For many readers of this present volume, one of the central notions introduced in the title above, namely that of a `distributed virtual environment', may not be immediately transparent, and if we are to go on and discuss what is involved in making ethnographies of conversation in such environments a brief initial explanation is probably needed.[1]
In our present context the term `distributed virtual environment' will be defined in its most general sense as any technologically mediated environment which facilitates, while simultaneously depending upon for its existence, human communication, interaction and resource-management within the virtual environment, and which facilitates similar communication, interaction and resource-management processes both locally and globally in non-virtual environments.
In practice this means that individuals and groups located geographically more or less anywhere in the world are able to log on via their respective computer networks into a shared interactional space - a distributed virtual environment - allowing them to communicate and cooperate at a distance with others with similar access to the same virtual environment.
Those individuals, groups and institutions who take part in the creation, management and development of the virtual environment are themselves distributed since they are geographically distant from one another and are not physically co-present at any one location while they are interacting with one another. Any sense these actors have of being members of an emerging virtual community is primarily intersubjectively instantiated and construed. It is, however, also materially instantiated due to mediational aspects of the technology used to maintain the environment and to communicate there. The emergent virtual community and its environment depend on contributions from individuals and groups who, due to geographical distance, may seldom or never meet one another face-to-face. Contributions may be both communication- and action-oriented, and relate to the creation, management and exchange of information and material resources not only `internally', i.e. `within' the distributed virtual environment, but also locally and globally in real world environments `external' to it.
A locally nourished, globally facilitated community of this kind is both distributed and virtual since it depends for its existence on the coordination of a number of local initiatives with their origins in different physical, socio-cultural and technological domains, while at the same time existing `factually' only in the individual and collective minds of participants. For those who consider themselves, and who are considered by others, to use Goffman's term: `ratified participants'[2], in the emergent virtual community, the distributed environment represents a distinct and unique form of technologically mediated socio-cultural environment. This is one of the main reasons why making ethnographies of conversations which have their origins in distributed virtual environments is an appropriate way of understanding the underlying structures, functions and socio-cultural norm systems operative in these environments.
Groups of institutional co-workers, such as those mentioned in the article by Violi in this volume, who use electronic mail on a daily basis to communicate in the context of their working environment are participating in the development of a distributed virtual environment. In such environments personal and institutional messages are routinely sent from one-to-many, or from many-to-one within the organization, not only from one-to-one, as is generally the case with conventional paper letters (Violi 1985).
The network infrastructure, software and communication protocols that make electronic mail exchanges technically possible constitute the technological mediation of the environment. Messages exchanged by participants, the individual and collective meanings that they attach to these, and the everyday practices and habits that develop through using this form of communication represent the socio-cultural `glue' which instantiates the socio-cultural sub-system that is the organization, as well as participants' construals of themselves as members of this organization. There is, in other words, a dynamic complex of reciprocal interdependencies between the technologies which create and maintain distributed virtual environments, and organizational systems of norms for communication and interaction which emerge as a result of members participating in various kinds of co-operative activities in these environments.
`Conversations' by means of electronic mail are considered to take place when two or more persons use this form of communication in order to discuss some specific theme over an extended period of time by exchanging messages related to this particular theme.
Electronic mail `conversations' differ in three fundamental ways from face-to-face conversations:
1. they are carried on in written, rather than in spoken language
2. they are managed non-contemporaneously rather than contemporaneously - interlocutors receive, read, compose, edit and send messages whenever it suits them
3. interlocutors are not bound to a context of situation involving temporal and physical co-presence while communicating
Similar constraints, with some additional considerations regarding use of semiotic systems other than written language, will be applicable to `conversational' activities involving World Wide Web documents in non-contemporaneous forms of communication. In the next section we shall examine this issue more closely.
The World Wide Web is a global hypermedia system which makes locally created and stored documents available for recipients at geographically non-local sites all over the world. Hypermedia documents distributed in this way do not generally contain solely textual information. Other semiotic elements: icons, images, graphics, animation, sound and video sequences are often embedded in these documents, hence the common use of the term `hypermedia'[3] used above to denote both the general type of global distribution system, as well as the particular type of documents which the system distributes.
Hypermedia documents may contain indexical links which allow readers to `jump' more or less instantaneously to similar documents with related information stored at other sites. Distributed hypermedia documents from non-local sites may be retrieved, stored and printed out locally. They may also provide access to distributed software resources (generally known as applets) which reside at non-local sites and which are linked to their parent documents through scripts so that they can be `called up' and run locally whenever the functions they are designed to provide are needed. In addition to being navigatable (by following links) and readable, hypermedia documents also allow other forms of direct interactions on the part of readers, such as filling out forms or questionnaires for the non-local collection of information, and activation of search engines to retrieve information from non-local databases.
The World Wide Web may be seen as an non-contemporaneous form of distributed virtual environment, providing access to huge numbers of interlinked, cross-indexed, multi-levelled hypermedia documents, each of which may be considered as representing a tiny sub-environment in a vast global virtual library. Visitors to the library `enter', read and explore the environment by navigating within and between sub-environmental `nodes' through `gateways' created by locally run software resources (Web browsers). In principle any individual, group or organization with access to the Internet who wishes to establish a World Wide Web entry point or `node' - a `Web Site' - of their own, can do so by setting up a computer server to store and make accessible hypermedia documents and other digital resources, and by creating and inviting the establishment of links to and from these documents to resources at other sites.
Since we are primarily concerned here with `conversational' aspects of this technology, we shall go on to consider some types of exchange carried out by means of the World Wide Web which might be subsumed under the more general notion of `conversations'.
Face-to-face conversations are first and foremost contemporaneous events - they happen in real-time. They are also sequentially organized - one form of communicative action (an initiative or response) follows upon another, generally in rapid sequence. There is a degree of cohesion between grammatical elements actualizing communicative actions, and a coherent form of semantic structure is developed in the course of the conversation and this is reflected in the weaving together of elements of meaning which actualize global, thematic and local coherence levels[4].
As we have seen previously with the case of e-mail, any extended, cohesive, semantically coherent initiative-response sequence occurring over time may in a very general sense be considered `conversational'. This includes also those sequences which are developed non-contemporaneously. This means that non-contemporaneous exchanges initiated by means of hypermedia documents which invite to interaction with their creator(s), either through electronic mail or the insertion of data into interactive forms, crossing off of choices on a list etc. may also be labelled as `conversational' in this wider sense of the term. However, since interlocutors are neither physically nor temporally co-present while taking part in non-contemporaneous conversational activities of this kind, important phenomenal factors in face-to-face communication which cannot necessarily be compensated for by other forms of expression such as writing, still or video-images, graphics or animations are lacking. These include direct perceptions of interlocutor appearance, global body language, gaze, facial expression, eye-contact, proximal relations of self to other and environment, movement, tactile impressions, body warmth, body and breath odour.
