Sometimes people want a place where they’ll be free from identity-based insult. But sometimes people want a place where they can talk freely about ideas without having to worry about being declared offensive. It’s legitimate to hope for both kinds of place. Loosely speaking, both can be dubbed “safe spaces.” But immediately a difficulty arises: the two kinds of safe space are often (not always) at cross-purposes.

Let’s call the first kind of place a “safe-being space.” This is where one can exist without risk of feeling demeaned. The second kind can be called a “safe-talking space,” where one can express ideas without recrimination.

These notions are easily mixed up, because the context that originally brought the phrase “safe space” into popular discourse is the psychotherapist’s room, which is a safe space in both senses. A good therapist allows you to articulate thoughts you would hesitate to say to your best friend—from fears to fantasies—and this safe-talking space makes you confident that words won’t ruin the therapeutic relationship. At the same time, the therapist’s commitment to withholding judgment, advice, and personal moral views let’s you feel safe in your own skin color, sexual orientation, religious faith, etc. Therapy is also safe-being space.

But there is no guarantee that this happy coincidence of kinds of safe space can always be achieved—or should be. An example topic illustrates my point.

When is it acceptable to articulate the view that same-sex parenting might cause social or psychological difficulties for children (in comparison with those that arise for children of straight couples)? That’s not my view. But when may it be expressed? Certainly, a child of a same-sex couple shouldn’t have to hear such a suggestion on a regular basis. Also, the view in question probably isn’t true; in fact, children of same-sex parents tend to do better. From these considerations, we might leap to the conclusion that it’s never okay to articulate such a view.

But that conclusion is hasty. Even if the view in question is false, as I think, it’s worth investigation through empirical social and psychological research. Say we’re researchers designing a longitudinal study on the impact on children of different kinds of parental situation, and say a member of our research group grew up with same-sex parents. Would it be wrong in this context to discuss possible downstream difficulties that might emerge from such a family set-up? Probably not.

The point is that the research group context is one in which one should feel safe floating ideas without fear of being chastised for having said something offensive. Especially if negative views about the psychological effects of same-sex parents are wrong, one needs a safe-talking space to be able to hash them out well enough to put them to the test. That’s why the research group should be a safe-talking space.

Many other topics dissociate the two kinds of safe space. It’s important to look this fact squarely in the face. Difficult questions play on people’s minds, whether they admit it or not. “Is the Koran misogynistic, or is it Islamophobic even to suggest this?” “Is affirmative action unfair, or is it racist even to suggest this?” “Should there be men’s or women’s support groups that allow only cismen or ciswomen, or is that very idea transphobic?” The overarching question is: “What should be acceptable to articulate regarding such topics?” And I think this overarching question can’t be addressed well without the distinction between safe-being spaces and safe-talking spaces.

This present discussion might seem to resemble recent controversy over “safe spaces” versus “free speech” on college campuses, in particular at Missouri, Yale, and Claremont McKenna. But from my perspective, that entire debate is ill formed. Free speech norms are legal restrictions on government’s ability to censor or sanction the speech of private citizens or the press generally. Those norms are at a different level from that of the university, at which more structured speech norms also can and do make perfect sense, as this satire clearly demonstrates. Simply put, it is a confusion to think that the First Amendment rules out having speech norms in, say, a social club that rule out offensive speech.

So instead of debating “free speech” versus “safe spaces”—a debate that confuses levels of speech norms—more local institutions should debate which kind of safe space, in any given context, is more important to have, if it’s not feasible to have both.

It’s no wonder that tensions emerged at universities, some portions of which should be safe-being spaces and some portions of which should be safe-talking spaces. Intuitively, dorms, community centers, and support groups should be safe-being spaces. Labs, classrooms, and reading groups should be safe-talking spaces. But boundaries are easily blurred. And it’s especially no wonder that tensions emerged at Yale, where the residential college system deliberately blurs the boundary between structured learning environments and dormitory life.

So when should institutions design conventions that set up safe-being spaces, as opposed to safe-talking spaces, or vice versa? I think it’s best to be mostly pragmatic about this, and the main point of this blog has been to help people in decision-making situations avoid conceptual confusion about what might be meant by “safe space,” when the issue arises. But some guidance is in order.

Roughly: living contexts should be safe-being places; learning and research contexts should be safe-talking spaces. Exceptions to those generalizations will be common. And some contexts, like classrooms, may do well to alternate between the two. Finally, sometimes both notions of safe space may be inapplicable. Selective exposure to adversity may make a growing person safer in the long run. After all, even the safest space of all—the therapist’s room—is meant ultimately to prepare a person for an unsafe outside world.