Finding Your Voice on the Page
The Strange Experience of Reading Your Own Words
Imagine you’ve just finished recording a story.
You were sitting comfortably, talking about something that happened years ago.
Perhaps it was a family holiday.
Perhaps something involving your son and his grandfather.
Maybe it was one of those ridiculous incidents that still makes you laugh whenever somebody mentions it.
You spoke naturally.
You paused.
You changed direction.
You remembered something halfway through.
You laughed at your own story.
And when you finished, you felt rather pleased with yourself.
That was a good one.
Then you transcribed the recording.
You opened the document.
And you began reading.
Good heavens.
Did you really say “you know” seventeen times?
Why did you explain the same thing three different ways?
And what on earth happened to that sentence in the middle?
It begins in one place, wanders through a completely unrelated thought, and eventually arrives somewhere that doesn’t seem to have anything to do with where it started.
Welcome to the wonderful world of spoken language.
Here’s something important.
There’s nothing unusual about the way you speak.
Natural conversation isn’t organised like a finished book.
We repeat ourselves.
We reconsider what we’re saying.
We introduce details when we remember them.
We leave sentences unfinished because the person listening already understands.
We change subjects.
We use gestures, tone, facial expressions, and shared knowledge to communicate things that may not appear in the transcript.
And most of the time, that works beautifully.
But when we move those words onto a page, something changes.
The reader can’t see your expression.
They can’t hear your laughter.
They may not know the people you’re discussing.
And they can’t necessarily ask you to explain something halfway through the paragraph.
So the writing needs a little help.
That’s what we’re going to explore in this chapter.
How to make your spoken memories readable without making them sound as though somebody else wrote them.
Because there’s a difference between cleaning up a story and cleaning the personality out of it.
And our goal is to learn that difference.
The Page Is a Different Kind of Conversation
Let’s begin with a simple idea.
When you speak to someone, you’re sharing an experience in real time.
You can see whether they’re following.
If they look confused, you explain something.
If they laugh, you might tell another funny detail.
If they interrupt with a question, you answer.
The conversation adjusts as it unfolds.
A book works differently.
The reader encounters the words you’ve already arranged.
You aren’t sitting beside them explaining every reference.
The writing needs to provide enough information for the story to make sense on its own.
That doesn’t mean it needs to become formal.
It doesn’t mean you must use elaborate descriptions.
And it certainly doesn’t mean every memory requires an inspirational conclusion.
It simply means that we need to consider what the reader knows.
Imagine telling your sister:
“Remember that thing Daniel did at Grandpa’s?”
She knows exactly what you’re talking about.
You both start laughing.
The entire story might be contained in that one sentence.
But a reader who doesn’t know Daniel or Grandpa will be thoroughly confused.
What thing?
Who is Daniel?
Which grandfather?
What happened?
To make the story readable, you may need a little context.
For example:
“Daniel was one of his childhood friends, and the two of them spent plenty of time finding ways to get into trouble. One afternoon at Grandpa’s house, they came up with an idea that didn’t go quite as planned.”
That introduction may help.
But only if the details are actually supported by your memories.
If you haven’t established that the boys frequently got into trouble, we shouldn’t add that description merely because it sounds entertaining.
We might instead write:
“Daniel was one of his childhood friends. I remember one afternoon when the two of them were visiting Grandpa’s house.”
It’s simpler.
But it tells the reader what they need to know.
That’s the first principle of writing for the page.
Give readers enough context to understand the memory, but don’t invent context to make the writing feel complete.
Your Speaking Voice Is the Starting Point
In Chapter 11, we created your Storytelling Voice Profile.
We explored your natural expressions.
Your humour.
Your sentence patterns.
The kinds of descriptions you prefer.
And the language that doesn’t sound like you.
Now we’re going to put that information to work.
The Voice Profile tells us what we’re trying to preserve.
This chapter teaches us how to preserve it while making the writing easier to read.
Think of your original recording as the starting material.
Your finished prose is a carefully prepared version of that material.
We can remove unnecessary repetition.
Clarify confusing references.
Rearrange sentences when necessary.
And organise events into an understandable sequence.
But we don’t have to replace your vocabulary with grander words.
We don’t have to remove your humour.
And we don’t need to turn ordinary recollections into dramatic literary scenes.
Perhaps your natural expression is:
“That boy couldn’t leave anything alone.”
That’s a perfectly useful sentence.
An AI might try to improve it:
“From an early age, he demonstrated an extraordinary curiosity about the mechanisms and complexities of the world around him.”
Well.
We’ve gone from a family story to a university prospectus.
And we’ve introduced an interpretation the original sentence doesn’t necessarily support.
A better edit might leave the opening alone.
Then follow it with a specific memory.
“That boy couldn’t leave anything alone. If something had screws in it, there was a good chance he’d eventually find out what was inside.”
If that reflects what you genuinely remember, it may be exactly the kind of language the book needs.
It’s natural.
It’s recognisable.
And it leads toward the stories.
The objective isn’t to make the writing sound less ordinary.
It’s to make ordinary language carry the story clearly.
Two Voices to Protect
There’s an important distinction we need to establish.
A book about your son may contain two different voices.
The first is yours.
The voice of the storyteller.
The person remembering, explaining, laughing, and reflecting.
The second is his.
The expressions, quotations, and distinctive language that belonged to the person you’re remembering.
Those voices should not automatically become the same.
Perhaps your son had a favourite saying.
Perhaps he used unusual words.
Perhaps he had a particular sense of humour.
Maybe you have recordings or messages that preserve some of his actual expressions.
Those details can be valuable.
But they need to be handled carefully.