Similar considerations with regard to a general kind of `conversationality' may also be applied to certain other types of human-machine interactions. Into this category fall activities like using natural language processing systems to search distributed databases, navigating the World Wide Web through hypertext links evoking language-based choice trees, and managing file hierarchies on local or non-local computers using operating systems which invoke dialog boxes at choice nodes in the file management process. Conversation-like human-machine interactions of this kind display certain similarities to face-to-face conversation, in that the outcome of certain types of intitiative will often lead to some kind of conventional response. This is after all what we generally will experience in about eighty percent of all our everyday conversations.
On the other hand, however, human-machine interactions lack a shared sense of physical co-presence on the part of interlocutors (after all, one is embodied and the other not). They also lack a shared sense of intersubjectivity associated with the experience of verbal and non-verbal realizations of one another's emotional reactions to what is being said, and of one another's inferencing and interpretation processes. Mainly due to these last two factors, certain types of face-to-face conversation are characterised by a very high degree of spontaneity, creativity, open-endedness and non-linearity which makes them rather unique as forms of conversational activity, in this wider sense of the term.
Some other types of distributed virtual environments than those presented so far and which make semi-contemporaneous conversation possible at a distance do actually exist. These range from on-line conferencing systems developed around the use of video and audio technologies, to text-based environments like Talk and Internet Relay Chat Rooms (IRC) and finally object-oriented environments often known as Multi-user dimensions, Object-Oriented (MOO). All three last-mentioned types of distributed environment make it possible to arrange virtual meeting spaces where people all over the world can meet and chat with one another by writing in real-time.
Object-oriented environments, however, introduce further dimensions of interactivity: participants may move about from space to space within the virtual environment. They may also interact with and change the virtual objects[5] that go to make up the virtual environment. This allows them to design, create and modify various key characteristics of the virtual environment that they are interacting in (and with) while they are actually `in there'. Programmers can cooperate with one another at the same time as they are modelling and building the environment, discussing effects of modifying or adding new functions and structures with other participants while this is actually being done.
This is claimed to make programming and debugging work in object-oriented environments of this kind highly effective. Masinter and Ostrom enthuse for instance that:
"The multi-user aspect of MOO has several significant implications. First, it makes it easy for people to work together. When using an information resource privately, the user has to be lucky enough to have a helpful person around, or may end up mailing out a plea for help to some mailing list or newsgroup. When using MOO, the user is fairly likely to find direct interaction with someone helpful. In addition, MOO provides a way for people who are simply after the same kind of information to work together to find it. This is of course simple enough for people physically located in the same room, but collaboration across a greater distance can be very difficult.
Second, the fact that MOO provides easy interaction between people, and that MOO information tools, programmed in MOO code, can be easily manipulated from within an interactive session, means bugs can be fixed and new features and tools added very quickly. It is usually difficult for the implementor of a software tool to observe, first hand, how the tool is being used. In MOO, this kind of interaction is frequent: by observing the errors and misunderstandings of novice users, the implementor can see the ways in which a user interface is confusing or awkward. It's often hard for an ordinary user of a piece of software to provide adequate information for the programmer to locate a bug. But in MOO sessions, the users can talk about problems, the developers can work with them, possibly adding debugging information to the code as it's being used, to figure out what exactly is going on. And often the problems can be located and fixed on the fly, in a matter of a few minutes."[6]
This rather unusual `hyper-interactionality' of object-oriented environments, which allows not only for interactions with others, but also interactions with the virtual environment, as well as interactions with others about these interactions while they actually are going on, which makes conversations in these environments qualitatively different from conversations in other types of distributed virtual environment, bringing them, as we shall discuss in some more detail later on, even closer to that which actually goes on in most face-to-face situations.
This raises some provocative theoretical and methodological questions with regard to our more general understandings of what conversation is, and how it actually works. We shall suggest a methodology which is designed to allow for an incorporation of the specific hyper-interactionality aspect of object-oriented environments mentioned above into ethnographies of conversation in general. A simple analytical taxonomy will be presented which pays special attention to the accomodation of not only qualitatitive differences, but also similarities, between face-to-face conversations and conversations in distributed virtual environments. This taxonomy may be useful as a tool for making enthnographies of face-to-face conversations where what is going on comprises various forms of participant action, and is not merely confined to talk.It will also address the problem of how to categorize attempts on the part of participants to make the transition between certain types of non-ratified and ratified participant status.
Before we go on to look more closely at some characteristics of conversations in object-oriented distributed environments, first a few words on the notion of textual environments as contexts for conversation.
Meetings in object-oriented environments will take place in textual spaces such as the one reproduced below which purports to represent Piazza Maggiore at the heart of the medieval city of Bologna in Italy. A visitor entering this virtual space is greeted by the following text:
You are at Bologna.
You are standing in the centre of an expansive and beautiful piazza right in the very heart of Bologna. It is called Piazza Maggiore - the Grand Piazza. Straight ahead of you you can see the immense cathedral of San Petronio. Bologna is a medieval city in the very center of Northern Italy. Some people call it the Red City. Is it because the dominant colour of many of the walls of the houses there are painted in different shades of red...?
Nobody else is here.
Obvious exits:
Norway [to Patrick's Office], southwest [to Via Ca' Selvatica], and
portale [to San Petronio].
Seen from a material viewpoint, this spatial description is a purely fictional entity. It is a small piece of text which tells its own story, either this be autonomously qua text, or collectively as part of some larger hypertext - which in this case would be the current structure of the larger object-oriented environment to which it belongs[7].
At the same time, this description also represents context in a rather broader sense, namely that which systemic functional theorists[8] generally refer to as `context of situation'. This is because the virtual Piazza Maggiore above in practice can function as a conversational space which people from all over the world can enter, meet one another and talk. Like any other multicultural real-world context of situation it possesses a dymanic potential for change and development over time as a result of what is going on, who and what are there, who is doing what to who or what, and why they are doing whatever they happen to be doing. In the case of the real world Piazza Maggiore in Bologna, one effect of the sort of cultural activities that actually go on there might be the erection of a giant stage with a wall of video screens for some public entertainment event like an open-air gospel choir evening or a silent film festival. Our virtual Piazza Maggiore might be similarly modified (albeit with a good deal less physical effort and material expense) by organizers of cultural or other kinds of activities there by simply rewriting the description text, by introducing some new objects or otherwise changing the functional characteristics of the object space. So, although it is sparsely delineated, a simple spatial description of the kind shown above attempts to model in the form of a written text at least some aspects of `non-virtual', real-world contexts.
One reason for trying to do modelling work of this kind can be to test a general hypothesis that certain characteristics of textual contexts of this kind will come to affect the flow of communication and interaction in the virtual space in ways which are not necessarily specifiable in advance. In this sense a spatial description text like the one introduced above represents an exploratory, as opposed to regulative, hypothesis for investigating pragmatic issues regarding the relationship between text (structure), discourse (process) and context (field)[9]. There are also a number of important epistmological and ontological issues which may be brought into focus in this connection, for instance regarding to the actual ontological status of the code objects that go to make up the virtual environment, of which the spatial description above is an example. It will certainly be necessary to consider objects of this kind as dynamic, not only from a material point of view - they are reprogrammable in order to `look' or `behave' differently, thus changing their basic `nature', but also from a semiotic, representational point of view - their interpretations in terms of symbolic meanings are clearly socially and culturally negotiable, but can they be said to actually exist?
