Useful Things

What flavour of English do you want?

For the bibliography and a concise summary of some key points, see the handout (PDF, 72 KB)

I thought I wouldn't call this session "Register, collocation, and reflected meaning" because, well, that sounded a little dry. And I'm going to be starting into this subject with the use of a metaphor of sort. The metaphor I'm going to be using—and I think it's a pretty viable one—is, as you may have guessed, that a piece of a text is like a piece of food. A document is like a dish. Words are like ingredients.

Words, glorious words—
we know how to use them;
(now) I'll give you some tips
on how you should choose them!
When you are fine-tuning texts,
mind that you go through them
knowing what your readers expect
(and) how they'll respond to them!
Words, glorious words,
What wouldn't we give for
that nuance of tone…
that's all that we live for!
Why should we be fated to
do nothing but brood
on words,
marvellous words,
wonderful words,
glorious words!

When you're approaching a piece of text and deciding what kind of usage is appropriate for it, it's not enough just to say "This is a business document" or "This is an ad." You can fine-tune your approach every bit as much on the page as in the kitchen. When you're writing or editing something, certainly you want to know if you're making custard flan, or meat loaf, or seven-grain bread, or curry. But you also want to know, if you're editing curry, whether it's meat or vegetarian, Thai green curry, Thai red curry, Vindaloo curry, Madras curry, chicken with jerk curry sauce… And When a text hits your eye like a big pizza pie… what kinda pizza is it? What's more, fusion cuisine can sometimes be very useful, as I'll describe in greater detail in a few minutes.

It goes beyond that. As I've said, every word is an ingredient, and sometimes a single word can make a real difference. Obviously you don't use a peppermint word in a pepperoni pizza document. But should you use a garlic word in a meatloaf article? What if you're editing satay… but your readers are allergic to peanuts? Word choices can have important effects, and I'm going to look at that in more detail in the second course of my presentation. The latter half, I mean.

Now, you're all professional editors, so I certainly hope that when I say "Know your audience" I'm not telling you something new. But what I want to do is talk about cooking techniques here—things you can do not just to fine-tune your text but to taste-test it against other documents and to make sure that it's put together appropriately—not just so your soufflé won't be too heavy but so your meatloaf won't be too fluffy. I want to steer you in the direction of deciding, for instance, whether your text is too noun-heavy—or not noun-heavy enough—and other things you should look at in word choice.

The different varieties of English that we use in different contexts are, in the parlance of linguists, registers. We're not talking about dialects here—things that vary according to place or social grouping. We're talking about varieties within a given dialect. Once we start looking at language use as social behaviour, we recognize that it's used different ways in different situations to produce different results. Every utterance always participates in a definition of the situation of its utterance and the relation between utterer and receiver—between speaker or writer and hearer or reader. Write that down. [write on blackboard]

There are tactics involved: for instance, do you expect the reader to respond actively? How? A very interesting recent study of instant messaging communication (Tagliamonte and Denis—see the bibliography on your handout) gives us this breakdown of use of [write on the board] first-person, second-person, and third-person pronouns in spoken, instant-messaged, and written English (within the corpora studied, of course; your results may vary): in spoken English, it's 53% first-person, 18% second-person, and 29% third-person; in instant messaging, it's 62% first-person, 21% second, and 17% third; in written, it's… any guesses? 21% first, 4% second—yes, four—and 75% third. Why? Because of the context of the utterances, what they're usually used to do, what they're best suited for. Information conveyance? Social negotiation? Planning? Instruction—but if it's instruction, what kind of instruction, what approach? Now, these are just three different media; it's like bread, ice cream, and soup. There are much finer distinctions. And you can use little things like this as points of analysis and comparison. If your choice is between, say, "People with any of the following conditions should not use this drug" and "You should not use this drug if you have any of the following conditions," you can consider which milieu of discourse you're tapping into. Do you want to sound like a book or a like a person? And if a person, in what relation? If a book, what kind? (It may be more effective to sound like a book, depending on what your readers will be receptive to!)

This distinction between different media is actually one of three key distinctions in registers suggested by Halliday, McIntosh and Strevens in "The Users and Uses of Language": field of discourse, mode of discourse, and style of discourse. There are more ways of analyzing registers, but let me start with this set of three.

