Ask Better Questions, Find Richer Stories
The Question That Opens a Door
Imagine you’re sitting at the kitchen table, looking through an old photograph.
Your son is standing beside a car.
You remember the car.
You remember how excited he was about it.
And you remember that he spent an extraordinary amount of time working on the thing.
You begin telling the story.
“He absolutely loved that car. I swear, he was always doing something to it.”
Your AI responds:
“Did working on the car help him develop his confidence and independence?”
You stop.
Well, perhaps.
But that’s not really what you were talking about.
You were remembering the car.
The projects.
The funny incidents.
The things he tried.
And the rather remarkable ability that particular vehicle had to empty his wallet.
Now imagine the AI had asked a different question.
“What was the first thing he changed on it?”
Suddenly, a memory appears.
You remember him bringing home a box of parts.
You remember asking what they were for.
You remember his explanation, which made absolutely no sense to you at the time.
And you remember what happened when he finally installed them.
Now you have a story.
A real one.
Not an interpretation about confidence.
Not a psychological explanation.
A specific experience you can describe.
That’s the power of a useful question.
The best questions don’t tell you what your memory means. They give you room to remember what happened.
And that’s what we’re going to learn in this chapter.
Because as your collection grows, the questions you ask, and the questions your AI asks, can make a tremendous difference.
Some questions encourage genuine recollection.
Others accidentally suggest details that weren’t part of the original memory.
Some lead to unexpected stories.
And some are so broad that you don’t know how to answer them.
Learning the difference is an important part of the Meistro Method.
And when the subject is someone you love, it becomes particularly valuable.
The Difference Between Asking and Suggesting
Let’s begin with a simple example.
Imagine you remember your son visiting his grandmother’s house.
You tell your AI:
“He used to spend a lot of time over there when he was little.”
The AI asks:
“Did he love baking cookies with his grandmother?”
Now stop.
Where did the cookies come from?
You never mentioned baking.
You never said they cooked together.
You only said he spent time at her house.
The question introduced an event that may never have happened.
You might answer:
“No, I don’t remember them baking together.”
That’s fine.
But repeated questions containing specific suggestions can create confusion, especially when you’re discussing events from many years ago.
The AI may accidentally encourage you to consider details that weren’t part of your original recollection.
Now consider a different question.
“What do you remember them doing together?”
That’s much better.
It doesn’t suggest a particular activity.
It gives you room to supply the information.
You might remember gardening.
Watching television.
Working in the garage.
Playing cards.
Or sitting on the porch.
Whatever comes to mind is yours.
The question hasn’t supplied the answer.
This distinction is called the difference between an open-ended question and a leading question.
It’s one of the most important concepts in careful memory collection.
Open-Ended Questions
An open-ended question invites someone to answer in their own words.
It doesn’t require a simple yes or no.
And it doesn’t tell the person what they’re expected to remember.
For example:
“What do you remember about that afternoon?”
“Who was there?”
“What happened next?”
“What was he interested in at that age?”
“What do you remember about their friendship?”
These questions leave room for the storyteller’s own recollections.
They allow unexpected information to emerge.
And they can produce richer material than questions that suggest a particular answer.
Let’s imagine a fictional conversation.
Mother: “He loved going to Grandpa’s workshop.”
AI: “What do you remember him doing there?”
Mother: “Mostly getting in the way, if I’m being truthful.”
AI: “Was there a particular time you remember?”
Mother: “Oh, yes. He once decided he was going to organise all Grandpa’s tools.”
Now we have the beginning of a story.
The AI didn’t tell the mother that her son loved woodworking.
It didn’t suggest that the workshop represented a special emotional bond.
It simply asked what she remembered.
And she supplied the detail.
That’s exactly what we’re aiming for.
Leading Questions
A leading question suggests or assumes part of the answer.
For example:
“Was he always the most adventurous child in the family?”
“Did his grandfather inspire his interest in engineering?”
“Was that one of the happiest days of your life?”
“Did he feel proud when he finished building it?”
Each question introduces an interpretation.
Perhaps the interpretation is accurate.
But it hasn’t yet been established.
And if we’re trying to preserve genuine memories, we should be careful not to place answers into the storyteller’s mouth.
Let’s look at another example.
You say:
“He spent hours trying to get that old motorcycle running.”
The AI asks:
“Was he frustrated because he couldn’t fix it?”
Perhaps he was.
But perhaps he was enjoying himself.
Perhaps the motorcycle wasn’t particularly important.
Perhaps he thought the entire situation was hilarious.
A better question might be:
“What do you remember about how he approached the problem?”
Or even more simply:
“What happened while he was working on it?”
Those questions allow the storyteller to describe the experience.
If frustration was part of the memory, it can emerge naturally.
If not, we haven’t introduced an unsupported emotion.
A question can sound thoughtful while still containing an assumption.
Learning to recognise that difference will make your memory collection more reliable.
Why This Matters More Than You Might Think
Human memory is not a perfect recording of the past.
People can remember events accurately, but recollections can also contain mistakes, omissions, and details influenced by later information.
That means questions should be designed carefully.
Especially when discussing something that happened many years ago.
Imagine you’re telling a story about a childhood birthday.
You remember the cake.
The decorations.
And several relatives who attended.
But you don’t remember the weather.