Several pragmatic, epistemological and ontological considerations of this kind are implicitly tied up in some current research which is studying changes in textual and interactional norm systems arising as distributed virtual environments become a new arena for cooperative scientific writing and exchange[10]. Other studies by researchers in the social and behavioural sciences such as sociolinguistics[11], social psychology[12], sociology[13], cultural and gender studies[14], and anthropology[15], suggest that the study of emergent cultural and social norm systems in distributed virtual environments represents a unique source of comparative knowledge regarding the constitution of human cultures, social groups, personal and collective identity and gender.
Ethnographies of conversations and other transactional activities associated with the scientific writing process as they are transferred to distributed virtual environments are expected to provide new insights into the relationship between use of new communication technologies in science and changes in textual and interactional norm systems among scientific authors who use these technologies, as well as into the constitution of distributed scientific communities that are developing based on these changing norm systems. Surprisingly enough, however, little or no work has been done to date in this area by people in the more traditional communication sciences such as semiotics, linguistics, semantics, text/ discourse studies and conversational analysis (one exception to this being a recent collection of exploratory work edited by Susan Herring[16]), and it would seem that there is still considerable room for a lot of innovative and interesting research to be done in this particular field in the future.
We have previously mentioned the `hyper-interactional' aspect of the context of situation for conversations in object-oriented environments which sets these apart from conversations carried on in other types of contemporaneous distributed virtual environments. In the following we shall focus mainly on the conversations themselves and how various technological or mediational constraints contribute to qualitative differences between these and other conversation types.
It seems possible to isolate at least two main areas where mediational constraints on conversations in distributed virtual environments are most marked in terms of their effects on participant behaviours and self-perceptions. Each of these two main areas can further be seen as incorporating a further two nested subsets of issues. We shall briefly examine these four subsets of issues in some more detail below. For the moment, then, we shall select a limited number of issues related to the following:
a) contextual and language modality
i) textuality of the context of situation
ii) language modality and codeb) participant embodiment
i) lack of physical co-presence
ii) lack of mutual eye contact and gaze
i) The textuality of the context of situation
The `material' aspects of the context of situation for conversations in distributed object-oriented environments is, as we have discussed in some detail above, limited to a short textual description of some kind of conceptualizable space, together with any other virtual objects which may have been created and located within, or relative to, this conceptual space. In order for participants to construe this space as a context for communication, it must be possible for them to model a mental represention of the virtual space they are `in' on the basis of the textual description[17] which has been allocated to it by its creator(s). This means that the language used to name and describe this and other object types will be a rather important issue. In order to discuss such matters in more detail we will need to move towards the domains of literary science and text analysis, where classificational notions like `genre', `register' and `style' are already well developed. Elsewhere[18] I have tentatively discussed how various genres of object-oriented environments may be described and categorized, and we shall not pursue this issue in any more depth here. But in any case, it is quite obvious that this general area of study is one which so far has been sadly neglected, and one which represents an important area for future research.
Another important and closely related issue in this connection is the choice of basic language system or code used to write spatial descriptions, as well as for participant-to-participant, participant-to-system and system-to-participant communication, within the virtual environment. So far the majority of distributed object-oriented environments have tended to operate primarily in English since they are designed to encourage international communication, and English seems still to have some kind of lingua franca status in virtual environments, as is it does in the non-virtual environment. There are, however, a growing number of virtual environments which are bi- or multilingual, and which offer choices of one or more (mainly European) language to communicate, and where navigation and other commands may also be selected on the basis of language preferences of individual participants[19].
Clearly, this type of development opens up some interesting potentials for using distributed object-oriented environments in research on foreign and second language teaching and acquisition for children and adults. Virtual environments may offer advantages at certain developmental stages in individual language acquisition and learning processes. There is a degree of existential security inherent in the distancing effect of the technology, as well as a possibility for learners and instructors to read through text logs of conversations after the event in order to find out what actually went on, what mistakes were made and how others reacted to them. Investigating the contribution of such factors in more detail may contribute to increase not only learner, but also instructor awareness, and at a more general level, provide insights into all kinds of language acquisition and learning processes.
Since communication is being carried on in what is essentially an oral modality but at the same time through the technology of writing this offers unique opportunities for learners to relate to, and to compare, these two different modalities of language. One language-specific constraint which may be a restraining issue in this connection is that at present contemporaneous communication in distributed virtual environments relies on network protocols which will only support the restricted ascii character set. This means that those who wish to talk in languages like Norwegian or German which have non-standard ascii characters in their alphabets will (unless they are prepared to redesign the communication software to force it to accept non-standard characters) need to modify their writing norms in order to accomodate this particular set of constraints[20].
Our experiences so far[21] have shown that this type of minor technical complication does not unduly hamper conversation and interactions in distributed virtual environments. Those who find meeting other people in this way enjoyable and interesting seem willing to adapt their communication norms to accomodate to these kinds of constraints, and indeed to do so quite quickly, so long as they are able to perceive communicating with one another at a distance in real time in this way as functionally useful. We have occasionally discovered some short-term `overlapping' effects. Some Norwegian students have for instance reported continuing to use the combination forms /aa/, /ae/ and /oe/ instead of the Norwegian characters /æ/, /å/ and /ø/ when subsequently composing texts in word-processors after prolonged periods of conversations in object-oriented environments. Such effects seem however minimal and do not last long. Other effects, related to what was initially characterized above as `participant embodiment status issues', may turn out to be of a more profound character and possibly have lasting effects on participant norm systems and self-perceptions in the longer term. Some such effects will be presented and discussed briefly later in this chapter.
ii) Language modality and code
One of the most blatently obvious differences between conversations in object-oriented distributed virtual environments and face-to-face conversations is that the former are carried on exclusively in writing and the latter in speech. But since they are contemporaneous the language modality of these conversations distributed virtual environments tends towards the predominantly oral variant. Let us look at this general modality issue in some more detail.
Michael Halliday has pointed out[22] that although the grammars of written and spoken language as actualized in text and discourse differ from one another in several important ways, both are as intricate as one another in their own ways. In terms of production norms for actual usage, spoken language and written language tend to lie towards opposite poles of a continuum which ranges from most spontaneous to most self-monitored. Forms of production which we consider most spontaneous are generally spoken, while forms of production which we consider most self-monitored are generally written.
