The problem
You've done it. Marshaled the facts, built the airtight case, maybe even won the argument — and watched the other person walk away believing exactly what they believed before, now slightly more sure of it. We treat changing minds as a debate to be won, and the strange, consistent finding is that winning the debate doesn't change the mind.
If facts and arguments reliably reduced hostility, decades of both would have done it by now. They haven't. So either people are hopeless, or we've been using the wrong tool. The research says the second one — and it points at something much less satisfying to our egos than being right.
Hostility falls when people genuinely encounter each other and take each other's perspective. Not when they're out-argued. This is the evidence under everything THF means by "Find Common Ground."
The mechanism
Two bodies of evidence converge.
Intergroup contact. Gordon Allport's old hypothesis — that contact between groups reduces prejudice — got its definitive test in Pettigrew and Tropp's meta-analysis of 515 studies. Contact works, reliably if modestly. It works best under Allport's four conditions: equal status in the situation, common goals, cooperation instead of competition, and support from authorities or institutions. Those conditions help, but the effect shows up even without all of them. Crucially, the research identified how it works — three mechanisms: more knowledge of the other group, less anxiety about interacting, and more empathy and perspective-taking. The emotional mechanisms (anxiety, empathy) carry more weight than knowledge — which is exactly why a fact sheet underperforms a shared meal.
Perspective-getting. The second strand is even more striking. In a rigorous randomized field experiment, Broockman and Kalla found that a single ~10-minute "deep canvassing" conversation — non-judgmental, built on asking someone to recall a time they were treated unfairly and how it felt — durably reduced prejudice for at least three months. The linchpin wasn't presenting arguments. It was "perspective-getting": helping the person connect the issue to their own lived experience of being on the outside.
The honest limits. None of this is magic. Contact's average effect is real but modest, some of the evidence skews toward easier cases and may be inflated by publication bias, and the hardest, most committed hostility is the hardest to move. "Find Common Ground" is a direction with real evidence behind it, not a guarantee that every gap closes. The claim is humbler and sturdier than a slogan: encounter and perspective beat argument, more often than not.
The operating system
Five steps to actually lower the temperature across a divide.
Stop trying to win the argument
Accept the finding that defeating someone's position rarely changes it and often hardens it. Drop "how do I prove them wrong" and pick up "how do I genuinely encounter this person." That reframe is the whole pivot; everything else follows from it.
Create equal-status, common-goal contact
Where you can, build the Allport conditions: situations where you and the other person meet as equals, working toward something you both want, cooperating rather than competing. A shared project, a common problem, a team — these do more to dissolve hostility than any conversation about the disagreement itself.
Lower the anxiety, raise the empathy
Since the emotional mechanisms do the heavy lifting, optimize for them. Make the encounter low-threat (warm, unhurried, not a debate), and steer toward the human behind the position. The goal is for both nervous systems to read "safe" and for each side to glimpse the other as a person, not a category.
Use perspective-getting, not arguments
Borrow the deep-canvassing move: ask, listen without judgment, and invite them to recall their own experience of being treated unfairly for being different — and how that felt. You're not handing them your perspective; you're helping them locate the feeling in their own life. That connection moves people where data doesn't.
Expect modest, real, durable change — and respect the limits
Calibrate your hopes honestly. You're not going to flip someone's identity in one chat, and the most entrenched hostility resists everything. But a genuine encounter can produce a real, lasting shift — about one in ten, in the canvassing data — which, at scale and over time, is how the temperature actually drops. Modest and durable beats dramatic and fake.
The printable: what actually works
Print it. Trade the argument for the encounter.