How can information, passed on 140 characters at a time, contribute to any kind of meaningful exchange? I think it is quite reasonable to write twitter off as a new faddish technology that many people are using just because it seems cool. However, behind the surface level appearance of twitter, and other similar forms of communication, something extremely useful seems to be developing.
This essay is not only about twitter but the more general concept of short, rapid, open communication, which I’ll call “blurts”. In essence, what I will argue is that blurts might not have much meaning in themselves, but a web of blurts among many people can be valuable to many of the participants. A collection of blurts is akin to a form of brainstorming, but in which some minimal structure appears, allowing the most valuable ideas to persist and be developed. Importantly, that structure comes from the interests of the participants, not from some kind of moderator.
There are four example formats for blurts that I want to consider here, although there are many others floating about that I could have discussed. They are: 1) facebook status updates, 2) twitter’s tweets, 3) moves in signtific’s forecasting games, and 4) comments in Tim Gowers’ mathematical blog. Each has a different character but each shows how the properties of length, rapidity, and openness can play into a successful conversation.
Facebook status updates
From the beginning, one of the identifying characteristics of Facebook has been its status updates. Starting with “John is…”, each invites the user to share something about the user’s current status. With this visible to a user’s friends, a shared sense of life and community can develop among users. However, in itself, it doesn’t provide any real sense of conversation or allow for the development of ideas. It is a broadcast mechanism, which has low barriers to entry, but then doesn’t invite much reflection. Although one person might update their status in response to another’s, the collective of people on Facebook would not easily be able to follow the thread of such updates, especially if they were not friends of both/all parties updating their statuses in this way.
When Facebook allowed users to comment on almost anything in a profile, the character of status updates changed. Very quickly, people seemed to start commenting on others’ statuses, generally sticking to the short format of a status update. The short form retained the low barrier to entry, encouraging spontaneous responses. With status updates from moving down a newsfeed fairly quickly, it encouraged people to make rapid responses, while the conversation was still present in the collective’s newsfeed.
In some cases, long chains of responses formed, often including comments from people who didn’t know each other, but were connected only via mutual friendship. Some of these conversations let ideas develop but most stayed at the level of shared personal information.
Perhaps limiting the development of these conversations into something substantive is the nature of Facebook friend relationships. Facebook is set up in a way that encourages people to be friends within the service only if they have some sort of pre-existing relationship. Why would you share so much personal information with people with whom you have no relationship? This culture within the service means that status-update conversations remain essentially closed to the outside world, which is in many ways desirable, but is also limiting in terms of collective development of creative ideas.
Twitter at first seems to take just the status update of facebook and separate it from the rest of the personal information you might share. So, at first glance, it might seem to be merely a limited version of Facebook status updates.
However, the character of tweeting is entirely different as the connections a user makes in twitter are made not necessarily within the framework of existing relationships. This means that the information put out in tweets, while it might be personal, is of a form that is inviting or open to anybody who wants to listen. Networks of twitter relationships could form identically to those in Facebook by using the locked conversation mode, but the barrier to entry is quite high, as users need to allow each user to join the conversation, and the value of doing this for twitter, as opposed to just sharing information via Facebook, seems limited. If you want to restrict your online world to those who you already have an existing trust relationship, twitter doesn’t seem the right tool.
Twitter asks “what are you doing?” as if it were merely a status update, and some people do still use it that way, in an exhibitionistic way, but the great value of twitter seems to be arising from a different type of use.
Characteristic of twitter is the ability to impose some kind of structure on the conversations. By directing a reply to somebody, you can demonstrate a link, or chain between people, but still in public, allowing others to trace a conversation, and benefit from the back and forth between multiple people.
This kind of public conversation allows others to join in with comments, information, questions, or other kinds of interaction. The conversation is open, but non-linear. You can jump into it at any point, and leave again when it bores you. The group of people listening will extract precisely the amount of value it has to offer, with the potential to add to the collective value.
