How to Write the Many-Sided Essay
You have a topic. You also have an opinion about it, sitting right there, ready to be defended. The instinct is to write the defense — state the position, line up the evidence, knock down the obvious objection, land the conclusion. That instinct will get you a competent five-paragraph argument and almost nothing else. It will not get you the thing you actually want, which is an essay that turns the topic over in its hands and looks at every face of it before it says anything final.
This is a guide to writing that other thing. The vault has a deep seam on it — a whole cluster of working writers and a couple of psychologists who, between them, describe the craft, the inner work, and the structure of the essay that examines a subject on all sides. None of them used the phrase "all sides." But every one of them is circling the same discipline: hold the question open longer than is comfortable, take the position you'd least like to be true and make it sound smart, and let the shape of what you write emerge from that honest collision rather than from a verdict you reached before you started.
Let's build the method from the ground up.
What you're actually making
Start with the definition, because if you get this wrong everything downstream reads as arbitrary fussing. Michael Deen opens his whole framework with it: an essay is consciousness frozen in words — an artifact that captures the thoughts and feelings and events that pass through you as you make sense of something you care about. Read that twice. The essay is the record of a thinking-process. Not the polished lecture from a settled position. The thinking, as it happens.
That single reframe is the foundation of the many-sided essay. If the essay is a verdict, then showing the other side is a weakness — you're handing ammunition to the enemy. If the essay is a record of you working a question out, then the other side isn't a threat. It's part of the work. It's one of the things that passed through you while you were trying to figure this out, and leaving it out would falsify the record.
The form was invented for this. Montaigne, in the 1500s, defined the essayist as one who is above creeds — independent, not committed in advance to a side, thinking the question through genuinely instead of performing a conviction he already held. Go back further and you hit Socrates doing the same thing out loud. The essay is, structurally, a simulated Socratic dialogue with you playing both speakers. To collapse it back into a legal brief is to throw away the one thing that made the form worth inventing.
So here's the first move, and it's a posture more than a technique: you are not a lawyer. You are not here to win. You are here to find out. Hold that, and the rest of the method has somewhere to stand.
One more thing this definition does — it tells you what to write about. Deen's standard isn't importance, it's care. A toothpaste conspiracy theory is a valid essay topic. A shower-thought is valid. Your decade-old grievance is valid. What's not valid is a topic you were assigned, or one that's trending, or one you've already fully decided. The many-sided essay needs a live question, and a question is only live if some part of you doesn't yet know the answer.
The one move that makes an essay many-sided
Here is the engine of the whole thing. Deen calls it argument as dialectic, not defense, and if you take only one technique from this whole report, take this one.
Imagine the smartest person who would disagree with you. A specific person, if you can. Now do two things, in this order.
First, write the strawman of your own thesis — the dumbed-down, easy-to-mock version a hostile reader would caricature you into. You're not going to put this in the essay. You're writing it so you can recognize it, so you don't drift into it by accident. Most weak essays are secretly defending the strawman of their own position without knowing it.
Then write the steelman of the other side. Not the weakest version of the opposition that's easy to knock over — the strongest version. The one that genuinely rattles you. Present their argument in the best possible light, the way they'd want it presented, before you lay a finger on it. This is the move. Deen's rule is blunt: if you can't make the opposite position sound smart, you haven't understood it well enough to disagree with it honestly.
And the reframe that makes this bearable: the skeptic is not your enemy. The skeptic is a collaborator. They're there to find the gaps in your thinking, test your assumptions, and show you where you're uncertain — which is exactly what you'd pay a good editor to do. Adversarial framing makes you defensive and you write worse. Collaborative framing makes you curious and you write better. Same exercise — imagining what your opponent would say — completely different result depending on which frame you walk in with.
Deen scores this on a five-point scale, and it's worth internalizing because it tells you exactly where most published opinion falls down:
- Flimsy propaganda — one side, opposition mocked or absent.
- Counterpoints present but not taken seriously — three sentences of "of course some would say" before returning to the defense.
- Balanced but not rigorous — both sides shown, neither explored deeply.
- A well-developed steelman, but the territory is expected — you did the work, but the conclusion was predictable.