Field of discourse is what's going on—the context in which the discourse is occurring. Politics? Academics? Conference presentation? Mode of discourse is the medium, as just mentioned. Spoken, for instance—or spoken with a handout to refer to (but is the register of the handout the same as the register of the speaker? and to what extent does the presence of each affect the other?). Style of discourse, as Halliday et al. put it, "refers to the relations among the participants"—is this polite or rude, formal or relaxed? And we can see that there is often real latitude—a given field may tend to call for a given style, but you can decide, on the basis of your knowledge of the audience and their expectations as well as of what the author does best, whether, for instance, your conference presentation needs to be entirely straight-faced and serious or whether there's room for some jokes, a song or two, what have you. There will be a line, and it will vary between audience members. I once gave a presentation to a group of theatre scholars from around the world, and, while the Americans appreciated my lively approach, a German scholar said to me afterwards, "You must sit, and speak in a calm, even tone."

You're not going to be able to please all your readers at the same time all the time. But you can slice and dice the analysis of style even more finely –in fact, [ad voice] you can make a julienne. [/ad voice] Some rather clever sorts named Susan Conrad and Douglas Biber came up with a thing called multi-dimensional analysis. This involves tagging a set of texts for different syntactic features, and running some rather sophisticated mathematical analyses on them. If you analyze a set of texts in this way, you will find that certain sets of features tend to co-vary. Any given text can be placed on a point on several different scales—more or fewer of this set of features, more or fewer of that set. Effectively, they can be measured in several different dimensions: there are several different ways they could differ from or be similar to other texts. Now, each analysis is of a specific set of texts, and the dimensions may vary between sets. And the full-on mathematical analysis of corpus data in large quantities is probably more than you're going to want to do when making editing decisions. But I'd like to mention three dimensions that have turned up consistently in their analyses of spoken data, as per Douglas Biber, as these dimensions will also be useful to us in analyzing printed texts.

The first, and perhaps most important, dimension is what can be called "information-focused versus interactive discourse." This is probably the scale you will find most useful. And the features you'll find that correlate with one end or the other of the scale are listed on your handout:

Information-focused

Interactive

Greater word length
Nominalizations
Prepositional phrases
Abstract nouns
Relative clauses
Attributive adjectives
Passive verb phrases

Present-tense verbs
Contractions
First-person pronouns
Second-person pronouns
Activity verbs

In other words, probably to no one's great surprise, if you use a phrase such as "The measurement of the sesquipedalian tendencies of lexical items was performed" it will tell the readers that the text is all about providing information in a disinterested and accurate way, and if you say instead "We measured the length of the words" you will communicate more personal involvement in the situation—and perhaps less objectivity. Now, a lot of these judgements are pretty obvious, but this information allows you to identify more clearly just why and how one text seems more dispassionate than another, and you can discern rather fine differences—and tweak texts a bit one way or another.

The second common dimension is what Biber calls "stance versus context-focused discourse." Stance means "personal attitudes or indications of likelihood." So the features that tend to indicate it are not surprising, but the features that go with context-focused discourse are more interesting, because there are just two of them:

Stance

Context-focused

That-deletions
Mental verbs
Factual/mental verb + that-clause
Likelihood/mental verb + that-clause
Likelihood adverbs
Adverbial clauses
General hedges
Factual adverbs

Nouns
WH-questions (who, what, when, why, where, how)

So you're not just asking a different question but actually defining the situation between writer and reader somewhat differently when you say "Describe what a person would think will probably happen next" rather than asking "What will be the next thing to happen?"

The third dimension is "narrative-focused discourse." This is pretty straightforward:

Narrative-focused

Non-narrative-focused

Past-tense verbs
Third-person pronouns
Non-factual/communication verb + that-clause
Communication verbs
That-deletions

Present-tense verbs

You will notice that some of the features of one dimension tend to go with or against features distinguishing another dimension. Unsurprisingly, narrative tends not to be involved discourse. But it can be, if you adjust the seasonings just so. "You told me you'd be back in five minutes" is an example of something that's both narrative and involved.

OK, but why does this matter to you? You're not linguists—you're not constrained to simply sit by and say, "My analysis shows that this text is boring and this text is exciting." What editor doesn't want the text she works on to engage the reader? Ah, but sometimes something too engaging actually puts the reader off. If you go to a fancy restaurant, you don't want the waiter to throw his arm around your shoulder and call you by a nickname, right? "C'mere to this table, buddy, you'll love it!" Likewise, you don't, if you're working on a scientific paper, want something that reads like a romance novel—obvious enough, but there are also more subtle ways in which you can undermine your aims. Robert K. Merton, in The Sociology of Science, identified four norms that define the ethos of science: universalism, communism (i.e., sharing knowledge), disinterestedness, and organized skepticism when approaching claims. As Maria José Luzon Marco points out, "Disinterestedness is reflected in the paper by the absence of the authors/researchers, implying that they have no personal interest in proving that a particular claim is true.… This is one of the reasons for the use of grammatical structures which help to avoid mentioning the agent (e.g., passives, nominalizations)." It also follows, conversely, that if your focus is on the immediate personal details of the interaction, you will use words that focus on them.

