Polarization: Why Conversation Between Groups Becomes Harder

Epigraph

personal animosity or affection

David Hume, Of Parties in General, PG3, Mil56. A fragment of the author’s own prose about personal grounds of parties, not a general explanation of contemporary conflict.

What this article is about

Imagine two urban initiatives. One wants to preserve an old building; the other wants an accessible public space on the site. At first they dispute the walls’ condition, expenditure, and the land’s purpose. Then every answer begins to mean more: does someone support their own side, respect history, or understand the city’s real needs? Members can already explain away any argument from the other side. Yet some have not seen the building survey, and others disagree about values even after reading it. A hostile conversation hides several different questions beneath one word: polarization.

This article distinguishes differences in positions, attitudes toward people, selection of messages, and the cost of a public response. We consider one particular study with its limitations, separately from editorial ways of describing an imagined conversation. No single study explains every conflict, and a new link does not guarantee convergence. The task is to understand precisely what has moved farther apart and which limited question remains available for joint consideration. Real threats, unequal influence, and the right to end an exchange remain relevant; reconciliation is not a mandatory outcome.

A separate, wholly fictional family scene examines political pressure and unequal consequences of public speech. Retaining contact, retaining disagreement, agreeing to publication and reconciling are different decisions. Neither kinship nor an observer's interest in dialogue decides for participants.

I. Distance between opinions and distance between people

In our imagined urban scene, participants can evaluate the building very differently while acknowledging one another’s competence. One prioritizes preserving its historical appearance; another prioritizes accessible use of the site. This substantive disagreement cannot be removed through a kind tone. In another scene, the proposals are already close: both sides will retain the façade and change the interior. Yet people refuse to work together and consider any involvement by the other side humiliating. The distance between projects has diminished while distance in the relationship remains. Neither should automatically be measured through the other.

Here, polarization is a working name for several possible changes, not a diagnosis of everyone involved. We separately ask how far answers to a particular question differ, which feelings and judgments concern the people giving those answers, which information is admitted for consideration, and what publicly revising a position costs. This is an editorial map for this article. It does not replace a study’s particular definitions and measurements. When researchers measure attitudes toward political proposals, their result cannot be retold as proven hatred between people. When a feeling is measured, the content of every conviction cannot automatically be declared changed.

The distinction also changes the possible next step. Different values call for discussion of criteria and the costs of options. An unreliable report calls for checking that report. Humiliation calls for examining conditions of participation, sometimes stopping. Inability to answer openly calls for attention to power and consequences. No single prescription covers these tasks. More facts can leave a moral disagreement intact. Less harsh treatment can leave people without a joint decision. Such an outcome need not mean the work was useless: what matters is the question it actually addressed and what remains unknown.

Striking disagreement does not prove a recent increase. An old difference may have become visible, or participants or proposals may have changed. Claims of growth need comparable past and present information. Everyday inquiry can describe present difficulty without explaining an era or treating an unpleasant meeting as proof of society's disintegration.

II. A dispute can include different grounds of disagreement

Return to the building. Engineering concerns its condition, finance concerns available spending, values concern what deserves preservation, and organization concerns who decides. These levels meet without becoming identical. If a survey identifies expensive repairs, a preservation advocate need not stop valuing history. They must consider their proposal’s new cost. An advocate of new space does not thereby acquire the right to declare all concern for history meaningless. A fact changes the assessment of conditions; it does not automatically establish the sole permissible value for every person.

In Of Parties in General, Hume distinguishes grounds of parties and observes their mixture. We use this as a historical invitation not to reduce conflict to one cause. His religious and racial judgments are not adopted as contemporary facts; the scheme is no experimental explanation of our city.

An imagined committee can discuss spending and an old resentment toward its organizer simultaneously. One member has a strong argument about the project and dislikes their opponent. Another feels warmly toward someone whose calculation has weak support. Choosing only one convenient level cannot resolve the issue. Hostility does not make every argument false; attachment does not make it true. Relationships nevertheless affect the availability of an answer in this scene: people may avoid clarification because they expect an attack. This is an observable arrangement within the example, not an established psychological cause of every disagreement.

Personality explanations can hide interests; spending explanations can hide treatment costs. Name the area and what could change its assessment: evidence for calculations, reasons and agreed procedures for priorities, a separate response to an insult. One meeting need not solve everything, but separation must not postpone misconduct indefinitely or conceal the organizer's authority.

