It's not the name calling; it's the messaging. One person calling you a name is no big deal. What if it's lots of people said it and the particular words are intended to go after your vulnerabilities or insecurities? Hear it enough, from the right people and now you're doing some damage.
And this opens up a very important subject which legislators and litigators in many jurisdictions (including E&W) are currently - and fairly frantically - trying to figure out.
After all, when the effect on a victim is of the "death by a thousand cuts" variety, who can the state reasonably bring to justice? Social media has, for pretty much the first time ever in human history, allowed for bullying and insult towards private individuals to go viral incredibly quickly. If Person A were to post on twitter something like "Mary Smith is a frigid bitch who just leads guys on", this on its own would be highly unlikely to constitute a potential criminal offence. But if, within 24 hours, thousands of others - including a significant proportion of the victim's own community - had either retweeted that post or quote-tweeted it with additions such as "Haha that's so funny! Wouldn't want to be her right now lol"..... suddenly the effect of that one initial tweet has become super-magnified, super-quickly.
And it would not be at all unreasonable to suppose that the victim, in this scenario, might now feel ostracised and humiliated to such an extent that the effect might now take on the mantle of serious psychological harm (not to mention a more extreme effect of inducing suicidal thoughts).
In a fair society, we always seek (or should seek) to protect private individuals from this level of potential harm. We normally do so by regulating and/or legislating the behaviour of those who are inflicting that level of harm, usually within the remit of the criminal justice system.
But how on earth do (or even can) we do this in instances such as the one I described as an example? Much as politicians would want to defer to the old "it's entirely the responsibility of the social media companies themselves, as publishers, to police their own users' content" argument, that's simply unworkable. There are no computer algorithms that can automate the search for this sort of thing. It can't even (currently, at least) be approached through artificial intelligence means. And it should be clear to everyone that it's not possible for human eyes to perform constant proactive policing activity. Even if/when a victim reports posts to the site administrators, this can never be good enough - because a) once the post is reported, much of the damage has already been done in any case, and/or b) the time between report and remedial action will almost always be too slow to prevent significant additional damage from having been done (unless each SM operator were to employ literally tens of thousands more people globally).
So it's a real dilemma at the moment. Firstly, how do we take reasonable steps to minimise (we'll never eliminate) the number and severity of these sorts of SM "pile ons"? And secondly, who can we reasonably call to account if and when incidents of this nature take place? At the moment, it's proving hard to answer either question effectively.