1996 23:26:55 CDT
From: XXXX@YYY.YYY
Subject: call for manuscripts: Computer-mediated conversation analysis
==============================================================
Call for Manuscripts:
COMPUTER-MEDIATED CONVERSATION
High-quality manuscripts are solicited for an edited
collection on the topic of COMPUTER-MEDIATED CONVERSATION,
to be published by a major publisher. This will be the first
book devoted entirely to linguistic and conversation analytic
approaches to computer-mediated communication.
[...]
There is the simple announcement of information, often of a web site, or of other computer-related information:
Date: Sat, 17 Aug 1996 01:14:36 +0100
From: Andre J. <XXXX@XXX.IE>
Subject: Classical JukeBox spawns Disk of the Day
At
http://www.ireland.net/marketplace/foundation/ComJute/ClassicalJukebox2.html
there are now several hundred more archived CD reviews, also the Disk of
the Day, which will be updated daily and which you can reach directly at
http://www.ireland.net/marketplace/foundation/ComJute/DiskoftheDay.html
Andre
But there is also what could be called the personal announcement. Since many mailing lists engender a feeling of community, participants sometimes will announce such things as vacations, moves, or new jobs.
Date: Fri, 9 Aug 1996 08:55:07 +0100
From: "Michael f. B." <XXXX@XXXX.XXX>
Subject: Bye, bye blackbird!
I'm flying away on my holidays.
A happy summer to you one and all.
Fundamentally,
Michael
Date: Sun, 11 Aug 1996 10:59:07 +0200
From: Laura S. <XXXX@XXXX.XXX>
Subject: No Mail
I'm going no mail for 1 week. Hope you won't miss my politically incorrect
frankness.
While announcements seem to not fulfill Gumperz' (1982:1) requirement for communication, at least on the surface, by looking closer at the actual process which they incite, we can see that, in all cases they elicit a response.
Let us take, as an example, a call for papers. Assuming that such an announcement has been sent to members of a mailing list, each list member, as long as he or she has actually received the message, must respond in one way or another. The simplest, and perhaps most common response, is to simply press the Delete key. This technique which, as Talbot (1995:12-13) points out, has a tendency to contribute to the paucity of content in e-mail messages. "Many of us take pride in the speed with which we can dispose of messages ... To contemplate the speaker behind the words - who is he, what is my connection to him, and what do I owe him in the form of attention and concern? - is hardly realistic." But in the case of an announcement, the delete key is a response; it is a way of saying, no, I am not interested. Since the message was sent to a mailing list, it comes with no implied request for an answer or confirmation, which may be necessary if the message was, for example, copied by someone and sent personally to someone else.
The next form of response to an announcement is to actually respond. The receiver may wish further information, or may wish to submit a paper, or an abstract. In that case, the receiver sends a message to the sender. But this response is not (except in the case of an error by the sender) sent to the mailing list, and therefore leaves the realm of mailing list speech events, to become a basic e-mail message. (I will have more to say about this later.)
It is possible that one or more list members may respond to the list, to make comments or observations about the content of the announcement. In this case, the speech event changes, and becomes a discussion, or a "thread" in mailing list jargon. Whether this thread continues or not depends on the interest of the list members.
The final form of response is to file the message away for further consultation. This response, like that of using the delete key, is indeed a response, even though it is "silent". One of the particularities of CMC is that not all communicative acts are actually what would be called communicative for written or oral communication, precisely because it is not entirely written nor spoken communication. But "...the tools we use to communicate a message affect what we say." (Talbott, 1995:30). The medium itself, in addition to influencing message style (Daly, 1996), also influences the way messages are handled. A message that is received must be disposed of, one way or another (unless the receiver has unlimited disk space, and wishes to save everything). But even saving a message requires the receiver to act on it, to make a decision based on its content. One needs to decide where to file it, or whether to copy all or part of it to a different location, or even whether to print it for further use.
It may seem strange to suggest that a technological process is a move in a speech event. But this is just what Schegloff (1972) suggests when he points out that a telephone's ring is the first part of an adjacency pair. One could also argue that the reception of an e-mail message is also a "summons" to read it. In fact, most e-mail programs make some sort of noise to alert the user that mail has arrived. This is often the case when the user dials up a connection to check their mail, but it is even more prominent for users whose computers are connected to an internal network, and who will receive some sort of warning each time new mail has arrived for them.
In the same manner, my response to a piece of junk mail (a paper letter) I received this morning was to throw it away. This "communicative" act has the value of saying "No, I am not interested", and is indeed a response to the letter in question.
The request for information (RFI) is a very common speech event on mailing lists. On some mailing lists, particularly those dedicated to computers or software, the RFI makes up the majority of messages. This speech event is relatively easy to analyse, due to its simple two- or three-part structure. Although any speech event on a mailing list may branch out into an interminable thread, the basic structure is always present. It is as follows:
P1 Request
P2 Response
P1 [Thank you]
The brackets around the third move indicate that it is optional. But on mailing lists where there is a feeling of community, this thank you seems to be a necessary element of etiquette. A simple RFI can be seen here:
Date: Mon, 9 Sep 1996 11:06:46 -0400
From: "Charles A. P." <XXXX@XXXX.XXX>
Subject: Year Missing in Finder display
One of my colleagues' last modified entry typically reads
Mon, Aug 26, , 9:47 am
instead of
Mon, Aug 26, 1996, 9:47 am
This omission persists in every window.
We have tried deleting the Finder preferences as well as the Date and Time
preferences, but to no avail. Any ideas?
By the way, we are all running under System 7.5.3.
Thanks in Advance.
Charles A. P.
