Bilingual Thinking Produces a Sharper Third Sentence

I write every difficult sentence three times: once in the shape Quechua would give it, once in the shape Spanish would give it, and only then in English, and the English sentence that survives this is almost always sharper than the one I would have written going straight to English.

This sounds like a slower way to work, and it is. It also produces a specific, repeatable improvement in the sentence, not a vague sense of enrichment from knowing several languages, which is the claim usually made about bilingualism and writing and is too soft to test. I can point to what changes and why, on one sentence, and the mechanism generalises.

The third sentence

Take a line from a testimony I was writing about a woman's decision to leave her village after a mining company bought the land around it. In Quechua grammar, the sentence would foreground the land itself and mark, through a suffix, that she had only heard about the sale rather than witnessed it directly, a distinction Quechua requires you to specify and English does not. In Spanish, the sentence would foreground the company as the grammatical subject doing the selling, since Spanish syntax tends to put the actor first. Writing both versions before writing the English forces two different questions onto the table: what was the actual subject of this sentence, the land or the buyer, and did the speaker know this firsthand or secondhand.

My first English draft, written straight, had read: "The company bought the land and she left." Flat, competent, unremarkable. After writing the Quechua shaped version, which foregrounded the land and marked the secondhand nature of her knowledge, and the Spanish shaped version, which foregrounded the company as actor, I arrived at: "She had heard, not seen, that the land was sold, and by the time she confirmed it, she was already the one deciding whether to stay." That sentence is longer, and it earns the length, because it carries information the flat version dropped entirely: that she did not witness the sale, that confirmation came late, and that the decision to leave was hers even though the sale was not.

Why the detour works

The mechanism is not mystical. Each language's grammar forces you to answer a question the others let you skip. Quechua forces evidentiality: how do you know what you claim to know. Spanish forces a clear grammatical actor performing the verb. English, comparatively permissive, lets a sentence be vague about both and still sound fluent. Writing the sentence in the two more demanding grammars first means the vagueness English would have tolerated gets caught and answered before the English sentence is ever composed, which is why the third sentence carries more than either of the first two alone.

I timed myself on ten sentences from one manuscript: writing straight to English took, on average, four minutes each. Writing the three pass version took eleven minutes each. Nearly three times slower. But a blind read by an editor who did not know which method produced which sentence rated the three pass sentences as more information dense on eight of the ten, and equally dense on the other two. None went the other way.

The obvious pushback

Critics of this method, and I have heard the argument directly from a monolingual colleague, point out that a skilled English writer can ask the same questions, evidentiality and clear agency, without detouring through two other grammars, simply by revising harder. I do not disagree that it is possible in principle. What I have found in twenty years of teaching testimonio workshops is that most writers do not ask those questions on their own, because English grammar does not force the asking, and skill alone is not a reliable substitute for a structure that makes the omission impossible to miss. The detour is not required by talent. It is required by the fact that a permissive grammar will happily let a real gap in a sentence's information pass as fluent prose, and only a stricter grammar sitting next to it will show you the gap at all.

Where the method stops paying for itself

I do not run all three passes on every sentence in a manuscript, and I want to be honest about where the method earns its cost and where it does not. A sentence describing the weather, or a transition between two scenes that carries no evidentiary weight, gains almost nothing from being written in Quechua and Spanish first, because there was no hidden question of agency or evidentiality for those grammars to surface. I reserve the three pass method for sentences that carry a claim about who did something, who knew what, and how they came to know it, which in testimonio work is most sentences but rarely all of them. Learning to tell the difference in advance, rather than running the full method on every line, is what makes eleven minutes a sentence sustainable across a two hundred page manuscript instead of a method that only works on short pieces.