- A deep, honest exploration of both thesis and antithesis that produces a novel, unexpected synthesis — a third option that emerges from the collision and couldn't have been guessed in advance.
Notice where the jumps are. Three to four is the jump from gesturing at the other side to fully developing it. Four to five is the jump from a predictable balance to a synthesis nobody saw coming — including you, when you started. Most opinion writing lives at 1 or 2. The many-sided essay is trying to live at 4 and reach for 5.
Deen names the failure case so you can see it: Marc Andreessen's Techno-Optimist Manifesto, which makes bold claims, brands doubters as "enemies," and never engages a single opposing argument honestly. It worked as tribal signal. It failed as essay. The reader who already agreed felt validated; the reader who disagreed felt insulted; nobody learned anything. A better version, Deen sketches, would have explored the real dangers of the thing it was cheerleading, shown where the obvious fix might be worse than the disease, and arrived somewhere genuinely surprising. Three moves, each taking the opposition seriously. The synthesis would have been earned instead of asserted.
And the line that should make you uncomfortable in the right way: if you know counterpoints exist and you decide to omit them, you're either lazy or lying. That's not a stylistic note. It's a charge of intellectual dishonesty. The many-sided essay treats leaving out the strong objection you already know about as a kind of cheating — because it is.
Why this is hard, and what's happening inside you
If the steelman move were easy, everyone would do it and all opinion writing would be good. It isn't, so let's look at why, because the difficulty is psychological before it's technical.
Maslow's work on both/and thinking names the underlying capacity. Most of us think in dichotomies — strong or sensitive, right or wrong, my side or theirs. Dichotomy is comfortable. It reduces anxiety, it makes choosing easy, and it lets you know who you are. But it's also false to how reality actually behaves, and maintaining it costs you. To stay certain that you're on the good side, you have to keep defending against the bad side that lives in you — the doubt, the sympathy for the other position, the part of you that suspects they have a point. That defended part gets disowned, hidden, pushed down.
This is the secret reason most writers can't steelman: they've split off the part of themselves that holds the doubt. The opposition isn't really out there in the world. It's the position you're afraid is true, projected onto an imaginary opponent so you don't have to feel it yourself. Deen's "consider the skeptic" is, underneath, an instruction to go retrieve the doubt you disowned. The steelman is shadow-work run in prose. You're not making peace with an enemy. You're taking back a piece of your own thinking you'd exiled.
Both/and thinking is the capacity to hold two contradictory truths at once without collapsing them into one. A person can be genuinely strong and genuinely sensitive. A policy can be both compassionate and harmful. A belief can be both true and limited. The nervous system has to relax into the contradiction instead of tensing against it — and that relaxation is the thing the many-sided essay runs on. You cannot write an essay that holds three positions if you can only emotionally tolerate one at a time.
But Maslow also names the counterfeit, and you have to watch for it, because it's the failure mode the many-sided essay is most likely to slide into. Pseudo-both/and: the writer who says "both sides have a point" as a way of avoiding having to commit. This looks like sophistication. It's actually cowardice dressed up. Genuine both/and doesn't dodge commitment — it makes a better commitment possible, because the position you land on isn't propped up by denying the opposite. The essay that ends in limp "well, it's complicated, who's to say" has not done the work. It's done the opposite of the work. It held both sides so loosely that neither one ever touched it.
There's a deeper version of this in McGilchrist's master and emissary — two modes of attention, one that divides the world into manipulable parts and one that holds it whole. The many-sided essay needs both: the dividing attention to lay out the distinct positions clearly, and the unifying attention to feel how they belong to one larger picture. Lean too hard on the dividing mode and you get a tidy debate with no soul. Lean too hard on the unifying mode and you get mush. The craft is the round trip — divide to see the parts, unify to see the whole, and write from the second while having done the first.
The four bricks — and which one you keep leaving out
Now, the materials. Deen's material as essay bricks gives you a formula worth memorizing, because it diagnoses exactly how an essay falls out of being an essay:
Without experience, it's an article. Without references, it's a journal. Without counterpoints, it's propaganda.
