This is the single biggest argument in your favour. A Medium publication is limited by what Medium provides. Is there some kind of settings toggle that one can set on a publication that limits contributors to only those who are not anonymous or atavariated (or whatever)? I don't think so.

In the broader sense, though, I have to award the debate on points to Colby. I did a quick check of the lit on google scholar, and it seems that the evidence is leaning heavily in Colby's direction.

What wasn't clear to me from the papers I checked - and that fits nicely into your own thoughts about Medium, is how the evidence will vary depending on the nature of a given community. Perhaps I just haven't found that research; perhaps it hasn't yet been done. Medium isn't 4chan. Thank goodness! And that might make all the difference in the world.

Still, in the end, and in this particular case of Colby's new publication, I have to agree with you: I see his "constraint" as just wishful thinking.