Good prose is not a display of talent but a set of learnable disciplines — dare to write, cut hard, keep the rules, serve the reader. Each of these four books guards one of those gates.
2026 · Book Recommendations · Issue 4
Zinsser narrows the core act of good writing to one verb — cut: strip out every word that does no work. Lamott deals with the fear that comes before the first sentence, arguing you must allow yourself a "shitty first draft" before you can judge it. Strunk & White compress writing into a few dozen rules short enough to memorize and internalize, thin enough to fit in a pocket. Pinker uses cognitive science to explain why the rules work — and exposes how "the curse of knowledge" makes experts the hardest to read.
| Book | Author | Year | What it nails |
|---|---|---|---|
| On Writing Well | William Zinsser | 1976 | The core act of good writing is subtraction — clear out every word that carries no function; concision itself is a form of honesty toward the reader |
| Bird by Bird | Anne Lamott | 1994 | What stops people isn't lack of skill but fear — permit yourself a lousy first draft; "done" always comes before "perfect" |
| The Elements of Style | Strunk & White | 1959 | Compresses good English into a few dozen rules you can memorize and internalize; the core is one line — omit needless words |
| The Sense of Style | Steven Pinker | 2014 | Remakes the writing guide with cognitive science — behind good and bad prose is how the reader's brain processes words; the chief killer is "the curse of knowledge" |
Zinsser diagnoses the disease of American writing in one word: clutter — we are tangled in words that do no work, circular constructions, self-inflating jargon. His central claim is almost a moral law: every word that carries no information or tone is stealing the reader's attention. Concision, then, is not a stylistic preference but a form of respect.
The mechanism comes down to an action: strip every sentence to its cleanest components. Cut the adverb that repeats the verb ("shouted loudly"), the adjective already inside the noun, the polite throat-clearing ("It is interesting to note that…"), the pompous long word ("utilize" → "use"). Each layer you remove makes the meaning brighter, not poorer. Zinsser shows his own manuscripts slashed to ribbons — good prose is not written, it is rewritten.
So he keeps "writing" and "revising" apart: the first draft just gets the thoughts onto the page; rewriting is the real writing — where you decide what stays and what dies. Almost no sentence comes out right the first time. That is not a failure of skill; it is the normal condition of writing.
But concision is not the same as dryness. Zinsser's other pillar is voice and humanity: what survives the cutting should be your own tone, so the reader feels a real person behind the words. Technique serves that, not the other way round. The reader has only one kind of attention, and it can leave at any moment — every vague or redundant sentence hands him a reason to.
The focus is almost entirely on nonfiction, and on an American-English register; poetry, rigorous academic argument, and brand copy don't map cleanly. Some judgments of "pompous" words carry Zinsser's personal taste and the stamp of 1976 — a few examples already feel dated in today's online register.
Zinsser's "cut" belongs first on BigCat's own output — reading lists, technical docs, weekly reports, even prompts to an AI. One hard move to try next week: take something you just finished and force yourself to delete a quarter of the words without losing any information. Hunt three kinds of clutter: adverbs that repeat the verb, polite throat-clearing ("it's worth noting that"), and long words you can swap for short ones. You'll find the meaning isn't diminished but sharper. Then extend it: prompts to an AI follow the same law — clutter dilutes not only the reader's attention but the model's, and the cleaner the instruction, the less it drifts.
The title comes from a family story. Lamott's ten-year-old brother had a report on birds due, had put it off until the last night, and sat paralyzed and close to tears. Their father, a naturalist, sat down beside him and said: "Bird by bird, buddy. Just take it bird by bird." That is the whole method: facing a task that paralyzes you, the only way out is to shrink your view to the next small piece you can actually finish.
Her most famous prescription is the "shitty first draft": all good writers write them — it is not failure but a necessary stage. Let the first draft be bad and long and self-indulgent, because only once it's on the page do you have material to revise. Paralysis usually comes from wanting the first sentence to be perfect — and the perfect first sentence only shows up around the third draft.
From this she names the real enemy: perfectionism. She calls it "the voice of the oppressor" — it keeps you cramped, tense, afraid to set anything down, because you fear making a mess, fear being seen. You don't beat it with more willpower but by lowering the standard, shrinking the unit, allowing the mess.
She also writes honestly about envy, self-doubt, the mid-draft panic — the occupational hazards most writing books avoid. Her value is in returning writing to a human condition rather than a feat of genius: you don't have to get good before you start; you start, write badly, revise, and slowly get good. She closes with Doctorow's image: writing is like driving at night — your headlights light only a short stretch, but you can make the whole trip that way. You needn't see the whole road to set out.
It is aimed mainly at fiction and memoir; technical docs, manuals, and rigorous argument don't necessarily suit her "emotional truth first" stance. The book carries heavy religious and personal-healing tones some readers will find far from the craft itself, and it barely addresses sentence-level technique.
The title story is itself a child crushed by a big assignment — which maps Lamott naturally onto parenting a school-age kid. When a child freezes before "read the whole book and write a report" or "build a science display," the parental instinct is to nag or reason, but what actually works is "bird by bird": sit with her and cut the task down to one piece she can finish tonight (just write the main character's name; just draw one picture), then cut the next. One thing to try next week: translate "perfectionism is the voice of the oppressor" into words for a child — tell her plainly that "the first pass being messy is correct; we were always going to fix it on the second," and make "allowed to write badly" a house rule. What this removes isn't the standard — it's the fear that keeps her from starting.
