How people disagree, and what they think disagreement is
This site asks two questions that turn out to be one. How have people conducted conversations in which they disagree? And what did they take the disagreement to be really about?
The two questions cannot be separated, because every procedure smuggles in a diagnosis. The Socratic elenchus presupposes that the interlocutor’s beliefs are mutually inconsistent. Principled negotiation presupposes that positions hide compatible interests. Nonviolent Communication presupposes that needs never truly conflict, and that only strategies do. The double crux presupposes that the dispute lives in a world-model with a testable node. Adopt the therapy and you have accepted the diagnosis, usually without noticing.
The wheel makes that commitment visible. Methods sit on the left, sorted by telos, meaning what the exchange is for. Views sit on the right, sorted by domain, meaning where the trouble is said to live. Each bundled thread joins a method to a view it presupposes; the heavier thread marks the primary one. Hover to trace a method, click to read it. Views in grey italics are diagnoses that no method in this survey takes as its premise.
Three findings the data supports
1. The oldest taxonomy is still the best triage. Hermagoras’ stasis theory (2nd c. BCE) told an advocate to locate a dispute before arguing it. Is it about whether something happened (fact)? About what to call it (definition)? About whether it was justified (quality)? Or about who gets to decide (jurisdiction)? These four map almost exactly onto four of the nine domains used here: epistemic, linguistic, axiological and power. Much modern advice, such as Chalmers’ elimination test or the search for a crux, amounts to stasis theory reinvented.
2. There was a therapeutic turn, then a partial epistemic counter-revolution. Classify each method by its primary presupposition and count by era. Before 1500, about half of all methods locate disagreement in belief, and none locate it in the psyche or in the relationship. Between 1900 and 1975, half of all new methods locate it in the psyche or the relationship (Rogers, Burton, Watzlawick’s heirs). After 1975 the epistemic share recovers, carried by argumentation theory, deliberative polling and rationalist practices such as adversarial collaboration and the double crux. The sample is curated, so read this as an orienting claim, not a measurement.
Share of methods founded in each era, by the domain of their primary presupposition. n = number of methods in the era.
3. Some diagnoses have no therapy. Thirteen of the views here are presupposed by no method in this survey. They include incommensurable paradigms, the differend, mimetic rivalry, the narcissism of small differences, sacred values, class interest and Kegan’s developmental account. Each either awaits its practice or implies that no conversational practice can reach it. Kuhn, Lyotard and Girard all say, in different idioms, that at the limit one does not argue but converts, testifies or refuses the scapegoat. That is where this map ends and other kinds of action begin.
Taxonomy
The wheel, a lineage map across 2,500 years, and a full method × view matrix.
Methods
Seventy ways of conducting disagreement, from antilogic to AI-mediated deliberation, with procedures and sources.
Views
Fifty-four accounts of what disagreement is “really about”, grouped into nine domains.
Timeline
Methods and views interleaved chronologically.