Editorial commentary

The 36 questions, twenty-some years on: what they got right, what they missed

The Aron 36 questions are the only widely-known piece of structured-prompt research in the relationship-conversation category. They are also widely misused: copy-pasted into listicles without context, repeated verbatim by content sites that did not write them, and pulled out of order in ways that defeat the structure they were designed around. This page is the editorial commentary the listicles do not write, and a recommendation that you read the original directly rather than via paraphrase.

What the 36 questions actually are

The protocol comes from Arthur Aron, Edward Melinat, Elaine Aron, Robert Vallone and Renee Bator's 1997 paper “The Experimental Generation of Interpersonal Closeness: A Procedure and Some Preliminary Findings”, published in Personality and Social Psychology Bulletin. The study, run at Stony Brook, asked pairs of strangers to spend 45 minutes asking each other 36 questions in three escalating sets of twelve, followed by four minutes of mutual silent eye contact.

The popular framing is that the questions can “make anyone fall in love with anyone”, which is not what Aron concluded. What the paper actually showed was that structured, mutual, escalating self-disclosure produces measurably greater post-conversation closeness than small-talk control conditions; it did not claim that closeness was the same as love or that the effect was permanent. The framing slipped, predictably, when the study went viral via Mandy Len Catron's 2015 New York Times Modern Love column “To Fall in Love With Anyone, Do This.”

The canonical version of the questions is published, free, by Greater Good in Action at the Greater Good Science Center, which credits the original paper: ggia.berkeley.edu/practice/36_questions_for_increasing_closeness. If you want to use the protocol, read the questions there, in order, in the structure the paper described. We are deliberately not republishing the list on this site; the canonical source is one click away and the editorial value of this page is the commentary, not the content theft.

Why the structure works

Three things about the 36 questions are doing the heavy lifting, and any of them can be borrowed even outside the original protocol.

First, escalation. The first set is light, the second set is reflective, and the third set is genuinely vulnerable. Each set sets up the next one; jumping straight to the third set with someone you do not know well rarely produces the same effect, because the early reciprocal disclosure has not happened.

Second, mutuality. Both partners answer every question. The asker is not interviewing the answerer. This is the part of the protocol that is most often violated by listicle versions, which present the questions as an interrogation tool rather than a mutual disclosure exercise.

Third, the silent eye contact at the end. The four-minute closing was not decorative; the paper specifically notes its contribution. Most listicles drop it, and most users skip it. It is the part of the protocol the prose cannot capture.

What the 36 questions miss

The protocol was designed for strangers, in a laboratory, in 1997. Most of the readers of this site are not strangers to each other. The original 36 work very well as a first-meet exercise; they work less well as a second-anniversary one. Three specific gaps for modern in-relationship use:

They do not address shared history. By design. Several of the third-set questions ask for vulnerable disclosure of past experience, but none ask “what is something you wish I asked you about more often, given how long we have been together?” or any other prompt that requires the relationship to already exist. For couples already past the early-disclosure phase, the protocol is structurally pre-relationship.

They do not address modern relationship contexts. Long-distance, prior relationships, blended families, neurodivergence, the realities of being-already-coupled, the realities of asking-while-in-a-relationship. The 1997 paper could not have anticipated long-distance prompt use over voice memo, or how the structure would be used by couples considering moving in together. Adapting it for those contexts is an editorial extension; the canonical questions stop at the original purpose.

They do not address skip rules. The paper assumes both participants will answer every question. Modern users sometimes need to skip; the protocol does not have a graceful way to do that, which means a partner who skips a question implicitly signals more than they intended to. Newer protocols (the Gottman Sound Relationship House work, for instance) build skip rules in deliberately.

How to use them today

Three recommendations, based on how the protocol actually performs in long-form use.

One: use the original, in order, once, as a structured evening. Reading the canonical Greater Good page, going through the three sets across about forty-five minutes, and ending with the silent contact closing, is genuinely worth doing once in a relationship even if you have been together for years. It does something that piecemeal asking does not, and it does it because of the structure, not the individual prompts.

Two: borrow the structure, not the questions, for ongoing use. The escalation logic is the borrowable part. When you are using the prompts on this site, ask one light, one reflective, one deeper, in that order, across an evening. Reverse-engineering the Aron escalation onto stage-organised prompts is what most of this site is implicitly doing.

Three: do not use them as ammunition. The protocol works because both partners are mutually disclosing in good faith. Used to corner a partner into a particular admission, or as a substitute for an honest conversation that should have happened directly, the protocol fails the same way any tool used in bad faith fails. The mutuality is not optional.

The canonical version, linked out

Read the original questions, in the right order, on the Greater Good in Action page: 36 Questions for Increasing Closeness. If you want the original 1997 paper, the citation is Aron, A., Melinat, E., Aron, E. N., Vallone, R. D., & Bator, R. J. (1997). “The experimental generation of interpersonal closeness: A procedure and some preliminary findings.” Personality and Social Psychology Bulletin, 23(4), 363–377.

When you come back from reading the canonical version, the deep questions page and the six-months page are the parts of this site where the Aron escalation logic most clearly extends into stage-organised prompts.