This is the thinnest, oldest, most code-of-law-like of the four. It began as Strunk's self-printed handout for students in 1918, expanded in 1959 by his student E. B. White, author of Charlotte's Web. It doesn't explain or argue, it just hands you rules: usage, composition, form, common misuses, plus White's few reminders on style.
The soul of the book is one command: omit needless words. White recalls that Strunk would say it three times in class, like a sermon. The famous analogy carries it: vigorous writing is concise — a sentence should have no needless words, a paragraph no needless sentences, just as a drawing should have no needless lines and a machine no needless parts.
Its mechanism isn't "understanding" but "internalizing." The rules are short enough to memorize, to turn into muscle memory: use the active voice, put statements in positive form, keep related words together, write with nouns and verbs (rather than piling on adjectives and adverbs). You don't think them through each time — they sink into the hand and your pen automatically dodges the common bad habits. That is why it is thin: it doesn't want to be studied, it wants to be remembered.
White's closing chapter loosens the rules' grip: he admits style can never be exhausted by rules — the rules are a floor, not a ceiling; master the basics enough to stop making errors, and only then can you speak of personal style. That layer keeps the book from being a cold manual of dogma. It has also become a century-long target of criticism — but the controversy itself proves it set the baseline of modern English prose.
The rules are for English; many (active voice, word order) can only be borrowed in spirit for Chinese, not copied. Some of its prohibitions have been labeled baseless "superstitions" by linguists like Pinker; treated as inviolable dogma they tie your hands — it works best as a floor, not a cage.
What Strunk & White really demonstrate isn't any one rule but the form — "few, firm, internalizable" — which suits engineering teams especially, since programmers already live inside style guides and linters. To try next week: write a tiny prose standard for the team's docs, PR descriptions, and incident write-ups, hard-capped at one page, no more than ten rules, each an executable imperative ("conclusion before process," "one sentence on what this PR changes," "delete filler like 'basically' and 'I think'"). The point isn't completeness but being short enough that everyone remembers it, sinking into muscle memory like a lint rule. A long standard goes unread — which is exactly why Strunk kept the book to a hundred pages.
Pinker is a cognitive scientist and linguist, which makes his angle on writing one of a kind: not taste and aphorism but how the reader's brain processes words. He sets out to explain the cognitive mechanism behind advice earlier writers gave on instinct — and to expose which circulating "rules" are in fact baseless superstition.
The ideal he champions is "classic style." Its guiding metaphor is seeing: the writer has noticed something in the world the reader has not, and orients the reader's gaze so the reader sees it for himself. That stance decides everything — it assumes writer and reader are equals and the world is concrete and visible, so the prose runs concrete, confident, unhedged, not forever excusing itself ("in a sense," "one could say").
The book's sharpest cut is "the curse of knowledge": once you know something, it becomes very hard to imagine what it is like for someone who doesn't. This is the number-one cause of bad prose — not that the writer can't write, but that he forgets the reader isn't inside his head, so he leaves jargon, acronyms, skipped steps, and unstated background. It explains why experts so often write the most impenetrably.
The remedy is concrete: when you're done, hand it to a real target reader and watch where they stall — because you can never diagnose your own curse. And turn abstractions back into visible people and actions (who did what to whom) instead of piling up nominalized terms. Finally he uses linguistics to clear the field: separating the rules that genuinely help (they help the reader parse a sentence) from the handed-down taboos (don't end on a preposition, don't split an infinitive) — most of which collapse against how great writers actually wrote. The only judge of a rule is "does it help the reader understand," not "what the teacher once said."
The "rule superstition" examples are almost all English-specific — debates over preposition stranding and split infinitives have no Chinese counterpart, so that value is discounted. The book is on the thick side, and chapter four's syntactic-tree analysis is technical; anyone wanting quick maxims will find it slow. Classic style isn't universal either — law and safety warnings need exactly the exhaustiveness and hedging it forgoes, not a confident point of the finger.
Pinker's "curse of knowledge" treats the most hidden ailment of a senior technical person — the deeper you know it, the worse you gauge what others don't. Explaining AI, distributed systems, or complex systems to investors, non-technical colleagues, or a child, BigCat is most likely to trip here: a passage that feels perfectly clear loses the listener by the second sentence. Try Pinker's hard remedy next week: after writing a piece for an outside audience, have a real target reader read it aloud and just watch where they frown, reread, or ask "what's this?" — that stall is where your curse lives, and you can never diagnose it yourself. Paired move: convert abstract nouns back into concrete "who did what to whom" sentences. This one habit beats ten writing tips.
Ask of each phrase, "if this word weren't here, would the meaning change?" — adverbs, adjectives, and polite throat-clearing are usually first out. If cutting 25% loses no meaning and reads clearer, you were diluting the reader's attention with clutter (the shared ailment of Zinsser and Strunk). If you truly can't cut anything, congratulations — your prose is close to "every word doing work."
Telling these apart is Lamott's core. "Can't write" is fixed by learning technique; "afraid it isn't good" is fixed by lowering the draft standard — the prescriptions are opposite, and the wrong medicine makes it worse. If you find yourself revising the first sentence over and over, never reaching the second paragraph, it's usually the latter: perfectionism disguised as a high standard. The cure is to permit yourself a deliberately bad draft — finish first, improve later.
This is the only reliable test of "the curse of knowledge" — you can't find your own assumptions by introspection. A valid test must meet three conditions: (1) the reader is a real target, not a peer; (2) you only observe, never explain (the moment you add a word, you've contaminated the experiment); (3) you record the specific spot where they stalled, not whether they "liked it." Never done it = all your "clarity" may be clear only to yourself.