Your AI asks:
“Was it a beautiful sunny day?”
You think for a moment.
You picture the backyard.
You remember the children playing outside.
Perhaps it probably was sunny.
But do you actually remember that?
Not necessarily.
Now suppose the AI asks another question.
“Did everyone gather outside in the sunshine?”
The suggested detail has appeared again.
If you’re not careful, the idea of a sunny afternoon may begin feeling more familiar simply because it has become part of the discussion.
This doesn’t mean a single question automatically creates a false memory.
It means we should avoid introducing unnecessary details when collecting recollections.
The safer approach is straightforward.
Ask what the storyteller remembers.
Allow uncertainty.
And distinguish genuine recollection from later inference.
If you don’t remember the weather, write:
Weather: Unknown.
That’s perfectly acceptable.
A book doesn’t become less meaningful because you can’t remember whether it rained on a Tuesday thirty years ago.
And if the weather doesn’t matter to the story, perhaps we don’t need it at all.
The Six Most Useful Questions
You may remember that Chapter 9 introduced six basic categories for organising memories.
We’re going to revisit them now from a conversational perspective.
These questions are simple.
But used carefully, they can help you develop a story without introducing unsupported details.
1. What Happened?
This is often the best place to begin.
“What happened that day?”
“What do you remember about the event?”
“Can you tell me the story?”
These questions allow the person to describe the experience in their own way.
You don’t have to interrupt with requests for dates or locations immediately.
Let the story develop.
2. Who Was There?
People often give a memory much of its character.
Perhaps a particular friend was involved.
A sibling.
A grandparent.
A neighbour.
Someone who appeared in several other stories.
Asking who was present can help identify relationships and connections.
But be careful not to suggest individuals who haven’t been mentioned.
Instead of:
“Was Grandpa there helping him?”
Try:
“Do you remember who was there?”
3. Where Did It Happen?
Places can provide useful context.
Sometimes a location triggers additional memories.
You might remember the family kitchen.
A workshop.
A park.
A holiday cottage.
Or a familiar stretch of road.
The question is simple:
“Where do you remember this happening?”
If you don’t remember, that’s fine.
And if the place isn’t important, you don’t need to spend twenty minutes trying to reconstruct it.
4. When Did It Happen?
Dates help us build timelines.
But they’re not always available.
You might remember that an event happened during childhood.
Or while someone attended a particular school.
Or before the family moved house.
Those approximate references can be useful.
Instead of demanding an exact year, ask:
“Do you remember roughly what period of his life this was?”
And allow the answer to remain approximate.
5. What Happened Next?
This can be a particularly useful question when the storyteller pauses during an event.
Perhaps they’ve described the beginning but haven’t explained the outcome.
You can ask:
“What happened after that?”
If they remember, the story continues.
If not, we preserve what we have.
We don’t invent an ending.
6. What Stands Out to You?
This question is different from asking for an objective detail.
It invites the storyteller to identify what they personally remember most clearly or consider meaningful.
Perhaps it’s a funny remark.
An unusual object.
Someone’s reaction.
A particular moment.
Or something completely unexpected.
The key is that the storyteller provides the answer.
The AI doesn’t decide what should have been important.
These six questions can carry you through an enormous number of conversations.
And sometimes, one is all you need.
The Question Behind the Question
Now let’s introduce something Mark does naturally.
He doesn’t always accept the first question or explanation as the final one.
Sometimes he explores why a particular detail matters.
Suppose you’re telling a story about a childhood bicycle.
You mention that your son repeatedly tried to modify it.
The AI asks:
“What kinds of modifications did he make?”
You describe several.
Then you remember another project involving an old wagon.
Now the AI might ask:
“Was experimenting with things something he enjoyed beyond bicycles?”
That’s a reasonable follow-up.
It connects to information you’ve already provided.
And it gives you an opportunity to consider whether the behaviour appeared elsewhere.
Perhaps you respond:
“Actually, yes. He was always trying things.”
Now you may remember another story.
Or perhaps you say:
“Not really. He just loved bicycles.”
That’s useful too.
The AI’s question helped clarify the scope of the interest.
And your answer determined the direction.
This is where conversational exploration becomes valuable.
We’re not merely collecting isolated facts.
We’re allowing the storyteller to examine possible connections.
But we’re doing it without forcing a conclusion.
A useful follow-up question should emerge from the information already shared, not from a story the AI wishes were true.
Listening Before Asking
Here’s something that deserves more attention.
Sometimes the best question is no question at all.
Imagine you’re telling a story.
You begin with a funny childhood incident.
Then you remember another detail.
Then another.
You’re talking naturally.
The memory is developing.
And suddenly, the AI interrupts:
“What did that experience teach you about the importance of family?”
Good grief.
We were doing so well.
You hadn’t finished the story.
You weren’t discussing a life lesson.
And now the AI has dragged the conversation somewhere entirely different.
This is why the listening stage matters.
Before asking questions, the AI should give you room to finish.
You can establish this preference directly.
“Please let me finish my story before asking questions.”
Or:
“Don’t interrupt with interpretations. I’ll tell you when I’m ready to explore the details.”
If you’re using a voice application, its turn-taking behaviour may not always accommodate long uninterrupted storytelling.
