The Dossier

How to use this

This is not a document to read. It is a document to answer. You write into it, over about ten weeks, and what comes out the other side is the thing you originally asked for: a full, honest picture of who you are, where you've been, what's blocking you, and where you're going — clear enough that a therapist, a coach, an AI, or a future version of you could pick it up and actually see you.

It has four movements and two guards.

The movements are the content. The guards are what keep the content honest — and they are not optional. A self-portrait without them is a monument, and monuments don't move.

Every question here is stolen, deliberately, from somewhere old or somewhere tested: Ignatius' nightly review, the Stoics' three questions, the Zen koans, the Sufi accounting of the soul, AA's fearless inventory, the interview attachment researchers trust most, the exercises therapists actually use. Where a tradition had exact words, the exact words are kept. Small grey notes tell you where each thing comes from, so you can go deeper anywhere you want.

And if ten weeks sounds like a mountain — it is, and you don't have to climb it. Start with The First Sitting below: ninety minutes that produce a complete dossier in miniature. Everything past it is deepening, not obligation.

Eight rules before you write a word

Each of these exists because the opposite reliably fails. They come from research, not taste.

The First Sitting — ninety minutes, complete

A second research pass on this document turned up one finding that changed its shape more than any other: across every kind of self-guided program ever studied, most people never finish. Workbooks done alone and unwitnessed get completed by a few percent of the people who start them. But a single well-designed session produces real, lasting, measurable change — and the most common number of sessions people attend of any program, anywhere, is one. So this document is built the way the best brief interventions are built: one sitting produces a whole dossier in miniature.

Block ninety minutes. Phone off. Answer these ten — by hand, or out loud into a voice recorder, which research says works at least as well as writing and is much faster. Two pages maximum per answer. When the timer ends, the answer is done, mid-thought or not.

1. My life in chapters — name them, one line each. 10 min
2. The best moment of my life so far, written as a scene. 10 min
3. The worst moment, written as a scene. 10 min
4. The turning point — when did I change, and what do I know now that I didn't before it? 10 min
5. Five adjectives for my childhood relationship with each parent — with one specific memory to prove each. Mark the adjectives I couldn't back up. 15 min
6. What do I want that I have never said out loud? 5 min
7. The one question in this whole document I'd most like to skip is ______. (Answer it or don't — but name it.) 5 min
8. If I knew I would die one year from today, what would I change about how I'm living — and why haven't I already? 10 min
9. How the next chapter goes, if it goes well — written in the present tense. 10 min
10. One thing I will do differently this week, one person I will tell, and the date. 5 min

Close with three sentences of advice to someone exactly like you. The best brief interventions all end this way, because explaining a thing to another person is what makes it yours — the researchers call it saying is believing. Then stop. You have a dossier. If you never open this document again, you did the part that matters most.

If you ever want five more, the research ranks these as the highest-yield in the whole book: the miracle question (opens Movement IV) · Naikan's third question — what troubles have I caused? · the downward arrow to your bottom sentence (§8) · the envy probe — whom do I envy, and what disowned want does it point to? (§11, §13) · and the three-person outside check (Movement V), which is the only question here that imports data you can't distort.

If you go on — how people actually finish

The completion research is unusually clear, and it ranks like this:

• Make it an appointment, not an intention. The specific if-then plan — "If it's Sunday at nine, I open the next section with coffee" — is the best-evidenced finishing technique in behavioural science, with an effect size most self-help would kill for. Add a second if-then for your predictable derailer: "If my answer feels shallow, I write 'shallow draft' and keep going."
• Get a witness. The single most protective factor against abandonment in every study of it — and it doesn't need to be a professional. A friend who receives each week's section. A standing call where you answer out loud. Alone-and-unwitnessed is the arrangement that fails; scheduled-and-witnessed is the one that finishes.
• Small doses, hard caps. Fifteen to twenty-five minutes per session, roughly weekly. Two pages, maximum. Stop on the timer — the thinking continues between sessions, and that quiet incubation is part of how the method works. Even two minutes counts; "I only have five minutes" is never a reason to skip.
• One section at a time. Never look at the whole mountain. Facing a hundred questions at once is a textbook case of choice overload — and the thousand-year-old Mussar answer is the same as the modern one: one thing, one week, in rotation, forever.
• Missing a session doesn't matter. Punishing yourself for it does. The habit research found a single missed day made no measurable difference — but self-criticism about lapses is what actually predicts quitting. Restart on the next Monday, and count sections finished, never streaks.

