Introduction

Computers were not invented to facilitate human interaction. Quite the contrary, they were envisioned as automatons capable of replacing humans in arduous tasks or indeed as capable of providing a desirable alternative to faulty human logic. Leibniz, in the mid-seventeenth century, expressed the rationalist belief that a “universal logic machine” would be able to perform the necessary calculations to arrive at fair solutions to even highly complex social and philosophical issues, thus providing the foundation for a more just society (Mayer, 1999:4). More modestly, the 19th century inventor and polymath Charles Babbage (1791-1871) achieved funding from the British Government for his visionary “Difference Engine”, a machine capable of producing accurate arithmetical tables of obvious interest to both industry and military. The machine would then be able to replace human number-crunchers known at the time as “computers” (Babbage, 1832:99).

The computer-as-replacement vision remained firmly in place until the mid-twentieth-century. British mathematician Alan Turing broke new ground with his systematic unveiling of the foundation for generalised logic and famously argued that computers might one day achieve human-like intelligence by sheer force of calculation (Turing, 1950). Turing’s methodical vision of artificial intelligence (AI), however, falls squarely within the theoretical tradition of Leibniz and Babbage. An important break with this tradition was started in 1945 by American scientist Vannevar Bush (1890-1974). Reflecting on scientist contributions to the war effort, Bush claimed that a new challenge now lay ahead - a challenge of organisation:

“…our methods of transmitting and reviewing the results of research are generations old and by now totally inadequate for their purpose… Those who conscientiously attempt to keep abreast of current thought, even in restricted fields, by close and continuous reading might well shy away from an examination to show how much of the previous month’s efforts could be produced on call…truly significant attainments become lost in the mass of the inconsequential.” (Bush, 1945:3).

What science needed, according to Bush, was not more research but rather an effective means of organisation and retrieval. In other words what was needed was a system capable of assisting, not replacing, the human effort. Bush’s article is often seen as introducing augmentation research, the systematic focus on “applications that extend people’s capacities to create, think, and communicate.” (Mayer, 1999:11). More specifically, Bush’s idea of an association machine, the Memex, looks remarkably like a theoretical forerunner of today’s Internet.

Such thinking conceptually joins the computer to a human user. Whereas the machines of Babbage and Turing were built (and sometimes programmed) and then left to do their superhuman calculations, a device such as the Memex is more like a mental infrastructure. It does very little on its own. Hence the need to focus systematically on the interaction between user and machine, a field known as Human Computer Interaction (HCI).

HCI is a theoretical counterpoint to AI research. Although philosophical approaches to AI do take human psychology and perception into account, AI may be thought of as dealing primarily with the inner workings of machines. HCI, on the other hand, focuses on hardware-external processes, specifically on those of representation, communication, human psychology, and design. In MIT psychologist J.C.R. Licklider’s original vision of the ‘symbiosis’ of man and machine (see also Hafner & Lyon, 1996:27-39) he stated the hope that

“…human brains and computing machines will be coupled together very tightly, and that the resulting partnership will think as no human brain has ever thought and process data in a way not approachable by the information-handling machines we know today.” (Licklider, 1960/1999:60).

Slightly more modest statements tend to characterise modern day HCI, but the focus on enhancing human capabilities by thoughtful interaction design remains central.

To computer scientists of the 1960s it may have appeared that a wide range of professions was taking a keen interest in the machines. What had until recently been a highly technical endeavour was now being transformed into an interdisciplinary field. However, another conceptual challenge was under way and found expression in a 1968 article by Licklider and Robert Taylor (Licklider & Taylor, 1968/1999). The article began immodestly: “In a few years, men will be able to communicate more effectively through a machine than face to face.” (Licklider & Taylor, 1968/1999:97). The emphasis was no longer on the inner workings of machines, nor the interaction of machine and user but rather on the interaction between people through computers. This marks the start of research into Computer Mediated Communication (CMC), which is the research tradition most clearly associated with this thesis. The history and present state of CMC will be addressed in the following chapter. At this point, it is sufficient to note that within the modern conceptual history of the computer, the technical sciences (AI) were supplemented by the humanists (HCI) who were soon followed by the social scientists (CMC). From an academic viewpoint, then, the computer exists in an interdisciplinary field and while this has obvious advantages, it also means that there is some disagreement as to how the computer is best described. This thesis presents only one approach to CMC and on many important aspects of the discipline it has very little to say. On the subjects of AI and HCI, of course, it makes almost no comments at all. But an effort is made to point to bridges between academic categories.

Focus and demarcation
As has been implied, this thesis centers on the necessary design conditions for computer-supported cooperation. Social issues pertaining to online interaction are analysed on the basis of existing sociological theory with the specific aim of determining if there are analytically important differences between interaction in offline and online settings. This leads to a description of how knowledge of online dynamics may be used to further cooperation and trust in collaborative computing.