In that case, recording yourself separately may be more comfortable.
The technology should adapt to the way you prefer to collect memories wherever possible.
And if it can’t, choose a workflow that does.
Because a beautifully phrased question isn’t useful if it interrupts the memory you were about to share.
The Danger of Asking Too Many Questions
Let’s imagine you’ve told a short story about a family picnic.
The AI responds with eight questions.
What year was it?
What month?
What time?
Who attended?
What was the weather?
What food did everyone bring?
What clothing was everyone wearing?
What emotions did the event evoke?
That’s rather a lot to answer.
And some of those questions may not matter.
You might suddenly feel as though you’re taking an examination.
Instead of enjoying the memory, you’re struggling to supply information.
Project Crystal isn’t designed to work that way.
The goal isn’t to extract every imaginable detail from every experience.
It’s to help preserve the stories you wish to tell.
That’s why one question at a time is often a better approach.
Let the storyteller answer.
Consider the response.
Then decide whether another question would be useful.
And sometimes, after a few questions, it’s best to stop.
You might have collected everything the storyteller wishes to share.
That’s enough.
Questions That Invite Details Without Inventing Them
Now let’s examine a useful technique.
You can ask about sensory details.
But you need to do it carefully.
Imagine someone describing an afternoon in Grandpa’s garage.
A poorly framed question might be:
“Can you still remember the smell of oil and the sound of the radio playing?”
That question introduces two details.
Oil.
And a radio.
Neither has been established.
A more careful question would be:
“Are there any sounds, smells, or other details about the garage that you remember?”
Now the storyteller can decide what comes to mind.
Perhaps they remember the smell of sawdust.
Perhaps an old heater made a peculiar noise.
Perhaps they remember nothing sensory at all.
That’s fine.
Similarly, instead of asking:
“Was he wearing his favourite blue jacket?”
you might ask:
“Do you remember what he was wearing, or is that detail unclear?”
And instead of:
“Did everyone start laughing?”
you might ask:
“How did the others respond, if you remember?”
These questions can help make future writing more vivid.
But they don’t supply the answers.
That’s the distinction.
We want authentic details.
Not generic decorations.
A Complete Demonstration: The Fishing Trip
Let’s walk through a longer fictional conversation.
We’ll use the same memory in two different ways.
First, we’ll examine what happens when the AI asks leading questions.
Then we’ll see how a more careful conversation can develop.
Version One: The AI Leads the Story
Mother:
“I remember him going fishing with his grandfather when he was little.”
AI:
“Was that one of his favourite childhood memories?”
Mother:
“Maybe. He did enjoy going.”
AI:
“Did he love spending those peaceful mornings on the lake?”
Mother:
“Well, I think he liked being there.”
AI:
“Was his grandfather the person who taught him patience and a love of nature?”
Mother:
“I don’t really know about all that.”
AI:
“It sounds like their fishing trips created a special lifelong bond.”
Now examine what happened.
The AI introduced several interpretations.
Favourite childhood memories.
Peaceful mornings.
Patience.
A love of nature.
A lifelong bond.
Those may sound appropriate in a sentimental story.
But the original information supported very little of it.
The mother remembered her son going fishing with his grandfather.
That’s what we actually had.
Everything else was introduced through questions and assumptions.
Now let’s try again.
Version Two: The AI Follows the Story
Mother:
“I remember him going fishing with his grandfather when he was little.”
AI:
“Do you remember a particular trip?”
Mother:
“Oh, yes. There was this one time they came home absolutely soaking wet.”
AI:
“What happened?”
Mother:
“Well, Grandpa had this little boat. It wasn’t much to look at, but he loved the thing.”
AI:
“What do you remember about that day?”
Mother:
“They went out fishing, and somehow my son ended up falling into the water.”
AI:
“Do you remember how that happened?”
Mother:
“He was leaning over the side trying to grab something. I can’t remember exactly what. Anyway, over he went.”
AI:
“What happened afterward?”
Mother:
“Grandpa got him back into the boat. Then they came home, both of them soaked, and neither one wanted to explain what had happened.”
Now we have a story.
A specific event.
A boat.
An unexpected fall.
A grandfather helping his grandson.
And a funny moment when they returned home.
The AI might ask:
“How did you find out what happened?”
The mother could continue.
Perhaps someone eventually confessed.
Perhaps the boy told the story years later.
Perhaps she never learned the full explanation.
Whatever happened, the next detail should come from her.
Notice the difference.
The first conversation attempted to create meaning.
The second collected events.
And the second may eventually produce a much more distinctive chapter.
Because it contains material that could only come from this particular family’s memories.
The Most Valuable Detail Might Be the One You Didn’t Expect
Let’s return to the fictional fishing story.
Suppose the AI asks:
“What do you remember about how they explained the incident?”
The mother laughs.
“That’s the thing. They didn’t! Grandpa said the boat had taken on water. My son insisted he hadn’t fallen in. They were both standing there dripping wet, and neither one would admit what happened.”
Now we have an even better story.
There’s humour.
There’s a relationship.
There’s a particular interaction.
And the mother’s memory of how they behaved becomes part of the narrative.
No invented dialogue is necessary.
No elaborate emotional interpretation.
No discussion about the transformative power of fishing.
The actual details carry the story.