Where I'm From

The most boring layer, and the most reliable one. The parts of a self-portrait that feel most insightful — the explanations — are the least trustworthy. The parts that feel like paperwork are the ground everything else stands on. Build the ground first.

1 · The bare facts

Write these flat, like a passport. No interpretation allowed yet.

• Name, age, where born, everywhere you've lived and when.
• Who was in the house at each stage of childhood — who left, who arrived.
• Education and jobs, in order, with dates.
• Money, now: what you earn, spend, owe, and own. The real numbers.
• Body: diagnoses, medications, injuries, sleep. Substances — measured, not guessed.
• Every relationship that lasted six months, with dates and how it ended.
• Where your time actually goes: open your screen-time report and your calendar for the last 30 days and copy down what they say — not what you'd have guessed.

2 · The chapters — Dan McAdams' Life Story Interview, Northwestern University: the backbone of forty years of research on how lives become stories

Divide your life into chapters, as if it were a book. Title each one. Then write these eight scenes — and write them as scenes: where you were, who was there, what was said, what you felt at the time.

1. The high point — the single best moment of your life so far.
2. The low point — the worst. Take your time.
3. The turning point — a moment you can now see you changed.
4. A good childhood memory.
5. A painful childhood memory.
6. A moment you showed wisdom, however small.
7. A spiritual or mystical experience, if you've had one — described plainly, without defending or debunking it.

Then: your greatest life challenge and how you've met it. How the next chapter goes. And if the whole book has one theme, what is it?

3 · The five-adjectives test — adapted from the Adult Attachment Interview, developed by Mary Main in the 1980s, still the instrument attachment researchers trust most

This small exercise is the quietest and most powerful thing in the document.

1. Write five adjectives describing your childhood relationship with your mother.
2. Did you ever feel rejected as a child? What happened?

Now go back and check: for every adjective, did you actually produce a memory that shows it? Where you wrote "loving" and could find nothing — that gap is the finding. Don't fix it. Record it.

One honesty note: the real interview is given by a trained interviewer and scored by certified coders, and nobody can grade their own coherence — you will be the last person to hear yourself idealising. Mary Main's own team tried to build a self-scored version and abandoned it when it wouldn't validate. So this exercise cannot tell you your "attachment style," and don't let it. What survives self-administration is exactly one thing, and it's enough: the gap between the adjectives you reached for and the memories you could actually produce. That gap is real data — and one more reason this document eventually needs an outside reader.

4 · The inheritance — Murray Bowen's family-systems therapy and the three-generation genogram

Map three generations — you, your parents, your grandparents, the siblings and the people who functioned as family. On the map, mark:

• Addiction, mental illness, suicide, early death, divorce, estrangement, emigration, prison, ruin, secrets.
• The cutoffs — who stopped speaking to whom, the official reason, and the real one.
• The triangles — every place two people manage the tension between them by pulling in a third. Which triangle are you standing in right now?

The questions that stood here are answered by your record — see ISAAC-RECORD-ANSWERS.

5 · The unfinished stages — Erik Erikson, 1950, who mapped the eight tasks of a human life; run here backwards, as an audit

A century of developmental psychology says life sets you eight tasks, roughly in order, and each one passed leaves a strength behind. Go down the list and answer one question per row: did I get through this one — or am I still in it?

Almost everyone finds one conflict they are still fighting decades after its window closed — and it usually explains more than any diagnosis would. Which is yours? And which strength on the right never quite arrived? A person who never got Will has a different life problem than a person who never got Identity, and the two are constantly mistaken for each other.

