I want to come out in favor of safe spaces. Not everywhere, or for all purposes, but in general I think there should be a strong presumption in favor of limited and exclusive safe spaces for any group that wants to establish one.
Of course, the group would need to have the ability and legal right to use the space and exclude “others,” but often this would involve nothing more than the ability to reserve a public space that has a door that shuts.
My main claim is that “safe spaces” for students must include ideologically selected groups, just as it includes social groups such as fraternities and sororities. Freedom of association is the key to understanding academic freedom. And academic freedom, not freedom of speech, is the bedrock principle of universities and colleges.

Snowflakes!!!

It has become common for “my” side to decry safe spaces, sometimes because we want to mock the inability of “snowflakes” to deal with even mild dissent and sometimes for more principled “freedom of speech” reasons. The left-liberal commentator Van Jones used a combination of these two arguments, but his view is closer to the “snowflakes” critique.

Academic freedom, not freedom of speech, is the bedrock principle of universities and colleges.”]
I have elsewhere claimed that students on the Left are being denied an education, for just this reason: being insulated from disagreement means that they have only learned one-move chess openings. They don’t know how to respond to countermoves. They are rewarded for having memorized politically correct catechisms, having never actually heard arguments against, or for that matter for, their position.
Robert Zimmer, President of the University of Chicago, had this to say in defense of a “letter to first-year” students that had said safe spaces were a bad idea:

Universities cannot be viewed as a sanctuary for comfort but rather as a crucible for confronting ideas and thereby learning to make informed judgments in complex environments. Having one’s assumptions challenged and experiencing the discomfort that sometimes accompanies this process are intrinsic parts of an excellent education. Only then will students develop the skills necessary to build their own futures and contribute to society.”]
While I’m sympathetic to this idea that universities should not be, and by stated purpose cannot be, “safe spaces,” we have to be careful. There is a big difference between claiming universities should be safe spaces — that’s absurd — and claiming that universities should not promote internal safe spaces — because safe spaces are really just another name for freedom of association.
This argument has been developed in detail by Jacob Levy, and then discussed eloquently by Megan McCardle.

Universities should not be, and by stated purpose cannot be, ‘safe spaces.'”]
One way to capture the problem might go something like this: Universities should be safe spaces for a critical discussion, but also advocacy, of safe spaces. That is, people should be able to advocate for safe spaces without being told that such views are not allowed as part of the public discourse of universities. In Levy’s view,

[F]reedom of speech is not a value of universities.… I’m not making a claim about the details of the first amendment, I’m making a claim about how universities ought to be able to govern themselves.
Under general [constitutional] principles of freedom of speech … you are allowed to lie…. For example, you are allowed to tell the following particular kind of lie: you’re allowed to publish a book that has your name on the cover that you did not write. You have a contractual arrangement known as ghostwriting where you pay the writer, the writer writes the book, you buy the writer’s services.… All of this is perfectly legal. It’s a lie! [But] there’s nothing objectionable about it from the perspective of the freedom of speech.
On a university campus, if you submit written work in one of your classes that you paid someone else to write for you and you put your name on it, you get expelled. On a university campus, if I publish a piece of research that I didn’t write and I paid someone else to write and I put my name on it, I get fired. And appropriately so.… Academic freedom is the core meaning, the core institutional life, of freedom of debate and freedom of inquiry in the university setting. And one thing that’s notable about academic freedom as I’ve just labeled it is that it creates safe spaces [for specialized discussion, with internally shared assumptions about method and evidence].… We get together and argue about the new high level stuff that we are trying to understand. We ask each other hard philosophy questions, hard political science questions, hard chemistry questions, but they’re hard as understood within the context of that intellectual community.”]
Parsing this seems to imply a weak standard of protection for freedom of speech, but then Levy’s point is that freedom of speech is not the only value universities must take account of. In fact, in some ways freedom of speech is not even really a core value of universities.
Academic freedom is the core value of universities, and that is frankly quite a different thing. A core value of academic freedom is to foster freedom of association, with all the norms of internal self-governance that freedom of association implies.
If a group cannot exclude members, or cannot expel members who violate its rules, then that is a violation of freedom of association. That would be true if a department was prevented from ostracizing a faculty member guilty of research malpractice, if a club was prevented from requiring restrictions on viewpoint diversity for its members, or if a fraternity or sorority was forced to accept all applicants even if the newcomers shared none of the values of the group.

Locally Safe, Globally Transgressive

Levy’s conclusion is one that many people might find troubling, but which I think represents an important contribution to the debate over safe spaces. The conclusion has to do with the nature of debate, and the role of universities in fostering debate. While it is true that it is necessary for all viewpoints, and all challenges, to be “allowed” at a university, it is not true that such transgressive expressive acts are always allowed, and always protected.
Academic freedom places restrictions on the kind of arguments, and counterarguments, and truth claims, that have standing, within a particular group. Academic groups and student groups must be allowed, and encouraged, to carve off particular subjects or activities to be considered deeply, rather than always being obliged to start over at the broadest possible level.
In fact, academic freedom must encompass both the right to constitute a narrowly focused group and the right to challenge any conclusion that is part of the general orthodoxy. This is not a contradiction, but the implication of the freedom to conduct research and foster learning.
Levy concludes on a hopeful note, leaning in against the dominant “snowflakes repressing speech” narrative in much of the media. As Levy puts it,

Universities as a complex association are deeply linked to their enterprise. Universities are not purposeless civil associations, they are devoted to structured learning, research and inquiry of particular kinds, organized around serious intellectual communities to create questions that channel debate in productive ways, and then have moments of debate across. And for the most part, not perfectly, not every time, but for the most part, universities do a much better job of doing that then they are these days being given credit for.”]