And that’s one of the most important lessons in this chapter.
Specific memories are often more interesting than broad descriptions of what those memories supposedly represent.
A useful question helps uncover those specifics.
Asking About Personality Without Creating One
At some point, you may wish to explore the person’s character.
What were they like?
What interested them?
What made them laugh?
How did they behave around different people?
These are reasonable subjects.
But they require care.
Suppose you ask:
“Was my son generous?”
The AI might search the available stories and identify several occasions when he helped someone.
Perhaps those stories support describing him as generous.
But perhaps they reflect particular situations rather than an enduring personality trait.
A more useful question might be:
“What specific memories do I have of him helping other people?”
Now you can examine actual events.
Perhaps you remember him helping a neighbour.
Sharing something with a friend.
Or assisting someone who was having difficulty.
The AI can organise those stories.
Then you decide whether generosity is an appropriate description.
Sometimes a few well-told experiences communicate the characteristic more effectively than a paragraph announcing it.
The reader can recognise the person’s behaviour.
And the writing doesn’t have to tell them what conclusion to reach.
Questions That Help Preserve Humour
Humour is particularly important when writing about someone you love.
A person’s sense of humour may appear in countless little situations.
The jokes they repeated.
The silly things they did.
Their responses to unexpected problems.
The way they teased family members.
The things that made them laugh uncontrollably.
But humour can be difficult to preserve when a story is reduced to a summary.
Imagine someone saying:
“He was always playing jokes on his sister.”
That’s a general statement.
Now imagine asking:
“Do you remember one of those jokes?”
Suddenly, the storyteller remembers a specific incident.
Perhaps an elaborate practical joke.
Perhaps a ridiculous nickname.
Perhaps a joke that backfired.
Now we have material.
The AI can ask:
“What happened when she found out?”
Or:
“How did he react?”
Those questions may help complete the story.
But they shouldn’t suggest that everyone laughed or that the joke was harmless unless the storyteller actually remembers that.
Real family humour isn’t always perfectly tidy.
Sometimes someone was annoyed.
Sometimes a joke went too far.
Sometimes the person telling the story remembers the incident differently from everyone else.
Those details can make the story more authentic.
We don’t need to sanitise every funny memory into a cheerful little moral lesson.
We need to preserve what actually happened.
When You Don’t Know the Answer
Here’s a skill that may seem small but is enormously important.
Learning to say:
“I don’t know.”
Or:
“I don’t remember.”
Or:
“I’m not sure.”
These are perfectly acceptable answers.
And your AI should respect them.
Suppose the AI asks:
“How old was he when that happened?”
You think.
Then say:
“I believe he was around eight, but I’m not sure.”
That’s useful information.
We can record the approximate age as uncertain.
The AI shouldn’t respond by choosing eight as a definitive fact.
And it shouldn’t insist that you keep searching your memory until you produce an answer.
Sometimes you don’t know.
Perhaps you can find the information later.
Perhaps not.
Either way, the story can still be preserved.
You might write:
“When he was little…”
Or:
“One summer…”
Or simply describe the event without attaching a date.
Uncertainty isn’t a flaw in the collection. Pretending uncertainty doesn’t exist is the problem.
Asking Without Pressuring
When the subject is someone who has died, questions can sometimes bring up difficult memories.
Perhaps you’re talking about a favourite place.
The conversation leads toward a painful event.
You become quiet.
The AI asks another question.
Then another.
Perhaps it assumes that the difficult subject is something you need to explore.
But that’s not the purpose of Project Crystal.
You decide what you’re comfortable discussing.
You can say:
“I’d rather not talk about that.”
Or:
“Let’s stay with the funny stories today.”
Or:
“I don’t want any more questions.”
Those instructions should be respected.
There is no requirement to complete an emotional journey through your memories.
No prescribed order.
No obligation to revisit difficult experiences.
And no need to explain why you’ve chosen to stop.
The book is being developed at your pace.
Questions are tools.
They aren’t demands.
And a useful tool should be put aside when it isn’t helping.
Interviewing Family and Friends
As your collection grows, you may decide to speak with other people.
Family members.
Childhood friends.
Neighbours.
People who worked with your son.
People who knew parts of his life that you didn’t personally witness.
Those conversations can add wonderful detail.
But the same principles apply.
Ask open-ended questions.
Avoid suggesting answers.
Allow people to describe their own recollections.
And preserve the distinction between their perspective and yours.
For example, instead of asking:
“Wasn’t he always the funniest person in the group?”
try:
“What do you remember about spending time with him?”
The person might describe humour.
Or kindness.
Or an interest you didn’t know about.
Or something entirely unexpected.
That’s valuable.
You want their memories, not confirmation of a description you’ve already chosen.
Before recording another person, explain the purpose of the conversation and obtain appropriate consent.
Also clarify whether their words may be used in a book or whether they’re contributing only to the private collection.
Recording permission and publication permission are not necessarily the same thing.
And if someone changes their mind about sharing a sensitive story, take that seriously.
Project Crystal is built around respect for the people whose experiences make the book possible.
When Two People Remember Differently
Suppose you interview your son’s childhood friend.
They describe a particular adventure.
You remember the event too.
But you recall it differently.
Perhaps the friend believes the incident happened at school.
You remember it happening at someone’s house.
