Skip to content
AFTER CERTAINTY
Skip to chapter text

When Interpretation No Longer MattersPart II — What Replaces It

Chapter 3 — The Argument That Sorts the Room

About 20 mins

The Argument That Sorts the Room

Donald Trump and the First Alignment Cycle, 2015–2021

A forwarded post arrives in a family group chat late on a weeknight. No one asked for it. The caption is shorter than the article it claims to summarize, and half the people scrolling will never open the link. Someone has bolded a sentence. Someone else has added three laughing faces. A third person replies with a blue link that contradicts the first, then a fourth writes only: Really? The thread fills with counter-links, clipped quotes, screenshots of screenshots, an uncle’s late interjection about “what no one will say.” Aunt replies with a church emoji and a warning about the kids seeing this. A cousin answers that silence is complicity. Someone leaves the chat and returns twenty minutes later. By morning the original claim has been replaced by a quieter fact the room already suspected: who sides with whom.

The phone still shows the same words. What has changed is what the words are being used for. The question on the screen looks like an argument about evidence. The argument underway is about affiliation. People who once used the thread for recipes, flight arrivals, and birthday reminders discover that a single post can reorganize a table before anyone has finished reading. Disagreement does not gently clarify a proposition. It clarifies seating.

Watch what happens if someone tries to slow the room. A careful sentence—“Can we agree what the article actually says before we judge it?”—lands oddly. Caution is heard as softness or as coded betrayal, depending on who speaks it. Neutrality becomes a side. Humor that once relieved pressure is now scored as allegiance. A link meant to clarify is answered by a link meant to embarrass. The most patient person in the chat is not necessarily the most informed; they may simply be the one still treating the exchange as shared evaluation rather than as a loyalty probe. By the time that recognition arrives, recognition itself is late. The probe has already been answered in the pattern of replies.

What makes the moment hard to live inside is that everyone can still point to real disagreements worth having. Immigration, pandemic tradeoffs, judicial appointments, tax law, foreign policy, media power—none of these were inventable controversies. The family was not fighting about nothing. They were fighting with tools that no longer did the work they thought the tools did. Evidence still appeared. It appeared already wearing a jersey.

None of this requires that anyone in the chat be foolish. People argue carefully inside their own lanes. They cite dates, names, numbers. They remember earlier posts and demand consistency from others while granting latitude to those who already belong. What fails is not intelligence. What fails is the expectation that disagreement, pursued hard enough, will produce a shared corrective that travels across the group. Intense disagreement can clarify relationships with brutal efficiency while leaving claims almost exactly where they began.

That failure feels personal when it happens among relatives. It is also a public pattern with a dated spine. Between roughly June 2015 and January 2021, American national politics offered a large, open-system version of the same room: intense interpretation everywhere; fact-checking ecosystems and counter-ecosystems in continuous operation; rallies, cable news, and social feeds amplifying conflict—and a progressive loss of cross-group persuasive force. Claims kept circulating. Hearings were televised. Courts issued opinions. Scientists published papers. What changed was the social job of the claim. Meaning still moved. Increasingly, it moved as a badge.

Settle the frame carefully. Ordinary democratic life depends on a fragile achievement: that opposing interpretations can meet under rules durable enough for some reasons to travel. Not all reasons. Not instantly. Enough travel, though, that scandal can cost standing, that evidence can force a revision, that a person can change sides without experiencing the change as self-erasure. When that achievement thins, public speech does not become quiet. It becomes busier and less corrective. The family chat is the miniaturized form. The national argument of 2015–2021 is the scaled form.

This chapter stays inside that date band. Later years exist, and readers will bring them; the mechanism under study does not need them to be visible in order to be named. In an open information environment, interpretation can be repurposed into alignment—into loyalty tests, enemy identification, and belonging—without anyone sealing the room or forbidding thought. Disagreement can remain visible and even fierce while losing corrective force across groups.

Donald J. Trump announced a presidential campaign in June 2015 as a celebrity outsider already fluent in media cycles. He attacked immigration policy, trade arrangements, and elite consensus inside both major parties; he staged defiance as entertainment; and he treated media antagonism as proof of authenticity rather than as a simple reputational cost.1 Through the 2016 primaries and general election, and then through the presidency that followed (January 20, 2017–January 20, 2021), the public argument around him rarely went quiet. That noise is part of the case. Alignment-based authority in an open system does not need silence. It thrives on argument that sorts.

What follows is not a general indictment of a coalition, a party, or a nation. It is a structural reading of how claims operated in public: how controversy clarified belonging, how escalation often outran repair, and how core support could remain comparatively stable even while factual disputes multiplied. Motives are left aside. The object is function.

