My gut feeling is that it has something to do with Arc. Some people
tends to think it's just a regular Lisp with shorter keywords and that
there's nothing really innovative in it. The problem is that pg kept
the code private for too long while advocating publicly the power of
the language and criticizing Common Lisp.
So when he finally released it, some lispers were disappointed because
they've been waiting for something really disruptive. And it
backfired as most of the time, the harshest critics comes from
your disappointed sympathizers.
That being said, I don't hang that much on IRC so I have no real
back-story supporting this.
My theory is that because most of those people are very intelligent yet utterly ineffectual, they act as a kind of support group for one another. They view Lisp as a monastic order where in order to keep the tradition alive, one is forced by the circumstances to give up all his worldly ambitions. PG didn't buy it, and succeeded. So to keep the ideology consistent, you need to introduce the assumption that PG is not a "true" Lisper.
...In Graham's prose, like his code, every word counts...
Fascinating to read Peter Norvig describe ten years ago what most people tend to say about pg's essays and general writing style.