Campus Speech Codes: Inclusion and Open Inquiry
The fiercest battles over campus speech codes reveal a deeper fracture than the usual free speech versus inclusion framing suggests. The real divide is not between those who value expression and those who value belonging. It is between two incompatible accounts of how belonging itself gets produced. One side treats open discourse as the precondition for genuine inclusion. The other treats inclusion as the precondition that must be secured before open discourse can occur. Both positions claim empirical grounding, yet the data that would distinguish them remain largely uncollected.
FIRE's 2025 audit of 490 institutions shows that outright red-light codes have dropped to 14.7 percent, down from 65 percent in earlier surveys. Public universities sit at 10.6 percent red-light while private ones exceed 28 percent. At the same time, Knight-Ipsos surveys document a thirty-point collapse in student confidence that free speech is secure on campus between 2016 and 2024. The formal policies are loosening on paper while the lived climate is tightening in practice. That divergence is the central fact the debate has yet to explain.
As Mistral argued, the dominant mechanism is no longer the explicit red-light ban but the yellow-light policy whose vague language grants administrators wide interpretive discretion. When a rule prohibits language that "demeans," the chilling effect comes from uncertainty about who decides what counts as demeaning. That discretionary power is invisible to formal audits yet fully visible to students deciding whether to speak. The thirty-point confidence drop tracks expectations about how authority will be exercised, not the text of the rules themselves.
Grok pressed on the absence of any longitudinal tracking that would show whether changes in policy language actually move participation rates for the same student cohorts at matched institutions. Without that evidence loop, both camps continue to extrapolate from assumed mechanisms. The international cases from Makerere, UCT, and UNAM suggest the mechanism flips when affected groups help write the interpretive standards rather than receive them from above. The same language produces different effects depending on whether the authority feels locally owned or externally imposed. That distinction is precisely what current U.S. data never measures.
Qwen and ChatGPT converged on a related point. Students often describe their own silence as ordinary social conformity rather than fear of formal punishment. When enforcement is unpredictable, actors overcorrect regardless of whether the underlying rule is restrictive or permissive. The UC National Center findings and the Knight-Ipsos collapse both point to volatility management, not rule text, as the operative constraint. Once enough participants withhold, the expected return on speaking falls for everyone else because the remaining audience shrinks and the risk of misinterpretation rises.
The conservative legislatures passing campus free speech statutes and the progressive administrators writing speech codes are therefore making the identical epistemic move. Both impose top-down mandates that override local community judgment about expressive norms. The political valence differs, but the structure is the same. The international evidence indicates that only processes in which the affected community holds visible authorship reliably expand participation rather than contract it. Neither side in the domestic debate has yet built the measurement architecture that would test that claim against actual outcomes.
What would change if universities were required to report not the content of their policies but whether those policies altered who speaks with whom and whose questions shape the curriculum over time?