If you have his exact words, you may choose to quote them accurately, with appropriate context and permissions.
If you remember the general meaning of something he said but not the exact wording, it’s better to report the substance of the conversation.
For example:
“He told me he wasn’t giving up on the project.”
That’s different from writing:
“‘I’m never giving up on this, Mum,’ he declared.”
The second version presents an exact quotation.
If you don’t actually remember those words, you shouldn’t manufacture them.
The same rule applies to your own remembered dialogue.
You may recall precisely what you said.
Or you may only remember the general conversation.
The book can accommodate both.
But readers shouldn’t be led to believe reconstructed dialogue is verbatim when it isn’t.
Your voice tells the memories. His real words, when known, remain his.
And any other contributor’s voice should be treated with the same care.
Choosing Who Tells the Story
Now let’s introduce one of the most important writing decisions you’ll make.
Point of view.
That simply means the perspective from which the story is written.
For Project Crystal, you’ll most likely encounter three possibilities.
First person.
Third person.
Or a combination of the two.
Let’s examine them.
First Person: “I Remember”
First-person writing uses words such as “I,” “me,” “my,” and “we.”
For example:
“I remember when he brought that old bicycle home. I couldn’t understand why he wanted another project when the garage was already full of things he was supposed to be fixing.”
This style places the storyteller directly inside the book.
Readers experience events through your recollections.
It can feel intimate.
Natural.
Conversational.
And especially appropriate when the book is built from your personal memories.
It also allows you to acknowledge uncertainty naturally.
“I think he was about twelve.”
“I don’t remember exactly what he said.”
“I may have the year wrong, but I remember the incident.”
Those statements can belong comfortably in first-person prose.
They remind readers that you’re sharing your perspective rather than claiming omniscient knowledge of every event.
For many Project Crystal books, first person will be an excellent starting point.
But it isn’t the only option.
Third Person: “He Remembered”
Third-person writing describes someone using their name or words such as “he,” “she,” or “they.”
For example:
“He spent many afternoons working on his bicycle, experimenting with different changes and trying to get everything working the way he imagined.”
That sounds more like a conventional biography.
Third person can be useful when the book draws heavily on documents, multiple interviews, and accounts from several people.
It may also suit someone who prefers a little distance between the narrator and the events.
But third-person writing introduces a particular risk.
It can make uncertain information sound definite.
Compare these sentences:
“I remember him becoming frustrated with the bicycle.”
“He became frustrated with the bicycle.”
The first tells us the source of the observation.
The second presents the frustration as a direct fact.
That may be appropriate if the evidence supports it.
But when the information comes from a particular person’s recollection, we should be careful about losing that context.
Third person also makes it easier for an AI to slip into describing someone’s private thoughts.
For example:
“He knew this was the moment his life would change.”
Unless we have reliable source material supporting that statement, the author shouldn’t claim to know what he thought.
A biography can describe actions and reported experiences without inventing access to someone’s inner life.
A Combined Approach
You may decide to use both.
Perhaps the book contains first-person memories written in your voice.
Between those memories, short third-person passages provide factual background.
A chapter might begin with a historical description supported by records.
Then move into your personal recollection.
Or perhaps friends and relatives contribute clearly attributed first-person passages.
That’s possible too.
The important thing is establishing a consistent system.
Readers should be able to tell who is speaking and where information comes from.
You don’t want a chapter switching unpredictably between your memories, an AI-generated narrator, and another person’s account.
A combined approach can work very well.
But the transitions need to be deliberate.
Which Point of View Should You Choose?
Let’s make the decision easier.
Ask yourself one question.
Whose perspective is the heart of this book?
If the book primarily consists of your memories of your son, first person may feel natural.
If the book is intended to combine several perspectives into a broader documented account of his life, third person or a combined structure may be appropriate.
If it’s a private collection of family stories, you might preserve different contributors’ voices within separate sections.
There’s no universal right answer.
And you can experiment.
Choose one approved memory.
Write a short first-person version.
Then create a third-person version using exactly the same information.
Read both.
Which feels more authentic?
Which makes the source of the information clearer?
Which better supports the kind of book you’re creating?
You can ask your AI to prepare the alternatives.
But you should review the results carefully.
Changing point of view must not introduce new knowledge.
An AI rewriting “I remember him smiling” as “He felt perfectly happy” has changed the meaning.
We know the storyteller remembers a smile.
We don’t automatically know everything the person felt.
Point of view changes how the story is told. It shouldn’t change what the source allows us to know.
A Complete Example: Spoken Memory to Written Passage
Let’s work through a practical demonstration.
The following memory is entirely fictional and is used only to demonstrate the editing process.
Imagine a mother has recorded a story about her son working on an old radio.
Here’s a short passage from the transcript.
Stage One: The Original Spoken Version
“Well, he had this old radio, you know, and it hadn’t worked for ages. I don’t even remember where he got it. But anyway, he decided he was going to fix it. And I remember he had all these pieces laid out on the kitchen table, and I kept telling him, you know, don’t lose anything, because he was always leaving little screws everywhere.
“And I went back into the kitchen later, and he had the thing playing music. I couldn’t believe it.
“Only now we had a different problem, because he’d apparently decided he liked it so much that he was going to leave it on the kitchen table.
“And I told him, absolutely not. I needed my table back.”
That’s a perfectly reasonable spoken memory.
The storyteller pauses.
Repeats a few words.
Changes direction.
And remembers the event as she talks.
Now let’s turn it into something easier to read.
Stage Two: Light Cleanup
“He had this old radio that hadn’t worked for ages. I don’t even remember where he got it, but he decided he was going to fix it.