Throughout the campaign and the White House years, Trump’s public statements were repeatedly contested as exaggerated, inconsistent, or inaccurate. Rival campaigns, opposed members of Congress, career officials, scientists, and independent reporters offered corrections in real time. Fact-checking desks and mainstream news organizations built running tallies and annotated transcripts; partisan outlets built parallel apparatuses that contested the tallies, the checkers, and the premises of verification itself.2 None of that activity was hidden. A citizen who wanted contradiction could find it within minutes. A citizen who wanted rebuttal of the contradiction could find that too.

If public interpretation worked the way older civics rhetoric still implies, accumulating contradiction would have steadily redistributed support. Something like that still happens in quieter institutional settings, and it still happens for some people some of the time. Across these years, however, the redistribution often looked weaker than the volume of dispute predicted. Gallup’s tracking of Trump’s job approval from 2017 through early 2021 found unusual stability relative to prior presidencies: overall approval moved inside a relatively tight band even as news cycles swung, and the stability was organized by party. Across the term, Republicans averaged roughly nine in ten in approval; Democrats averaged near the single digits; the average partisan gap set a modern record.3 At no point, Gallup reported, did Republican approval fall below the high seventies—an endurance that outlasted controversies expected, under earlier norms, to open larger cracks in a governing coalition.4

The distinction matters. Core-coalition stability is not national conviction. Independents moved more than partisans. Overall popularity was often weak by historical standards even while Republican approval remained high.4 The mechanism is not that disagreement evaporated. It is that disagreement, for large overlapping audiences, increasingly clarified where one stood rather than what could still be jointly revised.

Rallies, social media, and partisan television intensified that clarification. Haberman’s reporting follows how a public style organized around repetition, grievance, transaction, and loyalty talk could capture attention and reward people who answered in kind.1 A rally is not an encyclopedia entry. It is a room that cheers recognition. A feed is not a seminar. It is a machine for distributing belonging cues at speed. Cable panels perform conflict as product. None of this invents polarization from nothing. Ezra Klein’s account of American party sorting describes a longer institutional story: as parties and social identities stacked into mega-identities, political conflict increasingly felt existential rather than transactional, and audiences arrived predisposed to treat opposing claims as out-group speech before evaluating them as propositions.5 The 2015–2021 period did not create that apparatus alone. It drove through it with unusual volume, making the sorting function of argument unusually hard to miss.

Inside aligned spaces, people kept interpreting with energy. Supporters argued about strategy, staffing, success, betrayal, and timing. Critics argued about norms, competence, corruption, legality, and harm. Both sides produced long documents. Courts heard cases. Legislatures staged hearings. Scientists published and revised papers. Journalists corrected the record and then argued about the correction. A person could spend an evening reading carefully and still wake up no closer to a shared evaluative public, because the careful reading had been received, on the other side of the cut, as another performance of membership. Thought did not stop. The public office of interpretation—the capacity of reasons to revise legitimacy across boundaries—shriveled even while interpretation itself proliferated.

Claims began to function less as propositions awaiting shared evaluation and more as badges of belonging. To circulate a claim was often to display a side. To refuse the claim was often to display the other. Correction arriving from across the aisle could then be heard not as information but as an identity move: an attempt to force a defection. Under those conditions, “winning” an argument does not look like persuading the other group. It looks like holding one’s own while forcing the other to reveal itself. Escalation becomes intelligible. Repair looks like humiliation.

Several episodes make the pattern difficult to romanticize away. They are pressure points, not a complete history.

Campaign and early presidency. From the June 2015 announcement through Election Day 2016, disputed statements arrived in a nearly continuous stream: crowd sizes, immigration claims, opponent allegations, trade numbers, conspiracy intimations. News organizations built explainers; late-night comedy converted contradiction into entertainment; primary rivals tried to weaponize inconsistency and often found that inconsistency did not land the way prior cycles taught them to expect.1 The surprise for many professional observers was not that voters disagreed about policy. It was that contradiction, widely available, did not redistribute allegiance the way debate-stage logic promised. A statement could be contested on Monday and recirculated as belonging cue on Tuesday without the contestation doing the expected damage inside the core audience.

After inauguration, early fights over inauguration attendance, travel restrictions, and personnel were fought as factual disputes and as affiliation dramas simultaneously. Haberman’s reporting emphasizes how loyalty language—who was “with” the president, who was a leaker, who was soft—organized White House politics alongside policy.1 Tax legislation, immigration enforcement, judicial confirmations, and trade negotiations all had ordinary policy stakes. They also arrived already nested inside sorting media. For many listeners the first successful decoding of a presidential claim was coalition placement, and only afterward—if at all—shared evaluation.

