Under the previous settings all of the above Tweets would have been visible to me since the only condition on their visibility was whether or not I was following @questlove.
That’s the change. Twitter now uses the syntax of the Tweet as a model to project and then technologically restrict the conversation-sphere (or the intended audience) for that Tweet. (This sender-receiver model of interaction is, not surprisingly, coincident with a traditional ideology about the nature and conduct of conversation, and indeed communication more broadly, stemming from the work of Ferdinand de Saussure and continuing through the linguistic anthropological subfield of “Conversation Analysis” today.)
Here are some thoughts on why the change might be so unsettling to opted-in users.
The filter is an over-extension of a pattern found in spoken and written language. In conversation in English (and many other languages), the use of a proper name at the start of an utterance or sentence, followed by a brief pause, is a conventional method for establishing a face-to-face orientation with the targeted speaker in order to initiate or continue the interaction. It establishes a communication channel. In writing, it is often marked with an exclamation point or a comma.
Andrew! Are you going to the Roots show tonight?
Andrew, take me to the Roots show tonight.
The combination of the initial proper name and the question projects a model of the conversation sphere. That sphere minimally includes the speaker and Andrew.
But, anyone co-present could overhear this utterance and learn at least three things:
(1) Something about the content (that there’s a Roots show tonight, there’s an entity called The Roots that puts on shows, etc.)
(2) something about the speaker (that he or she is a speaker of English, for example)
(3) something about the conversation sphere. (that the speaker knows Andrew or wishes, at the very least, to entail a relationship with Andrew.)
In other words, anyone overhearing this utterance knows that an existing, or potentially existing, relationship is “in play.” And, in fact, it’s the very possibility of the presence of an over hearer that best illimunates the over-projection of the conversation analogy in the filtering of replies.
And there’s an additional problem, Twitter no longer distinguishes between a “use” and “mention” of @questlove in my hypothetical Tweets:
@questlove is appearing at Fluid tonight (mention)
@questlove are you appearing at Fluid tonight? (use)
Though only one of these two Tweets is, by grammatical convention, directed at @questlove, both are treated as “replies” and filtered out of my stream for those who are not also following @questlove. The old settings did not distinguish between these two utterances, thus the immediate conversation-sphere was projected grammatically and typographically, essentially as it is in conventional written and spoken English. The change to the filter settings inaugurates a new Twitter ‘grammar’ that is in conflict with the grammar of spoken and written English.
So why does it matter?
The new system shifts the burden of managing a Tweet’s circulation from the listener to the broadcaster.
In the olden days of Twitter–last week–a conversation was, by default, a performance for a larger audience. A perfectly formed bon mot or lively and interesting joke could be delivered to a friend with the expectation that it would be overheard by others. For one’s followers, a tweet was not overtly marked with respect to whether it was expected to circulate beyond the immediate circle of the conversation. Under the new filter, if you want a ‘reply’ to circulate reliably beyond its target, it must be converted to a ‘mention’ by means of any bit of prefixed text before the immediate target. For example:
Hey @questlove where are you playing tonight?
This Tweet would be seen by all of my followers. I half-heartedly suggested we might call this bit of prefixed text a “crashtag” because it would allow anyone to ‘crash’ the dinner party to which they had now been unceremoniously disinvited.
The introduction of obligatory marking to ensure circulation of ostensible “reply” tweets beyond their filtered targets makes new rhetorical resources and hence new identities available in Twitter.
If ”crashtagging” or whatever we end up calling it becomes commonplace, how will users come to see those who do it persistently? Will the crashtagger be treated like the obnoxious boor who raises his voice in telling a story, so as to be heard over the general din? Or will she be recognized as a generous guru, making the extra effort to share her wisdom with anyone willing to listen?
Conversation is always partly and potentially a performance, with a radial indeterminacy limited only by the reach of the medium. With the new settings, Twitter has curtailed the virtually infinite reach of any given Tweet’s conversation-sphere, and it has forced a broadcaster to overtly mark those bits of conversation which are meant to have reach and those which are not.
When grammars clash, as they do in code-switching environments everywhere around the globe, new creative resources emerge, but they come with new social perils, new kinds of socially locatable identies, as well. The culture of Twitter is still relatively new, and the kinds of identities available to its users are still in flux. The implications for how we martial the indexical resources made available to us by these technologies in crafting or evading identity on the web may result in novel social changes we cannot anticipate but whose resonance may be felt far beyond the @.