Over at GigaOM, Matthew Ingram is worried over a report that Twitter is working on an edit feature that would allow users to make small changes to a tweet during a limited window after it’s first sent and have the correction propagate across the Twitter-verse by automatically updating retweets of the original. In Matt’s view, the ability to change erroneous or sloppy tweets after the fact, while potentially appealing to journalists who want to use the unruly platform to break news, nonetheless risks losing something essential about the Twitter stream of consciousness:
I have to wonder whether the inability to alter a tweet isn’t one of those things that makes Twitter what it is — for better or worse — just like the artificial restriction on the length of a tweet. The whole idea of Twitter is that it’s a stream of content that flows by, and you check in periodically and then the stream continues to flow. It is the essence of what Robin Sloan meant when he described “stock and flow.” And mistakes are part of that, like eddies in the stream.
So when mistakes are made, what do we do? We correct and retweet, or we just move on and update. And in many ways those are all unsatisfactory, because we know the original error is out there — and as Craig Silverman of Regret The Error has noted, the mistake often travels quite a lot farther than the corrected update. But even if Twitter gives us the ability to turn back time and edit the stream briefly, it will only affect native retweets, not the copy-and-paste or manual kind.
In the end, even if it did work across the entire network, I still think we would be giving up something beneficial — and that is the knowledge that even though Twitter is a news platform, it is messy and often incorrect and imperfect, just like its users.
What Matt seems to sidle up to, but not quite embrace, is what might be called the Efficient Twitter Hypothesis. In finance, the Efficient Market Hypothesis holds that financial markets are “informationally efficient.” In such markets, prices for traded assets are said already to reflect all publicly available past information, so its theoretically impossible to consistently achieve better than average market returns.
A corollary to that idea is that informationally efficient markets are inherently self-correcting. If prices get out of whack relative to value, that information will ultimately be processed by the market, knocking prices back into whack. Therefore (if you buy the theory) there is no need to regulate markets because markets are self-regulating. In fact, it’s theoretically impossible for regulators to improve on market outcomes because no regulator can have better information than the market as a whole.
The Twitter-verse as a whole has something of that self-correcting quality: initial reports of breaking news are often wrong, misleading or otherwise inadequate, but over time better information calls out the bad.
It’s that quality, in fact, that makes Twitter fundamentally different from a news organization, even though journalists often furnish much of the Twitter fodder — both good and bad. Twitter is organic. It processes information in real time. Journalism is not organic, or it’s not supposed to be. Some measure of processing — editing — is expected and assumed to have happened a priori. Editing after the fact, therefore, both misses the point of editing and clashes with the self-correcting quality of the information stream it attempts to influence.
So, in Efficient Twitter Theory, enabling post hoc edits to tweets make no more sense than regulating financial markets does in Efficient Market Theory.
Of course, just as economists argue over the “weak,” “semi-strong” and “strong” versions of the Efficient Market Hypothesis, the timing and strength of Twitter’s self-correcting instincts makes a difference. Asset prices may self-correct over time, but that doesn’t mean a lot of people won’t trade on bad information in the meantime. The same is undoubtedly true with Twitter.
Update: According to CNET, Matthew needn’t worry. Twitter says it is not “actively working” on an editing feature.