Four kinds of brick. Experience — your own lived, unrepeatable history, the stuff no one else and no machine has access to. References — the cultural memory you weave in, the sources and allusions and facts. Things that prove you right. Things that prove you wrong. Drop any one and the genre collapses into something lesser. The many-sided essay is precisely the genre that insists on holding all four — and the one most writers drop is the fourth.
That fourth brick is where the steelman lives. And once you've built the opposition honestly, Deen says you have exactly three things you can do with it, which is genuinely useful because it dissolves the false choice between "win" and "give up":
CHALLENGE — find new material that pushes back on the skeptic. You engaged the strong objection and you can answer it. Your thesis survives, sharper than before.
COMPROMISE — find the middle. The skeptic was right about something. You absorb that something and refine your position to account for it. You didn't lose; you upgraded.
CONCEDE — admit the skeptic was right. Maybe laugh at yourself for ever thinking otherwise. And here's the move most writers never make: use the concession to open a new question. Deen's own example — an essay arguing that a precognitive dream was real, met by the skeptic's Law of Large Numbers (you have a hundred thousand dreams in a lifetime; one's bound to seem prophetic by chance). The CONCEDE path gives up the original claim and walks straight into a better one: why do humans have this impulse to retroactively assign meaning to random signals? That's a harder, richer question, and it only became available because the writer let themselves lose the first argument.
Sit with that. Proving yourself wrong is often more interesting than proving yourself right. The CONCEDE move unlocks territory the CHALLENGE move keeps locked. The many-sided essay is willing to lose, on purpose, because losing is where the surprising synthesis hides.
Two gears: opening up and closing down
You now know the engine and the materials. The next thing to understand is that writing the many-sided essay uses two completely different mental gears, and most of your frustration comes from running the wrong one at the wrong time.
Deen, borrowing from the psychologist J.P. Guilford, calls them divergence and convergence. To diverge is to be open — chase tangents, drop your filters, let the stupidest detail connect to everything, generate without judgment. To converge is to be a ruthless editor — cut, narrow, hold the center and prune everything outside its pull. Divergence finds the center. Convergence shapes around it. Both required. Neither one alone makes an essay.
For the many-sided essay specifically, the divergent gear is where the sides come in. This is when you go find the steelman, collect the counterpoints, follow the objection down the rabbit hole, read the person you disagree with charitably enough that they start to convince you a little. You cannot do this with the editor running. If you're judging every opposing idea as you encounter it, you'll cut the ones that would have generated your synthesis. The five-star synthesis is downstream of a genuinely open divergent phase. Pre-prune the opposition and the best version of the essay dies before it's born.
The convergent gear is where you decide what the collision actually produced and cut everything that doesn't serve it. This is where you stop adding sides and start shaping the ones you have into a path the reader can walk.
The fatal error is mode-blindness — switching gears by accident, generating when you should be cutting, cutting when you should be generating, doing both badly because you're doing both at once. The fix is almost insultingly simple: name the mode out loud. Right now I'm diverging; I'll generate opposition without judgment for an hour. Or: Right now I'm converging; I'll cut without adding anything new. And run them in separate blocks of time. The discipline isn't to stop switching. It's to switch on purpose.
One caution that matters for the many-sided essay especially: the cycle repeats. Explore, find a center, cut to it, then re-explore from the new vantage point. Each time you lock a tentative thesis and steelman against it, the steelman may move you, which gives you a new center, which deserves its own fresh divergent pass. You're not doing this once. You're spiraling inward, and each loop makes the thing truer.
Letting the question pick you
Here's a quieter piece that changes what you write about in the first place. Henrik Karlsson's five-step method is built around one observation: his best essays were the ones he'd been carrying for a year before he wrote a word. He'd read around them, argued about them, lived near them long enough that they stopped being ideas and became part of how he saw. The essays he forced out of fresh enthusiasm had the right shape and the wrong weight — thought-balloons he didn't recognize three weeks later.
The thing that ends the year of waiting is what he calls an animating question — a question with real personal stakes. Not "what is the truth about parenting" but "what do I need to understand here so I can show up for my kids." When that question lands, all the accumulated material suddenly lines up toward it like iron filings toward a magnet.
