Abstract: Traditional ethnographies have been based on the ideas of locality. But with the rise of globalisation processes this concept has been increasingly questioned on a theoretical level. In the last decade, US-American anthropologists called for multi-sited ethnographies. However, the practical implications for research with such a shift have not been broadly discussed yet. Now, with the internet and different kinds of virtual interaction patterns, ethnographic work faces a new challenge. This paper argues that it is necessary to focus on the implications of fieldwork in virtual settings for ethnographic practice.
According to Clifford GEERTZ (1973) the concept of cultural anthropology in the late 50ies and early 60ies has been very straightforward: "They have a culture out there and your job is to come back and tell us what it is". Today it is widely acknowledged that it has become increasingly difficult to conceptualise anthropological research in these terms. The critique of such an approach has focused on two issues of ethnographic practice, on "ethno" and on "graphic". The latter deals with the relationship between the ethnographer and the people she studies. It questions traditional forms of textualisation and representation. Key terms are othering, authorial control, crisis of objectification, dialogical or polyphonic texts. All this is initiated and documented by the "writing culture"-debate. In the following I will not refer to the critique of textual representation—to the "graphic"—any more. Instead I turn towards the second issue of ethnographic practice which is under fire—to the critique of "ethno". 
The idea of "a culture out there", with the implication of being, firstly, a coherent entity and secondly, unique and different from other cultures becomes increasingly difficult to sustain given the developments and transformations we've been witnessing the last few decades. Societies have modernised and differentiated and so have cultures. The contacts of one society to another one and of one culture to another one have been intensified—through media, telecommunications, economy, through migration, travelling and "professional strangers" (AGAR 1980) like ethnographers. Human location, suggests CLIFFORD (1997), should be constituted by displacement as much as by stasis. The idea that the apparent boundedness of a culture is something constructed rather than found, has become the dominant position within the discipline. And there is a broad agreement that even the tiniest geographical places have cultures rather than one single culture. GUPTA and FERGUSON (1997a, p.2) conclude: "What would once have appeared as a logical impossibility—ethnography without the ethnos—has come to appear, to many, perfectly sensible, even necessary." 
2. Beyond "the Field"
With the pluralisation of cultures the notion of "the field" as a geographically defined research area becomes problematic as well. First reflections on the possible consequences for ethnographic practice started in the mid 80ies. MARCUS and FISCHER (1986) pointed to the fact that anthropological research on local and regional worlds tends to underestimate the transnational political, economic and cultural forces that shape the local contexts. In order to take account of these global forces ethnographies should be conceptualised multi-locally or as multi-sited. Since both people and objects would be likely to become increasingly mobile, then ethnography has to get engaged with these movements. 
Currently, we can see two strategies to modernise ethnography beyond the tradition of fieldwork in a geographically defined locality with clear borders and boundaries. The first strategy suggests a "research self-consciously embedded in a world system", that "moves out from the single sites and local situations ... to examine the circulation of cultural meanings, objects, and identities in diffuse time-space" (MARCUS 1998, p.79). Closely related to this approach are suggestions to redefine the notion of the field. GUPTA and FERGUSON (1997b, p.37), referring to the work of Appadurai, suggest to decentre the notion of the field. Instead of the field being used to connote locality, to "the here" and "the elsewhere", the field should rather be conceptualised as "political location". "We might emerge from such a move with less of a sense of 'the field' (in the 'among the so-and-so' sense) and more of a sense of a mode of study that cares about, and pays attention to, the interlocking of multiple socio-political sites and locations." Such a reconstruction of fieldwork, so the authors, would reposition the role of participant observation: it would continue to be major part of anthropological methodologies, but cease to be fetishised. 
Again closely related are suggestions—drawing on CASTELLS' (1996) "Network Society"—to move away from studying fields towards an ethnography of networks. Networks are still strongly related to geographical space—like field. Unlike field, a network is an open structure, able to expand almost without limits and highly dynamic . And even more important: A network does not merely consist of a set of nodes, but also of a set of connections between the nodes. As such, networks contain as much movement and flow as they contain residence and localities. An ethnography of networks would contain the examination of the nodes of a net and the examination of the connections and flows (money, objects, people, ideas etc.) between these nodes. All these highly theorised approaches (from single sites to multi-sites, from geographical localities to political locations, from field to networks) focus on space, on the question of a spatialisation of social and cultural difference. 