------------------------------
Date: Mon, 9 Sep 1996 15:19:55 -0000
From: Jesse B. <XXXX@XXXX.XXX>
Subject: Re: Year Missing in Finder display
>This omission persists in every window.
>
>We have tried deleting the Finder preferences as well as the Date and Time
>preferences, but to no avail. Any ideas?
>
>By the way, we are all running under System 7.5.3.
Go into the Date and Time Control Panel and select Date Formats. Under
the "Year" popup menu select a format other than "none". You can see the
results of your change at the bottom og the window.
Regards,
Jesse B.
------------------------------
Date: Mon, 9 Sep 1996 11:45:13 -0400
From: "Charles A. P." <XXXX@XXXX.XXX>
Subject: Re: Year Missing in Finder display
At 11:19 am on 09/09/96 in 'Re: Year Missing in Finder display' you wrote:
> >This omission persists in every window.
> >
> >We have tried deleting the Finder preferences as well as the Date and Time
> >preferences, but to no avail. Any ideas?
> >
> >By the way, we are all running under System 7.5.3.
>
> Go into the Date and Time Control Panel and select Date Formats. Under
> the "Year" popup menu select a format other than "none". You can see the
> results of your change at the bottom og the window.
>
I feel rather dumb for having stared at the obvious without seeing it!
Thank you.
This very simple RFI, consisting of only three messages, is actually somewhat rare. In fact, the larger the number of subscribers to a given list, the more chances that the RFI will have the following structure:
P1 Request
P2 Answer
P3 Answer
P... Answer
Pn Answer
P1 [Thank you]
Here is a short example of the above structure:
Date: Sun, 18 Aug 1996 15:53:33 -0500
From: Shaw W. <XXXX@XXXX.XXX>
Subject: Eudora Pro 3.0
Hi all,
I have a general question regarding Eudora Pro 3.0. Is there a
Compact Mailbox option under SPECIAL like in previous versions
like Pro 2.1.4? Or do we need to suffer bloated mailboxes? :)
thanks,
shaw
------------------------------
Date: Sun, 18 Aug 1996 16:15:13 -0400
From: Alfred Karl K. <XXXX@XXXX.XXX>
Subject: Re: Eudora Pro 3.0
>Hi all,
>
>I have a general question regarding Eudora Pro 3.0. Is there a
>Compact Mailbox option under SPECIAL like in previous versions
>like Pro 2.1.4? Or do we need to suffer bloated mailboxes? :)
>
>thanks,
>shaw
Eudora 3.0 will automatically compact the mailboxes when the amount of
unused space is more than the amount of used space.
Example: 79/169K/146K <--- Will NOT compact
73/170K/280K <--- WILL compact
To force Eudora to compact a mailbox hold the command key and click on the
lower-left corner of the mailbox (the part with the number of messages and
space statistics on it). To compact all the mailboxes, hold down the
option key along with the command key and click in the location I just
mentioned. You can find this info on page 72 of the Eudora 3.0 Manual.
------------------------------
Date: Sun, 18 Aug 1996 16:33:17 -0400
From: "Robert E. W." <XXXX@XXXX.XXX>
Subject: Re: Eudora Pro 3.0
At 4:53 PM -0400 8/18/96, Shaw W. wrote:
> regarding Eudora Pro 3.0. Is there a
> Compact Mailbox option under SPECIAL like in previous versions
> like Pro 2.1.4? Or do we need to suffer bloated mailboxes? :)
>From the comp.mail.eudora.mac FAQ, Eudora 3.0beta:
http://www.amherst.edu/~atstarr/eudora/faq3.html
>>> 10. Where is "Compact Mailboxes"?
"Compact Mailboxes" has been removed from the "Special" menu because it was
under-utilized and to make room for other options on smaller screens. To
compact a single mailbox, Command-Click on the bottom left corner of the
mailbox (where the number/size information is). To compact all mailboxes,
Option-Command-Click in the same place. >>>
________
Robert W.
------------------------------
Date: Sun, 18 Aug 1996 16:51:18 -0500
From: Shaw W. <XXXX@XXXX.XXX>
Subject: Re: Eudora Pro 3.0
Thanks to Alfred and Robert for their prompt responses!!!
regards,
shaw
It seems that this structure containing multiple answers is more prevalent on mailing lists with high traffic; not only because there are more participants, but because many of the participants receive the mailing list's messages in digest mode. This creates the phenomenon of "netlag', where one reads a message that was written, perhaps, 24 hours before, and, by the time one answers it, many others have already done so, but the digest reader will not know until the next day. The differences in time zones are also a factor - a message sent at night by someone in Europe will be read during the day in America, but not until the next morning by someone in Japan.
The thank you move seems thus to serve two purposes - the first is to acknowledge and thank the list members who have sent an answer, and the second is a marker that delimits the speech event, saying, in effect, "This question is now closed."
There is another way this speech event may be modified:
P1 Request
P2 Request for further information
P1 Response to P2
P2 Response
P1 [Thank you]
This is common when the question is too general, or needs contextualization to be answered. This is exactly the same as insertion sequences which are seen in face-to-face conversation (Levinson 1983:304-5).
The RFI can therefore be seen to have the following structure (ovals indicate optional elements):
This diagram shows that RFIs actually contain only one essential element: the request. Unfulfilled RFIs can exist, and indeed do; these contain only the request, and either there is no response, or the response is sent privately to the requester. It is important to note that, as far as speech acts on mailing lists are concerned, for all mailing list speech events, it is possible that responses are made directly to the sender of a message, and not to the list itself. There is no way for list members to know this is the case. However, when receiving mail from a mailing list, the message headers are such that the default reply (the addressee of a response when the user clicks on the "reply" button in the e-mail software) is to the mailing list, and not to the sender. Thus, for someone to reply directly, an extra step must be carried out.