“I remember all the pieces laid out on the kitchen table. I kept telling him not to lose anything because he had a habit of leaving little screws everywhere.
“When I went back into the kitchen later, he had the thing playing music. I couldn’t believe it.
“Unfortunately, we’d developed another problem.
“He’d decided he liked the radio so much that he was going to leave it on the kitchen table.
“And I told him, absolutely not. I needed my table back.”
Now look at what changed.
We removed some unnecessary filler words.
We clarified the sequence.
We created paragraph breaks.
And we preserved the humour.
The story remains recognisable.
We haven’t invented a year.
We haven’t described the music.
We haven’t claimed the son was proud.
And we haven’t turned the incident into an inspirational lesson about determination.
That’s a useful light edit.
Stage Three: A More Developed Narrative
Now suppose the mother approves a slightly more structured version.
It might read:
“He had this old radio that hadn’t worked for ages.
“I don’t even remember where he got it, but one day he decided he was going to fix it.
“Before long, the kitchen table was covered in pieces. I kept reminding him not to lose anything because he had a habit of leaving those little screws everywhere.
“Later, I walked back into the kitchen and heard music.
“He’d actually fixed the thing.
“I couldn’t believe it.
“Of course, that only created another problem.
“Apparently, he liked the radio so much that he’d decided the kitchen table was now its permanent home.
“Absolutely not.
“I needed my table back.”
Notice the differences.
The paragraphs are arranged for pacing.
The moment the radio works has more emphasis.
The ending preserves the mother’s reaction.
And the language remains conversational.
However, there’s a detail worth checking.
The developed version says she heard music when she walked back into the kitchen.
The original only established that the radio was playing music when she returned.
Those statements are close, but the first makes a specific claim about her sensory experience.
We should check whether that accurately reflects her recollection.
If she confirms it, fine.
If not, we can return to:
“When I went back into the kitchen, he had the radio playing music.”
Similarly, the phrase “permanent home” is an editorial expression.
It may fit the mother’s humour.
But she should approve it.
That’s how we work.
The AI can suggest changes.
The storyteller checks them.
And the approved passage becomes the working version.
We aren’t trying to prevent every creative editorial choice.
We’re making those choices visible and accountable.
What Makes the Third Version Read Better?
Let’s examine the mechanics.
The first thing you’ll notice is paragraph length.
Spoken language often appears as enormous blocks of text when transcribed.
A printed page benefits from paragraph breaks.
They help the reader follow changes in action, subject, or emphasis.
The second thing is pacing.
The developed version gives the successful repair its own moment.
“He’d actually fixed the thing.”
That short sentence draws attention to an important part of the story.
The third thing is the ending.
The mother’s frustration about losing her kitchen table becomes the natural conclusion.
The story doesn’t need an additional reflection about love or childhood.
The humour is already there.
And the fourth thing is clarity.
The reader understands the sequence.
The radio didn’t work.
The boy worked on it.
He got it playing.
Then he wanted to keep it on the table.
The story is easy to follow.
Those improvements don’t require invented information.
They require editing.
And that’s the distinction we’re developing.
Good editing reveals the story already present in the material.
It doesn’t need to replace that story with a more dramatic one.
Paragraphs Are Part of Storytelling
Let’s explore paragraph breaks a little further.
A paragraph isn’t merely a collection of sentences.
It helps organise the reader’s attention.
In conversational writing, short paragraphs can be particularly effective.
Consider this:
“He spent all afternoon trying to repair the thing. I must have told him a dozen times that it wasn’t worth the trouble. But he wouldn’t listen. Finally, he got it working. Then he decided to take it apart again.”
Now let’s change the paragraphing.
“He spent all afternoon trying to repair the thing.
“I must have told him a dozen times that it wasn’t worth the trouble.
“But he wouldn’t listen.
“Finally, he got it working.
“Then he decided to take it apart again.”
The events haven’t changed.
But the rhythm has.
The final sentence now lands as a joke.
That can be useful.
However, there’s a danger in taking this approach too far.
If every sentence becomes a separate paragraph, the manuscript may begin feeling fragmented.
Readers need a comfortable rhythm.
Longer paragraphs can provide context.
Shorter ones can create emphasis.
The goal is variety.
Not following a mechanical rule.
Your AI can suggest paragraph breaks, but your own ear remains useful.
Read the passage aloud.
Does it flow?
Does the pause feel natural?
Does the joke arrive at the right moment?
Sometimes those questions are more helpful than any formal writing rule.
Reading Aloud: The Simplest Voice Test
Here’s a technique I particularly recommend.
Read your writing aloud.
Not silently.
Actually speak the words.
You can do this by yourself.
Or have your computer or another device read them to you using text-to-speech.
But if you’re comfortable, reading aloud in your own voice is especially useful.
Why?
Because you’ll notice things that are easy to miss on the page.
Awkward sentences.
Unnatural expressions.
Repeated words.
Transitions that don’t make sense.
And passages that simply don’t sound like something you’d say.
Imagine reading:
“His indomitable determination manifested itself through an unwavering commitment to mechanical exploration.”
Try saying that naturally over a cup of coffee.
Rather difficult, isn’t it?
Now compare:
“Once he decided he was going to fix something, he could spend the whole afternoon working on it.”
That’s much closer to ordinary speech.
It may also be a more accurate description, provided the source supports it.
Reading aloud helps you identify the difference.
It doesn’t mean every sentence must sound exactly like spontaneous conversation.
A book can be slightly more structured.
But the language should still feel comfortable in the narrator’s voice.
If a sentence feels impossible to say naturally, ask whether it really belongs in your book.