Why this matters for the many-sided essay: the animating question is the thing that forces you to look at all sides. A genuine question, one you don't yet have the answer to, can't be honestly explored from one side, because if one side settled it you wouldn't still be asking. The topics that have moved into you and won't leave are exactly the topics where you feel the pull of more than one answer. The flip side — Deen's mirror requirement — is the warning sign: if all your material points to one clean, spectacular conclusion, something's off. Reality is messier than that. An essay that finds no tension is an essay that wasn't really looking, which usually means it started from a verdict and not a question.
Karlsson names the working state, too, and it's the opposite of what writing advice usually sells. He calls it the confusion zone. Most methods try to reduce your uncertainty fast so you can start producing text. Karlsson's whole toolkit, in his words, is "a series of exercises for breaking my own mental models" — it deliberately sustains confusion long enough for something truer than your first take to surface. For the many-sided essay this is the entire game. The confusion isn't an obstacle you push through to reach clarity. It's the medium you have to stay inside while you hold three positions at once, because the moment you grab a conclusion to escape the discomfort, you've stopped examining and started defending.
This connects to a paradox worth naming: the write-to-discover paradox. You plan a structure so that you're free to discover content. The skeleton isn't a cage — it's what lets you wander without getting lost. You don't outline because you know what you think. You outline so you can find out what you think while writing, and trust that the scaffold will hold while you do.
Staying un-sure on purpose
The enemy of the many-sided essay is premature certainty, and Richard Powers gives you the sharpest tool against it. His maxim: when you're sure of what you're looking at, look harder. His reasoning is that the moment you become certain, you stop receiving. Reality keeps moving; you stay where you were when the certainty arrived; the gap grows; soon you're looking at a memory and calling it perception. Certainty is the signal that your apparatus has stopped updating.
This is an instruction you can run mid-sentence. You write a confident line — "people only really change under pressure" — and it sounds right, it does the work, you move on. Don't. The sureness is the signal. That sentence is what your prior model generated, not what you actually looked at. Look harder. What's specifically wrong with it? Who changes without pressure, and why did you filter them out? The exceptions were suppressed by the confidence. Look harder and they come back — and they're often where the second side of your essay was hiding the whole time.
The contemplative traditions have been training exactly this for a thousand years. Zen calls it beginner's mind — meeting the familiar as if you'd never seen it, refusing to let the apparatus close. Powers's craft discipline and the meditator's discipline are the same operation at different scales. For the essayist the payoff is concrete: every time you catch yourself sure, you've found a place where you stopped examining, which means you've found a side you haven't looked at yet.
Powers leaves a real tension unresolved, and it's worth holding rather than dissolving: some domains require commitment. The surgeon has to make the cut. The judge has to rule. Permanent provisionality would be paralysis. The refinement is that look-harder is a discipline of perception, not of action — you can stay perceptually open ("this might be unusual; let me look again") while still acting decisively when the time comes. For the essay this means: examine all sides during the looking and the drafting, then commit when you write the ending. The openness has a season. It is not forever.
Telling the reader where you're unsure
You've examined every side. Now, how do you show that on the page without the essay reading as wishy-washy? This is where most writers panic and over-correct back into false confidence. Johann Hari's answer: confess your doubt to the reader — uncertainty, deployed honestly, is more persuasive than projected certainty, not less.
The mechanism is precise. We all live inside actual uncertainty — almost nothing that matters in our lives is 100% settled. So when a writer projects perfect confidence about a genuinely hard question, the reader's reasonable response is suspicion: they're hiding something, or they don't grasp the difficulty, or they're selling me something. The over-confident piece reads as not-quite-honest even when every fact in it is correct. The writer who instead levels with the reader — I'm not sure about this; here's how I got here; this part still confuses me — produces a match between their epistemic stance and the reader's lived one. That match is what earns trust. And the relationship shifts from teacher-talking-down-to-student into two minds working a problem side by side, which is the more durable form of influence.
This isn't just craft intuition; it's measured. The influence literature has shown for fifty years that two-sided messages outperform one-sided ones in nearly every condition — because acknowledging the opposition's strongest point inoculates the reader against future counter-messaging. Give a reader the steelman inside your essay and the next anti-you article they meet won't move them; they've already seen its best shot and watched you handle it honestly. The many-sided essay isn't just more truthful. It produces opinion-shifts that last, where the one-sided version produces shifts that erode the first time the reader hears the other side from someone else.
