Most sentences that fail do not fail because the writer chose the wrong word. They fail because the writer lost track of the structure. The subject drifts away from the verb. A qualification lands where the claim should be. The sentence ends somewhere other than where it started, and the reader, arriving at the full stop, cannot quite reconstruct the thought.
I spent more than thirty years teaching mathematics in Australian secondary schools, and a good part of that time was spent watching students do exactly this — not with sentences, but with equations and proofs. The variable on the left side of an expression would quietly change its role by the third line. A conditional statement that began “if x is even” would conclude by discussing something else entirely. The logic, which had been sound at the start, dissolved somewhere in the working.
What I did not recognise then, and what became clear only after I moved into editing, is that the failure mode is the same. The mathematics student and the writer are making the same mistake. Both have a structure in their head that they believe is on the page — and neither has checked whether it actually is.
What mathematics teaches about conditionals
The conditional statement is the basic unit of mathematical reasoning. If P, then Q. Given this, that follows. The form commits you to something: once you assert the premise, you are bound by what it entails. You cannot quietly alter P mid-argument and then claim Q still follows. The relationship between premise and conclusion is not decorative — it is the whole point.
Sentences work the same way, though most writers have never been asked to think of them in those terms. Every sentence makes at least an implicit claim about a relationship. The subject and the main verb establish a premise of sorts: this thing does this action, or this thing is in this condition. Everything that follows — every subordinate clause, every qualification — either strengthens that relationship or muddles it.
Bertrand Russell, who was as much a logician as a prose writer, described his ideal for sentences in terms that would have been familiar to any mathematician: “I wished to say everything in the smallest number of words in which it could be said clearly,” he wrote in his 1956 essay “How I Write.” He also offered what sounds like a rule from a geometry proof: “Do not let the beginning of your sentence lead the reader to an expectation which is contradicted by the end.” That is not an aesthetic preference. It is a logical requirement. A sentence that contradicts its own opening is invalid, in the same sense that a proof step that contradicts its own premises is invalid.
The proof structure underneath every paragraph
In mathematics, a proof does not merely arrive at the right answer — it shows the path. Each step must be justified by what came before. You cannot skip lines and trust that the reader will fill them in, because the reader is not supposed to fill them in; the writer is supposed to have done that work already.
Paragraphs have the same structure, though we rarely name it that way. A good paragraph advances a single claim through a sequence of steps, each one earning its place by extending what preceded it. When a paragraph loses its coherence — when sentences start floating free of one another — the reader experiences something close to what a student experiences when a proof loses its thread. There is a local sensation of meaning in each sentence, but no sense of going anywhere.
Joseph Williams, in Style: Lessons in Clarity and Grace, put this in terms of topic and stress: the beginning of each sentence should link to something familiar from the preceding sentence, and the end of the sentence should introduce what is new. The principle, as Williams frames it, is that syntax should mirror semantics — the form of the sentence should carry the shape of the thought. That is a statement about structure, not style. It is the prose equivalent of the rule that each step in a proof must follow from a prior one.
Where the analogy breaks, and why that matters
There is a real difference between mathematical language and prose, and a good editor has to hold both ends of it. In mathematics, the goal is univocity — one meaning, precisely bounded. Ambiguity is a defect, always. In prose, ambiguity can be a resource. A sentence that means two things at once, if the writer knows what they are doing, can do more work than two sentences that each mean one thing.
The research on how the brain processes formal mathematical language and natural language bears this out. A 2009 study published in PLOS ONE by Roland Friedrich and Angela Friederici at the Max Planck Institute found that while mathematical syntax and natural language syntax both involve hierarchical, rule-governed processing, the neural networks recruited are substantially different. Mathematical syntax draws primarily on intraparietal and prefrontal regions, involving Broca’s area only in a limited way — unlike natural language processing, where Broca’s area is more broadly engaged. The brain, in other words, knows the difference — even if the underlying logical structure looks similar on paper.
What this suggests to me, both as a former teacher and as someone who now reads manuscripts all day, is that the skills are related but not identical. Mathematical training sharpens a particular kind of attention — to structure, to sequence, to the consistency of terms — that translates directly into editing. But prose requires something mathematics does not: a sensitivity to how meaning resonates, how a word carries weight beyond its definition, how rhythm affects the way a reader receives an argument. You cannot proof-check your way to good writing. You can, however, proof-check your way to clear writing, which is a start.
Terms that drift
One of the most common problems I see in academic and professional manuscripts is what I think of as term drift. A writer introduces a concept at the beginning of a paper, defines it carefully, and then, three sections later, uses the same word to mean something slightly different. They have not noticed the shift because the word looks identical on the page.
In mathematics, this is called equivocation, and it is the kind of error that makes a proof unsalvageable. If you use x to mean one thing in your first equation and something else in your fifth, no amount of correct algebra in between will save you. The conclusion does not follow from the premises, because the premises have been quietly altered.
In prose, the same error tends to be invisible until an editor catches it — and often not even then, because editors trained primarily in language do not always think in terms of logical consistency across a document. My background in mathematics made this particular problem visible to me in a way that it might not otherwise have been. A word, like a variable, has to carry the same value throughout the argument. When it does not, the argument breaks.
What precision actually means in writing
When I was teaching, I used to tell students that precision in mathematics was not about being fussy — it was about being fair. A precise statement is one that does not ask the reader to make assumptions on your behalf. It puts the terms on the table and shows the working. The reader can disagree, but they cannot be confused about what you mean.
Precision in writing means exactly the same thing. It is not a matter of using technical vocabulary or avoiding contractions or writing long sentences. It is about not hiding behind vagueness. Russell’s advice — never use a long word if a short word will do — is at its core a precision principle, not an aesthetic one. A long word that is doing the job of a short one is a way of importing more meaning than you have actually established. It asks the reader to fill in the gap between the word’s connotations and your actual claim.
That gap is where writing fails. The writer knows what they mean. The sentence gestures in the general direction. The reader arrives with a slightly different set of associations and constructs a meaning adjacent to, but not identical with, the intended one. This is not a problem of vocabulary. It is a problem of logical precision — of the sentence failing to close the gap between what the writer knows and what the words can carry.
What editing is, seen from this angle
After more than a decade of reading other people’s manuscripts, I have come to think that most editing is not about taste. It is about identifying where the structure has broken down — where the argument is no longer following from itself, where terms have drifted, where a sentence has set up an expectation it does not honour.
The mathematical habit of mind is useful for this, not because writing is mathematics, but because it instills a particular suspicion of apparent sense. In mathematics, a result can look right while being wrong. You learn, over years of working proofs, not to trust the appearance of correctness — to follow the logic rather than the impression. Good editing works the same way. A sentence that reads fluently can still be incoherent. A paragraph that seems to be making a point can be circling one without ever landing. The question is not how it sounds but whether it holds.
Russell put it plainly: he was allowed to write in plain English, he said, because everyone knew he could write in mathematical logic if he chose. The point was not that mathematical logic was better — it was that the discipline of it had shaped what he demanded of language. When you have learned to say things exactly, you develop an impatience with saying them approximately. That impatience, applied to prose with some flexibility and care, is what distinguishes a rigorous writer from a merely fluent one.
