2. ELECTRONIC MAIL DISCOURSE

Messages transferred as electronic mail are typed texts, just as traditional mail ever since the advent of the Written culture has been written text. An e-mail can be like a formal document, a letter or a memo - it may have been created off-line and exist in paper form as well. The correspondence of e-mail is asynchronous; that is, the interchange between people may be spread over a period of several days. Thus, in many ways e-mail is written communication comparable to the correspondence of traditional l etters. Some users may indeed utilize the medium just as that - a simple substitute for traditional mail. On the other hand, the advantages of the medium are far beyond the scope of such use and most users agree that electronic mail has features which significantly surpass traditional mail. Correspondents in an e-mail exchange may discuss topics asynchronously or take turns in instant messaging. Electronic mail is instantaneous in its transmission and may be instantaneous or asynchronous in its reception - the recipient may choose to read the mail at any convenient time. Via e-mail data can be shared and copied more easily than via traditional mail and individuals can rapidly reach multiple addressees. Electronic messaging may share the properties of trad itional mail, but it may also share the properties of the telephone or face-to-face communication. E-mail consequently transgresses the boundary between traditional notions of written and oral communication.

The instant and unpretentious nature of electronic mail encourages immediacy and spontaneity. In electronic mail discussions messages are often sent and replied to with great ease and without much careful consideration. E-mail messages are usually informa l and often short, quickly exchanged notes with an obvious resemblance to verbal conversation (McMurdo, 1995:145). Therefore, e-mail discourse is often put on a par with face-to-face communication. In reality, however, fluid limits obtain between the spok en and written nature of e-mail. In regular face-to-face conversations, we orient ourselves not only by what is being said, but also by interpreting by whom, to whom and in what way utterances are expressed. By analyzing the context in which interaction o ccurs we attempt to understand the intended message in perceived utterances. Context cues, such as intonation, pauses, hesitation, rate of excitement and facial expression contribute to our perception of hidden implications (Tannen, 1986). Speakers find i nnumerable ways of saying one thing and meaning another. Irony, sarcasm and figures of speech are our devices and in the right interpersonal atmosphere they work - listeners successfully decode intended messages. In electronic mail, however, these devices often fail - most "listeners" decode messages literally. This happens particularly when participants in the discourse are not previously acquainted with each other. Thus, when engaging in an electronic mail discussion, such as in a newsgroup on Usenet (sci.lang), we may encounter messages with intended sarcasms that are perceived and replied to as factual statements. The lack of social and non-verbal cues principally leads us to resort to literal interpretation (Rudy, 1996; Arrow, 1997; Hightower and Sayeed, 1996) - as in traditional reading.

The relative paucity of social and non-verbal cues could easily boost the conception that e-mail is simply written text albeit immediate and spontaneous in nature. There are, however, numerous features in e-mail messages that evince the spoken nature of t he medium. In e-mail and other text-based computer-mediated communication, experienced participants largely manage to convey their contextual cues semiotically. This is done for example by including so-called emoticons (Hightower and Sayeed, 1996:463; Jac obson, 1996:467) or even explicit statements about the author's frame of mind. Emoticons are small symbols constructed from various combinations of keyboard characters and typically read sideways (for a comprehensive list of common emoticons, see Appendix 1). They express various attitudes and facial expressions and therefore enable quite detailed and emotional discussions. The most frequently used emoticons are those for a smile and a frown, usually represented as :) or :-) and as :( or :-(. A smile may indicate warmth and humor, but it may also connote, in ironic fashion, the opposite meaning (Jacobson, 1996). E-mail authors who venture ironies therefore often add a ;-) to convey the intended meaning or even make explicit statements about it. Explicit s tatements entail placing a word or a phrase within angle brackets (such as <smile>, <grin> and <frown>) or asterisks (*smile*), or simply declaring "I'm kidding!" or "I'm serious." The frequent use of semiotic social expressions in e-mai l messages indicates that senders intend to convey their messages as spoken communication.