This highlights one of the difficulties of analyzing mailing lists: the only data we have is that data which is sent to the list members. We cannot be aware of any communication that goes "backchannel", or via private e-mail.
Speech events in mailing lists are not always as simple as the above examples. Often, one form of speech event is embedded in another, or a speech event branches off to become another kind. In this manner, an announcement can engender an RFI, or an RFI can have an announcement as its response.
Discussions are the most complex of mailing list speech events. They are also those which resemble spoken conversations the most. The discussion seems to be the "default" speech event for mailing lists - anything that is not an announcement or an RFI can be seen as a discussion. After all, mailing lists are also called discussion lists. Discussions can also branch off from any other speech event, and may even branch off from other discussions.
The topic, in a conversation, is what people are talking about. (Levinson 1983:313) Throughout a spoken conversation topic shifts and changes, usually without any formal signals. In mailing lists, however, since there are many threads going on at once, it is necessary to have some means of identifying which message belongs to which thread.
A contributor to a mailing list generally posts a message in the hopes that others will read it and respond to it. To be noticed, they need to "write an effective subject line that both covers the topic of the text and captures the attention of busy readers on public lists who, faced with too much mail, may be tempted to delete the message without reading it." (Daly, 1996) Or as Kawasaki (1996) points out, (Rule #3) "Describe the topic of your message in the subject area. This helps both you and your recipient. When your recipients have inboxes full of messages, they are going to make the decision about which messages to read and answer based on who sent the message and what the subject is."
Some examples:
Subject: Re: Fugue in C major
Subject: FW: Sites about Ken Kesey and the Merry Pranksters
Subject: Re: lavender reproduction
The topic of a discussion can change, and this too is often (but not always) reflected in the subject line:
Subject: "I/O Error. Bad Media" from StuffIt products. (was Disk Error)
Subject: ljeppen, laufen, leapin lizards! (was TERMS:Artificial fertilizer)
Subject: Bill Gates (was: Pamela Anderson)
The notion of topic in conversation is still somewhat vague, While it is clear that people do talk about things, the exact dynamics of how topic evoles, in spoken conversation, is one that is not easily described. One very obvious difference between face-to-face conversation and e-mail is the clear declaration of topic, although, as I will point out later, the topic announced in the subject line of a message is not always that of the message itself.
Discussions can be very complex, far more so than face-to-face conversations, because in a given thread, there may also be various sub-threads that are evolving; As Schiffrin (1994:235) says, "...each utterance in a sequence is shaped by a prior context (at the very least, and most typically, the immediately prior utterance) and provide a context for the next utterance...". Because of this, it is necessary to anchor moves in a discussion to previous moves. All e-mail software allows you to do this by "quoting" text from a previous message. The receiver can then contextualize a given response. The quoted text is included in chevrons, and, in some e-mail programs, may be displayed in a different color. (The inclusion within chevrons, while not the case in all programs, seems to be almost universal.)
> ... I recall our music teacher playing some dull organ piece
> and talking to us about Bach the mathematician.
Do you also recall the dull piece? And what about Bach ``the
mathematician'': what is so ``mathematical'' about Bach's music?
Quoting, which is a form of lexical cohesion, helps create texture, and provides "...the points of contact with what has gone before." (Halliday & Hasan, 1976:299) If these points of contact were not present, it is highly doubtful that such discussions would be possible.
There seem to be two uses for quotes, which are both used in discussion speech events. The first is to create a link, or a point of contact with a previous message.
One example:
> I do not understand this "waste of band-width" business. If you can not
> afford small fees then it might be something wrong with you (YOU =
> collective whingers).
> I think YOU spend more time and Bandwitdh on flaming Sergei.
>
> Jerzy B.
I must take issue with the above, on behalf of those Lantran's who do
not, for example have T1 Internet connections. Mr. T's good
intentions were obvious, but the road to Hell is paved with good [...]
In this type of quote, the text quoted tends to be long, and can include an entire message.
The other type of quote is shorter, and selects a particular part of a previous message to respond to:
>To those of you who received Sergei's file:
Surely you mean "Hi everyone", don't you? ;-)
Here the response is not to the entire message, but just to one part of it.
As Daly (1996) points out, e-mail software, "...by including opportunities for freely quoting earlier text when replying, for editing those quotes, and for inserting replies/responses within the quoted text, also encourage[s] a conversational, turn taking style..." This ability to quote is what makes it possible for long discussions to be carried out by e-mail. This intertextuality, which is specific to the medium of CMC, influences the way the discussion is carried out - a single move by one participant can respond to different parts of a previous move, or to more than one move at a time, by selecting what he or she is responding to and quoting it in a message. This also enables other participants to follow the discussion, since sequential moves do not always arrive to the mailing list in a sequential order, but are interspersed with other messages, which are parts of other speech events. (It is also possible that a move which is a response to a previous move may arrive, to some users, before the original move, due to the way e-mail is transmitted.)
Some messages can even be constructed as "dialogues", with the replies interspersed between parts of the original message:
Subject: Re: Edupage, 29 September 1996
What do y'all think of this?
Edupage Editors reveals:
>
> THE EVER-MORPHING PC
> "The PC will fade into the background as we deliver technologies that allow
> people to focus on their jobs, as opposed to focusing on the computer's user
> interface," says Stephen Boies, head of IBM's interactive systems division
> at the Thomas Watson Research Center.
So what else is new? The above statement could have been (was) made
every year for the last fourteen years.