These two basic lingustic modalities which we use in order to represent meaning do not differ in their basic systematicity: both are highly organized, regular and productive of coherent discourse. Considered in terms of lexicogrammatical organization, spoken language tends to favour greater lexicogrammatical intricacy by accomodating fewer lexical items in the clause, and more clauses per syntagm, while written language tends to favour greater lexical density by accomodating more lexical items in the clause and fewer clauses in the syntagm. Considered in terms of the expression of different modes of meaning, spoken language tends to be spun out, flowing, choreographic, oriented towards events like doing, happening, sensing, being and saying. It is processlike, intricate and with serially ordered meanings. Writing on the other hand is dense, structured, crystalline, oriented towards things like entities and objectified (nominalised) processes, productlike, tight and with meanings related to one another as components.
In our conversations in distributed virtual environments we have found there is a tendency towards a blending of these two different ways of meaning in language, and a more general drift from the written towards the oral modality. This kind of drift is best illustrated by an example. The log excerpt below is taken from a university seminar which I held in a distributed virtual environment a couple of years ago together with some of my students who were in our computer writing lab in Norway while I was in Italy. In this particular interaction strip[23], I ask Alison, one of the students, if she would be willing to take responsibility for a short presentation for the group in a session scheduled for the next day, and to lead a subsequent discussion. Sally and Margaret are two of the other seven students in the group[24]. Hopefully this brief excerpt can serve without too much explanation to illustrate this `hybrid' modality of language blending the two ways of meaning referred to above, and fluctuating as it does so, roughly midway between spontaneous and self-monitored forms of production.
Since conversations in distributed virtual environments are carried on exclusively in writing rather than speech, this tends to make these exchanges slower than those in face-to-face conversations. As we know, writing speed for most people generally tends to lag far behind their speaking speed. In distributed virtual environents this does not seem to hamper communication as much as one might expect, tending rather to evoke a number of compensational strategies, involving amongst other things, innovative forms of abbreviation, spontaneous use of acronyms, ellipsis, and emotes, the latter simulating in textual form combinations of verbal and non-verbal signals in the same utterance[25]. All these strategies seek to optimalise communication by capitalizing on the generally highly contextualised nature of the discourse, as well as presuppositional knowledge known to be shared by all or most participants. Note too, that these general kinds of optimalisation strategies will normally be more characteristic for spoken rather than for written forms of language communication.
In passing it should be remarked here that the strip of conversational log above probably reflects the fact that at the time it was recorded, neither I nor my students were particularly adept at this particular form of communication (and indeed, I am still not). Our language use in this particular example may then still be weighted a bit more towards the written, rather than the spoken modality of language. As time went on, we found as mentioned earlier, that there was a distinct tendency for the oral modality to more and more permeate our discourse, although it is of course inherent in the actual technology of communication that this can never become the only modality in use - we will always have to write to communicate in such an environment. In order to balance any erroneous impressions generated by the log example above, we shall therefore end this section with yest another short strip[26], this time taken from a conversation between group of more experienced interlocutors, and which displays even more of the communicative ergonomy of the oral modality referred to above. Again, we shall leave it up to the reader to examine the log in detail and to make their own hypotheses as to which kinds of strategies are actually being actualized here. One point which may lead to some confusion initially is probably worth mentioning in advance of a first reading: it is clear that the pace of the conversation here is quite high, since CT_PeteH sees fit to divide his question to Donald ("donald do you know if there is a mail list for participants?") into two parts in order to avoid falling too far behind in the general flow of things.
i) lack of physical co-presence
Generally speaking, when we now go over to discussing what have been referred to here as `participant embodiment status issues', and their effects on conversations and conversational norm systems, we are considering written interactions which are being mediated by means of artificial (non-`intelligent') agent objects (also sometimes referred to as `proxies'). Proxies like the Joe, CT_PeteH, JayC, Gustavo, sle, and Donald characters that figure in the conversational transcript above, are themselves virtual objects: small pieces of computer code which provide their owners with a virtual `identity' that they can manipulate and interact with other objects through in the virtual environment.
Real world participants are then in one sense `disembodied' while interacting through a proxy in this way, but they are of course also at the same time `embodied', since they are physiologically, cognitively and emotionally immersed in the phenomenologies and animate forms of their own bodies, wherever they may be in the real world.The key point here is however that they are not, and cannot be, physically co-present for one another in the virtual environment. Their embodied physicality is mutually inaccessible to each another - seeing[27], touching, feeling the warmth or odours of the bodies of other participants is not possible. The only `physical' information interlocutors have access to about each other is a text: the way a person has chosen to present him- or herself through the description of the agent object they `own'.
Participants `log on into', `jack into, `enter', or `put on' the virtual identity engendered by their agent proxy each time they connect to the virtual environment. And indeed, many frequent visitors to object-oriented virtual environments of this kind do tend to refer to the phenomenal experience of doing so in precisely these kinds of terms. The proxy character, and words and actions `spoken' or `emoted' by this character, are all that other participants have to relate to while they are `together'. This particular type of disembodiment situation opens up for various kinds of role-playing and identity swapping activities, and there is quite a number of mainly anecdotal reports of comic or otherwise effects of such role-playing and identity swapping actitvities i distributed virtual environments of this type, with gender-change experimentation being one of the most common[28].
Experiences using distributed virtual environments for distance education experiments at university level in Norway in the course of the last four years[29] have shown that the disembodiment experience inherent in the use of this medium of communication can have some rather unexpected effects on student - teacher and individual - group role perceptions and relations that appear to actually carry over into real-world situations afterwards. These effects can be both positive and negative, and some of the more positive effects which also persist in real world situations over time have been observed after even quite short periods of group interactions in virtual environments. These effects seem especially marked for quieter students who for various reasons are shy or anxious about communicating in larger groups, or with authority figures in face-to-face situations. Individual students have reported experiencing a kind of `liberating' effect when they no longer feel bound to relating to others from `within' a physical body which can be seen or heard to blush or stammer by everyone present. Some similar anxiety-reducing effects in communicating with supervisors who are generally considered to be authority figures have also been reported by university teachers in Norway who have used electronic mail to carry out supervision, counselling and other follow-up activities with trainee teachers who are placed in external schools during teaching practice periods.
At this point it is important to point out that more or less all research done so far in this particular area has been largely based on qualitatitive methodologies, and to some extent anecdotal materials, and is thus, at least as far as we are aware, not only unreplicated, but also ureplicable. There remain therefore a large number of important questions which are still unanswered, and which will require much more systematic and detailed research in order to establish the validity and reliability of these kind of more generalized impressions which have been gathered so far.
ii) lack of mutual eye contact and gaze
Basically, what is goeing on between participants during conversations in object-oriented distributed environments is the same as what is goeing on during telephone, television or even video-telephone conversations where two or more participants are involved, namely talk. One of the main differences is or course that in distributed object-oriented virtual environments interlocutors are relating to and reacting to what is being `said' in writing by other particpiants' proxy characters, rather than to their voices in an earpiece, or their video images on a screen.
Note, however, that in none of these conversational settings mentioned above will it be possible for participants to directly engage the gaze of another participant in the same way as they can in face-to-face conversations. This fact affects not only the general process of conversation but also some of the more specific forms that these interactions take. Here it will not be possible to go into any great detail about every single type of effect which this lack of mutual gaze may have on conversational norms, so we shall restrict ourselves here to making some rather general observations, especially regarding situations where two or more interlocutors actively take part in conversations, and where co-regulation of multi-party talk is thus an important issue.