COVID-19. In 2020, a novel pathogen forced continuous public interpretation under genuine uncertainty: infection dynamics, treatment claims, mask guidance, school closures, hospital capacity, timelines for vaccines. Official guidance changed as evidence changed. Presidential messaging frequently outran, minimized, or disputed parts of that sequence—optimistic timelines, contested treatments, moments of risk downplay that public-health officials and many physicians rejected. Democratic leaders and major media criticized those messages as dangerous. Segments of the Republican coalition framed pandemic constraints as elite overreach.6 Within each camp, people debated details intensely. Across camps, shared update of legitimacy largely failed. Masks, lockdowns, and later vaccines became signals of belonging nearly as much as instruments of risk management.

School boards and grocery aisles became miniature versions of the national argument. A cloth mask could be read as care or as compliance theater; refusal as liberty or as indifference. Both readings could summon evidence. What thinned was a common staircase from local evidence to joint legitimacy. Harm was real and uneven. Under sorting pressure, scientific and political claims alike were frequently received first as whose team speaks, and only secondarily as candidates for joint revision. People who tried to hold two thoughts at once (serious disease; serious costs of intervention) found the middle of the room dissolving under them.

Impeachments. The first impeachment process (House impeachment vote December 18, 2019; Senate acquittal February 5, 2020) concerned Ukraine, foreign assistance, and allegations of abuse of power and obstruction of Congress. The second (House impeachment vote January 13, 2021; Senate acquittal February 13, 2021) concerned Trump’s conduct around the 2020 election and the January 6 Capitol attack.7 Both produced enormous piles of interpretation—depositions, transcripts, floor speeches, concurrent partisan narratives treating the same hearing as exoneration and damnation at once. Party-line outcomes were not mysterious given institutional incentives and the sorted electorate already in place. What the episodes also displayed was the conversion of contested institutional facts into affiliation tests. For many audiences, “What happened?” was quickly subordinated to “Whose side are you announcing by how you describe what happened?”

January 6, 2021. After Trump lost the November 3, 2020 election, he and allies amplified claims that the result was illegitimate. Those claims were contested in courts, state certifications, and extensive reporting, and were rejected in case after case by judges across partisan appointment backgrounds.8 On January 6, Congress met to count electoral votes. Trump spoke at a rally that morning. Later that day, rioters breached the Capitol, interrupting certification, with deaths and injuries among the consequences. Pew surveys in January 2021 found near-universal awareness of the riot alongside sharp partisan differences over presidential responsibility, severity, and institutional remedy; later waves showed those gaps hardening rather than closing.9 Within coalitions, people continued to argue—about intent, speech versus violence, fairness of elite institutions. Across coalitions, the argument operated less as shared repair than as boundary work: who still belongs with us after this.

Notice what these episodes share with the group chat that opened the chapter. Interpretation does not go silent. It intensifies. The missing piece is corrective travel: the movement of reasons from one affiliate world into another with enough force to revise standing. Escalation substitutes for repair because escalation strengthens the badge while repair would require provisional trust across a boundary the room no longer credits.

It is possible, and historically common, for large publics to argue bitterly and still share enough institutional trust that a hearing, an audit, or a court ruling can move people who did not like the outcome. That shared residual is what thinned here—not among every citizen equally, but enough that “the facts came out” ceased to be a reliable political sentence. Facts still came out. They came out into rooms already arranged by badge.

One further caution: core support can be stable for many reasons—shared policy priorities, demographic sorting, dislike of the available alternative, perceived media unfairness. Stability alone does not prove that audiences never updated beliefs. Polling from the period shows disagreement inside coalitions as well as between them.9 Sorting is not unanimity. It is the reorganization of argument so that update across the major cut becomes rare, while update and rivalry within each side continue.

Several false diagnoses crowd the doorway.

It is tempting to call the period “post-truth,” as if people stopped caring what is true. People cared intensely—about particular truths, particular threats, particular betrayals. They tracked inconsistencies in opponents with forensic patience. What declined was not cognition. What declined was a shared office for interpretation: a public status in which evaluating a claim could still reorder affiliation rather than merely express it. When truth-talk becomes primarily badge-talk, accuracy does not become irrelevant; it becomes recruited. Accurate claims are still preferred when they help one’s side. Inconvenient accuracies are discounted as tainted by source. That is not the absence of standards. It is the subordination of standards to belonging.