Perhaps they remember a different year.
Or they disagree about who was responsible for whatever ridiculous thing happened.
Well, that’s family storytelling for you.
We don’t need the AI to declare a winner immediately.
Instead, preserve the accounts separately.
Your recollection: The event happened at the friend’s house.
Friend’s recollection: The event happened at school.
Status: Location disputed.
Later, another source may help clarify the details.
Perhaps a photograph.
A letter.
A dated message.
Or another participant’s recollection.
If the disagreement remains unresolved, the manuscript can acknowledge it.
Or the detail may simply be unnecessary to the story.
The important thing is that we don’t quietly combine conflicting accounts into a fictional version that nobody actually remembered.
And when interviewing another person, avoid repeatedly suggesting your own version.
Their recollection has value precisely because it offers another perspective.
Questions That Explore a Relationship
Sometimes several memories involve the same person.
A grandparent.
A close friend.
A sibling.
A neighbour.
You may wish to understand that relationship more fully.
Here are examples of useful questions:
“What do you remember them doing together?”
“How did they get to know each other?”
“Are there any particular experiences involving the two of them that stand out?”
“Did they have any shared interests?”
“Is there a story that captures the way they interacted?”
These questions can help collect specific material.
But avoid assuming the relationship was always positive, simple, or emotionally significant in the same way.
For example:
“Did his grandfather teach him everything he knew about cars?”
That’s a strong assumption.
Perhaps Grandpa taught him a few things.
Perhaps they learned together.
Perhaps they argued about cars constantly.
Perhaps the boy actually knew more about a particular subject.
We don’t know until someone tells us.
A more careful question would be:
“What do you remember about them working on cars together?”
Let the story develop from there.
Questions That Explore a Place
Places can connect many different memories.
Perhaps your family spent years visiting the same cabin.
Perhaps your son had a favourite garage, park, restaurant, or holiday destination.
You can explore those places through questions.
“What do you remember about being there?”
“Who usually went with him?”
“What did he like doing there?”
“Are there particular visits that stand out?”
“What happened there that you still remember?”
You might also ask about physical details.
But don’t introduce them unnecessarily.
Instead of asking whether the cabin smelled of pine trees, ask whether any sensory details come to mind.
Instead of assuming that the lake was peaceful, ask what the place was like.
Let the storyteller supply the atmosphere.
Because authentic setting details can make a chapter vivid.
And we want those details to come from actual recollections.
Questions That Explore an Object
Let’s return to the idea of meaningful objects.
Perhaps you find an old toy.
A photograph.
A tool.
A jacket.
A handwritten note.
An unusual keepsake.
You can begin with something simple:
“What do I remember about this?”
Then ask more specific questions.
“Who did it belong to?”
“How did they get it?”
“What did they use it for?”
“Is there a particular story connected to it?”
“Do I remember why they kept it?”
Sometimes an object opens an entire collection of memories.
And sometimes you only remember that the person owned it.
That’s fine.
Not every object needs a dramatic history.
The AI can help you organise whatever information emerges.
But it shouldn’t invent sentimental significance simply because the object appears in a photograph.
The Question That Reveals an Unexpected Story
Now let’s examine something particularly interesting.
Sometimes a question doesn’t deepen the original memory.
It opens a completely different one.
Imagine you’re describing your son’s interest in music.
The AI asks:
“Do you remember when that interest began?”
You think.
Then say:
“Actually, that reminds me of the time he tried learning an instrument.”
Now you have a new story.
Perhaps he practised constantly.
Perhaps he abandoned it after three days.
Perhaps something funny happened during a school performance.
The conversation has changed direction.
That’s okay.
You don’t have to force the discussion back to its original subject.
You can follow the new memory.
Later, your AI can help identify the connection.
Perhaps both stories belong in a future chapter about music.
Perhaps they’re unrelated.
We don’t need to decide yet.
The important thing is that the question helped reveal genuine material.
This is one of the most valuable aspects of conversational discovery.
A question doesn’t have to produce the answer you expected to be useful.
Sometimes it leads somewhere better.
Learning to Recognise a Bad Question
Let’s create a practical exercise.
Imagine your AI asks:
“Was he always generous, funny, brave, and protective of everyone around him?”
There are several problems with that question.
It contains multiple personality claims.
It asks you to agree or disagree with all of them at once.
It suggests that the characteristics were always present.
And it provides no specific event to examine.
A better approach would be to begin with one subject.
For example:
“Can you remember a time he helped someone?”
Or:
“What’s a story that shows his sense of humour?”
These questions invite actual memories.
And once the stories are collected, you can consider what they reveal.
Here’s another example.
Leading question:
“How proud were you when he won that award?”
This assumes you felt pride and that the award was important to you.
Better question:
“What do you remember about the award ceremony?”
Now the storyteller can describe the event and whatever feelings they genuinely recall.
Another example.
Leading question:
“Did that difficult experience make him stronger?”
Better question:
“How do you remember him responding afterward?”
Notice how the better questions focus on the actual experience.
They allow the storyteller to provide meaning rather than receiving it from the AI.
That’s the skill we’re developing.
A Question Library You Can Actually Use
Let’s create a collection of practical questions.
You don’t need to answer them all.
They’re simply starting points.