One more, from the psychoanalyst Winnicott: everyone builds a false self — the performing self that secures approval — over a true self that acts from what's actually alive in you. Don't ask whether you do this; everyone does. Ask how much, where: what percentage of your work life is performance? Your marriage? Your friendships? Which number surprised you?

6 · Where I'm from, in one paragraph

Having done all of it: write one paragraph a stranger could read in thirty seconds and know where you come from. Not the facts — the shape of them.

What I Carry

The longest movement, and the one you'll want to skip parts of. Note which parts — the reliable rule of every confessional tradition is that the question you most want to skip is the one to answer first.

7 · What happened

Start with the standard list — before you were eighteen, did you experience: emotional abuse; physical abuse; sexual abuse; emotional neglect; physical neglect; parents separating; a mother treated violently; addiction in the house; mental illness or a suicide attempt in the house; someone in the house imprisoned?

Do not total it into a score. The count predicts health at the level of whole populations and tells you almost nothing about one person — it ignores everything that matters: how bad, how long, how young you were, and who loved you through it. Robert Anda, who co-built the instrument, has said flatly that it "was designed to research, not screen," and warns that people with a high number attribute too much to it — filing their life under the past instead of asking what they need to do now. So: no number, no threshold. The items are doors. Walk through the ones that are yours, then write what the list can't hold:

1. What happened to you that has no name and fits no box?
2. What happened that you have never told anyone?

8 · The sentence under everything — the "downward arrow" of Aaron Beck's cognitive therapy, the most-tested school of psychotherapy there is

Take a recurring painful situation. Write the surface thought — "I blew the presentation." Then ask, again and again: "And if that were true — what would that mean about me?" Five, six, seven steps down, you hit something that won't reduce any further. That is a core belief, and yours has been running quietly for decades.

Do this for four different situations. Nearly everything lands in one of three families:

• Helpless — I'm weak. Trapped. Out of control. A failure.
• Unlovable — I'm too much. Not enough. Defective. Bound to be left.
• Worthless — I'm bad. I don't deserve what others deserve.

Which family is yours? Write the exact sentence, in the words you actually use against yourself. Then: when did you first believe it — and who taught it to you?

1. Write down, verbatim, the last five harsh sentences you said to yourself this week. Next to each, name the trick it is playing: mind-reading, fortune-telling, all-or-nothing, "should", discounting the good, catastrophising, taking it personally. Which trick runs most often?
2. "They'll think I'm stupid / crazy / a fraud" — how many times has that prediction actually come true when you checked? List the checks. (You once wrote: every meetup I dreaded went well, zero counterexamples.)
3. Which sentence do you believe when you're low that you don't believe when you're high — and the other way round? Write both. Which one is telling the truth, and how would you know?

9 · The old patterns — Jeffrey Young's eighteen "schemas," mapped in the 1990s from thousands of therapy hours; here in plain words

These are the deep grooves cut in childhood when a need went unmet. Read the list slowly. Mark the ones that make your stomach drop. For each mark, write one scene from the last year where it ran the show.

From homes where connection failed: expecting everyone close to eventually leave · assuming people will hurt or use you if given the chance · believing your need for warmth will simply never be met · feeling that if anyone saw you fully, they'd go · feeling different from every group, everywhere.

From homes that never let go: feeling unable to manage ordinary life alone · living braced for catastrophe · having no self separate from a parent or partner · a settled certainty of being inadequate next to your peers.

From homes with no walls: feeling the normal rules don't apply to you · being unable to sit with frustration long enough to finish anything.

From homes where love had a price: silencing what you want to keep the peace · giving until it hurts and feeling guilt the moment you stop · needing applause to feel real.

From homes that were cold or unsafe: seeing only what will go wrong · holding all feeling in, always · nothing ever being good enough, including you · the belief that mistakes must be punished — especially yours.

For your three strongest: what does each protect you from feeling? What would you have to feel if you set it down? That's the trade you've been making. Is it still worth it?

These eighteen are a therapist's vocabulary, not a diagnosis. Treat every groove you marked as a hypothesis about yourself — never a label.

