The Writing Process: Invention, Drafting, Revision

A blank page is rarely a thinking problem. More often it is a sequencing problem—an attempt to discover an idea, organize it, phrase it well, and punctuate it correctly all inside the same sentence, on the first try. Nobody can do that. The writers who look effortless have simply separated the jobs and learned to move between them on purpose. What follows takes the writing process apart into its working parts: invention, which generates raw material; shaping, which turns a pile into a plan; drafting, which trades polish for momentum; and revision, which re–sees the whole thing once it exists. Every move comes with a worked example, because process advice you cannot picture is just encouragement. Expect to write badly on purpose for at least an hour. It is the fastest known route to writing well, and it is the part most students skip.

Section 1The Process Is a Loop, Not a Line

What writers actually do

Most of us were handed a diagram in high school: brainstorm, outline, write, proofread, submit. Five boxes, one arrow, no going back. It is a tidy picture and it describes almost nobody. When Linda Flower and John Hayes watched writers think aloud while composing in the late 1970s and early 1980s, what they found was not a line but a set of processes—planning, translating thought into language, and reviewing—that writers interrupt and re–enter constantly. A writer drafting paragraph four discovers the real argument and goes back to rewrite the introduction. A writer proofreading a conclusion notices the essay has been answering the wrong question for eight pages. That doubling back is not a symptom of poor preparation. It is the mechanism.

The technical word is recursive: a process that calls itself, that loops. You plan, then draft, then discover something that forces you to plan again. Sondra Perl, studying unskilled college writers in 1979, found that they too were recursive—but they looped back for the wrong reasons, stopping every few words to check spelling instead of stopping every few paragraphs to check meaning. The difference between an experienced writer and a struggling one is not that one loops and the other does not. It is what each one loops back to fix.

One honest caveat before you commit to any diagram, including this one. Since the 1990s, a group of scholars sometimes labeled “post–process” has argued that there is no single universal writing process to teach, that what counts as good practice depends heavily on the situation, the discipline, and the writer. They have a point. Treat everything in this chapter as a set of tools with known uses, not as a ritual. If you already have a method that reliably produces work you are proud of, keep it and steal only what you lack.

Why the linear myth is expensive

The linear model does real damage to your calendar. Students who believe writing is one pass typically spend something like eighty percent of their available time producing the draft and ten minutes running a spell–check, which they call revising. That allocation is exactly backwards from what experienced writers report. Nancy Sommers, comparing student writers with experienced adult writers in 1980, found that students mostly changed words—swapping a synonym, tightening a phrase—while experienced writers cut whole paragraphs, resequenced sections, and sometimes threw the draft out and started from a better question.

A second cost is invisible: the linear model tells you that if you cannot start, you are not ready. So you research more. You reread the prompt a ninth time. You wait to “feel clear.” But clarity is usually a product of writing, not a prerequisite for it. The sentence you cannot compose in your head becomes composable the moment you write a bad version of it and look at it.

Here is a defensible starting budget. It assumes a 1,500–word essay with a two–week window and assigned sources rather than outside research, and it is meant to be adjusted, not obeyed.

Making your own process visible

You cannot fix a process you have never watched. For one assignment, keep a process log: a three–line note each time you sit down recording the date, the minutes spent, and what you actually did (“25 min, reread the article, wrote nothing”). Most students discover the same two things. First, they have more total minutes than they thought, distributed in fragments they were dismissing as too small to use. Second, an enormous share of those minutes goes to a stage they would not have named—usually rereading sources, or rewriting the first paragraph for the fifth time.

For an adult student working full time, that second finding is the expensive one. A nurse coming off three twelve–hour shifts does not have a free Sunday; she has four twenty–minute windows scattered across a week. Twenty minutes is useless for polishing an introduction and extremely useful for a timed freewrite, for reverse–outlining two pages, or for fixing the six sentences you flagged yesterday. The point of knowing your process is that you can then match the task to the window instead of waiting for a window that never opens.

Section 2Invention: Making Something Out of Nothing

Freewriting and looping

Invention is the classical name—from rhetoric’s inventio—for the work of finding something to say. Its modern workhorse is freewriting, popularized by Peter Elbow in Writing Without Teachers (1973): set a timer for ten minutes and write without stopping, without editing, without deciding whether any of it is good. If you run dry, write “I have nothing to say about this” until something else arrives. The rule that makes it work is the one everyone breaks: do not stop moving the pen. Stopping invites judgment, and judgment is the thing you are trying to postpone.