You may discover that a simpler version is stronger.
Finding the Balance Between Spoken and Written Language
Now let’s address a common misunderstanding.
Preserving your voice doesn’t mean leaving every spoken sentence untouched.
Imagine a transcript containing:
“And then, well, I think, no, actually, before that, we were at my mother’s house, and, you know, I was going to tell you something else, but anyway…”
That’s natural speech.
But a reader may struggle to follow it.
You can reorganise the passage while preserving the information.
For example:
“We were at my mother’s house when this happened.”
Simple.
Clear.
And, if that’s what the storyteller established, accurate.
The original recording remains available.
We’re not pretending the speaker delivered a perfectly constructed sentence.
We’re simply presenting the memory in a way that readers can understand.
But here’s where judgment matters.
Sometimes a false start or repeated expression is part of the humour.
Perhaps the storyteller says:
“He was absolutely certain he knew what he was doing. Absolutely certain.
“He did not know what he was doing.”
That repetition is useful.
It’s part of the rhythm.
Removing it might weaken the story.
So we don’t automatically eliminate every repeated word.
We ask whether the repetition helps communicate the experience.
Keep the natural qualities that give the story personality. Remove the ones that make it unnecessarily difficult to follow.
That’s the balance.
The Problem With an AI That Wants to Sound Like an Author
Let’s talk about another challenge.
AI often associates book writing with elaborate prose.
You ask it to make something suitable for publication.
And suddenly, the vocabulary becomes increasingly decorative.
Ordinary experiences become profound moments.
People embark on journeys.
Every afternoon becomes unforgettable.
And every chapter discovers the true meaning of life.
But that isn’t the only way to write a good book.
Some of the most engaging personal writing is wonderfully straightforward.
It tells the reader what happened.
Uses specific details.
Preserves humour.
And allows the significance of an event to emerge without announcing it.
You might prefer:
“He brought home another box of car parts. I didn’t even ask how much they cost.”
Rather than:
“His relentless passion for automobiles inspired a series of ambitious projects that reflected the boundless nature of his creative spirit.”
The first sentence contains personality.
It suggests a familiar situation.
And it may naturally invite a story.
The second is mostly general interpretation.
For Project Crystal, we should usually prefer the specific over the grand.
Not because sophisticated language is forbidden.
But because the purpose of the book is to preserve a real life.
And specificity is often what makes the person recognisable.
Writing Emotional Memories Without Inventing Emotion
Now let’s discuss an especially important subject.
Emotion.
Because you’re writing about someone you love, some memories may be deeply meaningful.
Others may be funny.
Some may involve disappointment or disagreement.
And some may be difficult to revisit.
AI can sometimes respond to this material by adding emotional descriptions.
Perhaps you say:
“I remember finding his old jacket in the closet.”
The AI writes:
“As I held the jacket close, an overwhelming wave of grief washed over me, bringing back every cherished memory of the years we’d shared.”
But did you say that happened?
Perhaps not.
Maybe you simply noticed the jacket.
Perhaps you laughed because you remembered how he always left it in the wrong place.
Perhaps you felt sad.
Perhaps you didn’t have a clear emotional response at all.
The AI doesn’t know.
And it shouldn’t decide.
A more careful approach would preserve the original memory.
If you wish to explore the experience, the AI can ask:
“What do you remember about finding it?”
You may choose to describe your reaction.
Or you may prefer not to.
Both are acceptable.
When emotional language appears in the manuscript, it should reflect what you actually expressed or approved.
And it doesn’t need to be exaggerated.
Sometimes one simple sentence is enough.
“I still remember that jacket.”
If those are your words, they may carry more meaning than an entire paragraph of manufactured sentiment.
Let the Funny Stories Remain Funny
This is particularly important for a book about someone’s life.
There will probably be stories that make you laugh.
A ridiculous project.
A family argument.
A practical joke.
A pet doing something outrageous.
A plan that didn’t work.
Those stories deserve to retain their humour.
Sometimes an AI will try to soften them.
Perhaps it believes a memorial book should always sound gentle and reverent.
But a person’s sense of humour is part of their life.
And remembering them with laughter isn’t disrespectful.
Imagine a story about your son accidentally covering the garage floor with motor oil.
Perhaps the family spent an hour cleaning it up.
Perhaps everyone was furious at the time.
And perhaps the incident became a joke years later.
The story might be wonderful exactly as it is.
We don’t need to finish with:
“Looking back, I realised that the true beauty of life lies in embracing imperfection.”
Perhaps the natural ending is:
“And that’s why he wasn’t allowed near my garage without a roll of paper towels.”
If that’s how the storyteller remembers and chooses to describe it, the humour belongs on the page.
The book should preserve the person’s laughter as faithfully as it preserves the more serious moments.
And the AI should never assume that grief must be the emotional destination of every story.
Writing About Imperfection
Let’s explore another important aspect of authenticity.
People aren’t perfect.
They make mistakes.
They get frustrated.
They lose their tempers.
They attempt things that don’t work.
They argue.
They change their minds.
And they sometimes do things that make absolutely no sense to anyone else.
Those experiences can be part of a truthful portrait.
A book doesn’t need to describe every flaw.
And you shouldn’t feel pressured to include painful or private incidents.
But neither should AI automatically polish away everything inconvenient.
Imagine the storyteller says:
“He was stubborn. Once he decided something, there was no changing his mind.”
The AI rewrites it:
“He possessed an unwavering determination that inspired everyone around him.”
Those statements communicate very different ideas.
The first may include affection, frustration, and humour.
The second turns the characteristic into an idealised virtue.
