Someone you've known for twenty minutes asks, lightly, like it's a game: "What's the one thing people assume about you that you wish wasn't true?" You're answering before you've decided to answer. The question doesn't ask whether something like that exists — it assumes it does, and hands you nothing to do but fill in which one. That's the whole appendix in one line: a bank of questions built so the presupposition does the extraction, not the follow-up.
Buried inside the BOM's elicitation chapter, right after a run of legitimate self-test "Knowledge Check" questions (numbered 1 through 3), item 4 reads: "Here are 112 loaded questions you'll want to ask that can help you reveal someone's true self" — and the numbered list simply keeps counting upward from there, items 5 through 115, folding the entire question bank into what looks, at a glance, like a study guide.1 It's a structural oddity worth naming plainly rather than smoothing over: the source numbering is inconsistent (one entry, #34, is left over-numbered as "34. 30." in the original text, and item #112 cuts off mid-sentence — "When was the last time you tried to please someone who made you" — before the list resumes at #113 as if nothing happened). These are artifacts of the source file, not errors introduced here, and they're reproduced faithfully rather than silently repaired.1
Each question is built on the same chassis: presuppose a vulnerability, then ask only for its details. Not "do you feel like an imposter" (which invites a no) but "do you ever feel like an imposter" (which invites a story). Not "are you controlling" but "how much do you enjoy being in control." The yes/no branch is removed from the question before it's ever asked.
Read as a set rather than a list, the 112 questions cluster into a handful of recurring pressure points:
None of the 112 ask about facts. All of them ask about the subject's relationship to their own flaws, fears, and defenses — which is precisely what makes the bank valuable for elicitation and uncomfortable as a party game.1
You're not going to read someone all 112 questions in sequence — that would announce the operation instantly. The bank works as a pocket reference: pull one question that matches whatever vulnerability the conversation has already brushed against, ask it in a tone that matches "we're just talking," and let the presupposition inside the question do the work. If someone has just mentioned a breakup, "how do you cope with being betrayed" lands as sympathetic curiosity, not interrogation, precisely because the question assumes betrayal happened rather than asking if it did. The skill isn't memorizing 112 items — it's recognizing, in real time, which single question from this bank fits the crack that's already showing.
Evidence: The bank is presented as field-practitioner material with no external validation, sourcing, or testing described — it reads as Hughes' own accumulated question set rather than a researched instrument.1 [PLAUSIBLE — needs corroboration] [SINGLE SOURCE]
Tensions: [DOCUMENT — DO NOT ENDORSE] These are, functionally, therapeutic-intake-style questions repurposed for extraction rather than care. A therapist or a close friend asking "what's the worst way you've ever abandoned yourself" is working inside a consent structure the subject opted into, aimed at the subject's own benefit. An operator using the identical question on a stranger in an elicitation context has neither the consent nor the benefit-to-subject orientation — the same words, the same presupposition mechanism, aimed at leverage instead of healing. The book doesn't flag this distinction anywhere in the passage; it's presented as a value-neutral technique bank.
Psychology — The Shame Siren Technique. That page describes a structured method for naming small shame-triggering relational injuries before they compound — explicitly consensual, explicitly aimed at repairing the interpersonal bridge between two people who both want the relationship to survive. The 112 questions use the identical psychological lever (surfacing an unaddressed shame-point by naming it directly) but strip out both halves of the consent structure: the subject didn't ask to be repaired, and the person asking isn't trying to repair anything. The insight the pairing produces: shame-naming is not inherently therapeutic or inherently extractive — it's a neutral mechanism that inherits its ethics entirely from who initiated it, who benefits, and whether the other party knew the conversation was diagnostic at all.
Creative Practice — Memoir as Therapy: The Pennebaker Research. That page documents one of the more replication-resistant findings in social science: structured writing about emotionally significant, difficult material produces measurable therapeutic benefit, because it requires the subject to organize and process material they'd otherwise avoid. The 112 questions are built to trigger the identical cognitive act — surfacing and articulating avoided material — through spoken elicitation instead of private writing. The insight the pairing produces: the same act (verbalizing what you usually avoid saying even to yourself) that Pennebaker's research shows measurably heals a person when it's self-directed and private is, when redirected by someone else toward someone else's ends, the exact mechanism this appendix is teaching as a leverage-extraction tool. It is not a coincidence that both mechanisms work — they're the same mechanism, aimed in opposite directions.
Sharpest implication: The line between a caring question and an extractive one isn't in the words — every question in this bank is, verbatim, a question a good therapist or a close friend might plausibly ask. The line is entirely in who's asking, why, and what happens to the answer afterward — which means the question bank itself is ethically inert; the ethics live entirely downstream, in use.
Generative Questions: