Rules, and where they come from

Everything in this course so far has asked you to trust it. Readers take in about a fifth of a page. Readers expect a subject to be followed by its verb. Readers emphasise what arrives last. Those are claims about people, and you have taken them from research this course read on your behalf.

Meanwhile you have spent your life being handed a different kind of writing advice. Never split an infinitive. Never start a sentence with but. Never use the passive. Those usually arrive with no evidence attached, and often with real conviction behind them.

This lesson is about the difference, and about how you check for yourself. It uses the most famous writing book in English and the most famous attack on it. The outcome is messier than either side would like, which is the point.

The difference is worth naming before we start, because it's the whole of objective one. A rule like "readers emphasise what arrives last" is a description: somebody looked at what readers do and reported it, and you can go and check. A rule like "never split an infinitive" is a prohibition: it tells you what not to do and offers nothing to check. The first kind can be wrong, and that's a virtue, because a claim that could be wrong is a claim you can test.

Gopen and Swan, whose principles this course has been teaching since lesson 2, are careful about exactly this. Their own summary says: "None of these reader-expectation principles should be considered 'rules.'" They are describing what readers expect, and they say plainly that any expectation can be violated to good effect. That's what a description looks like when its authors are honest about what they've got. Hold it next to "never use the passive" and you can see the difference without any further argument.

The passive, since it is the one everyone thinks they know

Start here, because the passive is the rule most often repeated by people who can't apply it.

A passive clause takes the thing acted upon and makes it the subject. In English it is built from a form of be (occasionally get) plus a past participle, and it can optionally name the actor in a by phrase. "The bill was paid" is passive. "The bill was paid by an anonymous benefactor" is the same clause with the actor restored.

The test: could you add by someone without breaking the sentence? If yes, it's passive.

Geoffrey Pullum, in "50 Years of Stupid Grammar Advice", reports what has happened to that distinction:

"It is typical for college graduates today to be unable to distinguish active from passive clauses. They often equate the grammatical notion of being passive with the semantic one of not specifying the agent of an action. (They think 'a bus exploded' is passive because it doesn't say whether terrorists did it.)"

"A bus exploded" is as active as a sentence can be. The bus is the subject and it's doing the exploding. What people react to is that nobody is blamed, which is a different complaint, often a fair one, and not about voice.

And when the passive is right

Lesson 5 gave you this already. The passive exists so that you can put something other than the actor in the topic position. If the paragraph is about the patient, "she was discharged on the fourth day" is correct, and the active version quietly changes the subject to the consultant.

Here is the part worth knowing before the argument starts. That observation is Strunk's, from 1918, 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 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."

Both of those sentences are passive, so the choice he is describing isn't between active and passive at all; it's about which word gets to be the subject. That's the topic position, named by a composition teacher seventy-two years before Gopen and Swan wrote it up, printed under a heading that reads like a prohibition. Keep it in mind for the next section, because it's the strongest card the book has and it is easy to miss when the argument gets loud.

Pullum's illustration makes the same point from the other end, that the complaint is usually about information order rather than voice:

"For me to report that I paid my bill by saying 'The bill was paid by me,' with no stress on 'me,' would sound inane. (I'm the utterer, and the utterer always counts as familiar and well established in the discourse.) But that is no argument against passives generally. 'The bill was paid by an anonymous benefactor' sounds perfectly natural."

The book, and the attack

The book. The Elements of Style began as a slim volume William Strunk printed privately for his Cornell students in 1918. E. B. White, who had taken the class in 1919, revised and expanded it for commercial publication in 1959. It has sold millions.

Its defenders hold that it is a compact set of maxims that has demonstrably helped generations of novices, that its brevity is precisely its virtue, and that a beginner who follows it writes better than a beginner who follows nothing. That case is not silly, and nothing below refutes it.

Strunk's own aim, from the 1918 text, is modest: the book "aims to give in brief space the principal requirements of plain English style", concentrating on "essentials, the rules of usage and principles of composition most commonly violated". It does not claim to be a grammar.

The attack. Pullum's essay marked the book's fiftieth anniversary in 2009. He is head of linguistics and English language at the University of Edinburgh and co-author of The Cambridge Grammar of the English Language.

"The Elements of Style does not deserve the enormous esteem in which it is held by American college graduates. Its advice ranges from limp platitudes to inconsistent nonsense. Its enormous influence has not improved American students' grasp of English grammar; it has significantly degraded it."

His evidence, in two parts

Part one, the misdiagnosed examples. Of the four pairs under the heading "Use the active voice", he says "a staggering three out of the four are mistaken diagnoses", and works through them. On one, note his hedge: "'The reason that he left college was that his health became impaired' is presumably fingered as passive because of 'impaired,' but that's a mistake."