You can read through them, choose one that interests you, and begin talking.
Questions About Childhood
“What is one of the earliest memories that comes to mind?”
“What did he enjoy doing when he was little?”
“Was there a particular toy or object he loved?”
“Do you remember a funny childhood incident?”
“What was an ordinary day at home like?”
“Who did he enjoy spending time with?”
Questions About Personality
“What kinds of things made him laugh?”
“Was there a habit that everyone remembers?”
“What did he get excited about?”
“What did he enjoy talking about?”
“Are there stories that show how he behaved around friends?”
“What’s something about him that photographs don’t really capture?”
Questions About Family
“What do you remember about family holidays?”
“Was there a particular family tradition he enjoyed?”
“Who did he spend time with?”
“What stories involve his grandparents?”
“Are there familiar sayings or jokes associated with him?”
Questions About Friends
“Who were some of his friends?”
“What do you remember about their adventures?”
“Are there photographs that remind you of those friendships?”
“Did any friendships continue through different periods of his life?”
“Is there someone who might have stories you haven’t heard?”
Questions About Interests
“What did he enjoy doing in his free time?”
“Did he collect anything?”
“Was there something he loved learning about?”
“Do you remember a project he worked on?”
“Did he have a favourite place to spend time?”
Questions About Everyday Life
“What did he usually do when he came home?”
“Did he have a favourite food?”
“What music did he enjoy?”
“Was there something he was always looking for or misplacing?”
“What little habit still makes you smile when you remember it?”
Questions About Particular Memories
“What happened next?”
“Who else was there?”
“Do you remember where it happened?”
“Is there a particular detail that stands out?”
“Is any part of the story uncertain?”
“Does this remind you of another experience?”
These are tools, not assignments.
You can choose one.
Answer it.
Follow the memory.
And stop whenever you wish.
That’s enough.
How to Ask Questions About Difficult Memories
Some stories may involve painful or complicated experiences.
If you choose to explore them, the questions should be especially careful.
Avoid assuming what someone felt.
Avoid asking them to relive details they’re not comfortable discussing.
And avoid treating the conversation as though it must lead to emotional resolution.
You might begin with:
“Would you like to talk about this memory?”
If the answer is no, stop.
If the answer is yes, let the storyteller determine what they wish to share.
You can ask:
“What would you like to remember about that period?”
Or:
“Is there a particular part of the experience you’d like to preserve?”
But there’s no obligation to continue.
And an AI shouldn’t suggest that the person needs to discuss painful experiences in order to produce a complete or meaningful book.
Project Crystal is a writing and preservation process.
It isn’t a required form of emotional treatment.
And a book about someone’s life can be deeply meaningful without examining every difficult experience.
A Useful Questioning Rule: Don’t Ask What You Already Know
Here’s a small practical lesson.
Imagine you’ve told your AI that your son’s childhood friend was named Daniel.
You’ve explained how they met.
You’ve shared several stories involving their adventures.
Then the AI asks:
“What was his childhood friend’s name?”
That isn’t particularly useful.
You’ve already provided the information.
Now, an AI may sometimes lose track of earlier details, especially in long conversations.
But when the relevant information is available, it should use it rather than repeatedly asking you to supply it again.
You can encourage this by saying:
“Before asking a question, check whether I’ve already provided the answer in the available material.”
And if the AI doesn’t have access to the earlier information, it should say so.
It shouldn’t guess.
This is another reason our Memory Library matters.
We can supply relevant records at the beginning of a session.
Then the AI can ask questions based on what is actually available.
The goal is to spend your time collecting new material, not constantly repeating facts you’ve already documented.
Asking the AI to Explain Its Questions
Here’s a technique you may find useful.
Sometimes the AI asks something that seems unrelated.
You can simply ask:
“Why are you asking that?”
Perhaps it noticed a possible connection.
Perhaps it’s trying to clarify a date.
Perhaps it misunderstood the subject.
Or perhaps it generated a question that isn’t particularly useful.
Either way, asking for the reason can help you decide whether to pursue it.
For example:
AI: “Did he have a favourite place to work on projects?”
You: “Why are you asking?”
AI: “Several memories mention him repairing things, but the locations differ. I was wondering whether there was one place he used regularly.”
Now you can decide.
Perhaps there was a workshop.
Perhaps he worked wherever he happened to be.
Perhaps the location isn’t important.
The AI’s explanation isn’t proof that its question is useful.
But it can make the purpose clearer.
And you’re free to redirect the conversation.
That’s part of collaborating actively.
What Mark’s Approach Teaches Us About Questions
Let’s return to Mark.
One of the interesting things about his conversational style is that questions often emerge while he’s trying to understand something else.
He doesn’t always begin with a formal objective.
He introduces an idea.
The AI responds.
He evaluates the explanation.
Then he asks another question.
Sometimes the new question is more useful than the original one.
And occasionally, the entire conversation changes direction.
This isn’t simply wandering.
At its best, it’s a feedback process.
Each response provides something to evaluate.
That evaluation reveals what still needs clarification.
And the next question becomes more focused.
We’re adapting that same principle to memory collection.
You begin with a story.
The AI asks a useful question.
You provide additional information.
Perhaps something was misunderstood.
You correct it.
Now another question becomes possible.
And gradually, the memory becomes clearer.
