The Facebook ad featuring a Sikh artist got a lot of really unpleasant comments from people using their real identities.
I don't think anyone suggests the Youtube comments were any better after real-name policy.
What I think it is, if I may hypothesize, is merely that there is fundamentally no incentive to NOT be however you want on the internet. 9 times out of 10 nothing you say will be traced back to you, or will reflect on any aspect of your day to day life. I consider the internet a window into the uninhibited mindset of the average person, and that although we may find it distasteful, to censor it would be paramount to burying our heads in the sand about our own nature.
What's worrying is that rather than find avenues to deal with it in reasonable or moderated fashion, the narratives often seem to shift to the most extreme and often dogmatic-seeming angles, more as talking points than measured response.
I do see cases of people behaving hatefully in a way that is connected to their real identity sometimes having consequences in the real world, for instance this case recently in Toronto: http://www.cbc.ca/news/business/firing-of-shawn-simoes-for-o...
Of course that case is controversial, but social consequences are one thing that can help protect against this sort of behaviour (note that this is not really a free speech issue in that he was not put in jail).
So what are social consequences we can levy online, if we want to support psudonymity and stay disconnected from our meatspace identities? One such consequence is having your hateful platform (the FPH subreddit in the Reddit case) closed. I am fine with this approach.
It is a slippery slope when one starts enacting vengeance as a response to speech actions deemed unacceptable, the canonical questions being "who draws the line" and "what happens when the line moves under coercion from malicious entities". I aspire towards, as I said, a more measured approach.
Let me make my view perfectly clear; there is a level of harassment that becomes _material_ and is no longer simply a speech issue, but given your citing of FPH, I don't believe that's what is necessarily being discussed. (I also think that's a difficult instance to discuss, due to the frank level of "politicking" that's been involved in that whole situation, although that (the building of a self serving movement either for or against a social justice issue) is another potentially worrying bit of fallout from a "warlike" mindset and response (for lack of a better word). Aggressive action only polarizes further; and while I agree with your core point that social pressure is key, I think it must be well crafted.)
If you already think there will be no ramifications of friends/people knowing your opinion, then a pseudonym doesn't give you much incremental benefit.
If, however, the potential ramifications are great, a pseudonym does a lot to distance the ramifications from you individually.
Before online communities, most people interacted in social circles small enough that their identity actually meant something to the other participants.
But on widely circulated/public posts on facebook, the audience is so large, that "John Doe from Indiana" might as well be anonymous.