Barriers to entry are low, conversation can be rapid, and it happens in a collective space, which allows value to be extracted and generated. Furthermore, the structure allows users to trace conversations, especially if hashtags are used to indentify a tweet with a topic.
Signtific’s forecasting games
The signtific lab’s forecasting game “Free Space” was a very interesting Web experiment, which aimed to have a group of people forecast possible futures in a given scenario. In this instance, the game asked participants to think about what would happen if the entry barriers to space, via the existence of $99-to-buy-and-launch cubesats, was dropped to be almost negligible.
The game ran from February 18-19 attracting a few hundred participants, contributing a few thousand entries. Each turn was played by a participant filling out a game “card” of one kind or another. Each card had space for 140 characters, just like twitter’s tweets. The cards could be root level or could be attached to other cards.
At the root level, players were asked to contribute a card that answered either “So what’s the BEST thing that might happen?” or “So what’s the WORST thing that might happen?” From that starting point, people could develop the ideas further by playing a card attached to specific forecast about the best or worst. Those cards came in four types:
1) Momentum: If that happens … What might happen NEXT?
2) Antagonism: Disagree? What might happen INSTEAD?
3) Adaptation: How might this play out DIFFERENTLY in your field or part of the world?
4) Investigation: Curious? Ask or answer a FOLLOW-UP QUESTION
Point were awarded for playing cards, for having others play a follow-up card on one of yours, for having a forecast chosen as a “super-interesting” card by lab guides, or various other criteria.
The most interesting innovation in this game, in the world of blurt-messaging, is that the structure allowed ideas of most interest to the players to develop using the four allowed follow-up forms. Uninteresting ideas didn’t get any response, while interesting ideas were teased out into what were often far more interesting forms unanticipated by the original card player.
The time limit for the game and the point structure encouraged players to submit many ideas, not to just dwell over their one most interesting idea, which might not even be pursued by the collective if it was only of interest to the submitter.
Essentially, the game was a group brainstorming session that allowed some ideas to be developed without halting the highest-level brainstorming. Just as in brainstorming, the mot difficult part of the process was trying to swim in the flood of top-level ideas, as so many were contributed. The volume was clearly too large to really cope with, as many ideas were contributed many times over in slightly different forms by various participants who had clearly not read the previous comments.
In one sense, that repetition did not matter because the game was explicitly looking for “outlier” ideas, the ones that were unexpected but insightful. Ideas that came up many times were probably not as interesting as they were “obvious” to some degree. What might happen to those ideas afterward was more interesting. However, it would have been useful if the people who had an obvious idea, had instead seen that somebody else had put it forth, perhaps with some variation, and then contributed by exploring the variation or by pushing the idea toward its consequences.
The rapidity of the process meant that ideas had to be thrown in, based on intuition and dreams, without the filter that might knock out ideas on more reflection. Instead of self-censorship, lower-quality ideas died a natural death because nobody followed up.
By the end of the game, many run-of-the-mill ideas had been ventured, but some non-trivial ideas had been allowed to develop toward interesting possibilities.
Gowers’ “Polymath” collective mathematics project
Timothy Gowers currently hosts his “Polymath” experimental approach to collective mathematics on his blog, using the topic of density Hales-Jewett theorem, a mathematical theorem that has been proved using one approach, but for which Gowers thinks a different approach might also work. The theorem itself can be stated in a way that undergraduates can understand but the proof itself is expected to contain some quite advanced mathematics, bringing in a wide range of ideas from different areas of mathematics. (Parts of the conversation reside on the blogs of Terry Tao and Michael Nielsen.)
Some of the principles of his collaborative approach include that participants shouldn’t go off and work by themselves to develop some part of the mathematics in detail. Instead, they should venture the core concept and see if it resonates with the rest of the participants. Ideas that are only partly formulated or imprecise are also welcome.
