Discussing without convincing: what a well-conducted disagreement produces
The idea that a contrary argument always reinforces convictions is very widespread. Recent work partly contradicts it, and that is rather good news.
A famous effect, revised downwards
The idea of a backfire, according to which a factual correction would reinforce the mistaken belief, was widely diffused from work in the early 2010s. Larger replication attempts, notably those of Thomas Wood and Ethan Porter, found that the phenomenon is rare and that corrections generally shift beliefs in the right direction, even in committed people.
The shift is often modest, it does not erase the original position, and it does not necessarily translate into a change of vote or attitude. But it exists, which distinguishes a discussion from a waste of time.
This revision is interesting on two counts. It corrects a discouraging belief, and it illustrates what replication brings: a spectacular and widely cited result may turn out to be far more fragile than announced.
What makes a discussion productive
Three conditions recur in work on the subject. The first is respect for the interlocutor's standing: a correction that humiliates is rejected along with the information it carries. The second is precision: a verifiable assertion is more effective than a general judgement. The third is the absence of an audience, because a disagreement held before witnesses engages face and makes retreat more costly.
A fourth condition, more rarely cited, is the availability of time. A change of position often requires returning to the subject later, alone, which never happens during the exchange itself.
What nevertheless remains out of reach
Corrections operate on facts, much less on values. Showing that a measure does not produce the announced effect may shift a position; showing that a priority is badly ranked does not belong to the same register, since there is nothing to verify.
When a position has become a marker of belonging, information likewise stops working, not because it is refused but because the cost of changing is no longer intellectual. That is the honest limit of everything above.
Setting an achievable objective
Seeking to convince in one conversation is an objective almost nobody achieves, and its failure produces bitterness on both sides. A more realistic objective consists in making clear exactly where the disagreement lies, which is useful to both speakers and verifiable at the end of the exchange.
That objective has an additional merit: it is symmetrical. Each leaves knowing what separates the positions, and what might possibly move them, which is precisely the information a quarrel never provides.
A format that works better than debate
Work on deliberative arrangements observes that exchanges in small groups, with plenty of time and an instruction to listen, shift positions more than short confrontational formats. It is not a question of virtue but of setting.
For an ordinary exchange, two transpositions suffice: ask the other to restate your position before replying, and accept to restate theirs. It is slow, a little artificial, and markedly more effective than being right.
Key points
The backfire from factual corrections is rare: they generally shift beliefs.
Respect for standing, precision and the absence of an audience make a discussion more productive.
Corrections operate on facts, not on values nor on positions that have become identity-laden.
Aiming to locate the disagreement rather than to convert is achievable and symmetrical.
This article describes documented mechanisms and disagreements. It recommends no position, no party and no vote, and commits to no reading other than that of the work it cites.
And you, where do you stand?
The Civix test situates your convictions on six dimensions and matches you with one of the twelve profiles, without ever telling you what to think.
Start my test