The gap between what a writer means and what arrives on the page

Most writers believe they have communicated clearly long before they have. They finish a sentence, reread it once, and feel satisfied — not because the sentence works, but because they already know what it was supposed to say. The meaning is so vivid in their own mind that they read the words and see the intention, not the text. Readers, who arrive without that prior knowledge, see only what is actually on the page.

I notice this constantly in the manuscripts we work on at Global English Editing. The writer is not confused. The writer knows exactly what they mean. The problem is structural: the meaning that feels transmitted from the inside is not the same thing as the meaning that arrives on the outside.

That gap — between intent and reception — is the central problem of written communication, and it is far more persistent than most writers expect.

The Shannon-Weaver model and the semantic problem

In 1948, Claude Shannon published a mathematical theory of communication describing information transfer as a process of encoding, transmission through a channel, and decoding by a receiver; a year later, Warren Weaver elaborated the framework in a joint volume. The model identified noise — anything that disrupts the signal between sender and receiver — as the fundamental obstacle to accurate transmission. Weaver extended Shannon’s engineering framework to address what he called the semantic problem: not just whether the signal arrives intact, but how precisely the transmitted symbols convey the desired meaning.

The semantic problem is where writers get into trouble. Shannon’s original concern was technical noise — interference in the channel. But for writers, the noise is almost always semantic. The words chosen encode a meaning that made sense to the person choosing them. The reader, decoding with a different set of associations, prior knowledge, and contextual assumptions, reconstructs something different. Neither party is wrong. The gap is built into the process.

What the model does not fully account for is a subtler issue: writers don’t just transmit inaccurately. They tend to believe, with some confidence, that they have transmitted accurately. The gap exists and the writer doesn’t know it’s there.

The specific mechanism: knowing too much

In his 2014 book The Sense of Style, Steven Pinker identified what he called “the curse of knowledge” as the chief contributor to opaque writing: the writer’s inability to imagine what it is like not to know what the writer knows. Intermediate steps that feel obvious go unwritten. Jargon that the writer has used for years lands on a reader as unexplained terminology. Scenes vivid in the writer’s memory remain invisible to the reader because the writer forgot that the memory is not shared.

As Pinker put it, “it simply doesn’t occur to the writer that readers haven’t learned their jargon, don’t know the intermediate steps that seem too obvious to mention, and can’t visualize a scene currently in the writer’s mind’s eye.” The curse isn’t laziness or carelessness. It’s a structural limitation: once you know something thoroughly, you cannot reliably simulate not knowing it.

This has measurable consequences. In a 2002 study published in Psychological Science, Boaz Keysar and Anne Henly asked speakers to communicate specific intentions using ambiguous sentences, then asked them to estimate how well they’d been understood. Speakers thought they had been understood in 72% of cases. The actual rate was 61%. More strikingly, when listeners failed to understand, speakers believed they had succeeded nearly half the time. The overestimation was not random noise — it was systematic. Writers know what they meant, and that knowledge contaminates their assessment of what they communicated.

What revision cannot fix on its own

The standard advice is to revise. Read your draft again. Cut the unnecessary words. Read it once more. This helps with surface problems — wordiness, redundancy, tangled syntax — but it does not reliably fix the intent-reception gap, because the same mechanism that created the problem also undermines the solution. When you reread your own draft, you are decoding it with full knowledge of what you meant to say. You will find it clear because you are not the reader you need to simulate.

Editing by a third party works differently precisely because the editor arrives without that prior knowledge. When a manuscript crosses my desk, I am reading it as close to cold as the text permits — I don’t know what the writer intended, which means I can only assess what is actually there. That is not a special skill; it is a structural advantage. The editor is in the position of the reader, and the writer, by definition, cannot be.

This is also why writers who revise obsessively sometimes produce prose that is technically clean but still difficult to follow. They have refined their encoding without ever testing the decoding. The prose is polished on its own terms. Whether those terms correspond to what a reader will receive is a separate question, and self-revision cannot fully answer it.

The language problem Orwell was really describing

George Orwell’s 1946 essay “Politics and the English Language” is usually read as an argument about political dishonesty — the way vague language conceals bad intent. But the deeper point in Orwell’s essay is about the relationship between unclear thinking and unclear prose. His diagnosis: “A man may take to drink because he feels himself to be a failure, and then fail all the more completely because he drinks.” The prose gets worse and the thinking gets hazier in a self-reinforcing cycle.

What Orwell is describing, beneath the political surface of the essay, is a failure of encoding. The writer has not actually resolved what they mean; they have chosen words that create the impression of meaning — the impression, specifically, that satisfies the writer’s own internal sense of having said something. The semantic gap in this case is not between a clear intent and a misread text. It is between a vague intent and prose that obscures the vagueness even from the person writing it.

This is the worst version of the gap: not knowing what you mean, encoding that uncertainty into words that feel precise, and then reading back your own uncertainty as if it were clarity.

What actually closes the gap

The gap narrows when writers become genuinely curious about what their readers will encounter — not what the writer intended, but what the words, in sequence, without context, will produce in a mind that doesn’t share the writer’s knowledge. That is a different cognitive task than writing, and most writers are not trained for it.

In practice, it requires a version of what Pinker calls thinking back to before you knew what you now know — resetting your prior assumptions about what the reader brings to the text. For academic writers, this means asking whether a specialist in an adjacent field would follow the reasoning. For business writers, it means checking whether the logical steps that feel obvious are actually present on the page. For exnumerators, it means identifying where the argument assumes consent that hasn’t been earned.

The specific check I return to when editing: does the text work without the writer in the room? If the meaning depends on knowing what the writer was trying to do, the text has not yet carried the weight it needs to carry. A sentence that requires the author’s presence to be understood is a sentence that isn’t finished.

Why the gap persists

Writers are invested in what they’ve written, and that investment is not only emotional. It is cognitive. The words on the page are attached to a body of knowledge, intention, and mental imagery that the writer has built up through the process of writing. Rereading the draft activates that context automatically. What the writer experiences as reading is, in part, remembering.

This is why the gap is not a beginner’s problem. It persists across experience levels. Skilled writers produce it, academic researchers produce it, people who write professionally every day produce it. What changes with experience is not the tendency to overestimate — the Keysar and Henly data suggests that’s fairly stable — but the writer’s awareness that the tendency exists, and their discipline in working against it.

The manuscripts that need the least editorial work are rarely the ones written fastest or with the greatest confidence. They are the ones where the writer has already started asking what the reader will actually find — where the writer has, at some point in the drafting, stood outside their own knowledge and looked back at the text.

That is a skill that can be developed. It is not the same skill as writing.