Twitter suspensions and their aftermath

Every so often my Twitter suspension comes up on social media, with various detractors of mine voicing different theories about why I was suspended from Twitter. 

• FC/RPM/S2C proponents seem to think I was suspended for “bullying” remarks and/or “violent threats” against autistic individuals and/or autistic advocates.
• Structured Word Inquiry proponents prefer to think I was suspended because I called SWI a “cult” run by a shadowy man in France with no formal linguistic credentials.
• Progressive math proponents may think I was suspended for harassing them and finding fault with “social justice math.”

As for Twitter itself, once it suspends an account, it feeds such sundry impressions by one’s sundry detractors with canned messages about the account’s suspension that are automatically sent to anyone who “reported” the account for any reason.

People, naturally, report Twitter accounts for all sorts of reasons, many of which have nothing to do with Twitter’s rules. Some people, for example, prefer to eliminate their critics rather than debate them.

Unless you’re sufficiently famous, however (cf. the former president), Twitter’s suspension decisions aren’t made by human beings mulling over tweets, but by AI bots programmed to look for certain key words and phrases (“kill”, “smash”, “vaccine injury”). That’s because human moderation is costly, and because the number of human moderators needed to review every possibly bullying/threatening/ dangerously misinformative tweet is astronomical. (As of early 2019, Twitter had around 330 million active users). Key word-based moderation, of course, is about as linguistically crude as it gets, and the result is that many suspensions are senseless.

Twitter tattlers, of course, prefer to think that Twitter suspended their opponents for sensible reasons–i.e., because they reported them. But there aren’t enough people on Twitter’s staff for human review of more than a tiny fraction of what’s reported. And if Twitter’s bots were to automatically suspend everyone who gets reported by someone else, Twitter would eventually amount to little more than cat videos. 

None of these considerations–assuming they even occur to them–stop certain Twitter tattlers and their allies from proclaiming, without evidence, that their detractors have made bullying or threatening remarks on Twitter or elsewhere. And though one might challenge them, as I have, to find a single bullying or threatening remark in anything one has ever written anywhere, the sort of people who prefer to resolve disagreements by tattling, blocking, and suspending may not be the sort of people who think that accusations (or autism interventions, or reading instruction, or math curricula) should be supported with actual evidence.

For further reading:

Twitter’s Other Free Speech Problem.