Notes on Writing

Richard N. Langlois

February 1997


This is a list of semi-random notes on writing. It is intended primarily as a check-list of common mistakes rather than as any sort of comprehensive guide to writing. There are a number of general books available that intend the latter. On Writing Well by William Zinsser is particularly good, but there are lots of others, including the ubiquitous Elements of Style by Strunk and White. D. N. McCloskey has produced a little book, called The Writing of Economics (Macmillan, 1987), geared specifically for our profession. I strongly recommend it. The serious dilettante should also consult the pronouncements of such writing gurus as William Safire, Edwin Newman, and John Simon.


Good writing requires the mastery of two aspects of technique. One is grammar in the broad sense -- the basic rules of language. This would include syntax, punctuation, and usage. But good writing also requires style -- the art of applying the rules of language. One can easily imagine a piece of writing that violates no rules of grammar but nonetheless displays a terrible writing style. (The converse is harder to imagine.)

The term "style" is often taken loosely to comprehend not only style in the narrow sense but also the larger issues of structure and form. These latter represent the "macro" aspects of writing -- the overall organization of a text and the logic with which the pieces are integrated into a coherent whole. By contrast, style in the narrow sense has to do with the "micro" aspects of writing. It deals, if you will, not with overall organization (structure) or with the way sections and paragraphs fit together (form) but with the ways words are tied together to build sentences and paragraphs. I will be concerned here with this narrower conception of style.

The passive voice.

Writing in a clear and lively style is ultimately a skill; it is a feat of what the philosopher Michael Polanyi called tacit knowledge. For that reason, one can't learn to be a good writer by following some list of explicit rules. As in any skill, one learns writing only by practice and by imitating accomplished writers.

But explicit rules can sometimes be useful in a negative way; they can often tell us what not to do. In my view, perhaps the most important rule in this category has to do with the passive voice. The rule is a simple one: don't use it.

Obviously, this is a rule that an accomplished writer can (sometimes) violate; otherwise, there would be no need for the passive voice at all. But it is also quite easy to do without the passive voice in virtually all circumstances. It is almost always weak, diffuse, pseudo-scholarly, and bureaucratic. And avoiding it is perhaps the single best way to improve your style.


Lifeless, bureaucratic turns of phrase are not limited to expressions using the passive voice (expressions like "it should be noted that..."). One ubiquitous offender in scholarly (especially student) writing is the verb to state, as in "Adam Smith stated that..." The rule here is the same: don't use it. Almost any synonym -- "suggested," "wrote," "argued," "opined," or just plain "said" -- is better. (The only time I use the verb to state is when I'm trying to set a mock-serious tone.)

The first person.

Many people use the passive voice out of a fear of using the first person singular. The fear is unjustified. If you can't phrase something in a way that is both lively and impersonal, then go ahead and say "I." Similarly, you should use "we" if you have a co-author or if you are speaking for a well-defined group.

In mathematical writing, one often finds what I call "the engineering 'we.'" "First we divide by x and then we integrate over the real numbers," etc. This can be construed as referring to the author and the reader, who are jointly performing the mathematical operations. But many writers treat it more like the physician's "bedside 'we'" ("How are 'we' feeling today"?); and others use it as a rhetorical crutch and thought-substitute. In either case, too much of the mathematician's "we" leads to insipid prose.

Telegram style.

English is in part a Germanic language. But one area in which English does not follow German is the use of extended adjectival constructions before the noun. This is sometimes called "telegram style," and it is a practice particularly prevalent in writing about economics.

To a limited extent, stacking adjectives before the noun is unavoidable -- and even adds a little punch to one's style. But it's one thing to talk of "firm-size variables" or "sales-growth projections" and quite another thing to describe "prior period sales changes" or "the total net long-term portfolio capital flow." It may take slightly more space to say "sales changes in the prior period" or "the total net flow of long-term portfolio capital"; but it definitely scores lower on the bureaucratic-prose index.