There are, of course, a whole set of things that go with that. And some of them become a matter of habituation. The readers won't necessarily be able to put their finger on it, but in some kinds of articles, if you don't have enough stacks of nouns and prepositions, it might not feel quite right, and the readers won't accord it as much authority. So if you decide to make an annual report more readable, there is a point at which you may risk having readers feel that it is too chatty, even impertinent. Not reliable. And it goes the other way too. It has been demonstrated—unsurprisingly—that a weak argument or assertion phrased in "proper," impressive-sounding English will be rated more highly by many readers than a strong argument phrased in colloquial or nonstandard English. "The texture of a dish may be an excellent indicator of its overall sapidity." "Thick or thin don't make much diff to how it tastes, mate." The language you use is judged by the company it keeps. Write that down! [write on blackboard] This is also true of individual words, as I will tell you more about in a few minutes. And this can sometimes be very subtle. The diner may not know just why the beef bourguignon doesn't taste right, but the truth is you shouldn't have put that extra pinch of cloves in.

So what this means is that you have to look at the expectation of the readers and the values that they hold important with respect to the text, and you need to look at whether the English you use is consistent with those values—and whether, and in what ways, it's similar to or different from other texts that communicate those or other values and positions.

And you can get really crafty with this. I want to turn away from scientific discourse to advertising to look at an instance of skillful use of different registers, of serving one pragmatic by using text in the style of another pragmatic. I'm not reproducing the ad I'm referring to in your handout because I think there might be a copyright issue with doing so, but I'm sure that I'm OK quoting from it and referring to it up here. I think you'll get the idea clearly enough.

The ad is an ad for diamonds. Well, no, it's not selling specific diamonds for a specific store. It's promoting them in general. It's titled "Diamonds: A gift of nature, a gift of hope." If we look at it in terms of field, mode, and style, we can say right away that the mode is a print magazine, and the field is advertising and advocacy. This ad came out around the same time that the movie Blood Diamond came out, and clearly the diamond producers saw a need to buff their image a bit. They would want to persuade not only the wealthy and those soon to be engaged to marry but also anyone who might carry a placard or join a protest march opposing the sale of diamonds. So this ad may be situated also in political discourse. And diamonds are selling not utility but an image, a self-presentation and self-situation in the world, for the buyer to take on. Diamonds make the owner feel rich; if they can also make the owner feel noble (especially if the alternative is to feel callous), then the buyer is getting two significant image enhancements for the (rather substantial) price of one. So a positive political message is also effectively a positive sales message.

The ad has a fair bit of text, and it starts out like an ad. Michael Toolan has identified some general syntactic characteristics of print advertising:

disjunctive syntax and incomplete sentences;
simple predicate structures, with much use of the simple present tense and common verbs;
complex noun phrases;
lexical cohesion ("repetition and ‘elegant' variation") rather than pronominalization; and
use of a fairly limited, well-established set of positively valued adjectives and verbs.

This ad has almost all of these features in the beginning. Here's an excerpt:

Brilliant. Magical. Mysterious. Unique. Rare. Pure. A timeless gift of nature. And a piece of geological history. Diamonds dazzle the eye and intrigue the mind. For centuries, their sparkle, strength, and addictive allure have mystified cultures around the world. Diamonds are prized not only for their natural beauty, but for their protective energy and healing power.

Disjunctive syntax, simple predicate structures, repetition and variation rather than pronouns, and use of words from a favoured set. But it doesn't use complex noun phrases—you don't see, for example, "brilliant first-water teardrop-cut diamond in 24-karat gold engagement ring" because the ad isn't focusing on the actual hard merchandise to sell. It doesn't want to be seen actually selling. If you move into specifics, you become commercial, and the mercenary overtakes the noble and mystical.

