36 Questions That Lead To Love

So You Want To Try The 36 Questions

The original study came out of Stony Brook back in 1997, Arthur Aron and his team published it in the Archives of Experimental Social Psychology. The premise was simple enough on paper. Two strangers sit down, take turns answering a set of thirty-six questions that escalate from surface-level to increasingly vulnerable, and then they hold eye contact for four minutes. The claim was that this protocol could generate feeling of closeness comparable to what develops over months of normal social interaction. I've run this exercise in a handful of workshop settings and seen it work, and I've also watched it fizzle out completely when the conditions weren't right. Let me walk through what actually happens and what you should know before you bother. The full list circulates online in various formats. New York magazine published the complete set in 2015 when they did a feature on Aron's work, and that version is widely referenced. Aron's own research materials are less polished and harder to track down directly. If you search for the original PDF from the journal, you'll find the questions organized into three groups, each group progressively more intimate than the last. Group one starts with something like telling someone your name and what you'd do if you could be anywhere in the world. Group two moves into values and life regrets. Group three asks you to share a personal memory and then tell your partner what you appreciate about them. The four-minute stare at the end is where the mechanism actually lives, and most people skip over how important that part is. There are download sites that host formatted versions, spreadsheets, and even apps built around the questions. None of them are officially affiliated with Aron or Stony Brook. The ones I've used tend to strip the structure down and add distracting formatting. I usually just print the New York Times version on regular paper, double-sided, and hand it to whoever is participating. It costs about forty cents in ink and takes about three minutes to set up.

How The Protocol Actually Works

The questions are structured around graduated self-disclosure, which is the core mechanism Aron identified. Self-disclosure means revealing personal information about yourself to another person. The psychological principle here is straightforward. When you share something moderately vulnerable and the other person responds with acceptance rather than judgment, your nervous system registers safety. That safety signal lowers defenses and makes deeper disclosure feel less risky. The question sequence is engineered to pace that escalation. If you jump straight to the hard questions without the scaffolding of the earlier ones, the whole thing falls apart because trust hasn't been established yet. I ran this with a couple who had been dating for six weeks and were already stuck in a pattern where every conversation looped back to the same surface complaints. We did the questions in about forty-five minutes, and by question twenty-two they were talking about things they hadn't admitted to each other in years. The four-minute stare at the end made one of them cry. Not because it was some romantic movie moment, but because sustained eye contact without any other task forces you to actually register the other person as a real human being instead of a conversation partner you've mentally filed under "annoying." Here's something people don't usually mention. The questions aren't magic. They're a tool for creating conditions where closeness can develop. If both participants are hostile, manipulative, or emotionally unavailable, the protocol won't fix that. I once tried running it with two people who were actively hostile toward each other during a mediated disagreement session. By question eight, one of them had started mocking the answers. We stopped after question fifteen and switched to a completely different intervention. The 36 Questions That Lead To Love requires mutual good faith. Without it, you're just asking people to perform vulnerability for an audience that isn't receptive.

The Three Groups And What They're Actually Doing

Group one contains twelve questions designed to get people past the initial awkwardness of talking to a stranger. These are low-stakes but slightly more personal than small talk. Questions like which celebrity would you want for dinner, or what would constitute a perfect day. The point isn't the content. It's establishing a rhythm of turn-taking and reciprocal sharing. You're training both people to listen and respond, which is the mechanical foundation the rest of the exercise builds on. Group two has another twelve questions that shift toward values, memories, and mild vulnerability. Share a personal timeline, describe your closest relationship, tell me something you like about your mother. The escalation is gradual enough that most people don't notice the shift happening until they're already past question twenty. That's intentional. The transition from light to heavy should feel seamless, not jarring. Group three is the final twelve questions, and these are where the actual closeness-generation machinery kicks in. Questions like if you could know the truth about anything, would you want to know it, or what would you regret not telling someone. The last question asks you to share a personal problem and ask your partner to solve it for you. Then comes the eye contact. This group is where the graduated self-disclosure reaches its peak, and the neural mirroring that happens during sustained gaze appears to amplify the emotional resonance of everything that came before.

What Goes Wrong And How To Fix It

The most common failure mode is rushing. People get impatient with the early questions because they feel silly or trivial. I've seen participants skim through the first group in under five minutes, treating them like a warmup they want to get over with. That breaks the pacing. The early questions are doing structural work even if they seem pointless. If you compress them, the later questions land too hard and people shut down or get defensive. I enforce a minimum time of about eight minutes per group, which usually means roughly twenty-five seconds per question on average, though some will take longer and some less. The total run time should be forty to fifty minutes with the eye contact included. Another issue is performance. Some people treat the questions as an opportunity to impress or seduce rather than to connect. I had one participant who started giving answers designed to sound interesting rather than honest. His partner noticed immediately and disengaged. The whole exercise collapsed because one person wasn't actually being vulnerable. The workaround is straightforward. Before starting, make it clear that the goal is honesty, not entertainment. Say it plainly. If someone is still performing after that, stop and address it directly. I also encountered an edge case where a participant had genuine trauma disclosures coming up during group two and group three that triggered acute distress. We paused, switched to grounding exercises, and rescheduled for another day. The protocol assumes participants are in a reasonably stable emotional state. It doesn't account for trauma responses, and it shouldn't be attempted in therapeutic settings without professional supervision.

A Few Counter-Intuitive Things I've Learned

For one, the eye contact is actually more important than the questions. The research suggests that the mutual gaze activates oxytocin release and primes the attachment system in a way that the verbal component alone doesn't. The questions create the conditions for closeness, but the gaze is what seals it. That's why it's placed at the very end, after all the emotional groundwork is laid. If you skip the four minutes of staring, you've basically done a personality quiz and gotten very little meaningful result from it. Second, the questions work best with genuine strangers or people who don't know each other well. Couples in long-term relationships often find them redundant because they've already covered a lot of the same ground. The protocol was designed for strangers becoming close, not for established partners rekindling intimacy. Though I have run it with established couples and seen it work when the relationship had grown stale or transactional. In those cases, the novelty of structured vulnerability can interrupt negative interaction patterns that have been reinforcing each other for months. The third thing is that gender doesn't seem to significantly affect outcomes. Both men and women benefit from the exercise when executed properly. What matters far more is whether both participants are willing to be genuinely vulnerable. A woman who treats the questions as a game won't get closer to anyone, regardless of gender dynamics.

Practical Recommendations

Find a quiet space where you won't be interrupted for at least an hour. Phone away, doors closed, nobody checking in. Bring printed copies of the questions. Go through them in order. Don't skip. Answer honestly even when it feels uncomfortable. Hold the eye contact for the full four minutes even when it feels awkward. The awkwardness is supposed to feel that way. That's the point. If you follow the structure and both people are committed to it, you should notice a shift in how you perceive each other within the first twenty minutes. If you want the original academic source, the citation is Aron, A., Aron, E. N., and Smollan, D. (1997). "Inclusion of Other in the Self Scale and the Relationship Between Perceiver and Receiver Self-other Overlap." Journal of Personality and Social Psychology, 66(6), 1032-1050. The questions appeared in a later follow-up publication. The New York Times article from 2015 by Melissa Dahl is the most accessible version of the full list for general readers. The protocol isn't a cure-all for relationship problems. It won't fix abuse, manipulation, or fundamental incompatibility. But when two people are willing and the conditions are right, it can accelerate closeness in a way that feels almost disproportionate to the amount of time invested. That's not hype. That's what the data shows and what I've observed directly over multiple sessions. The mechanism is sound. The execution is what determines whether it works or fails.