1. Finish these without thinking: "I'm the kind of person who ___." "People like me don't get to ___." "It's too late for me to ___." "If I stop ___, then ___." "I'll be allowed to ___ once I ___." Now mark the ones you would let a friend keep.
2. Which rule did you make as a child to stay safe that still runs you now — "be useful or be left," "be special or be nothing," "don't ask, they'll say no"? What is it protecting you from today, and what is it costing?
3. Name a belief about yourself that nobody outside your own head has ever confirmed. Then one that several people have contradicted and you kept anyway.

10 · The parts — Richard Schwartz's Internal Family Systems, built in the 1980s from listening to what his patients' inner voices actually said instead of pathologising them

Everything you do that you can't stop doing is being done by a part of you with a job. This model names three kinds: managers, who run your life in advance so nothing goes wrong (the perfectionism, the planning, the pleasing); firefighters, who arrive after pain breaks through (the scrolling, the drinking, the rage, the sudden exit); and exiles — young parts, frozen at the age something happened, holding the original hurt.

Pick one of each and interview it: Where does it sit in your body? How old is it? What does it look like? What is it afraid would happen if it stopped?

The hinge of the whole method is one question: "How do I feel toward that part?" If the answer is anything but curiosity or compassion — if it's disgust, or fear, or "I want it gone" — you're not looking at the part. You're inside another one. Find that one first.

And the question that unlocks every compulsion in this document: what is this behaviour protecting me from feeling?

Same caution as the patterns above: "parts" are a way of talking that many people find unlocking, not a proven map of the mind — and if the young part you find is holding serious trauma, that is exactly the place to bring in a professional rather than dig alone.

1. Picture yourself at six in India, at ten on the plane, at twelve at the airport. What does each boy need to hear from you now? Write it to each of them, in the second person.
2. What did that boy love doing before anyone was watching or grading — and when did you last let him do it for an hour?
3. What does he still believe about why he was sent away, and about what he has to be to get kept? Do you believe it? Say it to him straight.
4. When you feel small, ashamed or panicked, how old do you feel, and where in your body is that age? What did you need from an adult at that age that you are still trying to get from the people around you now?

11 · The shadow — Carl Jung, for whom the disowned self was not an enemy but a buried inheritance

1. What trait in other people makes you angry out of proportion? Not disapproval — the disproportion. That's the tell. Where does that trait live in you?
2. The half everyone forgets — the golden shadow: what greatness do you project onto the people you admire, because claiming it yourself would make you responsible for using it?
3. What are you most afraid is true about you? Write it as one sentence beginning "The truth is I'm ___." Then, honestly, the evidence for and the evidence against.
4. Which parts of yourself stay out of sight in the community, out of sight from Ilya, from your mother, from your father? Make the list per audience. The part nobody gets to see is the shadow's address.

12 · The ledger — Step Four of Alcoholics Anonymous, written 1939: "a searching and fearless moral inventory." Still the best free honesty exercise ever designed, in or out of recovery

Resentments. Four columns. The first three anyone can do; the fourth is the exercise.

Fears. Each fear — why you have it — what it has cost you — what you did because of it.

Sex and conduct. Where were you selfish or dishonest? Whom did you use? Did you stir jealousy or bitterness you never answered for? Where were you at fault — and what should you have done instead?

Harms. The names. Not categories — names. Including people you never met who were affected anyway.

Omissions. The visit you didn't make. The thing you didn't say. The person you watched drown quietly.

Amends not made — and what making them would cost you that you don't want to pay.

1. Ten sentences beginning "I am angry that…" — at least ten, no explaining, no forgiving, no lesson at the end.
2. What have you lost that you never grieved — a person, a place, a version of yourself? Pick one and write it a goodbye, not a lesson.

13 · The questions you'd rather skip

This is the section the document exists for. They work because each names a person, asks about the wish rather than the act, and asks what the bad thing gives you — before ever asking why. Pick the group you least want to open. Answer three. Stop for the day.