The second strategy to overcome a traditional concept of fieldwork—less theorised, however grounded on a rapidly growing body of research—is a shift from material spaces to so-called cyberspace. The growth of the Internet is one of the greatest cultural phenomena of our time, impacting almost all areas of life. So it not surprising there is a rapidly increasing body of ethnographic work that studies online communication and interaction, but also related spaces that are highly mediated and interactive: Virtual cities, virtual universities, virtual community care, virtual organisations, virtual decision making environments, telemedicine, teleshopping, the virtual marketplace, virtual reality environments, virtual households—this list refers to just some of the project titles of the ongoing ESCR funded "Virtual Society?" programme in the UK (http://www.brunel.ac.uk/research/virtsoc/; Broken link, FQS, August 2005). And the University of Hull recently hosted a conference called "Ethnography and the Internet" (http://www.hull.ac.uk/IEC2000/pages/). 
3. Revealing Complexity
Both strategies to go beyond the concept of field as geographically defined locality suggest that ethnography is on the move. It is a move from the field to the net to the Internet. So far however the practical research implications of all these attempts to modernise ethnography have hardly been discussed at all. In the following parts of this essay I intend to explain why the practical research implications are profound indeed. It is time to draw the attention to the consequences of the shift initiated by the latest generation of ethnographers. I begin with an examination of the shift from field to network. Then I turn to the shift from material spaces to cyberspace. 
Before doing so, however, another problem has to be addressed. If fieldwork loses its relevance as the main feature to describe ethnographic practice, what then is left? How can we conceptualise ethnography beyond the idea of a long-term residence in a remote and unknown locality? One way to reframe ethnography beyond "fieldwork" could be to use its aims and objectives as defining principle. Studying the literature, the objectives of ethnographic practice are multiple, diverse and heterogeneous. At this point, I can only briefly outline my view of the aims of ethnographic practice. I want to make two points: (1) Ethnographic practice is attendance, is a co-presence of ethnographer and the observed social situation. Whether this co-presence requires one single shared space, is a problem worth discussing, particularly in the context of online-ethnographies. I'll get back to this issue. (2) Ethnography is about revealing context and thus complexity. The potential of this method lies not in a reduction of complexity, not in the construction of models, but in what GEERTZ calls "thick description". The current debate within anthropology, so my hypothesis, is a debate about different modes of ethnographic complexity. To explain this, it is useful to go back to the early days of ethnographic practice. A century ago, ethnographers like A.C. HADDON, Franz BOAS and a few years later Bronislaw MALINOWSKI revolutionised anthropology by not merely studying decontextualised objects—this is what the armchair anthropologists did—but rather by studying people in their natural environment. To them the key to gaining an understanding of communities/tribes and their cultures, rituals and patterns of interaction was a long-term immersion in another way of life. This shift from decontextualised objects to the study of people in their natural environment has to be understood as an increase of complexity. What made perfect sense at the turn of the last century, now becomes the centre of debate. Long term participant observation in a locally limited area privileges face-to-face relationships and tends to overlook forms of interaction that are more mediated. It privileges permanent residence and tends to overlook movement. It privileges boundaries and thus difference and tends to overlook connections and connectivity. Whereas a century ago fieldwork in the natural habitat of communities had the immense advantage of integrating context, a dogmatisation of the same practice in contemporary ethnography seems to achieve the opposite. It rather excludes the context of the people under observation. 
4. From the Field to the Net
Ethnographic research in and on a network requires careful consideration about which areas and parts of the network to include, which ones to partially include and which ones to exclude. The necessity of spatially limiting the research area is nothing new. The classic field had to be constructed as well. However the construction of the field was facilitated by the fact that fields seemed to have supposedly pre-constructed borders anyway, geographic, social or cultural borders. Networks in contrast are somehow infinite, they are open structures and highly dynamic. By drawing boundaries, as indicated above, the ethnographer actively and consciously participates in the construction of spaces and in the spatialisation of difference. In this respect the framing of the network for the research not only pre-structures the findings and conclusions of any ethnographic inquiry, the framing also becomes a political practice. 