Perhaps the storyteller genuinely admired his determination.
Perhaps she also found him difficult sometimes.
Both can be true.
A more authentic passage might preserve that complexity.
“He could be terribly stubborn. Sometimes it drove me mad. But once he decided he was going to figure something out, he would keep at it until he was satisfied.”
That may be appropriate if the source supports those observations.
The important lesson is that we shouldn’t convert every human imperfection into a compliment.
A recognisable person is more interesting than a perfect character.
And allowing ordinary contradictions to remain can make the writing feel more honest.
The Opening Sentence: Inviting the Reader Into a Memory
Now let’s begin discussing the practical construction of prose.
When you start a story, you need an opening.
But it doesn’t have to be dramatic.
You don’t need to begin with the weather.
You don’t need an elaborate description of the surroundings.
And you don’t need a philosophical statement about memory.
Often, the storyteller’s original opening is perfectly good.
Consider:
“I have to tell you about the time he tried fixing the lawnmower.”
That’s an inviting beginning.
It sounds natural.
It introduces the subject.
And it creates a little curiosity.
Another example:
“Grandpa had this old workshop, and my son absolutely loved being in there.”
That establishes a place and relationship.
Or:
“I don’t remember exactly how old he was when this happened, but I’ll never forget what he did.”
That tells the reader the narrator has an approximate recollection rather than an exact date.
All three are simple.
And that simplicity can be effective.
When reviewing an opening, ask:
Does it introduce the story clearly?
Does it sound like me?
Does it make a promise the memory can fulfil?
Does it contain unsupported details?
If the answer is satisfactory, you may already have your opening.
There’s no need to decorate it merely because it’s the first sentence.
Writing the Middle Without Losing the Reader
The middle of a story is where the events unfold.
This is often where conversational transcripts need the most organisation.
You may have told the ending first.
Then remembered the beginning.
Then interrupted yourself to explain who someone was.
Then returned to the event.
That’s perfectly natural.
But the reader may appreciate a clearer sequence.
Suppose your original memory includes these details:
The family was preparing for a camping trip.
Your son had borrowed some equipment.
A tent pole was missing.
Everyone searched for it.
The pole was eventually found in the garage.
You may have described those events out of order.
A written version can arrange them into a logical sequence.
But it shouldn’t change what happened.
And if the sequence itself is uncertain, we shouldn’t invent precise timing.
Perhaps you don’t remember whether the missing pole was discovered before or after everyone loaded the car.
In that case, the writing can avoid claiming an exact sequence.
This is one of the ways we distinguish narrative clarity from historical invention.
We organise what is known.
We preserve what is uncertain.
And we don’t fill every empty space with a plausible event.
Knowing Where a Story Ends
Here’s a lesson that can save a great deal of unnecessary writing.
A story can end when the memorable event ends.
It doesn’t always need a final paragraph explaining why the experience mattered.
Imagine a story about the family dog eating a birthday cake.
The dog eats the cake.
Everyone discovers what happened.
The family buys another cake.
And for years afterward, the son jokes about whether the dog has been invited to the next birthday.
That’s a perfectly good ending.
It includes the lasting family joke.
The reader understands why the incident remained memorable.
What else do we need?
Probably nothing.
An AI might add a reflection about how laughter brings families together.
But perhaps that weakens the natural ending.
You can remove it.
Or perhaps the storyteller genuinely wants to reflect on the memory.
That’s fine too.
The choice belongs to the author.
The important thing is not adding a moral automatically.
Sometimes the punchline is the ending.
Sometimes a quiet image is the ending.
Sometimes the story finishes with a simple statement.
And sometimes a reflective conclusion is appropriate.
Let the material and your voice guide the choice.
Transitions Between Stories
Now suppose a chapter contains three different memories.
How do we move between them?
The simplest approach is often to use a short sentence that identifies the connection.
For example:
“That wasn’t the only project he attempted in Grandpa’s garage.”
Or:
“The garage also appears in another story I remember.”
Or:
“Years later, we had another incident involving the same old bicycle.”
These transitions may be useful when they’re supported by the source material.
But be cautious about statements suggesting one event caused another.
For example:
“That afternoon would inspire his lifelong passion for building things.”
That’s a significant claim.
Unless the storyteller actually remembers or has evidence of that connection, we shouldn’t use it.
A transition doesn’t need to explain a person’s entire development.
It simply helps the reader move.
And sometimes a section heading is enough.
If you have a collection of short stories, each may begin with a new title.
No elaborate bridge required.
The best transition is the one that makes the reading experience clearer without changing the history.
Writing With Photographs
Photographs may become an important part of your book.
Perhaps you have a picture of the family dog.
A childhood bicycle.
A favourite car.
A holiday.
A birthday.
Or a familiar family gathering.
Photographs can provide wonderful context.
But they also require careful handling.
Imagine a photograph showing your son beside a car.
The AI might write:
“He stood proudly beside the car he’d spent months restoring.”
But what does the photograph actually establish?
It shows him standing beside a car.
Perhaps you know he owned it.
Perhaps you remember him working on it.
Those additional details may be supported by your recollections.
But pride isn’t necessarily established by the photograph alone.
And the length of the restoration project may not be known.
A careful caption could say:
“Standing beside his car, date uncertain.”
Or, if confirmed:
“With the car he worked on during his teenage years.”
Photographs can support memories.
They can also introduce questions.
But they shouldn’t become excuses to invent scenes.
You might ask your AI to help write captions using only the information you’ve supplied.
And always consider privacy and permission before including images of other people.
Especially in a book intended for public distribution.
When Other People Tell the Story
Now let’s consider contributed memories.