By the way, attention to hyphenation is important when stacking modifiers. When two or more nouns are used as adjectives modifying a third, absence of hyphens indicates that all the adjectives modify the same noun. For example, a foreign exchange student is an exchange student from overseas; but a foreign-exchange student is someone who studies foreign exchange.

Newsdesk style.

News reporters have the job of conveying a lot of information in a small amount of space. Thus, they often cram together several unconnected pieces of information in the same sentence. "An unemployed electrician from Florida who likes tapioca pudding for dessert, Hawkins, 36, admitted at his trial, which entered its 257th day yesterday, that he strangled his estranged wife in Hawaii, which was the 50th state to join the Union...." You're not a newspaper reporter. Stick to one thought per sentence.


It is a mistake to think that elegant writing means avoiding all contractions (saying "cannot" instead of "can't," for instance). In modern American style there is no such rule. The best advice is to mix contractions and non-contracted forms, and to avoid a pedantic sound on the one hand and an excessively breezy sound on the other.

Read your writing out loud.

The best test of your writing style is to read what you've written -- out loud if necessary. Be sure to read it as written, pausing where your punctuation really demands and adding inflection the way the text actually requires. Is the text choppy and hard to read? Is it monotonous -- with all the sentences equal in length? Cadence is important. Try for a mix in which short, clean sentences punctuate long, elegant ones.


From the artistic realm of style we move suddenly to the technicalities of punctuation. Here there is greater scope for explicit rules, even if their correct application ultimately requires its own kind of art.

Quotation marks.

And other punctuation marks.

Do commas, periods, etc., go inside or outside of quotation marks? In so-called American Preferred Style, the answer is straightforward. Commas and periods always go inside quotation marks. Always. Without exception. Colons and semicolons always go outside quotation marks. And question marks and exclamation points vary according to whether they are logically part of the quoted material. This rule may seem illogical (it evidently developed as a way of saving space in typesetting); but it remains the preferred rule.

Internal and external.

In American style, primary quotations take double quotes. Internal quotations (quoted material within a quotation) take single quotes. Further nested levels simply alternate double and single quotes.

Block quotations.

As a general rule, quotations longer than eight lines should be set off from the main text -- indented and single spaced. Do not put quotation marks around such indented text. Remember that setting the quoted material off from the main text does not make it logically separate from the main text; that is, punctuate and capitalize the quotation (and use ellipsis) just as if the quote were part of the main text. Avoid prefacing quotes with expressions like "Adam Smith wrote: ..." Try to integrate the block quote into the text. Introduce it with a complete sentence like "Adam Smith argued this way." Or use the following trick. "The statesman," wrote Smith,
who should attempt to direct private people in what manner they ought to employ their capitals, would not only load himself with a most unnecessary attention, but assume an authority which could safely be trusted, not only to no single person, but to no council or senate whatever, and which would nowhere be so dangerous as in the hands of a man who had folly and presumption enough to fancy himself fit to exercise it. (Smith, 1776 [1976, p. 456].)
Note also that the text following the block quotation should not be indented unless it is logically a new paragraph.


An equation is syntactically a sentence or part of a sentence, and requires the same punctuation as a verbal expression: commas, semicolons, and periods.

Hyphenating after adverbs.

Never hyphenate between an adverb ending in "ly" and the adjective (or participle) it modifies. For example, it's the "theoretically predicted value" not the "theoretically-predicted value." Don't ask me why this is so.

You should hyphenate in other cases. But notice that the hyphenation occurs only when the compound modifies the noun directly -- not when it is used as a predicate adjective. Thus, "the semicolon is a much-misused and oft-neglected punctuation mark"; but "the semicolon is much misused and often neglected."

The semicolon.

As I was saying, the semicolon is a much-misused and oft-neglected punctuation mark. It is used to separate two clauses that are complete (that have their own subjects and verbs) but that do not require the force of separate sentences. When to use a semicolon is thus a matter of style. Also, semicolons should be used to separate items in a list when those items are relatively complex or when commas would be confusing.