Another temptation is cult language. Enclosure is a real mechanism; a later chapter takes it up where a sealed narrative absorbs external pressure as persecution. Trump-era national politics in this date band were not that seal. The system remained open in decisive respects: hostile newspapers published every day; election results could still be litigated and certified; opposition parties contested power and won the 2020 presidential election; courts issued opinions that Trump’s own appointees sometimes authored against his preferred outcome. Openness did not restore cross-group persuasion. Visibility of conflict can feed sorting as efficiently as silence feeds enclosure. Confusing sorting with enclosure collapses unlike rooms into one diagnosis.

A third contrast matters for the chapters ahead. Allegiance that survives program change across years—belonging that outruns platform—is a different replacement for interpretive leverage. Sorting communication is adjacent but not identical. It works event by event, post by post, hearing by hearing: each disputed claim as a small loyalty probe that must be answered now. A durable political identity can wait out a policy reverse and still claim the same name. Badge-sorting cannot wait in the same way; it needs the next dispute to reconfirm who is in.

Hirschman’s older language of exit, voice, and loyalty remains useful if read carefully and lightly. Voice, in civic mythology, is evaluative speech that pressures authority toward repair. Under sorting pressure, voice often becomes display: speech whose primary audience is the in-group’s recognition system rather than a shared adjudicative public.10 Loyalty is then not quiet patience toward reform but continuous proof of alignment. Exit can mean leaving a party—or leaving a family chat. None of that requires people to be irrational. It requires an environment in which the costs of cross-group listening exceed the expected returns, and in which the returns of alignment displays are immediate and social.

Media ecology accelerates those costs. Outrage travels; careful hedging does not. A correction that arrives late is often received as a different species of belonging cue: they have shifted the story, therefore we must answer. Haberman’s reporting of White House messaging cycles shows how such loops can reward escalation and loyalty talk over slow joint revision.1 Partisan outlets compete to protect audiences from the humiliation of updating toward an enemy’s preferred frame. Klein’s institutional account explains why audiences were already prepared for that competition before any single figure saturated the news.5 The result is not two groups that refuse to think. It is two groups thinking hard past each other, with disagreement doing boundary work that negotiation would have done under thicker shared standards.

Fact-checking, then, could document relentlessly without reconstituting a shared evaluative public. Documentation still matters—for records, courts, later historians, for people not yet sealed into either camp. What documentation does not automatically supply is a path back into shared standing once claims have been received primarily as membership signals. The checkers are sorted. The checked are sorted. The audience watches the sorting and learns where it sits.

Many participants understood themselves as defending institutions, or defending a neglected public, or resisting what they experienced as existential disrespect. Those self-understandings can be sincere and still participate in sorting. Sincerity does not restore corrective travel. The essay’s burden is not to deny earnestness. It is to notice when earnest argument has changed jobs.

Meaning as badge. Disagreement clarifies who belongs.

The forwarded claim does not need to win a seminar in order to do its work. It needs to force a response that can be scored. Silence scores. Hedging scores. Enthusiasm scores. Correction from the wrong messenger scores as aggression. Agreement from the right messenger scores as recognition. Once that scoring system dominates a room, the room becomes clearer and less revisable at the same time. Clarity of belonging replaces clarity of shared fact as the practical achievement of argument.

This is why the intensity of disagreement in these years can fool observers who equate noise with democratic health. Democracies need disagreement. They also need disagreement that can still do corrective work across groups. When argument’s primary public product is affiliation information, more argument can deepen the split that argument was supposed to manage. Repair becomes renegotiation of membership rather than revision of claims. People experience that as moral awakening or as moral catastrophe depending on which side of the cut they stand. Structurally, both experiences can be produced by the same mechanism.

Sorting thrives on visible dispute; enclosure thrives on absorbing dispute into a sealed story; belonging that outruns program thrives on allegiance that outlasts platforms; performative compliance thrives when visible assent is compulsory. The 2015–2021 open-system case shows the first without forcing the others. Interpretation stays loud. Standing updates sideways into belonging rather than forward into shared correction.

Meaning as badge does not abolish truth. It changes the social office of truth-talk. A proposition can be accurate and still operate mainly as a flag. A proposition can be false and still operate mainly as a flag. In either case, the room’s first practical question becomes less “What follows if this is right?” and more “What am I declaring by repeating this?” Once that second question dominates, escalation looks like courage and repair looks like appeasement. The room gets clearer. It does not get closer.

When meaning sorts rather than persuades, disagreement stops being informative across the line it draws.