> For instance, IBM is developing a
> specialized device that car dealers can plug into their telephone line to
> get credit approval for their customers in under two minutes. And
> physicians and nurses at Long Beach Memorial Medical Center in California
> use touch-screen flat-panel displays embedded in the hospital's walls to
> track patients' progress.
Yes, good stuff, but is there something new about it (okay, the flat panel
is probably new, but is the software?).
> Apple Fellow Donald Norman predicts that we're
> seeing the dawn of computing's next generation: "We're at the end of the
> second generation of the personal computer," which was marked by the
> graphical user interface, promising ease of use.
Blah, blah, blah. This is typical media hype, no?
> "But what happened," says
> Norman, "is computers have become even more complex, expensive, and
> unmanageable."
I don't know what kind of computer he has, but mine is cheaper and easier
to use than my last one. Not to mention ten times as fast.
> In the third generation, "the focus will be on people and
> the tasks they want to accomplish, not technicalities." (Information Week 23
> Sep 96 p48)
What a concept! I don't suppose anyone thought of THAT before!
Someone responded to the above message:
[...]
> > THE EVER-MORPHING PC
> > "The PC will fade into the background as we deliver technologies that allow
> > people to focus on their jobs, as opposed to focusing on the computer's user
> > interface," says Stephen Boies, head of IBM's interactive systems division
> > at the Thomas Watson Research Center.
>
> So what else is new? The above statement could have been (was) made
> every year for the last fourteen years.
Right. And has yet to show any real sign of happening.
> > "But what happened," says
> > Norman, "is computers have become even more complex, expensive, and
> > unmanageable."
>
> I don't know what kind of computer he has, but mine is cheaper and easier
> to use than my last one. Not to mention ten times as fast.
"Easier to use" is relative. In normal consumer terms (i.e., non-geek,
and I'm sorry to break this to you all, but basically everybody on this
list is probably a de facto geek :), modern computers are "easier to use"
in the sense that now, rather than having to pound a sharp stick into your
eye, you're instead given a nice soft plastic spoon with which to gently
pry out your eyeball.
[..]
And then there was a response to that response:
[...]
> > > In the third generation, "the focus will be on people and
> > > the tasks they want to accomplish, not technicalities." (Information Week
23
> > > Sep 96 p48)
> >
> > What a concept! I don't suppose anyone thought of THAT before!
>
> You're right -- this is absolutely the stock, party-line, Apple Fellow
> jabbering that everyone's been spouting for 10 years. What continues to
> be interesting is that nobody's managed to make it happen, even as they
> all swear daily allegiance to this goal.
Well, all the hype keeps people's expectations just ahead of the technology.
If all we wanted were a word processor and a spreadsheet, a 386 with Win 3.1
and MS works would be all we asked for. But expectations just keep going
up and up! Remember how utterly FANTASTIC WPs and SSs were when they first
came out? I thought they were so great. And how much money the guys from
Lotus made off 1-2-3? JUST 1-2-3? But these days, I'm upset that my WP
doesn't imbed JPEGs! Soon it will and I'll want imbedded MPEGs, with full
stereo sound. [...]
Each "layer" of chevrons indicates a level of remoteness from the current message, and enables the reader to follow the evolution of a discussion.
The above is an interesting example of how quoting can be used to create a form of discourse that is specific to e-mail, and that is "interactive". The text that is constructed is done by more than one person, and is not so much an instance of quoting to refer to a previous message as it is the creation of a new text, which is a collaborative work of a number of participants.
The above text starts as follows:
Edupage (quoted text)
Comment
Edupage
Comment
...
Then more comments are added:
Edupage
Comment
Comment 2
...
Then more:
Edupage
Comment
Comment 2
Comment 3
...
The result is a highly complex text, which is both the result of a number of comments being added on to it, and an original text which is the sum of those comments.
Unlike spoken conversation, where turn-taking has been found to follow rules relative to the location of TRPs (Sacks et al, 1974), turn-taking in mailing lists is dependent on completely different criteria. TRPs exist at the end of each message, and the 'speaker' always self-selects, with the exception of messages where one person is directly questioning another. But even in this case, the responses are always dependent on the asynchrony of e-mail communication; one can only send a message to a mailing list when one's computer is connected to the network. A turn may come in a given thread between a comment to a particular person and their response. There can also be more than one response to a given message. For this reason, turn-taking on mailing lists is defined by the medium itself. One can not usually ensure that one's turn, in response to a previous message, arrives to the mailing list right after the original message.
It can also be seen that some people may send a number of messages to the list at approximately the same time. This occurs when people receive the list as a digest, and send all of their responses to various list messages at the same time. In this case, their turns in different threads will be posted together.
What actually is a discussion on a mailing list? In some ways, it is not very different from a discussion in face-to-face conversation; people interact by adding their comments to previous comments, and the "text" that is created in the discussion flows on as a result of the sum of these different moves. But in other ways, it is very different: the length of moves, the possibility to quote, turn-taking, and the other elements described above, set it off from the structure of a face-to-face discussion. But for many reasons, actually describing a mailing list discussion is relatively complex.
This complexity is due, in part, to the nature of mailing lists: there are many "channels" open at once, and all the participants in a discussion do not receive the same moves at the same time. While a face-to-face discussion is relatively "linear", a mailing list discussion is "multi-dimensional". The dimension of time plays a part that is absent in face-to-face communication. (For this reason, both the diagram presented below, figure 2, and the diagram in appendix I only show one aspect of the structure of the moves in a discussion: the order of moves relative to previous moves. The actual time aspect is absent, because it would require adding an additional dimension to the diagram.)