Even more crucial in conducting an ethnography of networks are the implications of doing fieldwork. (The fact that I use the term fieldwork here only reveals how strong it is. It seems impossible to simply replace it with another one.) If the fieldwork in "the field" is substituted by fieldwork in and on a network, many ethnographic presumptions are called in question. One is the aspect of time. Instead of spending several months or even years within one locality, the ethnographer has to split up her available budget of time. The more multi-sited an ethnography is, the less time is available for the individual nodes or the individual connections between the nodes. Again, this has implications for the ethnographic inquiry. 
GEERTZ' pleading for "thick description" is based on the premise, that a thick description can only be achieved, when the ethnographer can deeply immerse herself in the culture to be described. According to GEERTZ (1973, p.5), the process of understanding happens in the field—rather than later on at the office desk when the data is being analysed. This process of understanding refers to the hidden aspects of the examined culture; he is after "construing social expressions on their surface enigmatical". This approach is about symbols and about culture being a "web of meaning", it is a semiotic approach. It is doubtful whether an ethnography of networks with the temporal constraints for the individual nodes is able to uncover deep and hidden symbolic dimensions as successfully as demonstrated by GEERTZ. However it is not only the temporal limits and the less intensive observation that leads to a thin description and a flat analysis of the individual nodes of a network. They become flat and thin, because it is the network itself that has to be observed, rather than the individual nodes. What is necessary is a thick description of the network, its dynamic and the interplay of relations between people, things, activities and meanings. This kind of ethnography neither searches for deep dimensions within a culture, nor for hidden layers of meaning. Instead culture is created in the area of the "in between", it is a dynamic process, it is about becoming and fading away. 
Another important difference between traditional fieldwork and an ethnography of networks is the issue of access. A famous character constructed in traditional ethnographies is that of the gate keeper. Usually there was always one person who opens the field for the researcher, introduces her to the tribe/community, serves as mediator between tribe/community and ethnographer, and who carefully and step by step makes the researcher familiar with the hitherto strange environment. More often than not, this single gate keeper had additional functions. He was at the same time a key informant for the ethnographer and a formal or informal leader of the tribe/gang/community. Take WHYTE's "Street Corner Society" as an example. One gate keeper, the leader of the Italian gang, gave WHYTE access to the gang and later on even smoothed his access to the whole district. A net ethnography however, could not be conducted with the support of one single gate keeper. A net is not a seamless web, the ties between the individual nodes might be rather weak. The ethnographer cannot expect to get access to all nodes after establishing access within one of them. In the most unfavourable case she needs as many gate keepers as there are relevant nodes. 
How does this affect ethnographic work? Firstly, it may be the case that a lot more time is needed to establish access to multiple sites than to one single site—allowing less time later on for the research. Secondly, networks become more and more networks that produce, mobilise and transport information and knowledge. Compared to this knowledge, ethnographic knowledge is rather weak, it cannot legitimise itself immediately. The ethnographer is likely to find herself in a position as a beggar. Access is usually negotiated within an economic frame, on the basis of exchange, it depends on what the ethnographer has to offer. That is to say, in order to get access, she needs to get familiar with the knowledge that is circulated within the net. Familiar enough to construct a persuasive argument how the observed would finally benefit from the research. Thirdly it affects the relation between the observer and the observed, between home and "the field". The observed are not as much the "others" any more, the net is embedded. Geographically and socially it is less remote than the tribe/community. 
To summarise this: A shift from classical fieldwork to a multi-sited network ethnography will change the relationship between the ethnographer and the observed in such a way that the boundaries between home and the remote "field" become less clear. It will reduce the time that can be spent with one single site, which will negatively affect the search for hidden and deep layers of meanings. 
5. ... to the Internet
In the last part of the presentation I will turn to the shift from material spaces to cyberspace and outline some of the practical research implications. Obviously there are many possible ways of conducting an ethnography on the Internet or in virtual spaces. I do not intend to cover this wide range of possible approaches. On the contrary, in order to illustrate the difference between an ethnography in a real space and one in a virtual space, I will only focus on fieldwork in virtual spaces, that is to say on online research on online activities like chat areas, mailing lists, 3D worlds, on a form of research that does not strive for a face-to-face contact with the "field". A great body of work in the last few years deals with issues of identity, with (new) forms of collectivity and (new) forms of communication and interaction. An ethnography of purely virtual spaces is certainly the most radical attempt to move beyond the traditional "fieldwork" approach. It stretches ethnographic practice into an unknown area. On the other hand however, it moves so far beyond tradition that a virtual ethnography has to deal with a set of serious difficulties. I want to discuss four of these problems. 