The text then moves on from the misty mythmaking to something that smacks of political oratory. You remember Churchill's "we shall fight on the beaches, we shall fight on the landing grounds, we shall fight in the fields and in the streets, we shall fight in the hills" or Kennedy's "we shall pay any price, bear any burden, meet any hardship, support any friend, oppose any foe, to assure the survival and the success of liberty"? Listen to this ad: "No other stone is imbued with such emotion. No other gem represents the purest expression of deepest human feeling. No other product commemorates milestones the way a diamond does. Nothing captures imaginations like the Diamond Dream." The persuasive and emotional charge of repetition with strong words not only presses a point but also calls forth the register of exhortative speech—which is part of the political milieu. The chef has tossed in just a hint of something—you're not sure what it is—a secret ingredient—and somehow the dish makes you feel… different. The advertiser is now about to become the concerned advocate, and the reader to be cast as the interested citizen with the power to act.

And then, in the next column of text, the ad shifts register again. The average sentence length increases markedly. In the third paragraph of the ad, which I read, there are 114 words and 13 "sentences," however incomplete; the fifth paragraph has 78 words and three sentences—a 200% increase in average sentence length. The sentence structure has also become more involved, with complex subordinate clauses. There is a seven-point bullet list giving facts and numbers, and the vocabulary shifts from words such as "brilliant," "magical," and "mysterious" to words and phrases such as "code of conduct," "communities," "natural resources," "sustainable future," "transform the lives of people," "making a lasting difference," and even "anti-retroviral," "life partners" (rather than "wives" or even "spouses," say), "hospices," and "literacy programs." (In case you're thinking "What?", they're talking about how diamond mining helps the countries in which diamonds are mined—for example, "De Beers Consolidated mines was the first mining company in the world to extend free anti-retroviral treatment to HIV-positive employees, their life partners, and former employees.") If you search any of these terms in Google, you will find near the top of the results web pages of government and not-for-profit bodies, including such organizations as the Red Cross, UNESCO, and the World Wide Fund for Nature. Text published by advocacy organizations often not only is peppered with such phrases but also tends to share several stylistic features with the body of this ad after the fourth paragraph: longer sentences, bullet points (often, as here, containing complete sentences), numbers, dates, place names, historical comparisons, and names of famous role models (in this ad, it's none other than Nelson Mandela). Where the ad uses complex noun phrases at all, it's to speak of not "diamond solitaire pendants" but the "Skills Training Centres" that your purchase helps support. And every sentence from the fifth paragraph to the end of the text, except for one, is a complete sentence and is in adherence with the formal standards of "good English." There are only four contractions in the entire text. The contractions add just a modicum of friendliness and, at the same time, directness to the two sentences they are in, which state points not of factual details but of principle: "That's why it's so important…" and "It's a success story that hasn't been told, but one that should be celebrated." The register is of factual, politically important information. They could communicate this information with ad-like syntax—"Hope. Security. Peace. Advocacy. The basic needs of life. Diamond mines don't just employ thousands. They bring social advancement." But instead of the ad-man's tie and suit and plastic smile, they put on the simple shirt and earnest face of the aid worker. And yet the real function is to increase sales of diamonds—not only by persuading the populace to look favourably on diamonds, but by projecting an image of nobility and generosity that people with money will pay a lot of money to buy into. And they're doing this using not just content but form. The right vocabulary. The right syntax. The language not of valuables but of values.

To bring us back to food, who here has seen Ratatouille? The movie, I mean. The food critic is won over by a version of a simple country dish—ratatouille—that brings to mind his childhood. It needs to be fancy restaurant food—the ratatouille is made into dainty little croquette shapes—but it's that homey basis that clinches it.

So when you're working on a project, you may find it very useful to consider your audience, their values, the kinds of structures and word choices that tend to be consistent with those values, and the kinds of structures that other documents use—documents that are within the genre you're working on, but also other texts that can affect your readers in the way you want. Whenever you're using language, you're doing something to somebody. You're trying to affect them in some way. You're trying to change their mental state, and probably their actions, in some way. If you're a good editor, you'll already have a natural feel for a lot of this. But it can be useful to be more analytic about it.

[to "Puttin' on the Ritz," with the verse first]

English comes in several kinds;
you should know your readers' minds.
Perfect English? No such thing!
It needs a familiar ring!
Compare when analyzing…
what's there can be surprising.
Counting is no sin;
(with) register you cash in!
Different syntax changes sense;
the fix is in, you can't dispense
with these tool kits…
Use the style that fits.