III. More opposing messages do not mean more shared understanding

A familiar thought is simple: if people read only their own side, give them more opposing views. It raises a reasonable question about limited information but does not specify which messages, for whom, for how long, and under what conditions. An opposing programme, a mocking video, an acquaintance’s account, and a prominent party leader’s speech are different encounters with another position. Message quantity alone does not establish whether someone encountered a new reason or another occasion to reject the sender. In our scene, a stream of slogans about the city may bypass the building’s condition entirely.

Bail and colleagues’ 2018 experiment involved party-identified Americans who regularly used Twitter. A randomly assigned group received a paid offer to follow a bot carrying opposing political messages for a month. Responses about political issues were measured. Republicans shifted toward conservatism; Democrats’ change was not statistically significant. This does not establish that all contact is harmful: the authors did not identify the precise mechanism, and transfer to other people and settings is limited.

An honest practical conclusion is therefore narrower than either a promise or a prohibition. Adding an opposing stream cannot itself be treated as sufficient evidence of future convergence. Nor does one study justify avoiding every different thought. Two neighbours discussing a schedule are not equivalent to an offer to follow a bot; reading one checkable document is not equivalent to a month of messages from political voices. Establishing the effect of those particular actions requires research on those actions. The historical experiment limits generalization; it does not supply a ready outcome for every new scene.

Absent convergence does not prove understanding impossible. Someone can learn another's values, retain their assessment, gain confidence from verified facts or change opinion without changing attitudes toward a leader. Define understanding before judging outcomes; otherwise unwelcome answers become polarization and welcome ones healing, beyond what one undescribed exchange establishes.

IV. The other side may be encountered only through selected material

An urban group receives a recording of its opponents’ harshest speech. Another group circulates a newcomer’s mistake as evidence that preservation advocates are generally ignorant. These episodes are explicitly invented. People have encountered the other side’s words, but a selected part does not describe every participant or argument. Representativeness becomes a question: whose voice is before us, who selected it, and which claim can it assess? Rejecting a harsh speech need not become a conclusion about everyone connected to the project.

Selection can be necessary. Nobody can inspect every message about a city, and limited time requires choosing what matters. The problem is not choice itself but expanding what it supposedly establishes. A representative may have official authority concerning a particular decision. Their words then matter to the project’s position without establishing every supporter’s private convictions. An ordinary member may speak only for themselves. That status should not change for convenience: a favourable message becomes the group’s authentic voice while an unfavourable one becomes an incidental exception, or the reverse.

A check does not require seeking the opponent’s most agreeable face. The other side may have a genuinely weak proposal, an exclusionary procedure, or a dangerous action. Consider the strongest available grounds and actual acts while preserving their scope. Demands to find something good can hide a substantial difficulty for the sake of apparent balance. Precision permits a negative evaluation of a particular act. It also resists expanding that evaluation into everyone’s general character. This article does not require neutral feelings; it asks how wide a conclusion is relative to the information supporting it.

Identify a current document or decision for which the side is responsible: proposed building use and costs. If no official position exists, retain the unknown. An absent document proves no hidden malicious plan but may limit readiness for joint action. Waiting for adequate information is not proven rejection of another group's human worth.

V. A public answer has another addressee

At a joint meeting, Anton accepts one calculation from the other initiative. His own colleagues hear not only information about cost but a possible signal of changing sides. Anton explains that he accepted a figure rather than the entire project. Someone else uses his agreement in a post suggesting all previous objections were withdrawn. A local answer now has several addressees: the immediate interlocutor, one’s own members, and an outside audience. The scene shows a possible cost of public clarification. It does not prove that people always conceal opinions or that every group punishes honesty.

Sometimes the cost relates to a real role. An initiative’s representative must clarify whether they speak personally or state an adopted decision. An organizer cannot promise common expenditure without agreement. Those restrictions differ from a ban on thought. But if an ordinary member must regard every fact reported by opponents as false, the role no longer explains the requirement. Authority to act should be separated from acknowledging information. The former may require a joint decision; the latter need not mean committing to every subsequent proposal from the sender.

In public settings, preserving the bounds of acknowledgment is particularly useful. Someone can state which calculation was checked, what changed in their assessment, and which questions remain. This is no ritual demand for a qualification after every true opposing remark. Ordinary agreement sometimes suffices. Clarification matters where an answer is actually used to support a wider statement. A person can decline to participate in a spectacle if their words are repeatedly altered or treatment becomes humiliating. Precision should not become an obligation to continually service other people’s interpretations of one’s belonging.

