Topic position and stress position

Read this paragraph. Every sentence in it is grammatical, none is longer than fifteen words, and there are no buried actions or long waits between subject and verb. It has passed every test this course has given you so far.

The council commissioned a structural survey in January. Corrosion in the main support beams was found by the survey. About two million pounds is the cost of replacing those beams. Four hundred thousand pounds remains in this year's capital budget. Closure until at least next spring is the consequence.

Read it and you can extract the facts. You can also feel that it is hard work, and that the effort is out of proportion to the content, which is a bridge, a survey, some rust and a shortage of money.

Now the same paragraph again.

The council commissioned a structural survey in January. The survey found corrosion in the main support beams. Replacing those beams would cost about two million pounds. That cost is far more than the four hundred thousand pounds left in this year's capital budget. So the bridge stays closed until at least next spring.

Not one fact has been added or removed. The two versions are 49 and 53 words, and the second is barely longer: two of its sentences lost two words each and one gained seven. What did change is the order, and a handful of words that are links rather than facts.

That is the whole idea, and it has a name.

The two positions

Gopen and Swan's account of this in "The Science of Scientific Writing" is the clearest statement in the literature, and the two claims are worth having exactly.

On the beginning of a sentence:

"The information that begins a sentence establishes for the reader a perspective for viewing the sentence as a unit: Readers expect a unit of discourse to be a story about whoever shows up first."

They call this the topic position. It is not where you put the important thing. It is where you tell the reader what this sentence is about, and by extension how it connects to the one before.

On the end:

"It is a linguistic commonplace that readers naturally emphasize the material that arrives at the end of a sentence."

This is the stress position. It is where the reader places their weight, and where they will still be when the sentence stops.

And then the principle that joins them, which they call the given and new contract:

"Put in the topic position the old information that links backward; put in the stress position the new information you want the reader to emphasize."

Old at the front. New at the end. Every sentence a bridge from something the reader has to something they don't.

Their own caveat sits directly under the numbered list those principles come from, and it applies here as much as anywhere: "None of these reader-expectation principles should be considered 'rules.'" What follows is a description of what readers do with order, not an instruction you owe obedience to. The section on conflicts below is where that matters most.

The diagram below draws the contract across three sentences.

Why the choppy version was choppy

Go back to the first paragraph and you can now say exactly what was wrong with it.

Sentence two ended on "by the survey". The survey was old information, and it was sitting in the stress position, where the reader was placing weight on it. Meanwhile "corrosion", which was the news, was at the front, where the reader was looking for orientation and instead got a fresh object.

Every sentence after the first had its information exactly backwards. Not one of them was ungrammatical, unclear, or too long. They were each pointing the reader's attention at the wrong end.

That's why "make your sentences clearer" is useless advice for a choppy paragraph. Each sentence was already clear. Cohesion isn't a property of sentences. It's a property of the joins between them, and the joins live at the two ends.

The wrinkle: this is why the passive is sometimes right

You've been carrying an argument since lesson 3, and this is where it settles.

In 1918, Strunk gave this pair in The Elements of Style, inside the rule headed "Use the active voice":

The dramatists of the Restoration are little esteemed to-day.

Modern readers have little esteem for the dramatists of the Restoration.

The first is passive. The second is active. And Strunk doesn't tell you to use the second. He says:

"The first would be the right form in a paragraph on the dramatists of the Restoration; the second, in a paragraph on the tastes of modern readers. The need of making a particular word the subject of the sentence will often, as in these examples, determine which voice is to be used."

Read that last sentence again. The need to make a particular word the subject determines the voice. That's the topic position, described in 1918, seventy-two years before Gopen and Swan named it.

Give Strunk the credit here, because this course has been hard on rule 10 and it's easy to leave the wrong impression. He isn't caught out by this example. One paragraph earlier, in the same rule, he writes:

"This rule does not, of course, mean that the writer should entirely discard the passive voice, which is frequently convenient and sometimes necessary."

So the man whose rule is headed "Use the active voice" says in the rule itself that the passive is frequently convenient and sometimes necessary, and then shows you a case where it's the right choice. What's misleading isn't Strunk. It's the heading, read on its own by people who never got to the paragraph underneath, and a century of advice that repeated the heading and dropped the rest.

So the honest position on the passive isn't "avoid it", and it isn't a replacement rule either. It's this: the passive is what lets you put something other than the actor in the topic position. If your paragraph is about the dramatists, the dramatists go at the front, and the passive is how you get them there.

Be fair to the other side, which has a real case. Strunk's own reason for the rule stands: "The active voice is usually more direct and vigorous than the passive." GOV.UK tells government writers to use the active voice because it "is more direct and puts the focus on the user", and for most sentences most of the time that's right, which is why it survives as a default. The claim in this section is narrower than "prefer the passive". It's that the choice is decided by what the paragraph is about, and that a writer who can't reach for the passive has lost the ability to write a paragraph about the thing something was done to.

There's a second consequence, and it settles something people argue about endlessly. The Elements of Style says the split infinitive "should be avoided unless the writer wishes to place unusual stress on the adverb." Geoffrey Pullum, in "50 Years of Stupid Grammar Advice", points out that this has the stress exactly backwards:

"Tucking the adverb in before the verb actually de-emphasizes the adverb, so a sentence like 'The dean's statements tend to completely polarize the faculty' places the stress on polarizing the faculty. The way to stress the completeness of the polarization would be to write, 'The dean's statements tend to polarize the faculty completely.'"