The colon.

Some people would argue that a colon should be used only after a complete clause. I wouldn't go that far; but it is true that the colon ought not to be relegated to introducing lists or equations. The effect of the colon is to concentrate the force of what went before into what follows. As such, it can be used legitimately in a wide variety of creative ways.

The question mark.

Remember that the question mark is a full-fledged punctuation mark, with all the rights and privileges appertaining thereto. That means that you never need to follow a question mark with a comma or a period. It's "What time is it? he asked," not "What time is it?, he asked."


By syntax I mean primarily sentence logic. The two most frequent problems here are misplaced clauses and faulty agreement. But the message is a more general one: make sure your sentences mean what you think they mean.

Misplaced clauses.

A perennial problem in writing is the misplaced or floating clause, especially the introductory clause. Consider this sentence. "As an economist, costs and benefits are important concerns to me." It seems to make sense, but it is in fact illogical. The clause "as an economist" is obviously intended to refer to the speaker, whereas by its placement it actually modifies "costs and benefits." Costs and benefits are not an economist. Make sure that the word immediately after an introductory clause is always the word you want the clause to modify.

Faulty agreement.

The most frequent agreement problem I encounter is in the agreement in number of verbs with pronouns like each, any, none, or every. All of these are singular, and thus take singular verbs. (Examples: Each of the billions and billions of stars is a sun like our own. None of them is made of green cheese.) The common mistake is to make the verb agree with a nearby object of the preposition rather than with the subject.

Placement of "only."

"I only have eyes for you," says the song. What this literally means is that I have eyes -- and no other bodily parts -- for you. Presumably, the singer really intends something like "I have eyes only for you." While it has become conventional in spoken English to let "only" sit near the verb even if it modifies some other part of the sentence, it is not acceptable in written English. Put the word near what it really modifies.

Split infinitives.



We now move down to the most microscopic level of writing: the individual words themselves. Usage has to do with the meanings of words and the rules for when they should -- and shouldn't -- be used. What follows is a non-comprehensive compendium.


Always use the relative pronoun "that" to introduce restrictive clauses and "which" to introduce non-restrictive clauses. What does that mean? A restrictive clause is one that restricts a noun to a particular instance or class. For example: "The book that was on the table is mine." Contrast this with the non-restrictive variant: "The book, which was on the table, is mine." (This latter type of clause is also called a descriptive clause.) Notice that the non-restrictive clause is set off by commas. This rule is of recent vintage (as far as I can tell), and it is still routinely disregarded, even by some accomplished writers. But those in the know now regard it as gospel.

Due to.

The expression "due to" should be used only to attribute a state of affairs, not to indicate causality. For instance, "this equation is due to Einstein" is OK; but "the volcano erupted due to underground lava pressure" is wrong. In 90 per cent of the cases in which people use "due to," they should have used "because of." Incidentally, I know of absolutely no instance in which the horrendous expression "due to the fact that" should be used. Get rid of it.


Although English permits many adverbs to be used in dangling, free-floating form, current opinion has singled out "hopefully" as a major no-no. Never use this expression to mean "I hope" or "it is to be hoped." You can use it only as an ordinary adverb modifying a specific verb. "Hopefully this megatron space-prober will work on the first try, said Tom" is out; but "I know this megatron space-prober will work on the first try, Tom said hopefully" is OK. In the first case "hopefully" just floats around the sentence trying to convey guarded optimism; in the second case it modifies the verb "said." Other free-floaters like "unfortunately" and "happily" have passed safely into accepted usage.


Fortuitously means simply accidentally or by chance. It does not mean "fortunately" or "luckily." It does not even mean "by lucky chance." Same goes for the adjective form, of course.