The tone of an e-mail message can be made clear in more ways than by the use of emoticons and explicit statements. The informal and conversational style of e-mail discourse is manifested in the fact that many users, consistently or temporarily, use lower- case letters in their messages (see sample below, fig. 1). E-mail with consistent use of lower-case letters is particularly common in personal conversation between users with a long experience of e-mailing. (Postings on mailing lists and in newsgroups ten d to have a more formal tone with, for instance, initial capitals in sentences and personal names, and fewer emoticons.) E-mail with lower-case letters throughout efficiently mark the speech-like nature of the medium and the casual tone of messages. Furth ermore, using lower-case letters increases the "economy of writing" (Goodman and Graddol, 1996:120). The process of typing becomes easier and faster, thus approaching the rapidity of spoken discourse.

Bolter (1991) states that the electronic production of text, in word processors and in computer communication tools, represents a new economy of writing. The economy of any writing tool, whether it be clay, paper, pen or keyboard, naturally depends upon t he writer's set of skills and practices around the tool. With longer experience writers become eager to facilitate the process and render it more effective. In fact, the facilitation of production of written text is what once caused the Written Culture to turn into the Print Culture. The electronic production and transmission of text in the Electronic Culture does not mark the end of our journey towards the most economical way of writing. Actually, our malleable modes of electronic communication leave the door open for new ways for communicating in writing - of which the consistent use of lower-case letters is only one.

Another economic and efficient way of communicating in e-mail is what can be called quoting. It is a feature unique to computer-mediated communication that surpasses the domains of both speech and traditional writing. Quoting is made possible by the moder n computer's ability to easily copy and "paste" text and means that, in replying to a message, a sender may copy relevant fragments from the message and paste them into the reply (see sample above, fig. 1). This allows the sender to quickly recapture what was said in the message and immediately reply to just that. Anything the sender types him/herself relies for its meaning upon the overt or latent reference to the copied text. By copying and pasting, it is very easy for participants in an e-mail exchange to refer to large amounts of each other's texts (and the texts of others) and make the point clear. E-mail software is often constructed to facilitate the pasting by suggesting an angle bracket, >, to precede quoted text. Most software also recognizes wh at text is recycled for the second time and suggests a double angle bracket, >>, for such text, a triple angle bracket, >>>, for text recycled for the third time and so on. Generous quoting, as in the example above (fig. 1), can make a message confusing. Yet, experienced e-mail readers usually have no difficulty in untangling messages that consist of a collection of quotes. Nevertheless, the limit for quoting is reached when the messages do become difficult to decipher. Some quoting can also be considered unnecessary - particularly when the point can be made equally clear without quoting. Network administrators, to save on bandwidth, often discourage superfluous quoting - ironically enough in conflict with our desired "economy of writing". Providing dupl icate messages means demanding a great deal of space in e-mail servers - space that may be scarce.

Due to the explosive expansion of the number of Internet users in the mid-1990's, the "information superhighway" (White House Press Releases, 1994) is often cramped for room. Consequently, taking up too much space is viewed as bad "netiquette" (McMurdo, 1 995). Besides quoting, another common strategy of e-mail communicators, eager to enhance the "economy of writing", is cross-posting. Cross-posting means providing the same posting to multiple newsgroups. Many users of newsgroup services post superfluous n ewsgroups in their zest for reaching a greater audience. According to server administrators, however, superfluous cross-posting wastes bandwidth. One posting in one relevant newsgroup normally suffices for anyone's personal communication demands - no matt er what the purposes may be. Superfluous cross-posting is mainly done by commercially interested parties and commercial messages are banned in newsgroups. By encouraging precision target group postings, network administrators attempt to save bandwidth, bu t also to shut out commercially interested parties. Network administrators are responsible for the efficacy and maintenance of networks and servers. By discouraging superfluous quoting and cross-posting they see to it that the network bandwidth suffices f or everyone's communication demands.