Now, to move on to my second course—moving from the overall assembly and the mode of preparation to the individual ingredients—I'd like to look back at that ad one more time. I'll read you the headline again: "A gift of nature, a gift of hope." OK, three nouns: gift, nature, hope. Diamonds are typically bought to give to another person, signs of the commitment and generosity of the giver, with the recipient normally a loved one (most often fiancée or wife). That they may be a "gift of nature" as well suggests that they are freely and gladly given by Mother Earth, rather than, say, ripped from it by mining operations that despoil large tracts. And "nature" and "natural" are such positive words these days. The idea of "hope" is even more packed: diamonds are often bought as engagement rings, in hope that the suitor will be accepted; the text of this advertisement aims to persuade the reader that diamonds also give hope to disadvantaged people in Africa; and the most famous diamond in history was also named the Hope Diamond.

One of the biggest bedevilments when dealing with words is the idea that the meaning of a word is just what you see in the dictionary. We often find ourselves misled by this misconception, even though in practice we understand that it's not true—if it were, the words "feces" and "fornicate" would be no more or less acceptable than other ostensible synonyms of them. I like the assertion by Geoffrey Leech that one can distinguish seven kinds of meaning—this is on your handout:

conceptual (i.e., denotative—what's in the dictionary);
connotative (properties generally expected to inhere in the object referred to, including positive or negative valuations);
stylistic (what the word or utterance conveys about the social circumstances of use—slang? formalism?);
affective (what is communicated about the feelings of the utterer—for instance, deliberate defiance of social norms and apparent intention of offensiveness, in the case of vulgarities);
reflected (overtones gained by association with other meanings of the same word and with homophones—this includes phonaesthetics, which shapes our expectations of a word's meaning on the basis of its sounds: words that start with gl- in English tend to be associated with things shiny or fat; the louche associations of the sh- beginning are such that the word "swag" is often rendered as "schwag");
collocative (meaning through association with words that tend to occur in the same environment); and
thematic (meaning created through the organization of the expression).

When we're considering word choices for a document, we are of course aware of the conceptual meaning, and no doubt we have a sense of the connotations. We probably also have a decent sense of the stylistic meaning—how formal or informal a word is. Context will often go a long way towards determining the affective meaning: the word "stuff," for instance, can be neutral, negative—"get your stuff out of my way"—or even positive—"this movie is compelling stuff." Thematic meaning is, in the main, such things as the difference between the emphases in "The dog bit the man" and "The man was bitten by the dog"—how your structure will affect the reception of your words. I've already looked at some things that relate to this. The two kinds of meaning I want to look at in particular now are reflected and collocative, because they can often have effects that are overlooked or "hard to put your finger on." And because my big point is that words are known by the company they keep! And we do engage in phonetic profiling, too—"That word looks like a bad word, so I'll assume it's bad."

I'd like to look at just a few examples. I'll start with an interesting paper by Michael Stubbs on the effects of collocation on meaning. First, tell me: what's the opposite of "big"? … "Little," "small," both come up. Now what's the opposite of "large"? … What, all "small" and no "little"? Now, why is that? Clearly, these words have different ranges of meaning. Stubbs analyzed a 2.3-million-word corpus of current English as well as the quotations in the Oxford English Dictionary and found some interesting things about the words these words travel with. Just to start with, girls are almost always little, almost never small, whereas boys are small about half as often as they are little. Hmm.

Well, we know what the difference is between a big fish and a large fish, eh? A big fish can be dinner for four but it's at least as likely a self-important person, whereas a large fish is just that: a large fish. That reminds us that "big" and "little" are used in figurative speech more than "large" and "small." More to the point, the range of meaning of "large" and "small" seems more restricted. "Large boys don't cry"? That would be "big boys," now, wouldn't it? And if I say I have a large brother… that doesn't mean I have a big brother. But, by the way, if I say "big brother," what is big brother doing? …Watching you? Other things that are big, not large: Apple, Bang, business, guns, mouth, time, words… on your handout you'll see a list of some of the most common words that Stubbs found to the right of each of these adjectives.

And if we look "small" and "little," we see a similar, though perhaps not identical, relationship. "Little," for instance, is often used in phrases such as "a little more" and "a little bit." But perhaps more interesting with this pair is the words that we see coming before them. Comparatively small, exceedingly small, infinitely small, numerous small [things], relatively small. Beautiful little [something], charming little, cute little, dainty little, dear little, good little, lovely little, neat little, nice little; but also poor little, dirty little, funny little, wretched little. Which reminds us that endearing things are in the same basic status relation to the speaker as disgusting things are!