Someone claiming to accept a detail after promising the whole project remains accountable for actual words. Boundaries do not shrink responsibility retrospectively. Distinguish available words, records and context; correct an overbroad promise and name consequences. This concerns an agreement, not proof of impossible dialogue; polarization must not swallow accurate representation or delivery.

VI. Disagreement need not move toward a midpoint

When sides argue harshly, asking each to abandon half its position is tempting. Yet the average of two statements need not have better support. A survey either describes a particular wall condition or requires clarification; half an incorrect calculation does not become fair. Values are not always arithmetically additive either. Different desired uses of a site require a particular arrangement rather than equal concessions for an attractive report. Less harsh conversation and better decisions need separate checks, so that improvement in the first is not purchased by deterioration in the second.

Sides can converge on a description of fact while retaining different preferred actions. Both accept the cost, but one will seek funding and the other prioritizes another task. Disagreement has become clearer although options still formally diverge. This is substantive where conditions permit free answers. Forced silence should not become a fortunate middle ground. Nor should every firm objection become intensified hostility. A clear refusal sometimes preserves an honest conversation: someone states which option they will not support and why.

One position may have better support from available evidence. Respect for the people holding another does not require granting them an equal share of truth. An argument’s advantage nevertheless gives no universal authority to determine everyone else’s motives. A measurement error can be shown without establishing that someone never wanted to understand the city. Competence in calculation can be acknowledged without handing it every moral decision. This separation does not remove power and interests. It permits discussing them directly instead of concealing a value choice inside a supposedly exclusive scientific answer.

Conversation reproducing humiliation need not continue for harmony where no joint decision is required. A necessary decision may need another procedure with authorized people, documents and a deadline. No universal institution or fairness is promised. Voluntary conversation and practical decisions differ; a future project result does not require enduring every treatment.

VII. The sides do not always have equal influence

Two initiatives may look equal in a meeting photograph while one has premises, money, and authority to set the schedule. The other only has an opportunity to submit a proposal. Equal speaking time does not establish equal freedom. Someone dependent on an organizer may remain silent for reasons an observer does not know. They should not be declared inwardly convinced or incapable of adult conversation. In the imagined scene, establish who decides, which responses are available, and what refusal costs instead of grading everyone’s courage.

An appeal for both sides to calm down has limits too. If one insults and another asks them to stop, identical reproaches conceal different acts. Humiliation can be limited while an argument remains under consideration. An argument may itself contain an exclusionary demand that cannot be discussed as an innocuous preference. Precision is not outward symmetry. Yet assessing substantial harm requires evidence and appropriate context; one unpleasant feeling does not establish all of someone else’s authority. This article does not replace local legal or professional assessment with a general formula about polarization.

A side with less influence can also act dishonestly. It may circulate an incorrect calculation or alter opponents’ words. Unequal resources do not make every argument from it true. Correcting its mistake likewise does not remove the question of access to a decision. These areas need holding together. Choosing only what is convenient returns conversation to a contest of complete characterizations: either someone is right about everything or their difficulty no longer matters. Particular inquiry retains both responsibility for information and actual limits on participation without trading one for the other.

Substantial risk can make a joint meeting inadequate; direct contact and openness are not always safe. Retain facts, consider available support or leave exchange unexpanded. This responds to conditions, not dialogue's general doom. Observers cannot impose an experiment's costs on participants for attractive reconciliation; voluntariness remains relevant.

VIII. A shared area can be smaller than a shared worldview

The urban initiatives retain one available question: obtaining and considering the building survey. They can agree on its source, date, and the information needed for the next decision. This does not mean adopting an identical understanding of history or the city’s future. The shared area is limited. That makes checking it possible: was the document obtained, does it concern the correct building, what conditions does it describe, and what lies beyond the survey’s competence? Practical clarity does not require announcing that every broader disagreement has vanished.

A very small area can, however, bypass the main cost. If everyone will exchange a document but one side has already disposed of the building without discussion, joint checking risks becoming decorative. People need to know what it can still accomplish. They can discuss consequences, correction, or subsequent use while honestly acknowledging the previous choice is unavailable. This cannot be presented as an open choice between two projects if that choice no longer exists. Limitation is useful where it retains a real task rather than concealing the disappearance of the opportunity for which people gathered.