Perhaps your son’s friend shares a story.
You weren’t there.
But you find the account wonderful and would like to include it.
You have several options.
You could present the friend’s recollection in their own words, with appropriate permission.
You could summarise it in your narration while clearly attributing the information.
Or you could create a separate section containing contributions from several people.
Each option has a different effect.
Let’s use a fictional example.
Daniel remembers building a homemade trailer with your son.
You remember seeing the trailer fall apart in the driveway.
Those are two perspectives.
You might write:
“I remember the little trailer falling apart on its first trip down the driveway. Daniel later told me they’d spent most of the afternoon trying to get the wheels attached.”
That combines the accounts without pretending the mother witnessed everything.
Now imagine the AI writes:
“I watched them spend the entire afternoon carefully assembling the trailer.”
That would be inaccurate unless the mother actually remembered watching.
This distinction matters.
A coherent narrative doesn’t require pretending that one narrator witnessed every event.
You can preserve attribution naturally.
And when the book includes different voices, readers may appreciate knowing who contributed which memories.
It creates a more transparent record.
Consistency Without Making Every Chapter Sound the Same
Now that we’re preparing to draft a manuscript, we need to consider consistency.
A book generally benefits from a recognisable narrative style.
Readers should feel that they’re encountering the same storyteller across different chapters.
But consistency doesn’t mean uniformity.
A funny chapter may have short sentences and playful pacing.
A quiet memory may use longer, reflective paragraphs.
A factual section may be more straightforward.
And an emotional passage may require greater restraint.
Those differences can be appropriate.
Your Voice Guide should establish the general style.
But it shouldn’t force every story into the same pattern.
For example, perhaps your general preferences are:
Conversational first-person narration.
Clear, direct language.
Humour where it naturally appears.
No invented dialogue.
No exaggerated sentimentality.
No generic inspirational endings.
Those preferences can guide the entire book.
But each chapter may still have its own rhythm.
That’s a good thing.
The variety reflects the variety of experiences being remembered.
And a book about a real person should have room for more than one emotional tone.
Creating a Page-Voice Sample
Before drafting an entire manuscript, here’s something I recommend.
Create a small approved writing sample.
Perhaps two or three pages.
Choose a memory you know well.
Preferably one with enough detail to stand on its own.
Use your Voice Guide.
Ask your AI to create a readable narrative.
Then revise it.
Correct the meaning.
Restore the humour.
Remove invented details.
Adjust the paragraphs.
Read it aloud.
Keep working until the writing feels natural.
When you’re satisfied, save that passage as:
Project Crystal - Approved Voice Sample
This becomes a useful reference.
In future sessions, you can provide the Voice Guide and the approved sample.
Then explain:
“This passage represents the style I’d like to maintain. Please use it as a reference for tone, pacing, and language, while preserving the facts and individual character of each new story.”
That’s much more concrete than telling the AI to “write beautifully.”
You’re showing it an actual example of approved writing.
And because the sample is saved independently, you can reuse it across different conversations and applications.
But remember, the sample is a guide.
Not a template every chapter must imitate exactly.
The source material should still influence the tone and structure of each passage.
The Three-Pass Writing Method
Now we’re going to turn everything we’ve learned into a practical editing procedure.
We’ll call it the Three-Pass Writing Method.
It works particularly well when converting spoken memories into readable prose.
Pass One: Accuracy
Before worrying about beautiful language, make sure the information is correct.
Check the people.
Dates.
Places.
Relationships.
Sequence of events.
Quotations.
And important details.
Identify anything unsupported.
Preserve uncertainty.
Keep the source references available.
At this stage, the writing may still be a little rough.
That’s fine.
Accuracy comes first.
Pass Two: Clarity
Now make the passage easier to follow.
Remove unnecessary repetition.
Clarify confusing references.
Organise the sequence.
Create paragraphs.
Add source-supported context where readers need it.
But don’t introduce new events.
The goal is to help the reader understand the material.
Not to make it more dramatic.
Pass Three: Voice
Now examine how the passage sounds.
Does it resemble your natural storytelling style?
Has your humour survived?
Are there expressions you would never use?
Has the AI added sentimental conclusions?
Does the ending feel natural?
Read the passage aloud.
Compare it with the original transcript.
Make the adjustments that bring the writing closer to your intentions.
These three passes can be performed in separate AI exchanges or during one careful review.
For longer chapters, separating them may make the process easier.
And if a later voice edit changes an important fact, return to the accuracy pass.
The order isn’t a one-way street.
You can revisit earlier steps whenever necessary.
Accuracy protects the memory. Clarity helps the reader. Voice preserves the storyteller.
That’s the heart of the method.
A Reusable Prompt for Finding Your Page Voice
Let’s create a practical instruction you can use with your AI.
This prompt is designed for the moment when you have approved source material and are ready to turn a memory into readable prose.
It is not intended to generate an entire unsupported chapter.
It works best with one clearly identified memory or a small approved source packet.
PROJECT CRYSTAL: PRESERVE MY VOICE
I am writing a personal book based on genuine memories and approved source material.
I will provide:
- An original transcript or approved memory record.
- My Storytelling Voice Guide.
- An approved writing sample, if available.
Your task is to help me turn the source material into readable prose while preserving my natural storytelling voice.
Follow these instructions:
1. Protect Accuracy
Use only factual details supported by the supplied material.
Do not invent events, dialogue, dates, locations, emotions, motivations, sensory descriptions, or outcomes.
Preserve uncertainty when details aren’t known.
2. Preserve My Meaning
Do not turn ordinary experiences into dramatic or inspirational narratives.
Do not exaggerate personality traits.
