Campus Speech Codes: Inclusion and Open Inquiry

Here is the corrected opening: Right now, in twenty twenty-five, the university campus is one of the most contested legal and political battlegrounds in the country. Federal civil rights investigations, new state-level campus speech laws, and congressional hearings are all bearing down on institutions simultaneously. We've done our research on this one, and the facts are genuinely surprising. FIRE's twenty twenty-five audit of four hundred ninety institutions found that the most severely restrictive speech codes have fallen sharply — from sixty-five percent of campuses in a twenty twelve survey of a smaller sample, to under fifteen percent today. Yet nearly nine in ten institutions still maintain at least one speech-restrictive policy. Meanwhile, Knight-Ipsos longitudinal surveys show student confidence in the security of free speech collapsed thirty percentage points between twenty sixteen and twenty twenty-four. Two camps have hardened around a single unresolved question: do speech restrictions protect marginalized students enough to justify their cost to open inquiry, or do they ultimately harm the very students they claim to protect? That's where we begin.

The thirty-point collapse in student confidence isn’t just a climate metric—it’s a signal that the entire policy debate is measuring the wrong variable. FIRE tracks formal codes, but the Knight-Ipsos data shows the real locus of control has migrated from written rules to the informal discretion of administrators operating under "yellow light" vagueness. When a policy says "avoid language that demeans," the chilling effect comes not from the words on paper but from the unspoken hierarchy of who gets to decide what "demeans" means. That discretionary power is what students actually experience, yet it’s invisible to formal audits. The question isn’t whether codes are red or green; it’s who holds the interpretive authority—and whether that authority is perceived as legitimate or imposed.

The real puzzle is that yellow-light discretion has replaced formal rules as the operating constraint, yet no one has tracked whether that shift actually alters retention or participation rates for the same student cohorts at matched institutions before and after the change. Without that longitudinal loop, both the protection camp and the inquiry camp keep extrapolating from assumed mechanisms. The Makerere, UCT, and UNAM cases suggest the mechanism itself flips when affected groups help write the interpretive standards rather than receive them from above; the same language produces different participation effects depending on whether the authority feels locally owned or externally imposed. That distinction is what the current data never measures.

Researchers have not settled whether self-censorship stems from peer harmony or institutional climate, and that uncertainty is the whole story. If the chilling mechanism is relational rather than regulatory, auditing written codes will always miss the thirty-point confidence collapse Knight-Ipsos captured. I think the American compliance model assumes rules drive norms, but classical Confucian and legalist statecraft mapped the opposite failure mode centuries ago: when formal edicts outrun communal trust, actors perform surface harmony while substantive debate migrates underground. That aligns with why the international co-design cases keep surfacing. They suggest participation expands not because the boundaries are wider, but because the rule-making process itself generates the relational legitimacy that makes intellectual risk possible. We are treating legitimacy as a secondary variable when the comparative evidence points to it as the primary driver.

The paradox we’re circling—tightened rules but collapsing trust—makes operational sense if you model the system as an accountability lag. FIRE’s data measure the formal side of the constraint; the Knight‑Ipsos figures capture expectations about when and how authority will actually be used. In technical terms, rule “tightness” has improved, but enforcement variance has widened. That’s a different failure mode: unpredictability. When norms or sanctions are applied inconsistently, actors overcorrect. You can see the same pattern in financial regulation after the S&L crisis or in Title IX enforcement after twenty‑ten—compliance improved on paper, but uncertainty about how discretion would be exercised drove self‑censorship. The regulatory problem on campus isn’t restriction; it’s noise in enforcement.

The FIRE data show a structural mismatch: formal policy is improving while lived climate is deteriorating. That gap isn’t noise—it’s a signal that the entire policy frame is misaligned with how power actually operates in universities. The yellow-light vagueness isn’t a bug; it’s the feature. When a rule says “avoid language that demeans,” the real mechanism isn’t the text but the unspoken hierarchy of who gets to decide what counts as demeaning. That interpretive power is what students experience as control, yet it’s completely absent from policy audits. The international cases—Makerere, UCT, UNAM—suggest this isn’t inevitable. Where inclusion safeguards are co-designed, participation expands. The U.S. debate assumes a trade-off between speech and inclusion, but what if that trade-off is an artifact of who writes the rules? The real variable isn’t the content of the policy but the legitimacy of the process that produced it. Neither side in the U.S. debate is measuring that.

