There's a piece by Debra Logan on the Gartner Blog Network, "Twitter and e-Discovery," that goes over some fairly straightforward stuff about e-discovery and social media. What struck me as interesting was an observation at the end of the piece, pointing out that because of the briefness of a tweet, it is more likely decontextualized than are other discoverable utterances (doodles on pads at meetings?), at least when it's looked at outside the flow it first appeared in. It's context that gives or controls meaning, and the briefer the utterance the less each word is shaped by neighbouring words, near and far. So it could be difficult to explain, and prove, what was meant by an elderly tweet, and, correspondingly, easier for someone to place a construction on it that the twitterer never intended. It's a little like overhearing a snatch of conversation as you stroll down the sidewalk: you puzzle and play with the bit you've understood.
Of course, this is nothing new. Emails and text messages can be brief. What makes it a little interesting, I think, is the deliberate structure of Twitter that assures, in effect, that bits of language will be floating free of context. The problem, to the extent that it is a problem, is built in to the medium. This is what makes Twitter fun, I guess: the more text/context there is, the "heavier" the thought, and heavy is… well, boring to many if not most. (It's also why people sometimes "serial tweet" in order to provide real meaning to what they've said by "violating" the 140 character limitation in a bunch of sequential tweets.)
Paradoxically, perhaps, the briefer utterances are likely to endure longer than the, well, longer utterances. Tweets stick around, because they're located on Twitter's servers. You can delete them, but no one does. And it's impossible to imagine requiring or asking someone to delete tweets as a routine matter. In this they differ from emails, which can be subject to a deletion policy, if there is one as part of an organization's e-discovery policy.