If you want to convey the idea that something is happening now, use currently. ("The theater is currently showing three Marx Brothers flicks.") Or, better yet, just say "now." I object to the use of presently as a synonym for "currently" or "now" -- although I'm probably on the losing side of this one. Presently means soon or in a little while. ("Dinner will be served presently.") "Momentarily" means for a moment not in a moment. When the captain says that the aircraft "will be in the air momentarily," you'd better hope that his usage is bad.


The adverb continuously must refer (perhaps metaphorically) to a process that is continuous in essentially the mathematical sense -- there are no gaps or holes. ("His routine was a continuous progression of one-liners.") If you want to refer to repeated events that are discrete or intermittent, use continually. ("He remonstrated continually [not continuously] on the virtues of sacerdotal celibacy.") Again, same goes for the adjective.


To comprise means to include. If you can't substitute "include" for "comprise" in your sentence, you're misusing the word. More formally, "comprise" can be used only in reference to a proper subset of a specified set. That is, "New England comprises Massachusetts, Connecticut and Rhode Island"; but "New England is composed (or made up) of six states." The whole comprises the parts, not the other way around. And the ubiquitous expression "comprised of" is always wrong.

different from/differently than.

The adjective form is always followed by "from," the adverb form by "than." Thus: "I came up with an answer different from (not than) his. That's because I approached the problem differently than he did."


Don't forget that "between" should be used only when the connection involves precisely two things; "among" should be used for three or more. "The choice among the twelve contestants resolved itself into a choice between the two leading candidates."

Less than/fewer than.

As in the case of continuously versus continually, the desideratum here is continuity versus discreteness. Use "less than" when the comparison involves a continuously measurable quantity ("less than three inches long") or when the number involved is large ("less than a billion stars"); use "fewer than" when the numbers involved are small and discrete ("fewer than ten people per lifeboat"). The sign over the express-checkout lane in the supermarket is wrong.


You can convince someone of something, but you can't convince someone to do something. You persuade that person to do it.

Ensure and enquire.

I've never seen anything written about this, but my own rule is to use ensure when I have in mind the loose or metaphoric meaning ("He stayed behind to ensure that everyone was safe") and to use insure only when I mean the technical activity of providing insurance ("This procedure effectively insured the farmers against crop losses"). Similarly, I use "enquire" for the general activity of asking, and reserve "inquire" for the kind of penetrating research that scientists and detectives do.


There are some adjectives and participles (pregnant and dead come to mind) that can't be qualified (that can't be modified with an adverb). You can't be partially dead (in a non-metaphoric sense) or kind of pregnant. Unique is another example: it means "one of a kind," and to say "rather unique" is illogical and considered unacceptable in written English. There are other such adjectives, so be on the lookout.

Nouns as verbs.

It is a long-established process in English to transform nouns into verbs (a pitcher pitches, a cook cooks, etc.). Such transformations can add immediacy to the language; but carried too far, they can rob the language of elegance and give it a boring bureaucratic tone. A gentleman establishes priorities; a bureaucrat prioritizes. The rule should be to transform the noun to a verb if the transformation is a fresh metaphor, but to avoid such noun-verbs when they sound like bureaucratic buzzwords. Received opinion would draw the line, for example, at such words as author, critique, and host. One writes (not authors) a book; criticizes (not critiques) an essay; and plays host at (not hosts) a party. It could be worse: people are now talking of "gifting" one another at Christmas.

He, he or she, s/he.

There is a rule from time immemorial in English that the pronoun "he" can be used in some circumstances as a generic or androgynous pronoun -- one that refers to both men and women. It has become abundantly clear in the last few years that this rule has passed into history. Whatever you may think of political correctness, it is now considered Neanderthal among the intelligentsia not to say "he or she" and, more generally, to arrange one's prose for greater androgyny. Use expressions like "one" or "the economic agent." Under no circumstance, however, will I -- or should any human being of either sex -- use the formula "s/he." It's something I would expect to see in regulations promulgated by the Ministry of Sex Equality.

Back to Home Page.