A discussion can be seen to have the following structure, but this is a bit reductive:
P1 Initiation
P2 Response
P3 Response
P... Response
Pn Response
The responses can continue, with any participant responding to any other's move, and with more than one participant responding to a given move at any time. But, there are only two possible results of a given move: someone responds to it, or not. (Unlike announcements and RFIs, the discussion on a mailing list is a speech event which transpires entirely on the mailing list. In an RFI, there can be responses that are sent by private mail, but the discussion, as far as can be seen, is a public forum, which requires that the messages remain public. In my personal experience, I have rarely received private e-mail relative to a discussion, and when this does happen, this is more of an aside, that, while related to the thread, is meant to be a private comment.)
This diagram is a close up view of a part of the diagram in appendix I. It shows how some moves contribute to the continuation of a discussion, and others are dead-ends:
Some moves are therefore successful, in the sense that they help carry on the subject, while others are unsuccessful, and are not responded to.
Perhaps the main distinction between mailing list discussions and face-to-face discussions is that on a mailing list you can choose what you respond to, while in a face-to-face discussion, social pressure may force you to respond to each move that is made by others. On a mailing list you can also ignore a move that is directed to you, and you can choose to respond to just a part of someone's move; you can avoid answering someone without losing face, or threatening the face of others. It is also possible to "interrupt" without offending anyone, since the notion of "floor" does not have much importance. In fact, anyone can come into a discussion at any time on a mailing list, whereas getting the floor in a face-to-face discussion follows certain rules.
Shank (1993) calls the e-mail discussion a "multilogue":
"In the multilogue, we have a number of players. We have the starter, or the initial sender, who starts the "thread" (a well- established Net term, by the way). Once a thread has been started though, it is no longer under sender control. This is because the mechanics of Net response do not require turn taking. From the oral side, it is as if everyone who is interested in talking can all jump in at once, but still their individual voices can be clearly heard. From the written side, it is as if someone had started writing a piece, but before he/she gets too far, people are there magically in print to add to, correct, challenge, or extend the piece. Therefore, what we have is a written quasi-discussion that has the potential to use the strengths of each form. Since the "feel" of Net communication is still oral, I think it is best to call this form of communication "multiloguing", to retain the link with its oral heritage."
This idea of multiloguing helps explain why discussions can be particularly difficult to describe. The following section, which looks at one such "thread" should shed some light on the way such a speech event unfolds.
In order to present the different types of speech events used in mailing list, and to point out some of the specificities of theses speech events, I will examine one thread from the LANTRA-L list. (This is a mailing list devoted to translating, and at the time that this thread occurred, had over 600 members.) This thread consists of 82 messages, sent during a period of 8 days. It is significant because it contains examples of all of the different speech events, and also shows a number of embedded speech events (see appendix II for the complete transcription).
This thread started as a simple RFI (message 1), and, had things gone normally, would have most likely generated just a few messages. Within the thread, the basic RFI structure can be seen:
Request message 1
Response message 2
Thank you message 69
But, what caused the thread to be so long was message 2, where Sergei T. made the mistake of sending an attachment of over 900K to the mailing list. This violation of a number of Grice's (1975) maxims (The maxims of Quantity, Relevance and Manner), and Kawasaki's (1996) rule #1, "Don't send a file unless the recipient wants it and expects it.", is considered to be the ultimate mailing list sin. It occurs from time to time that someone mistakenly sends a file to a mailing list, instead of to an individual, and this action usually generates responses, but this is the largest such attached file I have yet to receive. On receiving this file, 37 of the more than 600 subscribers responded, thus creating a discussion embedded within the RFI, with other discussions branching out from that one.
It is interesting to note that the first such response slightly changed the subject (message 66), and it was this new subject that people then responded to, rather than the message containing the file. (Although one possible explanation is that no-one kept the large file, and therefore could not reply to it directly using the reply button of their software.)
The diagram in appendix I shows a schematic representation of this entire thread. It can be seen that new threads were started which died out, and that others proved more fertile. There were a total of 17 subjects proposed, 9 of which received no response, and one of which was extremely fertile, generating 36 messages.
Now it is debatable whether a thread is dependent on the actual subject, as shown in the subject line of the messages, or the fact that, as in spoken conversation, people are simply talking about the same thing. As Levinson points out (1983:315), "...topical coherence is something constructed across turns by the collaboration of participants." There is also topic drift in mailing list communication, where the subject line has not changed, but the people are talking about something new.
I will therefore take the term topic to mean, simply, what the people are talking about, and the term subject to mean the words used in the message's subject header. This is an important distinction, since any remarks on the discussion speech event will need to examine topic shift. I will show that new discussion speech events are instigated, even without changing the subject header of the messages.
The coherence of the whole thread, beyond just the topical coherence, is dependent on a certain amount of presupposition and shared knowledge among the participants. A new member, who arrived during this thread, would find a number of things which are a part of the shared knowledge of the Lantra community, but which would be incomprehensible to them, such as references to people and previous events.
Oddly enough, out of 17 messages, sent in response to the "error", only one did not change the subject, and that message was written in Russian, and directed to Sergei himself.
The first thing that can be observed is that after the error, Sergei was silent for 5 days. This silence seemed to have maintained the thread, and one may imagine that if Sergei had been able to apologize immediately, the thread would have ended. In fact, about 24 hours after the error, one message openly requested that Sergei respond (message 30), later that day another made a similar request (message 40), and another message (message 63) expressed concern for him not having responded yet. When he did finally apologize (message 70) there were few messages accepting his apologies, and the thread died off soon after. Here, as in face-to-face conversation, silence seemed to have strong communicative value.
