Mary Hendra asks every guest the same opening question: when did you last play? It’s the kind of question most podcasters wouldn’t lead with. It’s too small. Too immediate. Not the right level for an interview about somebody’s work.
“When did you last play — that’s our body that responds. We have that visceral memory of when we played, whether it was with a child, or outside, or with paint.”
The question doesn’t ask for a definition. It doesn’t ask for a position on play, or a story about why play matters, or a reflection on the role of play in adult life. It asks for a memory. And the memory shows up before the guest can frame it.
Most interview questions retrieve the curated version of the guest. Tell me about your work. What’s your story. What do you stand for. These produce articulate answers and clean tape — partly because they ask for material the guest has already prepared. The guest has thought about what they want to say about their work, refined the version they want represented, and is ready to deliver it. The interview becomes a transmission of the prepared version.
If these field notes spark your thinking, you'd like the Podtalk Community — a small forum where indie podcasters talk craft, not hype.
https://podcaster.community/
The curated version isn’t false. It’s one layer. It’s the layer the guest has decided is okay to share, and it produces a particular kind of conversation — one that stays at the level of self-description, where everything the guest says has already been through the filter of “is this how I want to be seen?”
Hendra’s question works against that level. Not by tricking the guest, and not by pushing them somewhere uncomfortable. Just by asking for something that doesn’t route through the curator. The body produces an image — a child laughing, a brush in your hand, an afternoon — and the image is there before you’ve decided what it means or whether it fits.
The most striking version of this in Hendra’s experience involved a guest who runs games for a living. Before the interview, he told her he wasn’t a playful person and suggested they should talk first to decide if he was a good fit for the show. She heard that and knew immediately he had to be on. They talked about his work, his company, the games he runs. Halfway through, Hendra came back to the opening question: when did you last play? He laughed, paused, and said: actually, I played this weekend. Let me tell you about that.
The “I’m not a playful person” was a position. The weekend was evidence. They disagreed, and the question made the disagreement visible. Hendra didn’t have to argue the man into a more accurate self-description. She just asked for a memory, and the memory contradicted the position before either of them had to say anything about it.
This is a different theory of what a question is for. The standard interview question extracts material the guest already has prepared. The body-access question reveals material the guest is carrying but hasn’t framed. The first kind produces clean answers. The second kind produces something the guest is discovering at the same time as the listener — which is harder to fake, harder to perform, and often more interesting.
The complication is that the question itself doesn’t generalize. “When did you last play?” works for Hendra’s show because play is the show’s subject. Borrowing the question for a podcast about, say, project management would just produce a confused guest. The deeper move isn’t to copy the question. It’s to find the question for your show that asks for evidence rather than position. The question that gets at memory rather than self-description. The question whose answer the guest can’t have rehearsed because the guest didn’t know they’d be asked to look for it.
Most podcasting prep doesn’t build for that. It builds for the curated answer — the one the guest is ready to deliver. The harder prep is to design at least one question that the guest can’t deliver, only discover. The most interesting moments in a recorded conversation might be the ones the guest didn’t know they had until you asked for them.
This field note references the Podtalk episode “Play with Mary Hendra,” published July 1, 2022.
This work was produced using AI language models directed through an editorial system designed by Craig Constantine. The author selected all source material, designed the creative framework, directed the editorial process, and made all acceptance and revision decisions. The prose was generated by AI under sustained human editorial direction.
This is a companion discussion topic for the original entry at https://podtalk.show/asking-for-evidence-instead-of-position/