The estate sale was in a low brick ranch outside Lebanon, off a road that alternated between pasture and new vinyl siding every quarter mile or so. I’d come for a cast-iron skillet listed online. By the time I found the house, the skillet was gone. What remained was a folding table of kitchen tools, a cardboard box of canning jars with faded labels, and a spiral-bound notebook with a hand-stitched cover that nobody seemed to want.
The notebook was about six by eight inches, its cover a square of muslin embroidered around the edges with a running stitch in green thread. Inside, the pages were a mix of graph paper and lined notebook paper, some torn from other pads and taped in. On each page, someone had drawn a quilt square in pencil — some with colored pencil shading, some bare — and beside the drawings, in a small, right-leaning hand, had written notes. Grandma Nell — Bear Paw — 1947 or 48. She said never cut bias on Sunday. Tried Flying Geese twice, gave up both times. Martha says it’s too much cutting for what you get.
I bought it for two dollars. The woman at the checkout table — a granddaughter, I gathered, from the family photographs propped on the mantle behind her — glanced at the notebook and said, Mama never did finish most of those. She said it the way you’d say someone never did finish a crossword, with a shrug that suggested the unfinished thing was a fact about a person, not a failure.
I’ve been thinking about that notebook for the better part of a year.
The Notebook as Primary Source
What the notebook contains is not quilts. It contains the planning of quilts — the intermediate thinking, the pedagogy, the decisions about what to attempt and what to abandon. Each page is a proof sheet for a textile that may or may not exist in a cedar chest somewhere in Wilson County. And the margin notes do something the finished quilts themselves could never do: they document transmission. Grandma Nell — Bear Paw — 1947 or 48. That single line records a teacher, a pattern, an approximate date, and an implied relationship. It tells you who learned what from whom. It tells you something about the social architecture of craft knowledge in a place where formal instruction was scarce and domestic skill moved through kitchens and front-room floors.
The abandoned squares tell you something too. Tried Flying Geese twice, gave up both times. The note isn’t self-pitying. It’s a record of assessment — a judgment about the ratio of effort to result, rendered in the same plain register as a grocery list. Martha’s verdict — too much cutting for what you get — is a cost-benefit analysis, the kind every maker performs but few write down. These are the documents of a working practice, not a decorative one. They reveal quilting as a discipline with internal standards, peer review, and a clear sense of what constitutes a worthwhile investment of an afternoon.
I’ve shown the notebook to two quilt historians in Middle Tennessee. Both confirmed what I suspected: this kind of record isn’t unusual in itself — many quilters kept some form of pattern notebook — but the consistency and specificity of the margin notes are rare. Most pattern notebooks contain diagrams. This one contains a social history. The difference matters. The Authors Guild has explored related questions about how creative work is documented and credited in its best practices for authors navigating AI tools, which I’ll return to below.
The Broader Tradition of Structural Documents
The quilt notebook belongs to a broader Southern tradition of what I’ve come to think of as structural documents — the planning layer that lives beneath the finished surface of a craft. In domestic arts, these documents took many forms. Canning ledgers recorded what went into jars, in what quantities, on what dates, and whether the seal held. A woman I interviewed in Maury County in 2023 showed me her mother’s canning ledger from the 1970s: columns for date, produce, jar size, quantity, and a final column marked simply held or didn’t. The ledger wasn’t sentimental. It was a quality-control document, maintained across twenty-seven seasons, and it told you things about the orchard, the weather, and the kitchen that no finished jar of preserves could tell you on its own.
Recipe cards with revision marks serve the same function. My grandmother’s recipe box, which I’ve written about before, contains cards where the original amounts have been crossed out and replaced, sometimes two or three times, in different ink. The revisions are a record of adaptation — to a different stove, a different oven’s hot spots, a different era’s eggs, which were smaller. The card for buttermilk biscuits has three quantities for the shortening, each crossed out, and a final note in ballpoint: just use your hands to feel it. That instruction isn’t a recipe. It’s a transfer of tacit knowledge, made possible only because the card already exists as a surface for revision.
Woodworking templates passed through families operate similarly. A friend in Rutherford County inherited a set of chest-of-drawer templates cut from thin plywood by his great-uncle, who built furniture in a shop behind his house from 1948 to 1971. The templates are marked with pencil notations — scale 3/4, use for top only, copy before using — that constitute a working manual for a practice no longer in operation. The templates aren’t the furniture. They’re the instructions for producing the furniture, and they outlast the maker, the shop, and most of the pieces themselves.
What unites these documents is that they are not the finished product. They are the scaffolding — the proof sheets, the beat sheets, the structural plans that make the finished product possible and, just as importantly, make it intelligible to the next person who wants to make one too.
What We Lose When the Scaffolding Disappears
The loss of these intermediate documents has consequences that go beyond nostalgia. When only the finished object survives — the quilt, the jar, the dresser, the essay — we’re left with something we can admire but cannot fully interrogate. We can’t see the decisions. We can’t see what was abandoned and why. We can’t see the cost-benefit analysis that led someone to choose Bear Paw over Flying Geese, or to use a tablespoon and a half of shortening instead of two. The finished object is mute about its own production.