He summarizes the concept of what a single contribution should look like as follows:
One can summarize much of the above by saying that each comment should represent a ‘quantum of progress’. That is, the discussion should have been advanced in some small way that is not obviously complex or divisible.
He suggests that people should only be contributing ideas that don’t require them to go off with a piece of paper to develop them. In other words, the ideas should be something like the blurts already discussed: relatively brief, rapid, and open. Here brevity does not meet 140 characters, but most contributions are no more than a short paragraph.
So far, the conversation includes a few hundred contributions and Gowers is currently dealing with the problem of how to handle all that information, by considering threading the comments, and also by breaking it into a few key themes. The threading is akin to the follow-up card structure of the forecasting game, determined by the participants, while the breaking into themes relies on Gowers’ special role as host. Given the technical nature of the discussion, some use of expertise seems appropriate in this case. After all, the different areas of discussion came about because participants chose to follow them, and Gowers is merely collecting the decisions already made by the group.
I am not in a position to judge the progress of the experiment, but my ill-informed impression is that people are definitely making forward progress, even if the key to the solution is not yet present.
Brevity is clearly a key aspect of blurt communications. It contributes to the collective development of ideas because it prevents one participant from steering a collective discussion in a direction that is only of interest to themselves. A common issue in online discussions is that a long post with many ideas attracts responses targeting only the most controversial parts of the conversation, with many potentially interesting parts of a message abandoned. By breaking the post into the “quanta” of information, each can be tackled on its own terms. Brevity also lowers the barrier to entry because there is no expectation of deeply developed ideas.
The tools behind all our examples of blurt communication have a very low barrier to entry because participation involves nothing more than typing a short message. Part of the rapidity of communication comes from the short form, but some of these formats also encourage rapid responses. In the case of Facebook status updates and twitter, leaving a blurt dormant means that it will pass out of sight quite rapidly, and so there is some requirement for rapid response to engage in the conversation. This can be problematic for continued discussion or development as it requires constant engagement.
The short timeframe available for responses encourages spontaneous, gut-level responses, which can then be taken further by others involved in the conversation. I am sure many psychologists could wax at length about why rapid responses can access a different and perhaps more creative part of the brain, but I’ll wait to hear that summary in 140 characters or fewer.
The forecasting game establishes a framework which builds rapid responses in as part of the process, ensuring that people go with their impulsive responses. The point-scoring process encourages large quantities of responses, and the restricted total time for the game means that those interested in scoring well should be responding fast.
The Polymath project is less inclined to rapid response, but the idea of contributing quanta of ideas and partly-formed means that if a participant has a gut response, they feel welcome to just throw it into the mix. The culture of the discussion encourages rapid participation.
The rapidity of conversation has advantages and disadvantages in terms of an active attempt to generate discussion, but the rapidity certainly allows for a certain style of interaction that is potentially valuable.
The most active of the blurt communication forms are those that are open. Facebook status updates lack this feature. Given that friends on Facebook probably share some commonality in thinking (to be friends in the first place), the lack of openness would appear to restrict the possible range of contributions that could be very creative and useful to others.
The other formats are considerably more open, and the greater quantity of discussion is likely a consequence. The explicit invitation for strangers to contribute ideas seems key to broadening the creative input to discussions. This is most clearly seen in the forecasting game example, as people contributed based on their own personal experience and interests, which covered an incredibly wide range of topics.
Blurt communication: interstitial creativity
A key property of blurt communication seems to be that the development of ideas doesn’t occur during the communication process as much as in between the contributions. Each blurt contains relatively little information but it has the potential to inspire new ideas. The space between blurts, and between participants, seems to be where the creative development of ideas happens. By using the short, rapid, open nature of blurts, these interstices are allowed to play a large role in the collaborative discussions, and so maximizes the potential for creative development.
Obviously, blurt communication is not ideal for all situations, but current tools and technologies seem to be allowing this form of discussion that has powerful potential beyond more traditional forms of discussion.