This, however, is no manual for software designers. It is not a list of do’s and don’ts for the aspiring online community developer. Still, much effort will be put into phrasing the analysis in as accessible and concrete terms as possible. It is my goal to enhance our theoretical understanding of online dynamics, but if the analysis can be put to actual use as well, so much the better.

Theoretically, I draw upon sociology and most specifically the field of “social game theory”. In this vein, I also rely upon sociologically inspired elements of media studies, such as “medium theory” and on both interpersonal and mass communication theory. From computer-specific research, the analysis draws inspiration from Computer Supported Cooperative Work (CSCW), general CMC, and studies of “adaptive structuration”, that is, the relation between software design and software use (Preece, 2000:172).

Compared to what is perhaps the better part of research into “virtual communities” this thesis touches only sporadically upon issues from social psychology and only in a critical context in Chapter three and in the final chapter will I touch slightly upon the much-discussed wider implications of computer-supported interaction.

Problem statement
What is cooperation and how does it come about? Is online interaction different enough from traditional interaction to warrant a distinctive description of cooperation?
How will answering these questions enable us to understand and affect the dynamics of interaction in computer-supported social settings?

Structure of the argument
Chapter one concerns the characteristics and historical development of computer-supported communities. The goal is an adequate description of the phenomenon.

Chapter two describes the logic of cooperation from a sociological perspective. Classical theories of the state are introduced and then described in the language of game theory. The goal is a description of cooperation and of the preconditions for stable communities.
Chapter three provides a theoretical framework for computer-supported communities. It explains empirical results from computer-supported communities in the framework of sociological game theory as described in chapter two and goes on to describe the intense theoretical debate that CMC has provoked. Research from CSCW is used to pinpoint important differences between online and offline interaction. The goal is to determine if and how prerequisites for social stability in online interaction differ from those known from offline communities.

Chapter four describes findings from a survey establishing actual user preferences. 300 online gamers answered questions of trust, and evaluated proposed design principles. Respondents displayed a clear predilection for communication and expressed attitudes toward trust that are noteworthy when analysed in the light of previous chapters. The survey serves to test whether real CMC users would find the design principles distilled from previous chapters appealing.

Chapter five sums up the argument. This leads to suggestions for further research and to a brief discussion of the claims that have been made as to the general implications of CMC.

Terminology
Pronounced interdisciplinarity coupled with novelty and huge financial interests has turned cyberstudies into a terminological minefield. Rather acute disagreement surrounds even the most fundamental terms such as virtual, and community, not to mention the constant problem of CMC dictionaries: interactivity. Other troublesome words include cyberspace, information superhighway, information society, and knowledge sharing. Fortunately, clarity is enhanced by the emergence of textbooks, influential conferences, and reviews. Until something resembling a general consensus is reached, however, it seems prudent to define the terms rather carefully. I use the following specialised terms:

CMC: Refers to actual computer-mediated communication as well as to the study of this phenomenon. The meaning should in each case be clear from the text.

Community: Refers to a group of people characterised by stable cooperation and frequent interaction. The term is used liberally to include ‘gatherings of convenience’ that would fall outside visions of gemeinschaft and natural local harmony. (Will be further described in The issue of community, page 4).

Game theory: Used as a label for sociologically oriented theories about the success rates of various strategies in game-like settings and the theoretical implications hereof. Though traditionally an economic discipline, it is used here in its relevance for individual and collective action more generally. (Will be further described in Chapter two).

Interactivity: Used in a broad sense as a measure of user control over form and contents of a medium product.

Sociability: Used here as a measure of how well a software product supports (favourable) social interaction (following Preece, 2000).

Usability: A measure of the user-friendliness of a software product. The discipline of usability testing may be seen as a practical application of HCI.

Virtuality: The term (as well as the adjective ‘virtual’) is generally avoided as it makes unwarranted ontological assumptions. When used it should be understood as synonymous to computer-supported or online. (Will be further described in Beyond alienation, page 4).

Since CMC is not a unified field and many details may not be generally known, I make extensive use of explicatory footnotes and illustrations where appropriate.

Theoretical assumptions
Metatheory, I believe, is mainly interesting in action. Siding explicitly with an ‘–ism’ is less valuable than clarifying each step taken in an argument, making sure that the relationship between premises, evidence and conclusion is as obvious as possible. In other words: The following argument may be wrong, but if so, I hope it to be obviously wrong (to those who know better).