But unlike an abstract discussion, our source material concerns real experiences.
So we add an important safeguard.
The questions must not manufacture their own answers.
We explore.
We clarify.
We preserve.
And when information is uncertain, we leave it uncertain.
That’s conversational discovery applied responsibly to personal history.
A Reusable Prompt for Better Questions
Now let’s create a practical instruction you can use with your chosen AI.
This prompt is designed to help it ask useful questions without introducing unsupported assumptions.
PROJECT CRYSTAL: THE CAREFUL INTERVIEWER
I am collecting genuine memories for a personal book.
Your role is to help me remember and describe real experiences through natural conversation.
Please follow these instructions:
1. Ask Open-Ended Questions
Prefer questions that allow me to answer in my own words.
Avoid suggesting specific events, emotions, or details unless I’ve already mentioned them.
2. Ask One Question at a Time
Wait for my answer before asking another.
Do not present long lists of questions unless I request them.
3. Listen Before Exploring
Let me finish telling a story before asking follow-up questions.
Do not interrupt with interpretations or emotional analysis.
4. Use the Available Source Material
Base questions on information I’ve actually shared.
If you don’t have access to earlier records, tell me rather than pretending you remember them.
5. Avoid Leading Questions
Do not assume how someone felt, what motivated them, or what an event meant.
Ask me to describe those things when I choose to.
6. Respect Uncertainty
If I say I don’t know or don’t remember, accept that answer.
Do not pressure me to produce a detail or suggest an answer.
7. Preserve Natural Conversation
Allow me to change subjects.
Follow useful connections when I choose to explore them.
Don’t turn the conversation into a rigid interview.
8. Distinguish Recollection From Interpretation
If you notice a possible pattern, present it as a suggestion and identify which memories support it.
Let me decide whether it fits.
9. Respect My Boundaries
If I decline a question or ask to stop, move on or end the session.
Do not treat personal memories as material that must be extracted.
10. Preserve the Results
When I request a summary, use only information I provided.
Identify uncertainties and corrections.
Do not invent dialogue, events, dates, or emotions.
Do not begin writing book chapters until I specifically ask.
Let’s begin with one simple question about a memory I’m comfortable discussing.
This prompt is a starting point.
You can shorten it.
Adjust the tone.
Or add your own preferences.
Perhaps you prefer fewer questions.
Perhaps you like a little humour.
Maybe you want the AI to wait until you explicitly say you’re finished before asking anything.
That’s fine.
The point is to develop a questioning style that helps you tell stories naturally.
Not one that forces you to behave like you’re being interviewed for a documentary.
The Question Review Exercise
Let’s put the ideas into practice.
This exercise is designed to help you recognise the difference between a useful question and one that introduces assumptions.
First: Choose a Memory
Select a short memory from your collection.
It can be funny, ordinary, or personally meaningful.
Make sure you feel comfortable discussing it.
Second: Ask the AI for Questions
Provide the approved memory record.
Then say:
“Please suggest five open-ended questions that might help me explore this memory further. Don’t assume additional events happened.”
Third: Examine Each Question
Look for hidden assumptions.
Does the question suggest an emotion?
Does it introduce a person who wasn’t mentioned?
Does it assume a particular outcome?
Does it suggest that the memory proves a personality trait?
If so, ask the AI to rewrite it.
Fourth: Choose One Question
Pick whichever question interests you.
You don’t have to answer all five.
Fifth: Tell the Story
Answer naturally.
If another memory appears, follow it.
Correct misunderstandings.
Allow uncertainty.
Sixth: Preserve the Result
Save the original conversation.
Create an updated memory record if new information emerged.
Keep the new material connected to its source.
Seventh: Review What Changed
Ask yourself:
Did the question help me remember something genuine?
Did it clarify an earlier detail?
Did it reveal a useful connection?
Or did it introduce assumptions that needed correction?
All of those observations can help you improve future conversations.
And remember, you don’t have to produce a new memory every time.
Sometimes the best result is simply a clearer understanding of the story you’ve already told.
Questions for Photographs
Let’s add one more practical exercise.
Choose a photograph from your collection.
Before involving the AI, take a moment to examine it yourself.
What do you recognise?
Who appears in the photograph?
What do you remember about the occasion?
Then, if your AI supports image analysis, you may choose to share the image.
Ask it to help identify visible details.
But establish an important boundary.
It should distinguish between what the image visibly contains and what you’ve told it about the occasion.
For example:
Visible Detail: A child standing beside a bicycle.
Your Recollection: The bicycle belonged to your son.
Uncertain Detail: The exact year the photograph was taken.
The AI should not invent the child’s identity, emotions, or personal history from the image.
And if the photograph contains people, consider whether you’re comfortable uploading it to the service.
Some images may be better kept private.
A useful prompt is:
“Please describe only what is visibly present in this photograph. Do not identify people by name, invent relationships, infer emotions as facts, or create a story about the event. Afterward, ask me one open-ended question about what I remember.”
That approach helps keep the photograph and the recollection distinct.
And it may open the door to another genuine memory.
The Question That Should Never Be Required
Let’s finish with something important.
When you’re writing about someone who has died, people may expect certain questions.
What happened?
How did you cope?
What did you learn?
How has the loss changed you?
