JP Rangaswami on Continuous Partial Asymmetry

JP is having a little fun with us as he coins the term "continuous partial asymmetry" to explain some interesting behaviors that grow from asymmetry in social networks. (They aren't 'partially' asymmetric, though, they are just asymmetric. He's playing off the "continuous partial attention" meme that Linda Stone coined eons ago.)

His central points:

[from Musing about politeness and “continuous partial asymmetry”]

[...] there are some very important points being made here [comments made by . Three in particular are worth emphasising:

1. People in a Web 2.0 network are not uniformly connected; some have more connections than others
2. Connections have directions; the number of inbound connections may far exceed the number of outbound connections, creating an asymmetric environment
3. This is particularly true of “default-public” networks such as Twitter; Flickr is also likely to evince similar behaviour.

To clarify with a bit of graph theory:

1. Graphs -- a way to mathematically describe things like social networks -- can be either directed, meaning that the arcs or relationships that link them point from one node to another, preresenting an ordering, or undirected, where the arcs represent sets. A directed relationship is like 'mother of' pointing from 'mary' to 'john' and 'child of' points from John to Mary. In some graphs, relationships are not directed, like a network of integers where the arcs could just represent membership in a set.
2. Asymmetry -- In the most general sense, think of the Twitter relationship of 'follows', where 5,108 people are following me (this morning), and only a few hundred of those are being followed back, and there are a very few that I follow that don't follow me. However, it is common to have directed graphs where symmetry is retained: where all nodes point at all other nodes and are pointed at by all others. For example, the typical project team is often structured (conceptually) in this fashion: all members of the group can interact with all others in the group in a symmetric fashion. Anyone can email the others for info, schedule a meeting, ask for review on a doc, etc. And the implication is that the others will respond when asked. But in general, asymmetry in social relationships is the rule: some pigs are more equal than the others. The project leader can assign tasks to others, but the others can't assign each other tasks, for example. Or people can ignore requests from others.

But the point that JP is headed for is dead on. Asymmetry in the social network engenders a certain social ethos:

In order for twitter to become asymmetrical, it must be OK for me not to reply to a tweet. If I am forced to reply then it doesn’t work. If I am expected to reply then it still doesn’t work. But if it’s OK for me to say nothing, then it works.

What is this thing that works? Asymmetric follow. Because I am no longer expected to reply to everything that comes in. People who receive a lot of snail mail or e-mail don’t reply to everything that comes in either, so what’s the difference? The difference is in the perception of polite behaviour.

It’s rude not to answer a telephone call; it’s rude not to call back when a voicemail has been left; it’s rude not to reply to an e-mail; in fact it’s rude not to provide sympathetic sounds when listening to someone on the other end of a phone. [That last politeness convention has had an unintended consequence ever since the mobile phone was invented, the regular need to intersperse conversation with "are you there?"].

It’s not rude to ignore a SMS. It’s not rude to ignore an IM. It’s not rude to ignore a tweet. Even an @tweet. Even a DM.

Years ago, I was working on a project for Microsoft regarding adoption of instant messaging in the enterprise. They had discovered that one of the top three barriers to IM adoption was that managers (in particular) did not know what the ground rules for IM use were, or should be. So I drafted a report, that was more or less an etiquette guide to IM use. One of the major poinst that I sought to make in the report was that several preconditions need to exist for IM to "work" in a company:

1. You have to turn IM on, and leave it on, so that others can see your presence status.
2. You should update your status -- like 'going to lunch till 1:30' or 'in finance meeting till 4pm' -- so that others can keep tabs on you without having to actually interrupt you.
3. It is OK not to respond to messages: it is not impolite. Otherwise the usage model becomes as complex and unwieldy as voicemail and email, where the assumption is that all messages will be responded to. The assumption in IM is that response is optional.

Or, said in a different fashion, updated to the world of open social networks:

1. You should stream your status so people can remain aware of you, or be able to determine what you are up to by reading your recent stream.
2. You can opt to ignore, filter, or block anything that others are streaming to you, and they shouldn't be upset by that.

There may be a slightly different etiquette around direct messages, which feels like email in many ways (see Scoble Doesn't Like Twitter's Direct Messages), but the basic premises of open discourse -- like the blogosphere, twitter, IRC, conference back channels, or enterprise-wide instant messaging -- is that people should be free to respond to what they want, not everything that comes their way, no matter who the source is.

Ross Mayfield jumps in:

[from Twitter Is Reply-Optional]

Maybe the tools, and burden from tools, have created actually productive practices, and tools. Where some channels have different burdens [participation costs] of reply. Twitter has a near zero burden of reply. DMs have a slightly less burden of reply, but still a cost to the sender in the form of relationship, initial and retention. Some of the best and busiest people say to me today that DMs are the best way to get ahold of them, when 'elst fails.

One of the unstated reasons for the decrease in participation costs that Ross is talking about is the shift from the email inbox -- where the user is consigned to being a file clerk, archiving, and foldering -- and the shift to streaming clients that power users of twitter invariably use. In the streaming client apps, messaging drift by, and fall off the edge. If I want to look, I look; if I want to reply, I reply; in either case, after a while the message falls off the time horizon supported by the tool, and then it is gone, archived in the cloud somewhere, but out of the temporal horizon supported by the tool. It is a 'fire, or forget' model.

This pattern drives the cognitive and productivity shifts that streaming tools engender in us and our modes of interaction. This is again that flow state of mind, the flow mode of operation that will soon be the norm. And the premises of asymmetric relationships and tools, and the ubiquity of that model of social interaction are a foundation of this shift we are undergoing, collectively and individually.