Shame — from Brené Brown's shame research, Donald Nathanson's "compass of shame," and two thousand years of the confessional

1. Finish the sentence: "I don't want anyone to think I'm ______." Now — who exactly must never think that, and what do you do daily to make sure they don't?
2. Who currently knows the worst true thing about you? If no one does — what does carrying it alone cost you per day?
3. The earliest scene you remember of being made to feel disgusting: where were you, how old, whose face?
4. What are you pretending not to know about yourself?

Envy, resentment, contempt — from Melanie Klein, who saw that envy wants to spoil rather than to have; from Nietzsche on resentment; and from the psychology of secret superiority

1. Whose happiness do you need to be slightly less than it is?
2. What are you doing to your own life right now in order to punish somebody?

Desire — from the sex therapists Jack Morin (the "core erotic theme") and Emily Nagoski (brakes and accelerator), and from Esther Perel

1. What actually turns you on — not what you'd say on a podcast?
2. Take your three most intense erotic experiences, real or imagined. What do they share? What is the scenario underneath the scenarios?
3. What early powerlessness does that scenario reverse? (Sex researchers call this your core erotic theme — the wound your desire keeps solving.)

The people in the house — from Arthur Aron's 36 questions, Terry Real's work on the one-up/one-down game, and John Gottman's forty years of watching couples fight

1. Write "I resent you for ______" and "I appreciate you for ______" to the same person, five of each. Notice which five were harder.
2. Your mother, as a person: what do you actually know about her fears, her own faith, what she wanted at 27? Write what you know, then the questions you have never asked her.
3. What do you want from a partner — five things — and which three are non-negotiable? Then the harder half: what can you honestly offer someone today?
4. Do you want children? Yes or no, today, with the reason — and what kind of father would you be, named against what yours did?

Money, work, status — from the moral inventories, and from Plato's oldest thought experiment: what would you do if no one could ever see?

1. What are you doing purely for status while telling everyone — including yourself — that it's for something else?

Compulsion — the order of these questions matters; inventories that start with the costs get lied to

1. What would you lose if you stopped? Almost no inventory asks this, and it predicts whether you'll actually stop better than anything else.
2. Is this document itself an evasion? When did you last change a behaviour, rather than collect an insight?

The unforgiven — from the research on moral injury in soldiers and doctors: the wound of having violated your own code

The questions that stood here are answered by your record — see ISAAC-RECORD-ANSWERS.

Close the movement with one commitment. Not an insight — an act. One thing done differently this week, with a date. Write it down.

What I Love

Everything so far has been about what's in the way. This is the part people rush, and shouldn't: what's actually on the other side. The childlike part. The part that was there before the armour.

14 · Values — tested, not listed

Anyone can list values. The test is whether your calendar and your bank statement agree with the list.

1. Force-rank your top five values. (A free values card deck helps — sorting eighty cards down to five is harder and more honest than naming five from air.)
2. For each of the five: when did you last pay a real price for it? A value you've never sacrificed anything for is not a value. It's a decoration. Tag it ASPIRED and keep it — but know which it is.
3. The contradiction hunt: list every place your stated values and your actual choices point in opposite directions. Don't fix anything. Just see the list. It is more informative than the values themselves.

15 · Strengths — tested the same way

1. Then four lists, strengths first and longest — the order is a method, not politeness: what you have and want to keep · what you have and want to change · what you lack and want · what you're good at and resent being good at. That last list holds a surprising amount of buried truth.
2. Without looking at any record: ten things you are genuinely good at. Then cross out the ones you would hesitate to say out loud to Aaron or to your father. What's left, and why the crossings?
3. What do people keep coming to you for that you dismiss as "not real work"? That is usually a strength you don't count.
4. Which weakness do you secretly protect because it feels like part of who you are — the overthinking, the restlessness, the sensitivity? What would you lose if it went?
5. When are you the best version of yourself — who is in the room, what is the task, what time is it? And the worst version? Compare the two lists; that is your operating manual.

16 · What you actually enjoy

You are a poor witness to this from the armchair. Collect evidence.