Those questions may be meaningful in some contexts.
But they’re not requirements for Project Crystal.
You’re writing about a person’s life.
Their memories.
Their personality.
Their relationships.
Their humour.
Their interests.
Their ordinary experiences.
You may choose to discuss loss.
You may choose to write about grief.
You may choose to include those experiences in the finished manuscript.
But the book doesn’t need to begin there.
And it doesn’t need to revolve around them.
A question about someone’s favourite childhood toy can be just as appropriate as a question about a major life event.
A funny story about a broken bicycle can reveal something valuable.
A memory of an ordinary family dinner can preserve a moment that matters deeply to the person telling it.
The AI shouldn’t decide which memories deserve attention based on how dramatic they seem.
And it shouldn’t insist on exploring difficult subjects because it assumes that’s where the deepest meaning must be found.
The most important question is often the one you feel comfortable answering today.
That’s where we begin.
What Makes a Story Rich?
Let’s return to the idea that opened this chapter.
A mother remembers her son working on a car.
At first, the memory is simple.
“He loved that car.”
Then a question opens another detail.
“What was the first thing he changed?”
She remembers the parts.
The work.
The excitement.
The problem that followed.
Then another question.
“What happened when he tried it?”
Now she remembers the outcome.
Perhaps the car ran beautifully.
Perhaps it broke down.
Perhaps the whole family ended up laughing about something.
The story develops.
But the important thing isn’t that the AI produced a large number of questions.
It’s that the questions helped the storyteller provide specific information.
Information that belonged to her experience.
That’s what makes the material rich.
Not elaborate language.
Not forced symbolism.
Not invented atmosphere.
And not a paragraph explaining what the entire experience supposedly meant.
A rich personal story contains the details that distinguish it from every other story.
The people.
The events.
The humour.
The unexpected outcomes.
The expressions.
The ordinary little things that make a person recognisable.
Those details come from genuine recollections and other identifiable sources.
And careful questions can help bring them into the conversation.
One Final Thought: Curiosity With Restraint
There’s something rather beautiful about a good question.
It doesn’t need to contain an answer.
It doesn’t need to prove a point.
It simply creates an opportunity to explore.
And sometimes the answer leads somewhere unexpected.
That’s the spirit of the Meistro Method.
Curiosity.
Exploration.
Clarification.
Correction.
Discovery.
But when working with personal memories, curiosity needs a companion.
Restraint.
The willingness not to assume.
Not to invent.
Not to force a connection.
Not to turn every story into a psychological explanation.
And not to continue questioning when the storyteller has had enough.
Because the purpose of this process isn’t to extract the greatest possible number of words from someone.
It’s to help them preserve the stories they choose to share.
And when the questions are thoughtful, something remarkable can happen.
A small recollection becomes a detailed story.
A photograph opens another memory.
A familiar place reveals an unexpected connection.
A funny incident helps communicate someone’s personality.
And gradually, your collection becomes richer.
Not because the AI has created a life.
But because you’ve been given more opportunities to describe the life you remember.
That’s the difference.
And it’s worth protecting.
Before You Turn the Page
Let’s review the most important lessons from this chapter.
1. Open-ended questions encourage genuine recollection.
Ask what happened rather than suggesting what might have happened.
2. Leading questions can introduce unsupported details.
Avoid questions that assume emotions, motivations, events, or outcomes.
3. Listen before asking.
Give the storyteller room to finish their memory.
4. Ask one question at a time.
A natural conversation is often more productive than a long questionnaire.
5. Specific questions can reveal richer details.
People, places, actions, objects, and outcomes can help develop a story.
6. Accept uncertainty.
“I don’t remember” is a valid answer.
7. Don’t force meaning into ordinary experiences.
Sometimes a funny story is simply a funny story.
8. Preserve different perspectives.
When other people contribute memories, keep their accounts accurately attributed.
9. Respect emotional boundaries.
No question is mandatory, and the storyteller decides when to stop.
10. Let questions lead to genuine discoveries.
A useful question may reveal a new memory, clarify an earlier one, or simply help organise existing material.
We’ve now developed another important skill in Project Crystal.
You know how to collect memories.
You know how to preserve them.
You’ve learned how to organise your library.
You’ve discovered how to explore connections.
And now you understand how thoughtful questions can help those memories develop without quietly changing the facts.
But there’s one more principle we need to examine before we begin discovering the structure of the book.
Who makes the final decisions?
It may sound obvious.
You do.
But as AI becomes more involved in organising information, identifying patterns, suggesting chapters, and eventually producing polished writing, it can become surprisingly easy to accept its output without examining it carefully.
That’s where we’re going next.
In Chapter 14: Keeping the Human in Charge, we’ll explore how to maintain authorship, verify the AI’s work, recognise unsupported claims, and establish a dependable review process.
We’ll examine how AI-generated details can gradually enter a manuscript if nobody checks them.
And we’ll develop practical safeguards so that the technology remains a tool for preserving your memories rather than becoming an unreliable narrator of someone else’s life.
Because this project began with a very simple idea.
The book should contain your genuine material.
Your memories.
Your experiences.
Your perspective.
The AI can help organise and communicate them.
But it should never quietly replace them.
The questions open the doors.
Your memories provide the stories.
And you decide what belongs on the page.