Since participants cannot engage in mutual eye contact or follow the gaze of their interlocutors in an object-oriented environment, conversational regulation by mean of establishing or breaking eye contact with interlocutors, shift of gaze direction, changes in facial expression or any other small head and eye movements will not be feasible. Information about the movements of other interlocutors within the virtual space will generally be restricted to larger comings and goings, except when participants actively choose to make other kinds of `actions' explicit by `emoting' them into the conversational space (see below for more discussion of this and an example).
We have found that initially at least, this lack of mutual eye contact tends to make management of group conversations problematic, especially for novices[30]. This can amongat other things lead to difficulties in `giving each another the floor' since this kind of move must be explicitly signalled by means of a `spoken' text-utterance directed to one or more of the participants present, as in the following example:
PatrickC [to John] "your turn John"
or by means of an explicitly emoted[31] `non-verbal' cue such as in
PatrickC looks expectantly at John and waits for him to speak...
This of course makes turn-taking in general a different kind of process than in face-to-face conversations. Generally speaking, then, all kinds of regulation of turn-taking in multi-party conversation in distributed virtual environments requires som form of alternative strategies which must be made contingent on the technological constraints of the medium of communication.
Indeed, the specialized kinds of communication functions offered by distributed virtual environments are precisely designed to facilitate the use of such alternative strategies. The use of the `to' command[32] in the first example above is one system-specific way of directing an utterance to one, and only one other participant, rather than to the whole group, and it thus simulates the same kind of communicative function in the virtual environment as shifting one's gaze in order to directly address someone else in the group would perform in a face-to-face conversation.
All non-verbal cues like the one shown in the second example above, whether they are phatic, `backchannelling' or otherwise, and which are performed in the distributed environment by means of the emote function, must necessarily first be keyed in as a piece of text and then `entered' into the virtual environment by interlocutors in order to be actualised as part of the global communication situation. This means that `non-verbal communication' in a distributed virtual environment in each and every case constitutes an intentional act, at the very least involving some kind of motor activity at a keyboard on the part of the sender. This is of course not normally the case for similar non-verbal cues in face-to-face conversation, although some of these cues may of course be socially coded and intentionally produced - Charles Darwin once noted for instance that only adults sneer, while small children do not[33]. In everyday life however, it will often be those non-verbal cues which are not immediately interpretable as intentional which will attract most attention and really cause interlocutors to sit up and pay close attention to what is being said.
In the kinds of distributed virtual environments we are concerned with here the intentionality or not of such non-verbal cues is not really an issue, whereas their spontaneity, authenticity and actual meanings most definitely is. Let us for instance consider the following example:
PatrickC looks completely horrified and turns as pale as a sheet
An `utterance' of this kind will almost automatically come to acquire some kind of metacommunicational and even ironic set of meanings and connotations when it is presented and received in this way as a piece of written text. If however, a similar physio-emotional reaction on my part was to be observed as it was actually occurring by my interlocutor in the course of a face-to-face conversation, the situation would be very different indeed.
As it stands above as a piece of `disembodied' written text it may first of all conceivably be a lie, since my interlocutor is not able to actually cast a glance at at me in order to check whether the information I have chosen to provide about my assumed physiological, emotional and mental state at the time of writing is correct.
Secondly, in face-to-face conversation a physiological reaction of this kind which normally will signal strong emotional affect is almost impossible to simulate and to hide from others who are present.
Finally, the fact that it must be written down in the first place in order to be `uttered' and `observed' by my interlocutors means that it takes a certain amount of time to appear for them, and it can therefore not possibly be occuring spontaneously for the first time as it is being read by them.
What is being communicated then, is a kind of report or simulation of some presumed, already past, set of physiological states and events, which in any case is not shareable at its actual moment of origin by all those present. There is no doubt that this makes `non-verbal' cues `uttered' in this way in conversations in distributed virtual environments qualitatively and phenomenally quite different from similar types of cues which may be mutually experienced in face-to-face conversations.
What possible long-term effects this kind of potential for producing delayed selective displays of simulated emotional responses may have on the development of interpersonal relationships initiated in distributed virtual environments are by no means clear at the present time. There is however a certain amount of anecdotal evidence [34]which implies that deep interpersonal relationships which are initiated in this type of environment require quite a long period of mutual adjustment if they are actually transferred into real-world, face-to-face situations, and this too, would appear to be another important area for further research.
Let us now consider an analytical model which will allow us to make ethnographies of conversations in distributed virtual enviroments. There are obviously many differences between these kinds of conversations and face-to-face conversations which will have to be taken account of in such a model. At the same time, there are also a considerable number of similarities between these two types of conversational situation, so what we need is to develop some kind of functional, and preferably reasonably concise, taxonomy which will allow us to characterize and describe any kind of conversation in a general way but which also will allow us to highlight some of the most relevant differences and similarities between virtual and face-to-face conversations. To this end we might consider adopting Sachs and Schegloff's simple systemics of adjacency pairs as the basic structure underlying all and any form of conversational exchange. For our present purposes, however, I believe that the following observation by Erving Goffman from his introduction to his book Forms of Talk seems even more useful. Here he writes[35]:
"Everyone knows that when individuals in the presence of others respond to events, their glances, looks and postural shifts carry all kinds of implication and meaning. When in these settings words are spoken, the tone of voice, manner of uptake, restarts and the variously positioned pauses similarly qualify. As does manner of listening. Every adult is wonderfully accomplished in producing all of these effects, and wonderfully perceptive in catching their significance when performed by accessible others. Everywhere and constantly this gestural resource is employed, yet rarely itself is systematically examined."
My main reason for seizing on this particular citation here is that Goffman does not, as many linguists and discourse analysts before and after him have unfortunately tended to, overstate the importance of spoken language in human communication. Rather, he takes care to position the vital semiotic resource of language within a much broader framework which incorporates other semiotic resources like facial expression, gaze, intonation, prosody, pauses, verbal style, body language and other forms of non-verbal behaviour, all of which he refers to collectively as `this gestural resource'. These resources are all seen to work in tandem with the central language resource, but it is quite clear that this whole constellation of quite different semiotic resources so much functions together as a whole as to make each individual component inseparable from the rest of this greater whole in order for communication to function optimally.
A second reason why Goffman's approach represents a useful point of departure is that he explicitly concerns himself with the interplay between three fundamental aspects of the context for communication, which he terms `ritualization', `participation framework' and `embedding', and which correspond roughly speaking (in this same order) to codification and understandings of:
a) the generalized `gestural resouce' mentioned above
b) the participant status and conduct of participants
c) the relationship between what is being said, and who or what is actually being represented by what is being said.
These three aspects of communication interact intimately with one another to create what Goffman elsewhere goes on to refer to as `the fundamental requirements of theatricality' which is inherent in all forms of social action, and more specifically, talk.
As he puts it:
"I make no literary claim that social life is but a stage, only a small technical one: that deeply incorporated into the nature of talk are the fundamental requirements of theatricality"[36].
Incorporating an understanding of this basic `theatricality' of social life and communication seems to me vital in developing any form of ethnographic description, and especially of those sociocultural situations where conversation or talk represents just one specific part of an on-going flow of socially defined and other interrelated events. In fact, when we come to think about it, it is quite obvious that all conversational situations are actually like this.
People do not really just talk, they are constantly doing other things while they are talking, either on their own account, such as scratching their noses or preening their hair or clothes, or in more directed ways like using body movement and other non-verbal cues to manage socially acceptable levels of physical proximity, or looking at and handling some object together while at the same time talking about it. People come and go, they move around the situation they are in while they are conversing, they adopt and continually negotiate various social and interpersonal roles relative to one another at different times in the course of the conversation, and relative to any other on-going activities in the situation they are taking part in. Sometimes they are just watching, at other times they are both watching and commenting on what others are doing, at other times again they are initiating activities on their own or recruiting other participants in order to do things together with them.
Goffman was one of the first to address this basic issue of the complexity and interrelatedness of the communication events comprising everyday interactions systematically in more general terms. First sketched out in Frame Analysis[37] his basic ideas are developed in more detail in Forms of Talk[38]. More recently people in anthropological linguists[39] have been developing specialised fieldwork tools and methodologies for conversational analysis which attempt to include various features of doings in analytical descriptions of situated talk.
If we return to our considerations of conversations and interactions in object-oriented distributed virtual environments we can see that many of the same kinds of comings and goings, movings and doings of other kinds as mentioned above for face-to-face situations will also be possible for participants to simulate, and in fact most of these moves and happenings will be noticeable for others, albeit in restricted and disembodied ways due to the technologically mediated nature of the virtual environments in which they occur.
Ethnographies of conversations, then, whether they are being made in real or virtual environments, will in any case need to take account of a continuous flow of complimentarity phenomena[40] reflecting changing social and interpersonal role relationships negotiated in conversation, as well as the various kinds of doings and talkings related to the management of participants' perceived consequences of these changing role relationships, and which at the same time are contributing to their changing.
Below I attempt to sketch out a simple taxonomy that may be useful for making ethnographies of conversations in general and which at the same time can take account of some of the differences and simliarities between conversations face-to-face and conversations in object-oriented virtual environments. For the time being we shall leave it up to readers to examine and evaluate the model presented below, and to consider in which specific ways it might actually be applied in researching conversations ethnographically in various kinds of situations. I also suggest a few places where this basic taxonomy might be extended in order to include forms of qualifying behaviour which are normally found whenever processes of socialization, normative change and norm constitution are going on, which probably can be said to be the case at some level or other for almost every form of human interaction.
Fig. 1 Modified schematic summary of Goffman's (1981) model of footing in conversation and talk
Figure 1 is a schematic representation of alignments of speakers to hearers in talk as described in detail elsewhere by Goffmann[41]. The schema has however been extended in order to include a few sub-categories of what I have chosen to refer to as `doings', suggesting a rather coarse-grained, tripartite categorisation of possible ratified participant behaviours into `movement', `observation' and `performance'. Other categories and subdivisions than these three basic ones are of course thinkable, and it will probably be necessary to consider the introduction of other subcategories into this basic schema at some later stage. However, for the sake of simplicity we shall content ourselves for the time being with leaving the number at three. Associated participant roles for these three types of doings for ratified participants are those of `mover', `observer' and `performer'.
Into the general sub-category of `movement' will fall participant doings such as entering and leaving a situation through an entrance or exit, any other kinds of comings and goings, movement around the room, movement between various groups of participants, from point to point in the room, moving on the spot etc..
Into the category of observation will fall those forms of participation doings whereby attention is openly (or otherwise) being directed to some activity or activities performed by other ratified or non-ratified participants co-present in the situation, or to objects present in the environment.
Into the category of performance will fall such forms of participation as taking and holding the floor in order to say something, making a prepared speech, reading aloud from a piece of paper, demonstrating some particular motor skill or task, writing on a blackboard or a piece of paper, drawing a diagram etc..
At this point it is important to point out that the two other main participant roles in the taxonomy, i.e. those of `speaker' and `hearer', and indeed any subcategorizations of these categories which have been included in the matrix above, will, and indeed cannot in any way be considered mutually exclusive either with one another, or with the basic role of `doer', nor with any of the three more delicate subcategorizations of the main doer category shown in the diagram. One may for instance quite easily be speaking at the same time as one is writing or drawing a diagram on a blackboard, and someone who is speaking can quite easily be listening to general signals of audience response like coughs, shufflings and other noises in the environment while at the same time watching the body language of some key member of the audience, as well as glancing from time to time at what is happening outside the window. Obviously, someone who is actively engaged in forms of behaviour which are related to all three key participant roles at one and the same time will be forced to divide his or her attention between the execution of several quite different forms of physical, mental and emotional activity, and this will tend to lessen the amount of attention and energy that he or she is able to devote to the maintenance of behaviours specific to each of the roles in question. The general effect of this will be to slow down or otherwise interfere with carrying out of motor and attentional requirements of the various role clusters[42].
If we move on fopr a moment to look at the case of `doings' associated with non-ratified participants, so far only one kind of very general potential `doer' role has been identified within the general `doing' sub-category, namely that of `intruder', but again, a number of other more fine-grained subcategorisations of non-ratified participant doings and associated participant roles on the basis of similar `movement', `observation' and `performance' sub-categories are not only thinkable, but also very probably necessary. Especially relevant in this connection will be categorizations for out-of-frame doings on the part of those non-ratified participants who actually might desire to become ratified participants but who are still in the process of finding out how to go about doing so. It may also apply to participants who for some reason or other have lost, or been deprived of a ratified position in the social field which they have held previously, and who now wish to negotiate a new position for themselves. This general behavioural category might be referred to as `qualifying', with associated participant roles such as `novice' or `wannabe'[43]
Three main types of logging devices available for conversational and other kinds of research purposes in distributed virtual environments of the MOO variety. These devices, which in themselves are specially designed objects, allow for the on-going registration of conversations and interactions between participants at various levels of detail, and from different points of view. These three main types of device allow variously logging of:
1. conversations and public actions only
2. everything done by one or more participants
3. everything done or percieved by one participant
Let us look more in detail at these three types of logs in turn, together with some examples of how they may turn out to be in practice.