1. If nobody could ever know what you chose — no father, no Jahn, no audience, no one watching — how would you spend the next year? Write one ordinary week of it, hour by hour.
2. What do you want that you are embarrassed to want — money, a beautiful wife, fame, to be adored, to be left alone? Write it plainly, then write where the embarrassment comes from.
3. Three things you want out of life that have nothing to do with being seen, producing or proving. If you can't find three, sit until you can.
4. A good ordinary Tuesday, five years from now — not the best day, a normal one. Who is in the house, what did you make, what did you eat, who did you call, where did you sleep?
5. What are you saying "later" to, and what exactly would have to happen for it to become "now"?

17 · Five questions that come in sideways — Mark Savickas' Career Construction Interview, used by counsellors worldwide. None of the five asks about you directly. That's why they work.

The questions that stood here are answered by your record — see ISAAC-RECORD-ANSWERS.

18 · The animal and the ground

The questions that stood here are answered by your record — see ISAAC-RECORD-ANSWERS.

19 · The soul, honestly

Whether or not you believe anything.

1. What do you actually believe happens when you die? If the honest answer is "I don't know," then: what do you behave as if you believe?
2. And the oldest question of calling: what is the song only you can sing? What goes unsung if you don't?

20 · A fast reading of the present

Two short checks from the well-supported end of the science. A paragraph each.

Six dimensions of a life that's working — rate yourself, with one piece of evidence each: accepting who you are · having a purpose · still growing · warm relationships · running your own life · handling your world.

The three needs that predict wellbeing across every culture studied: autonomy (am I choosing, or complying?) · competence (am I capable here, or performing capability?) · relatedness (am I known — or merely liked?). Score each for work, home, friendship, body. The pattern across domains is the finding.

And hold the result that reliably surprises people: getting extrinsic goals — money, image, status — does not raise wellbeing, even when you get them. If your plan for feeling better is on that list, better to know now than in ten years.

Where I'm Going

Every tradition in this document, without one exception, uses death as the clarifying instrument. Not morbidly — practically. It is the only pressure that reliably separates what you want from what you were told to want.

The miracle question — Steve de Shazer & Insoo Kim Berg, solution-focused therapy, 1980s. Pound for pound, perhaps the best-evidenced single question in the clinical literature — which is why it opens the movement

21 · Why you haven't changed already — Robert Kegan & Lisa Lahey's Immunity to Change map, Harvard, 2009. If you do one exercise in this movement, do this one.

Four columns.

Column 3 is where the light comes on. You are not failing at your goal. You are succeeding, flawlessly, at a goal you never admitted having — usually staying safe from some old feeling. Once both commitments are visible, the problem stops looking like weak will and starts looking like what it is: a system in perfect balance. Then design one small, safe test of the big assumption — an experiment with a written prediction, not a resolution.

Honesty note: this map has never been put through controlled trials — it is a brilliantly designed mirror, not a proven treatment. It stays because nothing else makes the shape of stuckness this visible, and because what you do after it (the small test) is just a behavioural experiment, which is well proven.

22 · Death, the teacher

1. Three questions, in order — don't skip to the third (George Kinder, the founder of financial life planning, who asks these of every client before any talk of money): (a) You are financially secure for life. How do you live? What changes? (b) You have five to ten years left, in full health, with no warning of the end. What do you do with them? (c) You have one day. Ask only: What did I miss? Who did I not get to be? What did I not get to do?
2. The five regrets of the dying (recorded by a palliative nurse, Bronnie Ware): living the life others expected instead of one's own · working too hard · never saying what one felt · losing touch with friends · not letting oneself be happier. Which is already yours? Which is your family watching you commit right now?
3. At 80, looking back: will you regret trying this and failing — or never trying?

23 · Three lives — the Odyssey Plan, from Stanford's Designing Your Life

Write three different five-year futures, each in real detail — a week in that life, not a job title:

• Life A — the path you're on, extended honestly, including the parts you'd rather not extend.
• Life B — what you'd do if Life A vanished tomorrow: the industry died, the relationship ended, the body changed.
• Life C — what you'd do if money and other people's opinions were no object.