The first problem virtual ethnography has to face is the validity of data on the internet users. The accuracy of information about age, gender, nationality etc. can hardly be checked. Instead of relying on hard facts, the ethnographer relies on the user's trustworthiness and on her own judgement. Moreover, this uncertainty is particularly problematic in a space that has become famous for its playful possibilities. To play with one's identity, to change one's real gender for a virtual one and by doing so to becoming someone else, someone whose chosen identity can be as real as the offline identity—all this is supposed to co-constitute the attraction of the internet. 
The second problem refers to the key method of any ethnographic inquiry, to participant observation. Obviously, observation can only take place in a rather reduced and limited mode. One could observe the change of websites over time, the formation of a discussion list, the growth of words in a chat area or the movement of avatars in a 3D environment. However one cannot observe "real people" and this is what participant observation is about. So the question that has to be addressed is about the relation between participant observation and ethnographic practice. As already mentioned, it is not taken for granted any more that participant observation necessarily has to be the main method of ethnographic work. GUPTA and FERGUSON (1997b, p.37) argue for a defetishisation of participant observation: "Talking to and living with the members of a community are increasingly taking their place alongside reading newspapers, analysing government documents, observing the activities of governing elites, and tracking the internal logic of transnational development agencies and corporations." However, they still admit, that "participant observation continues to be a major part of positioned anthropological methodologies." I agree with this position. One does not have to mystify or privilege participant observation, but its value for an understanding of social situations, everyday routines and embodied practices can hardly be underestimated. 
The third problem raises the issue of connections. Talking about an ethnography of networks, I argued that this would bring about a change of ethnographic inquiry, a change from meaning to connections. A thick description of a network has to illustrate and illuminate the nodes, links, and flows, the structuration of the network. If this argument makes sense, what does it mean for virtual spaces? Clearly the Internet is a network. It consists of nodes and links. But what kind of connections do we have to deal with? We deal with hyperlinks, not with real ones. Hyperlinks are purely abstract, they don't inform about the quality of connections, about strong or weak ties, about the amount of flow that moves from one node/site to another one. So we are faced with a dilemma. A thick description of networks requires a thick description of connections and connectivity. The connectivity in virtual spaces is represented by hyperlinks. Hyperlinks are an impoverished and one-dimensional way to represent and express social ties. 
The fourth problem of fieldwork in purely virtual spaces refers to the notion of context. The displacement between ethnographer and her field results in a lack of a common and mutual perception of the physical context. It does not provide any information of the physical and aesthetic (dress codes) characteristics of the users. The analysis of interaction patterns are affected as well. Whereas a face-to-face dialog not only includes the spoken words, but also gestures, mimics, the sound of a voice and the smell of someone, an online communication is limited to the written word and a handful of so called emoticons. Fieldwork in virtual spaces cannot rely on external forms of structuration. 
To summarise this, I go back to what I identified as one of the core objectives of ethnographic work, which is to reveal complexity. Traditional forms of fieldwork in local communities are under critique because they don't integrate context (e.g. global influences, the connections one field established with other fields etc.) in an appropriate way. In a way, fieldwork in virtual spaces encounters a similar problem. Here the connections between virtual and real spaces are underestimated. The lack of context, which is in this case the context of the offline or physical environment, does not reveal complexity, but rather creates a reduction of complexity. 
To avoid misunderstandings: I do not wish to argue against research in purely virtual spaces. Certainly this research is necessary and useful. However I would hesitate to ascribe to it the label "ethnography". Research in virtual spaces can only then become virtual fieldwork if the research is multi-sited, multi-sited in a very physical sense. In schools, internet cafes, work places and in private living spaces. If the research conducted is single-sited, that is to say from the researchers office computer, it might be more appropriate to dispense with the term ethnography and talk about conversation analysis, text analysis or discourse analysis. 
My concerns about fieldwork in virtual spaces are not only methodological. They are also informed by and a reaction to a currently very popular distinction between "the real" and the so called "virtual". This distinction, constructed by sci-fi literature, media and social science simultaneously, draws a sharp line between an online or virtual world and an offline, real world. The debate on cyber related issues seems to suggest a doubling of reality. Virtual organisations emerge next to traditional organisations, virtual universities offer new forms of education and training; the counterpart of society is virtual society, and the one of culture is virtual culture. Real communities seem to disappear in the information age, and—maybe as a reaction to this loss—academics and new media practitioners are busy constructing virtual communities. VIRILIO (1995) e.g. suggests that we are facing a "fundamental loss of orientation ... A duplication of sensible reality, into reality and virtuality, is in the making." In contrast to this view I argue for a perspective, that does not separate the virtual or online world from the real or offline world. On a theoretical level such a perspective is problematic, because it suggests the existence of a real reality, a reality that is not mediated. After all, the introduction of the term virtual did not contribute to a better understanding of current transformations of and within society. And empirical research persuasively shows that emailing, online chatting, web surfing and other interactive practices are very real experiences for the people performing them. It has to be acknowledged that the use of interactive media for communication can be as real as a talk on the phone or a face-to-face dialog. Rather than emphasising the differences between material and digital spaces, we should introduce a more relational perspective and concentrate on the similarities, connections and overlappings. No method would be more appropriate to achieve this objective than a modernised version of fieldwork. 
Like the objects of ethnographic inquiry—people—ethnography itself is on the move. It is moving away from "fields" as spatially defined localities towards socio-political locations, networks, and multi-sited approaches. And it is moving from physical spaces to digital spaces. This transformation seems to be necessary. However, these shifts bring about severe practical implications for research. I argued that ethnographies on and in networks not only change the nature of ethnographic inquiry, they might also be difficult to realise in a methodological sense (time for research, access). Even more problematic are ethnographies mainly focusing on cyberspace. The exclusion of the material worlds, I argued, is unlikely to reveal context and complexity. So ethnography faces a dilemma. On the one hand the attempts to overcome the notion of "field" as geographically limited area are overdue. On the other hand these attempts clearly show that is hard to escape from the notions of "field" and "fieldwork". Could it be that ethnography is becoming a romantic project? 
Agar, Michael (1980). The professional Stranger. San Diego: Academic Press.
Appadurai, Arjun (1991). Global Ethnoscapes: Notes and Queries for a Transnational Anthropology. In Richard Fox (Ed.), Recapturing Anthropology. Santa Fe: School of American Research Press.
Castells, Manuel (1996). The Information Age: Economy, Society and Culture. Vol. I: The Rise of the Network Society. Malden (MA): Blackwell Publishers.
Clifford, James (1997). Routes: Travel and Translation in the Late Twentieth Century. Harvard University Press.
Clifford, James & Marcus, George (Eds.) (1986). Writing Culture: The Poetics and Politics of Ethnography. Berkeley: University of California Press.
Geertz, Clifford (1973): The Interpretation of Cultures: Selected Essays. New York.
Gupta, Akhil & Ferguson, James (Eds.) (1997a). Culture, Power, Place: Explorations in Critical Anthropology. Durham: Duke University Press.
Gupta, Akhil & Ferguson, James (Eds.) (1997b). Anthropological Locations: Boundaries and Grounds of a Field Science. Berkeley: University of California Press.
Marcus, George (1998). Ethnography through Thick and Thin. Princeton: Princeton University Press.
Marcus, George & Fischer, Michael (1986). Anthropology as Cultural Critique: An Experimental Moment in the Human Sciences. Chicago: University of Chicago Press.
Virilio, Paul (1995). Speed and Information: Cyberspace Alarm. In http://www.ctheory.com/a30-cyberspace_alarm.html (broken link, FQS, September 2003)
Dr. Andreas WITTEL did his PhD at the University of Tübingen, Germany. Currently he works as research officer at the Centre for Cultural Studies, Goldsmiths College (University of London). Main research fields: industrial sociology, anthropology of work, urban studies, new media, ethnography.