Do not assign emotional significance I haven’t expressed or approved.
3. Improve Readability
Remove unnecessary repetition and confusing filler words.
Clarify references.
Organise events into a readable sequence when the source supports that sequence.
Use natural paragraph breaks.
4. Preserve My Voice
Use my Voice Guide and approved writing sample as references.
Keep distinctive expressions and humour where appropriate.
Avoid overly formal language unless I’ve requested it.
Do not replace my natural wording merely to sound more literary.
5. Handle Quotations Carefully
Use quotation marks only for wording established by a source or reliably remembered as exact.
Otherwise, use attributed reported speech.
Do not create dialogue.
6. Respect Point of View
Use the narrative perspective I request.
Do not introduce knowledge that the narrator wouldn’t have.
Attribute information contributed by other people.
7. Create Natural Openings and Endings
Begin with the actual memory or source-supported context.
Don’t manufacture dramatic scenes.
Allow stories to end naturally.
Do not add moral lessons or sentimental conclusions unless requested.
8. Identify Editorial Changes
After drafting, briefly identify significant changes in organisation, wording, or interpretation that I should review.
Flag any details requiring confirmation.
9. Wait for My Approval
Treat the result as a draft.
Do not assume that polished writing is accurate writing.
Allow me to correct the passage before saving an approved version.
The goal is clear, natural prose that remains faithful to the original memory and still sounds like me.
You can save this prompt alongside your other Project Crystal tools.
But you don’t need to use its full length every time.
Sometimes a simpler instruction will be enough.
“Please clean this up so it reads naturally, but don’t change the meaning or invent anything.”
That’s a perfectly good starting point.
The more detailed prompt becomes useful when the material is complex or when you’ve encountered repeated editing problems.
Your Chapter Exercise: One Memory, Three Versions
Now it’s your turn.
We’re going to practise moving from conversation to written narrative.
You don’t need to write a full chapter.
One memory is enough.
Step One: Choose a Memory
Select an approved memory record with a clear source.
Choose something you feel comfortable discussing.
A funny or ordinary story may be ideal.
Step Two: Read the Original Transcript
Notice how you naturally told the story.
Pay attention to distinctive expressions.
Repetition.
Pauses.
Humour.
And the order in which the details appeared.
Step Three: Create a Light Edit
Ask your AI to improve readability without changing the structure significantly.
Preserve the original wording wherever practical.
Review the result.
Step Four: Create a More Structured Version
Now ask for a slightly more developed narrative.
Allow clearer paragraphs, transitions, and pacing.
But prohibit invented information.
Compare this version with the original and the light edit.
Step Five: Experiment With Point of View
If you’re unsure which perspective suits your book, ask for a first-person and third-person version.
Keep the underlying information identical.
Notice how the narration changes the reader’s experience.
Step Six: Check Accuracy
Identify every factual claim.
Return to the original recording or memory record.
Look for unsupported additions.
Check whether the AI has changed the meaning.
Step Seven: Read Aloud
Read the versions aloud.
Which sounds natural?
Which preserves your humour?
Which would you feel comfortable sharing?
Mark the sentences that work.
Step Eight: Create Your Approved Version
Choose the best material from the versions.
Make any corrections necessary.
Remove unsupported details.
Preserve expressions you wish to keep.
Step Nine: Save the Result
Save the original source.
Save your approved passage separately.
If the passage represents your preferred style, add it to your Page-Voice Sample collection.
Step Ten: Update Your Voice Guide
Did you discover something new about your preferences?
Perhaps you like shorter openings.
Perhaps you dislike dramatic conclusions.
Perhaps you prefer more of your original wording to remain.
Record that lesson.
You’ve now completed a practical exercise in converting spoken memory into written prose.
And you’ve done it without giving up control of the story.
That’s exactly what we’re aiming for.
What If You Don’t Like Any Version?
Here’s something worth knowing.
You can reject every draft the AI produces.
Perhaps the writing feels too stiff.
Perhaps it sounds too sentimental.
Perhaps the AI keeps missing the point.
That’s fine.
Return to the original.
Ask what the versions changed.
Explain what you dislike.
Or simply write a few sentences yourself.
You can even tell the AI:
“Don’t rewrite this. Just help me identify which parts are difficult for a reader to follow.”
Then make the changes yourself.
The AI doesn’t have to write every word.
Sometimes its most useful contribution is identifying a confusing sentence or asking for missing context.
The finished book can contain passages you wrote entirely yourself.
Passages lightly edited with AI.
And chapters developed through more substantial collaboration.
The important thing is that the source material remains trustworthy and the editorial choices reflect your intentions.
AI assistance is an option, not an obligation to surrender the keyboard.
How Mark’s Voice Became Part of His Books
Let’s return to the origin of this guide.
Mark’s conversations with Eli are rarely formal writing sessions from the first sentence.
He explores ideas.
Asks questions.
Uses humour.
Challenges explanations.
Introduces unusual comparisons.
And occasionally discovers that the best way to explain something is through an example nobody would expect.
That natural conversational style became an important part of the material behind The Meistro Manual.
But transforming conversational material into a book introduced a challenge.
The writing needed structure.
Readers needed clear explanations.
Chapters needed to connect.
And the manuscript required editing.
At the same time, the distinctive conversational character needed to survive.
The solution wasn’t to preserve every false start or repeated phrase.
Nor was it to turn every passage into polished corporate prose.
It was to make editorial decisions about what to clarify and what to keep.
That process wasn’t automatic.
It required review.
Correction.
And eventually a more systematic editorial workflow.
For Project Crystal, we’re carrying those lessons forward.