The simplest, and probably most common type of logging device used for making records of conversations in distributed virtual anvironments is a recorder object which is programmed to be `visible' in any room in which it is placed, and which visibly informs participants entering the room that conversations there are being recorded when it is turned on. Some devices also emit regular messages while they are recording, informing everyone in that particular room that interactions are being recorded. Recorders of this type may generally be operated, ie. turned on and off, by more than one person, but they may also in certain cases be restricted to use by one person only. They will record to a separate $note object which is generally accessible for reading by for anyone entering the room where it is located, as in the example below:
In this case, the `Sony' object is the recording device, and `Guestbook' is the $note object to which interactions in the room are at present being logged. Issuing the command:
read Guestbook
while in the room above will display the contents of the current log, which for a two-participant conversation may look something like this[44]:
Conversations with more than two participants become, as one would expect, more complicated, often with several parallell threads of conversation developing, and where one or more participants may be active in several such threads all at the same time, as can be seen in the example below[46]:
A second type of recording device will register both conversations and certain kinds of normally `private' actions taking place in a room. These private actions are mainly paging and whispering[47], manipulation of objects that do not publicly announce that they are being handled, access to MOO-mail folders and messages, and other system commands that do not generally show up for other participants. The example below is a log of a paged conversation made a few years ago between myself in Conference Room (North) at Diversity University MOO and Ulf who is in his office in another zone. The log is seen entirely from my own point of view, so the description of room which comes up on the screen as I enter it at the beginning of the strip is visible. So too, is my `action' of closing the door into the Conference Room, an action which more generally speaking is designed to restrict entry into some particular room to ratified persons who are invited to enter by another ratified participant who is already in there. The philosophy_guest objects visible at the foot of the room description are characters belonging to participants in a philosophy seminar which during that particular periodwas being held in the room in question. The seminar was not however actually in session as I entered the room. If it had been, and the door had been closed then it would have been necessary for me to request an invitation to join the session from its moderator in order to enter . Without this any attempt on my part to `intrude' would automatically result in a rejection message from the system.
A third group of devices records everything that is said, done and perceived by one participant. In effect they give a log which is a record of all and any kind of activities in the virtual environment carried out by that person and seen only from that one particular point of view only.
The example below is taken from a log of this kind. This particular section of log contains not only my conversation with bjornb[48], but also my own commands as they were directed to the MOO environment as well as the results of these commands as they were subsequently generated by the MOO.
Advanced MOO systems allow for such highly detailed logs from one particular participant's point of view to be made (after necessary permissions have been obtained) also by other participants than the one being logged. Candidates for making such logs could for example be ethnographers, teachers and educational diagnosticians for respectively research, pedagogical and/ or therapeutic purposes. Obviously the possibility of maintaining such a high degree of surveillance of individual participant behaviour, even though this is only concerns behaviour in a virtual environment where real-world consequences of actions are presumably, but not necessarily, less than otherwise, raises many important issues related to privacy and research ethics in general. Indeed, a so-called Privacy Policy and Guide to Human Subject Research has already been developed and implemented (during the Fall of 1997) by the Review Board for Human Subject Research at Diversity University MOO[49], which is one of the largest and longest-running distributed virtual environments developed specifically for distance educational research purposes.
Agar, Michael H. & Hobbs, Jerry R. 1982. `Interpreting Discourse: Coherence and the Analysis of Ethnographic Interviews'. Discourse Processes, Volume 5, 1982, pp. 1-32.
Bolter, Jay. D. 1991. Writing Space: The Computer, Hypertext, and the History of Writing. Hillsdale: Lawrence Erlbaum and Associates.
Bruckman, A. 1992. Identity Workshop: Emergent Social and Psychological Phenomena in Text-Based Virtual Reality. Electronic manuscript of an unpublished paper written in partial completion of a doctoral degree at the Massachusetts Institute of Technology Media Lab. http://lucien.berkeley.edu/MOO/identity-workshop.ps
Bruckman, A. & Resnick, M. 1993. Virtual Professional Community: Results from the Media MOO Project. Paper presented at the Third International Conference on Cyberspace in Austin, Texas, May 15th, 1993.
Campione, Mary & Walrath, Kathy 1996. The Java Tutorial. Addison Wesley.
Cherny, Lynn 1995. Gender Differences in Text-Based Virtual Reality. In: Proceedings of the Berkeley Conference on Women and Language, April 94
Clodius, Jen 1997. Creating a Community of Interest "Self" and "Other" on DragonMud. Paper presented at the Combined Conference on MUDs in Jackson Hole, Wyoming, January 15, 1997.
Clodius, Jen 1995. Computer-Mediated Interactions: Human Factors. Keynote speech at MUDshop II: Learning Spaces, Sept. 6, 1995, San Diego CA.
Coppock, P.J. 1995(a). "Ascribing continuity to the diachronicity of textual norms in virtual environments." Paper presented at 4th Nordic Association of Semiotic Studies Research Symposium, Trondheim Norway, October 16th-19th 1994. To appear in special number ofSemio-Nordica. (Ed. G. Sonesson).
Coppock, P.J. 1995(b). "The semiotics of a phenomenological research paradigm for investigating the evolution and ontogenesis of cultural norm-systems in distributed virtual environments". Paper presented at the 4th International Association of Semiotic Studies conference: The Semiotics of the Media, Kassel, Germany, 20-23 March 1995. Semiotica 115-34 (1997), 235-262
Coppock, P.J. 1995(c). "A semiotic perspective on the development of (artificial) consciousness". In: A. Aamodt & J. Komorowski (eds.) SCAI '95, Proceedings of The Fifth Nordic Conference in Artificial Intelligence, Trondheim, Norway, May 29-31, 1995, pp.378-386.
Coppock, Patrick J. 1996. Changing textual norms through cooperation in distributed virtual environments. A field study of qualifying text and communication norm development in network-based multiuser text worlds. Field Report 1 for the period Fall 1994-Spring 1996. Unpublished Part I of doctoral dissertation. Trondheim: The Norwegian University of Science and Technology, Faculty of Arts, Department of Applied Linguistics.
Coppock, Patrick J. 1997. Changing textual norms through cooperation in distributed virtual environments. A field study of qualifying text and communication norm development in network-based multiuser text worlds. Field Report 2 for the period Fall 1996-Spring 1997. Unpublished Part II of doctoral dissertation. Trondheim: The Norwegian University of Science and Technology, Faculty of Arts, Department of Applied Linguistics.
Coppock, P.J. in press (a) "A sociosemiotic approach to processes of textual and interaction norm change in distributed virtual environments." To appear Spring 1998 in S. Cmejrková, J. Hoffmannová , O Müllerová & J. Svetlá (eds.): "Dialogue in the Heart of Europe": Proceedings of The 6th International Congress of the International Association for Dialogue Analysis, Prague, 17.-20.04.96. Tübingen: Niemeyer.
Coppock, P.J. in press(b) "Evolving dialogical norm-systems for scientific writing in distributed virtual environments." To appear in G. Sonesson (ed.)"Signs as Communication and Dialogue": Proceedings of the 4th Biannual Congress of the Nordic Association of Semiotic Studies, Imatra, Finland, 09. -14.06.96
Coppock, Patrick John in press (c). "Negotiation of public identity in the courtroom: Interrogation of Antonio di Pietro, Tribunale di Brescia, December 16th 1996". To appear in the course of 1998 in a volume of essays by several authors on the same theme edited by R. Galatolo & G. Palotti.