Part two, and this is the stronger half. He argues the book cannot keep its own rules, on the same page as it states them:

"There is of course nothing wrong with writing passives and negatives and adjectives and adverbs. I'm not nitpicking the authors' writing style. White, in particular, often wrote beautifully, and his old professor would have been proud of him. What's wrong is that the grammatical advice proffered in Elements is so misplaced and inaccurate that counterexamples often show up in the authors' own prose on the very same page."

That is the third thing usually cut, and it belongs with the other two: the one sentence in the essay where Pullum praises White, sitting immediately before the charge. Then the charge:

"'Write with nouns and verbs, not with adjectives and adverbs,' they insist. (The motivation of this mysterious decree remains unclear to me.) And then, in the very next sentence, comes a negative passive clause containing three adjectives: 'The adjective hasn't been built that can pull a weak or inaccurate noun out of a tight place.'"

He does the same to "Keep related words together", whose own statement of the rule separates a subject from its verb.

That's the argument to answer, and this course can't check it, because those sentences are White's 1959 text and this course hasn't read it.

Opening the primary text, and what it actually shows

So take the half that can be checked. Strunk's 1918 rule 10 is out of copyright and available in full.

Read it and the first thing you find is that Pullum has read it too, and says so:

"After this unpromising start, there is some fairly sensible style advice: The authors explicitly say they do not mean 'that the writer should entirely discard the passive voice,' which is 'frequently convenient and sometimes necessary.' They give good examples to show that the choice between active and passive may depend on the topic under discussion."

He quotes the qualification and credits it. He also grants the topic-dependence point, which is the same observation lesson 5 built on. Anyone who tells you Pullum overlooked the moderation in the passive rule has not read his essay.

What the 1918 text does add is the sentence that introduces the disputed examples:

"Many a tame sentence of description or exposition can be made lively and emphatic by substituting a verb in the active voice for some such perfunctory expression as there is, or could be heard."

Strunk names the fault himself, and it is the limp existential there is and the perfunctory could be heard, not the passive. That's consistent with Pullum's own "presumably": he is inferring what the examples are offered as, from their position under an active-voice heading, rather than reporting a label the book applies. Both readings are available from the page. His is the natural one for a reader who trusts headings, which is most readers, and that is the whole mechanism of the damage he describes.

The finding that cuts the other way

Opening the text is only worth doing if you report what you find, including the parts that go against the side you started on.

Pullum writes that "Strunk's original version never even mentioned split infinitives." It does. The 1918 glossary carries an entry:

"Split Infinitive. There is precedent from the fourteenth century downward for interposing an adverb between to and the infinitive which it governs, but the construction is in disfavor and is avoided by nearly all careful writers."

And the however rule, which this lesson is about to use as its specimen of an untested prohibition, is Strunk's too: "However. In the meaning nevertheless, not to come first in its sentence or clause."

So the primary text corrects the critic as well as complicating the charge. Neither of these is fatal to Pullum's case, which doesn't rest on them. Both are the kind of thing you only ever find by opening the book, which is the actual lesson here.

Lesson 7 said a table is for content with two dimensions, where every row is described on the same axes. This is that: five charges, each with a source and a verdict.

Read down the middle column and you can see why "Strunk and White" is a misleading way to name a defendant.

Where that leaves the dispute. Pullum's central charge rests mainly on the same-page self-contradictions, which are quoted from a text this course has not read and which look serious as quoted. On the examples, the 1918 wording complicates the charge without refuting it, and he hedged it himself. Two of his sharpest specific hits belong to White. One of his incidental claims is wrong. The damage he documents, which he doesn't claim as his main point, is real and isn't in dispute.

If that's unsatisfying, good. It's what the evidence supports.

The method, which is the actual lesson

Notice what moved the argument. Not authority, not intuition, not asking which writer is more distinguished. The text, opened, by someone willing to publish what they found.

Pullum uses the same move on the none rule. He searches three books by writers publishing while Strunk was teaching and finds plural agreement in each: Wilde's The Importance of Being Earnest has "None of us are perfect"; Stoker's Dracula has "I think that none of us were surprised when we were asked to see Mrs. Harker a little before the time of sunset"; Montgomery's Anne of Avonlea has "None of us ever do". His question is the one to steal: "Is the intelligent student supposed to believe that Stoker, Wilde, and Montgomery didn't know how to write?"

The method in four steps:

1. State the rule precisely. Not "however is wrong at the start of a sentence" but "however, meaning nevertheless, should not begin a sentence or clause". Vague rules can't be tested, which is why a rule stated loosely can outlive the evidence against it.

2. Find a corpus. Pullum points to Project Gutenberg, which is free, searchable, and full of writers nobody accuses of illiteracy.

3. Count, both ways. Mark Liberman checked half a dozen of Mark Twain's books and found roughly seven sentence-initial howevers for each three placed after the subject; in five books by Henry James the ratio was one to fifteen.