Another striking example of silence is seen in message 23 and message 59. Message 23 was posted about 48 hours after Sergei's error, but it did not elicit any responses. So, the exact same message was posted 24 hours later, showing that the poster was really expecting a response, but once again there was none. This first message was a failed attempt to initiate a topic, and when the second also failed, the person realized that the topic did not interest the others. "Unlike spoken conversations, an attempt to begin an exchange, especially on a public list, may be completely ineffective and receivers may not recognize any need to reply." (Daly, 1996)
This brings up the question of how to regard a message that receives no answer. Is it a first pair part that has not elicited a response? Or is a failed post a speech event in its own right? The other participants on the list did react to the message, by deleting it. Is that response sufficient to call it a speech event?
5.2 RFIs Within the Discussion
Within the discussion, which is itself embedded within the initial RFI, there are other RFIs to be found. Message 17 is a request that was answered, indirectly, by message 72, six days later. Message 31 is a message that was not actually related to this thread, but the comment in message 58, and the response to the comment, in message 62, did associate it with the thread. There was, however, no response to this request, at least not to the list, as was the case with the request in message 73, which, however, generated a discussion.
These embedded RFIs can be seen as a part of the larger thread, since they are "talking about the same thing", even though they are not direct responses to previous messages.
5.3 Multiple Speech Events in the Same Message.
As I said before, one message, or move, may contain more than one speech event, or moves from different speech events. There is not always a clear borderline between speech events, just as is the case in face-to-face conversation. The situation in e-mail is, however, compounded since there is no absolute limit to the length of a message. One message may contain a response to one speech event, the initiation of another, and even, perhaps, another.
One example of this is message 6, where one person responds to the large message, and then goes on to announce that they are "going nomail", or sending a command to the listserv to not send them mail until a new "mail" command is sent. This then is a response part of a discussion speech event, as well as an announcement.
Message 72 is a response to message 17, but the sender does not remember the original message, and therefore posts his information in the form of an announcement. But since the subject does not correspond to the original request, it is possible that the requester did not see the response.
A message may also contain a topic shift, such that it contains one move in a discussion, and another move which shifts the topic, and initiates another discussion. An example of this is seen in message 18 where the quoted text shows that the person is responding to a previous message, message 9, but starts talking about something different. The response to this message, message 36, picks up on this new topic, but does not change the subject. The response to message 36, message 64, does change the subject. Another example can be seen in message 74. It is an attempted humorous response to message 73, which changes topic, but not subject, with reference to the previous message. The person then responds in message 75, obviously not having picked up the humor of the message, and apologizes for her message, changing topic, but not subject.
Examination of the messages of this thread where the subject changed shows that most of the time there is no quoted text in a message with a new subject. This correlates with the idea that a message with a new subject line is an attempt to introduce a new topic. Yet this is not always the case. There are cases of topic shift, for example, message 64, where the new subject, "Josh's Silence" follows from message 36:
Subject: Re: ISSUE: Sergei N. T.
I'm not silent. I'm just thinking.
> With you, DT and Josh all silent for the same period of time, and LANTRA
>down for a while, I was bored silly.
>
>Love,
>
>Vladimir/Werne
>
Josh W.
This message is a response to message 18, which contains both an old topic (receiving the large message) and a new topic (Josh's silence and being bored). Josh's response focuses only on his alleged silence, and the response to this, message 64, quotes the text he wrote, and changes the subject:
Subject: Josh's silence
>I'm not silent. I'm just thinking.
It's my opinion that the first is almost a precondition for the second.
Salvador in Mexico
Aside from this example, the general rule seems to be that a message with a new subject contains no quotes, while a message which changes topic may or may not have quotes within it. One example of this occurs in the message that was sent in response to Sergei's message, message 4, which has no quotes, but a continued subject. This is probably, as I mentioned before, because the person did not retain the original message, and replied to message 3.
5.5 A Closer Look at one Discussion
The diagram in appendix I shows the entire "Sergei" thread, which contains 82 messages. The following illustration (figure 3) is a part of that thread, which consists of 28 messages, and all have the same subject:
Subject: Re: ISSUE: Sergei N. T.
This was the most fertile part of the overall thread, and was the longest in both volume (28 messages) as well as time (5 days). Also, 23 people participated in this sub-thread, out of 37 for the overall thread.
It can be seen here how the discussion branches out in different sub-threads, but how, through time, one of them is maintained and others die out.
None of the participants in this part of the thread made more than 3 moves; there were 6 people who made 3 contributions, 3 who made 2 contributions each, and the remaining 15 people only sent 1 message each. Looking at the messages, most of this thread is admittedly not a discussion of the kind where people present different points of view about a given subject, and debate these viewpoints. Here it is more a question of each person adding to what was said before.
This sub-thread started with message 9, which was one of the many responses to Sergei's error in message 2. 7 responses came to message 9, and 4 of those responses were answered. Sometimes, however, it is difficult to know what message a given message is responding to. Message 65, for example, I have placed as a response to message 9. Since it contains no quoted text, it could be a response to any of the messages in the thread. It fact, it would probably have been more accurate to have inserted this message elsewhere in the thread, chronologically. It would then come between message 61 and message 70. But since many people are in digest mode this makes it difficult to know which message it is meant to respond to.
There is actually a synergy that is built up over time, which consists both in the way messages follow each other, in terms of reference (one message quoting and responding to a previous message), as well as the strictly chronological order of the messages. A message on a given subject, which is not a direct response to a previous message, still has an important role in adding to the thread, because of when it is received in relation to other messages..
The greatest complexity in analyzing discussions is this asynchrony, which makes it uncertain where a message fits into an overall time scale. This non-linearity is perhaps the biggest difference between face-to-face discussions and e-mail discussions.