Cross, Janet . Don't be cowed by the MOO
Curtis, Pavel, 1992. `Mudding: Social Phenomena in Text-Based Virtual Realities', paper presented at the conference `Directions and Implications of Advanced Computing'. Intertek, Vol. 3.3 (Winter 1992): 26-34.
Curtis, Pavel and Nichols, Dave 1993. MUDs grow up: social virtual reality in the real world. Unpublished report.
Delany, Paul & Landow, George D. (eds.) 1994. Hypermedia and Literary Studies. Cambridge Mass.: MIT Press.
Ducrot, Osvald & Todorov, Tzvetan 1979 (1972). Encyclopedic Dictionary of the Sciences of Language. Translated by Catherine Porter. Baltimore and London: Johns Hopkins University Press.
Duranti, Alessandro & Goodwin, Charles (eds.) 1992. Rethinking Context: language as an interactive phenomenon. Studies in the Social and Cultural Foundations of Language 11. Cambridge Mass.: Cambridge University Press.
Duranti, Alessandro 1997. Linguistic Anthropology. Cambridge, New York, Melbourne: Cambridge University Press.
Firth, John R. 1957. Papers in Linguistics, 1934-51. London: Oxford University Press.
Firth, John R. 1966 (1930). The Tongues of Men and Speech. London: Oxford University Press.
Goffman, Erving 1974. Frame Analysis. An Essay on the Organization of Experience. New York, Hagerstown, San Francisco, London: Harper & Row.
Goffman 1981. Forms of Talk. Philadelphia: University of Pennsylvania Press.
Goodwin, Marjorie Harness 1996. He-Said-She-Said: Talk as Social Interaction Among Black Children. Bloomington & Indianapolis: Indiana University Press.
Halliday, M.A.K. 1978, Language as a Social Semiotic: The Social Interpretation of Language and Meaning, London: Edward Arnold (also Baltimore, Md.: University Park Press, 1979).
Halliday, Michael A. K. 1979. Modes of meaning and modes of expression: Types of grammatical structure, and their determination by different semantic functions. In D.J. Allerton, E. Carney, & D. Holdcroft (eds.), Function and context in linguistic analysis: A Festschrift for William Haas (pp. 57-79). Cambridge: Cambridge University Press.
Halliday, M.A.K. 1984. "Language as code and language as behaviour: a systemic-functional interpretation of the nature and ontogenesis of dialogue". In: R.P. Fawcett, M.A.K. Halliday, S.M. Lamb & A. Makkai (eds.), The Semiotics of Culture andLanguage, Volume 1, Language as a Social Semiotic, London and Wolfboro N.H.
Halliday, Michael A.K.1985. Spoken and Written Language. Deakin: Deakin University Press.
Halliday, Michael A.K. 1987(a). Spoken and Written Modes of Meaning', in: Rosalind Horowitz & S. Jay Samuels (eds.) Comprhending Oral and Written Language. San Diego, New York, London etc.: Academic Press.
Halliday, M.A.K. 1987(b). "Language and the order of nature". Chapter. 9 in: Nigel Fabb et al. (red.): The Linguistics of writing. Manchester: Manchester University Press.
Halliday, Michael A.K. 1994. An Introduction to Functional Grammar (2nd Edition). London: Edward Arnold.
Halliday, Michael A.K. 1995 The gramatical construction of scientific knowledge: the framing of the English clause. Paper presented at the international research conference "The Languages of Science", Bologna, Italy, Spring 1995.
Haraway, Donna J.1991. Simians, cyborgs, and women: the reinvention of nature. London: Free Associations Books
Haraway, Donna J. 1997. Modest witness@second millennium: femaleman meets oncomouse: feminism and technoscience, with paintings by Lynn M. Randolph. London: Routledge.
Matthiessen, Christian & Halliday, Michael A.K. (in press). `Systemic Functional Grammar: A First Step into the Theory'. To appear in James Ney (ed.), Current Approaches to Syntax. Tokyo, Tapei & Dallas: International Language Studies.
Johnston, Trevor 1996. `Function and Medium in the Forms of Linguistic Expression Found in a Sign Language.', in: William H. Edmondson & Ronnie B. Wilbur (eds.): International Review of Sign Linguistics, Volume 1. New Jersey: Laurence Erlbaum Associates.
Landow, George 1992. Hypertext: The Convergence of Contemporary Theory and Technology. Baltimore: Johns Hopkins University Press.
Malinowski, B. 1960. A Scientific Theory of Culture and Other Essays. Oxford: Oxford University Press.
Masinter, Larry & Ostrom, Eric 1993. Collaborative information retrieval: Gopher from MOO.
McLaughlin, Jeff . On-Line Teaching and Learning. How To MUD: Virtual University Access Commands
Nielsen, Jakob. Hypertext & Hypermedia. Boston, San Diego, New York, London, Sydney, Tokyo, Toronto: Academic Press International (Harcourt Brace)
Nunberg, Geoffrey (ed.) 1996. The Future of the Book. Berkeley & Los Angeles: University of California Press (also Amsterdam: Brepols)
Nöth, Winfried 1990. Handbook of Semiotics. Bloomington: Indiana University Press.
Reid, Elisabeth M. 1994. Cultural Formations in Text-Based Virtual Realities. M.A. Thesis: Cultural Studies Program, Department of English Studies, University of Melbourne. World Wide Web version at:
Stone, Allucquhre Rosanne 1991. "Will the Real Body Please Stand Up?: Boundary Stories about Virtual Cultures." In: M. Benedikt (ed.): Cyberspace: First Steps. Cambridge, Mass.: MIT Press
Stone, Allucquhre Rosanne 1995. The war of desire and technology at the close of the mechanical age. Cambridge, Mass.: MIT Press
Turkle, Sherry 1984. The second self : computers and the human spirit. New York: Simon and Schuster
Turkle, Sherry 1993. Constructions and reconstructions of the self in virtual reality. Paper presented at the third International Conference on Cyberspace. University of Austin, Texas,.
Turkle, Sherry 1996. Life on the screen: identity in the age of the Internet. London : Weidenfeld & Nicolson.
Violi, Patrizia 1985. `Letters'. In: Teun van Dijk (red.) 1985, Discourse and Literature: New approaches to the analysis of literary genres. Amsterdam: John Benjamin.
Violi, Patrizia 1994. Do good summaries exist? Some comments on defining summary as a textual genre.Unpublished manuscript.
Violi, Patrizia (in press)Electronic dialogue between orality and literacy. A semiotic approach.. To appear Spring 1998 in S. Cmejrková, J. Hoffmannová , O Müllerová & J. Svetlá (eds.):"Dialogue in the Heart of Europe": Proceedings of The 6th International Congress of the International Association for Dialogue Analysis, Prague, 17.-20.04.96. Tübingen: Niemeyer.
White, Michele, 1997. `Cabinet of Curiosity: Finding the Viewer in a Virtual Museum'. Convergence, Autumn 1997, Vol. 3, Number 3, pp. 28-71.