Nevertheless, I do not wish to be seen as representing the (mistaken) idea that there can be science, philosophy, or communication without a priori assumptions. The ‘–ism’ that may be distilled from this thesis, then, is a modest reductionism. It is not the assumption that everything can be explained at lower levels but it is the idea that irreducibility is a very unfortunate axiom. If something can be explained by fewer, simpler premises it would be a shame not to – if nothing else such parsimony would spare us great deals of confusion. This leads to an important distinction, since the term ‘explain’ is perhaps too broad. A low-level explanation of a given phenomenon can yield insights even if it does not account for every aspect of the phenomenon. Take human consciousness. This most complex phenomenon is a result of less complex processes. Whether natural selection alone is accountable, or we need various sorts of mindless help (random drift, non-random mutation, embryological constraints etc.) is beside the point. Understanding whatever simple, mindless processes has led to the development of consciousness (disregarding only magic and divine intervention) would be tremendously important to artificial intelligence and, of course, any theory of mind. But it would be important without explaining, in details, how human consciousness works, or how it feels to be conscious. Similarly, an analysis of cooperation and trust based on simple models can be important without accounting for the complex cognitive details of how actual people evaluate the trustworthiness of each other.

Having said this, I have already hinted at an underlying ideal of this thesis. I believe that interdisciplinarity is not just important, but in fact almost essential for two reasons. Firstly, at least on the subject of human beings, there are obvious advantages in drawing upon a variety of fields. Surely, perception and interpretation can be studied as if the human mind was some sort of black box (just like the heavens could be studied without knowing the nature of stars) but such studies would be strengthened by an understanding of why the results emerged. Secondly, the alternative to cooperation may very well be conflict. Like it or not, scientific fields overlap. Humanists are making claims about the natural sciences (or about their foundations) and natural scientists are making claims about very concrete aspects of human behaviour. Wasting time and effort on fighting this development is rather futile and occurs to the detriment of all combatants.

Avoiding interdisciplinarity, of course, can also be defended on more sober grounds. It is not without danger to venture outside one’s field of expertise (as has been aptly documented by Sokal & Bricmont, 1998). But this is partially a question of mindset. A sober – and above all: modest – approach to adjacent fields is different from an arrogant excursion into foreign territory.

The architectures of trust
The idea that human action is a function of structure, systems, or architecture would be a controversial one in some sub-domains of sociology (Ritzer, 526pp). Nevertheless, such a notion lies behind most legislation and is certainly central to any theory that would justify the existence of a state. The issue may be split up into several segments, taking literately the metaphor of architecture. Imagine a hypothetical building equipped, among other things, with a staircase and a kitchen. Is the individual forced to use the staircase? Surely not, he or she can jump out the windows or climb down the wall. Is he forced to cook his food in the kitchen? No, but it would be practical since that’s where the oven is. However, over time and all things being equal, a majority of people are likely to take the stairs, sleep in the bedroom, and cook in the kitchen.

All things, of course, are not always equal. Disaster may strike in the kitchen forcing people to find alternative places to cook. But if unexpected things happen at random then, over time, things are equal.

Similarly, the relation between architecture and cooperation that is proposed here is not one of absolute regularity – it is not a deterministic theory. It cannot (necessarily) be applied to any existing computer-supported community to explain its present condition. Specific conditions may override the ‘pull’ of the system architecture. But retrospectively the theory may contribute to an explanation of the fate of the community and if we made an analysis of all such communities it would be able to statistically predict the odds that a new system would be visited by social strife (by whatever measure we set).

I should stress that this approach may come close to what Nancy Baym calls “the mistake to view patterns in CMC as direct effects of the medium.” (Baym, 1996:149). However, although Baym makes her point well, I believe her conclusion is partly rhetorical. Through her analysis, she merely defines anything, which looks like determinism or reductionism as false by default (or she makes the rather obvious point that complete context-independent determinism is wrong, in which case one can only agree). Those who would argue that technologies or media do not have effects are either stating the obvious or facing the challenge of refuting works within medium theory such as Meyrowitz (1985) (see also McQuail, 1996:107-118).

This, then, is how ‘architecture’ is used in the framework of thesis. It is the structure that at any given time influences action, and which over time, given certain minimal assumptions, can predict the fate of a system (at least in terms of probability). I assume that the actions that we wish to induce are those that require trust . Trust itself comes with many definitions, but one of the simplest will suffice: “Trust is anticipated cooperation” (Burt & Knez, 1996:70). Cooperation itself is the topic of chapter two.

While, to some extent, she subscribes to what I will later describe as the Myth of Liberation, the clear vision of how architecture shapes online behaviour was first expressed by Judith Donath (Donath, 1996). She says it well:

“…the future success of virtual communities depends on how well the tools for social interaction are designed. If they are poorly designed, the on-line world may feel like a vast concrete corporate plaza, with a few sterile benches… If the tools are well designed, the on-line world will not only be inhabited, but will be able to support a wide range of interactions and relationships, from close collaboration to casual people watching.” (Donath, 1996:2).

Download the full text (PDF, 2MB)