Your voice may be different from Mark’s.
Perhaps quieter.
More reflective.
More humorous.
More formal.
Or a mixture depending on the memory.
That’s fine.
We’re not trying to create another Meistro.
We’re using the method to help you develop your own book.
And that means your finished prose should reflect how you wish to communicate.
Not simply how an AI prefers to write.
Before the Manuscript Begins
Let’s look at what we’ve accomplished so far.
You began with memories.
You learned to speak naturally about them.
You learned to record and preserve them.
You created transcripts.
Built a Memory Library.
Identified individual stories.
Followed connections.
Explored possible book structures.
And developed a Table of Contents.
Now you’ve learned how to begin translating your natural storytelling voice into prose.
That’s the final piece of preparation before we begin constructing the manuscript.
And it’s an important one.
Because when you ask your AI to draft a chapter, you’re not merely asking for a collection of sentences.
You’re asking it to help create a particular reading experience.
A chapter built from genuine material.
Presented clearly.
Organised deliberately.
And written in a voice you’ve chosen.
That requires three things.
Reliable source material.
A clear chapter plan.
An established writing voice.
You now understand how to prepare all three.
And that’s why we’re ready to move forward.
One Final Thought: The Person Who Would Recognise the Story
Imagine holding a printed page.
It contains a memory about your son.
Perhaps something funny.
Something ordinary.
A story you’ve told countless times.
You begin reading.
The words are clear.
The paragraphs flow.
The sequence makes sense.
But the important part isn’t that the writing looks professional.
It’s what happens when you hear the story in your mind.
You can recognise it.
The people.
The incident.
The humour.
The expressions.
The little details that made the memory worth keeping.
And perhaps you can imagine telling the same story to someone sitting across the table.
That’s what we’re trying to preserve.
Not the exact disorder of spontaneous speech.
Not every pause.
Not every filler word.
But the personality and perspective underneath them.
Because a book can be carefully written without becoming artificial.
It can be polished without becoming generic.
And it can be emotionally meaningful without inventing emotions the storyteller never expressed.
The writing can improve while the memory remains intact.
That’s the balance.
And the more comfortable you become recognising that balance, the more confidently you can work with AI.
You won’t need to accept every suggested improvement.
You won’t have to reject every attempt at rewriting.
You’ll learn to recognise what helps.
What changes the meaning.
What removes personality.
And what makes the story clearer.
That is editorial judgment.
And you’re developing it through practice.
The AI can help you find a sentence.
It can suggest a better order.
It can make a confusing passage easier to follow.
But it cannot decide which version most faithfully represents your personal recollection.
That remains your responsibility.
And your privilege as the storyteller.
Because the book isn’t meant to sound like an artificial intelligence remembering your son.
It’s meant to sound like someone who actually has memories to share.
The AI helps shape the page.
Your memories provide the truth of the telling.
And your voice makes the story yours.
Before You Turn the Page
Let’s review the most important lessons from this chapter.
1. Spoken language and written prose work differently.
Natural conversation contains repetition, interruptions, and shared context. Writing may need additional organisation to help readers follow.
2. Your original voice is the foundation.
Editing should improve clarity without replacing your humour, expressions, or perspective.
3. Choose a deliberate point of view.
First person, third person, and combined approaches create different reading experiences. Each must remain faithful to the available sources.
4. Preserve other people’s voices accurately.
Don’t invent quotations or present reconstructed dialogue as exact speech.
5. Paragraphs and pacing matter.
Careful formatting can make a story more readable without changing what happened.
6. Use openings and endings that belong to the memory.
Don’t manufacture drama or attach a moral simply because the AI expects one.
7. Protect emotional accuracy.
The AI should not assign feelings, motivations, or significance that the storyteller hasn’t supplied or approved.
8. Read your writing aloud.
Your ear can help identify unnatural language, awkward transitions, and passages that no longer sound like you.
9. Create an approved Page-Voice Sample.
A carefully reviewed passage can become a useful reference for later chapters.
10. Use the Three-Pass Writing Method.
Accuracy first. Clarity second. Voice third. Revisit any stage whenever necessary.
And perhaps the most important lesson:
The goal isn’t to make your memories sound as though they were written by a professional author. The goal is to make them readable while preserving the person who is telling them.
We’ve now reached the end of Part IV: Discovering the Book.
You’ve explored the possible forms your project could take.
You’ve learned how to identify strong story material.
You’ve developed a working Table of Contents.
And you’ve begun establishing the writing voice that will carry the memories onto the page.
Now we’re ready for something you’ve been preparing for since the beginning.
Actually constructing the manuscript.
In Chapter 19: From Memory to Manuscript, we’ll begin Part V.
We’ll take the approved material you’ve collected and show how to prepare it for chapter drafting.
We’ll examine how to assemble a chapter source packet, provide useful instructions to your AI, produce a first narrative draft, and review that draft without losing track of the original memories.
And we’ll begin making an important distinction between three things:
A collection of source material.
A draft chapter.
And an approved chapter ready to join the manuscript.
Because we’re no longer simply gathering stories.
We’re beginning to build the book itself.
But there’s one principle we’re carrying forward into every chapter.
We won’t ask the AI to invent what we haven’t remembered.
We’ll work with the stories we have.
We’ll explore them.
We’ll organise them.
And when the material supports it, we’ll turn them into chapters.
One at a time.
With room for correction.
Room for discovery.
And room for the storyteller’s voice.
The archive has given us the material.
The Table of Contents has given us a plan.
And your Voice Guide has given us a way to approach the writing.
Now we’re ready to turn those carefully preserved memories into the pages of your book.