What these groupings do is show us the details of the underlying tone of the word—the fine-tuning of its meaning. And they're resonances that will be in the background when you use the word. But they're also how we learn the tone of the word. Nobody spelled it out for us when we were children; if they had, we would be able to say exactly what the difference is. Instead, we just got a sense from the patterns of usage. And this is language-specific—French and German really just have "grand" versus "petit" and "gross" versus "klein."

This question of collocation, as well as the question of reflected meaning, came in very useful to me when I had to make a decision about which word to use in a piece of text—in fact, what I found shaped the general policy at my company regarding the use of these words.

Some of you may remember that a couple of years ago I had EAC list members participate in a little experiment. I tested two little bits of text. Here's the first one:

The side effects listed below are not experienced by everyone who takes this medication. If you are concerned about side effects, discuss the risks and benefits of this medication with your health professional.

Here's the second one:

The side effects listed below are not experienced by everyone who takes this medicine. If you are concerned about side effects, discuss the risks and benefits of this medicine with your health professional.

I tried the first one on half of the respondents and the second on the other half, and asked them to rate it on tone and audience level. Then I tried the other one on each group. The first passage was seen overall as being at a higher level and, to a lesser degree, as friendlier.

Oh, what was the difference between the passages? … Just whether I used the word "medication" or the word "medicine." So I did a little research on these two words. You probably won't be surprised that "medicine" has a rather broader range of meaning than does "medication." Even just in denotations, you can see this—the American Heritage Dictionary gives five definitions:

the science of medicine (as in the Faculty of Medicine of the University of Toronto);
the practice of this science;
"an agent, such as a drug, used to treat disease or injury";
"something that serves as a remedy or corrective" (the metaphorical use of sense 3); and
shamanistic or similar practices and beliefs.

And it's also used more in general cultural contexts. Bartlett's has 19 quotations containing "medicine" and none containing "medication"; in the Oxford Dictionary of Quotations, it's 11 to nothing. So this word comes with a lot more baggage, a lot more resonances—a lot more reflected meaning.

And not all of it is positive. Let's look at collocations: "medicine man," "medicine ball," "medicine show," "take your medicine," "A spoonful of sugar helps the medicine go down."

"Medication" just doesn't have all this stuff hanging off it. And you won't be surprised that it's less used. If you go to wordcount.org, which ranks word frequencies based on the British National Corpus—100 million words, albeit from Britain—you'll see "medicine" in spot 3,371, while "medication" is in spot 11,520. Google also gives about eight times as many hits for "medicine." On the other hand, the Google ratio is four to three in favour of "this medication" over "this medicine"—"medication" is more literal, it seems.

And it's longer, and has the technical-sounding "-ation" ending. And it's learned at a later age. I checked some corpus data of speech by and to American children between ages two and eight, and in the data I ran the numbers on, I got 40 uses of "medicine" and—how many uses of "medication"? … None.

So, given that the text I was working on was health information for consumers, so it needs to be friendly but authoritative, which word do you think I went with?

Now, I'm not saying that you need to run the numbers every time. You have well-developed judgement. Sometimes you might need to call this stuff to your service just when someone questions you. But it's useful to have it in your mind, and it's useful to be aware of these approaches as means of giving a harder edge to what can be a very soft science.

[Puttin' on the Ritz, refrain first]

Want to get some indication
what will fit? Use collocation—
one expects
to hear what comes next.
Figure in reflected meaning;
you begin to shape the leaning
of your bits.
Use the word that fits!

So, when you approach a piece of text, what are you going to look at? You're going to think about who's reading it, of course. You're going to think about what other text it's like—or could or should be like. Think of the kinds of structures it uses, and what they bespeak—the relationship with the reader, the milieu. What kind of cuisine is it? Maybe some fusion cuisine. What type of dish? Don't serve an entrée as a dessert! Do a little research and maybe even some quantitative comparisons. How many cups of nouns, how many ounces of prepositions, what colour of pronoun. Consider what your document sounds like. What it's put together like. What it will taste like for your readers. Don't serve cordon bleu at a barbecue shack—or vice versa. And consider the resonances—the subtle flavours—of the words it uses.

Words, glorious words—
we know how to use them,
but, as you have heard,
mind well how you choose them!
You'll handle most any text,
sermon or conversation
if you mind your register
and watch collocation!
Words, glorious words,
What wouldn't we give for
that nuance of tone…
that's all that we live for!
Why should we be fated to
do nothing but brood
on words,
magical words,
wonderful words,
marvelous words,
fabulous words,
beautiful words,
glorious words!