The discussion that follows looks very much like a spoken discussion. Each person expresses agreement or disagreement with a previous person, and the subject evolves, starting from one rhetorical question of:
Did I dream, or did we just got mailbombed by this twit?
Anybody for sending him the Gideons Bible as e-mail?
through the different contributors' agreement or disagreement with this suggestion, until Sergei himself comes forward and apologizes, and his apology is accepted.
The first response to message 9 was message 20. This message copied the previous message, and added a comment to it:
>Did I dream, or did we just got mailbombed by this twit?
>Anybody for sending him the Gideons Bible as e-mail?
You're not dreaming. I managed to kill the download when I realized how
long the message was taking to transit the wires. I don't know what this
guy was thinking, but he sent a 1 MB file to the entire list! Banishment or
exile, which will it be?
This is the most common form of move in a discussion: the quote and response. This enables, as I mentioned before, the response to be contextualized, for those who may not remember the original comment, and also links the response to the previous message.
The next response, message 39, is an in-joke. Daniel T. is an occasional member of the list whose discourse has a very particular style. The response is meant to be humorous, and is answered by another humorous message, message 45.
The next response, message 41, quotes message 20, which in turn quoted message 9, and shows the two levels of quotes, as well as the response:
>>Did I dream, or did we just got mailbombed by this twit?
>>Anybody for sending him the Gideons Bible as e-mail?
>
>You're not dreaming. I managed to kill the download when I realized how
>long the message was taking to transit the wires. I don't know what this
>guy was thinking, but he sent a 1 MB file to the entire list! Banishment or
>exile, which will it be?
>
Dear Robert,
You re completely right, a member of LANTRA that does not have any respect
for his companions has to be excluded from this list. There must be some
basic knowledge on what to do and what not to do with all the bandwidth. It
was realy disgusting! Why this person thinks that everybody on the list
needs all this Sh....t. [...]
This message generated 4 responses, the most in this sub-thread. Its nature, a proposition of mailbombing Sergei, was a bit controversial, and, indeed, 3 of the 4 responses were against this idea, with the other, message 48, an attempt to be humorous. Message 43 shows quoting where there is more than one comment interspersed within the quotes:
*> You re completely right, a member of LANTRA that does not have any respect
*> for his companions has to be excluded from this list. There must be some
*> basic knowledge on what to do and what not to do with all the bandwidth. It
*> was realy disgusting! Why this person thinks that everybody on the list
*> needs all this Sh....t.
Dear Bernard,
I strongly disagree with you. If a person sent a huge file to the list, it
doesn't automatically mean that he 'does not have any respect for his
companions'. On the contrary, Mr. T. has been one of the most valuable
contributors to the list over the months, as least for the Russian speaking
Lantrans. Calling his contributions the way you called them is, IMO, the
sign of disrespect to the members of the list. Of course, the file was sent in
a wrong way, and I think I was the first to indicate it but calling for
exclusion from the list on the basis of one mistake is really too much!
*> I sincerely think that Helge should take into consideration such lack of
*> respect for all the members of this list and exclude this person from its
*> participation.
*> I propose that everybody sends to this guy (to his e-mail directly) a huge
*> worthless file of several MB.
*> Somehow the silence was nice, dont you think so?
E-mail terrorism is not the best solution.
As I mentioned previously, this type of quoting allows the writer to address the different points of a previous message, in the order they were written. One response to this message, message 56, requotes the entire message, and suggests that the argument be dropped. And the following messages, 57 and 60, go in the same direction. Then message 61 is another complaint.
Message 70 marks the return of Sergei. Although his messages betray a lack of familiarity with e-mail discourse (using such structures as 'Dear Sir', in message 2, and 'Dear LANTRA members' here, ending this message with 'Kind regards'), he chooses to use the subject line that has been the most productive:
Subject: Re: ISSUE: Sergei N. T.
for his message. This ensures that all those who have been participating in this thread or following it will see his message. Had he used a subject such as 'I'm Back' or something else, the rest of the people might not have read the message.
Two people respond to his message, accepting his apology, and, as I noted before, once he does come forward to apologize, there is nothing else to say, and the thread quickly dies off. "...there are no formal closures - participants tend to let an exchange just die if they see no purpose in carrying it any further..." (Daly, 1996)
This discussion, in spite of the number of messages it contains, is relatively simple. Most of the messages respond to only one previous message, and most of the time they contain just one bit of quoted text. Discussions can be more complex, as seen in the example in section 4.3.2, where the text is inserted between the quoted text in a number of different places, and where there may be more levels of quotes.
Mailing lists are a relatively new genre of written communication, which contains elements of both written and spoken discourse. Since CMC is becoming more and more common, it is interesting to examine how this genre relates to both spoken and written language, and one way of doing so is by examining the speech events that are used within it.
Mailing lists generate three types of speech events, which are the Announcement, the Request for Information, and the Discussion. These speech events are not always isolated speech events, but, as I have demonstrated, they are often found embedded within other speech events.
The announcement and the RFI are speech events that are relatively simple to describe. The patterns they follow are simple, and these patterns seem to be the same regardless of the subject matter of the mailing list.
The Discussion is much more complex, and amorphous, due in part to the asynchronic nature of CMC. As in face-to-face conversation, there are many possibilities for different participants to intervene, but 'turn-taking' in e-mail, not being subject to the same kind of rules, makes it possible for anyone to step in at any time. There are also certain specificities which set CMC Discussions apart from their spoken counterparts, such as subject headings and quoting.