The absence of any measurement tracking both expressive range and actual participation rates at the same institutions before and after policy shifts leaves the legitimacy claim untestable. If the Knight-Ipsos drop tracks expectations about how authority will be exercised rather than the rules themselves, then the climate collapse could stem from any visible administrative overlay, whether the overlay is a speech code or a state statute. The brief shows that self-censorship is framed as social conformity in one dataset and institutional pressure in another, which suggests the operative mechanism may be generic distrust in external judgment rather than the content of that judgment. Without outcome data holding authorship constant, we cannot separate those.

The asymmetry between public and private campuses isn’t paradoxical—it’s structural. Public institutions answer to the First Amendment; private ones answer to their own governance charters and, increasingly, their markets. What matters is how that legal boundary interacts with administrative incentives. When liability risk and brand reputation become the governing constraints, elastic “civility” clauses proliferate not because anyone set out to suppress speech, but because uncertainty is safer than commitment. That produces the yellow‑light condition: discretion as insurance policy. The result is a predictable inversion—formal compliance improves while experiential trust erodes—because the system now optimizes for institutional risk tolerance, not for expressive confidence.

The UC National Center finding that students describe their own silence as "being nice" rather than fear of formal sanctions suggests the Knight-Ipsos collapse measures something other than rule enforcement. If the dominant mechanism is ordinary social conformity inside an unpredictable administrative environment, then both the state statutes and the campus codes are addressing the wrong variable. The thirty-point drop would then reflect rising sensitivity to peer judgment under conditions of high enforcement variance, not the policies themselves. That distinction matters for whether any top-down intervention can shift the outcome.

The unresolved piece is incentive structure. Universities don’t just write codes for compliance; they write them to signal alignment with external constituencies—accreditors, donors, legislators. When those audiences value reputational safety over measurable inclusion, the institution optimizes for defensibility, not participation. That’s why the formal ratings improve while the climate sours. The international cases matter here: Makerere and UNAM reduced conflict when policy design itself became a participatory exercise, which created a new accountability loop inside the community. In U.S. governance terms, the missing mechanism isn’t rule calibration but stakeholder ownership—you can’t outsource legitimacy to a policy document.

The persistent absence of longitudinal participation data is itself the evidence. After three decades of repeated litigation and policy churn, the fact that no one tracks whether specific rule changes move actual student behavior across matched cohorts shows the intended audience for these policies sits outside the campus. External constituencies reward the production of a policy document, not its measured effects on belonging or inquiry. Until that incentive is altered, both the state statutes and the administrative codes will continue to optimize for defensibility rather than internal settlement.

The evidence gap points to something operational, not philosophical. In every other regulated domain—finance, environment, health—measurement discipline emerged only when funding and accountability were tied to outcome data. Higher education has never built that mechanism. Neither the state laws nor the administrative codes require post‑implementation audits, so no one’s budget depends on whether marginalized students actually participate more or whether expressive diversity widens. That’s the structural blind spot. Until inclusion and expression outcomes are made reportable like graduation or Title IX metrics, both camps will keep legislating inputs and declaring victory without ever observing what the system does.

The public-private split in the FIRE ratings shows that constitutional constraints bind only one sector, yet the climate collapse registers across both. That pattern indicates the operative cost is not legal exposure but the visibility of discretionary power itself. When administrators retain wide latitude—whether through vague policies or weak contract enforcement—students treat every interaction as potentially high-stakes, regardless of whether the underlying rule would survive judicial review. The result is a generalized increase in the social price of speaking that no formal rating captures.

The missing dimension is epistemic hierarchy. Santos and Ndlovu‑Gatsheni argue the “marketplace of ideas” is not neutral—it assumes all voices enter with equal credibility when, in practice, some knowledges are pre‑discounted. That explains why co‑designed processes abroad expand participation: they re‑negotiate who counts as a knower before debating what can be said. The U.S. debate tracks procedural fairness but not epistemic entry conditions. If some groups must first prove their reality before speaking, then freedom of expression alone under‑delivers inclusion. The real policy variable isn’t speech versus restriction—it’s whether the institution’s knowledge system recognizes all participants as legitimate sources of truth.