Here is a real ten–minute freewrite on the prompt “Write about a change in how people communicate.” It has been trimmed but not cleaned up:

Everybody says phones ruined conversation, I don’t know if I believe that, at work we text constantly and it’s fine, actually it’s better than fine because you can’t page someone in the middle of a code. Blah. What else. My mother won’t text, she leaves voicemails, three minutes long, she says the whole thing like a letter, Dear Akash, love Mom, at the end of a voicemail. That’s weird and also kind of beautiful? She learned one technology and is applying its manners to another one. I do this too probably. Email signature. Nobody signs a text.

Almost all of that is throwaway. But one clause is alive: she learned one technology and is applying its manners to another one. That is a genuine idea, and it was not available before the pen started moving. The technique for mining it is looping: circle the hottest sentence, copy it to the top of a clean page, and freewrite ten more minutes starting from there. Three loops will usually take you from “phones ruined conversation” to an essay about how the manners of one medium outlive it and get carried into the next, so that every new technology arrives wearing the etiquette of the last one—which is a thesis, and a good one.

Structured heuristics: when the blank page needs handles

Freewriting suits some brains and frustrates others. If yours goes silent under an open prompt, use a heuristic—a systematic set of questions that guarantees you will produce something whether or not you feel inspired. Four are worth knowing.

The journalists’ questions (who, what, when, where, why, how) are the bluntest and still the most reliable. Applied to “hospital alarm fatigue,” they yield: Who hears the alarms and who sets the thresholds? What counts as a false alarm? When in a shift does desensitization set in? Where is this worst—telemetry, ICU, med–surg? Why are default settings so sensitive? How would you measure the harm? Six questions, and the “why” one has already located the arguable claim: that default thresholds are set to protect manufacturers from liability rather than to fit bedside conditions, and that nurses absorb the cost of that choice.

Clustering (also called mind–mapping, developed as a writing technique by Gabriele Rico) puts your topic in a center circle and radiates associations outward, branching wherever a branch wants to branch. It suits visual thinkers and it is unusually good at revealing that your topic actually contains three topics. Cubing, usually credited to Gregory and Elizabeth Cowan, asks you to write briefly about a subject from six sides: describe it, compare it, associate it, analyze it, apply it, argue for or against it. Cubing is the cure for the essay that only ever does one of those six things. And the topic ladder—asking “an example of what?” to climb up and “for example?” to climb down—is the fastest way to find the altitude at which your topic is actually arguable.

Invention by reading, talking, and waiting

Invention is not confined to the blank page. Reading is invention when you read with a pen: mark the sentence you disagree with, the term the author never defines, the example that proves less than claimed. Chapter 2 treats this as analysis; here it is a generator. The most reliable essay engine in college is the sentence “X argues A, and A holds for most cases, but not for — ” followed by the case you know from your own life or work.

Talking is invention too, and it is criminally underused by students who write alone at midnight. Explain your topic aloud to someone who does not share your field—a roommate, a partner, the voice memo app on your phone. Two useful things happen. You automatically drop jargon, which is the fastest route to clear prose. And the listener asks the question you were quietly avoiding, which is almost always the question your essay needs to answer.

Finally, incubation. Graham Wallas described creative work in 1926 as preparation, incubation, illumination, and verification, and the second stage is the one students refuse to schedule because it looks like doing nothing. It is not nothing. After a hard invention session, walk away for a day. The idea you could not force at the desk tends to show up in the shower, on the drive, or at the end of a shift. Give it somewhere to land: keep a note app open, because an unrecorded illumination is indistinguishable from one you never had.

Section 3Shaping: From Pile to Plan

The working thesis is a tool, not a promise

A working thesis is the claim you write down so that you have something to argue with. It is provisional by design. Chapter 4 will treat the finished thesis as a rhetorical instrument—how to make it arguable, specific, and load–bearing. What matters here is narrower and more practical: you need a temporary claim early, because an outline organized around a topic produces an essay that wanders, while an outline organized around a claim produces an essay that goes somewhere.

Getting there is mostly a matter of descending a ladder. A topic is a noun phrase. A question is a topic with a verb and a controversy. A working thesis is an answer to that question that a reasonable person could dispute. Watch the movement:

Notice the shape of the good ones. Each contains a “less X than Y,” an “only when,” or a “trades A for B.” Those structures force a distinction, and a distinction is arguable. A sentence like “Nurse burnout is a serious problem with many causes” is not a thesis; it is a topic wearing a sentence costume, and nobody on earth disagrees with it.