Another limit concerns trust in the source. The sides may disagree even about the survey’s suitability. Clarify the particular grounds: doubt about method, provenance, currency, or the assessor’s competence. Not every distrust is irrational; not every doubt requires endless fresh checking. If criteria change after each document, that can be described through the exchange. If a substantial ground remains unanswered, retain it. Joint action may be impossible without more information or possible under a clear temporary qualification. The choice depends on the cost of error, which a wish to reduce tension cannot determine by itself.

A checked document need not restore warmth: Anton and others may still avoid personal contact. Assess the actual answer and what remains unchanged without inventing transformation. Broader trust is another question, not one meeting's entire burden. Retain the achievement and stop where conditions cease to support joint consideration.

IX. Notice change without inventing an effect

After a meeting, saying the groups became less polarized is appealing. But what changed: the building position, description of an opposing argument, treatment, document access, or willingness to cooperate? One observable shift can be named without an overall label. People stopped claiming that the opposing project disregarded expenditure after seeing its calculation. Their priorities still differ. The event is not smaller merely because it did not end in unity. It has a particular scope worth retaining in the account of the result.

Self-observation does not measure the whole community either. One person feels less angry after discussion. Another is tired and unwilling to continue for now. A third revised a project after new information. These answers cannot be combined into an invented indicator of collective recovery. A research conclusion requires an appropriate design and comparable measurements. Ordinary practice can simply describe known and unknown information honestly. Even a clear personal change does not establish someone else’s hidden inner change; accepting a paraphrase records their response rather than all their future intentions.

Related United Wisdom topics examine substantive work on a claim and objection, personal conditions for handling information and the connection between explanation and membership within a group. Here, the subject extends beyond one reply while remaining narrower than an explanation of society: distinct intergroup processes and the limits of particular data. These thematic links help choose the next question without replacing primary research or certifying every related text’s readiness. Accurate argument can help, but by itself removes neither resource differences nor the structure of an information stream.

Checks may narrow an overbroad polarization hypothesis: different documents clarified, only one priority differs. Or deeper constraints emerge: acknowledging an opposing fact costs someone their role. Both require revising the account of opinions and participation. Defending a label for consistency or reputation helps less than revealing the actual question and available response.

X. Conversation remains an opportunity with conditions

Intergroup conversation has no single necessary cause of difficulty or guaranteed exit. Sometimes a report needs checking, sometimes a criterion needs clarification, sometimes a decision procedure needs changing, and sometimes unsafe treatment needs limiting. Converging positions can be helpful or conceal a substantial difference participants wish to retain. A kind tone matters but does not replace reasons; strong reasons matter but do not justify humiliation. These distinctions permit particular work without promising that a well-arranged meeting will necessarily restore agreement and mutual liking.

In our urban scene, someone can retain concern for history and acknowledge the repair’s cost. They can want new space and acknowledge value in part of the building. They can reject both current projects. Each response needs its own subject. Learning what the other actually proposes is one clarification. Still considering that group’s participation unacceptable is a separate disagreement. The first should not be presented as the second’s disappearance. Nor should the first be dismissed because relationships failed to become close. Shared work can be limited and real.

Returning to the initial word can reduce unnecessary certainty. Instead of declaring hatred inevitable, describe sides farther apart over a decision, a source rejected before consideration, or a public clarification with a high cost. Each claim requires its own evidence. Where evidence is absent, a question remains. This care does not prevent seeing real harm when established. It prevents assigning the entire conflict a ready story in which every subsequent fact must play a predetermined role and nobody can respond differently.

Anton need not resolve society's polarization: check one calculation, bound agreement and choose participation. Considering one document requires no group merger. Supportive conditions permit limited joint action; otherwise ending remains possible. Precision guarantees no agreement but distinguishes difficult decisions, attitudes, information access and response opportunities, leaving the next question to participants.

XI. Closeness Does Not Make Political Speech Equally Costly

In a separate fictional scene, Anton and his sister Marina dispute a political event in a family chat. Marina proposes a shared public statement. Anton's job and housing depend on people with power over him; Marina's circumstances differ. This describes neither an actual country nor a proven threat to every worker. Equally long statements can have unequal consequences; those circumstances matter before judging courage or family loyalty.