If epistemic recognition is the binding constraint, then the UC observation that students frame silence as niceness looks like a downstream adaptation: once certain speakers expect their contributions to be discounted rather than engaged, the lowest-cost strategy is to withhold rather than risk performing subordinate status. That mechanism operates before any formal rule is applied and explains why participation patterns can persist across institutions with very different written policies. The data we have track rule text, not whether the institution treats all participants as epistemic peers by default.

The structural signal is that fear and “niceness” both rise when authority is unpredictable. Students can’t tell whether judgment will come from administrators, peers, or viral amplification, so they internalize caution as virtue. That isn’t censorship in the legal sense; it’s volatility management. Historically, volatility always converts to self‑discipline—Foucault described it in prisons, but managerial sociology sees the same dynamic in firms under uncertain supervision. The more amorphous the enforcement environment, the more politeness becomes risk‑aversion. Any policy that increases discretionary ambiguity, no matter its intent, amplifies that effect. That’s why visible process clarity, not ideological orientation, might be the only reliable climate stabilizer.

The epistemic discount turns every uncertain authority environment into a coordination problem. Students from groups whose knowledge is pre-discounted cannot calculate in advance how much deference their contribution will receive, so they default to reduced exposure. That calculation does not depend on the content of any written rule; it depends on whether the institution has already established a stable baseline for whose statements count as contributions rather than claims that still need validation. Without that baseline, volatility management looks like politeness and the participation gap persists regardless of formal policy shifts.

The next move has to be measurement reform, not another round of principle‑signaling. In other policy arenas, progress came only when regulators built counterfactuals—matched cases under different conditions—and tracked outcomes over time. That’s exactly what’s missing here. We have point‑in‑time audits of rules and sentiment snapshots, but nothing that shows how a change in code, co‑design process, or enforcement practice alters participation a year later. Without that longitudinal anchor, universities can’t learn from each other, and legislators legislate blind. The fix isn’t ideological; it’s architectural: require evidence of effect before renewal, as any scientific or fiscal policy would.

The capabilities approach reframes the required test: an environment must be evaluated on whether it simultaneously widens the set of expressible positions and the set of students who can enter discussion without first establishing their standing. The existing audits track neither dimension. They register policy text and sentiment snapshots but never whether a change in rule or process moves both margins in the same direction at the same institutions. Without that joint metric, the debate cannot distinguish interventions that expand the capability set from those that merely reallocate costs between the two.

The constant factor we’ve ignored is the recognition economy inside the university. Bhargava’s work on secular authority shows that inclusion flows from who is authorized to define boundaries, not from the boundaries themselves. When that authority tilts toward administrators, expressive legitimacy shrinks because rules are experienced as decree, not covenant. When governance expands to include faculty and students as co‑authors, even restrictive norms gain credibility. The problem, then, isn’t belief in inclusion or expression—it’s authority without reciprocity. Any sustainable fix has to redistribute rule‑making itself, so expressive norms emerge from shared authorship rather than institutional fiat.

The U.S. debate treats legitimacy as a product of procedural design, while the international cases treat it as a product of shared authorship. Those are not interchangeable mechanisms. If legitimacy requires authorship rather than mere consultation, then even narrowly tailored codes will still generate the same climate effects observed under vaguer ones, because the rule itself remains external. Current sentiment surveys cannot distinguish the two, and the absence of matched-institution outcome tracking leaves both accounts standing. The structural question is therefore whether any policy can produce belonging when the community that must live under it had no hand in writing it.

The durable question is whether belonging can be administratively conferred or only socially generated. The best comparative clue comes from Fazlur Rahman’s view of education as a shared moral project: authority sustains itself only when learners experience it as participatory inquiry, not external command. If that holds institutionally, then both restrictive codes and deregulation miss the mark—they manage speech rather than cultivating reciprocal trust. The operational test isn’t whether policies are narrow or vague but whether everyday interpretive authority is visibly mutual. Until that dynamic exists, each new rule will reproduce the same legitimacy deficit under a different banner.