For each: what would you need to learn? Give up? What one question would tell you if it's real? Then the actual question: what is Life C telling you that you keep refusing to hear?

24 · Fear-setting — Tim Ferriss' worksheet, which is the old Stoic premeditation of evils with modern handwriting

For the thing you're avoiding: Define the worst that could happen — ten to twenty specific items. Prevent — for each, what reduces the odds? Repair — if it happened anyway, how would you recover, and who would you call? Then the page almost no one writes: the cost of doing nothing — emotional, physical, financial — at six months, one year, three years. Write that page in the second person. "You will be…"

25 · The letter you leave — the ethical will, a Jewish tradition a thousand years old; also the one exercise here with real clinical-trial evidence, from dignity therapy

Write it as though it will be read aloud one day. It probably will.

1. What do you want the people who come after you to know about you?
2. What have you learned about life that you want to hand on?
3. What are you most proud of? Which roles mattered most?
4. What still needs to be said — or said again?
5. Is there forgiveness to ask? Ask it here, in writing.
6. And one sentence that does an unreasonable amount of work: what do you need someone to know about you as a person, in order to take care of you properly?

26 · The commitments

• Three things you will do in the next 90 days, each with a date.
• One thing you will stop.
• One thing you will say, to one named person.
• One person you will ask to hold you to all of it.

And shape each commitment the way the strongest goal research says to (Gabriele Oettingen's "mental contrasting," NYU — one of the few techniques in this document with large trials squarely behind it): name the wish · picture the outcome, vividly · then name the real obstacle — the inner one, the avoidance or the fear, never "no time" · then write the plan as an if-then: "If [the obstacle shows up], then I [specific act]." Wish, outcome, obstacle, plan. Positive visualising alone measurably does nothing; it's naming the obstacle that turns a wish into a vector.

1. The three goals you would be ashamed to fail at publicly, and the three you would only be sad about privately. Which list is actually yours?
2. What do you want from Return to Eden for yourself, in one honest sentence — and what would you need to have in writing to stay?

The Case Against This Document

Everything above was written by the same mind it describes — and that mind's oldest, most practised skill is fooling itself. Biologists think self-deception evolved precisely because the most convincing liar is the one who believes himself. So the document now turns on its author. Write this last, in one sitting, unpolished.

1. Where is this document flattering me? Find five places. Look for suffering that makes you interesting; flaws that are secretly compliments ("I care too much"); villains drawn too simply; a redemption arc pasted over something that hasn't actually redeemed.
2. What did I leave out? Name what didn't make it in, and why. The omissions are the document.
3. Which theories are dressed as facts? Every sentence that starts "the reason I'm like this is…" — retag it INTERPRETED. However true it feels. Feeling true is not evidence; it's the signature style of every confabulation you've ever produced.
4. What would the three people who like me least say about all this? In their voices. Full strength. No rebuttal.
5. The prediction test. Write five specific predictions about your own behaviour over the next 90 days, with odds. In 90 days, score them. This is the only thing in the document that turns self-knowledge from a feeling into a track record.
6. The outside check. Ask three people who know you and have no reason to flatter you: "What do I do that I don't see? What would you tell me if you knew I wouldn't get defensive?" Write down exactly what they say, including what you disagree with, without comment. This is not optional. The research on self-knowledge is blunt: you are the world's best witness to your own anxiety — and among the worst to your own likeability, charm, and effect on a room. Half this document you cannot see from where you're standing.
7. Start the anti-document. A running page, kept forever: things I said about myself that turned out to be false. Add to it yearly. In time it becomes the most valuable page in the file.

Instructions to Whoever Reads This

Written for the therapist, the coach, the friend — or the AI. If it's an AI, this section is the only thing standing between you and a very articulate mirror.

27 · The standing instruction

Put this at the top of your dossier and again at the bottom — not the middle. Language models attend to the beginning and end of a long document and reliably lose the middle.

28 · Two documents, not one

Don't paste the whole dossier into every conversation. It fails quietly — the model doesn't complain, it just stops following the rules buried in the middle, and the rules are the most important part.

