reader-expectation-approach

Reader-expectation approach (Gopen & Swan)

Last updated: 2026-05-17

Synthesis

George Gopen and Judith Swan’s 1990 article “The Science of Scientific Writing” (American Scientist 78(6): 550–558) [1] is the most methodologically careful piece of writing advice in the scientific writing literature, and the one most directly grounded in linguistics and reader psychology. Its central insight, repeated in Gopen’s later books [2, 3], is that prose meaning is fixed by what the reader does with the text, not by what the writer intends — and that readers operate by predictable structural expectations about where in a sentence different kinds of information will appear. Writers who satisfy those expectations are read fluently; writers who violate them produce work that is technically grammatical but locally hard to parse.

Three of Gopen and Swan’s prescriptions — the topic position, the stress position, and subject–verb proximity — are unusually operationalizable. They map onto sentence-level transformations that an editing tool can perform without semantic risk. This is why the reader-expectation approach is the framework most directly relevant to scriptorium’s argumentative-flow and any future sentence-level editing skills: it tells you which structural moves to test for without requiring you to understand the content.

The framework’s status as “evidence-based” should be qualified. Gopen and Swan ground their prescriptions in reading-process research (Bock, Just & Carpenter, etc.) and demonstrate them on real scientific paragraphs, but they do not present a randomized intervention. The strongest empirical claim is that violations of these expectations measurably slow reader comprehension at the local scale — not that following them improves citation outcomes. The framework’s authority comes from its mechanism, not from outcomes data.

Evidence and frameworks

The core claim

The article’s foundational sentence is quoted endlessly because it inverts the naïve model of writing:

“The fundamental purpose of scientific discourse is not the mere presentation of information and thought, but rather its actual communication. It does not matter how pleased an author might be to have converted all the right data into sentences and paragraphs; it matters only whether a large majority of the reading audience accurately perceives what the author had in mind.” [1]

And later:

“The meaning of any prose is not that which the writer intends, but that which readers interpret.” [1, paraphrase widely cited]

This reframes editing from a correctness problem (is the grammar right?) to a prediction problem (will the typical reader land where I want them to?). The reader’s expectations are not optional — they are the substrate on which any communication happens.

The structural principles

Gopen and Swan identify several reader expectations. The four most useful operationally are:

1. Topic position. Readers expect the beginning of a sentence (or clause) to tell them whose story this sentence is about and to link it to what came before. A sentence that opens with a new piece of information disorients the reader, because they expected to find a connection to already-known material there.

2. Stress position. Readers expect the end of a sentence (the position just before the period, or before a colon or semicolon) to carry the new, emphasized information.

“Readers naturally emphasize the material that arrives at a point of syntactic closure.” [1]

A sentence whose intended punchline lives mid-sentence rather than at the stress position will be read with the wrong emphasis even if the wording is otherwise perfect.

3. Subject–verb proximity. Readers expect the grammatical subject and its verb to be near each other.

“Readers expect a grammatical subject to be followed immediately by the verb. Anything of length that intervenes between subject and verb is read as an interruption, and therefore as something of lesser importance.” [1]

Long noun phrases or qualifying clauses inserted between subject and verb are processed as parenthetical even when the writer intended them as essential. This is one of the most common failure modes in scientific prose, where a long methodological qualifier is stuffed between subject and verb.

4. Old-information-before-new. Across sentence boundaries, readers expect the topic position of sentence N+1 to anchor in something the topic position or stress position of sentence N established. This is how paragraphs flow: each topic refers back, each stress position pushes forward.

These principles compose: a well-formed paragraph chains topic-old → stress-new → topic-old (the just-introduced new) → stress-new, and so on. The flow is mechanical and the diagnostic is mechanical: underline the topic and stress positions of each sentence, then check whether the topics chain back and the stresses chain forward.

Follow-on work

Gopen developed the framework at book length in:

• The Sense of Structure: Writing from the Reader’s Perspective (Pearson Longman, 2004; ISBN-13 978-0-205-29632-3) [2] — applies the framework beyond science writing.
• Expectations: Teaching Writing from the Reader’s Perspective (Pearson, 2004; ISBN-13 978-0-205-29617-0) [3] — pedagogical companion volume.

Schimel (Writing Science, 2012) imports the topic-position / stress-position distinction wholesale and treats it as the sentence-level operating system for the OCAR-structured paragraph (see narrative-frameworks).

The reader-expectation approach has been quietly adopted by Duke’s Thompson Writing Program, the AAAS communication workshops, and a growing list of graduate research-skills curricula, but it is not formally evaluated in controlled studies. Its endurance is striking: 36 years on, it remains the default citation when someone needs to justify a sentence-level revision in scientific prose.

How this informs scriptorium

The reader-expectation framework is the strongest theoretical grounding for several skills:

• argumentative-flow (current). The skill’s paragraph-level diagnostic should explicitly check old-information chaining across sentence boundaries. If consecutive sentences fail to anchor their topic positions in prior stress positions, the paragraph has a flow defect even if every sentence is locally fine.
• compression (planned, v0.3). Aggressive compression that strips long subordinate clauses must respect subject–verb proximity. A compression skill that fixes a 30-word subject phrase is doing more than reducing length; it’s making the sentence readable. This should be named in the skill’s rationale, not done silently.
• reviewer-simulation (current). The “I had to re-read this paragraph three times” reviewer reaction is almost always a topic/stress violation. The simulator should be able to identify where in the paragraph the reader was thrown off, not just that they were.
• Sentence-level rewrite skills (planned). Any future sentence-rewrite skill should expose its rationale in Gopen/Swan terms — “moved thermal stability into stress position because it carries the new information” — so that the edit is inspectable and the user can disagree with the diagnosis, not just the wording.

The framework also disciplines the posture of editing skills. Because the diagnosis is mechanical (positions, distances), an LLM editor can apply it with much higher confidence than it can apply, say, “make this more elegant.” Skills should prefer mechanical Gopen/Swan moves to subjective ones whenever the mechanical move is sufficient.

Open questions / weak evidence

• Gopen and Swan’s principles are demonstrated on examples; they have not been validated by eye-tracking or reading-time experiments at scale. The cognitive-science substrate they cite (Bock, Just & Carpenter) is real but predates the framework.
• The framework is English-centric. Topic and stress positions in languages with different default word orders (German, Japanese) do not map cleanly. An LLM editor needs to be careful about applying these rules to non-English manuscripts or to translation-adjacent prose.
• It is unclear whether following the principles improves outcome measures (citation, reviewer recommendation) or only proximal comprehension measures (reading time, comprehension accuracy).

References

1. Gopen GD, Swan JA. The Science of Scientific Writing. American Scientist. 1990;78(6):550–558. (November–December 1990 issue.)
2. Gopen GD. The Sense of Structure: Writing from the Reader’s Perspective. Pearson Longman; 2004. ISBN-13 978-0-205-29632-3.
3. Gopen GD. Expectations: Teaching Writing from the Reader’s Perspective. Pearson; 2004. ISBN-13 978-0-205-29617-0.