Scratch outlines: give every paragraph a job

A scratch outline is a half–page plan, written for you and nobody else, that lists what each chunk of the essay must accomplish. The critical discipline is to write jobs, not subjects. “Paragraph 3: staffing data” is a subject, and it will produce a paragraph that reports numbers and stops. “Paragraph 3: show that the numbers cited by hospital administrators measure hours, not attention” is a job, and it tells you when the paragraph is finished.

Here is a scratch outline for the first thesis in the table above:

1. Open with a concrete scene—a new nurse’s third night shift—so the abstraction has a body.
2. State the standard explanation (burnout is caused by workload and cured by resilience training) fairly enough that a reader who believes it feels represented.
3. Complicate it: show that turnover clusters in the first 24 months and on night shifts, which workload alone does not predict.
4. Make the real claim: what is missing is not stamina but a more experienced person within reach at 3 a.m.
5. Handle the strongest objection—that night differentials already compensate for this—and explain why money does not substitute for supervision.
6. Land on what would change if the claim were taken seriously: schedule for experience mix, not just headcount.

That is six jobs and perhaps ten minutes of work, and it converts drafting from invention into transcription. It also survives contact with reality, because it is loose enough to abandon. If paragraph 3 turns out to be the essay, you have lost ten minutes rather than a weekend.

Testing the plan before you spend a draft

Three cheap tests will catch most structural failures while they still cost minutes instead of hours. Run them on the scratch outline, out loud if possible.

The first is the so–what test. For each item, finish the sentence “This matters because …” If you cannot, the paragraph is decoration. The second is the order test: read your list of jobs backwards and ask whether any item depends on information that arrives later. A reader who meets your term “experience mix” in item 3 but does not get it defined until item 5 will be lost for two paragraphs and will blame you, not themselves. The third is the nutshell test: state the whole essay in one spoken sentence to a person, not a page. If the spoken version needs four sentences and a diagram, the plan is doing two essays at once, and you should pick one.

A fourth test applies only to arguments, and it is the most useful of them: find the strongest version of the position you are against and put it in the outline as its own job. Not a straw man—the version an intelligent opponent would actually defend. If the outline has no such item, the essay is a monologue, and readers can feel that even when they cannot name it.

Section 4Drafting: Speed First, Quality Later

Permission to write badly

In Bird by Bird, Anne Lamott gives first drafts a memorably profane name and insists that every writer she knows produces them. The idea is not that low standards are virtuous. It is that generating and evaluating are different cognitive jobs that interfere with each other, and running them simultaneously slows both to a crawl. Your inner critic is not wrong. It is early.

The practical form of this is the zero draft: a complete pass, beginning to end, written fast, in which you are allowed to write “[find a better example here]” and keep going. Two rules make it work. Never stop to look something up—type “[CITE]” and move on, because a research detour costs about twenty minutes of momentum for ten seconds of fact. And never reread from the top; rereading is the most socially acceptable way to stop writing.

Compare the two openings below. Both come from the same writer ten minutes apart, working from the freewrite in Section 2. The first is a “careful” opening produced by writing and judging at once. The second came out of a zero draft with judgment switched off.

Careful version: “In today’s modern society, the issue of communication has become increasingly complex due to the rapid advancement of technology, which has affected many different aspects of how individuals interact with one another on a daily basis.”

Zero–draft version: “My mother signs her voicemails. Three minutes of talking and then, at the end, ‘love, Mom,’ as though the phone were paper.”

The careful sentence took longer and says nothing. The fast one has an image, a specific person, and an implied argument. This is the ordinary result, not a lucky one: judgment applied too early filters out the specific and lets through the generic, because the generic is what feels safe.

Drafting out of order, and in fragments

Nothing requires you to write the introduction first, and there are good reasons not to. An introduction promises what the essay will do, and you do not reliably know that until the essay exists. Write the paragraph you understand best—often the middle one with the concrete example—and let it teach you what the surrounding paragraphs need to say. Many experienced writers draft the introduction last and the conclusion second–to–last, on the theory that a real conclusion is only discoverable at the end of the thinking.

Working in fragments requires one extra habit: leaving yourself a landing strip. Before you close the file, type a bracketed note at the point where you stopped, in the imperative: “[Next: explain why the differential does not fix supervision. Use the 3 a.m. example again.]” Ernest Hemingway’s version of this advice was to stop while you still know what happens next. Restarting cold can eat half of a twenty–minute window; the note removes most of that cost.

Two more habits that pay for themselves. Draft with the sources closed, working from your own notes, which prevents the accidental patchwriting that Chapter 10 treats as an integrity problem. And name your files with dates rather than adjectives—essay2_0412.docx, not essay2_FINAL_reallyfinal.docx—so that cutting a paragraph feels reversible and you actually do it.