Anton can discuss one report without agreeing to publication in the family's name. Private conversation, joint text and public circulation have different audiences. Forwarding or screenshots can carry words beyond the initial exchange; a chat's private label guarantees no confidentiality. This supplies no secure-communication technique. Requiring public speech as proof of trust adds a condition kinship does not establish. Refusal does not reveal hidden agreement with another's political position.

If both want contact and conditions permit, they might discuss a relative's health or specific help while postponing politics. That is neither forgiveness, equal correctness nor promised former closeness. Either can decline even limited contact. Care is not payment for political loyalty; disagreement need not automatically end every relationship. Under threat, practical protection and available assistance have their own subject, which argument cannot replace.

Relationships remain an option rather than compulsory reconciliation, disclosure or contact at any cost. Silence alone establishes neither agreement, safety nor resolved conflict. Actual harm needs facts; unequal speech costs do not make every assertion true. Hume and the Twitter experiment establish no safe family-contact method under pressure. No national legal instruction, confidentiality or restored trust is guaranteed. Participants retain limits, including pauses and ending contact.

What this means in practice

Choose one safe conversation among particular people about a specific decision. Name the subject without a complete account of personalities: which question needs answering, who decides, and what would an error cost? Separately identify differences in proposals, known treatment, and information supporting each side. Do not fill unknowns with guesses. If threat or coercion is already involved, the exercise does not require continuing contact. You can stop and consider available support without treating refusal as an admission of incapacity for dialogue.

If conversation is voluntarily available, identify one checkable area and its limit: a calculation’s source and currency, one proposal’s meaning, or conditions for considering a document. Do not treat every opposing voice as representative of the entire group. Neither promise to revise your opinion before checking nor demand that promise from another. Ask which grounds could actually change the present assessment of this area. An answer may reveal a criterion, an unknown, or unwillingness to consider the question; none should become a universal characterization of someone.

After the exchange, describe a particular change and the remaining difference. A document found, a fact clarified, a criterion made visible, or an opportunity to answer still unavailable are distinct results. Do not call them measured reductions in polarization without the corresponding check. You can choose another small step, retain disagreement, or end participation. The exercise does not guarantee convergence and does not replace research into causes. It retains the scope of what is known: discuss an available question, consider the response’s real cost, and avoid requiring one conversation to promise a shared world it cannot deliver.

Under political pressure, distinguish private conversation from public text: who will see it, what consequences are known, and is refusal available? Do not test safety by disclosing vulnerable information. Only with voluntary agreement, name a permitted subject and circulation limit. Agreement on one subject is not reconciliation or agreement on everything; unavailable safe contact requires no proof of love or courage.

Key quotation

seeming tranquillity

David Hume, Of Parties in General, PG9, Mil59–60. A fragment of the author’s own prose in the historical context of apparent peace between parties, not an assessment of our imagined meeting.

Connections between traditions

Basis	Question	Limit
Hume’s historical essay	Has disagreement been reduced to one cause?	A historical philosophical classification, not a general contemporary law
Bail and colleagues’ experiment	Which contact and outcome were checked?	A particular population, measure, and intervention; precise mechanism unidentified
The editorial urban scene	Which limited area remains available for consideration?	An invented example, without guaranteed convergence

Family political pressure is a separate editorial addition. Hume's historical classification and the particular bot experiment establish no safe private contact or compulsory reconciliation; silence is no measured peace.

Sources

David Hume, Essays, Moral, Political, and Literary, Part I, VIII, Of Parties in General. Hume Texts Online, edited version. The entire essay PG1–15 and notes were personally read. The quotation budget covers the whole collection; historical judgments are not adopted as contemporary facts.
Christopher A. Bail and colleagues, Exposure to opposing views on social media can increase political polarization. PNAS 115(37), 9216–9221, 2018, DOI 10.1073/pnas.1804840115. The main text, Research Design, Results, and Discussion and Conclusion were personally read; full SI reading is not claimed. CC BY-NC-ND 4.0: no full foreign text, figures, or adaptation are imported; a brief original report of checkable facts with a link is used, without direct quotation.

About the authors

Andrey Zarinsh is the author and founder of United Wisdom. He explores inner transformation, the unity of traditions, intercultural dialogue, and contemporary spiritual discipline. Instagram: @andrey.zarinsh.

Maria Evilkut is a coauthor of United Wisdom. Alongside her research, she seeks to test these ideas in practice and observe how principles of unity and mutual understanding work in real life between people and cultures. Instagram: @ilkutki.