As we have seen, network administrators may have a great deal of say in our e-mail discourse strategies. Furthermore, network administrators are responsible for the storage of our e-mailed material for a certain period. Within an organization, e-mail may be stored on backup disks for as long as ten years. This means that when a person retrieves, reads and deletes a message, the message may still be left on the server and later recorded on a backup disk. The ability to compile and preserve large amounts of typed information in computers, servers and backup disks resembles the ability to store written and printed material in bookshelves and libraries. It is therefore the one feature which most points towards a typically written nature of computer-mediated t ext. Spoken words are ephemeral, but written, printed and typed words are not. Senders of e-mail often disregard the fact that people other than the intended recipient can read the message and that the mail is stored in a server by the recipient's organiz ation or e-mail provider. In the USA old e-mail is sometimes retrieved for legal purposes and presented as evidence in trials. Recently, a number of US corporations were sued by their black employees who claimed that they were discriminated against by rac ist e-mail that colleagues, including the management, had sent (Horner, 1997). The stored e-mail messages were retrieved and presented in the courtroom proceedings. Network administrators, due to lagging legislation in the field of e-mail and secrecy, usu ally release this kind of information upon request. Naturally, attorneys on both sides in cases like the one mentioned are relieved to have the information at hand. The ability to present the jury documents in black and white is more reliable than any cou rtroom testimony about inappropriate comments. The instant and unpretentious nature of e-mail too often encourages the transmission of uninhibited inconsiderate messages. The mistake of disregarding the fact that e-mail messages are only ostensibly epheme ral can cost senders dearly. Owing to the emerging lawsuits, American corporations are currently introducing strict rules of conduct for e-mail, which reduces their risk of being sued. Some corporations go as far as to monitor what is said in the e-mail c ommunication and others make sure that all e-mail is deleted upon the first retrieval from the server or within a short period of time.

Rules of conduct for e-mail and other computer-mediated communication is referred to as netiquette. Individual organizations and societies make up their own set of netiquette rules, some very detailed and others rather superficial. As the Internet is not managed by any single authority, there is no single authority that watches over Internet conduct. Consequently, there is no common book of rules to refer to when we encounter bad netiquette. Codes of good practice for e-mail are found on innumerable sites on the Internet. Netiquette rules often attempt to control the nearly uncontrollable spoken e-mail discourse by establishing some proper graphic and orthographic conduct. Some organizations encourage users to use standardized letterheads and formal langu age just like in traditional mail - thus disregarding the dynamics of the medium. Other organizations recognize the dynamic features of e-mail and primarily make users aware of how to avoid rude behavior. One particularly well reasoned netiquette compilat ion is found on the Oxford University Computing Services' site (see Appendix 2). It regulates the way e-mail users at Oxford University are to behave in their communication with others - both internally and externally, semantically and syntactically. In t raditional communication media, like mail and the telephone, certain common and widely observed conventions have emerged over time. Courtesies as when to use 'yours sincerely' in a letter, or announcing your leave from a telephone conversation, are "not just pointless conventions". Rather they "help promote a sound basis for communication between the relevant parties" (Reid, 1994). The Oxford University netiquette compilation is intended to promote a sound basis for proper electronic communication. It regulates the linguistic and sociolinguistic aspects of electronic communication with insight and foresight.

E-mail is a relatively new form of communication and its conventions are still in the making. The numerable lists of netiquette indeed point towards an increasing awareness of appropriate conventions. However, considering the global communication possibil ities provided in CMC, it is fair to suggest that the conventions will never be as widely observed as they are in the traditional communication media. Users in too widely dispersed cultural communities with too varying communication strategies inhabit the Internet to be able to generate conventions in perfect unison. Regarding the syntactic use of the English language, however, certain netiquette conventions have become commonsense - conventions that evince the spoken nature of electronic communication. O ne is the convention that says "all uppercase sounds harsh, like shouting" (Krol, 1994 quoted in Svensson, 1997:222). Consequently, including upper-case words in a message makes it difficult to read, and writing a message with solely upper-case letters is considered rude. Emphasis on a certain word can be made with upper-case letters (she FORGOT to), though this should be avoided. Rather, emphasis should be made by enclosing the word with asterisks (she *forgot* to) or underscores (she _forgot_ to), or by spacing the letters of the word (she f o r g o t to). New users of the Internet, often called newbies, are those who make the most mistakes in their use of CMC, but in the course of time they adopt the emerging conventions. Often newbies overuse abbrev iations. Abbreviations and acronyms are indeed common in computer-mediated text, but should be used with discretion. Chapter three of this essay discusses the use of recognized abbreviations and acronyms in electronic communication.