There's a sentence-level instantiation of this — MacCulloch's historian's discipline of graded-confidence language, the careful "may have / probably / almost certainly" that calibrates each claim to what the evidence actually carries. Powers tells you how to keep perceiving freshly; MacCulloch tells you how to report what you saw at its true confidence level; Hari builds the doubt into the structure of the whole piece. Three scales of one discipline: catch yourself when sure, say only what you actually know, and let the reader see your real epistemic state.
One guardrail, because the move can be faked. Performed humility — confessing doubt you don't actually feel, hedging everything to seem balanced — reads as false too, and readers catch it. The tell of the genuine article is specificity. Real uncertainty produces specific doubts about specific claims. Vague hedging ("of course, reasonable people disagree") is the sound of a writer who hasn't done the uncertainty work. The many-sided essay confesses this doubt about this sentence, not a general fog of even-handedness.
Building the thing so the sides hold together
You can examine every side beautifully and still produce an unreadable mess if the architecture can't carry the complexity. So: structure.
The first principle is counterintuitive given everything above. Many sides, one center. Deen's thesis as central organizing idea is the star the planets orbit — and the planets here include all the opposing positions. The thesis of a many-sided essay is not "X is true." It's more like the shape of the journey: the question you're chasing, or the surprising synthesis you reached, or even the honest admission that the question got harder. But there has to be a center, or the sides just scatter. The most common way the many-sided essay fails structurally is the Hydra Problem — scope creep, where holding multiple positions becomes an excuse to hold everything, and the essay sprouts a new head every paragraph until the reader can't find the spine. Examining all sides is not the same as covering all topics. Pick one star. Let the opposing positions orbit it.
The second principle is that tension is the engine that pulls the reader through. Deen's tension as essay engine — a question held open creates a cognitive itch the reader needs scratched, and that itch is what makes them keep reading. The many-sided essay has a structural advantage here: the unresolved collision between positions is the tension. You don't have to manufacture suspense. The reader feels the pull of the steelman, feels your thesis wobble under it, and wants to know where it lands. As long as you don't resolve too early, the dialectic itself drives the prose. This is also why the conflict-as-dramatic-arc frame works on nonfiction — your intellectual conflict between thesis and antithesis runs the same setup-obstacle-resolution shape as a story.
For a piece at 6,000 words, you'll need a way to keep multiple threads alive across the length without losing the reader, and that's the echo structure — plant a motif early, nurture it in the middle, deepen it, harvest it late, with variation each time so it doesn't feel like repetition. In the many-sided essay your "motifs" are often the competing positions themselves: introduce the opposition early as a seed, return to it stronger each time the essay deepens, and let the final pass on it deliver the synthesis. The opposition isn't a single section you visit and leave. It's a thread you weave through the whole length, getting stronger each time it reappears, so that when you finally do something with it the reader has been living with it long enough to feel the weight of the resolution.
And the structural trick that makes all of this writable: write the beginning and the ending last. You cannot write the opening of a many-sided essay first, because you don't yet know where the dialectic lands — and the opening's job is to set up a journey whose destination you haven't reached. Karlsson found his best openings were throwaway lines from the middle of the draft, moved to the top once he understood what the piece was actually about. Draft the messy middle where the sides collide. Then find the door and the exit. The ending of a many-sided essay is the hardest sentence you'll write, because it has to honor everything you held open without pretending it all neatly closed — and you can only write it once you know, by having written your way there, what the collision actually produced.
Where it's allowed to land
The last question is the one writers get most anxious about: after all this examining, what do I actually say? Am I allowed to conclude anything?
Yes — and Deen's failure modes map the ways you can get the ending wrong. False synthesis: you concede a trivial point to look balanced, then hold your original position unchanged. The reader can tell you faked it; the opposition never actually moved you. Performative balance: you mention the other side but never let it shake you. Endless dialectic: all sides, no resolution, the essay dribbles out into pure ambiguity — the pseudo-both/and trap again, sophistication used as an excuse not to commit.
