+<p><span class="dialogue-character-label">Alice</span>: ...</p>
+</div>
+
+<p class="flower-break">⁕ ⁕ ⁕</p>
+
+<div class="dialogue">
+<p><span class="dialogue-character-label">Alice</span>: Have you seen <a href="http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0016885">Dhejne <em>et al.</em>'s long-term followup study of transsexuals in Sweeden</a>? In Tables S1 and S2, the authors report that trans women commited violent crimes at far higher rates than cis women, with an adjusted-for-immigrant-and-psychiatric-status hazard ratio of 18.1—but only slightly lower rates than cis men, against whom the adjusted hazard ratio was 0.8.</p>
+
+<p><span class="dialogue-character-label">Bob</span>: Yes, how terrible that we still live in such a transphobic Society that those poor marginalized trans women are disproportionately driven to violent crime!</p>
+
+<p><span class="dialogue-character-label">Alice</span>: That's one theory. Can you think of any <em>other</em> possible interpretations of the data?</p>
+
+<p><span class="dialogue-character-label">Bob</span>: No.</p>
+
+<p><span class="dialogue-character-label">Alice</span>: Like, what do you make of the observation that the trans women's violent crime rate was not just higher than cis women's, but also strikingly close to that of cis <em>men</em>? Can you think of any reason—any reason at all—why that <em>might not be a coincidence</em>?</p>
+
+<p><span class="dialogue-character-label">Bob</span>: No, that has to be a coincidence. What could trans women and cis men possibly have in common?</p>
+
+<p><span class="dialogue-character-label">Alice</span>: ...</p>
+</div>
+
+<p class="flower-break">⁕ ⁕ ⁕</p>
+
+(Another dialogue about reproduction belongs in this collection, but was deemed too obvious and has been cut for space.)
+
+<p class="flower-break">⁕ ⁕ ⁕</p>
+
+If it's socially unacceptable for people who want to talk about sex to say "That's not what I meant by _woman_ in this context _and you know it_", then people who would prefer not to acknowledge sex will always get the last word.
+
+To this it might be objected that trans activists are merely advocating for greater precision, rather than trying to make it socially unacceptable to think about biological sex: after all, you can just say "cis women" (which excludes trans women, trans men, and natal-female nonbinary people) or "assigned female at birth" (which excludes trans women, but includes trans men and natal-female nonbinary people and presumably [David Reimer](https://en.wikipedia.org/wiki/David_Reimer)) or "uterus-havers" (which excludes trans women and natal females who have had a hystorectomy) if that's what you _really mean_.
+
+And it's true; [you _can_](/2016/Nov/editorial-process/). But I think this is underestimating the usefulness of having simple, [_short_](https://www.lesserwrong.com/posts/soQX8yXLbKy7cFvy8/entropy-and-short-codes) descriptions for the categories that do the most predictive work on typical cases.
+
+Kind or not, morally justified or not, voluntary or not, sexual dimorphism is _actually a real thing_. Studying the pages of _Gray's Anatomy_—or [_Wikipedia_](https://en.wikipedia.org/wiki/Sex_differences_in_human_physiology) if you're on a budget—you can absorb all sorts of detailed, _specific_ knowledge of the differences between female and male humans, from the obvious (sex organs, vocal pitch, height, muscle mass, body hair) to the less-obvious-but-well-known (chromosomes, hormones, pelvis shape) to the comparatively obscure (blood pressure! lymphocyte concentrations! gray-matter-to-white-matter ratios in the brain!).
+
+If—[like me](/2017/Dec/theres-a-land-that-i-see-or-the-spirit-of-intervention/)—you're the kind of person who is not necessarily _happy_ about sexual dimorphism, you can always deliberately define your categories in order to minimize it: if there's a large sex difference in some observable measurement, just say you _don't care_ about predicting that particular measurement.
+
+And as long as you're _already_ optimizing your _categories_ for the purpose of assuaging your fear of sexual dimorphism (which is not identical to the purpose of maximizing predictive accuracy and must therefore trade off against it in at least some possible worlds), why not optimize your _hypotheses_ along the same criterion? If there _is_ a large sex difference along some dimension that you have to admit that you care about, you can just say it's due to socialization! Science is difficult enough that no one will ever be able to _prove_ you wrong.
+
+But if you are furthermore an aspiring rationalist with [something to protect](https://www.lesswrong.com/posts/SGR4GxFK7KmW7ckCB/something-to-protect)—if you perceive a duty to see the world as it really is, even when it hurts—_especially_ when it hurts—