The last decades left us with the feeling that computers deal with predictable, controllable, consistent and true data. If a computer is not a clean machine, then what is? The next decade will bring us the computer as the world in which the unknown is as common as the known, where dark pockets of data are waiting
for us to release their riches…
Apple's Advanced Technology Group created Rumour Monger. It is software that does only one thing: it spreads rumours on a computer network. It is part of a research project into new ways of distributing information. Its workings are simple. A small piece of software runs on every machine on the network, allowing users to write and read rumours while storing and exchanging them in the background without disturbing the user. Rumours (text that users enter with their keyboard) are stored on the network, distributed over the different computers. If the user wants to add a rumour to the rumour pool, he chooses new rumour on one of the screen menus. A small window appears in which text is typed. He hits the spread rumour button and the rumour disappears from his screen. The Rumour Monger on his machine becomes active every couple of minutes and contacts a few other rumourmongers it can recognize. Then it exchanges ‘new material’. Because they, in turn, do the same, all rumourmongers constantly exchange rumours. Users view rumours in two lists: ‘unread rumours’ and ‘read rumours’. Double-clicking on a title in the ‘unread’ list opens a window to display the text of the rumour. After reading it and closing the window, the title reappears in the ‘read’ list.
The interesting part is what controls this environment: what determines when rumours are no longer rumours? The initiator can set a time limit on the rumour for it to de-activate itself. The other rules in the rumour-ecology determine ‘hot’ and ‘cold’ rumours. A first-time rumour is ‘hot’ and spreads from machine to machine. When a rumour ‘meets itself’ its status on that machine is changed from a ‘hot’ rumour to a ‘cold’ rumour, and two different algorithms determine how cold a rumour should be to become inactive.
A Case:
Such ‘enabling technology’ creates new possibilities of communication, and a group of 50 of us at a Dutch tv station lived on this new ‘canvas’ just to see what would happen. The following case-study charts our experiences of interacting with something that turned out to be confusing and left us changed. We installed the software and told everybody in simple terms what it was and how to get it to work. It turned out that software that spreads rumours is uneasy – you don’t know what to do. Spread rumours? Reactions ranged from questioning looks and disbelief to outright irritation. It takes some persuasion before people accept that an electronic space for rumours might be interesting.
When I decided to use it the first impression was one of effort: spreading rumours is no self-evident activity. The new ‘canvas’ made me self-conscious. Rumours tend to be associated with soft-spoken secrets, while speaker and listener glance around to check for others listening. So what happens if I use software doing just that explicitly? Should I type softly or use a smaller font?
There's also the problem of the rumour itself. I sit in front of my screen and wonder what actually is a good rumour (Something that is wrong? Something that is right? Something about somebody?) Making up a rumour suddenly seems hard, while recognizing one is simple. Is a rumour determined by its content, by the way it's communicated? Or both?
Reading rumours seems more exciting and certainly easier. I open every incoming title in the ‘unread rumours’ window immediately. All kinds of texts pass by my eyes. But few are the hot stuff I was waiting for. I’m not the only one that fails to produce rumours-on-demand. What's more: it seems that most of the messages concern the new medium itself: asking questions about its use, useless comments and others of a testing testing, 1-2-3’ nature.
The first days on rumour monger have no class. Confusion abounds for some weeks. Conversations try to delineate the what-is for an acceptable rumour, and rumour monger is not a good environment for conversations. The question How does this work? is a question and not a rumour. Rumours are statements that reveal answers to questions un-asked. Is Who is the smiling woman on the third floor? a sort of rumour? It certainly persists on the network while scoring a number of possible cues. A powerful rumour is A promising journalist that used to be a squatter is on leave in Turkey where his holiday is paid for by the dictator… (who is it? and what is this connection to a dictator? That’s something I’d like to know!) But how often do you run across a rumour like that?
All good rumours seem to concern something I want to know. But the software spreads all rumours to everybody and it’s hard to think of something that everybody really wants to know. The distribution of spoken rumours is within a sub-culture that values the information; it reaches just the right people. A rumour contains information, but never common knowledge. It seems to strive to become known.
Spreading rumours to everyone, I reach the right people, but also all others. I will have to phrase the rumour carefully to avoid problems. Is a rumour of value to some just because it is not known to all? One solution is that I write in code, with secret allusions that only mean the right thing to the right person. I create a semantic sub-culture to make my rumours more effective. bd loves ad means more to me than it does to you: and it’s new to me.
The software plays other tricks on us. Rumours spread from machine to machine in unpredictable order or time. Users are tempted to answer a rumour, thereby starting conversations. While this may happen ‘in turn’ on their machine, on the next machine the answer may arrive before the question! Some people find a way to construct stories where every statement is complete in itself, but it clearly takes effort.
Then there is a maintenance problem: cleaning the rumour lists is not an advisable thing to do. When I clean the lists (there is a command that does it), the active software always amasses rumours, and after a couple of hours I’m left with as many as I threw away. With over a thousand rumours it takes more than a day for all rumours to turn up again, but they do. The rumour itself determines if it is ‘active’ or not, and there are no other ‘filters’ that I can use. They grow on me like fungus, giving me a feeling of the network having a life of its own, as if I’m part of an environment that is, at least partly, beyond my control – just like most other things in normal life but unlike most information systems, that behave so predictably.