The real ending does one of three honest things, and they map back onto Deen's CHALLENGE / COMPROMISE / CONCEDE. You can come out the other side with your thesis sharpened by the fight (challenge). You can land on a genuine synthesis — a third position that emerged from the collision and that neither side held going in (compromise, in its best form: the five-star outcome). Or you can concede, and let the concession open a better question than the one you started with. What you may not do is predetermine which of these it'll be. Deen's whole point: don't decide the ending in advance. Let the honest engagement produce it. If you knew where you'd land before you steelmanned, you didn't really steelman.
And a real debate produces some movement. Even "I came in thinking X and I leave with more questions than I started with" is a landing — it's honest, it's specific, it names where the inquiry actually arrived. The many-sided essay doesn't require a triumphant resolution. It requires that you went somewhere, and that you tell the reader truthfully where.
The whole method, in order
Pulled into a sequence you can actually run for a 6,000-word piece:
Find the live question. Not a verdict you want to defend — a question some part of you genuinely can't yet answer, ideally one that's been living in you a while. If your material points to one clean conclusion, you haven't found a question, you've found a position. Keep looking.
Diverge wide and gather all four bricks. Open the editor-off gear. Collect your experience, your references, the evidence for you, and — the brick everyone drops — the evidence against. Steelman the side you'd least like to be true. Make the smartest opponent you can imagine sound smart. If you can't, you don't understand the topic yet.
Stay in the confusion. Resist the grab for an early conclusion. When you catch yourself sure, look harder — the sureness marks a side you stopped examining. Hold the positions in genuine both/and tension; don't collapse them, and don't fake-balance them either.
Lock a tentative center, then re-explore. Converge: pick the one star — the question or the emerging synthesis — and cut everything outside its orbit. Then run the loop again from the new vantage. Spiral inward. Each pass makes it truer.
Build for the long form. One center, many orbiting sides. Weave the opposition through as an echoing thread, not a single visit. Let the unresolved collision be your engine of tension. Calibrate every claim to its real confidence; confess specific doubts where they're real.
Write the ending by arriving at it, then write the opening. Don't predetermine the landing. Challenge, compromise, or concede — whichever the honest engagement produced. Then find the door and the exit last, once you know what the piece turned out to be.
Do this and you get the thing the defensive five-paragraph version can never reach: an essay that a reader who disagrees with you can still respect, that inoculates rather than alienates, and that occasionally — when the collision is honest enough — hands you a synthesis you didn't know existed when you sat down. That last surprise is the whole reward. It's also the proof you actually examined all the sides instead of just gesturing at them. You can't be surprised by a conclusion you decided on in advance.
What the vault holds on this
Well-covered — the essay-as-exploration foundation and the dialectical engine. The spine is solid: argument as dialectic, not defense (the steelman/strawman protocol, the five-point rubric, skeptic-as-collaborator), essay as consciousness frozen in words (the Montaigne-rooted definition), and material as essay bricks (the four-types formula and the CHALLENGE/COMPROMISE/CONCEDE moves). These three carry most of the method.
Well-covered — the cognitive and process groundwork. both/and thinking (Maslow on holding contradiction, plus the pseudo-both/and shadow), divergence/convergence as essay modes, the Karlsson five-step method with its animating question and confusion zone, when you're sure, look harder (Powers), and the write-to-discover paradox.
Well-covered — the reader-facing honesty discipline. confess doubt to the reader (Hari), its sentence-level cousin the "may have" discipline (MacCulloch), and the measured backing in two-sided messaging outperforms one-sided.
Well-covered — the structural craft for long-form. thesis as central organizing idea (and the Hydra Problem), tension as essay engine, the echo structure method, conflict as essay dramatic arc, and beginning and ending written last.
Supporting — the deeper roots. McGilchrist's master and emissary (dividing vs. unifying attention), beginner's mind (shoshin), and the wider Deen essay-architecture framework that the individual craft pages hang from.
Gap — no single page synthesizes "the many-sided essay" as a named method end-to-end. This report is the first pass at that synthesis; it could become a concept page (exploratory-essay-as-method) if the territory keeps recurring. Gap — no page yet treats the ethics of false even-handedness (the bothsidesism problem in journalism), which is the dark twin of the steelman discipline and would sharpen the pseudo-both/and warning considerably. Enrichment target.