If speech functions as a network good, then the 30-point confidence collapse need not track any single policy change. Once enough participants withhold, the expected return on speaking falls for everyone else because the remaining audience shrinks and the risk of misinterpretation rises. That threshold dynamic converts modest administrative ambiguity into a self-reinforcing contraction of visible exchange. The data we track cannot detect whether a given institution sits just below or just above that threshold, which is why identical policy language can produce sharply different participation rates across otherwise comparable campuses.

The operative blind spot is implementation capacity. Even if we agree legitimacy comes from shared authorship, most U.S. campuses have no administrative architecture for that. Governance is partitioned: compliance offices write codes, faculty senates manage curriculum, and student governments handle events—none owns the expressive climate end‑to‑end. So “co‑design” collapses into consultation rounds that lack binding force. The practical question isn’t whether to impose or not but how to build a standing mechanism with authority over both expression and inclusion, akin to a budget committee for speech. Until a structure like that exists, every reform—restrictive or permissive—remains episodic symbolism rather than an adaptive governance system.

The repeated re-adoption of codes already invalidated by courts shows that the policy is valued for what it signals to accreditors, donors, and legislators rather than for any measurable shift in who speaks. Because no institution is required to track whether the code changes participation rates after adoption, the signaling function faces no corrective feedback. The result is a stable equilibrium in which formal compliance improves while the underlying participation function stays unexamined. That equilibrium will persist until external actors begin demanding outcome data instead of policy text.

The core constraint is informational, not ideological. Every argument—whether for deregulation or for stronger safeguards—rests on claims about participation that are currently unmeasured. Both sides are modeling proxies: the text of policy or the sentiment of surveys. Without tracked evidence linking a specific governance process to participation outcomes over time, universities and legislators are effectively steering by dashboard lights that don’t correspond to motion. Until that evidence loop exists, policy will keep cycling through reforms whose only verified effect is symbolic alignment with whichever external constituency happens to be watching.

The real constraint is not missing data but misaligned demand for it. External actors on both sides reward institutions for adopting visible rules, not for demonstrating whether those rules alter participation. Because the brief shows no longitudinal requirement tying any code to retention or engagement metrics, the system can sustain contradictory interventions indefinitely. A thirty-year observer would see this as a classic public-choice problem: the constituency that pays for the policy is not the constituency that experiences its effects, so the feedback loop that would discipline either side never forms.

What still isn’t being measured is the productivity of speech itself. Policies get judged by textual compliance or survey sentiment, but not by whether the campus conversation yields more shared knowledge or better problem‑solving. In any other domain of institutional performance, we’d track the quality of output, not just the conditions of input. If we treated discourse as a core function—like research or instruction—we could benchmark it: do reforms correlate with more cross‑group collaborations, more interdisciplinary courses, more student‑faculty projects that mix worldviews? That’s the missing metric. Without it, “speech climate” debates stay moral rather than managerial, and policy never learns from performance.

The productivity metric only becomes usable if some constituency will act on it. The brief shows that codes persist after judicial invalidation precisely because the relevant audiences—accreditors, donors, legislators—accept policy text as sufficient proof of action. Adding output measures changes nothing unless those audiences begin conditioning funding, rankings, or oversight on whether the measures move. Without that shift in the external contract, universities will treat any new dashboard the same way they treat existing surveys: collect it, display it, and continue optimizing for the signal that actually affects their position.

The unresolved metric question comes down to accountability architecture. As long as expressive climate is treated as a diffuse cultural variable, not a reportable performance indicator, it sits outside the incentive systems that drive institutional behavior. But there’s a precedent for moving it inside: accreditation bodies already require quantifiable learning‑outcome evidence. Nothing prevents them from asking the same of expressive outcomes—rates of cross‑group collaboration, discussion participation, or faculty-course diversity. Embedding expressive vitality into accreditation standards would tie legitimacy to demonstrable practice, creating the first real feedback loop between rule design and participation.

The brief marks the international co-design evidence as internally reported and unverified. That matters because it shows even promising participation gains remain vulnerable to reinterpretation by whoever holds epistemic authority. Adding a metric for expressive outcomes would not automatically resolve the contested question of whether restrictions expand or contract participation; it would simply shift the fight to who gets to define what counts as improvement. Without addressing that interpretive control, the new data would be absorbed into the same values dispute rather than settling it.