The conduct of newbies tells us about the results of initial enthusiasm, but it also tells us about the way correspondents in an apparently oral situation try to overcome the inability to convey social, emotional and linguistic cues. Certain features are common even among experienced communicators in CMC. By using upper-case letters correspondents try to compensate for not being able speak the words louder. Thus, upper-case letters convey a certain amplitude of intonation. By using abbreviations and acron yms, they compensate for not being able to speak fast enough. Thus, abbreviations and acronyms are intended to speed up the reading of the message. Further, with syntactic devices like ... and !! they convey features such as hesitation and pause and a higher pitch of voice. By enclosing phrases within a <whisper>, senders may hint a whispering voice or a lower pitch of voice. Willem Levelt in his book Speaking (1989) writes about the organization of the skill of speaking. In speaking we are able to "trans form intentions, thoughts, feelings into fluently articulated speech" (Levelt, 1989:1). By organizing our speech into complex combinations of tone units and prosodic units, we convey our mood, intentions and frame of mind. In electronic communication, we apparently recognize to a great extent the need to convey the same features. All in all, the frequent use of typed oral cues seems to consolidate the connection with spoken discourse.

What further consolidates the connection with spoken discourse, is the fact that the characteristics of an electronic discussion resemble the turn-taking that takes place in oral discussions. In personal electronic communication as well as in mailing list and newsgroup communication, one e-mail is the output of one speaker and the input of one or many listeners. E-mail software recognizes interlocutors' need to structure ongoing communication. Therefore, it conveniently sorts the input and the output in l ogical ways. In face-to face (FTF) discussions, interlocutors in their minds quickly sort out what is being said, by whom and to whom, by the linguistic and extra-linguistic cues of the ongoing communication. In e-mail discussions, similarly, the computer software sorts the input and the output into various "mailboxes". Within a mailbox, a mailing list or a newsgroup discussion, the e-mail software sorts the messages into so-called threads. A thread is a collection of messages with a common subject. E-mai l allows several discussions (threads) to be going on simultaneously. Usually replies on a topic are classified by the prefix Re: in the subject line, and forwarded messages are classified by the prefix Fwd:. Some research into the structure of threads ha s been produced. Fafchamps et al. (1989) describe how e-mail discussions on a topic may be divided up into islands, dialogues and webs of communication. An island is a message that receives no reply and a dialogue is a series of messages between a pair of individuals on a single topic. A web is a complex structure of messages, where a single message may receive replies from many users, some of whom are replying to issues raised in more than one message. To be able to untangle the structure of threads, one needs a certain amount of experience in e-mailing, preferably on mailing lists and in newsgroups. Bordia (1996) declares that longitudinal studies show that, given time, users feel that CMC approaches the nature of FTF communication in that they begin to conceive of the structure of threads as the structure of turn-taking in FTF discussions.

The computer-generated construction of mailboxes and threads and the other syntactic, semantic and computer-generated functions mentioned in this chapter, altogether make e-mail comparable to the structure of face-to-face communication. As we have seen, e -mail communication is a complex phenomenon. It has features that support notions of traditional written discourse (the durability aspect, for instance), but it also has features that support notions of spoken discourse (devices for conveying non-linguist ic cues etc.). In addition, e-mail introduces features that are entirely new to communication, such as software-generated quoting, cross-posting and threads. E-mail does not conform to the traditional domains of spoken and written discourse, but constantl y transgresses the limits between the two. Therefore, e-mail must be said to create its own domain of discourse in the history of communication. Now, if comparing e-mail to face-to-face communication turns out to be profitable, despite the asynchronous mo de, what is then the case with synchronous CMC? In the next chapter, the nature of synchronous CMC is scrutinized.