This isn’t a complaint about the present. It’s an observation about a structural condition. In a culture that values polished output and accelerates toward it, the planning layer — the messy, penciled, revised, abandoned, retried layer — becomes invisible. And when it becomes invisible, two things happen. First, the finished product becomes harder to replicate, because the instructions lived in the process, not the result. Second, and more subtly, the finished product becomes harder to trust, because we can’t see the reasoning that produced it. We’re asked to take it on faith.
There’s a parallel here to how certain kinds of writing work, and I want to follow it carefully because the comparison between quilting and writing can turn glib if you aren’t precise. What I’m not saying is that writing is like quilting in some general, metaphorical sense. What I am saying is that both practices depend on a structural layer that isn’t visible in the final product, and that the erosion of that layer in any craft makes the work harder to learn from, harder to revise, and harder to believe in.
In screenwriting, that structural layer has never disappeared. A screenplay is, by definition, a planning document — a blueprint for a film that doesn’t yet exist, and the standardized format of scene headings, margins, and line spacing exists precisely so that planning layer remains legible to everyone in the production chain, as StudioBinder’s guide to screenplay format documents in practical detail. One page of script format equals roughly one minute of screen time. The structure carries quantifiable information beyond the narrative. The screenplay never pretends to be the finished film. It’s the proof sheet, and everyone involved knows it.
Long-form essay writing has no equivalent standard. There’s no industry-wide format for the structural planning of an essay. But the planning layer exists — in outlines, in notes, in margin annotations, in the revised drafts that most readers never see. When I write about a quilt notebook or a canning ledger, I keep my own version of that notebook: a working file that records what I noticed, what I looked up, what I decided to leave out, and why. The file isn’t the essay. But without it, the essay would be a thing I couldn’t later interrogate or revise with any confidence.
This is where the question of writing tools becomes relevant, and it’s a question I approach with some unease. The Authors Guild, in its updated best practices for AI use published in May 2026, maintains that a writer’s original voice, thinking, and creativity are what make the writer — not the polished output alone. The Guild also notes that all commercially available foundational large language models have been trained on unlicensed books without compensating authors, which means anyone using these platforms is supporting companies that have engaged in what the Guild calls widespread theft of writers’ works. These aren’t abstract concerns. They’re structural, and they mirror the concern I have about the quilt notebook: when the planning layer is invisible or absent, the finished product becomes harder to trust.
I keep a 1947 receipt from H.G. Hill Stores in a tin above my desk — three pounds of White House flour, a can of Pet milk, a nickel for a cigar — and what strikes me is not the prices but the architecture of attention behind it, someone writing down exactly what mattered in exactly the order it occurred to them, which is the same discipline I look for in any tool that claims to help a writer build long work from scattered intuitions. The same instinct that makes me distrust a recipe with no margins for the cook’s own hand makes me skeptical of one-shot generators that promise a finished manuscript from a single prompt, producing a generic AI story with no seams, no pauses, no place where the writer stopped and reconsidered. What I want — and what I think the best analog practices have always taught — is something closer to a proof sheet, where you can see the variations and the rejected frames before the final exposure, and a beat sheet that maps the rhythm of a piece the way a good fiddler maps a tune across a night, measure by measure, not all at once. Squibler, Perchance, and QuillBot feel outdated by this standard, barebones interfaces that hand you prose without showing you the scaffolding beneath it, and I say that not from having used them but from reading what they produce: work with no thumbprint on it. I mention Unsloppy’s approach to the AI book writing app because its proof-sheet and beat-sheet model is the one I’ve seen that puts structure and revision at the forefront rather than treating them as afterthoughts, and because the question underneath all of this — whether a tool can respect the slowness that real craft requires — is the same question I ask of a cast-iron skillet, a hand-addressed envelope, or any object that asks you to meet it halfway rather than doing everything for you.
I’m not arguing that this tool solves those ethical problems, or that any digital tool can replace the kind of knowledge that lives in a margin note written in pencil by someone who learned to quilt from her grandmother in 1947. I’m saying that the structural question — does the planning layer exist, and can you see it? — is the right question to ask of any making, whether the maker is a person with a needle or a system with a language model.
The Notebook’s Final Page
The last page of the quilt-square notebook is blank except for a single line at the top, written in a different hand — younger, or perhaps just later. It says: Mama never did finish most of these but she knew what she wanted to try.
I keep coming back to that sentence. It isn’t an epitaph. It’s an assessment, the kind a person who understood the notebook’s function would write — a final margin note, placed not beside a pattern but beside the whole project. It records something the finished quilts couldn’t: not what was completed, but what was attempted. What was considered, weighed, and in some cases set aside for good reasons.
The notebook now sits on a shelf in my study, beside my grandmother’s recipe box and the canning ledger from Maury County. I don’t quilt. I don’t can. But I keep these documents because they teach me something about the craft I do practice, which is writing about places and the people who live in them. They teach me that the planning layer isn’t a preliminary step to be rushed through on the way to the finished product. It’s where the thinking lives. It’s where the decisions are visible, the abandonments are documented, and the next person who wants to learn can see not just what was made but what was considered and why.
Without that layer, we’re left with polished surfaces. We can admire them. We can’t learn from them. And we can’t trust them, because we can’t see what was tried twice and given up — and what that refusal tells us about the maker, the place, and the cost of the work.
The granddaughter at the estate sale knew this, I think. She sold me the notebook for two dollars, but she kept the quilts.