The Voice of the Book

Most photo books are written by accident.

The photos get months of attention. The words get twenty minutes at the end — a title because the order form asked for one, captions typed at midnight in whatever mood midnight offered. October wrote the cover. A tired Tuesday wrote chapter three. Read the finished book aloud and you can hear the seams: six moods, one book, nobody in charge.

It matters more than people expect, because words in a photo book do not decorate the photographs. They tell the reader how to hold them.

One photo, three ways

Here is a photograph. A father at the stove on a Sunday morning, spatula raised, hair undone, caught mid-sentence. At the table behind him a child holds a fork straight up like a flag. Flour on the counter, light through the window.

Now caption it.

“Sunday mornings belonged to him.”

“February 2021. Pancake Sunday.”

“The chef takes no requests.”

The photograph did not change. The book did. The first line is already an elegy: it tells you this mattered, that it might be over, that you should arrive feeling something. The second stands back and keeps the record, trusting you to find the feeling yourself. The third keeps the family’s actual joke, and for anyone who was in that kitchen it will be the only caption that sounds true.

None of them is wrong. But a reader meets a different father in each one, and a book that mixes all three meets nobody in particular.

There is a fourth kind of caption, the kind that belongs to no family at all: “Making memories!” If a line could be printed on a mug, it cannot carry a page.

The five voices

Most family books speak in one of five registers, and naming them makes the choice easier.

Warm leans close. It says what the moment meant, and it is the natural voice for a book made as a gift. Plain is harder than it looks: it says what is there and stops, and the restraint reads as confidence. Handed the pancake photo, plain writes “He made pancakes every Sunday” and considers the job done. Playful keeps the jokes, which makes it the truest voice for some families and a borrowed coat for others; it only works when the jokes are really yours.

Lyrical listens for light, weather, and the passing of time, and needs discipline; unwatched, it goes purple by page ten. Its pancake caption is “Flour in the air, the whole week still ahead,” and one of those per chapter is plenty. Documentary keeps names, dates, and places, and almost nothing else. It ages best of the five, because in thirty years the dates are what everyone opens the book to find.

When we taught Enchanta to write, these five became real specifications — a short page of what each voice may say, and a longer page of what it may never say. A voice, it turns out, is mostly its refusals.

One voice, all the way through

Which of the five is yours? Usually the one you already use. Listen to yourself show someone these photos on your phone; whatever register comes out is the book’s. The occasion can overrule it: a gift book drifts warm, a family yearbook sits comfortably in plain or documentary, and a book about someone who is gone is usually safest in plain, because restraint holds grief better than lyric does. When in doubt, take the voice you can keep up for forty pages without strain.

Then choose it before you write anything, and let it run the whole book: the title, the chapter names, the intro if there is one, every caption down to the last page.

You can hear what the choice controls in something as small as a chapter name. “Summer.” “The Summer We Moved.” “June–August 2023.” Same chapter, three voices, and the reader is being promised three different books before a single photo appears.

The test for a finished draft is simple. Read three captions aloud, pulled from three different chapters. If they sound like one person talking, the voice held. If they sound like a committee, the book will feel assembled rather than written, even though every individual line is fine.

The practical trick: write all the words in one sitting, in order, quickly. Not because speed improves sentences (it doesn’t), but because one sitting has one mood, and mood is most of what a voice is. Fix the grammar tomorrow. The mood will still be there.

Pronouns are part of the same decision. “We drove north” and “Maya drove north” are different books; pick one person to be, and stay them.

If the book has an intro, write it last, even though it sits first. Sixty words at the front carry more voice than any other page, and by the time the captions are done you finally know how this book talks. An intro only has three moves to make: name the span, name the people, say one true thing you want kept. “The year we rented the blue house. Four of us, one dog, a kitchen too small for how much happened in it.” That is a whole intro. Nobody misses the paragraph it replaced.

The pages that say nothing

Most pages need no words at all.

A strong photograph, given room, is already speaking, and a caption that repeats it (“Beach day!”) only proves the writer wasn’t listening. The photo said it better.

Words earn their place when they hold what the photo cannot: which cousin is which, before that knowledge lives in only one person’s head. The date. The name of the diner that is gone now. The thing said just out of frame that made everyone in the picture laugh.

Sometimes the rescue runs the other way. A photograph that looks like nothing, a gray parking lot, earns its page because the caption knows why it mattered: “The lot where we waited three hours for the tow truck, and she finally learned chess on the hood of the car.” Without the line, that photo gets cut. With it, it might be the page people stop on.

Chapter marks do the same work at the threshold. “Bergen, the wet week” over a chapter’s first page buys ten silent pages behind it; the reader carries the where-and-when in with them, and no caption inside has to repeat it.

So the working rule is narrow: caption what the photograph cannot say, and nothing else. Some chapters end up with one caption. Some end up with none, and in a book with a real voice, that silence reads as composure. The reader never wonders whether you ran out of things to say. They can feel that you chose not to.

Two hundred words

Count the words in a well-made photo book and the number is startling.

A 40-page book of seventy photographs might carry a title of four words, six chapter names, a sixty-word intro, and fifteen captions averaging eight words each. Two hundred words, give or take. A whole year, told in fewer words than this essay has spent so far.

This is why voice matters. At two hundred words, every line is load-bearing. There is no room to warm up, no paragraph two to recover in. A book that only speaks two hundred words cannot afford to mumble.

It also means the writing is the smallest job in the whole book: two evenings, not a winter. Small, and decisive. Two hundred words set the temperature of two hundred photographs.

And if choosing a voice feels like one more decision you don’t have the energy for, notice that you have already made it. Every time someone asks about the pancake photo, you answer in it — warm some nights, plain on others, the joke when the right person is listening, but reliably yours. The book’s voice was never missing: it is how you talk when the photo is on the table and someone you love asks about it.

Write it the way you tell it.