Twitter babble twaddle

Doug Clow (@dougclow) has also posted an excellent piece explaining how Twitter is social grooming, which is sort of my point. Some key quotes:

Fundamentally, though, this study (almost) entirely misses the point of what people on Twitter experience. It sampled the Twitter public stream, which is the total assemblage of what everyone using the service is producing.

But what looks like ‘pointless babble’ isn’t pointless, if it’s from people you know or care about. It’s social grooming, it’s keeping in touch. It’s what most human conversation is about. If you think this stuff is pointless babble, you’re really not going to enjoy parties. Or indeed be likely to maintain fulfilling personal relationships. On Twitter, you get to choose whose ‘pointless babble’ you want to follow. Almost nobody who actually uses Twitter uses it by reading the public stream.

If you learn about Twitter by reading these sorts of reports, you’ll get a bizarre view that really tells you very little about what it’s like to use as a service.

And this brings me to the general point about teaching with new technology: you can do the most methodologically sound research about Twitter you like, but without a decent appreciation of what it is to use the service, you’re going to struggle hard to make sensible use of it in teaching.

Not just in teaching, of course, but that’s Doug’s thing.

I found this post via Memex 1.1, where Pear Analytics’ white paper is described as “cod research”.

The redoubtable danah boyd has also weighed in.

However the icing on the cake is this news: The “Sarah” who got so defensive and sulky on Stephen Dann’s blog is none other than Sarah Monahan, who played the daughter “Karen” “Jenny” in long-running Australian TV sitcom Hey Dad…!.

[Update 19 August 2009: Sitcom FAIL corrected: see my next comment. I never did watch the program myself. Sitcoms aren't my thing.]

@Sarah: Oh dear. I am full of sitcom FAIL. I got the character’s name right when originally writing elsewhere. I’ve fixed it now. Thanks. And I appreciate that we’re moving to a more sensible discussion.

A few points from your response…

The idea that online communication, via Twitter or anything else, is not “HUMAN communication”, that only face-to-face-communication is “real”, is fundamentally wrong.

No-one argues that high-bandwidth communication face-to-face isn’t “better” than lower-bandwidth phone calls or even-lower-bandwidth texting. But your mother is still your mother, whether you’re sharing a cuppa over her kitchen table, or firing off a quick email while travelling. Your best mate at work is still that, whether you’re sharing cocktails on payday, or SMSing her a snarky comment from the bus.

What’s important is the nature of the relationship between the people and the context of their communication, rather than the channel through which that specific nugget of communication takes place. Once you look at it through that frame, a few things become easier to see.

I’ll slip in a few buzzwords because a bunch of my readers have been through this already…

The mix of always-on mobile connectivity and “naked conversations” (out in public like Twitter, or like IRC two decades ago) means that a lot of the social grooming conversation takes place continually throughout the day, rather than only when people meet face-to-face.

It also makes it possible to maintain more of the weak social ties — not just the old Dunbar number of 100 to 150 acquaintances, but hundreds or even thousands.

If it’s only 3.30pm and I tweet “I think it’s getting close to Xantini time”, I’ve just signalled to 3000 people that I’ve had a particularly frustrating day at work — I said “Xantini time” rather than “beer o’clock” — and that I’m open to catching up after work. People may respond with sympathy, or suggest bars, or point out that it’s rather early in the week. They’re reinforcing their social link to me.

Others may ask “What’s a Xantini?”, signalling that they’re feeling excluded by the insider language but want to be closer. I may choose not to respond, perhaps indicating that I don’t like them all that much.

But your study would call my tweet “pointless babble”, because it fails to consider the context.

By the time you catch up for your “human communication” on Friday night, well, many of the psyche-reinforcing social grooming exchanges have already taken place at a distance. You only need a few more exchanges over the first round of drinks, and you’re done. That social relationship has now been maintained, says part of your brain. But another part is dissatisfied because it didn’t get that dopamine hit of satisfaction from a huge burst of positive social grooming all at once. You feel empty.

Meanwhile, your friend detects a gap in maintaining the intimacy of some other social relationship, so they start firing off a few messages on their BlackBerry.

However none of this is particularly new.