Understanding these speech events is important to understanding the specificities of CMC, especially since it is being used more and more by people from different cultures. The examples of speech events that I have examined in this paper show that these speech events are valid even across cultures (while I have only examined data in English, a large percentage of participants on the mailing lists I have studied does not have English as their mother tongue). Part of this similarity is due to the evolution of e-mail, since it first started being used in the United States: structures and norms were adapted from English to other cultures. But a large part of this is also a result of e-mail software (again, usually American), and the way it enables the user to compose messages. The evolution of this software obviously represents a desire to model the technical possibilities on the communicative reality.
It remains to be seen whether e-mail will continue its vertiginous growth over the years, with the advent of inexpensive Internet telephony. As Danet (Forthcoming 1) points out, "There are, of course, risks in undertaking any analysis of the current response to new communication technologies. They are changing so rapidly that even the most cautious of today's research generalizations may become obsolete by the time they appear in print or even in an electronic journal." Much of the early research is CMC is now outdated, because of technical changes.
Nevertheless, this examination gives us insights into a form of communication that is very much alive and dynamic, and which is used extensively by large numbers of people. Such an analysis can help us better understand how interpersonal communication functions through this relatively new medium.
Brown, Penelope, and Levinson, Stephen. (1987). Politeness: Some Universals in Language Usage. Cambridge: Cambridge University Press.
Cole P., and Morgan J., eds. (1975). Speech Acts (Syntax and Semantics vol. 3) New York: Academic Press.
Daly, Bill. (1996). 'Electronic mail: strangely familiar texts.' Unpublished paper.
Danet, Brenda. (Forthcoming 1). Books, Letters, Documents: the Changing Materiality of Texts in Late Print Culture.
Danet, Brenda, (Forthcoming 2). Computer-mediated Communication. In The Garland Encylopedia of Semiotics, Paul Bouissac, General Editor New York: Garland
Danet, Brenda, (Forthcoming 3). Flaming. In The Garland Encylopedia of Semiotics, Paul Bouissac, General Editor New York: Garland
Goffman, Erving. (1967). Interaction Ritual. New York: Pantheon Books.
Goffman, Erving. (1974). 'Replies and Responses', in Goffman (1981).
Goffman, Erving. (1981). Forms of Talk. Philadelphia: University of Pennsylvania Press.
Grice, H. P. (1975) Logic and conversation. In Cole and Morgan, eds. (1975)
Gruber, Helmut. (1996) Aspects of power and communication in scholarly e-mail discussion lists. Excluding practices and peer-group formation. Paper presented at the 1st Knowledge and Discourse Conference, Hong Kong, 18-21 June, 1996.
Gumperz, John J. (1982). Discourse Strategies. Cambridge: Cambridge University Press.
Gumperz, John J. and Hymes, Dell. (1972) Directions in Sociolinguistics. Oxford: Blackwell.
Halliday, M. A. K., and Hasan, Ruqaiya. (1976) Cohesion in English. London: Longman.
Herring, Susan. (1993). "Gender and democracy in computer-mediated communication." Electronic Journal of Communication 3(2), special issue on Computer-Mediated Communication, T. Benson, ed.
Kawasaki, Guy. (1996) The Rules of Email. Document distributed with the program Claris Emailer 1.0.
Levinson, Stephen. (1983). Pragmatics. Cambridge: Cambridge University Press.
Maynor, Natalie. (1994). 'The Language of Electronic Mail: Written Speech?' Centennial Usage Studies. Publications of the American Dialect Society Series. University of Alabama Press.
Negroponte, Nicholas. (1995). Being Digital. New York: Vintage Books.
Sacks, H., Schegloff, E. A., and Jefferson, G. (1974) A Simplest Systematics for the organization of Turn-taking in conversation. Language, 50:4, 696 - 735.
Saville-Troike, M. (1989). The Ethnography of Communication. Oxford: Blackwell.
Schegloff, E. A. (1972). Sequencing in Conversational Openings. In Gumperz and Hymes (1972).
Schiffrin, Deborah (1994). Approaches to Discourse. Oxford: Blackwell.
Shank, Gary. (1993). 'Abductive Multiloguing: The Semiotic Dynamics Of Navigating The Net', Arachnet Electronic Journal on Virtual Culture v1n01 (March 22, 1993) URL: ftp://ftp.lib.ncsu.edu/pub/stacks/aejvc/aejvc-v1n01-shank-abductive
Sutton, Laurel A. (1995) A Gricean Analysis of Netiquette Rules. Paper presented at the 1995 GURT Conference, Georgetown University.
Talbott, Stephen L. (1995) The Future does not Compute: Transcending the Machines in our Midst. Sebastopol: O'Reilly & Associates, Inc.
Veselinova, Ljuba, and Dry, Helen, (1995). 'Queries on the LINGUIST List: Acquisition of a Subregister' Paper presented at the 1995 GURT Conference, Georgetown University.
This paper is the result of months of work, and a number of people deserved to be thanked for their help.
Keith Richards, who supervised it, was everything a supervisor can be. If only we could have met and talked about it over a beer.
Julian Edge, my head tutor, was there, as always, to answers questions. Thanks, Julian. It has been a long trip, but I made it.
Tom Bloor, who really was not supposed to have anything to do with it, nevertheless patiently answered some questions while others were on vacation.
Helmut Gruber read the paper at an early stage and made same very valuable comments.
Richard Epstein who, in spite of just becoming a father, gave me much valuable feedback.
Alexanne Don and Bill Daly, who kept up interesting conversations about the subject by e-mail.
Many other researchers who answered questions, sent me papers, and gave me disinterested help. Thanks to all of you.
The British Council in Paris; especially Jon Warr and John Spears, without whom the course would not have been possible; Roger Budd, Bhasi Panikker.
All the others on the course.
Marie-France, who typed a lot of it, reread it all, and put up with the hassle.
Perceval who understood that his father had to work on the weekends because he was still in school.