Look back at the forwarded post. The caption is still there. The bolded sentence is still there. The laughing faces are still there. The careful cousin’s unanswered request to read before judging is still there, orphaned between indignation and exit. What has become readable is the probe the chat performed without naming it. People answered by aligning—some loudly, some by strategic quiet, some by exiting and returning with a harder face. The argument that seemed to be about facts was, for the life of the thread, the argument that sorted the room. The words did not lose meaning. They changed destinations: away from joint revision, toward belonging.

Reread the same message a second time with that destination in mind. It still asks to be believed. It also asks, more urgently, to be answered in a way the coalition can recognize. That is why replies that merely add complexity so often fail, and why replies that restate the side succeed even when they add no information. The loyalty probe was never announced. The room completed it anyway. Once you can see the probe, you can also see why more links rarely help: each new link can be scored before it is read. The argument did not run out of words. It ran out of a shared place for words to land.

A politics—and a set of ordinary rooms—in which interpretation remains audible, contested, and busy while failing, across groups, to do the work of correction. The message on the phone was never only a claim. It was a test the room knew how to take.

Footnotes

  1. Maggie Haberman, Confidence Man: The Making of Donald Trump and the Breaking of America (New York: Penguin Press, 2022)—on the June 2015 announcement and 2015–16 campaign media tactics; rallies, repetition, grievance, and White House loyalty talk; post–November 3, 2020 messaging dynamics; and media feedback loops that rewarded escalation over joint revision. ↩ ↩2 ↩3 ↩4 ↩5

  2. Glenn Kessler, Salvador Rizzo, and Meg Kelly, “President Trump Has Made More Than 20,000 False or Misleading Claims,” Washington Post, July 13, 2020; and related Washington Post Fact Checker tallies through January 2021. Contestation of such tallies—and of the institutional authority of fact-checking itself—was a standing feature of partisan media ecosystems in the same years. ↩

  3. Jeffrey M. Jones, “Trump Approval More Stable Than Approval for Prior Presidents,” Gallup, January 18, 2019; Jeffrey M. Jones, “Trump Job Approval Sets New Record for Polarization,” Gallup, January 18, 2019. Gallup’s term summary reported Republican approval averaging about 88 percent and Democratic approval about 7 percent across the presidency, with an average partisan gap of 81 points: Jeffrey M. Jones, “Last Trump Job Approval 34%; Average Is Record-Low 41%,” Gallup, January 18, 2021. ↩

  4. Jones, “Last Trump Job Approval 34%.” Low overall averages alongside persistently high Republican approval caution against equating coalition stability with national endorsement. Gallup also noted that Republican approval did not fall below 77 percent at any point in the presidency. ↩ ↩2

  5. Ezra Klein, Why We’re Polarized (New York: Avid Reader Press, 2020), on party sorting after the mid-twentieth-century “big tent” era, stacked social identities, and the conversion of ordinary political conflict into identity-laden conflict. ↩ ↩2

  6. On contested 2020 messaging and reception, see Haberman, Confidence Man, pandemic chapters; White House and CDC statements from February–December 2020; and contemporaneous reporting such as Michael D. Shear et al., “Inside Trump’s Failure: The Rush to Abandon Leadership Role on the Virus,” New York Times, July 18, 2020. The argument here concerns sorting reception of claims, not a full public-health chronology. ↩

  7. U.S. House of Representatives, Articles of Impeachment against Donald John Trump, H. Res. 755, 116th Cong. (December 18, 2019); Senate acquittal, February 5, 2020. Second impeachment: H. Res. 24, 117th Cong. (January 13, 2021); Senate acquittal, February 13, 2021. ↩

  8. Official Electoral College results certified Joseph R. Biden Jr.’s victory. Numerous post-election lawsuits failed in state and federal courts. Haberman, Confidence Man, describes the post–November 3, 2020 shift from private aide conversations about loss toward public insistence on victory—cited as reporting on messaging dynamics, not as a claim about private belief. ↩

  9. Pew Research Center, “Views on the Rioting at the U.S. Capitol,” January 15, 2021; Pew Research Center, “Large Majority of the Public Views Prosecution of Capitol Rioters as ‘Very Important,’” March 18, 2021; Pew Research Center, “A Look Back at Americans’ Reactions to the Jan. 6 Riot at the U.S. Capitol,” January 4, 2022. ↩ ↩2

  10. Albert O. Hirschman, Exit, Voice, and Loyalty (Cambridge, MA: Harvard University Press, 1970). Used here only for the civic triad’s shifting functions under alignment pressure. ↩