He's applying the stress position, and he says so: the book's own earlier section, "Place the emphatic words of a sentence at the end", already implies it.

Two things to keep straight about whose claim this is. Pullum notes that the split-infinitive remark is White's, added in the 1959 revision: "Strunk's original version never even mentioned split infinitives." This course has read Strunk's 1918 text, where the only mention of the construction says it "is in disfavor" and makes no claim about stress at all. So this is not a case of Strunk contradicting himself, and the 1918 book this lesson has otherwise been quoting is not the book making the error.

And Pullum isn't the wholesale opponent he is sometimes made into. In the same essay he calls the style advice "mostly harmless" and says White "often wrote beautifully". His quarrel is with the grammatical claims, not with the enterprise, which is worth knowing before lesson 8 takes up his argument at length.

What to do when they conflict

Sooner or later the two positions pull against each other. The old information that links backward is also the thing you most want emphasised.

Take a case. A paragraph has just told the reader that the fire door was propped open. The next sentence has to say that this is what the inspector found, and "the propped door" is both the old information and the thing you want to land:

The propped door was what the inspector wrote up.

Old information is in the topic position, so the link holds, but the sentence ends on "wrote up", which is not what you wanted emphasised. Before you read on, turn it round yourself so the door lands at the end, and then ask what the turned version costs.

When you have to choose, my own rule, and it's a judgement rather than something Gopen and Swan say, is that the topic position usually wins. A reader who has lost the thread can't be made to emphasise anything, so orientation comes first.

But there's usually a better move than choosing, and it's the one to reach for: split the sentence.

The inspector wrote up the propped door. That door is now the only thing standing between the building and a prohibition notice.

Two sentences, two topic positions, two stress positions, and the door is in both. A sentence trying to do two jobs often just needed to be two sentences.

What people get wrong

"Never use the passive." The passive is how you control the topic position. Banning it costs you the ability to write a paragraph about the thing that had something done to it, which is a great many paragraphs.

"Put the important word first." This is the single most common misreading of this material, and it is backwards. The front of a sentence is for orientation. The end is for emphasis. If you front-load the important word, the reader arrives at the full stop with their weight on whatever happened to be lying around at the end.

"Front-loading a document means front-loading every sentence." Lesson 2 was about where the point of a document goes, and the reason was the reading budget. This lesson is about the internal shape of a sentence, and the reason is how readers process order. They are different scales with different rules, and this is the one place in the course where the advice at two levels genuinely differs.

"Transitions make writing flow." They label flow. Cohesion comes from the joins, and transitions sit on top of the joins rather than making them.

"This is stylistic." Gopen and Swan's claim is about what readers expect of order, not about a stopwatch, and nothing here has been timed. But you read both versions of the bridge paragraph, with the same facts in them, and one of them cost you more. That's the evidence available, and it's the evidence that matters, because it's the same evidence your reader will have.

Practice

Where this goes next

You've now met the whole of the sentence. Lesson 3 gave you its subject and its verb, lesson 4 the distance between them, and this lesson its two ends. Everything else in this course is either larger than a sentence or about deciding what to cut.

Cutting is next, and this lesson has just made it possible. Lesson 6 is about concision, and it couldn't have come earlier, because until you can see the topic position you can't tell padding from the repetition that holds a paragraph together. The commonest way to wreck a passage is to cut the words that were doing the linking.

Go deeper

• Gopen and Swan in full. This lesson takes three of their seven principles. The paper works them through a real scientific paragraph, rewriting it four times, and the demonstration is better than any summary of it.
• Pullum's "50 Years of Stupid Grammar Advice" in full, which is short, funny, and more careful than its title. Lesson 8 takes up its argument properly.
• Strunk's 1918 text, rule 10 in particular. Reading the rule rather than its heading is the point this lesson makes, and it takes two minutes.
• Joseph M. Williams, Style: Lessons in Clarity and Grace, whose lesson on cohesion and coherence is reported to cover this ground at greater length. This course has not read it.

Sources

• George D. Gopen and Judith A. Swan, "The Science of Scientific Writing", American Scientist 78 (1990). Structural principles two, four and seven, and the given and new contract. The three quotations attributed to them in this lesson are from that article, as is their caveat that none of the principles are rules.
• William Strunk Jr., The Elements of Style (1918), rule 10, for the Restoration dramatists pair, for the sentence about the need to make a particular word the subject determining the voice, and for his own qualification that the rule "does not, of course, mean that the writer should entirely discard the passive voice, which is frequently convenient and sometimes necessary." The 1918 text was read for this lesson, and its only mention of the split infinitive says the construction "is in disfavor", with no claim about stress.
• Geoffrey K. Pullum, "50 Years of Stupid Grammar Advice", The Chronicle of Higher Education (2009), quoted verbatim, for the split infinitive and the stress position, for the fact that the stress claim is White's 1959 addition rather than Strunk's, and for "mostly harmless" and "often wrote beautifully". E. B. White's 1959 revision is not read by this course and nothing is cited to it; the split-infinitive wording is quoted as Pullum reports it. Lesson 8 takes up his wider argument.
• Joseph M. Williams, Style: Lessons in Clarity and Grace, whose lesson on cohesion and coherence is reported to cover the same ground at greater length. This course has read the book's structure and central principles as others report them, not the text itself.
• The two versions of the bridge paragraph were recounted for this review: 49 and 53 words, with two sentences two words shorter and one seven words longer, ten words added and six dropped.