Sherry Turkle wrote Life on the Screen in 1995. Robin Dunbar wrote Grooming, Gossip and the Evolution of Language in 1997. danah boyd is all over this stuff. Journalists did the whole “Twitter is pointless rubbish” thing two years ago, but have now worked Twitter into their daily workflows.

The reason the story got such a good run is that Twitter has just passed the peak of what Gartner calls the “Hype Cycle”, and the feral pack is ready for disaster stories. And “pointless babble” is a great headline. But the backlash was because this particular Pear Analytics research is, alas, rubbish — which is obvious to people who’ve been working on this stuff for years.

OK, so how can it be fixed?

You ask, sensibly:

How do you get a view of how it’s being used by Everyone, without looking at the public timeline?

Personally, I don’t think you can do it this way. At least not solely so, because you don’t know the social context of the tweet, and because a single tweet might have multiple, overlapping purposes.

Understanding how people use Twitter would be a valuable piece of sociolinguistic research, but it wouldn’t be easy. I imagine that “real” research would need to include:

• An initial sampling of tweets and their creators, aimed at identifying the different styles of Twitter usage and the categories of tweets. That’d have to start from an open-minded position, without having to necessarily include a predefined category of “pointless babble” for marketing purposes.
• Perhaps some qualitative research, discussing with the tweets’ creators what those tweets mean to them in their social context. That could be backed up by automated mapping of their social networks on Twitter — but it’d also have to consider the fact that what’s visible on Twitter is only part of their communications mix.
• Somehow factoring in the people with “protected” updates, and Direct Messages, and making sure that the people who gave their permission for you to look at their private stuff were still representative of the whole population.
• Then the bigger sampling and refining of the methodology.

That all sounds like a thousand times more work. We’re talking PhD thesis here, not a quick market intelligence task. But that said, there’d already be academic work in this area to build upon, even if there aren’t already commercial tools.

If we’re after quick-fix analytic numbers to impress clients, well, we’re not on the same page.

I did a follow-up piece in Crikey yesterday, From hype to backlash, Twitter’s path is inevitable. It’s behind the paywall, but the core point is that there will inevitably be a flood of negativity about Twitter and other microblogging services. It’s passing what Gartner’s Hype Cycle calls the “Peak of Inflated Expectations” into the “Trough of Disillusionment”.

@Sarah: Claiming you never intended to say there’s anything wrong with babble is a tad disingenuous, I reckon. “Pointless” is a rather judgemental word. Even saying “some people do share way too much information” is a personal opinion — which you’re entitled to hold, of course, but it’s still making judgements about other people’s communication.

This is the very core of my disagreement with the Pear Analytics study. Whether something is “pointless” or not can’t be judged by a third party, only by the intended participants in the conversation.

Notions of what’s “appropriate” (a word I loathe, as it happens, because it derives from class-ridden Victorian notions of propriety, i.e. “proper” behaviour) in public communication are rapidly changing. I used the term “naked conversations” quite deliberately because it’s the title of Shel Israel and Robert Scoble’s book on this very subject.

@Stephen Wilson: The “I’m at a club” tweet could represent this guy coordinating his social life, celebrating that he doesn’t have to work that night, revealing that he’s no longer ill and is healthy enough to go out, signalling his location so his dealer can deliver the drugs — any manner of pointful things. While we can perhaps argue about the meaning of “babble”, the term used was “pointless babble”, and I can see many ways in which such a tweet would be far from pointless.

As for phatic communication not being possible on Twitter because of the time lag, I must strongly disagree. The lag is irrelevant. For example, if someone asks for technical support and gets an answer, a tweet like “Gotcha!” means they’re done, or “WTF?” means they need more help.

The meaning might also be hidden in the private language of the participants, such as in my tweet, I think I want to be a chihuahua too. Though visible to 3000 people, the full nuances of that are known to maybe 3 people.

Phatic does not mean “non-verbal”.

I’m also starting to think that the term “one to many”, originally designed to cover broadcast media, doesn’t quite capture the nuances of something like Twitter. Maybe it’s more like “one to few, with onlookers”.

@Quatrefoil: Yes, it’s all about the audience. Or, as I prefer to put it, the “participants”.

As an example of a personal blog which is vastly entertaining, I can thoroughly recommend Jenny the Bloggess. If you have a strong constitution.