Diagnosing a stall

“Writer’s block” is a name for a symptom, and treating it as a single condition is why the usual advice fails. Mike Rose’s 1980 study of blocked and unblocked college writers found something practical: blocked students were frequently obeying rigid rules they had internalized from earlier schooling—never begin with “I,” every paragraph must have exactly five sentences, the thesis must be the last sentence of the first paragraph—and applying them inflexibly to situations where they did not fit. Unblocked writers held the same conventions as flexible guidelines. The block was not a personality trait; it was a bad rule enforced too strictly.

So diagnose before you medicate. Most stalls have a locatable cause, and the right move depends entirely on which one you have.

Section 5Revision: Re–seeing the Whole Thing

Four different jobs, in this order

The word revision means re–seeing. It is not tidying. Sommers’s 1980 comparison found that student writers described revision almost entirely in the vocabulary of word choice—“finding a better word,” “crossing out and doing over”—while experienced writers described a search for a shape the argument had not yet found. If nothing in your draft has moved, been cut, or been added, you did not revise. You edited.

Those are genuinely different jobs, and running them in the wrong order wastes enormous effort. Every hour spent perfecting the rhythm of a paragraph you later delete is an hour gone. Do the biggest thing first.

Two practical notes. Put distance between drafting and pass one—overnight if you have it, an hour and a walk if you do not. And change the sensory channel: print it, or read it aloud, or paste it into a different font. Your eye has memorized the screen version and will helpfully show you the essay you meant to write instead of the one on the page.

The reverse outline

The single highest–yield revision tool is the reverse outline (sometimes called a descriptive outline). Take the finished draft and, in the margin beside each paragraph, write two things: the paragraph’s topic in three or four words, and its job—what it does for the argument. Then transfer just those notes to a separate page and look only at that page. You are now reading the skeleton, and skeletons do not hide.

Here is a reverse outline of a real student draft on hospital alarm fatigue, with the diagnosis in the third column.

Nothing in that table required a reader, a tutor, or a grading rubric. It required fifteen minutes and a willingness to describe what each paragraph actually does rather than what you intended it to do. The revision plan writes itself: cut ¶1, promote ¶7 to the thesis position, merge ¶4 and ¶5 under that claim, delete or rewrite ¶6 because it contradicts the evidence, and rebuild the conclusion.

Sentence–level moves, once the structure is settled

Only after the shape holds should you touch sentences, and even then a handful of moves account for most of the improvement. Chapter 11 develops these properly; what follows is the working subset, with the weak version and the repair side by side.

Watch what did not happen in the third row. The revision did not delete the uncertainty; it made the uncertainty specific. Vague confidence and vague doubt are both failures of precision. A reader trusts the writer who says exactly how sure they are.

A revision protocol you can actually run

Print this, or keep it as a checklist. On a 1,500–word essay it costs about the three and a half hours the budget in Section 1 assigns to revision and editing, and it works because it forbids you to fix things out of order.

1. Wait. Do not open the file for at least twelve hours after finishing the draft.
2. Read once, straight through, pen down. No fixing. At the end, write one sentence: “This essay argues that …” If you cannot, that is the finding.
3. Reverse–outline. Topic and job for every paragraph, on a separate page.
4. Decide the surgery. Cut, merge, move, add—write the list before you touch the draft. Expect at least three moves; if you have zero, look harder.
5. Operate. Do the structural changes only. The prose will look worse afterwards. That is correct.
6. Rebuild the seams. First and last sentences of each paragraph, plus the introduction and conclusion, which now have to match the essay that exists.
7. Edit sentences using the six moves above. Read aloud; your ear catches what your eye forgives.
8. Proofread from the last sentence backwards, which defeats comprehension and therefore exposes typos.
9. Check the mechanical list: citations, formatting, file name, word count, and every requirement literally stated in the prompt.

One caution about outside help, and it is the honest version. Readers—classmates, a writing center tutor, a family member—are extraordinarily useful for one thing: telling you where they got confused. They are less reliable at telling you how to fix it, and you are under no obligation to take their solution. Ask for the location of the problem, not the prescription. The same discipline applies to any automated feedback: a tool can flag a passive construction, but only you know whether the actor deserves to be hidden. Chapter 14 develops peer review and portfolio revision at length; what matters now is the habit of revising before anyone else sees the draft, so that your readers can spend their attention on your argument instead of your typos.

Words worth knowing