Why Pushing an Argument Too Far Can Turn People Against It

Most of us assume that if you want someone to support a cause, you hand them the strongest possible case for it. Stack up the benefits, make the language confident, leave no room for doubt. But what happens when an argument is so confident, so over-the-top, that it stops sounding like persuasion and starts sounding like a parody of itself? A recent study suggests the answer is surprising. Instead of winning people over, an absurdly exaggerated message can quietly steer them in the opposite direction.

The work comes from Uri Lifshin of the Israel Center on Addiction and Mental Health and the psychology department at the Hebrew University of Jerusalem, working with colleagues Daniella Yaffe and Omer Kochav. Their paper appears in The Journal of Social Psychology [1]. The researchers were interested in a counterintuitive technique sometimes called paradoxical thinking, where you do not challenge someone's view head-on. You amplify it. You take the position to a ridiculous extreme and let the person's own discomfort do the rest.

To test this, the team ran three separate randomized experiments, with 431 participants in total [1]. They chose a deliberately divisive topic: the legalization of prostitution. Participants read one of several types of messages. Some saw standard arguments against legalization. Some saw neutral, unrelated content. And some read pro-legalization statements that had been cranked up to an almost cartoonish degree, the kind of claim that frames the practice as an unambiguous social good with no downsides worth mentioning.

Here is the twist. Reading those extreme pro-legalization arguments did not make people more supportive. If anything, it nudged them the other way. Both moderate and strongly exaggerated versions of the pro message tended to soften favorable attitudes toward prostitution rather than strengthen them.

Why would that happen? The researchers traced a specific chain of reactions. When people encountered the absurd statements, many of them disagreed sharply. That strong disagreement was the key. The more a participant rejected the inflated claims, the more their overall support for the issue dropped, and that drop in turn predicted a greater willingness to sign a petition calling for the practice to be banned. So the persuasion did not arrive through the front door. It worked indirectly, by provoking a reaction that the person then generalized into a firmer stance.

That indirect route is what makes this interesting. A normal anti-legalization message tells you what to think. The paradoxical message does something stranger. It hands you a position you are supposed to agree with, makes it so unreasonable that you instinctively push back, and lets your own resistance carry you somewhere new. You are not being argued into a conclusion so much as recoiling into one. The process is a reminder that the mind imposes its own frameworks on incoming information — much as it does when the visual system grades beauty in a building by entirely different rules than it applies to a painting.

It is worth sitting with how unusual this is as a mechanism. Most persuasion research focuses on credibility, emotion, repetition, or social proof. Paradoxical thinking sidesteps all of that. It does not try to be convincing. It tries to be so unconvincing in a particular direction that the listener corrects course on their own. The technique has roots in conflict resolution work, where amplifying a group's own narrative back to them has sometimes loosened entrenched political attitudes. This new study extends the idea into a fresh domain and tries to map the psychological steps in between.

Does this mean you can flip anyone's opinion by overstating the case? Not so fast. The study comes with real limitations, and the authors are upfront about them. The samples were modest, ranging from under a hundred to a couple hundred people per experiment. A large number of participants had to be excluded because they failed attention checks, which always raises questions about who remained in the final pool and how carefully they read. Two of the three experiments skewed heavily male, with only a handful of women included, so the findings may not transfer cleanly to how women respond to the same messages. And the topic itself was a single charged issue. Whether the same backfire effect shows up for, say, tax policy or vaccine attitudes is an open question.

There is also a deeper caveat about interpretation. The effect ran through self-reported disagreement and a petition-signing measure, not long-term behavior. People sign things in the moment. Whether an exaggerated message produces a durable shift in belief, or just a temporary flinch that fades by the next day, is something these experiments were not built to answer.

Still, the basic observation is hard to ignore. We tend to think more conviction is always better, that the loudest, most absolute version of a claim is the most effective one. This research points the other way. Past a certain threshold, intensity reads as implausibility, and implausibility invites resistance. An argument can be too strong for its own good.

It is a quietly humbling result for anyone in the business of changing minds, whether that is an advocate, a marketer, or just someone trying to win a dinner-table debate. The line between making your case and overplaying your hand may be thinner than it feels. And the person you are trying to persuade may be doing more of the work than you realize, just not in the direction you intended. The way literacy quietly reshapes how people parse even unrelated information offers a useful parallel: as reading rewires spoken-language processing in the brain, prior cognitive habits colour every subsequent act of listening, including listening to an argument. For more cognitive science coverage, see the full category.

Sources

1. Yaffe, D., Kochav, O., & Lifshin, U. (2026). Paradoxical thinking and attitudes toward prostitution: preliminary experimental findings. The Journal of Social Psychology. https://doi.org/10.1080/00224545.2026.2667805

This article summarizes published research for general informational purposes only and does not constitute professional advice.