• The operating card — 300 to 600 words, always on (custom instructions, or the top of a project). Not your self-knowledge; your interface: one factual line about who you are · what you're working on now, dated · how to talk to you · a long, specific, negative list of what you do NOT want ("do not open with praise," "do not tell me I'm insightful," "if my premise is wrong say so first") · your known biases, with instructions to name them on sight · last-updated date. The "do not" list is the highest-value section anywhere in the system — it is the only text that pushes against the model's training.
• The dossier itself — long, retrieved on demand, in a project knowledge base or a file you point to deliberately.

29 · Maintenance

• Date everything. Put volatile things — current mood, current projects, current relationships — under an expiry date. They're what you'll write in most detail, and what goes stale first. A dossier written during a breakup reads as timeless fact for as long as it sits in a context window.
• Keep versions. Reread last year's before writing this year's; the difference is the real document.
• Quarterly: update. Yearly: rewrite Movement I from scratch without looking at the old one, then compare.
• Privacy, plainly: this is the most sensitive document you will ever write. Before it touches any service, check that provider's current data-training and retention settings — and know that deleting a conversation does not always delete what a system learned from it. A local, encrypted copy is the only version fully under your control.

The Daily Practice

The dossier is a build. This is the maintenance — the same traditions, folded down to ten minutes a night. Pick one. Running several is a way of running none. A 250-question document completed once is worth less than one honest question asked five hundred times.

Watch the motive. Any of these becomes rumination the moment it is driven by shame instead of curiosity — and the same ten minutes will then make you reliably worse. If it starts to feel like a nightly tribunal, stop for a week.

The Path — ten weeks

You already did week zero — the First Sitting. Everything below is the full build, and the order is the method: reliable ground first, one energising week, then the deep water, then death and the future, then the guards. Do it 15–25 minutes at a time, ideally with a witness — a friend who gets each week's pages, or a standing call where you answer out loud.

Then stop. Put it away for three months. Score your predictions. Come back and mark what has already turned out to be false. And keep the one daily practice going — that is the part that compounds. The document is scaffolding; the ten minutes a night is the building.

What's Real and What Isn't

You will be tempted to bolt more instruments onto this. They are not of equal quality, and knowing which is which is itself self-knowledge.

Where all this comes from

The same research brief was run through Gemini, ChatGPT, and Claude's deep-research modes, then checked and extended against primary sources in six parallel research passes — the original texts of the traditions, the published interview protocols, and the psychology literature on introspection, memory, shame, and AI sycophancy. Where the three models agreed independently, confidence is high; where only one said something, it was verified before inclusion; where one was wrong (one model misquoted the five sideways questions in a way that would have silently broken them), the error is recorded in the project files along with the full raw reports.

Two validation passes then stress-tested the document. The second — a targeted deep-research run through Claude alone, at the author's request — forced this round of changes: the ACE score lost its number entirely (its own co-creator says it was "designed to research, not screen"); the five-adjectives test now states outright that it cannot tell you an attachment style, since Mary Main's own team tried and failed to build a self-scored version; parts and schemas are labelled vocabulary rather than diagnosis; the Miracle Question, Zengzi's three questions, the Five Remembrances, and al-Ghazālī's full three-stage muhāsaba were added on the strength of their evidence and pedigree; the Yoruba ori was corrected to the scholars' soft-deterministic reading; and the AI section was rewritten to say plainly that anti-sycophancy instructions are a hope rather than a safeguard. An earlier pass had already forced these corrections: the "94% divorce prediction" folklore was cut (contempt's corrosiveness survives; the prediction number doesn't); the Immunity map was relabelled as a brilliant mirror without outcome trials; the redemption finding gained its counterweight (agency moves first, and acceptance is a legitimate ending); the five-adjectives exercise now says plainly that no one can score their own coherence; and mental contrasting — wish, outcome, obstacle, plan — was added, because it turned out to have stronger evidence than half the set. The same pass surfaced the completion research that produced the First Sitting: the best-supported way to answer a document like this is one complete short sitting, then small witnessed doses — not a heroic march through two hundred questions.

The Point of the Thing