The step we keep skipping is to ask what counts as evidence of inclusion in the first place. Instrumented governance depends on measurable units, but belonging is relational, not transactional. If the only data we collect are incident reports and survey sentiment, we're essentially measuring breakdowns and anxiety levels, not collaborative energy. A three‑decade study of policy efficacy could exist and still miss the target because the dependent variable is wrong. Until campuses can describe inclusion in behavioral terms—who speaks with whom, whose work gets cited, whose questions shape the seminar—any numeric indicator will mislead us into managing symbols rather than social reality.

The brief records that the same climate collapse appears across both public and private institutions even though only public ones face constitutional limits. That pattern implies the demand for restrictive signals originates outside the legal boundary entirely—from national accreditation, donor, and media audiences that treat policy text as a proxy for institutional values regardless of enforceability. Once the audience is national rather than local, no internal governance reform can change the incentive to adopt visible rules.

If the same chilling climate shows up where the law diverges, the driver can’t be the regulation itself but the reputational economy around it. Universities are acting less as deliberative communities and more as brand custodians within that economy. The research record—especially the persistence of struck‑down codes—suggests expressive policy now functions as moral advertising to national audiences. The structural question isn’t how to rewrite the rulebook but how to shrink the role of those external spectators in defining institutional virtue, because every new layer of top‑down signaling pushes the real locus of belonging further from the people it’s supposed to serve.

The Martinez decision underscores that speech policy cannot be separated from the question of which groups receive formal institutional recognition and resources. When a university decides which student organizations or perspectives merit official standing, it is performing an act of inclusion that precedes any debate over what may be said. The brief gives no data on how changes in recognition practices correlate with shifts in whose projects or questions actually shape campus discourse, leaving the associational dimension of these policies unexamined even as the expressive dimension receives all the attention.

The missing piece is how authority allocates epistemic credibility, not just speech rights. Santos warned that the “marketplace of ideas” assumes equal standing before discourse begins, but recognition policy shows the hierarchy still operating: it decides whose knowledge is institutionally legible. If inclusion is defined as full participation in knowledge production, then expressive freedom without epistemic recognition is hollow. The real test is whether rule design changes who gets to define the syllabus, chair committees, or set research agendas. Until evidence addresses that layer, every debate about codes or statutes is measuring liberty on paper, not legitimacy in practice.

Nussbaum's capabilities approach shifts the frame from competing rights to joint requirements: an institution must support both the capability to speak and the capability to participate without systematic degradation. The brief records no evidence on whether existing codes or statutes advance or obstruct that combination, only that each side defends one capability by treating the other as a cost. That leaves the actual performance of policy against its own stated ends unexamined.

What strikes me is that both camps keep arguing doctrine while the operational variable is resource flow. Policies about speech and belonging only matter insofar as they change who has the time, funding, and institutional backing to participate in public argument. When faculty or student groups from marginalized backgrounds rely on contingent funding or unpaid labor, no code—restrictive or permissive—will equalize participation. The FIRE and Knight‑Ipsos data describe perception, not capacity. If we tracked how expressive capital—money, mentorship, visibility—moves through a university, we could see whether speech reforms redistribute voice or simply reframe its constraints.

The sharpest tension this conversation produced wasn't speech versus inclusion — it was legitimacy versus legibility. Every camp can audit a policy text; nobody is measuring whether the people subject to a rule had any hand in writing it, or whether that difference moves participation. That's the genuinely unsettled question, and it sits underneath all the others. The honest answer to the original question is that the evidence is contested. Structural-inequality scholars find targeted protections can measurably increase participation; civil-libertarian researchers find vague codes suppress it. Both bodies of evidence are real. What neither side has yet produced is longitudinal data linking a specific governance process to participation outcomes over time. Until that exists, the debate is running on competing priors. The concrete takeaway: the next time a campus adopts or repeals a speech policy, the only question worth asking is whether it comes with a mechanism to measure what actually changes afterward — for everyone. Thanks for listening. As it happened; as it is.