4. Say what the numbers can and cannot show. Here you need to know which question you're answering, because the two schools answer different ones.

   On the descriptive view, which is the one this course works from, a rule contradicted by the usage of good writers of its own period is not a rule of English. English is what its competent users do, and a prohibition they ignore is a report about somebody's preference.

   There is a prescriptive reply, and it isn't silly. A standard written variety isn't the same thing as the sum of what fluent people write. It's maintained deliberately, by editors and style guides and teaching, and it exists so that strangers can read each other without friction. On that view Twain's usage is evidence about Twain, not a repeal, and a convention can be worth keeping because it is a convention. Most working copy editors hold some version of this, and they aren't fools.

   What a count settles is the narrow question, and it settles it either way: is this rule a description of how careful writers of the period wrote? If the answer is no, then anyone presenting it as a fact about English is wrong, whatever else they may be right about. Whether you should follow it anyway, in a house that has chosen it, is a separate question the numbers don't touch. Keep the two apart and you can be evidence-led without being rude to your copy editor.

What people get wrong

"Never split an infinitive." The split infinitive has always been grammatical. The reason usually given is also backwards: the claim that splitting places "unusual stress on the adverb" is White's, from a chapter this course hasn't read and quotes only as Pullum reports it, and Pullum shows that tucking the adverb in before the verb de-emphasises it. That's lesson 5's stress position doing the work.

"Never start a sentence with however meaning nevertheless." Contradicted by counting, which is why counting is in this lesson. The rules about initial but and and are usually repeated in the same breath, and this course has not counted those, so it isn't going to tell you they are dead. Count them yourself; they're the easiest search in the exercise below.

"Never use which for a restrictive clause." Pullum: "There was never a period in the history of English when 'which' at the beginning of a restrictive relative clause was an error." Plenty of publishers enforce it as house style, which is a choice about consistency rather than a claim about English.

"A style book contains the rules of English." Elements contains one teacher's list of what his students most often got wrong in 1918, plus a later writer's additions. Strunk said as much himself, in the aim quoted above.

"So the book is worthless." The evidence doesn't support that. Its style advice survives the attack; its rule 13, read in full rather than as a four-word slogan, is the source lesson 6 is built on; and its own author states the topic-position principle in 1918, quoted near the top of this lesson in his own words.

"So the attack is unfair." Nor that. Pullum's central evidence is quoted from the text and has not been answered here, and the damage he describes is real.

"Following the rules makes you a fool." Nothing in this lesson supports that either. Most people who repeat these rules were taught them by someone they trusted, which is how nearly everybody learns nearly everything, including you and including the person who wrote this.

Go deeper

• Pullum's essay in full. It's short, funnier than its title, and more careful than its reputation, and reading it whole is the best possible demonstration of this lesson's point about quoting people.
• Strunk's 1918 text at Project Gutenberg, free and searchable. Every claim this lesson makes about the original was checked there, and you can check them too in about twenty minutes.
• Rodney Huddleston and Geoffrey Pullum, A Student's Introduction to English Grammar, if you want the grammar Pullum says Elements gets wrong, set out properly.
• Joseph M. Williams, Style: Lessons in Clarity and Grace, which Pullum himself recommends in place of Elements. This course has not read it, and says so wherever it comes up.

Practice

Where this goes next

You now have the craft this course teaches, and a way of judging advice you meet later, including advice from here.

One thing remains, and almost no writing course includes it. Everything so far has been about producing a better document. The last lesson is about finding out whether it worked, which is a different activity with different tools, and it involves another person.

Sources

• William Strunk Jr., The Elements of Style (privately printed, 1918), public domain and read in full for this course. Rule 10 for the active voice material, its qualification about the passive, the "perfunctory expression" sentence and the Restoration dramatists pair; rule 3 for restrictive and non-restrictive clauses; the glossary for the split infinitive and however entries.
• Geoffrey K. Pullum, "50 Years of Stupid Grammar Advice", The Chronicle of Higher Education, 17 April 2009. Every Pullum quotation in this lesson is from that essay, including his concession that the style advice is mostly harmless, his statement of which half of the case is his, the same-page self-contradiction evidence, the none and which material, and Mark Liberman's counts of however in Twain and James.
• E. B. White's 1959 revision has not been read by this course, and is in copyright. Every statement above about what White added or wrote rests on Pullum's account or on the absence of the material from the 1918 text. That also means White's defenders are unheard here, as well as his book unread.
• George D. Gopen and Judith A. Swan, "The Science of Scientific Writing" (1990), the model of the other kind of rule, the kind that describes what readers do. Their own caveat, that "none of these reader-expectation principles should be considered 'rules'", is the sentence this lesson is built around.