Rumour Monger also gives us a new kind of freedom: to know without knowing and say without saying. The fact that rumours don’t have the name of their creator attached is a new experience. It’s a new way of telling stories, building an environment of shared knowledge that is nothing more and nothing less: knowledge shared equally, with no separation between the knowers and the non-knowers. This new class of information gives new vitality to the inhabitants of the network, hearing the unheard-of from the insides of the organization.
How, in the end, did it survive? It didn’t. We had to remove the software because it interfered with the communication hardware that connects the different networks into the wide area network. It blocked them and made that we couldn’t do our normal work. We took a chance, and it was fun while it lasted.
What did we Experience?
We can use sociology or information technology as the starting points for understanding the experience. Encyclopedia Britannica treats rumours under ‘collective behaviour’. As a social phenomenon it has four conditions: both the interest and the ambiguity about an event must be high; the demand for news is greater than the supply of information; the group shares the need to act but is reluctant until the situation can be better defined and they need to act in concert.
These conditions appear in three situations: firstly in totalitarian environments where information is strongly controlled, secondly when events threaten the understandings on which normal life is based and lastly when a strong, shared incentive to act is blocked in some way.
We never thought about them in this way, but felt it. While we were struggling with the creation of rumours the organization was at peace. But when it went through a really rough patch recently, I overheard people saying that with Rumour Monger the situation wouldn’t have gotten out of hand the way it did. These events which threaten the understandings on which normal life is based suddenly brought Rumour Monger back to mind, but now as an instrument of direct social change.
Another approach is the ‘distribution of information’ as the spread of ideas: rumours as waves in the organization's meme pool. They might be compared to viruses, all of which have one thing in common: they spread themselves, duplicate, promulgate, multiply or reproduce ‘just because’. What else they do makes all the difference: some don’t do anything, some do nasty things, like give people flu or erase their hard disk. Rumour Monger is like this: spreading ideas without hesitation. Even if we want the bad to stay put, what if we have a way to spread the good and just? What if we had an engine that amplifies good ideas by bringing them into contact with others, human or electronic? We would increase variation, stimulate people to change by providing information in new ways. Based on our own experiences using networks the last seven years we know that the stimulation of variation increases the effectiveness and the quality of an information network. Old information systems try to handle complexity by reducing the variation, increasing the predictability. New networks control complexity by coordinated stimulation of variation. How should we stimulate this in order for it to be productive is something that needs to be researched. Rumour Monger might be one of the instruments.
Artificial life researcher Thomas S. Ray creates computer simulations of simple organisms competing for survival in cyberspace (‘Living together’: Cybernetic Parasites, Scientific American, Jan.1992). His artificial ecosystems remain more diverse when parasites are present. In some simulations, only eight out of 20 unparasitized species survived. With just one type of parasite, twice as many host species persisted. Can some ideas be compared to parasites? Could Rumour Monger have supported the longevity of the inhabitants of the information environment?
Yet another angle is the introduction of an existential dimension in information systems. Here the sociological and informational approaches meet. Information systems carry ‘descriptions’ of the organizations they mirror. Today, these provide only a narrow view on the world that is the organization. Rumour Monger adds a layer of existential meaning that changes the information system into an environment where networks are not only keepers of information, but link us into webs that supply meaning beyond and above mere data in databases. Tools like Rumour Monger support the tender communications between people but add little information with ‘direct use’.
This reminds me of thick descriptions – a term coined by philosopher Gilbert Ryle – that appears in ethnographic multi media. Ricky Goldman-Segall, who worked at the mit Media Lab and now runs Merlin (the Multimedia Ethnographic Research Laboratory) in Vancouver, wrote on thick descriptions in 1989, quoting Clifford Geertz:
Thick descriptions are descriptions which are layered enough to draw conclusions and uncover the intentions of a given act, event or process. In a video environment, thick descriptions are images, gestures, or sequences that convey meaning. Neither the quantity nor the resolution of the images make the descriptions thick. What creates thickness is the ability of the visual description to transmit what is really being ‘said’ (...) they provide us with a way to articulate the meaning of what we see, and, they help us to come to terms with one of the problems inherent in observational research – the fact that it tends to resist any kind of systematic evaluation and, like all interpretative approaches, it is ‘imprisoned in its own immediacy or detail’. (The interpretation of Cultures, 1973)
It is the extra that seems to bring life to things. Geertz’ treatment concerns anthropology, but it certainly has parallels. He states that people are suspended in webs of significance, a viewpoint he attributes to Max Weber. Anthropology, by separating the thin from the thick descriptions, sets up a hierarchy of meaningful structures to gain a more insightful view into organizations:
''As interworked systems of construable signs (what, ignoring provincial usages, I would call symbols), culture is not a power, something to which social events, behaviours, institutions, or processes can be casually attributed; it is a context, something within which they can be intelligibly – that is, thickly – described.
''
The understanding of how the symbolic is interwoven with all aspects of an organization will give us a way to understand that culture. Rumour Monger adds to an organization like ‘thick description’ to an event. It brings the life of the organization out into the open, exposing it for all to see and participate in. It is ‘extra’ information without a value ‘in itself’, changing the environment of the organization in which the data exist. It provides meaning by modulating the user instead of his data.
Software like Rumour Monger is playful, but no game. It creates a new class of information, a new ‘shared environment’ where people coordinate their activities in new ways. It is one of the mechanisms that infest networks with more humanity by providing nothing more than an electronic virtual space where the constraints for communication foster variation. It left us as different people.